This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
2.0 P,=P,
where: T, = Travel time of sonic waves in the anomalous zone T, = Travel time of sonic waves in the sound concrete Tk= Thickness of the anomalous zone P, = Number of measurement paths affected by sonic anomalies P, = Total number of travel paths for each pile D =Pile diameter
174
Srivanavit et al. (1999) reported the interpretation of test signals in comparison with the actual integrity of 9 model tests conducted in Bangkok. Model piles are of 800mm diameter with l m length. The summary of the model test results including the description of model piles as well as discussions by authors are illustrated in Table 4. Table 4. The summary of sonic logging model test results
7.3 Factors to be considered in interpretation of sonic logging test Turner (1997) pointed out the need of both theoretical knowledge and practical experience in interpretation of the test results in CIRIA Report 144. Suggestion is also made in this report that the anomaly shown in sonic test signals can be caused not only by changing in physical properties of the materials, but also by factors within the measuring system itself. These factors are; e e e e e e
175
Free movement of the probes within access tubes Mismatched probe positions especially at pile toe Measurement resolution Incorrect position of access tube Air gaps or different material around access tubes Aggregate variation (in the case of base grouting)
It is to be noted from the author’s suggestion based on the above factors that only variations of transmit time more than 15 to 20% of the norm should be regarded as warranting further investigation.
fect by other integrity tests. Figure 5 shows the photo of dynamic load test carried out on barrette of size (1 .Ox2.7m) with toe depth 48.94m located at the bank of water supply canal. A 20 ton hammer was used to activate the 2050 ton test load in this project.
8 HIGH STRAIN DYNAMIC LOAD TEST 8.1 Overview of the test in Bangkok High strain dynamic integrity test has become a well-accepted method especially for evaluating the pile capacity in the foundation industry today. A large number of related technical papers and case histories of the test have been published and it is a part of standards and specifications such as ASTM D4945-89 (Standard Method for High-strain Dynamic Testing of Piles). In Bangkok, dynamic load test is applied for both driven and bored piles. A summary of the number of dynamic load test carried out during 1991 to 1997 is shown in Table 5. It is to be noted that quantities shown are collected from the available sources and the actual tested numbers are likely to be slightly more than those indicated in the table. As can be seen in the table, application of dynamic load test increases year by year in Thailand. The pile driving analyzer (PDA) with computer software by Pile Dynamic INC, USA is mainly used in Bangkok. Table 5. Yearly minimum quantity of dynamic load test conducted in Thailand during 1991 to 1997 Year No. of dynamic load test done 1991 84 134 1992 242 1993 369 1994 464 1995 473 1996 695 1997
For driven pile it is usually performed at two stages, during driving (initial driving monitoring) and some period after installation of pile (restrike test). For bored piles, a specially designed pile cap is normally required as an integral part of the pile head to avoid pile head damage. Ram weights of 4 ton to 20 ton are commonly used in Bangkok. The common ram weights used for different pile sizes are shown in Table 6. Table 6. Common ram weights used for different size of piles Ram weight (ton) Pile size (m) 5 0.50 4.00 0.50 - 0.80 8.00 1 .oo - 1 s o 20.00
The dynamic load test is also occasionally used as an additional investigation when pile is found with de-
Figure 5. Dynamic load test carried out on a barrette (1 .Om x 2.7mx 48.94m).
8.2 Comparison and correlation of dynamic and static pile load test A comparison between dynamic and static load test results has been reported by various researchers. Seidel and Rausche (1984) reported the results of dynamic and static load tests performed on drilled shafts of the West Gate Freeway in Melbourne, Australia. A 20 ton hammer with drop heights between 1.6 and 2.5m was used for the dynamic tests of 12 shafts ranging from 1 lOOmm and 1500mm in diameter and 35m to 64m in length. The authors reported that dynamic activation of static pile resistance forces exceeded 3000 ton for some 1500mm diameter shafts. Skin friction predictions from dynamic load tests and values obtained from instrumented shafts under static load tests were remarkably similar and pile head load-movement relationships obtained from both test methods were comparable as reported by the authors. Prebaharan et al. (1990) reported the results of dynamic and static tests conducted on bored piles at the Marina Bay Station of the Singapore Rapid Transit System. According to the report, dynamic load tests were carried out for bored piles of lOOOmm diameter and 25m to 50m length founded in Old Alluvium of Singapore Island. Agreement between dynamic and static load tests results permitted replacement of 44 out of 51 tests with dynamic load test from originally planned static load tests as stated by the authors. Vasinvarthana and Kampananon (I 997) presented the efficiency and reliability of dynamic load test carried out in Bangkok. Figures 6 (a) and 6 (b) show the load settlement curves of dynamic and static load tests conducted in Bangkok on large and small di176
ameter piles respectively. As can be seen in the figures, load-settlement characteristics obtained from dynamic load test agree well with those of static load test for both large and small diameter bored piles.
ACKNOWLEDGEMENT The authors wish to express their appreciation to EDE Co., Ltd. for providing the test data. Thanks are also given to Mr. Veera Vasinvarthana and Mr. Natomon Kampananon STS Engineering Consultants Co., Ltd. for giving permission for the use of their previous works in this paper. REFERENCES
Figure 6 (a). Load-settlement curves of dynamic and static load tests on large diameter piles ($I 0.80m x 50.0m).
Figure 6 (b). Load-settlement curves of dynamic and static load tests on small diameter piles ($I 0.35m x 18.91m).
9 CONCLUSIONS Use of sonic integrity, sonic logging and high strain dynamic tests in Bangkok has been presented. Though sonic integrity test is a simple and cost effective method, the reliability of the test results is highly dependent on the experience of the person in both field-testing and interpretation. The results from sonic logging test conducted on model piles in Bangkok helps to extend the knowledge of the signal characteristics and interpretation. High strain dynamic test has become a wellaccepted method especially for load testing of piles. Agreement between static and dynamic load test results enhances the confidence in using the dynamic load test as shown by rapid increment in number of tests conducted in seven years in Thailand, mainly in Bangkok.
American Society for Testing and Materials 1989. ASTM 04945-89 Standard inethod for high-strain dynamic testing of piles, Philadelphia: American Society for Testing and Materials. Balasubramaniam, A. S. 199 1. Inaugural lecture on contribution in geotechnical engineering soil mechanics and fozrndation engineering, AIT, Bangkok, 14 March 199 1. Faiella, D. & Superbo, S. 1998. Integrity non destructive tests of deep foundations by means of sonic methods - Analysis of the results collected on 37 sites in Italy, Van Impe & Haegeman (eds), Proc. 3rd Int. Geotechnical Seminar on Deep Foundations on Bored and Auger Piles, BAP 111, Ghent, Belgium, 19-2 1 October 1998: 209-2 13. Rotterdam: Balkema. Prebaharan, N., Broms, B., Yu, R. & Li, S. 1990. Dynamic testing of bored piles, Proc. of the Tenth Southeast Asian Geotechnical Conference, Volunw I , Taipei, ROC, 16-20 April 1990: 373-378. Seidel, J. & Rausche, F. 1984. Correlation of static and dynamic pile tests on large diameter drilled shafts, Proc. 2"" Int. Col$ on the Application of Stress Wave Theory on Piles, Stockholm, Sweden, 27-30 May 1984: 313-3 18. Stockholm: Swedish Pile Commission. Srivanavit P., Hiranoon J. & Pimsarn S., 1999. Integrity tests on deep bored piles constructed in Bangkok subsoil by sonic logging method & interpretation, The engineering rechnology exhibition and syinposiurn, Organized by Engineering Institute of Thailand, Bangkok, 1-2 November 1999: 1-8. (in Thai). Stain R. T. & Williams, H. T. 1991. Interpretation of sonic coring results: A research Project, Proc. 4"' Int. DFI ConJ on Piling and Deep Foundations, Stresa, Italy, 7-12 April 1991: 633-640. Rotterdam: Balkema. Thasnanipan, N., Baskaran, G., & Anwar, M.A. 1998a. Effect of construction time and bentonite viscosity on shaft capacity of bored piles, Deep Foundations on Bored and Auger Piles, Van Impe & Haegeman (eds), Proc. 3"' Int. Geotechnical Seminar on Deep Foundations on Bored and Auger Piles, BAP 111, Ghent, Belgium, 19-21 October 1998: 171177. Rotterdam : Balkema. Thasnanipan, N., Maung, A. W. & Baskaran, G. 1998b. Sonic integrity test on pile founded in Bangkok subsoil - signal characteristics and their interpretations, Proc. 4"' Int. Cot$ on Case Histories in Geotechnical Engineering, St. Louis, Missouri, 9-12 March 1998: 1086-1092. Turner, M. J. 1997. Integrity testing in piling practice, CIRIA Report 144, London: Construction Industry Research and Information Association. Vasinvarthana, V. & Kampananon, N. 1997. Efficiency and reliability of dynamic load test), Seminar on Foundation ' 97, Organized by Engineering Institute of Thailand, Bangkok: 79-94. (in Thai).
177
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Estimation of pile head stiffness using sonic integrtty testing K. Irnada & Y. Nakata Fuji Engineering Consultant Company Limited, Osaku, Japan
T. Matsumoto Department of Civil Engineering, Kanazawa University,Japan
ABSTRACT: This paper presents a method for estimating the initial stifhess of the load-displacement of a pile (pile head stifkess) fiom the Sonic Integrity Test (SIT). Basically, the proposed method is identical to wave form matching commonly used for the interpretation of dynamic load testing. Validity of the proposed method was examined through a series of SITs and static load tests (SLTs) of a model pile, 800 mm in length and 25 mm in diameter, embedded in model grounds having different profiles of shear modulus of the soils. The proposed method was further applied to the field tests of two cast-in-situ concrete piles where SITs and SLTs were conducted. It is shown fi-om the model tests and the field tests that the proposed method gives us a good estimate of the initial stifhess of the load-displacement curve of a pile. having different profiles of shear modulus of the soils. The proposed method is further applied to the field tests of two cast-in-situ concrete piles where SITs and SLTs were conducted. It is shown from the model tests and the field tests that the proposed method gives us a good estimate of the initial stifhess of the load-displacement curve of a pile.
1 INTRODUCTION Pile foundations should satis@ required qualities and performances. It will be of vital importance to ensure these requirements are met for constructed piles in the limit state design and the performance based design. For cast-in-situ concrete piles, the sonic integrity testing (SIT) has been widely used to assess qualities and mechanical properties such as configurations, Young's modulus and the strength of piles after hardening of the concrete. The static load test (SLT) and rapid load tests such as the Statnamic test has been employed to obtain the performances of a pile such as the load-displacement relation as well as the ultimate pile capacity. However, these load tests are conducted on only a few piles at a construction site, because these tests are time and cost consuming compared to the SIT that may be conducted on to all piles constructed at a site in a time and cost efficient manner. This paper presents a method to estimate the initial pile head stiilbess (initial tangential stiffness of the load-displacement curve at the earliest stages of loading) of a pile by conducting a sonic integrity test with the interpretation based on one-dimensional stress-wave theory. Basically, the proposed method is identical to the wave form matching commonly used for the interpretation of dynamic load testing, except that FEM is adopted to estimate the loaddisplacement relation for the tested pile. Validity of the proposed method is examined through a series of SITs and SLTs of a model pile, 800 mm in length and 25 rnm in diameter, embedded in model grounds
2 INTEGRITY TEST AND STATIC LOAD TEST OF MODEL PILE 2.1 Model ground, model pile and test procedure The test apparatus for model tests is shown in Figure 1. The model grounds were made in a cylindrical
Figure 1. Model test apparatus.
179
length, L, of 8 0 0 Young's ~ modulus, E, of 1.65.1 O6 kN/m2, a density, p, of 2.1 8ton/m3and a bar wave velocity, c, of 8OOm/s, was placed in the soil box after the completion of the bottom soil section. Thereafter the middle and upper soil sections were made. The number of tamping for each soil layer of 100 mm thickness in the three soil sections is listed in Table 1. A total of six different model grounds were made. The longitudinal wave velocity, VP,of each soil layer was estimated by generating a sonic pulse on the ground surface, and cone penetration test was conducted to obtain a variation with depth of the tip resistance, 9.. In Cases-5 and 6, the sonic integrity and the static load test were conducted on the model pile. This paper will present and discuss the test results of Cases-5 and 6 only. Figure 2 shows the variations with the the depth of the cone tip resistance, q., and the soil density, p. The shear wave velocity, VS,was estimated fiom the measured VPby means of Equation (1).
acrylic box having an inner diameter of 250 mm and a height of 950mm. The soil of the model grounds was a dry Toyoura sand mixed with lime by a weight ratio of 15% (weight of lime/weight of Toyoura sand = 0.15). Table 1. Conditions of model grounds. Case No Number of tamping for each soil layer of 1OOmm thickness Upper Middle Lower section section section Case- 1 0 0 0 Case-2 8 8 8 Case-3 0 0 40 8 8 40 Case4 0 8 40 Case-5 8 0 40 Case-6
Number of soil sections
1 1 2 2 3 3
The soil was compacted in the cylindrical box by letting a ram having a weight of 58.8N fall fiom a height of 20mm to make a soil layer of 100 mm thickness. This procedure was iterated 8 times to obtain the model ground of 800 mm height. Accelerometers (acoustic emission sensors) were placed on the top of each soil layer of 100 mm thickness to measure the longitudinal wave velocity, VP,of each layer. The model ground consisted of three soil sections as indicated in Figure 1. A model pile made of Teflon, having a diameter, d, of 2 5 m , a
1- 2v, v -- v4 2______ ( 1 -Vs) where v. is Poisson's ratio of the soil which was assumed to be 0.3.
Figure 3. Measured variations in shear wave velocity with depth of model grounds for Case-5 and Case-6.
(b) model ground for Case-6 Figure 2. Variations in cone tip resistance and density with depth.
2.2 Wave analysis of the integrity tests A computer program KWAVE developed by Matsumoto & Takei (1991) was used for the wave propagation analysis for the sonic integrity tests of the model pile. The pilehoil system adopted in KWAVE is shown in Figure 4. Rational soil resistance models proposed by Randolph & Simons (1986) and by Deeks (1992) have been incorporated in KWAVE for the shaft and the base resistance, respectively. The pile displacement and the displacement of the soil adjacent to the pile is calculated separately so that the viscous damping and the radiation damping are allowed for individually (Figure 5). Relative displacement between the pile and the soil does not 180
occur until the mobilized shear stress reaches the maximum shear stress, zm,, that is expressed by the slider in Figure 5 . According to Novak et al. (1978), the spring constant, k,, and the radiation damping, er, of the outer shaft resistance are estimated as follows:
k,
= 2.756 l(d),
c, = G I V,,
d= pile diameter
G = shear modulus of the soil
For the base model (Figure 6), the spring constant, C b , and the lumped mass, Mb, are given as follows (Deeks 1992): kb, the damping constant,
kb =-
8G ~ ( 1 v)d -
(4)
(2) 3.2p,VS n(1- v )
-
C b = p
(3)
M~
= 2d3
~
3.2G n(1- V)V,
0.1 - v 4 1-v
In the wave matching analyses of the sonic integrity test signals, the values of the maximum shaft resistance, zmas, and the maximum base resistance, q b , were taken as very large, because it was judged that rmxand q b were not reached in the sonic integrity tests fi-om the fact that permanent settlement of the pile did not occur after the sonic integrity tests. The wave propagation analyses were conducted assuming various profiles of VSof the model grounds until a good matching of the calculated and measured pile head velocities were obtained. Note that all the soil resistance parameters are determined from the assumed VSand the measured ps. Figure 4. Modeling of pile and soil used in KWAVE.
Figure 5. Shaft model (after Randolph & Simons 1986).
Figure 7. Calculated and measured pile head velocity.
Figure 6. Base model (after Deeks 1992).
181
The calculated and measured pile head velocities for Cases-5 and 6 are shown in Figure 7. The assumed profiles of VS of the model grounds are shown in Figure 8, together with the measured VS. The profiles of VS estimated f?om the sonic integrity tests seems to represent the measured results well, although the estimated VS underestimate the measured VS of shallower depths down to 350 mrn. The underestimation of V . for the shallower depths
might be due to loose contacts between the pile and the surrounding soils.
2.3 FEM analyses of the static load tests FEM analyses were conducted to estimate the loaddisplacement curves for the model pile in Cases-5 and 6. The axi-symmetric FEM mesh for the model ground and the model pile is shown in Figure 9. The lateral displacements along the center axis, the vertical and lateral displacements at the bottom and side of the ground were fixed to reflect the model test conditions. Both the pile and the soil were model as elastic materials. The shear moduli, G, of the soils estimated from the sonic integrity tests (G = pYVs2) and vs = 0.3 were used in the FEM analyses. The load-displacement curves estimated from FEM analyses for Cases-5 and 6 are compared with the curves obtained from the static load tests in Figure 10. The estimated load-displacement curves are good measures for the measured results. It should be noted that the measured load-displacement curves shown in Figure 10 are initial portions where an almost linear relation between the displacement and the load was obtained.
Figure 8. Variations in shear wave velocity with depth estimated from wave matching analysis, together with measured results.
Figure 9. Axi-symmetric FEM mesh for model ground and pile.
Figure 10. Comparison of estimated and measured loaddisplacement curves.
182
3 INTEGRITY TESTS AND STATIC LOAD TESTS OF CAST-IN-SITU CONCRETE PILES Above, a method for estimating the initial stifThess of load-displacement of a pile fiom the sonic integrity test was demonstrated against the model pile. Below, the proposed method is applied to two actual cast-insitu concrete piles. 3.1 Test description The sonic integrity test and the static load test were performed on two cast-in-situ concrete piles, designated as pile D and pile E, constructed in two different sites. The soil profile and the SPT N-values at the test sites of piles D and E are shown in Figures 11 and 12, respectively. The specifications of the concrete piles are listed in Table 2.
Figure 1 1. Soil profile and SPT N-values at the test site of pile D, together with the pile seating.
Figure 12. Soil profile and SPT N-values at the test site of pile E, together with the pile seating.
Table 2. Specifications of test piles and principal results of static load tests. Item Pile D Pile I 1.2 1.5 Diameter (m) Pile 23.5 44.9 Length at SLT(m) 22.5 43.9 Embedded length at SLT (m) 17.6 36.4 Length at SIT (m) O7 2.60~10' 2.27~1 Young's modulus (kN/m*) 4200 Bar wave velocitv (m/s) 4000 16.7 19.5 Principal Maximum load (MN) results Pile head displacement at 200 52 maximum load (mm) Yield load (MN) 12.7 16.7
A static the load test was performed on pile D, having an original Iength of 23.5m, 4 weeks after the pile construction. The top 5.9m of soil was excavated aRer the static load test, and the corresponding section of pile D was cut, so that pile D had a length of 17.6m. Thereafter, the sonic integrity test was conducted on pile D of 17.6m length. Pile E had a similar situation such that the original length of 43.9m at the static load test was shortened to 36.4m before conducting the sonic integrity test. The measured load-displacement curves of piles D and E are shown in Figure 13. In order to compare the load-displacement curve to be estimated from the sonic integrity test with the static load-displacement curve, the measured load-displacement curve of pile D with the original length of 23.5m was corrected to estimate a load-displacement curve of pile D with the shortened length of 17.6m, as follows. As the axial forces of pile D were measured, the shaft friction along the top 5.9m of pile D was subtracted fiom the measured pile head load, and the elastic deformation of a pile length of 5.9m was subtracted fiom the measured pile head displacement. The loaddisplacement curve of pie E with the shortened length of 36.4m was also estimated similarly.
Figure 13. The measured load-displacement curves of piles D and E of full length, and the estimated load-displacement curves for shortened lengths at integrity tests.
The corrected load-displacement curves of piles D and E are also shown in Figure 13. Each corrected curve will be compared with the load-displacement curve estimated from the sonic integrity test. 3.2 Wave analyses of integrity test signals analyses of the sonic integrity test Wave signals of piles D and E were performed using the KWAVE program. In the analyses, Poisson’s ratio
Figure 15. Axi-symmetric FEM meshes for piles D and E.
Figure 16a. Load displacement curves and estimated initial pile head stiffness of piles D and E.
was assumed to be 0.4 for all soils, and density was assumed to be 1.9todm’ for sandy soils and 1.7todm3 for clayey soils. Wave propagation analyses were iterated assuming various profiles for the shear modulus, G. The results of the final
Figure 14. Calculated and measured pile head velocities.
184
matching are shown in Figure 14. In Figure 14, the pile head velcity, v, is normalized by the maximum velocity, VO, induced by the impact of hammer. The profile of G assumed in the final matching have been shown in Figures 11 and 12 for pile D and pile E, respectively.
Osaka area of Japan. Pile diameters range from 1.2m to 2.0m and the pile lengths range fiom 11.8m to 43.9m. In Figure 16, the pile head load, P, is normalized by the ultimate pile capacity, P,, and the pile head displacement, w, is normalized by the yield pile head displacement, 4, at the yield load, Py. Therefore, PIP, corresponding to w l 4 = 1 is defined as PJP,,. Here, let the secant stifhess, m,, be defined as P,/6;,.
3.3 FEM analyses of the static load tests FEM analyses of static loading of piles D and E were conducted to estimate the initial stiffness of piles D and E. hi-symmetric FEM meshes for piles D and E are shown in Figure 15. The piles and the soils were modeled as elastic materials. The profiles of the shear modulus, G, identified from the sonic integrity tests and Poisson's ratio of 0.4 were used for the elastic parameters of the soils. The calculated load-displacement curves of piles D and E are compared with the load-displacement curves which have been corrected from the static load test results. The calculated and corrected measured initial stiffness of piles D and E are listed in Table 3. The initial stiffness, m,,, estimated from the SITS
Figure 16b. Load displacement curves of cast-in-situ concrete piles constructed at different sites in Osaka area.
Table 3 . Comparison of estimated and corrected measured initial pile head stiffness of piles D and E. Pile Initial stiffness Initial stiffness m,,lm,,n from SIT from SLT m,, (MN/mm) m,, (MN/mm) D 8.4 1 7.22 1.16 F 14 41 12.56 1.15
overestimate the corrected measured initial stiffness, rn,,,,, by 15% only. The overestimation of the initial s t f i e s s may be attributed to the non-linearity of the soils. This is because the shear modulus, G, estimated from the SIT corresponds to a very small strain level, which would be higher than G of the soils surrounding the piles under the static load tests. The authors are aware that some assumptions concerning Poisson's ratio and soil density made in the wave matching analyses, also be reasons for the overestimation of the initial stifhess. 3.4 Discussion on the uses of estimated initial st qfness
As mentioned earlier, it is common practice in Japan t o perform only one static load test at a pile construction site. Estimation of initial pile stifhess of the other piles at the site by means of the SIT could serve for assessment of the performance of the other piles, by comparing the estimated initial stiffhess with the initial stifhess obtained from the static load test. Another use of the estimated initial stiffness is discussed here. Figure 16 shows the loaddisplacement curves of a total of 11 cast-in-situ concrete piles constructed at different sites in the
Figure 17. Frequency distribution of m,lm, of cast-in-situ concrete piles shown in Figure 16.
The frequency distribution of the ratio of the initial stiffness, m,, to the secant stifiess, m,, of the piles in Figure 16 is shown in Figure 17. The ratio, m,lm,, ranges from 0.07 to 0.36 with an average of 0.21. Although the number of the piles is limited, the results of Figures 16 and 17 indicate that the initial stifhess can be used for a rough but useh1 estimation of the secant stiffness. 4 CONCLUSIONS A method for estimating the initial stiffness of the load-displacement curve of a pile (initial pile head stiflkess) fiom the sonic integrity test has been proposed. Validity of the proposed method was 185
examined through the sonic integrity testing and static load testing of a model pile in different ground conditions. The proposed method was further applied to two cast-in-situ concrete piles constructed at two different sites at which the sonic integrity tests and the static load tests were conducted. The proposed method estimated the initial stifftiness of the piles with a good accuracy, 15% overestimation of the measured values. The uses of the initial stifftiness were discussed and it is demonstrated that the initial stifhess can be used for a rough estimate of the secant stifhess.
REFERENCES Deeks, A.J. 1992. Numerical analysis of pile driving dynamics. PhD Thesis, The University of Western Australia. Matsumoto, T. & M. Takei 1991. Effects of soil plug on behaviour of driven pipe piles. Soils & Foundations, JSSMFE, 3 l(2): 14-34. Novak, M., T. Nogami & F. Aboul-Ella 1978. Dynamic soil reactions for plane strain case. J. Mech. Eng. Div., ASCE, 104(EM4): 953-959. Randolph, M.F. & H.A. Simons 1986. An improved soil model for one-dimensional pile driving analysis. Proc. 3rd Int. Con$ on Num. Methods in Offshore Piling, Nuntes: 1 17. Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil models for axial pile response. Proc. 4th Int. Con$ on the Appl. of Stress- Wave Theory to Piles, The Hague: 3- 14.
186
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Integrity testing of cast-in-situ concrete piles associated with the construction of New Haccho Bridge Yuji Michi Yoshimitsu Company Incorporated, Komatsu, Japan
Tatsunori Matsumoto Department of Civil Engineering, Kanazawa University,Japan
Yoichiro Matsuda Ishikawa Prefectural Government, Kunazawa, Japan
ABSTRACT: Cast-in-situ concrete piles were constructed for the foundations of New Haccho Bridge at Komatsu City, Ishikawa Prefecture, Japan in 1997. The requirements for quality control during construction of the piles were the quantity of concrete used for each pile, the outer diameter of the casing pipe and the excavation length. For quality assessment of the concrete piles after hardening, Sonic Integrity Testing was conducted on a total of 8 piles of one abutment after curing periods of 17 to 28 days. Wave analysis of the Sonic Integrity Test signals showed that all the piles were constructed as designed. These signals will play the role of the reference signals, if a disaster such as a great earthquake should occur in the future, The reference signals would be compared with the Sonic Integrity Test signals that would be obtained after the disaster, in order to detect the occurrence of pile damage easily and quickly. m over Haccho River in Komatsu City, Ishikawa Prefecture. The bridge is sumorted bv two abutments with cast-in-situ concrete piles.
1 INTRODUCTION Non-destructive inspection of piles such as the Sonic Integrity Test (SIT) was used to detect the existence of damage to piles after the Hyogoken-Nambu Earthquake in 1995. The SIT was widely used, because the SIT is easy to conduct on piles when their heads are exposed and because the measured signal (pile head velocity) is interpreted based on one-dimensional stress-wave theory. However, there were cases in which it was difficult to determine whether a pile had damage or not by means of the SIT (Hayashi 1996). This may be attributed to the fact that the influence of soil resistance was not taken into account in the interpretation of the measured signals and that SIT signals were not measured before the earthquake for the purpose of comparison. The New Haccho Bridge was constructed in 1997 after the Hyogoken-Nambu Earthquake. Cast-insitu concrete piles were adopted for the foundations o f the bridge. Another objective of conducting SITS was to obtain reference SIT signals for the constructed piles, which would be compared with Sonic Integrity Test signals that would be obtained after a disaster in the future.
2.1 Site condition Figure 1 shows the profiles of the soil layers and the SPT N-values at the construction site. The clay and sand of the alluvium deposits are distributing alternately to an elevation of -18.2 m underlain by stiff layers of gravel and sand having SPT N-values greater than 50. End-bearing cast-in-situ concrete piles seated in the gravel layer were adopted for the foundation piles. The piles had a length of 23 m and a diameter of 1 m. The design ultimate pile capacity, R,, of the piles is 2.3 MN which was derived fiom the following empirical rule (Specifications 1990): Ru = q d A p q d = 2.94
I
m / m 2 for N > 30
in which q d is the ultimate toe resistance and A, is the cross-sectional area of the pile. A total of 8 piles were constructed at each abutment. Figure 2 shows the layout of the piles at one of the abutments, designated as Al. The date indicated beneath each pile represents the date of concreting of the pile construction process. The requirements for quality control during the construction of the piles were the quantity of concrete used for each pile, the outer diameter of the
2 OUTLINE OF THE CONSTRUCTION WORKS The New Haccho Bridge is a highway bridge on an access route to Komatsu Airport, with a span of 26 187
Figure 1. Soil profile and SPT N-values at the construction site, together with estimated soil parameters.
2.2 Integrity testing The SIT was conducted on all the piles at abutment AI 17 to 28 days after the concreting of the piles, using the IFCO SIT system. Treatment of the pile head of each pile had been finished when the SIT was conducted. The concrete of the top 1.2 m of the pile was removed and 20 steel reinforcement bars, 32 mm in diameter, were exposed as shown in Photograph 1. Measurements of the bar wave velocities, c, of a concrete test piece, 200 mm length and 100 rnm diameter, were done after different curing periods of the concrete by means of the ultrasonic wave test. The test results are summarized in Table 1. It can be seen that the bar wave velocity increased with increasing curing period until 17 days, and almost leveled off after that. Young’s modulus, E,, of the concrete was estimated from the following relation:
Figure 2. The layout of the piles in abutment A,.
E, = P C C
2
in which pc is the density of the concrete. E, at the SIT was estimated as E, = 36.5 GN/m2 fiom the measured p, = 2.28 ton/m3 and c = 4000m/s. Table 1. Results of the ultrasonic wave tests conducted on the concrete test piece.
Photograph 1. Situation of the integrity testing.
Test date June 26, 1997 June 30’ l997 3, 1997 July 4, 1997 June 11, 1997
casing pipe and the excavation length as designed. The requirements were achieved under inspection of a piling engineer of experience, and it was judged that all the piles had been constructed as designed.
188
Curing period (days) 2 6 9 10 17
Wave velocity (ds) 3590 3690 3802 3891 4000
3 RESULTS OF THE SONIC INTEGRITY TESTS 3.1 Measured SIT signals The measured SIT signals (time vs. pile head velocity) of all the piles at abutment A1 are shown in Figure 3. All the SIT signals are similar, indicating uniformitv of the d e s .
Figure 4. Modeling of pile and soil used in KWAVE.
Figure 3. Measured SIT signals.
It can be seen from Figure 3 that the amplitude of the pile head velocity decays with time, showing a large influence of radiation damping of the soil. Reflection waves of the incident waves from the pile toe are detected at a time around 11.5 ms. The bar wave velocities, c, of the piles are estimated as 4000 m / s from this return traveling time of the incident waves. The bar wave velocity, c, estimated from the SITS is identical to that of the concrete test piece measured by the ultrasonic wave test. The pile head velocities oscillate periodically. Although the soil resistance is a major influence for this, another reasons for this may be considered to be the influence of wave propagation in the exposed steel reinforcement bars above the pile heads and the changes in the properties of the cross sections of the piles due to the different volumes of the steel reinforcement bars along the pile shafts.
Figure 5. Shaft model (after Randolph & Simons 1986).
3.2 Method of wave propagation andysis Wave propagation analyses of the SIT of pile PI were conducted. A computer program KWAVE developed by Matsumoto & Takei (1991) was used for the wave propagation analyses of the sonic integrity test signals. The pile/soil system adopted in KWAVE is shown in Figure 4. Rational soil resistance models proposed by Randolph & Simons (1986) and by Deeks (1992) have been incorporated in KWAVE for the shaft and the base resistance, respectively. The pile displacement and the displacement of the soil adjacent to the pile are calculated separately so that the viscous damping and the radiation damping are allowed for individually (Figure 5). Relative displacement between the pile and the soil does not
Figure 6 . Base model (after Deeks 1992).
occur until the mobilized shear stress reaches the maximum shear stress, zmax,that is expressed by the slider in Figure 5. According to Novak et a1 (1978), the spring constant, k,, and the radiation damping, c,, of the shaft resistance are estimated as follows: 189
Table 2. Specifications of each section in the pile. Section No. Section 1 Distances from pile head 0.0 to 6.1 m Length (m) 6.10 Diameter (mm) 1000 Sectional area (m2) 0.785 Area of steel bars (m2) 0.016 Young’s modulus (GN/m2) 39.93 Wave velocity ( d s ) 4000 Density (ton/m ) 2.496 Impedance (kNs/m) 7836
k,
c,
= 2.756 /(d), d = pile
=G/
diameter
V,, G = shear modulus of the soil
Section 2 6.1 to 7.4m 1.30 1000 0.785 0.024 41.66 4000 2.604 8176
(3) (4)
For the base model (Figure 6), the spring constant, kb, the damping constant, L‘b, and the lumped mass, Mb, are given as follows (Deeks 1992): (5) 3.2p,VS - 3.2G n(1- v) n(1- V)V,
cb =--
(6)
(7) In the wave propagation analysis of the sonic integrity tests, the values of the maximum shaft resistance, zmax,and the maximum base resistance, q b , were taken as very large, because it was judged that z ,, and q b were not attained in the sonic integrity tests from the fact that permanent settlements of the piles did not occur after the sonic integrity tests.
Section 3 7.4 to 11.3m 3.90 1000 0.785 0.008 38.21 4000 2.388 7499
Section 4 11.3 to 12.6m 1.30 1000 0.785 0.012 39.07 4000 2.442 7667
Section 5 12.6 to 23.0m 10.40 1000 0.785 0.004 37.34 4000 2.334 7328
Poisson’s ratio of the soil, v, was taken to be 0.49, because the undrained condition was assumed to be present during the SITS. The parameters of the soil models were determined by means of Equations (3) to (7), using the soil properties thus estimated. The soil properties and the soil model parameters have been listed in Figure 1. The pile was divided into a total of 5 sections to take into account the different volumes of the steel reinforcement bars, as listed in Table 2. Young’s modulus, Ep, of the composite pile of concrete and steel bars was estimated fiom the following equation: Ep =(E,A, +E,A,)/(A, +A,)
(10)
in which Ec: Young’s modulus of concrete (36.5 GN/m2) A,: the total cross sectional area of concrete E,: Young’s modulus of steel (206 GN/m2) A,: the total cross sectional area of steel bars Assuming c = 4000 m/s for each pile section, density, p p , of each pile section was estimated as follows:
3.3 Estimation of mechanical properties of the soils and the pile In order to utilize the above-mentioned soil models in the wave analysis of the SIT signals, the soil properties were estimated as follows. The soil density, p,, was assumed to be 1.8 ton/m3 for the clay, 1.9 ton/m3 for the sand and silt, and 2.0 ton/m3 for the gravel. The shear wave velocity, V,, of the soil was estimated from the following empirical equation using the SPT N-value (Imai & Yoshimura 1968):
Thereafter, the shear modulus, G, of the soil was estimated based on the elastic theory of shear wave propagation:
G = psV:
(9)
190
Impedance, 2,of each pile section is defined as follows: Z = ( E p A ) / c = ( E p ( A ,+ A p ) } / c
(12)
A criticism of this approximate procedure for estimating the pile properties may be made concerning the assumption that the bar wave velocity is uniform along the total length of the pile. The bar wave velocity varies according to the changes in the properties of the pile sections. However, in the common practice of the SIT, only an average bar wave velocity is estimated fiom the return traveling time of the incident wave along the total length of the pile. Therefore, the bar wave velocity was assumed to be constant along the total pile length in this particular case.
The measured incident velocity at the pile head was input as the boundary condition in the wave propagation analysis of the SIT signals. As the data sampling time interval was 0.025 ms, the pile was divided into elements having a length of 100 mm.
3.4 Results of wave propagation analyses Two wave propagation analyses of the SIT were conducted with and without taking the influence of the soil resistance into account. The calculated pile head velocity in the later analysis is compared with the measured one in Figure 7. Some bumps detected in the calculated velocity are caused by the changes in the cross sections of the pile, of which the amplitudes are relatively small compared to the amplitudes of the incident velocity and the velocity at t = 11.5 ms caused by the reflection wave fiom the pile toe. It can also be seen fi-om Figure 7 that the calculated velocity caused by the reflection wave from the pile toe is very large compared with the measured velocity, strongly suggesting the importance of consideration of the soil resistance even in the wave analysis of SIT signals where amplitudes of accelerations, velocities and displacements in the pile are very small compared to dynamic load testing. The pile head velocity calculated fiom the wave analysis with consideration of the soil resistance is compared with the measured velocity in Figure 8. The calculated velocity is comparable with the trend of the measured velocity. The periodic oscillation found in the measured velocity is not reproduced in the calculated velocity. It should be noted that the wave propagation in the steel bars exposed above the pile head was not taken into account in this wave analysis of the SIT signals. It may be judged again fiom the wave analysis of the SIT signals that all the piles were constructed as designed.
Figure 7. Comparison of the calculated and the measured pile head velocities in the wave analysis without the influence of the soil resistance.
Figure 8. Comparison of the calculated and the measured pile head velocities in the wave analysis with the influence of the soil resistance.
resistance is indispensable even in the wave matching analysis of the sonic integrity test signals. In order to take the influence of the soil resistance into account, it is preferable to conduct soil investigations such as seismic cone penetration testing in parallel with the sonic integrity test to estimate the shear wave velocity of the soil.
4 CONCLUSIONS
If the possibility of the existence of piles not satisfying design requirements is found with SIT, other types of tests such as borehole camera inspection and/or the Statnamic test and/or the lateral load test should be performed, depending on the position of the pile defects and the pile performance required in the design, or change of design should be considered.
The sonic integrity tests performed on cast-in-situ concrete piles for the foundation of New Haccho Bridge have been described for the purposes of quality assessment of the constructed piles and obtaining integrity test signals from the piles as reference signals which would be compared with sonic integrity test signals that would be obtained after a disaster in the fbture, in order to detect occurrence of pile damages easily and quickly. The principal conclusions from this study are as follows :
REFERENCES Deeks, A.J. 1992. Numerical analysis of pile driving dynamics. PhD Thesis, The University of Western Australia. Hayashi, T. 1996. Fundamental study on qulity control methods for concrete foundation piles. Dr. Dissertation, Aichi Institute of Technology.
1. All foundation piles for the New Haccho Bridge were constructed as designed. 2. Consideration of the influence of the soil 191
Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil models for axial pile response. Proc. 4th Int. Con$ Appl. of Stress-Wme Theoly to Piles: 3-14. Matsumoto, T. & M. Takei 1991. Effects of soil plug on behaviour of driven pipe piles, Soils & Foundations, JSSMFE, 31(2): 14-34. Matsui, T. & K. Oda 1996. Foundation damages of structures. Special Issue of Soil & Faundations: 189-200. Novak, M., T. Nogami & F. Aboul-Ella 1978. Dynamic soil reactions for plane strain case. J. Mech. Erg. Div., ASCE, 104(EM4): 953-959. Ohtsu, H., Y . Hatsuyama, A. Takeishi & K. Horikoshi 1997. A study on pile foundations damaged by the 1995 Hyogoken Nambu Earthquake. Proc. Int. Con$ On Deformation and Progressive Failure in Geomechanics (IS-NAGOYA '97) ,Pergamon: 583-588. Randolph, M.F. & H.A. Simons 1986. An improved soil model for one-dimensional pile driving analysis. Proc. 3rd Int. Con$ on Num. Methods in Ofshore Piling, Nantes: 1 - 17. Specijkations 1990. Specification for highway bridges Part IV: Substructures, Japan Road Association.
192
Application of Stress-Wave Theory to Piles, Niyama 8,Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
resentation of low strain integrity testing in the time-frequency domain J. €? Seidel Monash Universi&, Melbourne, Vic.,Australia
ABSTRACT: The results of low strain integrity tests are usually presented either in the time doniaiii or in the frequency domain. The information presented in either form is essentially the same, however, the two methods of display can highlight different aspects of the response. An alternative method of display in the timefrequency domain is presented. This form of presentation derives from the mathematical concept of wavelets. The time-frequency display does not present new information, but is an interesting alternative to the time domain and frequency domain, which can enhance the user’s understanding and interpretation of the pile-top response. 1 INTRODUCTION
Low strain integrity testing techniques are coniiiioiily used to assess the construction quality of cast-in-situ piles. In such techniques, the motion generated at the pile-head due to a transient impact is measured and interpreted. The motion may be determined directly from geophone measurements. Alternatively an accelerometer can be used, and the signal integrated to derive the pile-head velocity. The transient impact is typically imparted by a sinall hand-held hammer, although alternative methods have been used. As the impact is generated by a relatively small device. the resulting acoustic wave which transmits along the pile length has low energy, and a relatively high frequency content. It generates only low-strain values in the pile, hence the term low-strain integrity testing. The wave transmission and reflection is typically interpreted using the theory of onedimensional wave mechanics. It is assumed that the pile reasonably represents a rod with length significantly greater than its diameter. This interpretation may not be appropriate near the pile head where significant dispersion of the wave from the impact occurs. In one-dimensional wave mechanics, the pile impedance, Z, is defined as the product of mass density, p, pile area. A, and the speed of wave transmission, c. The impedance of a pile will be affected by the cross-sectional area of the pile and by the quality and composition of the pile material at any section. If a pile with non-uniform impedance is impacted at the pile head, the induced stress wave will be re-
flected from each change in impedance. The measured pile response is affected at the time each reflection returns to the pile-head. The nature and location of impedance changes may therefore be interpreted from the pile-head velocity response. Because of the relationship between the physical properties of the concrete or grout and its impedance, the existence of anomalies in the pile section may be inferred. Two general caveats apply 1. That individual pile responses are judged against a wide sample, or tlie complete population of piles; 2. That the inferred anomalies be confirmed by physical inspection. It should be mentioned that energy is also radiated from the pile shaft into the surrounding soil. The effect of the soil is to absorb energy, and therefore to dampen the measured response. The greater the soil stiffness, the greater will be the damping effect. This limits the penetration of the input wave into the pile, and the effective length of shaft which can be verified (Turner, 1997). In order to compensate for the degradation of the input wave. amplification routines are sometimes applied to the response to assist with identification of features which would otherwise be hidden.
2 CLASSICAL METHODS OF DISPLAY
The pile-head responses are typically displayed h one of two forms - in the time domain or the fre193
(1 iic nc 1 do ilia i n . Gen era11y , proprietary systems d i splal. data in only one or tlie other of these forms. Testers tend to form their judgements based only on that particular view of the data. A third form of displa) of the data will be presented in the latter sections of this paper. 2.1 Time doiiiciin ’fhe time domain display plots pile-top velocity as the vertical axis against elapsed time as the horizontal axis (see Figure 1). Records are interpreted particularly by focussing on the response during the first passage of tlie input wave down and then up the pile (i.e. to a time of 2L/c, where L is the pile length and c is the speed of the compression wave in the pilc). This is not to say that the response beyond 2L/c is not used to assist in interpretation.
Figure 1. Typical response in time domain for pile with disCrete impedance reduction
In Figure 1, the initial impact is indicated. Subsequent features are all reflections from the p,ile shal’t and the surrounding soil. From one diinensional mave mechanics theory, it can be established that a discrete reduction in pile impedance will generate a tensile reflection from the con~n~encement of the reduction. A compression reflection will follow where the pile returns to its normal dimension. This can be seen as a characteristic positivehegative cycle i n the time domain. The amplitude of this cycle will be a fhction of the relative reduction in pile impedance; of the length of the impedance reduction, and the wavelength of the input wave. By tlie same logic, a discrete increase in pile impedance will cause a negative/positive cycle in the response viewed in the time domain. The toe response is also a key feature which is always sought in the time domain. If the toe response can be identified, it is possible to establish whether or not the pile is free of acoustic reflectors. Assuming that energy reaches the pile toe different reflections can be expected. If the pile toe approximates a “free-end”, the reflected wave would be tensile. and velocity increase will be detected at 2L/c. A fixed condition will generate a compressive reflection, and a velocity decrease. Depending on the stiffness of the toe condition, therefore, a spectrum of responses from large tensile to large compressive could theoretically be generated. In practice. at such low strain levels, a tensile response is
more common. The magnitude of tlie response will depend not only on the end-condition, but also on tlie attenuation of the incident and reflected waves. The toe response will be a half-cycle rather than the full cycle resulting from discrete impedance reductions along the shaft. If an amplification routine is applied to the time domain signal, it is not possible to make an absolute determination on the toe condition from the amplified response. However, the toe condition may be judged relative to like piles on a project fi-om a comparison of responses as long as the piles are tested in the same manner and transformed using the same amplification level. Although the indications of the impedance reduction and the toe response shown in Figure 1 are clear, it is not always possible to locate these features with sufficient accuracy. It may, for instance be unclear whether a particular response emanates from the pile toe or a defect just above the toe. This may have a critical effect on the pile evaluation. Part of this ambiguity arises from uncertainty about the true speed of wave transmission. Values of 3600 m/s to 4000 m/s (and above) are typical for grout and concrete. Stated accuracies in length determination of 10% are therefore commonly quoted. There is another reason which can cause difficulties in interpretation. When the hammer strikes the pile-top, the energy that enters the pile is composed of a spectrum of individual frequencies. The composite of these frequencies results in the summed input waveform which is indicated by the measured velocity response. Each of these individual frequency components travels down the pile. Elasticity theory actually shows that the transmission speeds for the different frequencies are very marginally different. More importantly. because the high frequency components contain relatively little energy, these components are attenuated faster than the low frequency components which contain the bulk of the input energy. The consequence is that the response received from the pile has a different frequency spectrum than the incident wave, being more dominated by the low frequency components. The composite of the individual frequency reflections will be a different shape to the incident wave, complicating interpretation. A shortcoming of the time domain method therefore is that it does not allow discrimination of the frequency response.
2.2 Freqirency domain As noted above, the hammer impact generates a spectrum of waves with different frequencies and amplitudes or energy contents. In order to view the response of the pile in the frequency domain, it is customary to measure not only the pile-head veloc194
ity. but also the pile-top force generated by the hammer. Both the force and velocity responses are deconstructed to establish the frequency spectrum of the hammer input and the pile/soil response respectively. Depending primarily on the pile geometry, the pile Lvill tend to resonate at particular frequencies when the reflection from the pile toe or an intermediate feature is "in phase" with the input wave. The result will be a heightened velocity response at such frequencies. The ratio of peak velocity response to peak force (called the mobility) at these part i c 111ar i 11 put frequencies w i 11 be a inay i mum. 1 he frequency domain display plots pile-top mobility as the vertical axis against applied frequency as the horizontal axis (see Figure 3). ~~
As discussed earlier. this method of display is dependent on the determination of the frequency spectrum of both the velocity and force records. This is normally done by a technique called Fast Fourier Transformation (FFT). It is implicit in this technique that the recorded transient signal is actually part of an infinitely recurring response. I n this way, the signal can be transformed into an equivalent set of infinite sinusoidal functions, each with a defined amplitude and phase shift. As the analysis is then based on infinite series, the frequency response analysis does not contain any temporal (time) information. To highlight this, Figure 3 shows two signals - the first consists of the superposition of two frequencies (Sin1Ot and SidOt) and the second contains these two frequencies in separate succession. Although the resulting wave traces are quite different, the FFT of both signals is identical. This is a limitation of this method of analysis.
Figure 2. Typical response in frequency domain for pile with discrere impedance increase (after Ellway, 1987).
These records are primarily interpreted by establishing the frequencies at which local maxima of the mobility occur. The frequency difference, Af, between successive maxima (harmonics) correspond to intervals of c/2L. As noted previously, given the uncertainties in establishing the exact value of the wavespeed, the pile length, or the length to a feature within the pile can be established to within 10% i.e. the same limitation as for the time domain analysis. For the same reasons discussed earlier about the greater persistence of the low frequency components, the overall pile length is most likely to be evident i n the low frequency area of the mobility plot, whereas features higher in the pile will be more evident in the high frequency area (see Figure 2). Another feature of the mobility plot is the pilehead dynamic stiffness, determined as the inverse of the slope of the initial (low frequency) portion. At very low frequencies the inertial effects are said to be negligible. Hence, the stiffness is suggested (Davis and Dunn, 1974) to correlate to the initial elastic loading stiffiiess of a static load test. This suggestion is not universally accepted. However, j List as conclusions may be drawn with respect to the comparative toe responses of a large population of piles tested using time domain analysis, a comparatively low pile-head dynamic stiffness may justify further investigation.
Figure 3 . Comparison of FFTs for two distinctly different wave forms (after I
An alternative method of display, which simultaneously provides information in the time domain for the individual component frequencies is presented.
3 WAVELET ANALYSIS Fourier series are evaluated by the use of mathematical functions called integral transforms. These transforms are the result of multiplying the original function with a kernel function integrated from -CO to +CO. The kernel for a Fourier transform is (3x1"'.exp(-ixy>. An alternative mathematical technique, known as Wavelet analysis, is based on a similar technique of integral transformation. Wavelet analysis is being applied to a vast array of applications in science and engineering. Kumar and Foufoula-Georgiou ( 1994) provide a good introduction to the use of Wavelets in the related field of Geophysics. The multiplying kernel function in this case is called the mother wavelet, and the limits of integration are finite values at or slightly beyond the measured response. That is there is no implied assump195
A. is a scale parameter which has the effect of dilating ( 0 1 ) or contracting (3,<1) the wavelet the function ( ~ ( f ) t is a localization parameter which translates the wavelet so that the function j [ r ) can be analyzed around the point (or time), t. In order to preserve unit energy at all frequencies, and therefore not bias the integrand, the amplitude of the wavelet must be adjusted in accordance with Equation 1 as the mother wavelet is dilated or contracted. This concept is shown for the Mexican hat wavelet for the cases of h=l, h
tion of’the signal being an infinite series. Transient s i p a l s can be analyzed. Lee and Yaniamoto (1994) claim that every application using FFT can be formulated using wavelets to provide more localized temporal (or spatial) and frequency information. It is this ability of wavelets to obtain frequency content of a process locally in time (time-frequency locali za t i o n ) that is particular 1y import ant . There are many different wavelet transforms available - that is, it is not a unique function. A simple one-dimensional wavelet called the Mexican hat wavelet, is shown in Figure 4. However. the selection of wavelets is not arbitrary. A wavelet 111us t sat i sfy t liree conditions : 1 . I t inust have unit energy 2 . I t must provide compact support. or sufficiently fast decay so that it satisfies the requirement of space (or time) localization. 3. It must have a zero mean - that is the integral of the wavelet function from -cc to +cc is zero. This is called the admissibility condition, and ensures that there are both positive and negative components to the mother wavelet. The Mexican hat wavelet in Figure 4 can be seen to satisfy these conditions.
Figure 5. Dilation and contraction of the mother wavelet
3.2 Time locnlizntion
The temporal variation of frequency content is determined by progressive translation of the wavelet across the transient response using the parameter t. For each value of h, the value of is varied so that the wavelet passes across the entire signal, and the integrand, Wf(h,t)determined for each combination of h and t. This process is sliown scheniatically in Figure 6 for a single frequency of wavelet at 3 arbitrary positions.
Figure 4. The Mexican hat wavelet
The interested reader is referred to Kurnar and Foufoula (1 994) and the extensive list of references contained in that paper for a formal description of the mathematical requirements that apply to wavelets. However, the general procedures which allow wavelets to establish frequency content and time localization are described hereafter: 3.1 Estahlishing frequency content The general form of a wavelet transform of a functionf(t) with finite energy is as follows: I,Jq(h.t)
= I,flZl) = S.j(u)
\\.liere:
yJ&)
n‘tf
h>0
\+,f([zr-t]1.2)ClZl
(’)
Figure 6. Wavelet translation for time-frequency localization.
Lastly. it is noted that when the parameters h and t in the wavelet transform take on continuous values, it is called a continuous wavelet transform. For practical applications, it is necessary for the wavelFtt and the signal to be discretized so that the integra196
tion can be performed. This is then referred to as a discrete wavelet transform. For the discrete transform, the time steps are expanded in relation to the dilation of the wavelet.
Depending on the purpose of the time-frequency analysis, it may not be necessary to impose the strict mathematical requirements of wavelet analysis. If the purpose is only to evaluate the relative time-frequency behaviour, it may be possible, or even preferable to apply a hybrid wavelet/crosscorrelation technique. In cross-correlation techniques, the correlating function should have a similar characteristic to the feature to be extracted from the signal. Such a choice will lead to the highest correlation values. For low-strain pile integrity testing, the input waveform is reasonably well approximated by a halfcycle of a sine wave function. It is evidently that a half-cycle sine wave does not meet all the criteria required for wavelets. 1. The energy may not be unity. However, if an appropriate amplitude of the “mother” sine wave is chosen, a unit energy could result. Furthermore, if the resulting integrands are all normalized prior to plotting, the energy in the mother sine wave is irrelevant. It is only necessary to ensure that the sine wave is appropriately scaled during contraction or dilation so that energy remains constant under all transformations. 2. The half-cycle sine wave does provide compact support and fast decay, thus satisfying the requirement of time localization. 3. As it has only positive components, the halfcycle sine wave does not meet the criterion of a zero integral in the interval -us to +us. The author does not believe this to be critical for the purpose of analysis of the low-strain signals. As suggested before, the advantage of applying the dilational and translation techniques of wavelet analysis to a sine wave form rather than a true wavelet, is in providing an optimum opportunity for correlation with the measured signal. It is believed that the fact that all the criteria for wavelets are not satisfied is not relevant for a relative assessment of time-frequency localization. If, however, the purpose of the analysis is to allow deconstruction, modification and reconstruction of the signal, as wavelet analysis allows, the use of the sine-wave form could not be assumed to satisfy these requirements.
4 WAVELET ANALYSIS OF LOW STRAIN INTEGRITY TEST SIGNALS 4.1 Previozu investigations
The application of wavelet theory to low-strain integrity testing has been pioneered by researchers at Napier University, Edinburgh (Addison et al., 1997; Adclison and Watson, 1997, Watson et al., 1998). In Addison et al. (1997), the authors present finite-element generated responses for imaginary piles 10m in length, with and without a 20% impedance increase at mid-length. They indicate that the 20% iinpedance cliange is difficult to spot in the time domain, whereas the defect can be observed clearly in tile wavelet transformation. The author believes that in practice a 20% impedance change would be clearly discernible in the time domain if a toe response as large as that indicated were obtained. It is assumed that the poor reproduction of the paper does not allow their contention that the change is clearly evident in the wavelet plot to be confirmed. At the time of these papers, Addison and his coworkers appeared to favour use of the Daubechies wavelet, possibly because its shape provides high resolution. The resulting wavelet transformation has a somewhat chaotic appearance, which is somewhat disturbing to the uninitiated. Their data is typically presented as one-dimensional transforms plotted against time, although time-frequency fields are also presented. It appears that at that time, a preference for coiitinuous or discrete wavelet transformations was not clear. The discrete wavelet approach has significant computational advantages and does not senerate redundant data. By comparison, the continuous form does not impose restrictions on dilation incizments, but generates significant data redundancy, and hence takes longer to perform. The team at Napier University gives every indication of continuing with their interesting research in this field. 4.2 Current Investigations
The author has also experimented with application of wavelet analysis, exclusively to actual data collected from field measurements. The similarity between the wavelet technique and cross-correlation technique, in terms of passing a defined wave across the signal to be analyzed are evideni. In wavelet analysis, however, the waveform is dilated or contracted and an appropriate scaling factor applied to ensure constant energy.
5 RESULTS In the following discussions, the term wavelet analysis is used advisedly, taking into account the differences with classical wavelet analysis already referred to. The outcomes of this study, W P W presented in one of two ways.
197
5.2 Time-fi-eqiiencydomciin preseimlion
5.1 Time-doinciin presentation
As the result of time-stepping using a spectrum of wavelet frequencies varying from approxiinately 300 Hz to 15 kHz. The wavelet integrands are normalized to a peak maximum value of 1 for PVJ(h,r). The array of results are then presented as a contour diagram of the wavelet integrands with frequency as the vertical axis and time as the horizontal axis. Examples of this presentation, also compared with the original data in the time domain are shown in Figures 8, 9 and 10.
As the result of time-stepping a wavelet of a single frequency through the original signal. A modified response curve, being the normalized integrand values of the signal and the wavelet results. The results are presented in the time domain, and have a similar form to the original data. This is shown in Figure 7, where a set of data, and the wavelet-modified data are compared. It can be seen that the original data has been smoothened by the wavelet transforrnation.. because the frequencj of the chosen wavelet was lower than the high frequency perturbations in the data. It is, if used in this manner, a useful filtering technique. As always, filtering of data must be undertaken with care to eiisure that significant features are not removed. This algorithm has been incorporated in coinniercially available software (Likins and Rausclie, 2000).
5.2.1 Pile with dcimnge Figure 8 shows the time-frequency response plot for a pile with a significant defect close to tlie pile mid-length. The horizontal axis is time, at the same scale as the time-domain plot also shown. The vertical axis is frequency in Hz, ranging from approximately 300 Hz‘ to 11000 Hz. This is a descending logarithmic scale - low frequency is at the top of the plot and high frequency at the base. The plot shows contours of equal PVflh,t). Areas between the contour lines are shaded so that largest values of PV‘h,t) are darkest. The contoured values vary from +l to -1; the corresponding shades of grey vary from black to white. The plot shows that at approximately 0.92ms, there is a “ridge” of maximum values generated across all frequencies. This corresponds to the impact. which can be seen directly underneath in the time domain plot. As discussed previously, the impact generates a spectrum of frequencies which enter the pile. The peak value for Pvflh,t) occiirs at a peak frequency of about 5500 Hz, corresponding to a wavelength of approximately 0.7m. Another ridge can be seen at approximately 3.00ms. This corre-
Figtire 7. Comparison of original signal arid modified signal after wavelet anaiysis using a single fi-equency
Figure S. Comparison of time-domain data and \vavelet analysis in tlie time-frequency domain for pile with inidlength defect
198
Figure 9. Comparison of time-domain data and wavelet analysis in the time-frequency domain for 0.9kg (2 pound) haininer
Figure 10. Comparison of time-domain data and wavelet analysis in the time-frequency domain for 53kg (12 pound) hammer
sponds to the reflection from the pile toe. In this case, the peak frequency occurs at approximately 3600 Hz. This downward frequency shift is characteristic of this type of plot. As noted previously, the higli frequency components of the incident wave are iiiore quicltly absorbed, and the reflected wave is expected to be more dominated by low frequency components. The 2L/c time is determined by measuring from ridge to ridge, giving a wave return time of 2.1 1ms.
Another feature can be seen emanating from the base corresponding to the high frequency components at about 1.9ms. This is the damage indicator. As stated by Ellway (1987), structural defects in the pile preferentially reflect high frequency components of the incident wave. Low frequency components pass through with little reflection. The damage indication cannot be seen in the low-frequency range of the plot. It can be seen that pile damage indications will be 199
Wavelet analysis is a mathematical technique which allows the data to be viewed in the timefrequency domain. Little research or investigation of the application of wavelets to pile integrity testing has been undertaken to date. However, it appears that this approach may provide a more complete w y! of viewing the data. Significant further work 1s needed to refine our understanding of the application of wavelets to low strain test data, and to its interpretation. The author has undertaken “hybrid” wavelet analysis using some of the principles of wavelet application using a half-cycle sine wave function. The time-frequency responses presented provide deeper insight into the pile-head response, and demonstrate some physical principles previously established by others. At this stage, it is a qualitatively useful method of presenting low-strain data. It has potential to provide useful quantitative analysis with further research.
rcstricted to the high frequency range, whereas soil resi btance, which generates damping responses will dominate the low frequency range.
5.2.2 Pile slruck with difji;r.ent haiiuners Figures 9 and 10 show both time domain and time-frequency domain graphs for the same pile struck lvitli a 0.9kg (2 pound) and 5.5 kg (12 pound) han: mer, respectively. A comparison of the two tests is interesting. The time domain results are first coin pare d. The pile was initially tested with the lighter liammer. Figure 9 indicates the possibility of some impedance reduction above the pile toe. The reflections are not strong, and in the time-domain there is overlap between this “damage” response and the later response presumed to be the pile toe. This makes it more difficult to make a definitive determination of the pile wavespeed. In order to confirm (or otherwise) tlie damage, the pile was struck with the larger hammer so that more energy would reach the toe. Figure 10 shows, howevei.. that there is no clear reflection from above the toe response, and fk-thermore, the toe response has a very long rise time. Again, determination of the pile wavespeed is unclear, and possibly in conflict with the data from the smaller hammer, if tlie rise-to-rise method is adopted. The time-frequency responses are now examined. First. it is clear that the input frequency of the signal generated by the larger hammer (about 300 Hz) is significantly lower than that generated by the smaller hammer (about 800 Hz). The lower frequency content of the returning wave is also evident, particularly for tlie smaller hammer which is subject to greater degradation. As noted before, pile defects are more likely to be detected with high frequency waves. The ridge at about 7.3ms in the high frequency range of Figure 9 for the smaller hammer is the damage indication. No such ridge can be seen in Figure 10, because of the limited high frequency content of this signal. The pile toe is seen clearly and definitively in both plots, and the 2L/c time can be firmly established as 8.1 ms for both tests, thus overcoming the confusion from the time domain plots.
6 CONCLUSION Lo\\ -strain pile integrity testing data is normally plotted in either the time-domain or the frequency domain. The information provided by both sets of data is essentially the same, however, each method highlights different features of the response. However, both methods of data presentation have limitations. As discussed, frequency information is absent from the time domain plots, and the frequency response contains no temporal information.
REFERENCES Addison, P.S., Sibbald, A. and Watson, J.N. (1997). Wavelet analysis : a matliematical microscope with civil engineering applications. Insight Vol 39 No. 7, July, 1997 : 493 - 497. Addison, P.S. and Watson, J.N. (1997). Wavelet analysis for low strain integrity testing of foundation piles. 5“’ Intl. Conf. On Inspection, Appraisal, Repairs, Maintenance of Buildings and Structures. 15-16 Mat, 1997, Singapore. Davis, A.G. and D i m , C.S. (1974). From theory to field experience with the non-destructive vibration testing of piles. Proc. Institution of Civil Engineers, Vol. 57, Part 2, December : 57 1-593 Ellway, I<. (1987). Practical guidance on the use of integrity tests for the quality control of cast-in-situ piles. Ground Engineering, Vol. 20, No. 7, October : 5-13, I
200
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
First experiences in the application of the stress wave theory to foundations in Uruguay A.Gutikrrez, L.Abreu, Ch. Hoffmann & D. Hasard Laborutorio de Control de Culidud de Fundaciones, Instituto de Estructurus y Transporte, Facultad de Ingenieriu, Universidud de La RepL2dicu,Montevideo, Uruguay
ABSTRACT: Since 1995 the LCCF (Foundations Quality Control Laboratory) of the Engineering School has incorporated new advances in addition to the traditional foundation quality control technology combining soil studies, pile driving diagrams, the extraction of core drills for concrete sample testing, static load test with sonic integrity test, and lately the dynamic load test. Here, some real cases are presented, in which sonic integrity testing (SIT) has been used successfully for different purposes .
1 BRIDGE OVER YERBAL STREAM
2 720 ANCAP TANK - LA TEJA REFINERY
This bridge suffered the effects of a serious flooding; its foundations were exposed, and a part of its superstructure was damaged. During the reconstruction of the bridge, one of the support pillars fell over onto an adjacent pillar, damaging it, in such a way that it also needed to be repaired. In order to determine the effects of this on the piles, SIT was used on the possibly damaged piles and on the piles of unaffected adjacent pillars that were not affected. These tests thus differ from the usual pile test in that the caps were already built. It was decided to place the sensor and run the test on the upper surface of each pile cap directly above the pile to be tested. Figure 1 shows the signal obtained by placing the sensor and exiting cavities made especially below the pile cap in order to help obtain clearer signals, that were used in the data interpretation.
Sonic integrity testing was used on a floating tank roof in La Teja refinery of the National Administration of Fuels, Alcohol, and Portland (ANCAP). The case deals with 44 pile driven foundations, all of approximately 10m in length, located on a grid, covering a circle of approximately 16m in diameter. The decision to use a foundation of piles was based on the problems generated by differential settlements in these types of tanks. The problem arises when the lateral wall strains impede the descent of the floating roof which gets stuck, causing firther deformations that can lead to the cracking of the tank walls and extensive fuel loss with the consequent economic and ecological problems. Although specific ground studies haven’t been done for this site, information about the soil properties was available fkom previous studies carried out for surrounding tanks and sites as a reference. According to this information, there is a heterogeneous filling of essentially granulated soil, with low strength zones and a water table at a depth of about 2m. The piling company supplied the pile driving diagram for the first piles that shows there is a higher strength layer between the depths of 5.0 and 6.0 m. The strength then dropped when the pile reached 6 . 0 ~until the depth of approximately lO.Om, where the strength increased again. See Figure 2.
Figure 1. Signal registration with the sensor placed o n the pile cap
20 1
Figure 2. Pile 1 driving diagram
In this case it is seen, in agreement with the integrity signals, that 6 piles broke off at a depth of approximately 5.0 m. The piling company said there were no signs of problems during the construction. In addition, there were detected 5 more piles with a negative reflection at a depth between 5.0 m and 7.0 m, but the signals did not show an actual break. The remaining 33 piles gave very similar and normal signals, fkom which the average signal was taken. The average is shown for the pile 1 in Figure 3 , as a reference signal. In one of the 5 piles with a negative reflection, pile 1, the integrity testing was repeated after a reinforcing pile was made, pile lB, alongside the first one. In this case it was found that the new signal did not have the negative reflection, and it was also compatible with the average signal of the piles without integrity problems. It was inferred to be a minor problem of local impedance reduction at the depth of 5 . 8 a most likely caused by a minor fissure, that was sealed when the adjacent pile was placed.
Due to this experience, in another one of the 6 defective piles, pile 42, a stepped static load was applied, to see if the defect could be sealed with preloading. A maximum load of 150 kN was reached reacting against the piledriver, an integrity test was carried out after each load step. In this case no important modification of the signal was seen. In the attempt to reach higher loads the piledriver lost stability, which is the reason why an important part of the strength was not achieved. Figure 4 shows the signal registration before and after the application of the maximum load reached, as well as the signal after the construction of the adjacent pile, pile 42B. In this case, the same type of signal obtained earlier was maintained, with insignificant changes, showing a narrowed section in the pile.
Figure 4. Sequence of signals obtained on pile 42 before and after the application of the static load and after the adjacent pile construction, 42B.
Figure 3 . Signal registration made on pile 1 before after the construct~onofthe adjacent pile 1 ~in, comparison with the average site signal. 202
In fact, the high number of piles having a defect at the same depth (14%) in a relatively reduced area of ground led to the conclusion that a possible systematic error in the piles construction may have been committed*After contacts with the piling company the conclusion was reached that the piles were reinforced only half way down their length (5m). This
3 GRAIN UNLOADING WHARF EXTENSION COLONIA
way the company could use 12m bars easily available on the market by just cutting them in half without the necessity of connections. Where the bars ended, they were closed off by bending the bars, forming a “nest” that, because of the quantity of bars, left very little space for a proper concrete passage. It is important to keep in mind that the concrete used for these types of piles is a dry concrete that contains thick aggregates. As mentioned above, the zone where the narrow section appeared, corresponded to the passage fiom a higher lateral ground kiction to another with less resistance. For this reason, if during the concreting the case was not carefully raised, and was done abruptly, the pile can be cut off with the invasion of saturated ground. The high number of defective piles found at this depth can then be explained by the difficulties resulting iiom the passage of the thick aggregate concrete through the folded bars, and also due to the transition from resistant soil to a lower lateral iiiction during the case raising. The foundation was reinforced by building several new piles, following the proposal from the planner, all made with continuous reinforcing bars up to the pile toe. See Figure 5.
0
0
0 0 0
0
0 0 O0 0
0
0 0 0
0 O0 42
0 0
0
During the piling phase of a grain unloading wharf in the Nueva Palmira port, a loaded, 20,000 ton barge accidentally hit a group of piles constructed only 10 days before, reacting elastically. SIT studies were done to see if the piles involved had suffered structural damage. In agreement with the information given by the contractor, the piles constructed were drilled piles made of reinforced concrete, all about 27m in length, 0.8m in diameter, and fixed in about 4m of a high resistance ground layer, belonging to the Fray Bentos formation. The piles have steel casings in their upper section of 20m. The signals obtained from the integrity testing presented a clear hammer impact, and a good reflection of the pile toes. The lengths obtained corresponded reasonably with the ones given by the construction company. All signals indicated a lateral fiiction increase below the depth of 20m, that corresponded generally with the end of the steel casing and the direct contact between the ground and the concrete. None of the signals showed a negative polarity reflection, which would suggest the presence of any breaks or narrowings in the shaft of the studied piles. To evaluate the possible damages on pile number 11, its signal was compared with the average signal taken from the adjacent and unaffected piles (9,10, and 12). As shown in figure 6, the signal of pile 11 and the average signal present very similar behaviors, giving reason to affim that the pile did not suffer important damages.
0 0 0 0 OO 0 0 0 0 0 42B
0
References: Damaged pile Reinforcing pile
0
Pile without problems
Figure 6. The signal of affected pile number 11, and the comparison with the average signal Figure 5. Arrangement of the foundation piles of tank 720.
203
shortly. Although there are still no official standards for the tests, many state organizations like the Transportation and Public Works Ministry (MTOP) and the Petroleum National Company (ANCAP), require these controls in the sites under their administration.
The tests allowed the integrity control of the affected piles and the comparison of their signals with the signals of the nearest piles. In this way the theory of structural damage to the piles could be rejected.
4 MACHINE FOUNDATION IN THE TECHNOLOGICAL LABORATORY OF URUGUAY (LATU) As there were no plans or any information about this site, SIT tests were done to determine the normative conditions and thickness of the existing foundation blocks of 2 dynamic test machines. On both foundation blocks, the approximate dimensions were 1.5m by 2.0m. Five signal registrations were taken, a central one, and four peripherals. Because of the contractor’s interest in estimating the block’s default thickness, and the inability to test the concrete, it was decided to use a wave propagation velocity of 3,800 m/s to test the concrete. All the signals presented similar behavior. An average signal was defined for each block and the thickness was determined in each case. In one of the blocks the thickness was estimated at 2.0 m k 0.2 m and in the other at 1.6 m It 0.2 m. Figure 7 shows the average signal of block number 2.
Figure 7 number 2.
Average signal of foundation block
5 DYNAMIC LOAD TEST (DLT) After many difficulties in fine tuning the DLT, the first investigation tests were done recently in piles built in the Libertad - Dolores formation clays. The short term objective is to incorporate this test into customary foundation controls. 6 FUTURE PROSPECTS The foundation quality control and the use of sonic methods to determine the integrity and the bearing capacity of piles, promises to become a reality 204
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Detection and prevention of anomalies for augercast piling 6.Likins & G.Piscsalko Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche & C. M. Morgano Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio, USA
ABSTRACT: On many sites, augercast piles have significant advantages including cost and low vibration. They also have some uncertainty due to the construction methods. Visual inspection during grouting is often difficult. Static test piles have been used to confirm installation procedures and soil conditions, but are restricted to only a small sample of piles. Furthermore, load test piles receive special attention during the installation compared with typical production piling. T o verify installation and increase confidence, Non-Destructive Evaluation (NDE) methods like low strain Pile Integrity Testing are often specified for some percentage of production piles. Because such NDE methods require testing the pile after the grout or concrete has hardened, the repair or replacement can be expensive if problems are found. However, preventing defects through thorough monitoring during installation is preferred over Pile Integrity Testing. Augercast piles can be automatically monitored for grout volume pumped versus depth. With this information available during installation, if a low grout volume is measured for any depth increment, the pile can be repaired immediately while the grout is still fluid. Installation monitoring reduces the need for subsequent NDE tests. While installation monitoring and NDE tests assess pile structural integrity, the bearing capacity cannot be assessed by these methods. Bearing capacity of augercast piles can be evaluated economically by high strain Dynamic Pile Testing methods used to test driven piles if a suitable drop weight is available.
Keywords: augercast pile, installation monitoring, integrity testing, dynamic pile testing 1 INTRODUCTION
The quality of augercast piles often depends on the contractor’s skill. The most critical is the control of auger withdrawal during grout placement (Roberts 1998). However, for a visual inspector, it is difficult to accurately assess grout volume and auger depth simultaneously. Automated measurement systems can accurately monitor volume pumped as a fimction of depth (Likins 1998). This information can guide the operator during installation to produce a shaft exceeding any minimum requirements. Low Strain Pile Integrity Testing of selected piles verifies the installation monitoring effectiveness. This requires striking the pile top with a small hammer and measuring both input and reflections from pile non-uniformities or the pile end (Rausche 1992). Pile Integrity Testing is applied to any pile or
even all piles after the grout is hardened. There are some limitations for long or jointed piles. Bearing capacity can be evaluated on any pile on site using High Strain Dynamic Pile Testing (Likins 2000a). Test Piles can be selected after installation based on their site location (to assess soil variations), on specific pile installation records, or at random. The capability of installation monitoring and subsequent low and high strain dynamic testing allows the engineer to completely investigate the foundation. 2 CURRENT INSTALLATION PROBLEMS MONITORING AUGERED PILES
205
Visual inspection counts pump strokes as a fhction of estimated auger depth to estimate volume versus
depth. Due to a large volume of information, there are many possibilities for deficiencies or errors in recording all the necessary information. Further, the volume per pump stroke can be easily vary by as much as 20% (Likins et al. 1998). Further, pump strokes are usually manually counted per five foot interval; this interval has marginal precision and even the visual determination of the depth may have one foot errors. Alternative methods of more accurate inspection are desirable.
the grout line. During augering, the auger torque is measured. In practice, after the operator inputs the pile name, the PIR-A Control Unit handles virtually everything else. During drilling, the operator observes the auger's current depth and torque. Higher torque (just below the crane stall torque) makes drilling more efficient and reduces the spoils brought to the surface. Geotechnically, the engineer can assess the torque at auger rehsal to distinguish strong soils or low torque.
2. I Automated monitoring of augercast piles during Installation Automated monitoring (Likins et al. 1998) of grout volume pumped versus auger position provides information to the piling crew which can guide the operator during auger withdrawal. The schematic in Figure 1 shows the overall PIR-A configuration. A small Control Unit conveniently located for the crane operator acquires and processes all measured data. The PIR-A depth monitor has a self-retracting cable attached to the auger gear box. As the auger advances or withdraws, the depth is measured by a rotary encoder tracking the cable. A Magnetic Flow Meter accurately measures pumped grout volume to an accuracy within 2%. A Grout Pressure Transducer installed in the grout line (usually near the Flow Meter) continuously measures pressure in
Figure 1. PIR-A schematic.
Figure 2. P R display during grouting phase. At the pile design depth, the operator initiates auger withdrawal and grout injection. During grouting, the measured grout volume pumped per unit depth is displayed graphically as in Figure 2 for pile 162. The operator can adjust the auger withdrawal rate to keep the grout above the target line. This can also be guided by keeping the grouthnc display slightly above the depthhnc display. If the grout per increment is low, the pile can be re-augered immediately through this increment and the pile re-grouted to eliminate defects in the pile and subsequent remedial measures. The operator presses "return" when grout is observed at the surface and "done" when the pile is completed. Upon completion of the pile, monitoring results for each depth increment are printed in the field. Auger time, torque during drilling, and grout volume and pressure during grouting are listed for every depth increment. Summary information shows pumped volume for the auger stem, grout "head" (for the given diameter the head is the equivalent length of extra grout pumped prior to withdrawal), grout volume per pile shaft increment (the most important information), and spill (grout pumped above ground level). Printed results are available prior to moving to the next pile location. When the operator uses the PIR-A results to guide the pile installation, the installed shaR should meet any minimum guidelines established by the engineer. Thus, the quality assurance of the pile is 206
at the right for the pile bottom at 25 meters (example assumes a stress wave velocity of 4150 d s ) . The bottom plot shows a clear pile bottom reflection with a steady velocity signal between the impact and pile bottom, indicating a good pile shaft. The upper plot for another pile on the same site shows a pronounced velocity increase at about 16 m which indicates a reduction in pile cross section or concrete quality. In general, sharply defined changes in the velocity are attributed to impedance changes, while slow changes are usually caused by soil resistance. If soil resistance effects are known from reference piles, then unusual shafts can be identified. Figure 3. Pile integrity tester system. improved. By thorough and accurate inspection, the need for further NDE testing such as low strain Pile Integrity Testing is reduced. Recording of other construction operations such as grout arrival time on site and collection of grout specimens for strength testing cannot be done automatically. The construction process also includes other activities such as screening, installation of reinforcement, maintaining adequate waiting times between casting of neighboring shafts, observing for subsidence for previously grouted piles, site excavation, etc. which require some human supervision. If problems are observed in any post grouting phase, the piles can be subjected to low strain Pile Integrity Testing to assure integrity.
Figure 4. PIT velocity records of a deficient pile (top) and a normal pile (bottom).
3 PILE INTEGRITY TESTING
4 COMPARATIVE RESULTS FOR AUGERCAST PILES
Pile Integrity Testing (Rausche et al. 1988, Likins et al. 2000b) uses a hand held hammer to impact the pile top and generate a compressive stress wave in the pile. Stress wave inputs and reflections (from non-uniformities or the pile toe) are measured by an accelerometer. Figure 3 shows a Pile Integrity Tester (PIT), accelerometer and hammer. This method can be applied to almost any solid concrete shaft. The pile top is prepared by removing the upper concrete if it has been contaminated during construction, and making a smooth location and attaching an accelerometer with a thin layer of bonding material. Accelerations from several hammer blows are normalized, integrated, averaged, digitally filtered and displayed as velocities. Further processing applies an exponential magnification function which restores reflection details diminished by soil resistance, pile material damping or pile nonuniformities. The resulting signal is interpreted by the skilled test engineer. Figure 4 shows an example output for a pile with LengthDiameter ratio of 40. An exponential magnification is applied, increasing from unity value at the left or pile top to a maximum multiplier (40x)
Both Pile Installation Recording for Augered Piles and Pile Integrity Testing have been employed on recent construction projects. Records from these methods are compared and interpreted. 4.1 Case histoly I
The site had 500 mm (20 in) augercast piles ranging in length from approximately 18 to 20 meters (58 to 66 ft). Soil borings generally indicated medium dense sands transitioning into dense sands at a depth of approximately 13.7 to 15 m (45 to 50 ft). The shafts were socketed at least one meter into the weak bedrock formation. Following an initial static test failure on this site (static test pile not tested by PIT or PIR-A), Pile Integrity Testing was applied to several other completed piles and the Pile Installation Recorder (PIR-A) was specified for use on all remaining production piles to improve quality assurance. 207
The Pile Integrity Tester's records of all the shafts in Pier B18 indicated a characteristic increase of impedance in the upper weak soils (evident by a negative velocity to a depth of about 45 R) followed by a decrease of impedance (evident by an increase in velocity) beginning at this depth. Although the impedance decrease cannot be precisely quantified by PIT, it seems likely that the decrease is due to a return to nominal diameter in the lower denser soils from a larger shaft diameter in the upper less dense soils. The PIT velocity record for Pile BlSA, Figure 5, indicates a velocity increase or reduction in either cross sectional area or modulus at approximately 14.6 m (48 R). The PIT toe reflection in the rock socket resulted in a negative velocity (compression) at 17.7 m (58 R). This is typical for shaft sections with either a high soil resistance or an increased pile impedance.
Figure 5. PIT data (case 1). The PIR-A result for the same Pile 18A showed that between 15.2 and 14.6 m (50 and 48 R) below the top that the grout volume per 0.6 m (2 A) increment decreased from an oversized average of about 0.18 m3 to 0.13 m3 (6.5 R3 to 4.66 ft3). The minimum required volume was 0.154 m3 (5.45 R3 ) per increment. Although this one increment was slightly low, the volume for the low increment still had a grout ratio of 1.07 over the nominal hole volume and adjacent increments were higher than required; some redistribution of the fluid grout between neighboring increments is likely. Examination of numerous PIR-A records demonstrated fairly uniform grouting rates, and by implication no significant problems. In addition, selected piles were tested by Pile Integrity Testing for integrity after installation. Because no hrther difficulty was experienced due to adequate PIR-A quality control and confirmed by Pile Integrity Testing, there was no hrther static testing.
record along the shaft is relatively flat and thus indicates a relatively uniform shaft and no defects.
Figure 6. PIT data for pile 349. The PIR-A result is given in Figure 7. The volumes printed are for 610 mm (2 R) depth increments. This depth increment was selected to provide sufficient precision taking into consideration that typically about 8 pump strokes would be required for such a depth increment. As shown, the grout pumped was relatively uniform and generally above the volume required of 4.421 R3 per depth increment representing the 1.25 grout ratio requested. The grout ratio for two segments in the middle to lower part of the pile were slightly below the 1.25 grout ratio desired for the project, but still well above the factor 1.0. Further, the adjacent segments were above the volume required. The low volumes pumped near the pile top (bottom two lines in Figure 7) are acceptable since the grout return was already observed. This early grout return is due to
depth 44 42 s10
38 36 34
63
"59 I45 4.84 4-63
0
7
0
8 0 8
0 0 0 0 0 0 0 0 0
30
4.49
2s 26
1.273
4.73
9
24
4.34
e
22 20
3.01
9 7
0
18 16 14
4.70 1-45
9
0
e
0
.U
8
It
4.56 4.49 4.66
7 8
4.91 1-342 0.991:
8 2
0 0 0 0 0 0
L
0
10
Shaft 349 is a 450 mm (18 in) diameter shaft 13.9 m (45.5 ft) in length and is one of several hundred piles on this project. All piles were installed using a PIR-A. This pile was requested to be tested to assure the PIR-A results were achieving the desired result and as a random check on the pile. The PIT record for pile 349 is shown in Figure 6. The pile toe is very clearly visible at the expected time after a magnification factor of 35 is applies and the velocity
L
22
8 8 7 8
32
4.2 Case history 2
7-88 4.a
8 6 4
2
4-70
9
< 1999-07-26-M~S5 1 -
Figure 7. PIR-A data for pile 349.
208
a
the extra grout head pumped prior to auger withdraw al. 5 HIGH STRAIN DYNAMIC TESTING Monitoring during installation does not assure sufficient pile bearing capacity. Thus, either additional static testing or high strain dynamic pile testing (Hussein et al. 1996, Likins et al. 2000a) are required. High strain dynamic pile testing with the Pile Driving Analyzer@(PDA) has also been applied since 1974 to drilled shafts and augercast piles with increasing frequency using drop weights to generate the loading. In some countries extensive dynamic pile testing is now routinely performed on drilled shafts and augercast piles. The pile top may be temporarily extended above the ground surface t o encase protruding reinforcement or to facilitate attaching the strain and acceleration sensors two diameters below the pile top. A steel striker plate and minimal plywood cushion are placed on the pile top to distribute the impact over the entire top surface. The drilled shaft or augercast pile is then impacted by a simple drop weight guided by a short set of leads. A low drop height is first applied to assess signal quality and alignment of the weight with the shaft. After each impact, the net permanent displacement or "set per blow" is carehlly measured. The drop height is then increased until either the set per blow exceeds 2 or 3 mm (to activate the h l l capacity), or until the indicated capacity exceeds the required ultimate capacity, The PDA monitors stresses directly and additional cushion material can be inserted to avoid overstressing. The measured strain and velocity data can be analyzed on site by the signal matching software CAPWAP@ to independently check if the activated capacity exceeds the desired test load. 6 CONCLUSIONS L o w strain Pile Integrity Testing (PIT) can detect major defects in the pile shaft at low cost and with little effort. However, particularly for very long piles, PIT results may be difficult to interpret and should not be the only means to verify the quality of the foundation. As a minimum, geotechnical borings and field installation observations should be included in the evaluation process of the foundation. For very long piles the method may not provide conclusive evidence of integrity of the whole shaft. Where the Pile Installation Recorder (PIR-A) is installed on an augercast rig, it automatically records the installation of all piles on a job. Grout volume and grout pressure information from the PIR-A during installation guides the contractor into installation of quality piles. The PIR-A records are
used to judge pile consistency and acceptance. With more accurate information available, augercast piles are more readily specified and accepted by designing engineers. If automated PIR-A installation records of grout versus depth indicate a good shaft, then this may reduce the need for hrther PIT tests. Thus, PIT testing can then be restricted to shafts with questionable PIR-A records or shafts with problems observed after installation, during installation of subsequent piles or during excavation. In addition, a small percentage of randomly selected production piles may also be tested by PIT for quality control. Capacity of augercast test piles can be determined by a static load test. However, dynamic pile tests followed by CAPWAP analysis is a well proven alternative for augercast piles when minimum installation time is critical or if multiple tests are desirable to evaluate site variability. Dynamic tests on augercast piles usually require some pile preparation and a drop weight to apply the impact.
REFERENCES Hussein, M., Likins, G. & Rausche, F. 1996. Selection of a hanmer for high-strain dynamic testing of cast-in-place shafts. Proceedings,j$h international conference on the application of stress-wave theory to piles. Orlando, FL, USA. Likins, G., Rausche, F. & Goble, G. 2000a. High strain dynamic testing, equipment and practice. Proceedings, sixth international confirence on the application of stress-wave theory to piles. Sao Paulo, Brazil. Likins, G., Rausche, F., Morgano, C.M., & Piscsalko, G., 2000b. Advances in PDI low strain integrity test methods and equipment. Proceedings, sixth international conference on the ayplication of stress-wave theoiy to piles. Sao Paulo, Brazil. Likins, G., Piscsalko, G. & Cole, C. 1998. Pile installation recorder tests for ACIPJCFA piles. Proceedings, 7ke deep foundations institute 's aiigered cast-in place piles specialty seminar: aiigered cast-in-place piles. Houston, TX, USA. Rausche, F., Likins, G. & Slicn, R.K. 1992. Pile integrity testing and analysis. Proceedings. &iirth international conference on the application ?f stress-wave theory to piles. The Hague, the Netherlands. Roberts, T. 199X. Quality control for augered cast-in-place piles in Texas and Louisiana. Proceecling.y the deep fixinclations institute '.c nrtgeretl cusf-in place piles specialty serninar: aiigered cast-in-place piles. Houston, TX, USA.
209
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama (a Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
cent advances and proper use of P 1low strain pile intepty testing G. Likins Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio,USA
ABSTRACT: Low strain Pile Integrity Testing (PIT) is a valuable tool to locate major defects in drilled shafts o r concrete piles. It has demonstrated its worth in detecting defective piles when other methods are not practical, It requires no advance planning or access tubes so can be applied to any concrete pile after installation. Since the testing is quick and simple, the cost per test is quite reasonable. However, proper application and interpretation are essential to obtain reliable results. The advent of digital processing has opened up a wide variety of possibilities for enhancing the information contained in the data. Many of these procedures will be discussed and examples shown depicting their application. Certain rules of processing are presented and should be followed to assure good interpretation. INTRODUCTION
TESTING P RACT 1C E
Structural engineers are often faced with the question of pile integrity. Difficulties during installation, excavation procedures, slope failures or lateral movements due to accidental impact can create doubt about the integrity of any pile. In some cases, doubt also arises due to lack of information about existing piles when a new use or increased design loads are required. In some cases lack of inspection on current installations leads the structural or geotechnical engineer to question the foundation adequacy.
The pulse echo method (Rausche et al 1992) uses a hand held hammer to impact the pile top and generate a compr-essive stress wave in the pile Figure 1 shows a Pile Integrity Tester (PIT) for pulse echo testing. The small battery powered main unit is highly portable from pile to pile and extrcmel! rugged. It uses a touch screen for input rather than ii computer keyboard. Experience has proven this to be a highly reliable method and extremely user friendly. The screen displays intuitive menus to guide the user and shows the graphical signals for field interpretation. Stress wave inputs and reflection?; (from non-uniformities or the pile toe) are measured as a hnction of time by an accelerometer attached to the pile top. The acceleration is integrated to velocity by PIT, and then interpreted by the test engineer. The pile top surface is prepared by removing the upper concrete if it has been contaminated with soil, bentonite slurry or other foreign materials during construction, and finding or making a smooth location when a driven pile top may have been rough cut. An accelerometer is then attached to the smooth pile top surface as shown in Figure 1 with a thin layer of a soft paste like vaseline, petro wax. etc. Accelerations from several hammer blows are normalized, integrated, averaged and displayed as velocities. Further data processing includes
If access tubes have not been installed in the questioned piles (required preplanning and extra cost) to allow for Cross Hole Sonic (CSL) testing, then the costs for evaluation can quickly become large. Static loading for every pile is obviously prohibitively expensive. Coring every pile is likewise not feasible, and even if it were it would only provide a small sample that could easily miss a local defect at some depth. Therefore, the only practical alternative is often PIT. Since PIT is simple and low cost. it i s conceivable to inspect every pile on a site if required Often a statistical sample is made and if no problems are found, then testing may end. If defective piles ar-e located, then additional tests of neighboring piles ar-e then justified. The selection of which piles to test can be based on various observations. or made considering redundancy of the pile, or at random
21 1
relatively sharply defined reflections are attributed to impedance changes, while slowly changing retlections are usually caused by soil resistance If the etfect of soil resistance is known from reference piles. then unusual shafts can be identified This method can be applied to almost any shaft DATA ENHANCEMEhT AND Ah.4LYSIS How do we obtain a clear graph such as the one in Figure 3? The actual raw data for the defecti\.e pile is shown in Fiyure 3 Figure I . Pile Integrity Tester (PIT) with hammer aid accelerometer for the application of the pulse echo method. application of an exponentially increasing magnification function. The magnification restores reflection details which are diminished by soil resistance, pile material damping or pile nonuniformities. To demonstrate the usefulness of PIT in distinguishing a good pile from a defective one, Figure 2 shows an example output with an exponential magnification increasing from zero at the left or pile top to a maximum multiplier (40 times in this example) at the expected time of reflection from the pile bottom on the right; the pile length is 25 meters (82 ft), and the stress wave velocity is 4 150 m / s (13,610 ft/s) in this example. The bottom plot shows a clear signal from the pile bottom together with a steady velocity signal benbeen the time ot' impact and pile bottom reflection, indicating a good pile shaft. The upper plot for another pile or1 the same site shows a pronounced velocity increase about 16 m (52 5 ft) which indicates a reduction in pile cross section or concrete quality I n general.
Figure 3 "Raw data" from a PIT test Interpretation would be difficult at best since the soil resistance has greatly dampened the reflecting signals. The standard method to compensate for the soil damping is to apply a magnification function that begins shortly after the input pulse with a value of unity and increases exponentially to some factor (40 in this example) at the pile bottom. The amplification function should be shown graphically. The result of the application of this exponential magnification function is shown in Figure 4. In this case the curve
Figure 4: With only magnification
Figure 2 PIT velocity \ S depth records o f a deticienr pile (top) and norind pile (bottom)
also gradually wanders fi-om the zero axis and i n part resembles the magnification function Slow changes are difficult to interpret Generally they are the result of soil resistance effects However, the real goal for integrity testing to evaluate structural pile integrity so soil resistance is 212
not important. Therefore, a high pass filter (I-IPF) is applied which removes low frequency events usually caused by soil resistance. In the case of the extreme example in Figure 4 (most data is generally near zero even after magnification), the data is subjected to high pass filtering which removes components with less than a defined filter frequency. In general, care must be taken to not apply E-IPF with too high a filter frequency which could eliminate the frequencies of interest. The most important frequency components
vibration of reinforcement bars protruding fi-or11the pile top cracked pile top concrete or other- surtice effects. Since the “noise” frequency is much hisheithan the input frequency, if we select a wa\,elet mother function with similar frequency content to the input pulse, unwanted frequency content will be suppressed as shown by the wavelet filtered Figur-e 7
Figure 7: PIT record ofFigure 6 with applied \ v a \ w l e ~ filter Figure 5: With 40 magnification and 25 high pass filter are those that make up the input pulse and the reflections from changes of pile pi opei LIP, Reflections from defects or the pile bottom h o u l d have frequencies which are similar to that of the fo rcing input func t I o n The results shown in Figure 5 were further impr-o\,cd t o the top curve in Figure 2 by use of the wabelet analysis, although in this case the inipr-o\,ement I S slight. The wavelet analysis (Rioul 199 1. Seidel 2000) is a specialized filter that et‘fecti\,el\ st i-engthens t he signa1 freq u en cy CO in p o ii e 11t s t h a t match the input pulse and removes undesirable frequency coin p o n en t s re su 1t i n 2 fir0 m 11o i s e To more clearly demonstrate the effectiveness ot‘ Figure 8: Data with different locations for hammer impact and accelerometer. which is much more easily interpreted. Another consideration in PIT is the location of the accelerometer and the applied hammer impact. The three records shown in Figure 8 have different characteristics. The top curve has the hammer input applied at the pile center while the accelerometer is attached at the edge. The middle cuive has the same accelerometer location but the force is applied at the eytreiile opposite pile edge; obviously this configuration i s less desirable. In the bottom graph. the hammei-
Figure 6: PIT record with significant “noise” wavelet analysis, consider the record in Figure 6 which includes significant high frequency “noise” components in the data caused by PerhaPS by the 213
location. The hammer impact location h a s La[-itxi around the pile at the four compass directions (specifically Nor-th, South, West. and East. respectively). It is relatively clear that there is a local defect near the pile top in the NW quadrant For piles installed i n a soil bearing layer- lie i w reflection is almost always tension because the soil stiffness is less than the pile stif-fness Tension reflections are observed in PIT data as a velocit!, reflection at the toe \uith the same sign as thc impact signal. In all example cases presented in this pqw the pile toe produced such a tension reflections 'l'lic. question often arises as to what magnification factor to apply. In general one should try to make the strongest reflection as high as the impact signal. For similar piles in similar soils on the same site, it is logical that the signals obtained by PIT should be processed with similar parameters and result in
Figure 9: Same data as Figure S but with Wavelet processing input and accelei onieter loc(ttions
Figure 10 Accelerometer at center of SO0 iiiin pile, hammer impact applied at N, S, W and E locations
Figure 1 I : PIT records for two piles of different lengths similar curves. It is relatively simple to identi5 piles with unusual records and thus properties. However, what if the piles are different? For example, the results of tests on two piles installed in similar soil conditions on the same site but with different lengths are shown in Figure 1 1 The toe reflection for the shorter pile is the same amplitude as the input using a magnification factor of 5 However, the longer pile has a smaller reflection using the same magnification This is logical due to the effect of extra soil damping on the longer pile The binary magnification factors (2, 4, 8, 16, 32 ) are shown as the vertical markers inside the graphical representation of the magnification ciirve For the shorter pile in Figure 1 1 with magnification at the pile toe, a magnification of 4 is at about 3 1 ft depth However, for the longer pile, at 31 ft the 214
magnification factor is only about 2.5 while the factor 4 is only achieved at a depth of 47 fi. Thus the use of similar magnification factors for different length piles does not result in “consistent” niagnification functions.
In Figure 12 a magnification of 15 has been applied t o the longer pile resulting in a factor of 4 at about 31 fi so this then is a consistent magnification as a function of depth. The toe response is now of similar
Figure 12 Pit records &ith consistent different length piles
frequency domain, then it is even more apparent when presented in the standard time record The major advantage for the frequency aiialj,sis i n our opinion is to obtain a quantification of shaft qualit), by calculating the relative pile stiffriess (piles v, it11 lower stiffness are more likely to be defective.) One other tool associated with PIT is to obtain an impedance profile. This makes a shape visualization as in Figure 13. However, experience has shown that this is not often required and piles can be classified without this extra effort for most piles.
iiicigiiificdioii i;,i
amplitude for both piles even though the lenyth\ ni c different It appears that the longer pile I S only about 13 7 m (45 f t ) long rather than the planned I 5 S iii (52 ft) While there IS some uncertaintv ( u p to 10” U ) in the wave speed of concrete for the wiie concrete suppliei, the same wave speed i s usunllq u\ed t i ) i ‘111 piles at one site to plot the records If the roe reflection occurs at a time that indicates more than 10% variation of wave speed, the pile length I \ questioned I n this case, the length of the \Iic)i t e i ~ I I C matches well the toe signdl with the iioimnllj assunied wave speed of 3,960 in/\ ( 1 i 000 ti/\) I 0 1 the longer pile the wave speed ~ ~ o u lhdn \ e t o tw 4,580 m/s ( 1 5,000 ft/s) to indtch the toe s ~ g n ~l ibli t h (1 pile length of 15 8 in (52 ft) Since this wnve speed exceeds the average one bv inore t h n n 10” i t \iotiI(I be concluded that the pile was indeed t o o h i t There are further processing options for data Data can be analyzed i n the frequency domain for iiiobilit! and dynamic stiffness. However, determination of‘ length or shape is better made i n the time domain a s presented in the exainples in this paper- since the magnification function can only be used in the tinw domain. Lack of a magnification severely limits the effective length that can be analyzed i n the frecluenc:), domain. Further, if something is clear in the 215
Figure 13 : Processed signal and calculated profile
RECORD CLASSlFICATION Before embarking on any type of pile integrity test, there should be a clear understanding among all parties involved as to the consequences of the test and the action to be taken if the test generates a question in the integrity of the foundation. The test results can generally be placed into the following 4 categories and each one of them may require a different course of action.
A -
Clear toe response. No apparent defect.
B
Clear indication of serious defect. reflection is usually not apparent.
-
C -
Toe
lndication of possible defective shaft, although toe response is apparent.
D - lnconclusive data
Category A piles are good piles within the accuracy limits of the method. For example, a it is generally agreed that a defect that affects less than 20’10 01‘ the pile cross section cannot be detected with certaint!, (actually this a strength of the method since it does not create questions where only minor problems exist.) In this category pile integrity is satisfactoiy and unless there is some othei- reason to suspect the shafts, they are generally acceptable Of course, this assumes that also the pile length indication i h satisfactory. Category B piles are somehow defective and some contingency plan must be used Extra tests could be made. If the defect is near the ground sur-t’ace. excavation to expose and repair the dettctile pile portion is generally possible. The shaft could a l s , ~be cored and the defective are pressure grouted, a subsequent PIT could confirm the effectiveness of‘ any repaii- measure The pile could be abandoned and simply replaced The course of action inav depend on the cost of a new pile vei-sus the cost to repaiiCategory C piles may also be assigned a reduccd capacity. Also, other pile tests may be considered. or excavation if the defect is located near the y-ound surface. If the pile is a friction pile and the det’ect is located far down the pile. the upper s o i l resistailcc above the defect may render- the h i - c e in thc p i l e 21; the defect as acceptable structur-ally and the ciclilii inay in some cases be not serious Category D piles may have poor data dcic to pooi. concrete quality at the pile top. Tr-imming the pile to a lowei- elevation where good quality concrete IS assured and retesting the pile is an alternatiie Ollici reasons for inconclusive results such as a long oi irregular pile are discussed in the next section I t is reasonable to accept a certain percentage of piles with inconclusive records at a site where soil and pile conditions produce complex recoi-ds.
concern where there are localized defects of small extent. The method works particularly well on solid pile cross sections, such as concrete or timber piles. It has only limited applicability to steel piles. Of course, concrete filled pipes can be tested. In cases where there are mechanical joints or full section cracks, the wave cannot usually cross the resulting gap. Therefore only the portion above the gap is really tested. Defects which are very short compared with the input pulse width and do not cover the entire section can be misleading due to superposition of the waves generated by both initial reduction and following increase. A similar problem exists for gradual changes which create relatively slow changes in the record. Normal processing of PIT data then removes the reflections from gradual changes since they are similar in frequency content as the soil resistance effects. Therefore if a pile gradually increases and then quickly returns to the nominal cross section as it enters a new strong soil layer, the sudden reduction inay produce a record that indicates a defect. I t is, therefore, important to compare any PIT result with the general soil profile to assess the likelihood of these situations. There is a general rule of thumb that the test effectiveness is limited to piles with lengths less than 30 times their diameter (L/D ratio) This is really applicable to the length embedded in soil since there is little signal degradation for shafts acting as columns in water until below the mudline. In some cases it might not be possible to “see” even this far down the pile. However, with recent improvements in electronics including lower noise and higher precision 16 bit sampling (which gives 8 times more resolution to older 12 bit samplers), PIT now regularly tests piles with L/D ratios far in excess of 30.
LlMITATIONS
PIT is a very useful test for many sitiiatioiis I n ~ i i i e cases it is the only practical alternative to doins nothing which may not be acceptable i z hci-e questions arise during constr-uction I t is certaInl\ ;i simple and relatively inexpensive test and thus i s widely used. It is intended to locate major defects and to that put-pose it generally is adeqiratc l r i h ii strength of the method that it does not generate
216
The pile length effectively investigated by PIT depends not only on soil damping, but also on the non-uniformity of the pile, because any change in pile quality or shape produces reflections. The first nonuniformity detected will be more reliably analyzed than additional sources of reflection farther down the shaft. Highly non-uniform piles produce complex records which are dif’ficult to analvze I n such c a w h the records collected under comparable circumstances can be compared Records thitt dill;,i
are obtained by avoiding the test on contaminated pile top concrete and by assuring meaninghl records where the pile top is large relative to the hammer size. In many instances it is helpful to vary hammer size and impact and measurement location at the pile top. The former to produce records that have sufficient resolution and the proper frequency content for a particular pile-soil system, the latter to find any defects near the pile top but only at one side.
significantly from the “normal signature“ should Ix hrther investigated. PIT gives no information about load carrying capacity To confirm capacity a static loaad test or a dynamic pile test are required (Hussein 1996, Likins 2 000a). OTHER APPLLCATLONS, SOLUTIONS PIT is essentially an after-the-fact test. The pile must be already installed and the concrete hardened I f a major defect is present in the pile, it can be detected. but the consequences and costs of repair- are relatively high and it is obviously better to avoid an)’ defects. An installation monitor for ContinuousFlight- Auger Piles also called Auger-Cast Piles is an instrument that inspects the pile during installation when correction of the defect is both simple and relatively inexpensive (Likins 2000b). Even U i t h such installation monitors, the need for PIT rernains. for example, when problems occur after the installation such as lateral impacts, slope failures. or often merely the need to evaluate existing piles Piles have also been often successfully tested iificithey had already been put to service under a structure (Hussein et al., 1992). However, the more complex the structure-foundation system the more difficult will be the data interpretation. For example. it is possible to tests piles located below a pile cap. but the bottom of the pile cap will produce a mquireflection. This is acceptable if the pile below the c a p is uniform and of reasonable L/D ratio. Testing U I I ii pile below a structure will produce waves that travel up in the structure as well as down the pile To clearly separate reflections of pile from those of the structure it is often necessary to install two accelerometers on the side of the exposed pile z t t\\v different levels.
CONCLUSIONS Modern digital electronics and signal processing techniques have made it possible to build equipment and develop analysis methods that extend the natural limits of the simple Pulse Echo Method Although the PIT equipment meets all the demands that can be put on a modern equipment, successhl testing still requires experience both during the test process itself and when analyzing the data. The field engineer has to make sure that clear records
The analysis and data interpretation process also has to be done carefully, particularly when choosing magnification and filter parameters. Fully automated data processing is usually not advisable: each step of the data processing procedure should be reviewed as to its effects on the records. The processed data should be plotted and the records classified. Depending on the data interpretation certain additional actions may be required on the construction site. It is important that such possible actions are clearly established prior to conducting PIT.
REFERENCES Hussein, M., Likins, G., and Rausche, F., (1 996) Selection of a Hammer for High-Strain Dynamic Testing of Cast-In-Place Shafts.. Proceedings, Fifth International Conference on the Application of Stress- Wave Theory to Piles, Orlando, Florida, USA. Hussein, M., Likins, G. and Goble, G., (1992) Determination of pile length under existing structures, Proceedings, Deep Foundation Institute annual Meeting, New Orleans. Likins, G , Rausche, F., and Goble, G., (2000a) High Strain Dynamic Pile Testing, Equipment and Practice, Proceedings, Sixth International Conference on the Application of Stress- Wave Theory to Piles, Sao Paulo, Brazil. Likins, G , Rausche, F , Morgano. C M , and Piscsalko, G (2000b) Detection and Prevention of Anomalies for Augercast Piling, l’rocrcw’iiicy\. Sixth International Cwference 011 he A p p f i u u i o i t of Stress- Wave Theory to Piles, Sao Paulo, Brazil Rausche, F , Likins, G and Shen, R K ( 1 992) Pile Integrity Testing and Analysis Protecrdinp Fourth Infernational (’onftciznte o t i t h e
217
Applicafioti of Stress- W m e 7heoq: t o 1)ile.s. The Hague, The Netherlands.
Rioul, 0 and Vetterli, M (1991) Wabelets and Signal Processing IliLL S i p i d Proco\ jiiig Magmine. Vol 8, no 4 October, p 14-38 Seidel, J.P., (2000). Presentation of low strain integrity testing in the time-frequency dornain. Proceedings, Sixfh Iuiferticrtiouial ( ' o i ? f i . l - ~ ~ i i c011 ~e fhe Applicatioii of Stress- Wave 7hc.o~~: to 1'ilcJ.s. Sao Paulo, Brazil.
218
Application of Stress-Wave Theory to Piles, Niyama 13Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Stress wave propagation velocity at early ages C. Restrepo Esninesa and Restrepo, Bogota, Colonzbia
ABSTRACT: 1 INTRODUCTION
Table. 1
In order to ensure the continuity of a cast-in-place concrete pile Pile Integrity Tests (PIT) can be performed with certain advantages such as low cost and non difficult pile head preparation. However, given that the stress wave propagation velocity is assumed during the PIT analysis, the ASTM standard requires that concrete must have at least seven days of age, moment what is considered that concrete’s wave propagation velocity vs time curve starts to flatten. Given the construction processes, sometimes an early PIT analysis on a certain pile could be desirable to help the decision making in the construction process. In order to investigate the stress wave velocity at early ages (younger that the seven days given by the ASTM standard) the geotechnical instrumentation Department at Espinosa & Restrepo ltda with the collaboration of senior undergraduate students of the Universidad Javeriana in Bogota Colombia planned and performed a laboratory testing program. The present paper describes the methodology, the results and conclusions of the tests performed. 2
TESTING PROGRAM
2.1 Probes preparation Using concrete or mortar mixed by a commercial concrete plant owned by Metroconcreto S.A based in Bogota, Colombia; 4 concrete probes 5 meters long and square section 0.20 wide were casted for each of the following compressive strength mixtures: 3000 psi, (20.68 Mpa), 3500 psi (24.13 Mpa) 4000 psi (25.58 Mpa) and mortars 3000 psi (20.68 Mpa) and 3500 psi (24.13 Mpa). The waterlcement ratio for the concrete used were a s follows:
I Type Concrete Concrete Concrete Mortar Mortar
I Nominal compressive strength 4000 3000 3500
IW/Cratio
0.60
During the probe’s cast, concrete or mortar the materials were sampled for compressive strength tests at ages 3 days, 7 days, 14 days and 28 days. The probes were sampled and tested according ASTM standards. To ensure proper curing all samples were covered with wet fabrics to guarantee moisture content during the curing process.
2.2 PIT testing Over all 4 probes for each material type PIT tests were performed every day from day 1 to day 28 to monitor the stress wave propagation velocity change. During testing enough PIT registries were taken for proper statistic analysis. In order to obtain appropriate PIT curves a 450 grams hammer was used given that with such mass a the pile tip was registered with good definition and the length of initial peak was minimized; all PIT measurements were performed by the same technnitian in order to achieve the highest homogeneity in the PIT registries. 2.3 Error analysis During the analysis 3 error types where statistically analyzed using Pile Dynamics software and basing the analysis on velocity or length variations, and the results obtained are revised as follows:
219
2.3.1 Heterogenity between probes within a concrete strength Includes variations introduced at the casting moment by the fluid concrete mix vibrating of the probes. To measure this error the coefficient of variation of the velocities obtained using PITPC software.
REF. RANGE AVERAGE STAT DISTRIBUTION OCURRENCE PROBABILITY (note Nol) Whole sample 1.0% - 3.0% 2 CHI SQUARE 65%
REF
RANGE
AVERAGE STAT DISTRIBUTION
3
RESULTS
Given the curves obtained for every ages within a concrete or mortar strength vs wave propagation velocity two intervals can be clearly differenced. The first includes mixtures of ages between 1 to 7 days were an strong dependancy between age and velocity. All four probes data velocity vs mix age at early ages (1 to 7 days) were averaged and regression analysis were performed to modeled the age-velocity relationship during early. The following chart resumes the regression analysis performed and the results obtained at ages from 1 to 7 days. REGRESSION ANALYSIS
Concrete type
OCURRENCE PROBABILITY (note Nol)
EQUATION
Type CONCRETE 3000 prl CONCRETE 3500 prl CONCRETE 4000 pd MORTAR 3000 psi MORTAR 3500 prl
Log. Polinomlal Pollnomlal Log. Polinomial
V=234.99Ln(t)+2864.2 V=O.40581‘.6.4744t’+ 15.931t’+ 138.16t+3018.9 V = l.2937t‘-18.688~’+56.102t’+209.44t+2374.9 V=256.05Ln(t)t2966.4 V = 15.452r’-228.6t*+ 1083.8t+ 1863.9
R’ 0.980 0.977 0.998 0.981 0.992
4 CONCLUSIONS The possible errors within the testing program correspond to normal conditions during the construction or PIT performance and they were reduced to acceptable ranges. The wave velocity is directly proportional to the concrete‘s mix age during the first 7 days of curing process. According to R2 obtained the regression analysis performed characterizes accurately the stress wave velocity at early ages; for ages older than 7 days the stress wave velocity reduces its dependancy of age and becomes a property related to the compressive strength of the concrete or mortar.
2.3-2 Heterogeneity during PIT curves acquisition Heterogeneity introduced by variations of the PIT technician criteria, accelerometer disposition etc. Such error was statically analyzed in terms of measured pile length based on the results obtained with PITPLOT software. The variation of length error’s magnitude obtained is as follows: Table.3
According to the results the investigators conclude that the error introduced was reduced within acceptable ranges.
2.3.3 Errors within the PIT analysis. It is given by the variations generated by tests performance variations, variabilty within the analysis given variations on the technitian’s criteria who employs the software tools for a particular test. For this particular testing program most of the software tools such as high pass, magnification and pivoting were not used given that the data acquisition was closely controlled to obtain the best possible curves.
The correlation equations obtained between age and stress wave velocity for every concrete or mortar tested fits well reflection R2 values close to unity. In general the result obtained with the mortar reflects a more consistent behavior; in the opposite the concrete reflects results slightly more scatter. The researcher attributes this behavior to the fact that in the case of mortar probes the ratio between aggregates particule size/cross sectional side is much lower than for the concrete. i.e in concrete the ratio varies from 0.10 to 0.19 and for mortars it rages around 0.05 or lower obtaining a much more uniform media for the stress wave to travel The results obtained with the 24.13 Mpa(3500 psi) reflects a reduction of the stress wave velocity at 28 days of age. The researcher concludes that external effects could have influence such behavior. There is a clear tendency for mortar to get higher stress wave velocity than concretes. 220
22 1
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Examination of a new cross-hole sonic logging system for integnty testing of drilled shafts Samuel G. Paikowsky Geotechnical Engineering Research Laboratov, University of Massuchusetts, Lowell, Mass., USA
Les R.Chernauskas & Leo J. Hart Geosciences Testing and Research Incorporated, North Chelmsford, Mass., USA
Car1 D. Ealy & Albert E DiMillio Federal Highway Administration, McLean, Va., USA
ABSTRACT Drilled shafts and other mixed or cast-in-place concrete deep foundation elements can be costly solutions These foundations usually carry very high design loads, and often serve as a non-redundant, single load-carrying unit These conditions have created a need for a high-level of quality assurance and control applied to each in-place constructed deep foundation element The non-destructive testing method, Cross-Hole Sonic Logging (CSL), currently offers the most reliable technique for assessing the integrity of in-place constructed deep foundation elements Recent years have seen progress in CSL instrumentation, taking advantage of the available computer technology The software applications, however, have greatly fallen behind, thereby limiting the effectiveness and potential of the CSL method and deep foundations integrity testing in general A new, original CSL testing system by the name of PISA (Pile Integrity Sonic Analyzer) makes use of an innovative software and data acquisition system, hence representing the state-of-the-art in deep foundation integrity testing The PISA has the capability to show real-time graphical information during logging, including planar tomography, which can identify the boundark! of a compromised zone within the foundation element The equipment operates completely in a Windows graphical environment allowing alphanumeric and graphical reports to be generated directly into word processing software The real-time graphical representation during logging and the ease of reporting enables immediate, extensive on-site evaluation and decision-making The PISA system was evaluated on different construction sites The case history presented in this paper relates to a class ‘A’ prediction as tests were carried out on shafts in which defects were intentionally planted The test results were submitted before the defects locations were known, both presented I the paper The obtained results demonstrate the ease of use, accuracy of measurements and enhanced capabilities of the PISA The systems’ abilities are shown to be superior to any other currently available commercial system Cross-Hole Sonic Logging (CSL) is a common testing methods for determining the integrity of inplace constructed deep foundation elements, such as drilled shafts and caissons. A minor variation of this method, called Single-Hole Sonic Logging (SSL) can also be used on smaller diameter drilled mini-piles and augercast piles. These methods are both non-destructive testing (NDT) methods and involve generating a sonic pulse with one transducer (transmitter) and picking the signal up with another transducer (receiver). The transducers typically consist of a geophone or accelerometer The methods differ only in the number of tests per pile and the locatiodorientation of the transducers within the pile. Significant improvements and advances in instrumentation, data acquisition hardware, and
1 INTRODUCTION
Deep foundations integrity testing mostly applies to foundations constructed on-site from concrete or grout, such as drilled shafts, drilled mini piles, pressure-injected footings, and pre-cast concrete piles. Drilled shaft foundations usually carry very high design loads, and often serve as nonredundant, single load-carrying units. The integrity testing is required for quality control during construction to detect flaws in the pile (e.g. necking, cracking, void, poor quality material, etc.) common in these cast-in-place concrete piles. As a result of the increasing design requirements on these foundations, a need for a high-level of quality assurance and control has been created.
223
computer technology have been made in recent years. The software applications, however, have greatly fallen behind and have not taken full advantage of the existing technological advances, thereby limiting the effectiveness and potential of the CSL method, as well as other deep foundations integrity testing methods (Chernauskas and Paikowsky, 1999). A new state-of-the-art CSL testing system has recently been developed that utilizes unique soft. ware to take advantage of the new hardware (Amir and Amir, 1998a) This system is called the PlSA (Pile Integrity Sonic Analyzer). The PlSA is based on a lightweight, portable, pen touch, computer that operates in a Windows graphical environment. This system is easy to use and efficient with regard to its ability to make the collected data available in a real-time manner. The following paper provides the basic background theory on the CSL integrity testing method, a description of the PISA system, and a summary of a recent case history including large size rock socketed drilled shafts, defects detection, and verification. 2 OVERVIEW OF ULTRASONIC INTEGRITY TESTING METHODS
2. I
< iwss-Hole Sorm h g g ~ ~ g
Cross-Hole Sonic Logging (CSL) is the most common integrity testing method for drilled or cast-in-place foundations A piezoelectric transducer is used to generate a signal that propagates as a sound (compression) wave within the concrete, while another transducer is used to detect the signal Each transducer is placed into a vertical PVC or steel tube that has been attached to the reinforcement cage and filled with water prior to the concrete placement The water acts as a coupling medium between the transducer and the tube A typical tube arrangement and testing principles are presented in Figure 1 The source and receiver transducers are lowered to the bottom of their respective tubes and placed such that they are in the same horizontal plane. The emitter transducer generates a sonic pulse (on the order of 10 pulses per second), which is detected by the receiver in the adjacent tube The two transducers are simultaneously raised at a rate of around 300 mm/sec (1 ft/sec) until they reach the top of the drilled shaft Typically this process is repeated for each possible tube pair combination (perimeter and diagonals) Figure lb shows the six tube combinations that can be tested (logged) using a configuration of 4 tubes within a drilled shaft Increased shaft diameter calls for a larger number of tubes, which increases the number of
Figure 1 Typical CSL testing setup showing (a) transmitter and rccen er at dffcreiit depths. aiid (b) plaii 1 ien of thc CSL tubes 111111 possible test coinbiiiatioiis
combinations and thereby the resolution of the testing zone. In homogeneous, good quality concrete, the stress/sound wave speed, C, is typically around 3,800 m / s (12,000 to 13,000 fi/s) and is related to the modulus, E, and bulk density (unit weight, y, and gravitational acceleration, g) as follows:
If for any reason the condition of the concrete is compromised, the wave speed will be reduced relative to that of the "good or sound" concrete value. Figure 2 presents a typical sonic signal for which the propagation time between the transducers is measured. The vertical axis is the signal amplitude (microvolts) and the horizontal axis is the time (microseconds). The point where the amplitude begins to rapidly fluctuate indicates the arrival time of the signal to the receiver (a.k.a. threshold time) Since the distance between the two tubes is known, the wave speed of the con-
Crete between the tubes can be evaluated by the fo 110wing relationship:
tion the transducers in different elevations to create more signals, allowing the development of a tomographic presentation of the investigated zone. The limitations of the method include detection of defects only when they exist between the tubes. The testing can be performed only on drilled shafts for which access tubes were installed. Debonding between the tubes and concrete is common if testing occurs long after the concrete placement. Testing in fresh concrete is also difficult as certain zones may cure at a lower rate, creating difficulties in the interpretation of the threshold time and energy. These zones may therefore be interpreted as poor quality concrete.
C = -L t
The wave speed in equation 2 is only an estimate, as the identification of the arrival time, t, is subjective and the distance between the tubes, L, is known only at the top of the shaft. The signal arrival times can then be plotted with depth to generate a log for the particular tube combination as presented in Figure 3. In addition to the threshold times, the energy of each signal may also be plotted with depth. This information can be used to compare signals of one zone to another where lower energy and/or later arrival times correspond to a compromised concrete quality andlor defect. Advantages to this method include the direct assessment of pile integrity and the ability to posi-
2.2 Single-Hole Sonic Logging Single-hole sonic logging (SSL) is a variation of the direct transmission CSL method in which the source and receiver are placed in the same tube and the signal travels in a vertical direction (refer to Figure 4). For drilled shafts and caissons, the method is limited to defects adjacent to the tube and is usually used only when a drilled shaft requires integrity assessment after construction. Due
Figure 2. CSL typical testing signal.
Figure 4. Typical SSL testing set-up showing transmitter and receiver at different depths.
Figure 3 Presentations of CSL test results in the form of threshold time and energy with depth.
225
zone is detected in this stage and the tomographv option is enabled, the probes are lowered and raised relative to each other around the suspect zone, to fkrther investigate and delineate the area. The signals can be examined and adjusted by manually picking the points or using preset algorithms to automatically determine the first arrival time (FAT) as shown in Figure 7.
to high coring costs, a single hole is advanced (often down the middle) to the bottom of the shaft or slightly below the depth where a defect is anticipated. It may also be desirable to perform SSL during CSL testing to isolate the location of a defect at a certain depth (i.e. distinguishing whether the defect identified using CSL is adjacent to the tube or in between the tubes). Recently SSL has been performed within smaller diameter drilled mini-piles and augercast piles (Amir and Amir, 1998b) The use of SSL in these foundation types may become more commonplace in the near f& ture, as research and experience provide insight for the most efficient vertical placement of the tubes to assess the lateral integrity. Brettman and Frank (1996) describe a comparison between CSL and SSL tests 3 THE PISA CSL/SSL TESTING SYSTEM
3. I Gene~crl
The PISA (Pile Integrity Sonic Analyzer) is a modular system allowing for adoption, upgrade and incorporation of additional integrity testing technologies The current integrity testing options available in the PISA include cross-hole sonic logging (CSL) and single-hole sonic logging (SSL) using CHUM (Cross-Hole Ultra Sonic Module) and sonic echo (a k a small strain propagation) using PET (Pile Echo Tester) module Additional modules are currently under development In addition to its modularity, two advantages of the PISA integrity testing system over other systems include its software and portability The PISA is the only Windows 95/98 based system and is also compatible with Word 2000 The software is updated periodically to incorporate new developments and algorithms that make data collection, interpretation, and report preparation easier and efficient The PISA is lightweight (only 42 3 N (91/2 Ib)) and self powered, hence can be easily carried around from shaft to shaft or site to site This feature is also beneficial for air travel The system can be also used as a standard laptop, saving the cost and space required for an additional personal computer (PC) when using a dedicated CSL testing system Figure 5 presents a photograph of the PISA system, including computer and sensors As a scale, the width of the computer screen is 23cm (9 inches) Figure 6 presents the layout of the pile screen, where one can enter the pile information and select the tube orientatiodlocations Selection of the desired tube combinations is accomplished by drawing a line between any two tubes Realtime graphical presentation of the concrete integrity is provided during data collection If a suspect
Figure 5 . PISA system components.
4 CASE HISTORY
4 I Rnckgrormd Four drilled shafts were constructed at the Auburn University in Auburn, Alabama as part of a research study The shafts are 914mm (36in) d]ameter and approximately 11 Om (36fi) long Each shaft was equipped with 4 access tubes and various defects were installed during construction The defects were constructed as soil inclusions formed by sand bags made of a tough material and tied to the rebar cage The cross sectional area of the defect was based on the measured perimeter of the bags following their installation The concrete was poured into a dry hole which was cased the full length of the shaft Over a year after construction, Geosciences Testing and Research, Inc (GTR) personnel tested the shafts using the PISA system and submitted the results to Prof Dan Brown of Auburn University The actual "manufactured" defects were then revealed and comparisons were held between the predicted and actual defects 4.2 Detected w. Planned Defects
Figures 8 and 9 summarize the defects as detected by the PISA testing (on the right hand side) versus those planned/manufactured during construction (on the left hand side).
226
Figure 6. Layout of the pile screen.
Figure 7. Data collection screen.
227
Figure 8. Presentation of manufactured defects with predicted results of PISA CSL testing system for Shaft 4 at the Auburn University test site.
trated in two zones which may be the case following the casting The lower defect was identified in the right location, however was marked as approximately 25 6% of the cross-sectional area versus the manufactured defect planned as 10% of the cross-sectional area. Shaft 9: Two zones of defects were identified in shaft 9 (Figure 9). The upper defect was identified as the right size but at a location approximately 0.6m (2ft) above the center of the actual location. The lower defect was identified at the right location and with the correct size Shaft 2: A soft bottom in part of the cross-
The description of the manufactured defects were for example: "Shaft 4 - soil inclusion at 11 to 13ft below the ground surface on the north side (towards shaft 9), covers 20% of the crosssection." As such, the defects outlined in Figures 8 and 9 are a reasonable approximation of the descriptions. Shaft 4: Two zones of defects were identified in shaft 4 (Figure 8). The center of the upper defect was suggested to be about 0.6m (2ft) above the center of the actual defect with overlapping margins. The area of the defects was correctly identified though the test suggested that it is concen-
228
Figure 9. Presentation of manufactured defects with predicted results of PISA CSL testing system for Shaft 9 at the Auburn University test site.
5 SUMMARY AND CONCLUSIONS
section was identified in Shaft 2 This was not an intentional defect and may have resulted from the regular construction process Shaft 7 A weaker zone was identified between 1 5 to 3 7m (5 to 12fi) along one segment of the shaft (tube 2) No intended defect was installed in this shaft The shaft, however, was loaded to failure in bendins and extensive tension cracks were expected to be developed on the south side around 3 7m (12ft) This information, like all other information, was provided after the tests results were submitted
Four relatively small size sand inclusions were installed in two shafts out of four constructed shafts. All the defects were identified in the tests conducted over a year after construction. Three out of four defects were identified in their approximately correct size, the fourth defect was assumed to be about 3 times the size of the actual defect. Two of the defects were also identified within approximately O.Gm (2ft) of their actual locations. Overall, the test results of Class 'A' prediction provided accurate and reliable evaluation. The to-
229
mography feature of the testing equipment certainly allows an operator to estimate the extent of the defected zone with a higher accuracy than ever before. The challenge of finding the defect seems to be smaller than its accurate description. The latter, however, is of great importance in order to be able to conduct structural evaluation of the defected shafi and hence to assess the need of remedial action. In summary, the PISA system represents a new generation of CSL equipment capable of conductins non-destructive testing with ease and accuracy not available before. A C K N O K E DEMENT S The evaluation of the PISA system was possible with the support of the Federal Highway Administration (FHWA) The cooperation of Dr. Dan Brown of Auburn University enabled the testing of the drilled shafts
WFERENCES Amr. E I and Aiiiir. J M . (1998a). "Testing Piles rvith Virtual Iiistrunieiits". Proceediigs of the DFI conference. Vieiuia. Austna. pp 4 1 1 - 1 1 4 Amir. E I aiid Ainir. J M . (1998b) "Rcceiit AdLances i i i Uitrasonic Pile Testing.' 3"' Iiiternatioiial Geotechiiical Seminar 011 Bored and Augured piles. Ghent, Belgium Baker. C N . Dnimnglit. E E . Bnaud, J . Mensah-Dnuiiiali. F aiid Parikh. G . (1992). "Dnlled Shafts for Bndge Fouiidatioiis". Final Report Under Contract No DTFH6 1-88-2-00040, Federal High\%a) Administration. Fcbruaq Brcttmaii. T aiid Frank. M . (1996). "Coiiipansoii of Crosshole and Single Hole Sonic Iiitcgrit? Logging Methods". Procecdlngs of the 5'" International Coiifcreiicc 011 t~ic Application of Stress W ~ eLTheoq to Piles. Orlaiido. Florida. September. pp 698-707 Clicmauskas. L R aiid Paikov slq . S G . (1 999). "Deep Foundations Iiitcgnt! Testing Techniques aiid Case Histories". Cir 11 Eiigmeeniig Practice Spriiig/Suiiimer 1999. ~ o l14. no 1. Boston Soclet) of Ciril Engneers. SectiodASCE. pp 39-56
230
4 Pile-soil interaction
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Keynote lecture: Identification of soil-pile model interaction parameters from recorded time-displacement signals Abdallah I. Husein Malkawi & Izzaldin M.Ayasrah Civil Engineering Department,Jordan University of Science and Technology,Irbid, Jordan
ABSTRACT: A series of pile load tests were conducted and time-displacement signals were obtained at combinations of pile installation depths and sand’s initial densities. The signals were successfdly analyzed employing the Logarithmic Decrement and a semi-iterative procedure for under-damped and critically damped signals, respectively, in signal matching processes. The proposed signal matching procedure was shown to be a powerful tool in evaluating soil-pile model parameters, more specifically damping constant and stiffness Damping constants and damping coefficients were then plotted against pile depth, sand density, and pile static bearing capacity that was known for every test soil-pile condition. Based on this study, damping constants and damping coefficients were found to decrease with pile installation depth, sand density and static resistance The static pile bearing capacity was also shown to stand for the combined roles of depth and density. and the roots of Eq.(2) are:
1 INTRODUCTION
Damping of a material is defined as “The mechanism by which the energy of vibration is gradually converted to other form(s) of energy such as sound, heat, noise, etc” The damping type prevailing in pile driving problem is believed to be viscous damping, in which the damping force is proportional t o the velocity of the moving body, and the viscous damper is modeled by a piston-cylinder dashpot The dominating equation of free motion in a viscous media is
3
where, (c ) is the damping constant, (k) is stiffness, defines the damping ratio as (c/ccr), and (a,,)is the natural frequency. The roots (bl, b,) are either real equal, real distinct, or complex, defining the three possible types of free damped motion, namely, Critically damped, Over-damped, and Under-damped motions, respectively.
(c)
1
where: in is the mass of the moving object, c is the viscous darnping constant of the surrounding media, k is the stiffness, y is the displacement, and its first and second time derivatives are velocity and acceleration (v, a) respectively. The solution of the above equation is in the form:
when substituted in Eq.(l) yields the following characteristic equation: mb2+cb+k=0
2 ANALYTICAL PROCESURE
Described below are the analytical procedures for under-damped and critically damped pile motions, over-damped motion is far beyond our concern Over-damped and even critically damped motions are so impractical in pile engineering problems, and the appearance of motions near the critical can only be attributed to the relatively very small pile bearing capacities and the reduced test pile scale It is known, however, that critically damped motion is quite similar to the over-damped with the only
2
233
difference in damping period, that the over-damped motion endures more than the critically damped as shown in Figure (1).
2.1 Under-damped Motion: For such cases two methods can be used to calculate the damping ratio, the first utilizes Hilbert transform techniques either derived by Thrane (1 984), Mohammad and Tomlinson (1988), or by Husein Malkawi and Mohammad (1996), while the second is the logarithmic decremental approach. The analytical procedure for the signals in this research will be based on the logarithmic decrement due to its simplicity that it requires little calculations. The logarithmic decrement (6) is defined as:
Y ) 6= In( 1
and then plotted over the experimental signal to check for the matching between the derived and experimental time-displacement signal. As the value of damping constant, c is being verified, and with the static pile capacity, R, in hands, the viscous damping- Smith damping, coefficient (J) can be computed as: 8
2.2 Csitically Damped Motion
For the critical damping condition the solution algorithm is iterative; given the initial conditions (yo, v,), the time displacement hnction becomes: 9
4 The natural frequency (a,) is adjusted in eq.(9) until it fits best with the experimental signal, then for the selected fitting equation:
Y2
hence, the damping ratio is calculated as.
6 5 where;yl and y2 are the amplitudes of the first and second peaks of the oscillation, respectively. Measuring the damping period- from the recorded signal, between the first two successive peaks (zd), then: 6 now the damping constant (C) is computed as:
c=cc
Cl'
= 217143
7
3 TEST MODEL AND MATERIAL
3.1 Model Pile
The test pile is a steel, close ended, smooth pipe pile with a geometry chosen so that its slenderness is comparable to real pipe piles. The test pile is (90-cm) long, having lengtwdiameter (L/D) ratio of (90/6.1 = 14.75), the (5 mm) thick pile-wall is capable of sustaining piles section undeformed even after pile driving. The test pile was also provided with a steel pile cap, the cap has a semi-spherical notch on its upper face where a steel ball (ball bearing) is placed to insure pure axial loads, that it prevents bending moments from occurring. 3 2 SU??dPrcpr~tres
Figure 1 Comparison between critically damped and over-damped motions. The values computed above along with the initial conditions (yo, v,) are now used to obtain the parameters of the under-damped equation of motion,
The sand used in this study is fine to medium, poorly graded sand (SP) named (Sweileh Sand), having the gradation curve shown in Figure (2) This type of sand has been chosen because it is easy to control especially when it is a matter of density due to the shape of particles (uniformity) Using the sand bed preparation technique described later, the dry density of the sand in the loosest test state is (1 55 g/cc) and the angle of internal friction varied in the range (26 - 30") depending on the level of the effective stress Whereas, in the densest, the dry density of the sand is (1 72 g/cc) and the corresponding friction angle Those friction angles laid in the range (35-40")
234
to a maximum of (100-cm) because the resulting dust may seriously affect the instrumentation of the laboratory. Consequently, for the case of (70 %) relative density, gentle vibrations are necessary to achieve such high density.
were obtained from direct shear test conducted on loose and dense sand specimens at different representative normal stress levels
Table 1 Selected Relative Densities and the Corresponding Heights of Fall Initial Sand D, P tl Falling (%) (g/Cm3) Height (cm) Cond. Loose 35 1.55 30 Medium
50
1.63
80
Dense
70
1.75
100”
* Vibration employed. Figure 2 Grain size distribution curve
4 TESTPROGRAM
3.3 Sand Container
A general view for the test elements, instrumentation and setup is shown in Figure (3). The first step in all pile’s tests is the preparation of sand bed at the desired test density (or relative density). The sand is filled in the silo which is then elevated to an elevation corresponding to the desired test relative density obtained from Table ( l ) , and as a double check, tank full weight is measured by the dartec machine, and knowing the depth of sand in the container the density is insured.
The sand container is a square (100 x 100 cm2) sectioned, (125 cm) high steel box, with (5mm) thick wall, and an empty weight of 2.73 kN (278.37 kg). Those dimensions were chosen in such away to prevent the interaction of pile driving stresses with the container’s either wall. This container was manufactured in the engineering workshops of Jordan University of Science and Technology. To obtain precise measurements of the in tank sand depths the tank was marked at corners and walls mid-lengths with a metric graduation to the nearest (0.5 cm) 3.4 S m d Bed Preparation Techiiiqire
The model pile is to be installed in sand number of times at combinations of installation depths and sand initial densities Prior to every test a mechanical technique named “Sprinkling Method” will be employed to get the required sand density. In this method sand is allowed to fall freely from a silo elevated over the sand container, the desired sand density will be function of the silo’s elevation For every elevation, the corresponding density is measured then the relationship between the falling height and sand density or relative density is determined Table (1) summarizes the selected test relative densities, and corresponding densities for different silo elevations. The density is directly calculated when final sand height in the sand container is measured, and container’s empty and full weights obtained from the DARTEC machine- when operating in the tension mode Silo elevation (height of fall) has been restricted
Figure 3 Test elements and experimental setup First, test pile is driven by a square load wave of (7.5 kN) amplitude and a frequency of (20 Hz), see Figure (4) The dynamic impact is applied to the pile via DARTEC machine ram which is now set to operate on the dynamic- strike mode, ram stroke (displacement) and ram frequency are the only inputs for those kinds of tests, chosen to be (8 mm, 20 Hz) respectively.
235
along with a semi-iterative parametric optimization process The optimum parameters are fhrther judged and checked through their match with the experimental signals “Signal Matching” For the purpose of evaluating damping coefficients it was necessary to conduct static pile load tests at the same test conditions Constant Rate of Penetration (CRP) tests were conducted at a rate of (0 03 mm/s) according to the ASTM (D1143-81) Ultimate pile static resistances at combinations of pile installation depths and sand initial density are shown in Table (2) Damping constants and damping coefficients are then studied with different test parameters, pile installation depth, sand’s initial density, weight of soil column at the level of pile’s toe, and ultimate static soil resistance The under-damped motion is clear since it is characterized by the oscillation (wave like motion), while critically and over damped motions can only differentiated by their damping period, as mentioned earlier Below are detailed examples for signal matching processes along with parametric evaluation for two damping cases, under-damped, and critically damped or near Each of which is analyzed separately
Figure 4 Pile driving input, square load-time wave. To measure pile’s displacement with time a Linear Variable Differential Transducer (LVDT) was placed under the pile cap thus allowed to simulate pile’s actual movement, the (LVDT) is connected to an analogue storage oscilloscope from which the final Time-Displacement signal was obtained. The (LVDT) is one primary, two secondary coils and a magnet, the motion of the magnet induces an electrical current, this current is transformed into displacement by the (LVDT) control panel, and the signal is shown on the screen of the oscilloscope The pile is forced to move for a definite forced displacement depending on the strike of the machine ram, then continues to move freely for a little displacement upon which analytic works are later conducted For the case in hands, forced displacement can be easily measured as that corresponding to constant Time-Displacement (t-d) slope, i.e , for the initial linear part of the (t-d) signal where the slope is a constant value Even when running the machine in the dynamic mode, the ram will move down as fast as specified by the frequency input until it reaches a maximum forced displacement at a high constant jacking rate, then rebounds up to allow free damped motion of the pile The same sand preparation technique, and test setup were utilized in the static pile load tests An XY adoptei- (Pen Recorder) was employed instead of the LVDT to indicate pile loads and displacements plotted on a millimeteric paper. Pile capacity is taken as that corresponding to a displacement of 10% of pile’s diameter As for the expected experimental errors, they may be comparatively insignificant The machine is capable of measuring within 0 5 N (0 05 kg), sand level in the container is precisely achieved by applying gentle vibrations to the container’s walls, and adjusted via a levelmeter allowing the achievement of the desired sand depth with a tolerance of few millimeters, and the oscilloscope expected errors are probably the least. 5 DISCUSSION Measured time-displacement signals are to be analyzed utilizing the “Logarithmic Decrement”,
Table 2 Static pile bearing capacity in (kN) as obtained from CRP tests. 30 45 60 Depth D,% 35 1.5 2.4 3.25 so 2.2 4.5 56 4.3 6.1 8.7 70 5 1 I I I i ~ e r . - ~ ~ ~ ~ i y e u (Refer ~ ( j t i otoi iFigure 5 ) The first step is to plot the initial tangent for the curve starting from (0 0,O 0), this tangent match with the initial linear portion that corresponds to constant pile velocity, thus indicating the region of forced motion The last point on the linear initial part of the curve, is the first in the free movement region The axes is now transformed to the point (t‘=O 0, x‘=O 0) in the free motion part This point corresponds to (t=75, x=6 4) The initial linear part of curve is extracted form Figure ( S ) , the remaining part represents pile’s free motion and is re-plotted on the new axes ordinates (t’, x‘) as shown in Figure (6) below The solid line represents the recorded transformed signal, while the dashed is the deduced Based on the Figures ( 5 and 6), yo=O 3 min , v,,=89 inm/sR,= 8 7 kN yl=O S inin, y2=0 22 mm, at z,1=36 ms The logarithmic decrement, 6 = In yl/y2= 0 82 1 Damping ratio, = 0 1296 Damped frequency, atl = 0 174 rad /ms = 0 175 rad h i s Natural fi-equency, a,,
<
236
To check for the obtained parameters, the deduced pile displacement fhnction is plotted against the actual (experimental) signal to check for their match. The under-damped equation of motion is:
Similarly, Figures (7, through 10) show underdamped signals that were analyzed using the logarithmic decrements and the equations described above.
y ( t )= 0.63 e- 0.02268r cos(0.174t- 61.4) Now that the equation of free pile damped motion is verified and checked, the damping constant for the system is computed as:
c= 2m
Figure 6 Free under-damped pile motion at 60-crn pile installation depth in sand prepared at 70% D,.
Figure 10 Time-Displacement signal for pile installed at 60-cm, in sand prepared at 35 % relative density.
237
5.2 Critically DanipedMofron (Refer to Figure 11)
Figures (13, through 15) are signals recorded and suspected to be at Or near the condition, and were all analyzed according by similar approach.
The first step is the transformation of axes to the first non-linear point, which corresponds to the beginning of the free motion The origin of the transformed axes has the ordinates ( t= 55, x= 4 3) as shown in Figure ( I 2) Initial conditions for the signal are y,, = 0 2 mm, v, = 76 m d s , R,=4 5 kN The equation of motion for the critical damping condition is y(f) = ( y , + (v, + w,,y,)f)e-' =
w,lf
Figure 13 Time-Displacement signal for pile installed at 45-cm, in sand prepared at 35 % relative density.
0.2 + (0.076 + 0 . 2 ~ , , ) t ) e - " ~ ~ '
Now, the natural frequency which is the only unknown that has to be varied until the above equation fits best the experimental signal. The optimal frequency for this case is ( a = 0.08 rad./ms. = 80 rad/s), then: =1280 N s/m = 1 28 kN s/m c = 21n a,, and k = 51 2 kN.s/m and, J = 0 284 s/m where (R,) is obtained froin constant rate of penetration static pile Tests
Figure 14 Time-Displacement signal for pile installed at 30-cm, in sand prepared at 50 % relative density
Figure 15 Time-Displacement signal for pile installed at 30-cm, in sand prepared at 35 % relative density
Figure I 1 Time-Displacement signal for pile installed at 45-cm, in sand prepared at 50 % relative density
It is noted, however, that increasing the depth from 30 cm to 45 cm in 35% and 50 % relative densities (% Dr), have almost produced no effect on the mode of motion and even on the model parameters, that for those four combinations the same damping constant, and stiffness were obtained Table (3) summarizes the parameters obtained by the analytical approach introduced herein The variations of stiffness with depth and density are not noticeable, specially for critically damped motion (Figures 1 1 , and 13) One possible justification is that the stiffness is very sensitive to the precision by which the natural frequency is computed, and treating those two signals as critically damped endures some deviation from the real frequency values, the error encountered in the
Figure 12 Free criticallydamped pile motion at 45cm Pile in sand Prepared at 50% Dr
238
frequency produce larger deviation in the value of stiffness from the real because frequency is raised to the power of two. The influence of the precision of calculated frequency is less significant on the errors of calculated damping constant. Table 3 Results of tile load tests depth k C
45 60 30 45 60 30 45 60
180.0 2.40 115.2 0.512 (1.63 g/cc) 5O%D, 53.8 1.296 51.2 1.280 121.0 0.480 (1.76 g/cc) 70%D, 123.0 0.464 133.1 0.432 0.363 295.0
J
seems to conclude the roles of pile installation depth and sand density.
I
0.967 0.1575 0.64 0.284 0.0857
Figure 17 Pile’s total damping constant Vs. pile installation depth for different test sand’s initial densities.
0.108 0.0708 0.042
Figures (16, through 19) may further help in the study of those parameters with available test variables. According to Figure (1 6), increasing pile installation depth non-linearly reduces damping constant for all test densities, and Figure (1 7) shows that the role of pile installation depth and soil density on damping coefficient, and damping constant can not be considered through soil stresses (soil column weight) at the level of pile’s toe, each one of which has to be considered separately, or look for an alternative measure. This may be attributed to the fact that, both-density and depth, affect many variables other than the level of stress, such as soil shear/compressive strengths, stiffness, and shear modulus, so treating variables separately may make the problem easier to handle. Static pile bearing capacity is another important factor in correlating to damping constants and coefficients. Static pile resistance is obtained from constant rate of penetration tests (as indicated earlier), as shown in Figure (1 S), damping constant is reduced by increasing the static capacity of the pile. Yet, and as clear from this figure, the amount of reduction is larger for looser sands than for the denser. As for the relationship between pile bearing capacity and damping coefficient, it can be seen that at any distinct sand density the relation is perfectly linear, while for the whole set of data the damping coefficient is a power function (ax”) of bearing capacity, J = 4/R? Similar to damping coefficient, the damping constant is a power hnction of Static pile bearing capacity as shown in Figure (19), staic pile capacity
Figure 18 Influence of soil overburden pressure on darnping coefficient, for all test densities.
Figure 19 Pile’s total damping coefficient Vs. static pile capacity for different test sand’s initial densities. 6 CONCLUSIONS
On the basis of the experimental work, the signal process introduced in this study as well as the correlative study, the following conclusions are made: The use of logarithmic decrement in underdamped motion and the semi-iterative signal
239
matching in critically damped motions are verified to be easy and powerfd in evaluating some of the soilpile model parameters such as; damping constant, darnping ratio, and stiffness. Both, damping coefficient and damping constant decrease with pile depth of installation, sand’s density, and static pile bearing capacity. Static pile bearing capacity concludes pile installation depth and sand’s initial density, while soil pressure at the pile’s toe level does not. The predictability of stiffness is dominated by the accuracy of predicting the natural frequency. ACKNOWLEDEMENT This research is part of the second authors MS thesis supervised by the first author and it was supported by the Dean of Research at Jordan University of Science and Technology. REFERENCES ASTM D 1 143-8 1, (1998) “Standard Method of Testing Piles Under Static Axial Compressive Loads”, Vol. 04.08, Philadelphia, pp. 179-1 89. Clough, R. W., and Penzien, J., Dynamics of Structures, McGraw-Hill, Inc., New York, N. Y., 1975. Husein Malkawi A. I. & Mohammad, K. S., (1996) “Estimating Damping Constant of the Pile-Soil System Directly From Measured Displacement Using Hilbert Transform Technique”, Fifth International Conference on the Application of the Stress-Wave Theory to Piles, Orlando Florida, September, pp. 37-54. Litkouhi, S. & Poskitts, T. J., (1980) “Darnping Constants for Pile Driveability Calculations”, Geotechnique, 30(1), pp.77-86. Mohammad, K. S. & Tomlinson, G. R., (1988) “A Simple Method of Accurately Determining the Apparent Damping of Non-Linear Structures”, Proceedings of the 6th International Modal Analysis Conference, Florida. Rao, S. S., (1995) “Mechanical Vibrations”, Addison-Wesley Publishing Company, New York, N. Y., Third Edition, 1995., pp. 97-140. Thrane, N., (1984) “The Hilbert Transform”, Technical Review No.3, Bruel and Kjaer.
240
Applicatjon of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Load transfer analysis from increasing energy dynamic load tests in concrete piles driven in very soft clay formation J. Balech & N.Aoki University of Siio Paulo, Siio Carlos, Brazil
ABSTRACT: This paper presents the results of dynamic loading tests performed on concrete piles driven in very soft clay formation. The load transfer mechanism under blows of growing energy is described in terms of mobilized resistances (friction and point) and displacements at the top and at the base of the piles. A comparison between measured static and dynamic load - displacement curves shows that they are in good agreement. correlation with the resistance-displacement curve obtained in the dynamic test. The measured load transfer for all the piles are also presented (Balech,
1 INTRODUCTION 1.1 Constant energy dynamic loading test The dynamic loading test is usually performed with constant energy blows and each blow is analyzed by CASE and CAPWAP methods when the monitoring is done with PDA system (Gobble et al., 1996). The following results are obtained: a) total, friction and point resistances; b) load distribution along depth; c) friction and point quakes; d) maximum displacements, velocities, acceleration at various sections; e) maximum force and energy.
2000).
2 SOIL DESCRIPTION The soil profile for concrete hollow piles at sites A and B are presented in tables 1 and 2. Table ........1......Soil ....,. ...profile ............. at ~....site ................A ..,,........~..........."".', .................. .,... ........................ Soil characteristics i Thickness j N
.....A........
I
.........................................................................
1.2 Gradual increasing energy dynamic test The dynamic loading test used in Brazil is related to the gradual increasing energy blows procedure (Aoki 1989, Bernardes 1989). Each blow is analyzed according to with the constant energy dynamic loading test. From this analysis it is possible to obtain the variation of the many variables described in section 1.1, under growing energy levels. The rupture load can be characterized when the mobilized resistance remains constant after two blows of increasing energy (Aoki, 1997, Aoki & Cintra, 1997)
Soft, dark brown fine to medium 1.80 i 2/30 silty sand with roots ................................................................................................................................... : 14.20 17/746 Very soft, gray, organic micaceous silty clay ................................................................................................................................... Dense, gray, clayey fine and me- ! 2.20 10/30 dium sand with gravels ................................................................................................................................... Very stiff, mottled, micaceous: 0.20 j 2 ..c l a ~ e ~ .~s l~~l n, .a ~ s..es.dual.soll s~c .... ~................................i....... .. Table 2. Soil profile at site B .................................................................................................................................... Soil characteristics i Thickness
:
N
i ............Im>.............;..Blows!cm.. Sofi, dark brown fine to medium I 1.00 1 1/64
silty clay with many roots .................................................................................................................................. Very soft, gray, organic
i
Medium, gray, sandy medium !
clay with gravels ........................................................................
This paper presents the analysis of typical results of gradual increasing energy dynamic loading test performed on six hollow precast reinforced concrete piles, driven into gnaissic residual soil underlain by a thick soft clay formation. All these piles are driven with free fall hammers and the dynamic loading tests were monitored and analyzed by using the PDA system methodology. Static SML loading test and the load-displacement curve for site B pile show a good
:
..
..........................................................................
micaceous silty clay .........................................................................
1.3 Concretepiles in very soft clayformation
i.. ........... I?>........... .:. . B!OW!!C.~.
16.36
............................
0.64
:.................................
12/834
:........................
1
5/15
> .........................
Water level depth = 0.0 m (sites A and B) *impenetrable soil
3 PILE LOAD TRANSFER DIAGRAMS 3.1 Piles andpile driving system
The pile driving system is presented in table 3:
241
Pile driving system
Pile type
Pile type
Hammer type Hammer weight (kN) Pile cap weight (kN) Cushion thick. -hard wood (m) Cushion thick. - plywood (m)
Free fall 50 3 0.30 0.06
Free fall 34.2 2 0.30 0.06
_A_
R
I
The main characteristics of the hollow reinforced concrete piles are presented in tables 4 and 5 . Table 4. Main pile characteristics
Concrete fck Steel fvk Pile A1
Diameter (cm> 50
500
Driven length (m) 19.3
500
Time after driving (days) 20
3.2 Mobilized normalforces diagrams
When the hammer hits the pile the soil resistance is mobilized and the maximum skin friction and point resistance vary in accordance with the nature and rheological characteristics of different soil layers found along the depth. The normal force diagram resulting from this impact, obtained from CAPWAP analyses performed for each blow of increasing energy applied to the piles, is presented in figures 1 to 6. Except for pile A5, the normal force diagram for all the piles is approximately constant along the soft clay layer thickness, due to the low
Figure 4. Mobilized normal forces diagram: pile A4
Figure 1. Mobilized normal forces diagram: pile A1
242
of the measured displacement at the top of the pile, for the same energies of figures 1 to 6. In the case of piles A l , A3, A4 and A5 (Figures 7, 9, 10 and 1 l), it was observed that: a) in order to mobilize the maximum unitary local friction in the soft clay layer a maximum displacement of about 10 mm, is necessary; b) in the sandy layer, the maximum local friction is mobilized for displacement of about 20 mm. For the piles A2 and A3 (Figures 8 and 9), the same trend is observed but, in the last blows, the mobilized skin friction diminishes and becomes equal to zero in the case of pile A2. This fact can explained by the presence of a very soft clay (muddy soil) in the upper layer.
Figure 6. Mobilized normal forces diagram: pile B 1
mobilized soil resistance along the pile shaft. Although being in the same site A, the different behavior of pile A5 can be explained by a local variation of the soil profile. In the case of piles Al, A2,A3 e A4 (Figures 1,2,3 and 4) the measured normal force diagram can be explained by the transition between the soft clay and the residual sandy soil layer. A sudden loss of friction resistance can also be observed in the case of last blow applied in pile A2,where all the resistance was transferred to the point. For pile B1 embedded in very soft clay (Figure 6) the friction resistance was negligible and only point resistance was mobilized for all the blows. 3.3 Mobilized unitary localfriction resistances
Figures 7 to 12 present the mobilized local friction resistances, measured at various depths below ground (DBG) indicated in the figures, as a hnction
243
Figure 8.Unitary local friction diagram: pile A2
Figure 12. Unitary local friction diagram: pile B 1
3.4 Local friction resistance aiong depth The measured unitary local friction resistances at each depth below ground (DBG), for the indicated growing energy blows, are presented in figures 13 to 18. These diagrams show how the local skin friction is mobilized along depth. As a general trend, at each applied energy, the 10cal friction varies according to the nature and resistance of the soil layer. It can be observed that, in the case of piles A5 and B 1, the unitary local friction resistance decreases with depth instead of increasing as in the case of piles AI, A2,A3 and A4. In the case of pile B1 the maximum skin friction resistance is as low as 8 kPa corresponding to a very soft clay with an SPT resistance equal to zero (actually 50 to 80 centimeters of sampler penetration, under the static action of SPT hammer + rod weight).
Figure 13. Local friction diagram along depth: pile AI Figure 11. Unitary local friction diagram: pile A5
244
Figure 18. Local friction diagram along depth: pile B1
nent (set) displacements of the point of the piles, for growing energy level. It can be observed that the elastic displacement (quake of pile point) is dependent on the applied kinetic energy level. The shape of the total point displacement curve suggests that the point of piles A1 and A4 has apparently reached the layer which is impenetrable to percussion. 3.6 Point percentage on the total resistance
Figure 16. Local friction diagram along depth: pile A4
It is emphasized that the CAPWAP analysis for each blow of increasing energy was done without any attempt to adjust the friction resistance distribution t o the known soil conditions. 3.5 Point resistance x point displacement and set
Figures 19 to 24 present the relation between the mobilized point resistance and the total and perma245
The percentage of point resistance on the total mobilized resistance, for each blow of growing energy, is shown in figures 25 to 30. Piles AI, A3 and A4 present the usual behavior where the lateral friction is mobilized before the point resistance for small system deformations and, then, the pile resistance is gradually transferred to the base (Kezdi, 1975). Pile A2 shows the same usual behavior until the energy level correspondent with the last blow. In this event the residual lateral friction is reduced to zero and all the resistance is represented by the point. Pile A5 shows that the lateral friction is much
greater than the point resistance in all sequence of growing applied energy. Finally in pile B1 for all blows of growing energy, the point resistance predominates.
The so-called quake of the point (elastic displacement resulting from the elastic deformations of the soil under the base of the pile) is directly dependent on the amount of energy reaching the soil under the pile point. It can not be considered a soil parameter. The ultimate load resulting from the static load is in good agreement with the ultimate load obtained in the gradually increasing energy dynamic load test. The load transfer diagrams obtained in dynamic load test show a good resemblance to those obtained in instrumented static load tests. In the soft clay layer the lateral skin friction grows up to a maximum and thereafter diminishes to a residual value when the pile is gradually subjected to increasing applied energy.
4 CORRELATION BETWEEN STATIC AND DYNAMIC PILE LOADING RESPONSE All the equivalent static load transfer diagrams obtained in this site, from dynamic loading test performed with gradual increasing energy, show a general trend that is in good resemblance to those obtained with instrumented piles subjected to static load test. The reliability on the application of a dynamic loading test procedure to obtain information about the pile behavior under static loading has grown with the increasing utilization of this type of test all around the world. Otherwise the correlation between static and dynamic load test results in the same pile can be made and, in doing so, it would be possible to improve our understanding of when different soil conditions occur. Such correlation was done in this site and the results are presented in figure 3 1 .
6 REFERENCES Aoki, N. 1989. A new dynamic load test concept. In Japanese Society of Soil Mechanics and Foundation Engineering (eds): International conference on Soil nzechanics and foundation engineering, 12. Drivability of piles, 2: 1-4. Rio de Janeiro, Brazil Aoki, N. 1997. The evaluation of the ultimate bearing capacity of driven piles by using increasing energy dynamic load tests (in Portuguese). Doctorate thesis. Sa"o Carlos School of Engineering, University of Sa"o Paulo, Brazil. Aoki N. & J.C.A. Cintra 1997. New interpretations of the dynamic loading curves for driven piles based on the energy approach. In Almeida (eds), Recent developnzents in soil and pavement nzechanics: 467-472. Rotterdam: Balkema. Balech, J. 2000. Analysis of dynamic loading tests performed in concrete piles driven in soft clay formation. (to be published in Portuguese). Master of Science dissertation. Siio Carlos School of Engineering, University of Sa"o Paulo, Brazil. Bernardes, G.P. 1989. Dynamic and static testing of large models piles in sand. Doctorate thesis. Departnzent of Civil Engineering. Norwegian h t i t u t e of Technologv, Trondheim, Norway. Gobble, G.G. & Likins Jr. 1996. On the application of PDA dynamic pile testing. In F.C. Townsend (eds): International Conference on the Application of Stress- Wave Theory to Piles, 5: 263-272. Florida. Kezdi, A. 1975. Pile foundations. In H.F. Winterkorn & H. Fang (eds), Foundation Engineering Handbook: 556-600. New York: Van Nostrand. Van der Veen, C. 1953. The bearing capacity of a pile. International Conference on Soil Mechanics and Foundation Engineering, 3. v. 2: 84-90. Switzerland.
Figure 3 1. Resistance versus displacements curves: pile B 1
The dynamic test was performed in 1 day and the static test was carried out three weeks after the dynamic one. The ultimate load of the pile is greater than 1000 kN. The extrapolation done by the usual method (Van Der Veen, 1953) indicated an ultimate load of about 1050 kN in both cases. 5 CONCLUSIONS The constant energy dynamic load test is inadequate to measure the behavior of a driven pile as the only piece of information obtained from this test is a point in the resistance - displacement curve. 248
Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Dynamic pile testing and finite element calculations for the bearing capacity of a quay wall foundation - Container terminal Altenwerder, Port of Hamburg E Kirsch, B. Plapmann, T. Huch & W. Rodatz Institute for Foundation Engineering and Soil Mechanics, Technical University of Braunschweig, Germany
ABSTRACT: In April 1999 the construction of a new quay wall in the port of Hamburg started. In a first stage two new berths with an overall length of 1400 m will be built. In order to proof the bearing capacity of the foundation dynamic pile testing is performed to a great extent. To date more than thirty different piles were tested by the Institute for Foundation Engineering and Soil Mechanics of the Technical University of Braunschweig. In order to check for possible set-up effects redriven tests were performed. For comparison reasons static load tests were also carried out, one of which was done with an instrumented pile to allow skin friction and end bearing to be evaluated separately. The CAPWAP results proofed to be very helpfid especially in those cases where the testing situation differed from the later service condition. Special questions called for detailed finite element analysis of the load-settlement behaviour of single piles in certain construction stages. 1.2 Sire conditions
1 CONTAINER TERMlNAL ALTENWERDER 1.1 The project The port of Hamburg is one of the biggest and most important amongst the harbours in Europe. Its geographical and infrastructural position makes it very interesting as a gate from the North Sea to Europe. This stands especially for the east of the continent since the Baltic Sea cannot be shipped with the big container ships of the latest generation. Since 1980 the total freight handling in Hamburg has grown from 63,l million tons to 75,8 million tons in 1998. In the same time the share of contained freight has grown from 1 1 % to 48 % (Freie und Hansestadt Hamburg 1999). This underlines the importance to provide of sufficient space for containers and container ship berths As early as 1973 first planning started to realise the extension of the port of Hamburg in the area of Altenwerder, which is situated south of the well known Kohlbrand-Bridge, one of the landmarks of Hamburg. It did cost a lot of political endeavor until in April 1999 the construction of the quay wall itself could commence. In a first phase two new berths with a length of approximatetely 800 m will be built until the year 200 1. Finally a total length of 1400 m will give four modern container ships the opportunity t o unload their goods.
249
The construction of the quay wall is done from the dry. The 24,5 m draught necessary for modern container ships will be achieved only after major dredging works (Fig. 1). Therefore it is necessary to install all load bearing elements of the quay wall by an extensive pile driving operation with depths to approximately 30 m.
Figure 1. Cross section (Freie und Hansestadt Hamburg 1999).
the site investigation programme and as chosen for design purposes are given in Table 1. Approximately half way along the quay wall axis the depth of the boulder clay reduces and the structural elements of the foundation are embedded into it. This was not addressed as a problem since the soil investigation gave reason to expect sufficient bearing capacity. Figure 3 gives the geological situation along the axis of the quay wall. Table 1. Soil parameters Twc ’{ y’ kN/mi kN/in3 Landfill 19 11 Sand 18 10 Clay 17 7 Moulmeat 14 4 Gravel 19 10 Boulder clay 22 12
9‘ O
27.5 32.5 25.0 17.5 35.0 30.0
C’ Es kN/ni’ MN/iii’ 0 8.0 0 40.0 10 2.0 10 1.5 0 100.0 20 25.0
Figure 3 . Geological situation (Miller 1999)
1.3 Construction details Figure 2. View on the construction site (Freie und Hansestadt Hamburg 1999).
In the first phase approximately 1300 cast-in-place concrete piles with a total length of 32 km are driven into the ground to support the quay slab and the beam of the gantry crane tracks. In order to speed up the construction of the wall and in particular the front piles these structural elements are placed into 32 m deep trenches and are then driven a firther five meters to final position. Figure 2 gives a view on the construction site in Summer 1999. Subsoil conditions in the north of Germany are characterized by the effects of the last glacial period. Under deposits of marine sand and gravel with layers of silt tertiary boulder clay with different stiffnesses is encountered. Typical soil parameters as revealed by
The quay slab rests on tubular steel piles (0 1 2 1 9 16 ~ mm) and a mixed sheet pile wall with double planks (€32 975 A) and intermediate planks (AZ18-10) as well as on a total of 1300 driven castin-place concrete piles 0 51 cm. Horizontal stability is ensured by 46 m long inclined steel piles (HTM 600/136) driven at 2,27 m intervals. The rear crane rail, at a distance of 35 m from the front rail, is founded on a row of inclined driven cast-in-place concrete piles (Wittwer & Krefi 1999). 1.4 Pile instaiialioiz
The steel pipes and the double planks are placed into a slurry trench filled with a cement-bentonite suspension. This mixture should hardens and achieves approximately the same qualities as the surrounding 250
soil. Afterwards the steel profiles are driven by a 15 tons Menck MHF 10-15 hammer for a hrther five meters into the subsoil to ensure sufficient bearing capacity. For the driving of the anchor piles and the cast-in-place concrete piles hydraulic hammers of appropriate capacity are used.
The tests at the driven steel profiles and tubular steel piles respectively were done at the end of the driving operation and also as a restrike test three weeks after installation. Up to now a total of 23 tests with CAPWAP evaluation were carried out. Figure 6 shows an example.
2 DYNAMIC PILE TESTING 2.1 lest r.eslrlls
Up to now 24 restrike tests were carried out on different concrete piles All test results were obtained by the CAPWAP procedure. Figure 4 gives a graphical overview of the load calculated capacities split into skin friction and end bearing.
Figure 6. CAPWAP result of a tubular steel pile
Figure 4. CAPWAP results of the driven cast-in-place piles
A field test allowed the comparison between the CAPWAP results and the results of a previously performed instrumented static load test, where skin friction and end bearing were measured separately. The instrumentation consisted of a load cell at the end of the pile and stress measurement sections in different depths along the pile axis. Figure 5 shows the load settlement curves calculated by CAPWAP and those measured during the static load test.
The reason for the need to gain information about the skin friction distribution along the pile axis by CAPWAP analysis lies in the specific construction of this quay wall foundation and the regulations of the competent authorities. Each individual pile has to be able to carry its load with an appropriate factor of safety even in the case of a failure of the wall. Due to possible horizontal movement of the wall the soil behind fails by developing shear bands. Thus it may loosen and subsequently reduce its frictional capacity. Therefore all resisting forces acting along the pile above an assumed line of failure must not be taken into account for the determination of the bearing capacity. This assumed line of failure starts at the toe of the sheet pile wall with an inclination of approximately 1.2 and goes in upward direction into the backing soil. It is obviously important to distinguish between skin fiction and end bearing in cases where piles are being excavated and exposed to water later on. 2.2 Special soil condition As shown in Figure 3 parts of the foundation are embedded in the boulder clay which was addressed by the soil investigation as a relatively stiff clay. In contrast the systematic pile tests revealed a significant reduction in the bearing capacity of all structural elements in that area. It became obvious that the clay was far softer than expected. Carefbl adaptation of the piles design became necessary and different
Figure 5. Load-settlement response of a concrete pile.
25 1
structural solutions were investigated by means of dynamic pile testing. In case of the concrete piles their lengths were extended as well as the redrive method was adopted. For the steel piles a series of different pile shoes with wings and filling sheets was tried to cope with the situation (Fig. 7).
Figure 8. Set-up of the driven cast-in-place piles
4 FINITE ELEMENT ANALYSIS 4.1 General remarks
Figure 7. Different pile shoe designs
In this critical phase the dynamic pile testing proofed its usefulness. Fast performance and the relatively low costs of dynamic pile tests made the redesign of the foundation during construction possible. In the end this observational method using the different CAPWAP results produced an optimised foundation.
Numerical simulation in geotechnics today is object of profound investigation. The characteristics of the soil as a material led to special solutions on the side of constitutive laws for geomaterials as well as in the practice of finite element calculations. Numerous publications deal with the numerical simulation of piled foundations and their problems (e.g. ElMossallamy 1999, Maybaum et al. 1999). For the finite element analysis of the subject case the ANSYS program was used. In the last years specific implementations into the program were realised at the Institute for Foundation Engineering and Soil Mechanics of the Technical University of Braunschweig to adjust it to geotechnical problems (e.g. Vittinghoff et al. 1997, Planmann et. a1 1999). 4.2 Niimerical modelling
3 SET-UP EFFECTS
For all types of piles a certain time dependent development of the bearing capacity was expected. As the design especially for piles embedded in the boulder clay turned out to be difficult, more effort was spent in the investigation of possible set-up effects. Since the bentonite slurry hardens with time the increase of the skin friction along the steel profiles is evident. In fact the resisting forces became so high in some cases that it was impossible to gain any vertical displacement during the restrike tests. For the concrete piles a gain in bearing capacity of 10% to 20% within two month after pile installation was encountered (Fig. 8). The reduction in the bearing capacity and its subsequent regain of pile D182 was object of intense studies. In order to proof that the reduction could be explained not by a general loss of strength of the surrounding soil but by the disturbance of the ground adjacent to that special pile by the recent cutting of the slurry trench, numerical simulation of the construction process was carried out.
The geometrical modelling of the problem was guided by the use of specially designed macros, which allow the easy variation of different parameters. Due to the spatial nature of the situation a three-dimensional discretisation was necessary. Using the symmetrical nature of the problem the inclined pile D182 was cut in half and the adjacent slurry trench with the surrounding soil and appropriate boundary conditions was modelled as shown in Figure 9. To allow for the calculation of the primary stress field as well as for the different construction stages double elementation became necessary. To model the soil-pile interaction properly areas with high stress gradients were meshed finer than other zones. The finite element mesh consisted of a total of 1988 brick elements with second order shape functions. The soil behaviour is simulated by an elastoplastic constitutive law. The Drucker-Prager yield criterion governs the plastic deformations. The influence of the choice of the constitutive law was subject of former studies (Maybaum et. a1 1999). Material properties of the slurry and the concrete were chosen according to prior investigation (Heinrich 1998).
252
The simulation of the construction process is done by a step-by-step analysis beginning with the in-situ stresses, then modelling trench and pile installation and finally the stepwise loading of the pile The load was applied in steps of 200 kN up to failure Two different situations were investigated and the findings were compared with the results of two dynamic pile tests. One took place before and one after the cutting of the slurry trench adjacent to pile D 182
the horizontal displacements of the soil towards the slurry filled trench before pile installation It is apparent that the pile is located in the area of influence Figure 1 1 shows the vertical displacement of the soil around the axially loaded pile in the state of failure
Figurc 11 Vcrtical displaccmcnt duc to loadcd pilc
The gained by finite element analyses as well as obtained by the dynamic pile tests before and after the slurry trench installation are given in Figure 12
Figure 9 Finitc clcrncnt mcsh for pilc D182 with adjacent trench
4 3 Results
With the finite element analysis it was possible to proof that the reduction in the bearing capacity was solely caused by the installation of the slurry trench, which occurred in this particular case after the pile driving and is therefore only a local effect in time In general all pile installation took place after the cutting of the trench and therefore no reduction of the bearing capacity had to be expected Figure 10 shows the influence of the open trench o n the surrounding soil The shaded areas represent
Figurc 12 Load-scttlcincnt ciin cs of pilc D 182
5 CONCLUSIONS
The subject project is an example of extensive use of dynamic pile testing Adaptation of the pile geometry due to the change of the soil properties became necessary when pile tests revealed insufficient bearing capacity The quick and economical nature of the dynamic pile testing allows a relatively easy redesign of a piled foundation The results of dynamic pile tests can be evaluated using a three-dimensional finite This the Of local effects due to different construction stages
Figpie 10 Horizontal displacement due to open trench bcforc pile installdtion
253
REFERENCES Freie und Hansestadt Hamburg 1999. Hamburger Hafen Zahlenspiegel. Freie und Hansestadt Hamburg, Wirtschaftsbehorde Stromund Hafenbau 1999. Container Terminal Altenwerder Neubau Kaimauer, I. Bauabschnitt. Miller, C. 1999. Aspekte zur Planung der Kaianlage. Hunsa 10/1999. Wittwer, G. & Kref€t, R. 1999. Bau der Liegeplatze 1 und 2. Hansu 10/1999. El-Mossallamy, Y. 1999. Load settlement behaviour of large diameter bored piles in over-consolidated clay. In G.N. Pande et al. (eds), Numerical hdodels in Geomechanics NUWOG J87I. Rotterdam: Balkema. Maybaum, G., Vittinghoff, T. & Rodatz, W. 1999. On the bearing capacity and the servicability of piled rafts. In G.N. Pande et al. (eds), Numerical Models in Geoniechanics NUMOG L7I. Rotterdam: Balkema. Vittinghoff, T., Rodatz, W. & PlaRmann. B. 1997. Integration of a Numerical Model in an Observation Programme. In Jian-Xin Yuan (ed.), Coniputer Adethods and Advances in Geoniecl~anics.Rotterdam: Balkema. PlaBmann. B.. Kirsch, F.: Vittinghoff, T. & Lohr, M. 1999. Implementierung infiniter Elemente im ANSYS Open System und deren Anwendung bei Halbraumberechnungen in der Geotechnik. In CAD-FEM GmbH (eds). 17. CADFEN users' meeting. Heinrich. 0. 1998. Einfld des Bodenkontinuums a& die Entwicklung von Verschiebungsgrofien bei FE-Analysen. Thesis, unpublished.
254
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
The shaft dynamic response of a pile in clay: Induced pore pressure A.Benamar Laboratoire de Me'canique, Universite' du Havre, France
ABSTRACT: Driving of pile generates the transient shear in soil and can induce excess pore water pressure if the soil is cohesionless. The shaft dynamic response of a single pile-model in clay is examined in a laboratory study. Using the principles of stress waves propagating in a one-dimensional media, stress and velocity data have been analyzed. A pore water pressure sensor is placed in the soil in order to assess the influence of such parameter in the mobilization of shaft resistance during driving. The experimental study concerns both drained and undrained clay. The parametric study show that the most significant effects associated with pore water pressure are the reduction of the maximum skin fiiction allowed at the interface. This reduction being less important than induced excess pore pressure.
1 INTRODUCTION Pile foundations usually are necessary for most of situations requiring construction of buildings on soft ground. If driving is possible and allowed, driven piles mostly are prefened. Due to pile driving in cohesive, low permeable soils high excess pore water pressure can be induced around the pile. The state of soil stress after driving a pile results from the addition of the following states: initial state of stress of the soil before pile driving, stress change during pile driving, stress change after pile driving (reconsolidation). The change of stress in the soil during and after pile driving must be known for the estimation of the shaft capacity of piles being driven in cohesive bearing soils. In practice use estimation of excess pore water pressure in advance represent a cheep help to avoid damages at buildings, at pile driven before, or at slopes in the vicinity of the driving area. This paper deals with the development of excess pore water pressure induced by pile driving in cohesive soil.
test) are performed. Table 1 give the physical and mechanical properties of the material obtained under a pressure of consolidation ( T ~= 500 kPa. The sample is 0.5 m in height and 0.2 m of diameter. The clay is confined in a triaxial cell under a controlled isotropic pressure ranging from 200 kPa to 500 kPa. The hammer blow is obtained by various available falling masses. The impact of such masses on the rod produces a rectangular pulse with various amplitude and duration (Benamar 1992).
2. LABORATORY DRIVING TESTS In order to study the mobilization of skin friction along driven pile in laboratory, a steel rod, 9 m long, is driven through a clay sample reconstituted around the pile model (figure 1). The clay samples are made with kaolin and laboratory tests (triaxial and vane
Figure 1: Experimental pile driving set-up 255
1
During driving tests the pile-soil interface can be or not drained at the top and bottom sample. The drainage allows seepage of water induced by excess pore pressure along the steel rod.
I
3. EFFECT OF PILE DRIVING IN CLAY
Table1:Soil characteristics (oC= 500 Wa)
I Plasticity limit
I
23
Plasticity index
0.70
Compression index
I Water content
I Shear strength
31 %
I
39 % 65 W a
3.1. Displacement of soil due to pile driving Driving pile through a soil layer is always combined with deformation of the soil adjacent to the pile. In non-cohesive soils voids decrease due to compaction effects of driving, while the volume of cohesive soils remains constant due to lower permeability and the dynamic nature of solicitation. Near the pile partially high excess pore water pressure arises. It decreases after the end of driving process due to reconsolidation of soil. According to Steenfelt et al. (1981), during pile installation significant horizontal movement of soil occurs even in a distance up to ten pile radius f?om the pile axis. At this distance movements of the magnitude of 0.05 times the pile radius were measured. In the experiments reported the pile displacement is measured but not soil displacement. Results obtained for pile displacement are discussed in further section.
Before each driving tests series ten hammer blows are performed in order to avoid the "healing" effect in the pile-soil interface and making the soil under current driving conditions. The piling is operated at a steady blow rate of one blow per minute approximately. Several clay samples are tested in various conditions. At the end tests, water content and undrained shear strength are measured using for the second parameter the laboratory vane test. The evolution of this characteristic with water content is illustrated on figure 2. This evolution is quite linear for the homogeneous reconstituted clay: the undrained shear strength decreases linearly when the water content increases.
3.2.Excesspore water pressure due to pile driving
Figure 2: Soil shear strength versus vater content. A pore water pressure sensor is placed into the soil sample against the steel rod in order to measure excess pore water pressure at the interface during driving. But the sensor is moved during the consolidation about 2 centimeters from the interface. Nevertheless, the measurements of excess pore water pressure in the soil near the pile interface can give interesting information concerning the evolution of effective lateral stress during driving. Strain gages are glued on the steel rod and stress waves are measured in order to derive pile shaft resistance.
Pile driving generates transient loading in the soil. When the rate of loading exceeds the rate of porewater pressure dissipation (controlled by the permeability), the dilative or contractive tendency of a soil does not manifest, and shear takes place at constant g lobal vo lume. The vo lume-confinement-shear domain of soil behavior prescribes that shear at constant volume can only occur if the effective confinement changes, therefore, there will be porepressure changes. In non cohesive, high permeable soils pile driving induces void reduction since the drainage is easy. On the other hand, in cohesive, low permeable soils the deformation during pile driving occurs at constant void and then develop excess pore pressure, leading to a reduction of effective shear resistance of the soil. Due to the fact that pore pressure measurements are difficult and expensive different methods to estimate excess pore water pressure due to pile driving were given by several authors (Vesic 1972, Randolph et al. 1979), using different stress-strain soil theories, and assuming the pile driving being modeled as the expansion of a cylindrical cavity. Few authors suppose that the excess pore pressure is constant within the plastic zone around the pile. But local measurements showed a relatively wide difference between measured and calculated values. The method according to Randolph et al. (1979) was
256
modified by Cunze (1989) in consideration the soil properties, stress state, changes of stress and of geometrical conditions. The excess pore water pressure is expressed as a general function of the initial stress state, geometrical conditions and soil conditions. He proposed the following equation for prediction of excess pore water pressure within the soil plastic zone around the pile, due to the volume expansion caused by introducing the pile volume into the soil: = 1.l(p’nc- p
* c s ) -~ P p ’ +1. rp1
-ro
2 c u E
(1)
raj<
Where G/C, is the rigidity index. The change of effective stress due to pile driving can be estimated (Cunze 1989) using the following equation: p ’ n c - p ’ c s zz p’nc(O.83-0.51L.L.) (3) With: 1
1
s
2K0) (4) 3 c ~ = overburden ’ ~ ~ load L.L.= liquid limit IQ, = coefficient of earth pressure at rest. The excess pore water pressure expected due to pile driving is determined by equation (1) using four driving tests as follows: soil: normally consolidated clay liquid limit LL = 0.54 plastic limit PL = 0.31 water content W=35% coefficient of earth pressure at rest KO= 0.60 undrained shear strength C, = 57 W a rigidity index G/C, = 47 confining pressure P = 300kPa ro = 0.01 m Pile: radius inspection point (distance) p = 0.03 m nc
Table 2. Estimated excess pore pressure for 4 tests
v I (m/s)
With: P ’ , ~= mean effective stress of normally consolidated soil. p’cs = mean effective stress of soil reaching the critical state. C, = undrained shear strength rpl= radius of the plastic zone r = radius of the pile p = distance to the pile axis. Vesic (1 972) expressed the radius of the plastic zone as: $I = G
p
tip resistance, the undrained Young’s modulus is replaced in the formula proposed by Cunze (1989) by the shear modulus G. The overburden load being derived as the difference between the dynamic load stress (product of impedance pile Z = 12.7 kN.s/m and pile velocity) and the confining pressure.
= -0-YO(1+
From equation (2) we obtain rP1 = 0.07 and equations (1) is used .to calculate excess pore pressure values represented on table 2 below, where v is the pile velocity. Because in the experiments reported here the pile is sheared through the soil without any
0.99 1.21 1.57 2.21
G’VO
(Wa) 105.6 195.7 343.2 605.4
I
P’nc (kPa) 77.4 142.8 250.5 441.9
I P’nc-P’cs I (@a)
42.6 78.6 137.8 243.1
Au (kPa) 135.7 161.8 205.1 281.9
1
The excess pore pressure measured during these four tests does not exceed 66 kPa. This large difference between measured and estimated values is due to the fact that Cunze equation is derived for both shaft and pile tip interaction and mainly based on expansion of cylindrical cavity. These results show in one hand that the equation proposed (1989) is more appropriate to estimate the excess pore pressure at the immediate vicinity of pile tip, and in an other hand that the excess pore pressure developed around pile shaft during driving is very low in comparison with pile tip. The presented theory of pore pressure prediction does not integrate all parameters affect the excess pore pressure, particularly the water content and the steady state water pressure in the soil before piling. Note that the example treated by Cunze (1989) concerns a clay with 90 ‘Yowater content, while the tests reported in this study are performed on clay with a maximal water content of 37%. Figure 3 below shows the evolution of excess pore pressure for ten successive identical hammer blows. There are large values of excess pore pressure at the two first blows and following blows induce a relative low excess pressure. Because of no drainage in the soil the pore pressure reaches rapidly a maximum value. The first blow corresponds in this situation to restrike causes a weak displacement before allowing large displacement at the two following blows. As excess pore pressure do not varies significantly after the three first blows, the displacement remains constant. As long as excess pore pressure develop near the pile-soil interface, the effective stress must diminish. But this behavior is not observed (figure 3). This situation can be explained by the fact that no such large excess pressure develop at the interface because of transverse drainage in the cylindrical clay sample. This effect is verified by water content measures in the sample after tests. There is more water at midradius from the pile that at interface.
257
3.3. Influence of excess pore pressure on pile shaft resistance The large capacity increases with time (set-up) are considered unusual for sandy soil. In cohesive soil the fine grain content seems to control the permeability preventing normal rapid drainage conditions. Low unit shaft friction was observed during driving and was higher on restrike after reduction in pore water pressure (Likins et al. 1992). The soil set-up is the result of dissipation of the piling induced excess pore pressure in the soil after the initial driving. Figure 5 shows the evolution of both excess pore pressure and unit shaft friction during a driving sequence. The shaft resistance is derived using stress wave measurements near the pile top (Benamar, 1996). Such quantity is calculated as the difference between incident and transmitted stress wave in the pile-soi] interface. There is a large decrease of excess pore pressure at the first blow before reaching a steady state of constant excess pore pressure. But the unit shaft friction does not follow the same evolution, showing a low variation. When the excess pore pressure decreases the effective stress in the soil must increase. The measured unit shaft friction does not increase by the same amplitude than excess pore pressure decreases. In clayey soil the excess pore pressures are still dissipating. Therefore, the unit shaft friction is smaller than the value expected after some fixther wait. There is no direct relation between the variation of such two parameters. Thus, the shaft resistance can not be directly related to the effective stress in the soil as defined by Terzaghi's concept. Note that the average pile displacement corresponding to these measurements is 4.85 mm.
"We 3: Of Pore Pressure and pile displacement during driving. It is evident that in piling the displacement increases with hammer energy. Using driving tests resuits involving several hammer energies, the evolution of excess pore pressure and pile displacement with driving energy is investigated (figure 4).
Figure 4: Influence of driving energy on pile displacement and excess pore pressure Contrary to the pile displacement, the evolution of excess pore pressure is not affected by hammer energy and the same behavior as in figure 3 is presented here. The amplitude of pore water pressure being more large for high energy. The excess pore pressure reaches rapidly an equilibrium state and no more important increase is possible.
258
Figure 5 : effect of excess pore pressure on pile shaft resistance
4. EFFECT OF SURCONSOLIDATION ON INDUCED PORE PRESSURE DURING DRIVING The clay reconstituted sample was consolidated under a confrning pressure of 500 kPa and driving tests were performed under an isotropic stress of 300 kPa. The variation of shaft resistance with the overconsolidation ratio (OCR) was investigated (Benamar et al. 1993) and an increase of 30 % in unit shaft resistance is noted for a value 1.67 of OCR. Driving piles is than more hard in overconsolidated soils. The influence of overconsolidation on the induced pore pressure in soil is carried out (figure 6). For an OCR value of 1.67, the excess pore pressure is higher in normally consolidated soil. This difference is much more marked at the first blow, excess pore pressure being four times more important in normally consolidated soil. In the rest of driving sequence the difference of induced pressure is less important, the ratio being approximately equal to 2. The driving tests are performed in each case after a wait, leading consolidation to be made. The surconsolidation makes the soil sample in low stresses conditions and wave stress is not sufficient to reach stress statelevel as in normally consolidated soil. These results explain perfectly the conclusions concerning the shaft fiiction mentioned above. Nevertheless the increase of unit shaft resistance does not reache the same proportions as excess pore pressure. The such parameter can not be directly related to the shaft resistance. An effective stress method of analysis for estimating dynamic pile skin fiiction taking into consideration the induced pore pressure can not be adjusted, using effective stress theory.
Figure 6: Induced pore pressure in normally consolidated and overconsolidated clay
5. CONCLUSIONS The experimental set up in laboratory allowed an investigation of the influence of pile driving in induced pore pressure. Such experiments are difficult to perform in field piling. The dimensions of driving laboratory model can affect the interpretation of induced pore pressure measures, but the general conclusions concerning the influence of various parameters seems to agree. The variation of pore water pressure in the soil during driving do not induce the same variation on the dynamic shaft resistance. The experimental results show that very small excess pore pressures are developed due to dynamic shearing of the pile (pore pressures less than 2 % of the consolidation stress). The pile driveability is not strongly affected by the excess pore pressure which partially dissipate in the pile-soil interface. The measured pile displacements seem very low by comparison with the applied dynamic energy, implying high shaft fiiction. Development of excess pore pressure depends on the initial state of stress in the soil before driving. Parametric studies show that the most significant effects associated with pore water pressure are the reduction of the maximum skin fiiction allowed at the interface and the alteration of the pile response due to slippage. The shaft resistance can be fully mobilized easily without pore water pressure than with it. REFERENCES Benamar, A., P. Lepert & D. Levacher 1992. Physical and numerical simulation of lateral shaft fiiction along a driven pile, Proc. 4th Int. Con$ on the Application of Stress-wave Theory to piles, The Hague, The Netherlands, Sept 1992., pp. 3742, Balkema. Benamar, A. & D. Levacher 1993. CaractCristiques de l'interaction latirale argile-pieu durant le battage. Annales de 1 YTBTP, juin 1993, No 5 14, pp 55- 65. Benamar A. 1996. Dynamic soil resistance from pile driving analysis. Proceedings of the Europeafi conference on Structural Dynamics, Eurodyn '96 Florence, Italy, 5-8 June 1996, Vol. 2, pp. 1039. 1043, Ed. Balkema. Cunze, G. 1989. A modified method for the estimation of excess pore pressure generated bj pile driving. Proc. Int. Con$ on Soil Mechanic: and Foundation Engineering, Rio de Janeiro August 1989, vol. 2, pp. 1097 - 1100, Balkema . Likins, G., J. DiMaggio, F. rausche & W. teferra 1992. A solution for high damping constants ir sands. Proc. 4th Int, Con$ on the Application o, Stress-wave Theory to piles, The Hague, Thc Netherlands, Sept 1992., pp. 117-120. Balkerna.
259
Randolph, M.F., J.P. Carter & C.P. Wroth. 1979. Driven piles in clay - The effects of installation and subsequent consolidation Geotechnique,vol. 29, Nr. 4, pp. 361 - 393. Steenfelt, J.S., M.F. Randolph & C.P. Wroth 1981. Instrumented model piles jacked into clay. Proc. 10. Int. ConJ:on Soil Mechanics and Foundation Engineering, Stockholm, vol. 2, pp. 857 - 864. Vesic, A. 1972. Expansion of cavities in infinite soil mass. Journal of geotechnical engineering. ASCE, vol. 98, SM2.
260
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
An investigation of pile diameter influence in the bearing capacity on Dynamic Load Test (DLT) G.F? Bernardes Department of Civil Engineering, U N E S t Guurutinguetci,Brazil
C.S.Andreo & C.GonGalves Benuton Pile Foundation Company,SGo Puulo, Brazil
ABSTRACT: During the construction of five residential buildings in the city of Taubate, State of S3o Paulo, it was possible to carry out one comprehensive investigation of the behavior of precast concrete piles in clay shales. This paper describes the results of Dynamic Load Tests (DLT's) executed in three piles with different diameters and with the same embedded length. The tests were monitored using the PDA@(Pile Driving Analyzer) and the pile top displacement was measured by pencil and paper procedure. From the curves of RMX versus DMX resulted fkom CASE@method, CAPWAPC@analyses were made for signals where the maximum mobilized soil resistance was verified. The results were compared with the predicted bearing capacity using the semi-empirical method of Dkcourt & Quaresma (1978) and DCcourt (1982) based on SPT values and the description of the soil profile. Some comments related to the values of quake and damping used for clay shales in the analyses are also presented the high stress levels developed during driving which might reach the pile element strength. Results of Dynamic Load Test in soft clays (Gonqalves at al. 1996) indicate that even for profiles with NI less than 3 a pile can have considerable bearing capacity if the effect of set up is taken into account. The point unit stress resistance can be estimated by the expression
1 INTRODUCTION The semi-empirical methods of Decourt-Quaresma (1978) and D6court (1982) for prediction of the bearing capacity of single precast concrete piles based on the SPT values is eequently used in Brazil because of its simplicity. For design purposes the ultimate load (&It) is usually separated into two components; the lateral or skin load (Rlat) and the point load (hoe) as follow:
qO=C.N,
(kPa)
(3)
where N, represents the average of SPT measured at base level and the ones measured one meter above and bellow the pile point. The characteristic factor "C" can be estimated from Table 1 as a fkction of the soil type.
where Alat and At, are respectively the lateral and base bearing areas and (fs) and (qo) the Unit shaR and point resistance expressed in stress units (Ha). According to Decourt (1982) the lateral unit stress (fs) can be calculated by the equation
Table 1 : Characteristic factors "C"
Soil Type Clay Clayey Silt Sandy Silt Sand
where NI is the average of SPT values along the pile shaft. In the above equation there is no distinction of soil type and it is recommended NI values of 3 and 50 as lower and upper limits of SPT values. Cintra and Aoki (1 999) suggest for precast concrete piles a value between 15 to 35 for upper limit depending on pile diameter. These values take into consideration
"C" (Ha) 120 200 250 400
The factors presented in Table 1 represent average values obtained in static load tests performed in different geological formations. During the investigation of the behavior of precast concrete
261
depth between 3.80 to 11.00 meters it was horizontally bedded clay shale formation, plastic, dark gray to graphite. Underneath, the soil layer consist of stiff clay shale with no sign of lamination, high plasticity, green, until the end of SPT boring at 19.00 meters deep. The ground water table was measured at one meter below the ground surface. The bearing capacity predictions for the clay shale layer based on DCcourt & Quaresma Method (Fig.1b), gives 54.0 kPa and 4720.0 kPa for average unit shaft and point resistance respectively.
pile in clay shales formed in Cenozoic age, Taubate Basin, S.P. (Bernardes et a1.1999) it was possible to v e r that ~ the bearing capacity prediction was too conservative while the prediction for steel piles in clay shales from "Ilhas" Formation in Bahia (Velloso and Hammers 1982), showed exactly the opposite situation. Therefore, it is recommend to propose factors and constants to be used in this semi-empirical method related to a specific geological formation. In this paper, the influence of the pile diameter in the bearing capacity is investigated. The predictions using DCcourt & Quaresma Method is compared with the dynamic load test results executed in precast concrete pile in clay shale.
2 GEOLOGICAL AND CONDITION OF THE SITE
3 DYNAMIC TEST PILE PROGRAM
GEOTECHNICAL
The area of study is located in Taubat6 City founded over Terciary sediments formed in Cenozoic age known as Taubat6 Basin. These sediments cover an area of approximately 170 km in length and 20 km wide forming the Paraiba River Valley. The representative soil profile of the sediments near the test area is presented in Figure la.
Figure 1. Soil characteristics and bearing capacity predictions with depth.
The upper layer of sandy clay fill with dirt is 2.70 m thick overlying a 1.10 m thick layer of very soft organic clay, black, with fragments of roots. For
Three pile with different diameters (23, 26 and 33 cm) were driven down to 9.00 meters deep. These test piles were precast concrete piles having a yield stress of 35.0 MPa, Elastic Young Modulus of 30.15 GPa and speed wave propagation of 3510.0 m/s. The driving equipment was a free fall drop hammer with 20 kN ram weight for the 23 and 26 cm pile diameters and a 28 kN ram weight for the 33 crn pile diameter. A capblock of hard wood and a cushion made with Masonite fiber plate were used together with a steel helmet to reduce the high impact stresses. This system guarantees a driving efficiency between 45 to 55%. The piles were monitored using a Pile Driving Analyzer (PDA@) during pile installation and restrike after six days. The pile top movement was also measured by pencil and paper technique. First, the dynamic load test method (Aoki 1989) at restrike was used to plot the mobilized resistance by CASE@ method (RMX) versus the maximum displacement of the pile top (DMX), considering damping factor of 0.4 s/m, in other to select the best signals that guarantee full mobilization of the soil resistance (Fig. 2). Further analyses of the mobilized soil resistance are related to the three tests indicated in Figure 2. One complete investigation of the dynamic load test for 23 cm pile diameter is presented elsewhere (Bernardes et al. 1999). The signals of force (F) and velocity multiplied by pile impedance (ZV) that will by use are presented in Figure 3. As a simple procedure to check the measured signals and the CAPWAPC? analyses, the displacement from double integration of the accelerometer was compared with the direct measurement of the pile top displacement by paper and pencil. For all three piles, it can be seen from Figure 3 that after the time L/c there is a significant increase of F and corresponding decrease of ZV related to a high shaft resistance in the clay shale. The results of the CApWApC@analysis made with the signals from Figure 3 are presented in Table 2.
262
Figure 2. ( b x ) versus (Dmx) from Case@Method with J=0,4
It can be seen from Table 2 that the analyses gave a very little lateral quake (0.4 mm), probably related to the high stiffbess of the clay shale, and point quake in order of D/60. The lateral damping was practically constant (0.240 s/m) and the point damping was very small which as expected for failure condition (Gonqalves et al. 1998). The quality match ("Mqno") varies between 3.30 and 3.49.
4 ANALYSIS OF THE MOBILIZED SOIL RESISTANCE The simulated static load test from CAPWAPC for the test piles is presented in Figure 4.It can be seen that increasing the pile diameter the s t s e s s of the initial part of the load-displacement curve increase as the maximum mobilized soil resistance. This behavior might contribute for the evaluation of pile displacement under working condition. For one specific foundation, reducing the number of piles and increasing theirs diameter can cause a significant reduction in the foundation displacement. The soil resistance distribution mobilized along each pile element is presented in Table 3 . The last five elements, embedded in the clay shale, support most of the lateral resistance, between 80 to 90%. This behavior was already expected based on the dynamic signals from Figure 2. According to Table 2, the lateral unit resistance is equal to 127.0 kPa that is 135 YO higher than predicted in Figure I-b. For this particular investigation, it seems that the increase of pile diameter and consequently the lateral soil @
Figure 3. The measured force and velocity multiplied by impedance.
displacement during pile installation, do not alter the lateral unit resistance and the bearing capacity is only proportional to the lateral area. According to Eq. 2, a constant factor of 20 could be use instead of 10. For unit point resistance, the measured average values of 4330.0 kPa is 10% lower than the prediction using Eq. 3 with characteristic factor of 120 and the bearing capacity is also proportional to the base area. Based on the effective stress analysis, the skin bearing coefficient (Ns) and the bearing capacity factor (Nq) are 2.2 and 55 respectively which are congruent for s t Sclay.
263
Table 2: Summary of the CAPWAPCO analyses
Pile Wp Wr H Ef K i t DMX Qt Qs Js Jt MQno Diam. (%) (kN/m) (kN) (cm) (kN.m) (%) (kN) (mm) (mm) (mm) (s/m) (s/m) (cm) 23 0.95 20 80 8.7 54.4 687 14.7 3.9 0.39 0.249 0.114 3.46 26 1.25 20 100 9.9 49.5 813 13.8 4.0 0.40 0.239 0.098 3.30 33 2.05 28 140 18.4 46.9 1176 16.8 5.3 0.43 0.210 0.010 3.41 Obs.: W, = weight of pile; Wr = weight of ram; H = ram fall height; EMX = measured energy; Ef = efficiency; Rult= ultimate resistance; DMX = top displacement; Qt = point quake; Qs = lateral quake; J, = skin damping; J, = toe damping; MQno = signal match. Table 3: Summary of the mobilized soil resistance
1
Skin Resistance Distribution / pile segment (kN) 2 3 4 5 6 7 8 9
Rlat (m)
Diam.
Rlat (shale)
f, (shale)
Rtoe
(W (W) (W
qo (@a)
~ i l t
(W
4 23 cm 0.0 3.0 17.5 28.0 44.0 76.0 97.5 114.0 128.0 508.0 459.0 127.2 179.0 4308.0 687.0 4 2 6 cm 0.0 10.0 25.0 45.0 64.0 85.0 107.0 124.0 128.0 583.0 508.0 125.6 230.0 4332.0 813.0 4 33 cm 0.0 22.0 38.0 82.0 93.0 133.0 141.0 147.0 148.0 804.0 662.0 127.8 372.0 4349.0 1176.0 Obs.: Rlat = lateral resistance; Rlat(shale)= lateral resistance supported by the shale deposit; f, (shale) = unit skin resistance; Rt, = point resistance; qo = unit point resistance; = ultimate resistance.
is proportional to the lateral and base areas. The Decourt & Quaresma Method predicted very well the point resistance and under estimated the shaft resistance. For fbture foundations construction in the same region where site investigations identifjr this clay shale deposit, it can be use a factor of 20 can be used in Eq. 2.
Pile Top and Point Load (kN) 300
600
900
1200
h
E E
W
5.0
.w
5
?
3a
REFERENCES 10.0
Aoki, N.1989. A new dynamic load test concept. Drivability of Piles, Proceedings of the Discussions Session 14, XII ICSMFE, Technical Committes on Pile Driving, 1:1-4.
.,- 15.0 3
p1
Bernardes, G.P.; Gonqalves, C.; Andreo, C.S. and Fortunato, S.G.S. 1999. Evaluation of bearing capacity of precast concrete pile in shale fiom Dynamic Load Test (DLT). XI Panamerican Con$ on Soil Mechanics and Geotechnical Engineering: 3 :1491- 1496. Foz do Iguap: Brazil.
20.0
Figure 4. Simulated load deflection curves from Capwapc@.
5 CONCLUSIONS
Cintra, J.C.A. and Aoki, N. 1999. Carga admissive1 em fimdag6es pro fundas. Internal Publication EESC-USP: 61p. Siio Carlos: Siio Paulo.
Here, the dynamic load test with energy increasing (DLT) was used to investigate the influence of the pile diameter in the bearing capacity of precast concrete piles in clay shales. The DLT tests were also used to compare the resulted mobilized soil resistance fiom CAPWAPCO analyses with the predictions using the semi-empirical method of DCcourt & Quaresma based on SPT values. These analyses indicated that the unit shaft and point resistance is constant and the bearing capacity
Decourt, L. 1982. Prediction of the bearing capacity of piles based exclusively on values of the SPT. Second European Symposium on Penetrating Test: 1:29-34. Amsterdam: Holanda. Decourt, L. and Quaresma, A.R. 1978. Capacidade de Carga de Estacas a partir de valores de SPT. VI Congress0 Brasileiro da Mecbnica dos Solos e 264
Engenharia de FundaqGes, 1 :45-54. Rio de Janeiro: Rio de Janeiro. Gonqalves, C.; Andreo, C.S.; Bernardes, G.P. and Fortunato, S.G.S. 1998. Prova de carga dinhica em estaca pre-fabricada apoiada em areia argilosa densa. XI COBRAMSEG, 3:1535-1541, Brasilia: Siio Paulo Gonqalves, C.; Andreo, C.S.; Bernardes, G.P. and Fortunato, S.G.S. 1996. Estudo da reduqiio do comprimento de estacas pre-fabricadas atraves da m a k e do "set-up" corn base na instrumenta$io dinAmica com o PDA@. Terceiro Seminririo de Engenharia de FundaqGes Especiais e Geotecnia: 2:35-46. Siio Paulo: Siio Paulo. Velloso, P.P.C. and Hammers, M. 1982. Estudo da cravaqiio de estacas metalicas em folhelhos da formaqiio ilhas, na Bahia. VII Congress0 Brasileiro de Mecbnica dos Solos e Engenharia de Fundaqbes: 3:325 - 338. Olinda: Recife.
265
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Pile set-up in sands J. lc!Seidel Monash University,Melbourne, Vic., AustraLia
M. Kalinowski Baulderstone Hornibrook, Sydney,N S.K, AustraLiu
ABSTRACT: This paper describes the results of dynamic and static load testing at a site where open-ended steel tube piles were driven into a dense to very dense sand layer. The pile capacities on driving were significantly less than anticipated fiom the design calculations, with shaft resistances of 5 to 20kPa measured even at 40m depth. Later testing confirmed development of shaft resistances in excess of 15OkPa. A mechanism of pile whip or flexure is hypothesized. Resistance fiom the internal soil column was also investigated.
1 PROJECT DESCRIPTION
which occurred in the Tertiary and Quaternary Ages. An inferred geotechnical profile is shown in Figure 1. The subsurface sequence encountered at Pier 4 consisted of the following sequence: 0 3m of recent fill 0 6m of medium dense sand (Port Melbourne Sand, PMS) 0 6m of slightly overconsolidated marine sediments (Coode Island Silt, CTS) 0 l l m of highly overconsolidated silty clay (Fisherman’s Bend Silt, FBS)
The Bolte bridge is a new crossing of the Yarra river in Melbourne Australia. The bridge is a six-span structure which provides a new north-south fieeway link to the west of the central business district. The two river crossing spans were constructed by balanced cantilever method fi-om the Piers 2, 3 and 4. The construction for the outer spans used conventional closure pours. The geology at this site is a complex alluvial sequence of deposition, erosion and redeposition
Figure 1. Inferred stratigraphy at the Bolte Bridge site, showing approximate formation boundaries and in-situ strength test results
267
8m of dense clayey and gravelly sands (Moray Street Gravels, MSG - the lowest Quaternary stratum) e 3m of dense sands (Newport Formation, NF -the highest Tertiary stratum) e 13m of dense to very dense sands and sandy gravels (Werribee Formation, WF) e bedrock - Melbourne mudstone of Silurian Age (Silurian, S) The sequence at Pier 3 was generally similar, although the upper layers are predominantly absent because of the dredged river channel. A thin, extremely weathered Tertiary Age Basalt layer (TOV) was also intersected at Pier 3. Estimates of undrained shear strength and the results of SPT testing at Piers 3 and 4 are also shown in Figure 1. All piers are supported on large groups of driven open-ended steel tube piles. The piles are 1200mm diameter with a wall thickness of 20mm. No additional reinforcing plate was used at the pile toe. Design pile founding levels (by others) were RL-43m (43m below mean sea level) for the 56 piles at Pier 3, and RL-42m for the 18 piles at Pier 4 on the south bank of the river. In both cases the target founding stratum was the Werribee Formation comprising dense to very dense sands. An ultimate capacity of 25MN was computed for piles at both piers. Allowance had been made for downdrag in the design. e
37
CLAYEY GRAVELLY SAND (SC-GC) very dense, greyN>50 (33.45, green, fine to coarse sand, subanbdar gravel
38
Increasing sand content
5135mm)
N>50 ( I 7,28,
39
40
221100mrn)
SAND (SM-SP) very dense, green-my, some silt, trace clay; gravel medium to m a n e grained
2 TRIAL PILE DRIVING Driving of the first test pile commenced at Pier 4. This pile, denoted TPl penetrated very easily to the target depth. A hammer stroke of only 400 to 500mm was required - significantly less than the maximum stroke of 1200mm. This was unexpected, given previous wave equation predictions. The hammer operated very well, with energy transfers consistently in excess of 90%. Table 1 shows the driving record for the last 3m of driving of TPl to target depth. Depth intervals are 250mm. The energy figures are those determined from proximity switches built into the hammer. Table I Driving record above target depth for TPI
41
Gravelly band noted at 41.4m Occasional geen-grey silty clay layers
42
were the soil conditions in the founding Werribee Formation. A borelog close to the initial piling location at Pier 4 is shown in part in Figure 2 at the depth of the Werribee formation. Note in particular that the SPT N values were 50 or greater throughout the layer. An automatic trip hammer device was used. It is also emphasized that the formation at this location is predominantly sand, with evidence noted of some (<5%) silt and traces (4%) of clay. Occasional layers of slightly clayey sand were found. Piles were driven with a purpose-built 16 tonne Juntann hydraulic drop hammer. Preliminary Wave Equation analysis had indicated that this hammer, with a maximum stroke of 1 . 2 would ~ be required to reach the target founding depths. A hammer with a larger energy rating would be required to demonstrate the required ultimate capacity. Testing at the site consisted of monitoring 15% of piles using a Pile Driving Analyzer@and 3 static pile load tests.
(SM-SC) slightly clayey, some gravel
From(m)
To(m)
Blows
30185mm)
43
44
4s
Gravel band comprising anbdarquam and mudstone noted to 43.5171 (SP-SM) my, fine to medium grained, some silt with fri able mudstone gravel
Energy (t-m)
Set(mm)
39.75
40.00
(No) 23
8.2
10.9
40.00
40.25
27
6.7
9.3
42.50
42.75
24
7.2
10.4
Silty clay layer to 43.1rn
(SP) light grey, some silt, rare quartz gravel
46 Grey-brow1 clayey sand layer to 46.4m
47
(SP) Grey, coarse-grained, slightly gravelly
48 Light grey, trace of fines
Figure 2. Borelog excerpt from the founding Werribee Formation at Pier 4
The pile was allowed to set-up at this depth for a matter of some hours. As the indicated capacity was still low, and the penetration per blow still large, the
Of particular importance, both to the pile design and to the dynamic pile test data discussed hereafter 268
pile was driven a further 4 metres beyond the target founding depth. The pile was restruck at 48 hours (2 days) and 96 hours (4days), and demonstrated some capacity increase, but insufficient to satisfy design requirements. CAPWAP@ analysis of the data indicated shaft resistances typically 5 to 10 kPa in the dense to very dense sands on driving, increasing to between 30 and 40kPa after 96 hours. These were significantly less than the resistances predicted by geotechnical analysis, as indicated in Figure 3.
Figure 4. Comparison between CAPWAP@ prediction and static load test for TP 1
3 OBSERVATIONS DURING CONTRACT DRIVING
Figure 3. Comparison of inferred shaft resistance distributions for TPI at driving and restrike with design pile shaft resistance
Test-pile TPI was subsequently driven to a depth of 48.5m, at which time the capacity indicated by dynamic pile testing was still significantly lower than expected. Despite this, a decision was made to perform a static test on the pile at this depth. Reaction was provided by anchors founded in the mudstone rock a significant depth below the pile toe. A prediction of the pile capacity was made fiom C A P W Aanalysis. P~ The prediction and the results of the load test are shown in Figure 4. Plunging failure commenced at a load of 5.5 MN, and a load of more than 6MN could not be sustained. The comparison between prediction and static test is excellent. Due to program requirements a decision was made to drive all subsequent piles for the project to the basement mudstone rock underlying the Werribee Formation sands. It was not feasible to leave piles in the sands in the hope of further and SUEcient set-up effects.
269
The majority of further pile driving and testing occurred at Pier 3 , the central river pier. Pile driving was undertaken from a large sheet-piled cofferdam which had been filled to above high water level. Results fiom testing of Piles 301 and 307 will be further described in this section. It is noted that the Pier 3 pile cap was approximately 12.5m x 52m in plan. Pile 301 was a corner pile, and had two adjacent neighbor piles at 2.5m centre-spacing. Pile 307 was an edge pile, again with two adjacent piles at 2.5m centre-spacing.
3.1 Pile 301 test sequence Pile 301 was tested on initial drive on 16 November, 1996. Subsequent restrikes were performed on 25 November (9 days), 30 November (14 days), and 8 December (22 days). In order to ascertain the relative amounts of internal and external shaft resistance on these open-ended tube piles, Pile 301 had been drilled out to within 6 metres of the pile toe between the last two restrikes. The force-time and velocity-time responses for each case, including two blows for the final restrike of this pile are compared in Figure 5. Note that the pile was installed with the Juntann 16 tonne hydraulic drop hammer. All restrikes were performed with a 20 tonne trigger release drop h a m e r capable of 3.5m maximum drop height.
The different hammer characteristics are evident from comparing the input waveform of the fxst record against all the other records. 3.1.1 Pile 301 - “Set-up” It is evident to a person trained in the analysis of dynamic pile testing records that there is a significant increase in capacity, and particularly in shaft resistance from the driving case to even the frst restrike after 9 days. A capacity increase of this magnitude had not been observed for the driving of the test pile at Pier 4 in very similar stratigraphy. Table 2 summarizes the total capacity and the shaft and toe capacity contributions for Pile 301 at the time of each test. Table 2. CAPWAP@ mobilized capacity summary Pile 301
End of Drive
Shaft Toe Total
Restrike #I
Restrike #2
Restrike #3 B10w2 mobilized resistances (in MN) 15.5 15.9 21.4 10.4 1.5 .5 2.7 1.1 13.1 17.0 17.0 21.9
Restrike #3 Blow6 17.4 4.0 21.4
The shaft resistance distribution estimated by CAPWAP@ analysis for all tests is shown in Table 3 for each 2m segment of the 5 l m of pile penetration. It is noted that the shaft resistance distributions are not purported to be exactly correct. However, the total magnitude of shaft resistance is expected to be reasonably accurate, and the distribution is expected to be reasonably indicative of the variation of shaft resistance with depth at the time of testing. The availability of energy to hlly mobilize capacity can also affect the apparent shaft resistance distribution, particularly near the pile toe. It is furthermore noted that these shaft resistance values are computed on the basis of the external perimeter alone. No account has been taken of the internal shaft resistance or the sum of internal and external perimeters. That is, the implication is that all resistance is external. The authors believe that this is normal convention for dynamic pile testing. To the authors’ knowledge, there is no substantive evidence on the relative contributions of internal and external shaft resistance. In examining this table, it is evident that shaft resistances at the end of driving were extraordinarily low over most of the pile length, excepting the bottom 6 metres (last 3 rows of Table 3), which correspond to penetration into the weathered Silurian rock. To a depth of 47 metres, pile shaft resistances were uniformly less than 20 @a, and as low as 2 kPa in deposits that were found to be dense to very dense.
270
Table 3. CAPWAPO mobilized shaft resistances for Pile 301 Depth below ground 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51
End of Drive
As can be seen fiom Table 3, the estimated shaft resistances had generally, and substantially increased between the second and third restrikes. In the lower 24m of pile shaft, the inferred average shaft resistance increased from 155 kPa to 198 kPa over this period. Even by the 6'h blow, at which time there was a noticeable reduction of shaft resistance, the average shaft resistance over the lower 24m had decreased to 16OkPa - still above the inferred shaft friction at the time of the second restrike. It would appear that the loss of resistance due to removal of the internal soil column was more than compensated for by other factors which served to increase the external resistance in the intervening period. In addition, the average shaft resistance values computed for Pile 301 are not noticeably less, and in some cases even exceed the shaft resistance values for the other corner piles 315, 342 and 356 which had the same configuration of surrounding piles. The internal soil columns of these piles remained intact. Although no definitive recommendation can be made from this data, it would appear that the internal shaft resistance was not a major component of the total shaft resistance for this project. If it had been a significant component, some noticeable loss of shaft resistance would have been expected, and the shaft resistance for this pile would have been less than that measured for equivalent piles in the group.
Restrike
Restrike Restrike Restrike #1 #2 #3 #3 B10w2 Blow6 mobilized shaft resistance by depth (in kPa) 5 5 4 2 2 5 5 13 13 6 6 7 22 22 9 7 II 25 26 10 9 11 27 16 10 10 I1 26 16 9 12 16 18 19 10 10 15 10 II 14 10 17 15 15 19 10 40 47 17 20 10 24 35 53 18 79 26 94 73 13 76 65 71 65 8 76 106 113 36 4 67 I06 81 39 2 89 160 165 150 4 105 178 216 167 9 103 188 216 167 16 17 93 216 167 13 17 161 216 167 13 17 186 216 196 13 17 162 216 196 91 20 186 216 196 200 91 169 255 222 167 159 169 255 222
For the first restrike, 9 days after driving, the shaft resistance increased by approximately 50%, however, this increase was very uneven. Almost no increase was observed to a depth of 25m or beyond a depth of 37m. Very large increases were, however, observed in the Moray Street Gravel and Newport Formations. Although these formations contain some thick clayey bands, they are predominantly gravelly sands with little or no fines content. The second restrike at 14 days after driving suggests continuing increases in shaft resistances below 29m, and a sudden and substantial rise in shaft resistance in the Werribee Formation sands. At this time, shaft resistances in the Werribee Formation were indicated as 160 to 180 kPa - well in excess of the design assumptions. 3.1.2 Pile 301 - Assessment of internal shaft resistance After the second restrike was performed, a decision was made to try to estimate the relative proportions of internal and external shaft resistance, for the purpose of estimating the reliability of the shaft resistance predictions. The internal soil column within Pile 301 was therefore removed to a depth of 6m above toe level. The internal water level within the pile was maintained close to the pile head. The third restrike test was performed within 1 or 2 days after removal of the internal soil column.
3.2 Pile 307 - Load-settlement prediction A fkrther static pile load test was undertaken at Pier 3 on Pile 307 in order to confirm the validity of the dynamic pile testing at this site, particularly in view of the unexpectedly low shaft resistance values measured during driving. Driving of Pile 307 was completed on 14 December, 1996. Restrike testing of the pile was -performed on 8 January, 1997 - 25 days later. Programming constraints only allowed the pile to be statically load tested on 4 February - 52 days after installat ion. Two analyses were undertaken on the restrike data for Pile 307 - on the first and fifth blows delivered. The capacity estimates for the two blows were similar, however, CAPWAP@analysis indicated a trend of decreasing shaft resistance and increasing toe mobilization for the test sequence. The predicted load-deflection responses for both blows were provided in advance of the static load testing as suggested upper- and lower-bounds to the expected test response. A comparison of predicted and measured responses is shown in Figure 6. The two predictions bound the load test measurement. It is noted that the ultimate pile capacity was not actually measured in either dynamic or static test, so no direct comparison 27 1
the "cookie-cutter" effect. As a result, the dynamic response will include a distributed internal shaft resistance response along the pile shaft-soil column interface. The inertial effect is not present during static loading, and the soil column is therefore more likely to move with the pile, in a plugged manner. Internal soil-column resistance in this case will be equal and opposite to the resistance developed at the pile base. The internal resistance is concentrated above the pile base. It is interesting that despite these different behaviors, the predicted load-settlement responses were in very good agreement.
of ultimate capacity is possible. However, the loaddeflection prediction is, as stated, very good.
4.2 Low driving resistances The reason for the extremely low installation shaft resistances measured in dense to very dense sands (see Table 3) can only be speculated. The local shaft resistance at the pile-soil interface, z, is a hnction of the effective horizontal stress at the interface, d h , and the interface fiiction angle, 6. Thus
Figure 6. Predicted and measured load-movement responses for Pile 307.
z = ort1 . tan 6
4 DISCUSSION
Consider the driving of Pile 301, and in particular the inferred shaft resistance of approximately 13 kPa at a depth of 43m. Assuming an average bulk density of 19 kN/m3 to this depth, and a water table at ground surface, the vertical effective stress at this depth would be approximately 400 kPa. According to Kulhawy (1 984), the horizontal effective stress for this type of low-displacement pile should be approximately equal (i.e. K = 1.O, and (rrll= 400 H a ) . The pile-soil fiiction angle, 6, is a hnction of the soil fiiction angle, $, and the pile material, in this case steel. Kulhawy and others recommend values of 6/+ of approximately 0.7. For the dense to very dense sands at this depth, a Eriction angle of at least 40" should be expected. From which 6 = 28" is estimated. From Equation 1, T should be 210 kPa, although design convention is to limit values to approximately 120 to 15OkPa. Even if the proposition of a critical depth is accepted, the expected shaft resistance would be at least 90 kPa at this depth. it is noted that the value of 210 kPa actually agrees reasonably with the shaft resistance values determined for the third restrike of Pile 301. This suggests that the long-term response of these piles can be justified by geotechnical theory. The key question is what caused the severe reduction (to 13 kPa) during pile installation. The possibility of a reduction in interface fi-iction angle is discounted. Such changes have not been reported. The presence of mica is known to reduce
The project is notable for a number of reasons: 1 It provides additional documented 'Class A' comparisons between dynamic and static load testing, in particular for open-ended steel tube piles. 2 The shaft resistances of the tube piles during installation was significantly lower than anticipated in design, or than would be considered reasonable by conventional geotechnical analysis. 3 The project provides some evidence, albeit indirect, regarding the proportion of internal and external shaft resistance of steel tube piles 4 The shaft resistance of piles at this project increased dramatically (but inconsistently) over time, even in deposits that only contained a small fiaction of fines 4.1 Load test comparisons Predicting the behavior of open-ended steel tube piles can be complicated because of physical differences in pile response during dynamic and static loading. In particular the behavior of the soil column, as plugged or unplugged can be quite different during static and dynamic loading. During a dynamic event, the pile will tend to move past the internal soil column because of the inertia of the soil column. The pile therefore acts as an unplugged or partially plugged pile. This has been referred to as 272
interface friction angle, but this effect would be permanent; would not cause a reduction of this magnitude; and in any case, mica was not reported in these sands. Temporary increases in pore water pressure can reduce the effective horizontal stress in fine-grained soils, resulting in substantially lower installation resistance, however, in this case, the sands are too coarse to sustain elevated pore pressures. Note the capacity of test pile 1 only recovered marginally over a period of 4 days. Another mechanism for reduction of lateral stress on the pile must be found. Tomlinson and Poskitt have both previously observed the phenomenon of pile whipping, which can create a cavity around a pile. This cavity can only be sustained if the soil possesses some amount of binder which gives the soil a temporary or permanent cohesion. A dense soil in particular may be capable of self-support and arching f?om the pile surface. For the want of a more plausible explanation, the phenomenon of pile-whip is believed to have caused low contact stresses between the pile and the soil, and hence the low observed driving resistances. Given that the pile hammer and assembly had a combined weight of 28 tonne; that the hammer was not restrained by any leader system; and that free pile lengths above ground of up to 20m were observed during driving, sway and pile whip could be feasible. Support for this conjecture may also be found in the driving shaft resistance distribution. The resistance distributions for test-pile 1 shown in Figure 3 at the completion of driving and for various stages of restrike show a local maximum at mid-pile length (depth 20m) and local minima at the !A and % positions (8m and 32m respectively), especially at the end of drive. This resistance distribution was characteristic of the initial driving behavior of most of the piles tested, and this characteristic also persisted in some of the restrike data. Figure 7 shows the resistance distributions for Pile 301 at the end of driving, and also at a penetration of 43.7m, 7m above final toe level.
It is interesting to observe that the characteristic resistance distribution is linked to the pile penetration rather than the adjacent soil, suggesting a structurally-induced geotechnical phenomenon. It is proposed that the pile was flexing in a Mode 2 Euler buckling shape. With this type of flexure, a local minimum in movement (and maximum in shaft resistance) could be expected at mid-length. Lateral movement would be greatest at the !A and % lengths, causing the greatest tendency for formation of a cavity, and hence least shaft resistance. Further research is required to substantiate this hypothesis.
4.3 Internal shaft resistance As noted earlier, the evidence at this site suggests that internal shaft resistance was a small percentage of the shaft resistance measured during dynamic testing. The results of removing the internal soil column were obscured by other effects which caused the external shaft resistance to increase in the period between the restrikes prior to and after the soil removal. Construction programming prevented restrike testing immediately before and immediately after removal of the soil column. Such a procedure would have minimized other effects and made the comparisons of greater value. Further such testing is recommended in the fkture to provide some reliable international data on internal shaft resistance for open-ended steel tube piles. 4.4 Shaft resistance increases The shaft resistance increases measured at this site would appear to be linked to the very low initial shaft resistances measured during installation and discussed at length in Section 4.2. In that discussion generation of locally elevated pore pressures was discounted as a mechanism for the low shaft resistances measured in the sands. Equalization of pore pressures must logically therefore be discounted as a mechanism for the observed subsequent shaft resistance increases. It was hypothesized in Section 4.2 that the low shaft resistances were due to a tendency for a cavity, or zone of low contact pressure to develop around the pile as a result of pile flexure during driving. It was also shown that the final shaft resistance values measured for Pile 301 were consistent with geotechnical theory. The mechanism of pile “set-up” in this case then needs to be one which will restore the contact pressure at the interface, removing the arching of the sands away from the pile. It is believed that both the densification and vibration effects of driving adjacent piles resulted in increased horizontal stress at the pile-soil interface. It is quite possible that these stress levels would de-
Figure 7. Shaft resistance distributions for Pile 301 during installation
273
velop naturally over time with creep phenomena, and that adjacent driving and vibration simply accelerated the process. Test-pile 1 which failed at a low capacity, and did not demonstrate a substantial “setup” effect was the first pile driven at site, and would therefore not have experienced the same vibration and densification effects that occurred in the large pile group of Pier 3.
5 CONCLUSIONS This project provided some interesting and unusual opportunities. It is concluded that dynamic pile testing was able to predict static load behavior for these driven steel tube piles extremely well. This was demonstrated by two Class A predictions. Piles were installed through relatively clean sands in a dense to very dense condition, and at significant depth with unusually low shaft resistances. A mechanism of pile flexure has been hypothesized as responsible for this phenomenon. Subsequent pile ‘set-up’ is believed to be due to restoration of normal stress conditions around the pile, promoted by vibration and driving of adjacent piles. Tests to determine the proportion of measured shaft resistance which can be attributed to the internal soil column were inconclusive. However, there is a reasonable suggestion that the internal shaft resistance was not significant. Further research is recommended. REFERENCES Kulhawy, F.H. (1984) Limiting tip and side resistance : fact or fallacy? ASCE Symposium on Analysis and Design of Piled Foundations, San Francisco, October 1984. Ed J. Meyer.
274
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Modeling pore pressure generation during dynamic testing of deep foundations I? L. Pinto Department of Civil Engineering, Universidude de Coimbru, Portugal
ABSTRACT: Dynamic testing of deep foundations is becoming a common method to determine the static capacity of piles and drilled shafts. Many of these foundations are placed underwater and it is expected that the generation of pore pressure will affect the resulting capacity. The objective of the current work is to present a numerical method to predict pore-pressure generation in this situation. An axisymmetric finite element formulation is used with consideration of a continuous pore pressure field. A pile installed in saturated soil is loaded by a sine wave pulse and the pore pressure is computed at the tip. 1 INTRODUCTION
porous media was proposed by Zienkiewicz and Shiomi (1 984). This formulation considers continuous solid and fluid displacement fields as well as the pore pressure field. The governing equations are
Large-strain dynamic testing of deep foundations is gaining popularity as a method to determine the static capacity of piles and drilled shafts. There has been a strong evolution in understating the phenomena involved since Smith (1960) proposed a rational approach to the problem. The work of Rausche et al. (1985) among others has provided useful tools for the interpretation of test results. Nevertheless some aspects require more research. Deep foundations are often installed in saturated soils and it is expected that the pore pressure generated by the compression wave will have a direct impact on the interpretation of the test results. The model currently used by many researchers and field engineers is based on the discretization of the pile on an assemblage of masses connected by elastic springs. The soil-pile interaction is considered through linear-perfectly plastic springs and dashpots. This model has limitations on the consideration of pore pressure generation. Alternatively, the finite element method presents great potential to perform such task (Pinto 1998).
where us is the displacement of the solid phase, uf is the displacement of the fluid phase, p is the pore pressure, CT"' is the effective stress tensor, n is the porosity, pSand pf are the microscopic mass densities of the solid and fluid phases respectively, K is the permeability tensor, Kf is the fluid bulk modulus and b is the body force tensor. A dot over a vector denotes time derivative. 3 FINITE ELEMENT FORMULATION
2 BACKGROUND
The solution of the governing equations through the finite element method yields the following set of equations:
The theory developed by Biot (1955, 1956 and 1962) is used to provide the framework for the dynamic behavior of saturated porous material. Saturated soil is considered as a two-phase mixture, where each phase has its own motion. The most general formulation to solve the differential equations governing the dynamic behavior of saturated
M a + C v + K d=F(t)
275
(4)
where M is the mass matrix, C is the damping and K the stiffness matrices. Each array is composed by the following sub-matrices:
The time integration follows the Alpha scheme introduced by Hilber et a1 (1977). If the integration parameters are chosen correctly, it preserves unconditional stability and second order accuracy. The numerical damping granted by this method affects less severely the natural modes with period T slightly larger than the time step At than Newmark’s counterpart. It can also dissipate more effectively the higher frequency (lower period) modes. In the current case, the deep foundation and surrounding soil were modeled using axisymmetric 9/9/4c elements. One of such elements is depicted on Figure I . A continuous, quadratic approximation is assumed for the displacement fields of the solids and fluid. Pore pressure variation is assumed bilinear within the element.
The pile-soil interface was modeled as impervious by slaving the horizontal displacements of the fluid phase to the solids phase, along the shaft and tip. This prevents the fluid from penetrating the pile. A limitation of the current model is that no interface elements were used.
Figure 2. FEM mesh
The soil is saturated, with the ground water table at the surface. The water was considered incompressible (Kf = 1.E 1 1 kN/m’) Two soil types were studied. The first is representative of a medium loose sand and the second of a phosphatic clay. For comparative purpose, analyses were performed with different reologic models. The properties of the sandy soil are presented on Table 1. Two extreme permeability coefficients were considered to study its influence on the porepressure generation.
Figure 1 . Element 9/9/4c
Table I . Soil propertics for the elastic analysis
This element is capable of providing accurate solutions without incurring in problems such as spurious pressure modes, m-sh locking or poor convergence rate.
Es MPa 20 20
v
n
KD
Ps
Pf
2.65 2.65
1.00 1.00
mls
0.37 0.37
0.3 0.3
1.0E-2 1.0E-10
A single pulse of harmonic load was applied at the pile head using the following expression:
4 NUMERICAL MODEL
q(r)= A(l-
The numerical model shown in Figure 2 represents a 38.7-m long concrete pile, 0.52 m in diameter. It has 33.5 m embedded in the ground. The concrete modulus is 41.6 GPa and Poisson’s ratio is assumed to be equal to 0.2. The time that is required for the compression wave to travel to the toe and back to the pile head (2L/c) is equal to 18.6 msec. For the analysis, the pile was modeled with 82 axisymmetric elements and 470 elements were used to represent the soil mass.
COS(
~ t ) )
(6)
Two load types were considered: a quasi-static with o = 50 rad/s and a typical dynamic pulse with o = 1000 rad/s. The first load is comparable to the duration of a “Statnamic” pulse. The time increment used was equal to 5 msec, during 250 msec. On the second load At = 0.25 msec and the response was studied for 25 msec. The maximum load, corresponding to 2A, was equal to 5.0 MN for the sandy soil and to 2.0 MN for the clayey soil.
276
5 RESULTS Results for the quasi-static load are presented and discussed first. The results show that for the elastic solution, without interface elements, the skin friction provides most of the resistance to the applied load and only a small percentage of the load reaches the tip (close to 3.5%). This load generates pore pressure in phase with it. Figure 3 depicts the excess of pressure, normalized by the hydrostatic pore pressure, for soil with high and low permeability. As should be expected, the less permeable soil has the highest pore pressure increase.
used has an associated flow law, with the consequent dilative behavior of the plastified points in the soil. This behavior causes higher confining stress around the shaft, providing higher skin friction and limiting the force reaching the pile tip. Also, as the soil dilates, the pore pressure is reduced. The combined effects explain the small departure of the plastic curve from the elastic solution. It is expected that a more complete model, with non-associated flow law and a cap to control the volumetric plastic strains would perform better. On a second set of runs, the same problem was studied on a clayey soil. The loading conditions were A = 1000 kN and w = 50 rad/s. The soil properties, displayed on tables 2 and 3, are based on lab and field tests performed on soft phosphatic waste clay found in Polk County, Florida. The Cam-Clay model was used in the analysis. The model is limited in its ability to represent the behavior of soil under cyclic load, but provides valuable insight on the differences from a linear elastic representation. Table 2. Elastic soil properties for clayey soil
Es MPa 0.456
v
n
0.45
0.75
KD m/s 6.9E-5
P<
Pf
2.65
1.00
Table 3. Soil properties for the CarnCIay model
Figure 3. Pore pressure ratio at the pile tip. Elastic soil and slow load.
The elastic model is over-simplistic, but points out the importance of the permeability in the pore pressure generation. Later the soil was modeled as elastic-perfectly plastic. The Drucker-Prager yield surface was used, with c’=O and 4 ’=30”.
CO
M
K
h
P’co (kPa)
3.0
1.0
0.147
1.570
r(z)
The elastic solution shows a 10% increase of pore pressure at the pile tip, resulting from the lower permeability and high deformability of the soil.
Figure 5. Pore pressure ratio at the pile tip for the clayey soil
Figure 4. Pore pressure ratio at the pile tip. KD=lE-2 m/s.
For this particular case, Figure 4 shows that the consideration of the plastic model has a very limited influence on the pore pressure at the tip. The model
The impact of the plastic model is obvious in Figure 5. As the soil surrounding the shaft plastifies, more load is carried to the tip. The reduction of volume imposed by the elastic and plastic volumetric strains causes the pore-pressure to raise more than on the elastic representation. 277
The results presented so far relate to a quasi-static load. It is of great interest to study the importance of the loading rate on the pile-soil response. For the following set of analysis, the pile was loaded by a single pulse, with the duration of 6.3 msec and the response was followed for 25 msec. This scenario is closer to a typical dynamic load test. On the case of the sandy soil, the Drucker-Prager model was used with the same properties as before.
6 CONCLUSIONS The purpose of the current research is to evaluate the potential of a FEM coupled dynamic formulation to predict pore pressure generation on piles installed in saturated soil under dynamic axial loads. A small study to determine the influence of permeability, soil’s reologic models and loading rate is presented. Results showed the importance of the permeability on the pore pressure. For the elastic solution and the particular soil and load conditions, a decrease of KD from 1 .E-2 to I .E-I0 m/s caused the excess of pore pressure to raise from 5% to 30% of the hydrostatic pressure. The Drucker-Prager model with an associated flow law did not show significant differences from the elastic solution. On the clayey soil, the CamClay model was able to better represent the pore pressure generated when the soil plastifies at the region of the tip. Loading rate and inertial effects are visible for both soils. The current approach seems to have potential to model pore pressure generated during the performance of a dynamic load test. Some improvements are required. A dynamic interface element may be implemented, to better model the load transfer through skin friction. Reologic models capable of modeling cyclic and dynamic soil behavior are also of interest. Most importantly, calibration with values measured in field and laboratory tests is required.
Figure 6. Pore pressure response for dynamic loading conditions on sandy soil.
The influence of the loading rate is captured on figure 6. The excess pore pressure raises to close to 80% of the hydrostatic value, a significant increase from the 5 % computed with the quasi-static load. The finding strengthens the importance this pressure may have on the interpretation of dynamic test results. Similar findings are shown on Figure 7. Here the clayey soil was used along with the CamClay model. A significant increase of pressure is computed but now it is quickly damped.
REFERENCES Biot, M.A. 1955. Theory of Elasticity and Consolidation for a Porous Anisotropic Solid, J. of Applied Physics, 26, 182185. Biot, M A . 1956. Thcory of Propagation of Elastic Waves i n a Fluid Saturated Porous Solid - I Low Frequency Range, J. Acoust. Soc. A m , 28, 168-178. Biot, M.A. 1962. Mechanics of Deformation and Acoustic Propagation in Porous Media, J. of Applied Physics, 33, 1482- 1498. Smith, E.L. 1960. Pile Driving Analysis by the Wave Equation, ASCE Journal of the Soil Mechunics and Foundation Division, 86:4, 35-61. Pinto, P.L. 1998. Coupled Finite Element Forinu1ution.s for Dynariiic Soil-Structure Iiiternction, Ph.D. Dissertation, University of Florida, Gainesville, USA. Zienkiewicz, O.C. & Shiomi, T. 1984. Dynamic Behavior of Saturated Porous Media; The Generalized Biot Formulation and its Numerical Solution, Iiit. J . Num. Anul. Methods it7 Geomechanics, 8,7 1-96. Hilber, H.M., Hughes, T.J. & Taylor, R.L. 1977. Improved Numerical Dissipation for Time Integration Algorithms in Structural Dynamics, Ecirthquake Engineering Nt7d Structurul Dytiatiiics, 5 , 283-292. Rausche, F., Goble, G.G. & Likins, G.E. 1985. Dynamic Determination of Pile Capacity, Journal of Geoteclznicrzl Engineering, ASCE, 11 1 (GT3), March, 367-383.
Figure 7. Pore pressure response for dynamic loading conditions on clayey soil.
278
5 Codes
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Keynote lecture: The performance of the dynamic methods, their controlling parameters and deep foundation specifications Samuel G.Paikowsky & Kirk L.Stenersen Geotechnical Engineering Research Laboratory, Department of Civil and Environmental Engineering, University of Massachusetts, Lowell, Mass., USA
ABSTRACT: An ongoing project supported by the USA National Cooperative Highway Research Program (NCHW) under the Transportation Research Board (TRB) of the National Academy of Science (NAS), is aimed at rewriting AASHTO Deep Foundation Specifications for the year 200 1 . The AASHTO specifications are traditionally observed on all federally aided projects and generally viewed as a iiatioiial code of US Highway practice, hence iiifluenciiig the construction of all the deep foundations of highway bridges throughout the USA. The new code is based 011 Load and Resistance Factor Design (LRFD) principles with resistance factors obtained from probabilistic analysis of data. A large database (PD/LT2000) is the backbone of the dynamic methods’ performance evaluation. This database originated with the work presented by Paikowsky et al. (1 994), Pailtowsky and LaBelle (1994), and additional information acquired since. A summary aiid careful evaluation of the large database is presented, detailing the performance of various dynamic methods when compared to static load testing to failure. The parameters that control the accuracy of the dynamic predictions are analyzed, suggesting the importance of certain mechanisms associated with the pile penetration aiid the dynamic simulatioiis. The controlling parameters and the statistical analyses are then utilized for the development of resistance factors to be recommended for the new specifications.
1 BACKGROUND National Cooperative Highway Research Program, project, NCHRP 24-17, “LRFD Deep Foundations Design” was initiated to: (i) Provide recommended revisions to the driven pile and drilled shaft portions of section 10 of AASHTO Specifications and (ii) Provide a detailed procedure for calibrating deep foundation resistance factors. The current AASHTO specifications as well as other existing codes based on Load and Resistance Factor Design (LRFD) principles were developed using insufficient data, hence they utilized mostly back-calculated factors. The main challenges of the project are therefore: (a) Compilation of large, high quality databases and (b) Framework for a procedure and data management to enable: (i) LRFD parameter evaluation and (ii) Future updates. These challenges include two requiremeiits: (i) Organization of the factors following the design - construction - quality control sequence (i.e. independence in resistance factors according to the chronological stage and the evaluation proce-
dure) and (ii) Overcome the generic difficulties of applying the L W D methodology to geotechnical app 1i cat i ons , i .e . incorporation of indirect var i ab i 1ity , (e.g. site or parameters interpretation), judgment (e.g. previous experience), and other similar factors The project team, headed by the first author, is divided into three major groups dealing with static analyses (University of Florida), probabilistic approaches and structural analyses (University of Maryland), and dynamic analyses (University of Massachusetts at Lowell). The present paper provides a background for design methodologies and the LRFD. Database PD/LT2000 is presented and analyzed. The state of practice and the selected dynamic methods are described, followed by an initial evaluation of the signal matching technique and examination of the controlling parameters. The performance of the dynamic methods is then provided, categorized according to the controlling parameters. The obtained results are used for the development of resistance factors to be recommended for the new specifications.
28 1
Limit States Design (LSD). The two types of limit states are the Ultimate Limit State (ULS) and the Serviceability Limit State (SLS) both of which must be satisfied when using the LSD method. Regardless of the complexity of the analysis and calculation, all limit states designs are carried out to satisfy the following criteria for ULS and SLS, respectively,
2 STATE OF STRESS DESIGN METHODOLOGIES 2.1’ Working stress design Working Stress Design (WSD) method, also called the Allowable Stress Design (ASD) method, has been used traditionally in Civil Engineering since it was introduced in the early 1800s. The design loads, Q, consist of the actual forces estimated to be applied directly to the structure.
Factored resistance 2 Factored load effects
(2)
Deformation 2 Tolerable deformation to remain ser(3) viceable ULS pertains to structural safety and involves the collapsing of the structure, or i n relation to piles, when the ultimate bcariiig capacity of the soil is exceeded. SLS represents the conditions affecting the function or service requirements (performance) of the structure under expected service/working loads. These conditions can include excessive deformations and settlement or deterioration of the structure pile(s).
where Q = Design load (F); Qr,// = Allowable design load (F); R,,= Ql,lt= Ultimate geotechnical pile force resistance; and FS = Factor of safety. The factor of safety is commonly defined as the ratio of the resistance of the structure (RJ to the load effects (Q) acting on the structure. Table 1, last used in AASHTO (1997), presents common practice in which different factors of safety are used in conjunction with the level of control in the analysis of the pile’s ultimate geotechnical resistance. Presumably, when a more reliable and consistent level of construction control is used, a smaller factor of safety can be implemented, which leads to a more economical design. Practically, however, the proposed factors of safety in Table 1 (excluding the static load test to failure) do not necessarily consider the conservatism (i.e. under-prediction) of the examined method and hence their economical basis is questionable, (to be flirther discussed in section 9.3).
3 LOAD AND RESISTANCE FACTOR DESIGN (LRW 3. I Principles The design of a pile depends upon predicted loads and the pile’s capacity to resist them. Both loads and capacity have various sources and leveis of uncertainty that engineering design methods and processes have historically compensated for by experience and subjective judgment. These uncertainties can be quantified using probability-based design, or safety check expressions, aimed at achieving engineered designs with consistent levels of reliability. The intent of the Load and Resistance Factor Design (LRFD) method is to separate uncertainties in loading from uncertainties in resistance and to assure a prescribed margin of safety. Figure 1 shows the probability density functions for the load effect, Q, and the resistance, R. “Load Effect” refers to the load calculated to act on a particular element, (e.g. a specific pile). As the loads are more deterministic than the resistances, the load effect has smaller variability (smaller coefficient of variation, translating to a narrower probability density function) than the resistance. The region where the two curves overlap is where the load effect is greater than the resistance, indicating a high probability of failure. The shaded area under the probability density function for the resistance represents the probability that the resistance will have a value between a and b.
Table 1. Factor of safety on ultimate axial geotechnical capacity based on level of construction control (AASHTO 1997). Basis for Design and Type Increasing Design/Coiistructioii of Construction Control Control Subsurface Exploration
X
X
X
X
X
Static Calculation
x
x
x
x
x
Dynamic Formula
X
x
x
x
x
Wave Equation CAPWAP Analysis Static Load Test Factor of Safety (FS)
X
X
x
x
3.50 2.75 2.25 2.00* 1.90
*For any combination of construction control that includes a static load test, FS =2.0.
2.2 Limit states design In the 1950’s the demand for a more econoinical design of piles brought about the use of Limit States or 282
Figure 1 . An illustration of probability density functions for load effect and resistance.
In LRFD, partial safety factors are separately applied to the resistance and load components. A factored (reduced) strength (capacity) of a pile is larger than a linear combination of factored (magnified) load effects. In this format, the strength is reduced and the load effects are increased, by multiplying the corresponding characteristic (noniinal) values with factors, which are called strength (resistance) aiid load factors, respectively. The nominal values (e.g. the nominal strength, R,,) are those calculated by the specific calibrated design method and are not necessarily the mean strength or resistance, and (see
late 1920s by the Russians N.F. Khotsialov and N. S. Streletskii (see Ayy~ibet al. 1998b) and was first introduced in the US by Freudenthal (1 947). Recent development of LRFD in civil engineering was initiated in structural engineering (e.g. Ellingwood et al. 1980). Reliability-Based Design codes utilizing the LRFD format are used by the American Institute of Steel Construction (AISC 1994, Galambos & Ravindra 1978) and by the American Coiicrete Institute (ACI 1995). An effort was made by the National Standards Institute (ANSI) to develop probabilitybased load criteria for buildings (Ellingwood et al. 1982a,b) and ASCE 7-93 (ASCE 1993). The American Petroleum Institute (API) extrapolated LRFD technology for use in fixed offshore platforms API (1989), and Moses (1985, 1986). Additional comprehensive summaries for the implementation of modern probabilistic design theory into design codes include Siu et al. (1975) for the National Building Code of Canada (1 977), Ellingwood et al. (1 980) for the National Bureau of Standards, and the CIRIA 63 (1977). Ayyub et al. (1998a) provide details on LRFD rules for ship structures that were developed for the US Navy. The AASHTO LRFD Bridge Design and Construction Specifications (1994) resulted from work in NCHRP Project 12-33 (Nowak 1993, 1999) provide design guidance for girders. 3.3 LRFD performance and advantages
3
Figure 1). For example, while and the distribution represent the mean and performance of signal matching analyses predictions from many case histories, R,, is the predicted value for the specific analyzed pile. Based on considerations ranging from analyzed case histories to existing design practices, a prescribed value is chosen for the probability of failure. The probability for failure of the designed pile following the application of the resistance and load factors should be smaller than the prescribed value (the overlapping area in Figure 1). Practically, the resistance factors are calculated to provide the prescribed probability of failure.
3.2 Background information The concept of using the probability of failure as a criterion for structural design was presented in the 283
Experience has shown that adopting a probabilitybased design code results in significant savings and/or efficient use of materials. Reliability improvements are still under evaluation even though the new codes are designed for reliability to be equal to or better than the older codes. Experiences are not yet well documented, but designers have commented that, relative to the conventional working stress code, the new AISC-LRFD requirements are saving from 5% to 30% stecl weight, with a 10% typical savings (Ayyub 1999). This may or may not be the case for other industries. Specific benefits in pile design include the following: 1 . A more efficiently balanced design results in cost savings and/or improved reliability. 2. Uncertainties in the design are treated more rationally and rigorously. 3. An improved perspective of the overall design and construction processes (sub- and superstructures), and the development of probabilitybased design procedures can stimulate advances in pile analysis aiid design. 4. The codes become a living document that can be easily revised to include new iiiformatioii reflecting statistical data on design factors. 5 . The partial safety factor format used herein also provides a framework for extrapolating existing design practice to new foundation concepts and materials where experience is limited.
3.4 LRFD in geotechnical engineering Early use of limit state design for geotechnical applications was examined by the Danish geotechnical Institute (Hansen 1953, 1956) and was later formulated into code (Hansen, 1966). Independent load and resistance factors were used, with the rcsistance factors applied directly to the soil properties rather than to the nominal resistance. Considerable effort has been directed over the past decade to the application of LRFD in geoteclinical engineering. LRFD approaches have been developed in offshore engineering (e.g. Tang 1993, Hamilton & Murff 1992), in general foundation design (e.g. Kulhawy et al. 1988, 1994, 1996), and in pile design for transportation structures (Barker et al. 199 1, O’Neill 1995). In geotechnical practice, uncertainties in resistance principally manifest in site characterization, soil behavior, and construction quality. The uncertainties have to do with interpreting site conditions, understanding soil behavior, and accounting for construction effects. Uncertainties in extzrnal loads are small compared with uncertainties in soil and water loads and the strength-deformation bchaviors of soils. The approach for selecting load and resistance factors developed in structural practice, though a useful starting point for geoteclinical applications, is not sufficient. Work is needed to incorporate in the LRFD formulation factors that are unique to geotechnical design. Philosophically, the selection of load and resistance factors need not be made probabilistically, although in current structural practice a reliabilitytheory-based calibration is commonly used. This approach focuses more on load uncertainties than resistance uncertainties, and does not include many subjective factors unique to geotechnical practice. An expanded approach is needed if the full benefits of LRFD are to be achieved for foundation design. The National Research Council reported, “this.. . subjective approach reflects the general lack of robust data sources from which a more objective set of factors can be derived” (NRC 1995). The report continues, “. . .realistically, because of the tremendous range of property values and site conditions that one niay encounter, it is unlikely that completely objective factors can be developed in the foreseeable future.” Today, the situation has changed somewhat, but not entirely. The research team of NCHRP 24-17 gathered robust data sources on pile capacity from which a more objective calibration of resistance factors can be made. Nonetheless, there remain uncertainties associated with (1) site conditions, (2) soil behavior and interpretation of soil parameters, and (3) construction methods and their quality. These factors are difficult or impossible to understand from
the pile databases alone. Knowledge factors need therefore to be developed and combined with the reliability-theory-based calibration of the database records to achieve a meaningful LRFD approach.
3.5 LRFD forpiles 3.5.1 1999 AASHTO LRFD Bridge Design Spec$cations for Foundations The ultimate resistance (I?,,) multiplied by a resistance factor (4, which is equal to the factored resistance (R,.), must be greater than or equal to the summation of the loads (Q,) multiplied by a modifier (71,) and load factor (x)for the strength limits states:
(4) where:
11, = factors to account for effects of ductility ( q ~ ) , redundancy ( q ~ and ) , operational importance (71). The 1999 AASHTO LRFD Bridge Design Specifications for Foundations provides the following equations for the determination of the factored bearing resistance of piles, QR:
for which:
Q,>= q,A,
(7)
where $q = resistance factor for the bearing resistance of a single pile specified for methods that do not distinguish between total resistance and the individual contributions of tip resistance and shaft resis= bearing resistance of a single pile (F); tance; Q,,[, Q,= pile tip resistance (F); Qs = pile shaft resistance (F); qr, = unit tip resistance of pile (F/L’); qs = unit shaft resistance of pile (F/L’); A, = surface area of pile shaft (L2); A, = area of pile tip (L2); and #q)qp,#qs = resistance factor for tip and shaft resistance, respectively, for those methods that separate the resistance of a pile into contributions from tip resistance and shaft resistance, The resistance factors for use in the above equations are presented in Table 10.5.5-2 of the specifications. The table incorporates a factor h,, to account for the level of field capacity verification. As an example, if a wave equation analysis is used to determine the pile capacity with an assumed driving resistance, a resistance factor of 0.65 is recotnmended. If pile driving formulas, e.g. ENR equation, are used to verify the pile capacity without stress wave measurements during driving, a h, factor
284
of 0.80 is recommended. This results in an actual resistance factor of 0.52 to be used in the above analysis/verification sequence (0.65 x 0.80).
3.5.2 Worldwide LRFD for dynamic methods Review of the development of LRFD applications for Geotechnical Engineering is presented by Goble (1999). A short description of LRFD codes that include dynamic methods follow. The Australian Standard for Piling-Design and Installation ( 1995) provides detailed recommendations for resistance factors to be used with the dynamic methods. The recommended resistance factors range betwecn 0.45 to 0.65 for methods without dynamic measurements (including WEAP) and between 0.50 to 0.85 when utilizing dynamic ineasurements with signal matching analysis. Selection of the appropriate resistance factor depends on driving conditions, geotechnical factors (e.g. extent of site investigation), and extent of testing (e.g. low range for ~ 3 of % the pile tested and high range for >15%). In traditional structural design specifications, a nominal value is given and the value used is based primarily on engineering judgment and cannot exceed the nominal value. The Australian Standard is therefore unique by providing a guide for choosing the appropriate resistance factor. Interestingly, no distinction is made regarding either soil type or time of driving (i.e. EOD, BOR) when referring to the signal matching based on dynamic ineasureinents. The method by which the resistance factors were generated is not provided in the code. The AUSTROADS Bridge Design Code (1992) provides resistance factors for four categories of dynamic methods. The range of resistance factors is quite large and there is no explanation as to how the resistance factors were obtained. Goble (1 999) postulates that the resistance factors were calibrated via the working stress design method. The Ontario Bridge Code (1 992) recommends relatively low resistance factors with no differentiation between the dynamic analyses. No information is provided on how the resistance factors were obtained. The Canadian Bridge Code (1992) is brief in its design requireinents for deep foundations. For dynamic load testing, resistance factors of 0.4 and 0.5 are recommended for routine testing and analyses based 011 dynamic nieasurements, respectively. 3.5.3 Dificu1tie.s of the Existing LRFD Codes All existing codes suffer from two major difficulties. One is the application of LRFD to geotechnical problems as described in section 3.4 (e.g. site variability, construction effects, etc.). The other problem is lack of data. None of the reviewed codes and associated resistance factors were developed based on databases enabling the calculation of resistance factors from case histories.
The current AASHTO specifications encounter additional difficulty due to the irrational multiplication of the resistance factor by the modifier h,. This procedure requires the interaction of two independent evaluations (e.g. static analysis and dynamic methods) and results in unnecessary and confusing conservatism. A clear separation of the resistance factors on the basis of design and construction is required. 4 PD/LT2000 DATABASE The database PD/LT2000 contains information related to 210 driven piles that have been statically load tested to failure and dynamically monitored during driving and/or restrike (389 analyzed measurements). PD/LT2000 is comprised of the integration of databases PD/LT (Pailtowsky et al. 1994) and PD/LT2 (Paikowsky & LaBelle 1994) with expansion by an additional 57 pile cases (Stenersen 2000). The data of PD/LT2000 were carefully examined and analyzed following procedures described by Paikowsky et al. (1994), resulting in detailed static and dynamic pile capacity evaluations. Table 7 presents a summary of the data contained in PDiLT2000 broken down according to site location and soil type, pile type and capacity, driving behavior and time of driving.
5 REFERENCE STATIC CAPACITY 5.1 Overview The dynamic methods are assessed by comparing the pile capacity of the evaluated method with a reference static capacity of the pile. The determination of the pile’s static capacity based on load- displacement relations is not unique. The test results depend on the load testing procedures and the applied interpretation method, both of which are subjective. The following sections examine each of these factors and its influence on the reference static capacity.
5.2 Failure Criterion Past work (Paikowsky et al. 1994) have resorted to a “representative” static pile capacity based on the assessment of five interpretation methods; Davisson’s Criterion (Davisson 1972), Shape of Curve (similar to the procedure proposed by Butler and Hoy 1977), Limiting Total Settlement to 25.4 mm and to 0.1B (Terzaghi 1942), and the DeBeer log-log niethod (DeBeer 1970). A single representative capacity value was then calculated for the analyzed case as the average of the methods considered relevant (i.e. provided reasonable value). The development of a framework for future modifications requires that the evaluated resistance factors be based on an objec285
Table 2. Summary of the data in the PD/LT2000 database. Pile Types
Soil Types
Location Pile p p e
No.
-Pile
OEP
-
CEp
6,
Voided Cont ?te
35
No.
Location
.
Northeast USA Soiitheust USA North USA South USA
254
9
305
5
356
8
Australia
sq.
406
1
Biwswick
L
457
8
508
8
Soil Inertia
Soil
I
I
Pile Capacities
1
Time EOD
44 69
Type of Data
&
BOR
Clay /Till
24
EOD
l0
BOR ‘s
&
EOD
I
I
Rock 2
BOR Conc (171 !?I)
EOR 610 762
Octagonal Cono.et e
Sand /Sill 16 I
Kon
5
Ontario
3
TinibeiMonotuhe
2
Total
2 10
1
DD
22
Sweden
1
NA
6
DR ALT
2 210
Notes: 1. OEI’ 2. CEP 3. EOD 4. BOR 5 . EOR 6. DD 7. DR 8. ALT 9. Blow Cts. 10. AI<
-
-
Open Ended Pipe Pile Close Ended Pipe Pile End of Driving Beginning of Restrike End of Restrike During Driving During Restrike Alternate measurement Blow Count Area Ratio
No. sq. Conc. NA USA Northeast USA Southeast USA North USA 18. South USA 19. Noi-thwcst USA 20. Southwest USA 11. 12. 13. 14. 15. 16. 17.
tive, repetitive procedure. 111 order to do so, the static capacity of each pile in database PD/LT2000 was evaluated utilizing all five aforementioned criteria and a representative capacity was assigned for each pile. A statistical analysis was then carried out by determining the mean and standard deviations of the ratio of the representative pile capacity to the method being evaluated. Figure 2 shows the histogram and calculated distributions (normal and lognormal) for Davisson’s failure criterion. Davisson’s criterion was found to perform the best overall and was therefore chosen as the single method to be used when analyzing load-displacement curves. The method provides an objective failure criterion and was also found to perform well for piles exceeding a
-
-
-
-
Number of Piles / Cases Square Concrete Non Applicablc / Unknown United States of America Federal Highway Regions 1, 2 & 3 Federal Highway Region 4 Federal Highway Regions 5 , 7, & 8 Federal I-Iighway Region 6 Federal Highway Region 10 Federal IHiahwav Region 9
diameter of‘ 610nini (examined through 30 pile cases). The data presented in Figure 2 demonstrates, however, that though small, a bias exists in the static capacity used as a reference for the evaluation of the dynamic methods’ performance.
5.3 Load Test Procedure An additional factor to examine is the influence of the static load testing procedure (loading rate) on the designated pile capacity. This influence was examined in two ways. Two detailed case histories from a research site in Newburyport, Massachusetts were evaluated. A pipe and pre-stressed concrete heavily instrumented 286
Figure 2. I-listograni and frequency distributions of KsD for 186 PDiLT2000 pile-cases in all types of soils.
friction piles were tested over a lengthy period of time at a bridge reconstruction site. Both piIes were tested using three types of static load testing procedures: slow maintained (testing duration of about 45 hrs), short duration (testing duration of about 6-8 hrs), and static cyclic (testing duration of about 15 min.). Dctails about the piles and the testing are presented by Paikowsky and Hajduk (1999, 2000) and Paikowsky et al. (1999). Thc interpretation of the load-displacement relationships in both cases suggested that the test type had an insignificant influence on the pile capacity, (referring to a failure criterion irrespective of the displacement). The effect of the test type was further investigated utilizing a databasc containing information related to 75 piles tested under slow maintained arid staticcyclic load testing procedures. In the static-cyclic procedure, the piles were loaded to failure using a high loading rate and then unloaded. The process was repeated for four cycles. The testing procedure and its interpretation method are presented by Paikowsky et al. (1 999). A comparison between the pile capacity based on Davisson’s failure criterion for the slow maintained tests and the static-cyclic capacity is presented in Figure 3. The obtained relations and the associated statistical information suggest that there is no significant influence on the static pile capacity based on the applied static load rate. The static cyclic load test results were also compared to the representative static pile capacity (based on the aforementioned five methods) resulting in a mean Ksc of 1.023 and a standard deviation of 0.057.
Figure 3. Comparison between pile capacity based on Davisson’s criterion for slow maintained load tests and static cyclic load test capacity for 75 piles, (Paikowsky et al. 1999).
These evaluations led to the conclusions that Davisson’s pile failure criterion can be used as a method to determine the reference pile capacity irrespective of the static load-testing procedure.
6 DRIVEN PILES AND DYNAMIC METHODS STATE OF PRACTICE IN USA HIGHWAY CONSTRUCTION To address the state of practice in the USA highway construction, a questionnaire was sent to the Department of Transportation offices of all fifty states, requesting information on the use, design and construction of deep foundations. The main results from 45 states and 2 Federal Highway officials regarding driven piles are: 0 Driven Piles constitute 75% of all bridge foundations. From the people who responded, 64% prefer the use of driven piles to any other foundation type. In the driven pile categories: 21% PPC (Pre-cast Pre-stressed Concrete), 52% H-Piles, 25% CEP (Closed Ended Pipe) and 2% OEP (Open Ended Pipe). 0 Dynamic Equations usage: 45% ENR (Engineering News Record, Wellington 1892) and 16% Gates (Gates 1957). 0 WEAP (Wave Equation Analysis Program) is used by 80%. 0 84% perform dynamic testing during construction on approximately 1% to 10% of the piles. 0 82% use EOD (End Of Driving) conditions to set the piles’ length, 18% use BOR (Beginning Of Re-
strike) only, and an additional 34% use both EOD and BOR. Some specifically use EOD for end bearing and BOR for friction piles (NH). 36% do not consider pile freeze or relaxation when assigning driving criteria. The questionnaire's responses suggest an extensive use of dynamic measurements, particularly at the end of driving, extensive use of WEAP, and significant use of the dynamic equations. Q
7 THE CHOSEN DYNAMIC METHODS 7.1 Oveiview Prior to detailed analyses leading to the deterniination of resistance factors, two coniponents must be established: (a) type of dynamic methods to be evaluated and (b) the conditions under which these methods need to be examined. Sections 7 and 8 address these issues, respectively. 7.2 Methods of Analysis 7.2.1 Geiier-al Table 3 presents a summary of the major available dynamic methods for evaluating pile capacity. The methods are subdivided according to the project stage (i.e. design vs. construction) and the need for data obtained through dynamic measurements. The incorporation of dynamic equations and WEAP reflects the previously described state of practice. The methods that require dynamic measurements can be broadly categorized as those that utilize a simplified analysis providing an instantaneous pile capacity evaluation for each hammer blow, and those that require elaborate calculations (i.e. signal matching), traditionally carried out in the office. 7.2.2 WEAP Based on Smith (1 960), the use of the WEAP (Goble & Rausche 1976) during design is of great importance for achieving compatibility between the driving system, the pile, and the soil conditions. Drivability study and pile stress analysis often determine the pile type and geometry and tlie adequacy of the proposed equipment. The evaluation of WEAP effectiveness for capacity predictions is difficult to assess as a large range of input parameters is possible and the results are greatly affected by tlie actual field conditions. Examination of the method through analyses making use of default values is probably the best avenue. Other evaluations, including WEAP analysis adjustments following dynamic measurements (e.g. matching energy) seem to be impractical in light of the other available methods, and lead to questionable results regarding their quality and meaning (Rausche et al. 1997, Rausche 2000). The WEAP analyses are evaluated in this study in two ways: (i) as a dynamic method for pile capacity pre-
diction, using WEAP default input values and the pile's driving resistance at the end of driving (EOD) compared to the static load test results and (ii) WEAP as a pile design method examining the analyzed stresses at the design stage with the measured stresses during construction. Such evaluation leads to a strength factor (related to the allowed structural stresses in the pile) that is beyond the scope of this manuscript. 7.2.3 Dyiiamic Equations The choice of dynamic equations came to address the state of practice and reflect a range in equation type and performance. While the Engineering News Record (Wellington 1892) was proven to be unreliable through the years (e.g. Olsen & Flaate 1967) it was founded on a solid theoretical basis and is used in construction in about half of the states in the USA. The equation's traditional foimulation (e.g. Massachusetts State Building Code 1997) includes a factor of safety of 6, which needs to be omitted when developing a resistance factor. The evaluation of the ENR equation is carried out, therefore, utilizing its basic formulation. The Gates equation (Gates 1957), while empirical, was found to provide reasonable results in the past (e.g. Olsen & Flaate 1967, Long et al. 1998). The equation was further enhanced by Richard Cheney of the Federal Highway Administration, (FHWA 1988), based on a statistical correlation provided by Olsen and Flaate (1967). 7.2.4 The Case Method The Case method (Goble et al. 1970, Rausche et al. 1975) is often used in field evaluations, as it is built into Pile Dynamics Inc.'s Pile Driving Aiialyzer (PDA), the most coinnionly used in the USA. The method is based on a simplified pile and soil behavior assumptions (free end and plastic soil), resulting in a closed forin solution related to the impact and its reflection from the tip. With the years, the method evolved to be implemented into at least five different variations (GRL 1999). The Case method utilizes a damping coefficient (Jc) that is assumed to be associated with soil type. The influence of this factor on the predicted static capacity depends on the reflected wave from the pile's tip, and hence on the driving resistance. The case-damping coefficient was investigated through a back calculation (to match the measured static capacity). Results, described in the next section, suggest no correlation between the soil type and the case-damping coeffcient. The coninion recommended practice suggests the use of the method based on a specific site/area calibration (GRL 1999). This approach, in conjunction with the application of the method for maximum resistance (RMX), has proven effective. Accumulated experience on extensive jobs in the Boston area (e.g. GTR 1997, 1998) has denionstrated the effectiveness of the Case method, when 288
Table 3. Summary of the available dynamic methods.
Categoiy
I
WEAP 1960, Goble et al. 1976) ENR (Wellington 1892)
Design Stage
D):t?nmic Eqziutions
I
Advai7 rages
Metlrod
Gates (Gates 1957)
Disadvnt 1rages
-
Equipment Match Drivability Study Structural Stresses
-
Sound Principles Common use
-
Empirical Comnlon use
~ ~ FHWA 1988 Signal Matching (e.g. CAPWAP) (Goble et al. 1970) Case Method (Goble et al. D):namic Measvrements 1970, Rdusche et al. 1975) Energy Approach (Pailtowsky 1982) "Resuired by the research Dane1 andior
Correction ~ based additional data
- Non unique Analysis Performance sensitive to field conditions
-
-
capacity predictions" Needs to be examined without a built in FS.
-
Found to be more reliable than other equations
Depends~on database ~
-
Was found to be reliable l ~
Stationa'y 'Oil forces Expensive Requires time
-
Office Method Found reliable a! BOR
Requires local calibration Presumed dependency of soil conditions found baseless
-
Was found reliable with local calibration How to obtain national or international calibration?
- Depends on original database ~
On
Solid principle of matching calculations to nieasurernents by iimosine nisd. B.C.
-
-
Simplified Analysis Field Method
-
Simplified Analysis Field Method
Required for Construction
- Required Evaluation foz-
Unreliable
-
Cotntiiell/
Shows capacity which may not be present at EOD
-
n
- Ideal foi construction
DOT's state codes
calibrated. A statistical examination of local calibration was perfoinied in Florida by McVay et al. (2000). The results of this analysis suggested that for 48 cases, the ratio between the static pile capacity to the Case method prediction at EOD was 1.344 f 0.443 (mean f 1 S.D.). As no generic conditions exist for the use of the Case method, international or national calibrations are unrealistic. As the projection of local calibration (of good experience and practice) beyond the geographical location may be unwise and/or unsafe, the Case method was excluded from the examined dynamic analyses.
7.2.5 Tlie Energy Approach The Energy Approach uses basic energy relations in conjunction with dynamic measurements to determine pile capacity; the concept was presented by Paikowsky (1982) and was examined on a limited scale by Paikowsky and Chernauskas (1992). Extensive studies of the Energy Approach niethod were carried out by Paikowsky et al. (1994), and Paikowsky and LaBelle (1 994). The underlying concept o f this approach is the energy balance between the total energy delivered to the pile and the work done by the pile/soil system. The basic Energy Approach equation is:
c
(9)
where R,, = maximum pile resistance, E,,,,, = measured maximum energy delivered to the pile, D,,,, = measured maximum pile top displacement, and Set = permanent displacement of the pile at the end of the aiialyzed blow, or I/nieasured blow count. For further details regarding the Energy Approach niethod see Pailtowsky et al. (1994) and Paikowsky (1 995).
7.2.6 The Signal Matcliing Techniques The response of the niodeled pile-soil system (e.g. force at the pile top) under a given boundary condition (e.g. measured velocity at the pile top) is compared to the measured response (force measured). The inodeled pile-soil system or, more accurately, the niodeled soil that brings about the best match (visual graphical match) between the calculated and measured responses, is assumed to represent the actual soil resistance. The static component of that resistance is assumed to be the pile's capacity at the time o f the driving. The 289
procedure of signal matching was first suggested by Goble et al. (1970), utilizing the computer program CAPWAP. Others developed similar analyses, (e.g. Paikowsky 1982, Paikowsky & Whitman 1990) utilizing the computer code TEPWAP. The TNO program was developed by Middendrop and van Wee1 (1986), which led to improvements and the CAPWAPC program, which is in coininon use to date. The signal matching technique is often refei-red to as post driving analysis or office method. With the availability of faster, portable computers, it became reasonably simple to conduct the analysis in the field as well. In difference from the ficld method however, the signal matching analyses cannot be carried out for each blow during driving.
8
THE CONTROLLING PARAMETERS
8.1 Overview
Preliminary examination of the parameters controlling the performance of the dynamic analyses is carried out prior to a final detailed evaluation of these methods, leading to resistance factors. Such examination iiifluences the sub categorization of the dynamic methods (according to the important controlling parameters), hence, directing the user to utilize the appropiate resistance factor according to the relevant conditions of the employed method. For example, if soil type is a controlling factor and the accuracy of the signal matching method is largely affected by soil type, evaluation of the method for different soil types will result in the development of resistance factors, depending on the soil type. Conversely, if soil type does not control the accuracy of the spccific dynamic method, categorization based on soil type is neither desired nor perused. The following sections outline the logic used for the preliminary examination of the controlling parameters, the analyses and the results. The rationale presented in this section follow previous studies by Paikowsky et al. (1994), Paikowsky (1995), Paikowsky et al. (1995), and Paikowsky and Chernauskas (1 996). The evaluation of static capacity from pile driving is based on the concept that the driving operation induces failure in the pile-soil system, (i.e. a very fast load test is carried out under each blow). Dynamic analyses encounter three fundamental difficulties: (i) correct formulation of the penetration process (e.g. soil motion, soil plugging etc.), (ii) evaluation of the static resistance out of the total resistance encountered during penetration, and (iii) time dependent pile capacity, (Paikowsky 1995). The parameters controlling the accuracy of the dynamic predictions therefore reflect the ability of each method to address the above difficulties. Based on the concept of a pile loading to failure under each blow, it has been traditionally assumed
that during high driving resistance (i.e. refusal) there is not sufficient pile penetration to mobilize the full pile capacity (e.g. Chellis 1961). As such, the dynamic methods are deficient under high driving resistance, categorized as equal or above 12BPI (Blows Per Inch) or approximately 5BPcm (Blows Per cm), (MHD 1988). The soil type is also believed to constitute a major factor as soil damping parameters are commonly employed to represent viscous resistance in the modeling of the soil’s dynamic behavior. This viscosity is assumed to be soil type dependent, and associated with intrinsic soil properties. As such, high viscosity values are expected for cohesive soils and low viscosity values are expected for cohesionlcss soils. Naturally, under a given velocity, high viscous values are associated with higher dynamic resistance, and logically should prove more difficult to accurately define the static resistance. The effect of time is well recognized but poorly quantified. Piles undergo a decrease or increase of capacity with time, also known as relaxation and setup, respectively. While the resistance during driving and its static component represent the conditions encountered during penetration, the major interest remains the long-term ability of the pile to carry load during its service life. The examination of the dynamic method predictions with static load tests (often carried out long after the driving) therefore remains valid. The predictions can be assessed in relation to the time in which the driving and/or the dynamic data have been obtained (e.g. EOD or BOR). The sections below examine the importance of each of the above three assumed controlling parameters. The results are used to evaluate additional possible controlling factors, laying down the framework for the detailed evaluation of the dynamic methods.
8.2The Eflect oJ’Soil Type The effect of soil type is examined in two ways: (i) the correlation between the parameters assumed to be soil type dependent and soil type, i.e. damping parameters and (ii) the accuracy of the predictive methods relative to the soil type. Figures 4 and 5 present the relationship between soil type and Smith damping parameters used in approximately 370 CAPWAP analyses, from PD/LT2000, for the tip and side pile resistances, respectively. Figure 6 presents the back-calculated case damping factors for 290 cases, from the PD/LT database (Paikowsky et al. 1994), required to obtain a match between the predicted capacity and the measured static capacity. All three figures clearly indicate that no unique relationship exist between soil type and damping parameters, suggesting that other mechanism control the value required for a damping factor rather than the soil type.
Figure 6. Tip soil conditions versus calculated Case damping coefficient, Jc, based 011 static load test results for 290 PD/LT pile-cases (Paikowsky et al. 1994).
A summary of the statistics obtained when examining the accuracy of the signal matching technique (specifically CAPWAP) based on soil type is presented in Table 4. The statistics shown are the mean and standard deviation of a normal distribution function for the ratio of the pile’s static capacity (based on Davisson’s failure criterion) to the pile capacity obtained in the CAPWAP analysis. The major categorization based on three soil types at the pile’s tip show no significant differeiices between clay arid till vs. sand and silt. Although the cases for piles found on rock provide different values, the
numbers are based on a small subset of 15 pilecases, compared to 100 and 265 pile-cases for the other categories. Table 4 provides further examination of time of driving and driving resistances as subsets of the soil type categorization. Two sets are examined based on the time of driving: EOD and last BOR, (i.e. in the case of niultiple restrikes, only the last restrike is considered for the analysis). The presented results suggest that the time of driving significantly affect the performance of the CAPWAP prediction, regardless of soil type. The mean values for the BOR sets are closer to one, while the mean values for the EOD are closer to two. The coefficients of variation (the ratio of the standard deviation to the mean) show values of 0.33 and 0.39 for BOR, while the EOD ratios are 0.55 and 0.85, which indicates the existence of substantial scatter. Further evaluation of the records is carried out on the basis of driving resistance. The division between cases for which the driving resistance is smaller or greater than SBPcm, examines the aforementioned notion of refusal and the expected accuracy of the dynamic methods. The results shown in Table 4 indicate that the cases for which the driving resistance is smaller than 5BPcni result in less accurate analyses with larger scatter, compared to the cases for which the driving resistance was above SBPcm. Though driving resistance seems to be an important factor, clear understanding of its influence on the accuracy of the dynamic methods calls for additional investigation, which is presented in a following section.
291
Table 4. CAPWAP analysis based on soil type categorization. Clay & Till
Scind & Silt
Rock
Meat1
1.352
1.5 17
0.930
Standard Deviatioii
0.723
1.085
0.172
Niiniher of Cases
100
265
Time of Driving
EOD
15
BOR(last)
EOD
BOR(1ast)
EOD
B OR(last)
Meotz
1.634
1.133
2.068
1.193
0.968
0.925
Sianda rd Deviation
0.899
0.444
1.765
0.391
0.132
0.203
Nziivher of Cuses
45
40
77
116
7
7
Blolv Count (BPcni)
<5
25
<5
25
<5
25
<5
25
<5
25
< j
25
Meaii
1.127
1.725
0.750
1.315
2.191
1.458
1.126
1.283
1.070
0.952
0.671
0.879
Sia 11 durd Deviation
0.637
0.807
0.241
1.160
1.901
0.512
0.386
0.355
-----
0.136
0.163
0.230
Nuinher ojCases
35
35
11
10
64
13
74
40
1
6
3
3
Notes:
1. EOD = End of Driving 2. BOR(last) = Beginning of the last rcstrike
3. BI’cni
=
Blows uer centiineter
Table 5. Summary of static and dynamic based capacity gain data sets (Paikowsky et al. 1996).
In summary, while the performance of CAPWAP is not well correlated to soil type, other factors associated with soil type may be important (e.g. low driving resistance in soft cohesive soils or gain of capacity with time) but soil type itself does not appear to be important. The data of Table 4 suggests that time of driving must be considered and driving resistance need to be further examined.
Static Daia Sets L TT atid
C,,
t,s*
Dytiamic Data Set PD/LTT CEld
tX**
All Data
c,,
t,5**
No.
15
5
7
6
22
11
8.3 The efect of Time
Avg
0.389
385.0
0.348
21.3
0.376
186.6
Penetration of piles in fine-grained soils causes compression and disturbance, resulting in soil resistance during driving that differs from the long-term pile capacity. Although factors such as thixotropy and aging contribute to this phenomenon, the most significant cause for gain of capacity with time is associated with the migration of pore water. Measurements carried out on a model (Paikowsky & Hart 1998) and full-scale piles (Paikowsky & Hajduk 1999, 2000) show that pore pressure at magnitudes similar to the total soil pressures create in clays around the pile’s shaft zones of about zero effective stresses, resulting in almost a complete loss of frictional resistance. Pailtowsky et al. (1995, 1996) examined the static and dynamic gain of capacity with time based on radial consolidation; a normalization process was followed, allowing for a comparison between different pile sizes. Table 5 presents a summary of parameters describing the pile capacity gain with time based on static and dynamic testing, The presented data shows that while the rate of capacity gain (normalized to the maximum capacity vs. time
Stdev.
0.1 19
226.3
0.068
7.9
0.106
237.9
* *‘F
closed-ended piles only excluding the case from Denmark
on log scale) is similar based on static and dynamic measurements (Cg, = 0.389, Cgtd= 0.348) the associated time for achieving 75% of the maximum capacity (normalized for all piles to 254mm (lft) diameter) is about 20 times greater. In other words, dynamic testing (namely CAPWAP) while following the physical behavior of capacity gain, exhibit this gain much faster than the actual gain monitored by the static load test results. The ramifications of these conclusions are: (i) actual gain of capacity is much slower than that exhibited by the dynamic methods, (ii) scheduling of construction or testing based on capacity gain should consider the reason for time evaluation (i.e. actual loading in construction or dynamic testing as part of quality control), and (iii) at present, the dynamic methods evaluation should concentrate on the long term pile capacity.
292
8.4 The Effect ojSoil Motion
8.4.1 Overview Paikowsky and Chernauskas (1 996) had shown that the stationary soil assumption, under which the soillpile interaction models were developed, does not reflect the physical plienomenon that occLirs during pile driving. The use of pseudo viscous damping serves as a inechanisin to absorb energy, but, as it does not reflect the actual phenomenon, its correlation to physical properties (e.g. soil type) or time of driving cannot be achieved. If the motion of the displaced soil is a major factor contributing to the energy loss during driving, a substantial portion of the dynaniic resistance should be a function of two parameters: (i) mass/volume of the displaced soil that is a function of the pile geometry, namely, sillall vs. large displacement piles, and (ii) acceleration of the displaced soil, (especially at the tip) that can be conveniently examined as a function of the driving resi st an ce . 8.4.2 Soil Acceleration The energy loss through the work performed by the displaced soil inass at the tip is directly related to the acceleration of this inass. The detailed evaluation of the soil's motion at the tip is beyond the scope of the present paper and is described by Holscher (1995), Holscher and Barends (1996), and Hajduk et al. (2000). The indirect evaluation of these accelerations can be performed through the driving resistance, which is the measure of the pile's final displacement under each hammer blow. With low driving resistance, high acceleration and velocity (i.e. free-end analogy) are developed at the tip. In the case of high driving resistance (hard driving), there is small acceleration at the tip, resulting in little, if any, mobilization of the soil inass beyond a radiating elastic wave. The corresponding energy loss due to soil motion is, tlierefore, small. To evaluate the blow count that identifies the transition between 'easy' driving with high soil acceleration values and 'hard' driving with low values, the ratio between the static capacity and the CAPWAP prediction (Kw) is presented in Figure 7 against the measured blow count values. The data were also separated into intervals of 8 BPlOcm (2BPI) with the ineaii and standard deviation of each group graphed as a point and an error bar against the mid point blow count of the interval. For example, for driving resistance between 0 and 8BP 1 Ocm there were 42 cases with a mean of 2.506 and a standard deviation of 2.217 plotted at the center of the interval, i.e. at 4BPlOcm. The data presented in Figure 7 shows that for the first two intervals (up to l6BPI Ocm) the predicted capacity was substantially lower than for all other intervals with a significantly higher scatter. After approximately 16 blows per 10cm, the mean and standard deviation of the indi-
Figure 7. KSW versiis blow count for all pile-cases in PD/LT2000.
vidual intervals fall well within the range of all cases. The boundary of the dynamic method evaluation based on driving resistance was defined, therefore, as 16BPlOcm (4BPI).
8.4.3 Soil Displacerneiit and Pile Areu Ratio The volume of the displaced soil is identical to the volume of the penetrating pile, except when pile plugging takes place (Paikowsky & Whitman 1990). The piles, therefore, can be classified as small (e.g. H and unplugged open pipe) and large (e.g. closed pipe and concrete) displaceinent piles. Additional classification of open-pipe piles can take place according to a tip-area ratio similar to that used for soil samplers (Paikowsky et al. 1989). As most soil displacement takes place at the tip area, the classification of piles can be better served by looking at the ratio between the pile's embedded surface area and the area of the pile tip (Paikowsky et al. 1994): A = & = Szirface area in contact with soil (1 1) A,,,> Area of pile tip
According to this ratio, a pile that is traditionally referred to as a “large displacement” pile can behavc like a small displacement pile if it is driven deep enough. Because the frictional resistance of a pile increases as the pile skin area in contact with soil increases, the effect of the soil niobilizcd at the tip decreases. As the pile’s embedded surface area and the skin friction increases, the energy losses resulting from the mobilization of the soil mass at the pile tip will decrease relative to the energy losses along the side of the pile. For example, the area ratio for cylindrical (closed-end) piles is:
In which D = penetration depth and R = pile radius. For the same pile diameter, this area ratio increases linearly with depth. For example, a 356-mm (14inch) diameter pile will have an arca ratio of 69 at the depth of 6.1 m (20 ft) and an area ratio of 360 at the depth of 32 m (105 ft). It is clear that the cffect of soil inertia at the tip in the second case will be substantially snialler than in the first case, and the pile may be classified as a “small displacement pile”. A quantitative boundary of AR = 350 between “small” and “largc” displacement piles was proposed by Paikowsky et al. (1994). Figure 8 presents the relationship between the arca ratio and the ratio of the static capacity over CAPWAP prediction (Ksw) for 382 cases. The data were separated into area ratio intervals of 175, with the mean and standard deviation of each group graphed as a point and an error bar against the mid point area ratio of the interval. For example, for the 139 piles with area ratio between 175 to 350, the mean was 1.656 and the standard deviation was 1.425 plotted at thc center of the interval (i.c. at tlie arca ratio of 262.5). Figure 8 suggests that piles with area ratios smaller than 350 present less accurate predictions and larger scatters compared to the mean and the scatter of all cases. Above an area ratio of 350, the mean and standard deviation of the individual intervals fall well within the range of all cases. As the driving resistance may also affect the data in Figure 8, the influence of the area ratio was fkrther examined for piles with a driving resistance greater than 16 BPlOcm at the EOD. The 71 cases answering to these criteria are presented in Figure 9. The data in Figure 9 suggests again that when excluding the easy driving resistance effects, the accuracy of tlie dynamic predictions are still lower with a largc scatter for pilcs with area ratios smaller than 350. The boundary of AR = 350 between “small” and “large” displacement piles was therefore confirmed based on database PD/LT2000.
294
Table 6. Summary of the Performance of'the Dynamic Methods.
.z -5
Fh'WA
c
Illoclified
EOD
GLlles
EOD
BI.Ct.
EOD
1WAP
Notes:
16BPl O c i ~
EOD
=
AR ENR I<SS
=
= =
135
1.073
0.573
0.534
0.45
0.35
0.27
62
1.306
0.643
0.492
0.60
0.47
0.37
99
1.656
0.724
0.48
0.34
0.25
End of Driving Area Ratio Engineering News Record Equation Ratio ofthe Static Load Test Results to the predicted capacity
9 TIHE PERFORMANCE OF THE DYNAMIC METIHODS CATEGORIZED ACCORDING TO THE CONTROLLING PARAMETERS 9.1 The Amlysed Cases The time of driving, driving resistance, aiid area ratio proved to be the major controlling parameters of the dynamic niethods. To facilitate the codes' separation between design and construction, the database was organized into these categories, followed by subcategories of methods that use aiid do not use dy ii aiii1c in eas uremeii t s, with subsets fo 11o w i 11g the controlling parameters. Figure 10 presents for the aiialyzed subsets, the number of cases in the set aiid the normal distribution mean aiid standard deviation. WEAP is utilized in the design stage. The analysis is carried out for driving stress evaluation, leading to a load fs-ctor, not included i n this paper. The use of the method for the evaluation of pile capacity was examined through WEAP results for default input values and the blow count at the EOD compared to the static load test results. The presented data was provided by GRL Inc. (Haiinigan et al., 1996). For the construction category, tlie cases without dynamic nieasureiiients evaluate the dynamic equations, specifically tlie ENR, the Gates, and the
1.199
BOR B1 Ct BP 1 Ociii COV
= = = =
Beginning of Restrike Blow Count Blows per 1Ocni Coefficient of Variation
FIHWA version of the Gates equation. The cases with dynamic measurements evaluates both, CAPWAP and Energy Approach methods. The dynamic methods are brolam down into subsets based on time of driving, driving resistance, and area ratios. Judgment arid statistics guidelines were used for the inclusion or exclusion of a few cases. For example, extreme CAPWAP under-predictions (beyond 2 S.D.) were observed at the EOD in one site. All cases 111chided easy driving and large area ratios and if included in the general population of data, the EOD statistics would have become 1.861 i 1.483 (niean+lS.D.). This site is included only in the subcategory of blow count < 16 BP 1 Ocin and AR < 350.
9.2 The critical cases The statistical analyses presented in Figure 10 allows for the identification of the critical cases that require calibration aiid developnient into resistance factors. For exaiiiple, the CAPWAP cases include (i) all data, (ii) EOD, (iii) BOR, and (iv) the worst combination of soil motion effect (Blow count < 16 BPlOciii aiid AR < 350). Table 6 presents a summary of the major categories of the dynamic methods that are identified from 295
I I
I I
Dynamic, Analysis
Construction
Design
< 16 BPlOcm 2 16 BPlOciii 1.306 f 0.643 0.876 0.4 I9 No. = 62 No. = 73
*
< 16 BP 1Ocm 2 16 BP 1Ocm 0.929 f 0.688 0.809 f 0.290 No. = 32 No. = 127
I I
i Signal Matching (CAPWAP) 1.368 f 0.620 No. = 377 I
1.626 + 0.797
Field Evaluation Energy Approach 0.894 f 0367 N A
BOR (last) 1.158 f 0.393 No. = 162
= 171 I
1
I BOR (last) 0.785 f 0.290 No. = 153
1.084 f 0.43 1
< 16 BPIOciii 2 16 BPIOcni 0.830 f 0.352 0.775 f 0.274 No. = 29 No. = 124 I
AR < 350 AR 2 350 1.181 1.717 k 0.841 i 0.468 No. = 37 No. = 34
AR < 350 AR 2 350 1 . 1 10 1.178 f 0.379 i 0.303 No. = 83 No. = 47
AR < 350 AR L 350 1.054 0.926 4 0.459 f 0.320 No. = 39 No. = 34
I
AR < 350 AR 2 350 0.736 0.851 f 0.249 f 0.305 No. = 82 No. = 42
"All values represent the ratio of the static capacity based 011 Davisson's failure criterion over the dynamic methods prediction, mean i: 1 S.D.
Figure 10. Flow chart presenting the sub-grouping of the dynamic analyses according to the controlling parameters and the resulting, statistical parameters for a normal distribution function.
296
Figure 18. Iiistogram and Frequency Distributions for EOD default value GRLWEAP pile-cases, (99),data provided by GRL (see Mannigaii et al. 1996).
ugh 18 present tlie data along with the calculated iioi-inal and lognormal distributions. 9.3 Ititetviediate Conclusions The data presented in Table 6 and Figures 1 1 through 1 8 lead to several preliminary conclusions: (i) The sigiial matching procedure generally underpredicts the pile’s capacity. Tlie method performs very well for the BOR (last restrike) cases. (ii) The simple Energy Approach provides excellent prediction for evaluating the pile’s capacity during driving (EOD). (iii) The above suggests that construction delays due to restrike and costly signal matching analyses need to be examined iii light of capacity time dependency and economical factors. (iv) The FHWA modified Gates equation provides very reasonable predictions for evaluating the pile’s capacity when dynamic methods are not carried out. (v) A reasonably good match exists for most cases between the calculated lognormal distribution and the observed data. For this reason lognormal functions were used in calibration procedures of the resistance factors. (vi) The traditional factors of safety presented in Table 1 can now be evaluated in light of the available data. For example, the coefficient of variation for the WEAP analysis is 0.724, which practically means that the method is unsuitable for the purpose of capacity prediction. The reduction in the factor of safety from 3.50 to 2.75 in Table 1 when adding WEAP analysis to static calculations is therefore unfounded. Tlie use of unspecified CAPWAP (general case) again does not justify the reduction of the factor of safety to 2.25 even though the average prediction is conservative and hence the ineaii case with a FS = 2.25 relates to an over prediction ratio of 3.1 (1.368 x 2.25). hi comparison, the use of FS = 2.25 with a specified CAPWAP at tlie BOR is reasonable and is associated with an acceptable probability of failure for single pile application (approximately 3 cases out of 162, i.e. 1.85%, see Figure 15).
Figure 17. Histogram and Frequency Distributions for all (384) FH WA modified Gates equation pilc-cases in PDiLT2000.
Figure 10 as the cases that require calibration for a resistance factor. Histogram and frequency distributions were prepared for tlie identified critical cases in order to examine the match between the actual data and the probability distribution functions. Figures 1 1 thro298
10 RESISTANCE FACTORS 10.1Metlzodology
Following Ayyub and Assakkaf (1999) and Ayyub et al. (2000) the present project calibrates LRFD partial safety factors using the First-Order Reliability Method (FORM). FORM can be used to assess the reliability of a pile with respect to specified limit states, and provides a means for calculating partial safety factors # and y, for resistance and loads, respectively, against target reliability levels, PO. FORM requires only the first and second moment information on resistances and loads (i.e. means and standard deviations), and an assumption of distribution type (e.g. normal, logiiormal, etc.). The franiework of the calibration process is presented in Figure 19. In design practice, there are usually two types of limit states: ultimate limit states and serviceability limit state. Each can be represented by a perforinance function of the foi-in:
in which X is a vector of basic random variables ( X I , X?,. . ., X,,) for strength and loads. The perfoniiance function g(X) is sometimes called the limit state ftinction. It relates the random variables for the limit-state of interest. The limit state is defined when g ( x ) = 0, and therefore, failure occurs when g(x> < 0 (see Figure 19). The reliability index P is defined as the distance from the origin of the space of basic random variables (XI,X Z , . . , X,,)to the failure surface at the most probable point. The most probable failure point is that point on the limit state fLinction at which the probability density of the basic random variables is greatest. This is also called the design point. This relationship can also be used to back calculate representative values of the reliability index p from the current design practice. In developing design code provisions for piles, it is necessary to follow the current design practice to ensure consistent levels of reliability over various pile types. Calibrations of existing design codes are needed to make the new design formats as simple as possible and to put them in a form that is familiar to users or designers. For a given reliability index p and probability characteristics for the resistance and load effects, the partial safety factors determined by the FORM approach might be different for different failure modes for the same or differing component. For this reason, calibration of the calculated partial safety factors (PSF’s) is important in order to maintain the same values for all loads at different failure modes. In the case of geoteclinical codes, the calibration of resistance factors is performed for a set of load factors already specific in the structural code. Thus, the load factors are fixed. In this case, the fol-
299
lowing algorithm is used to determine resistance factors: (1) For a given value of the reliability index p, probability distributions and moments of the load variables, and tlie coefficient of variation for the resistance, compute mean resistance R using FORM. (2) With the mean value for R computed in step 1, the partial safety factor 4 is revised as:
are tlie mean values of tlie loads where p~.,aiid and strength variables, respectively and y,,i = 1, 2,.. ., ii, are the given set of load factors.
10.2 Cnlctrlated factors Figure 20 presents examples of the resistance factors calculated based on the above procedure for the general CAPWAP and Eiiergy Approach cases. As no exact target reliabjlity has yet been established, the factors were evaluated for target reliability values of 2, 2.5, and 3.0 associated with probability of failure values of 2.3%, 0.62%, and 0.14%, respectively. The factors were evaluated using load factors of 1.25 and 1.75 for Dead Load (DL) aiid Live Load (LL), respectively, and for DL to LL ratios ranging from 1 to 4. The obtained results presented in Figure 20 suggest very little sensitivity to tlie DL to LL ratio. A parametric study was carried out for a generic coefficient of variation of 0.40 and dead to live load ratios ranging froin 1 to 10. No significant influence of the dead to live load ratio on the calculated resistance factors was found. The large dead to live ratios represent a wide possibility of bridge construction, typically associated with very long bridge spans. A summary of the calculated resistance factors for all tlirec targct reliability levels for the idcntified dynamic methods is provided in Table 6.
Figure 20. Calculated resistance factors for tlie CAPWAP and Energy Approach general cases showing the influence of the dead to live load ratio.
fall into the general Eiiergy Approach case in Table 7. As a reasonable probability of failure for a siiiglc pile can be estimated to be approximately 1% to 2% (considering tlie redundancy of pile groups), Table 7 presents representative resistance factors obtained from tlie average calculated from the target reliability level (p) of 2 and 2.5. These are prelimiiiary values demonstrating the inethodology and by iio nicaiis represent the recoinmended or final approved AASHTO specifications.
111.4Evaluution of the dyriarnic methods efliciericy
10 3 Preliniirzniy Resr.s!aiice Factors Table 6 was developed based on the cases presented
The resistaiice factors alone do not provide a measure for the evaluation of tlie efiicieiicy of tlie dynamic methods. Such efficiency can be evaluated through the bias factor (mean of the ratio of the measured over predicted), its coefficient of variation (see Table 6) or the ratio of the resistance factor to the bias factor, i.e. flnieaii Ksx, as proposed by McVay et al. (2000). This ratio is provided for the final selected cases iii Table 7. The efficiency values in Table 7 suggest that overall the higher efficiency is obtained by the signal matching analyses for the last restrike, followed by the Energy Approach at the eiid of driving (0.581 vs. 0.507).
Figure 10 and using the resistance factors presented in Table 6, the critical dyiiaiiiic cases were reevaluated. The selected methods, their important categories and reconimended preliiiiinary factors are presented 111 Table 7. When reevaluating tlie calculated resistance factors and developing Table 7, the ENR equation was omitted due to very poor performalice The cases within a specific dyiiaiiiic method that had similar resistance factors were combined iiito one category. For example, the ROR and general Eiiergy Approach cases had the same calculated resistailcc factors so the BOR cases would in
300
Table 7. Preliminary resistance factors for the critical dynamic cascs.
Case
Method
4/A4em KJ,V
0.54
0.61
0.446
BC
0'52
0.35
0.44
0.170
BOX
0.73
0.6 1
0.67
0.581
General
0.48
0.39
0.44
0.492
EOD
0.60
0.49
0.55
0.507
Gates
General
0.85
0.67
0.76
0.425
F H WA iiiodified
Geiieral
0.42
0.33
0.38
0.404
EOD
0.48
0.34
0.41
0.248
Matcliiiig
Dy ainic Mea.siii.eiiieiits
Elieigy Approacli
~
Repi.eseiztalive Resistaiice juctor, #
0.68
Geiieral
Dynaiiiic ~qiiatioiis
P=
EODc3.50
Sigiiul
,
Resistiince.fuctoi., 4 2.5 p i = 2.3% [ J i = 0.6%
f l = 2.0
WEAP
.
on restrikes seem to be marginal compared to the Energy Approach at the End of Driving (EOD). These conclusions though representative of most cases, cannot be based on statistical data alone. For example, sites that exhibit a significant but highly variable setup may economically justify consistent and long-term restrikes along with signal matching analysis. Tlie field application of the Energy Approach provides an exceptionally efficient evaluation of' pile capacity during driving. The framework for the development of resistance factors as part of the LRFD methodology seem to facilitate a design wliicli is better suitable for geoteclinical applications. The presented work is only an initial stage in that direction. A more complete code based on LRFD needs to consider factors associated with subsurface variability, site-specific technology and previous experience, as well as amount and type of testing during construction.
11 CONCLUSIONS Based on tlie presented data the following conclusions are derived: The compilation of a large database allows for tlie evaluation of the dynamic methods, tlie examination of tlie Working Stress Design (WSD) methodology (e.g. validity of tlie assigned factors of safety), and tlie development of new methodologies such as the Load and Resistance Factor Design (LRFD). The dyiianiic methods performance is controlled by the time of' driving and soil inertia, wliicli in turn is controlled by tlie driving resistance and the ratio of the soil displaced by the pile's tip to the area of the soil along the shaft of tlie pile. Tlie most commonly known dynamic equation, the Engineering News Record (ENR) equation, is shown to be completely unreliable and unreasonable for use. In contrast, tlie Gates equation and its variation, modified by the Federal Highway Administration (FHWA), seen7 to provide a reasonable assessment of tlie pile's capacity considering tlie absence of dynamic measurements. The wave equation analysis performs poorly when used for pile capacity evaluation. However, this conclusion should not be niistakcn with the importance of the wave equation analysis during the design stage. Tlie drivability study and pile stress analysis often detei-niine tlie pile type, geometry and tlie adequacy of the proposed equipment. Signal matching techniques (e.g. CAPWAP) prove to be most reliable for long-term restrike measurements. However, when evaluated on efficiency, the application of the signal matching
ACKNOWLEDGEMENTS The presented research was sponsored by the American Association of State Highway and Transportation Officials (AASHTO), under project 24- 17, in cooperation with the Federal Highway Administration (FHWA). Tlie panel of the research project and Mr. David Beal of the NCHRP are acknowledged for their stimulating demands. Messrs. Jerry DiMaggio, A1 DiMillio and Car1 Ealy of the FHWA are acknowledged for their interest, concern and support. Drs. Gregory Baecher and Bilal Ayyub from tlie University of Maryland contributed to sections 3.1 through 3.4, section 10.1 and performed tlie 301
calculations of the presented resistance factors. Drs. Mile McVay and Bjorii Birgissoii of the University of Florida and Dr. Ching Kuo of GeoStructures contributed to a questionairre resulting in the information presented in section 6. Dr. Frank Rausche of Goble, Rausche, Likins (GRL) and Associates provided the data pertaining to the evaluation of GRLWEAP. DISCLAIMER The research presented here is part of an ongoing project, which is aimed at rewriting the AASHTO Deep Foundation Specifications for the year 200 1. The presented parameters are preliminary and do not reflect in aiiyn'ay the recoiiiiiiended or approved parameters. The opinions and conclusions expressed or implied in the paper are those of the eiitities/iiidividuals performing the research and are not necessarily those of the Transportation Research Board, the National Research Council, tlie Federal IHighway Administration, tlie American Association of State Highway and Transportation Officials, or the individual states participating in the National Cooperative Highway Research Program. IGFERENCES AASHTO, 1994. LRFD Bridge ntid Consrruclion Sjiec(/icirtio17.s. American Association of State Highway and Transportat i o n 0ffic i a 1s, Washington, DC. AASHTO, 1 997. Sfaiirlwci Specif;cutioti.s.fbl- Higlz~vuyBridges. American Association of State Highway and Transportation Officials, Washington, DC., 16"' Edition (1996 with 1997 interims). AASHTO, 1999. LRFD Bridge atid Cotisr~uclio~i Speci/icntioizs. American Association of State Highway and Transportation Officials, Washington, DC. ACI, 1995, Builriiizg Code Requireiiients ,for Reiufol-ced Coticrefe. American Concrete Institute, Detroit. AISC, 1994. Load and Resistance Factor Design. Muiiuul of' Steel Cotistt*iicfioii, American Institute of Steel Construction, Chicago, IL. API, 1989. Draft Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms - Load and Resistance Factor Design. API RP2A-LRFD, American Petroleum Institute, Dallas, TX. ASCE, 1993. Minimum Design Loads for Buildings and Other Structures. ASCE 7-93 (formerly ANSI A58.1). AUSTROADS, 1992. A USTROADS Bridge Design Code, National Office, AUSTROADS, Surry Hills, NSW, Australia. Ayyub, B.M. 1999. Cost-Benefit Analysis of Risk-Infornied Design of Structural Systems. American Society of Mechanical Engineers, the Research Committee on Risk Technology, Washington, DC. Ayyub, B.& Assakkaf, I. 1999. LRFD Rules for Naval Surface Ship Structures: Reliability-Based Load and Resistance Factor Design Rules. Naval Surface Warfare Center, Navy. Carderock Division, U.S. Ayyub, B., Assakkaf, I., & Atua, K. 199Sa. Development of LRFD Rules for Naval Surface Ship Structures: Reliability-Based Load and Resistance Factor Design Rules, Part Ill - Stiffened and Gross Panels. Naval Surface Warfare Center, Carderock Division, US.Navy.
Ayyub, B., Assakkaf, I. & Atua, K. 2000. Reliability-Based Load and Resistance Factor Design (LRFD) of Hull Girders for Surface Ships. N u i d Engineers Joui.1~~1, ASNE, May 2000. Ayyub, B., Assakkaf, I., Atua, K., Engle, A., Iiess, P., Icaraszewski, Z., Kihl, D., Melton, W., Sielski, R., Sieve, M., Waldman, J . , 61 White, G. 1998b. Reliability-based Design of Ship Structures: Current Practice and Emerging Technologies. T&R Report R-53, Society of Naval Architects and Marine Engineers. Barker, R., Duncan, J.: Rojiani, K., Ooi, P., Tan, C., & Kini. S. 1991 . A4anuals j b i . (lie Design of Bridge Foutidations. NCHRP Report 343. Transportation Research Board, National Research Council, Washington, DC. Butler, H.D., & Hoy, H.E. 1977. Users Munuu1,foi~the T a u s Q u i ~ l i - L ~Metlzocl,ji,r ~d Fouiiclatioiz Load Testing. Federal Highway Administration, Office of Development, Report No. FHWA-IP-77-8, Washington, DC. Canadian Geotechnical Society. 1992. Caiiadiuti Founckiliot? Etigi~eer-ingA4~irziia1,3'd Edition. Bi-Tech publishers, Ltd., Richmond, British Columbia, Canada. CIRIA 63. 1977. Rationalization of Safety and Serviceability Factors in Structural Codes. Construction Industry Research and Information Association, SWIP 3AU, Report 63, London. Chellis, R.D. 1961. Pile Foundations, McGraw M i l l 2'ld ed. NY. Davisson, M.T. 1972. Iiigh Capacity Piles. P~oceeditigs,Soil A4ecliutiics Lecficre Series 011 Itiiiovatioris it7 Fourzdafiotz Coristrucfioii. American Society of Civil Engineers, lllinois Section, Chicago, 8 1- 1 12. DeBeer, E.E. 1970. I'roefondervindellijke bijdrage tot de studie van het grandsdraagvermogeii van zand onder funderinger op staal. English version, Geofech~ique,Vol. 20, No. 4, 387-41 1 . Ellingwood, B., Galambos, T., MacGregor, J., & Cornell C. 1980. Developnient of a Probability-Based Load Criterion for American National A58. National Bureau of Standards Publication 577, Washington, DC. Ellingwood, B., Galambos, T., MacGregor, J., & Coniell, C. 1982a. Probability Based Load Criteria - Assessment of Cui-rent Design Practices. Jour-nal of [lie S r t u c t i d Division, ASCE, Vol. 108, No. ST5, 959-977. Ellingwood, B., Galambos, T., MacGregor, J., & Cornell, C. 1982b. Probability Based Load Criteria - Load Factors and Load Combinations. Jourmd 03''the Stnictut.al Division, ASCE, Vol. 108, ST5, 978-997. FIHWA, 1988. FHWA Guide Specifications for Driven Piles. Federal Highway Administration. Freudenthal, A.M. 1947. Safety of Structures. T~ansactionsof' the ASCE. Vol. 1 12, 125-1 80. Galambos, T. & Ravindra, M. 1978. Properties of Steel for use i n LRFD. Joul-tin1 U / S t ~ u c t w a Eiigiiieei.ing. l ASCE, Vol. 104, NO. 9, 1459-3468. Gates 1957. Empirical Formula for Predicting Pile Bearing Capacity. Civil Eiigiiiewirig, Vol. 27, No. 3, 65-66. Goble, G . 1999. Geotechnical Related Development and 1111plementation of Load and Resistance Factor Design (LRFD) Methods. NCHRP Synthesis 276, Transportation Research Board, National Research Council, National Academy Press, Washington, DC. Goble, G . , Likens, G.,& Rausche, F. 1970. Dynamic Studies on the Bearing Capacity of Piles - Phase 111, Report No. 48. Division of Solid A4echunic.s, Sfruclures, atid Mechanical Desigiz. Case Western Reserve University. Goble, G. & Rausche, F. 1976. Wave Equation Analysis of Pile Driving-WEAP Program. Vol. 1-4, FHWA #IP-76- 14.1 through #IP-76- 14.4. GRL. 1999. Pile-Driviizg Ana/yzet., PAK Users A4unual. Goble, Rauschc, Likins and Associates, Inc.
302
GTR. 1997. Dynamic Pile Testing Report, Central Artery/Tunnel Project C07D2, I-90/Airport Ititerchange Arrivals Tunnel - Phase I, East Boston, Massachusetts. Geosciences Testing and Research, Inc. North Chelmsford, MA. GTR. 1998. Dynamic Pile Testing Report, Central Arteiy/Tunnel Project C07D2, I-9O/Airport Intel-clznnge Toll Plaza, East Boston, Massachusetts. Geosciences Testing and Research, Inc., North Chelmsford, MA. Hajduk, E., Paikowsky, S., Holscher, P., & Barends, F. 2000. Accelerations of a Driven Pile and the Surrounding Soil. Proceedings, 6'" International Conjerence on the Applications of Stress- Wave Theory to Piles, September 11-13, 2000. S2n Paulo City, Brazil. Hamilton, J.M. & Murff, J. 1992. Selection of LRFD resistance factors for pile foundation design. Proceedings Structures Congress '92. American Society of Civil Engineers, San Antonio, ASCE, NY. Iiannigan, R., Goble, G., Thendean, G., Likins, G., & Rausche, F. 1996. Design and Construction of Driven Piles Foundations - Volume I1 (Working Draft). U.S. DOT Federal Highway Administration. Washington, DC. Hansen, B. 1953. Earth Pressure Calculation. Danish Technical Press. Copenhagen. Hansen, B. 1956. Limit Design and Safety Factors in Soil Mechanics. Bulletin No. 1, Danish Geotechnical Institute, Copenhagen. Iiansen, B. 1966. Code of Practice for Foundation Engineering. Bulletin No. 22, Danish Geotechnical Institute, Copenhagen. Holscher, P. 1995. Dynamical Response ojSaturated and Diy Soils. Delft: Delft University Press. Holscher, P. and Barends, F. 1996. In-sihi Measurement of Soil-Motion near the Toe of a Dynainically Loaded Pile. In F.C. Townsend, M. Iiussein & McVay, M. (ed), Proceedings, 5"' In tern n tion a / Coizfeimce of the AQp/icalion of Stress- Wave Theory to Piles, September 11-13, 1996. _nrla.1ld5, F1 76-36
Kulhawy, F.H. et al. 1988-1994. Reliablilty-Bascd Foundation Design of Transmission Line Structures. Report EL-5507. Electrical Power Research Institute, Palo Alto. Kulhawy, F.H. et al. 1996. Proceedings, Uncertainty in the Geological Environnzent. Madison, ASCE, NY. Long, J., Bozkurt, D. & Kerrigan, J. 1998. Value of Methods for Predicting Axial Pile Capacity. Ti~ansportation Research Board Paper No. 99-1333, 76'" Annual Meeting, Januury, 1999. Washington, DC. McVay, M., Birgisson, Bjorn., Zhang, L., Pcrez., A & Putcha S. 2000. Load and Resistance Factor Design (LRFD) for Driven Piles Using Dynamic Methods - A Florida Perspective. Geotectinicul Testing Joio-nul, ASTM, Vol. 23, NO. 1, 55-66. MHD. 1988. Standard Specifications for Highways and Bridges. Massachusetts Highway Department. Middendoip, P. & van Weel, P. 1986. Application of Characteristic Stress-Wave Method in Offshore Practice. Proceeclings oftlie 3'" Interriationu1 Coiljierence on Numerical A4ethods in Oflshore Piling. Moses, F. 1985. Implementation of a Reliability-Based API RP2A Format, Final Report. API PRAC 83-22. American Petroleum Institute. Moses, F. 1986. Development of Prelimiilary Load and Resistance Factor Design Document for Fixed Offshore Platforms, Final Report. API-PRAC 95-22, American Petroleum Institute. National Research Cot~ncil. 1995. Probabilistic Methods i n Geotechnical Engineering. Washington. National Research Council of Canada. 1977. National Building Code of Canada. Ottawa.
Nowak, A. 1993. Calibration of LRFD Bridge Design Code. Department of Civil and Environmental Engineering Report UMCE 92-25. University of Michigan, NCHRP 1233. Nowak, A. 1999. Calibration of LRFD Bridge Design Code. Department of Civil and Environmental Engineering Report UMCE 92-25. University of Michigan, NCHRP 1233. Olsen, R. & Flaate, K. 1967. Pile Driving Formulas for Friction Piles in Sand. ASCE JSMFC, Vol. 93, SM6, Nov. 1967, 279-297. Ontario Ministry of Transportation and Communication. 1992. Ontario Highway Bridge Design Code and Comnientaiy, 3'" ed. Paikowsky, S. 1982. Use of Dynamic Measurements to Predict Pile Capacity Under Local Conditions. M.Sc. Thesis, Dept. of Civil Engineering Technion-Israel Institute of Technology. Paikowsky, S. 1984. Use of Dynamic Measurements for Pile Analysis. Including PDAP-Pile-Driving Analysis Program, GZA Inc., Newton, Massachusetts. Paikowsky, S. 1995. Using Dynamic Measurements for the Capacity Evaluation of Driven Piles. Civil Engineering Practice, Journal of the Boston Society of Civil Engineers SectiodASCE. Vol. 10, No. 2, 61-76. Paikowsky, S. & Chernauskas, L. 1992. Energy Approach for Capacity Evaluation o f Driven Piles. 4'" International Coizference 011 the Application of Stress- Wave Theoty to Piles. The Hague, Netherlands. 595-601. Paikowsky, S. & Chernauskas, L. 1996. Soil Inertia and the Use of Pseudo Viscous Damping Parameters. 5'" International Coiiference on the Application of Stress- Wave Theory to Piles. Orlando, FL. 203-2 16. Paikowsky, S. & Hajduk, E. 1999. Design and Construction of an lnstruiiiented Test Pile Cluster. Research Report submitted to the Massachusetts Highway Department, September, 1999. Boston, Massachusetts. Paikowsky, S. & Hajduk, E. 2000. Theoretical Evaluation and Full Scale Field Testing Examination of Pile Capacity Gain with Time. Reseat-ch R e p o i ~to be submitted to the Massachusefts Highway Department, December, 2000. Boston, Massachusetts. Paikowsky, S. & Hart, L. 2000. Development and Field Testing of Multiple Deployment Model Pile (MDMP). Research Report to be submitted to the Federal Highway Administration, April, 2000. FHWA-RD-99- 194 Washington, DC. Paikowsky, S. & LaBelle, V. 1994. Examination of the Energy Approach for Capacity Evaluation of Driven Piles. US FHWA International Conference on Design and Consfructioii of Deep Foundations, December 6-8, 1994. Orlando, FL. Vol. 11, 1133-1 149. Paikowsky, S., LaBelle, V. & Hourani, N. 1996. Dynamic Analyses and Time Dependent Pile Capacity. 5'" Inlernafional Confer.eiice on the Application of Stress- Wave Theory to Piles. Orlando, FL. 325-339. Paikowsky, S., LaBelle, V. & Mynampaty, R. 1995. Static and Dynamic Tinie Dependent Pile Behavior. Research Report submitted to the Massachusetts Highwqy Department, November, 1995. Boston, Massachusetts. Paikowsky, S., Operstein, V., and Bachand, M. 1999. Express Method of Pile Testing by Static Cyclic Loading. Research Report subriiilted to the Mussachusetts Highwaji Departnient, October., 1999. Boston, Massachusetts. Paikowsky, S., Regan, J., and McDonnell, J. 1994. A Simplified Field Method for Capacity Evaluation of Driven Piles. FHWA Report No. FHWA-RD-94-042, September, 1994. Paikowsky, S. & Whitman, R. 1990. The Effect of Plugging on Pile Performance and Design. Canadian Geotechnical Journal, Vol. 27, No. 4,429-440.
303
Paikowsky, S., Whitman, R. & Baligh, M. 1989. A New Look at the Phenoinenon of Offshore Pile Plugging. Marine Geotechnology,Vol. 8, No. 3, 2 13-230. Rausche, F. 2000. Personal communication. February 2000. Rausche, F., Goble, G. & Likens, G. 1975. Bearing Capacity of Piles from Dynamic Measurements, Final Report. Ohio Department of Transportation, Ohio DOT-05-75. Rausche, F., Thendean, F., Abou-matar, H., Likins, G., and Goble, G. 1997. Determination of Pile Driveability and Capacity From Penctration Tests. Vol. 1-3, Final Report. FH WA-RD-96- 179, Federal Highway Administration. Siu, w., Parinii, S. & Lind, N. 1975. Practical Approach to Code Calibration. Joumal of the Structural Division. ASCE, Vol. 101, NO. ST7, 1469-1480. Smith, E. 1960. Pile Driving Analysis by the Wave Equation. Journal of Soil Mechanics atid Foundations,Aniericaii Society of Civil Engineers,August 1960. 35-61. Standards Association of Australia. 1995. Australiam Standards, Piling-Design and Installation. Homebush, NS W. Tang, W. 1993. Recent development i n geotechnical reliability. Probubilistic tiielhods in Geotechnical Eiigineeriiig. Li, K. & Lo, S-C. (Eds.), Balkema, Rotterdam. 3-28. Terzaghi, K. 1942. Discussion of the Progress Report of the Committee on the Bearing Value of Pile Foundations. Proceediiigs, ASCE. Vol. 68: 3 1 1-323. Stenersen, K. 2000. Load and Resistance Factor Design (LRFD) for Dyiiainic Analyses of Driven Piles. To be siibniittetl in Pnrlial Fiilfillimizt of the Requii-merits for. CI Muster of Scieuce Degree. University of Massachusetts Lowell, Lowell, Massachusetts. Wellington 1892. Discussion of “the Iron Wharf at Fort Monroe, VA.” By J.B. Cuncklee. Transactions, ASCE Vol. 27, paper No. 543, Aug. 1892, 129-137.
304
design codes for pile foundations - A review G.G.Goble GobLe Rausclze Likins and Associates Incorporated and Pile Dynamics Incorporated, University of Colorado, Boulder, Colo., USA
ABSTRACT: Several Load and Resistance Factor Design (LRFD) codes for geotechnical applications have been developed over the past few years. Some of these codes have covered all aspects of geotechnical design while others have been primarily concerned with foundations. The basic structure of the driven pile parts of these codes will be reviewed and critically compared, particularly with regard to the way capacity determination and dynamic testing is handled. In some cases dynamic testing is not even mentioned while in the other extreme the resistance factor is governed by the amount of dynamic testing used. Also, the resistance factors used in the various codes will be compared including a comparison of the load factors
1 INTRODUCTION
The basic strength requirement of LRFD can be stated as
The first Load and Resistance Factor Design codes (LRFD) to be used were adopted in the early 60’s in the United States and Denmark. They were developed independently and for quite different reasons and neither of the developing organizations were aware of the activity of the other. In the United States the American Concrete Institute adopted a code for the design of reinforced concrete buildings (ACI 1963) that had a LRFD format. It covered all aspects of the design of reinforced concrete elements except foundations. Foundation design was ignored except for the structural design of the reinforced concrete footing. The footing size was selected to satisfy the geotechnical requirements based on allowable stresses, or in the case of deep foundations, allowable loads. The ACI’s motivation for the change was the problem with the design of reinforced concrete elements based on an assumed linear elastic stress distribution on the cross section. In cases where reinforcement was in compression, this analysis was very conservative. The economic advantages of the new procedure were so great that the new code was widely adopted within two years. The drive for this change came from structural engineers. The code change in Denmark (Hansen 1966) was driven by geotechnical engineers. The motivation was somewhat different than in the United States and while it produced a similar the load and resistance factors, were quite different.
where (Pk is the resistance factor for the k’” failure mode, Rk is the nominal resistance in the kth failure mode, yij is the load factor and Qij the load of the ith load type in the jtil load combination. This notation will be used throughout this paper. In addition to the strength requirements covered by Equation (l), there are other requirements that must be considered. They include such behavior aspects as settlement, vibration, lateral displacement, etc. In this paper the discussion will be limited to axial pile strength since it is usually the primary design consideration. Many other factors must be considered in performing a complete design. It is usehl to review the hrther development of LRFD in the United States. The methods contained in the original ACI LRFD Code were known as ultimate strength design at the time they were adopted. After the new code appeared it was suggested that the load and resistance factors could be generated rationally using a probabilistic analysis (Cornell 1969). The resulting procedures were used in an extensive study to develop load and resistance factors (Ellingwood et a1 1980). However, the original load and resistance factors used in the ACE Code were not determined in this manner but were selected based on engineering judgement. There have 305
been some limited changes in the ACI load and resistance factors since the 1963 version but those changes have been small. They were made in the time shortly after the initial adoption. When probabilistic methods are used the statistical distribution of both the various types of loads and the strength is used to select load and resistance factors such that a desired probability of failure is achieved. The two primary problems with this approach are the lack of uniqueness when a single condition is used to generate three constants and the limited data available on the statistical distribution of the loads and the strength. It is not completely convincing that a probabilistic analysis is superior to the use of judgement in selecting load and resistance factors. The load and resistance factors of LRFD replaced the single safety factor that foundation designers have traditionally used in allowable stress design (ASD). In structural design by ASD, allowable stresses were normally used instead of a safety factor and these stresses were selected and codified based on tradition and experience. In ASD structural design, the factor of safety was not visible to the designer. But for most aspects of geotechnical design, allowable stresses were not (and cannot realistically be) established. Instead strengths were determined and a factor of safety was applied. Thus, in the geotechnical area there has been a tradition of evaluating the acceptability of a design based on a calculated strength together with a factor of safety. The factor of safety should be based on the requirements of the strength variability and load variability. When the conversion is made to LRFD, it is desirable that the design process produce designs that are similar to those produced by ASD using a safety factor. This implies that the sum of the influence of the LRFD load and resistance factors should be equivalent to the ASD factor of safety. Safety factors used in driven pile design have traditionally been selected by the geotechnical engineer based on the perceived need for safety in the pile design. This is an inappropriate procedure since the same safety factor provides a margin of safety against variability of the loads. Therefore, the structural engineer should have a voice in the selections of the safety.factor. As an illustration of this problem imagine two different design examples. The first case is the pile foundation for a large multistory reinforced concrete building where the dominant load comes from the dead load of the structure. Compare this case with the pile design for a dock structure. The dock has very little dead load but may be required to carry large live loads. In the first case there is little requirement for factor of safety on 306
the loads while in the second case the loads are highly variable and must have a large factor of safety. This paper will review and compare the load and resistance factors used in several recently adopted LRFD Codes. It should be noted that in order to meet the length limits it was necessary to simplify the presentation of the various factors. These codes are discussed in much greater detail in an NCHRP Synthesis (Goble 1999). 2 LOADFACTORS During the examination of several different LRFD codes it was observed that some of the resistance factors for a particular load combination had values varying over a considerable range among the different codes. Some of this variability may have its source in differences in the load factors between codes. To make the comparison of resistance factors easier, the load factors have been collected for the basic gravity load combination and they are given in Table 1. There are several additional load combinations in the codes but this one is usually the most important and a comparison between the values provides a good general view. Table 1. Load factors for selected codes. Code Dead load Live load AASHTO Bridge Code 1.25' 1.75 ACI 3 18-95 1.40 1.70 AISC and ANSI 577 1.20 1.60 1.201 1.40 Ontario Bridge Code Canadian Bridge Code 1.20 1.60 Eurocode2 1S O 1.35 Danish Code 1 .oo 1.30 1S O Australian Code 1.25 API Code3 1.30 1S O There is a variety of load factors for the various dead load types. The quantity 1.20 is the most common value. This is a considerable simplification of the Eurocode specification on loads and load factors. The definition of loads, load combinations, and load factors is quite complex. The code of the American Petroleum Institute is concerned with the design of offshore petroleum recovery platforms. The above brief summary shows the rather large variation of Load factors now in use in the LRFD Codes that were reviewed. The load factor on dead load ranges from 1.OO in the Danish Code to 1.40 in the ACI Code. The most common value is 1.20 but the Eurocode uses 1.35, almost as large as the ACI
Code. The range of Values for load Factors on live load extends from 1.30 in the Danish code to 1.75 in the AASHTO LRFD Bridge Code. It is quite understandable that load factors for live loads could be quite different among the various codes since the live loads could be more variable in one application than another. The same conclusion cannot be reached with regard to structural dead load. A difference of 1.00 to 1.40, a factor of 40 percent, between the Danish and the ACI Code seems quite large. 3 RESISTANCE FACTORS
Some research has.been performed on the use of rational methods in determining resistance factors for single driven piles. . Goble, et al. (1980) reported a set of resistance factors consistent with the requirements of pile design. These factors were obtained by calibration with existing factors of safety. Goble and Berger (1994) and Berger (1989) reported on resistance factors obtained using available data bases from Dennis and Olson (1983), Pennsylvania DOT (Unpublished), and Rausche et a1 (1997) together with a probabilistic analysis. These calibrations were obtained using the software developed during the National Bureau of Standards study (Ellingwood, et al 1980) and, therefore, they are related to the NBS load factors that were developed for building loads. These load factors are given in Table 1 as AISC and ANSI 577. Some of the results obtained by Berger are summarized in Table 2. A large volume of pile capacity data has been assembled since the work of Berger, and it would be desirable to repeat this analysis with the new data. In the work of Berger the accepted probability of failure was 3.0 standard deviations from the mean. This quantity is known as the safety index.
3 . I AASHTO LRFD Bridge Code (199 7)
Resistance factors for the soil strength limit state for driven piles are tabulated in the LRFD Bridge Code and that information is summarized here as Table 3 and 4. In this table h, is a multplier defined by the installation control method. It is applied to a quantity given by method of capacity determination. The use of these resistance factors has proven to be unclear to foundation designers. Consider some examples: Suppose that the capacity is determined by the SPT-method at a site with piles having end bearing and friction in sand. The specified resistance factor is 0.45 L. If a bearing graph is used for verifying capacity during driving, h, is 0.85, giving a resistance factor of 0.38, a value associated with a factor of safety of about 3.7, a considerably larger value than would be used in current practice. In the same case where PDA testing with static load test, the resistance factor is 0.45, a value associated with a factor of safety of about 3.0 again much greater than is normally used. Table 3. Resistance factors from the AASHTO LRFD Bridge Code Method/Soil/Condition Resistance factor Skin friction: Clay 0.70 a, Alpha method 0.50 a, Beta method 0.55 A, Lambda method End bearing: Clay 0.70 A, Rock 0.50 a, Skin friction and end bearing: Sand SPT-method 0.45 a, CPT-method 0.55 A V Wave equation 0.65 a, Load test 0.80 a,
Table 2. Resistance factors for a safety index of 3.0 Method Resistance factor Static analysis 0.42 ENR 0.42 Wave equation 0.50 CAPWAP 0.73
Table 4. AASHTO h, values Installation control methods Pile driving formulas Wave equation PDA testing - 2% to 5% of piles PDA testing - 2% to 5% of piles with Static test PDA testing - 2% to 5% of piles with Signal matching PDA testing - 10% to 70% of piles
Barker et a1 (1991) generated the resistance factors for the static analysis methods contained in the AASHTO LRFD Bridge Code (1994) using the rational probabilistic approach on available estimates of basic soil property variability. This approach does not include the model variability, site variability, or the fact that other means are normally used to determine capacity. They do not report any calibration analyses of the other nominal strength determination methods nor do they mention the use of available pile capacity data bases.
A V
0.80 0.85 0.90 1.oo 1 .oo 1.oo
If the SPT-method were used together with wave equation analysis and static load test, then what should be used for the multiplier on L . There is no guidance in the text of the LRFD Bridge Code. Which resistance factor should be selected?
307
The concept of having the resistance factor related to the static analysis method is inconsistent with current practice.
of elasticity of the pile material, C is the velocity of stress wave propagation and the other two terms are not defined. The source of this approach is not given.
3.2 Ontario Bridge Code (1992)
The Ontario Highway Bridge Code defines the limit states for strength and specifies resistance factors for deep foundations subjected to axial compression load as given in Table 5. The resistance factors seem to be quite small when compared with the LRFD Bridge Code. There is no discussion of the methods used to obtain these values. Table 5. Ontario Bridge Code resistance factors for deep foundations Axial load Factor Static analysis 0.4 Static test 0.6 Dynamic analysis 0.4 PDA test 0.5 3.3 Canadian Bridge Code (I 988) The design requirements given in the Canadian Bridge Code are quite brief. The soil limit state is dealt with in three separate categories. For load tests the resistance factor is specified as 0.5 for routine testing and 0.6 for high-level testing and the two levels of testing are defined. However, the definitions are quite qualitative. If dynamic testing is used the resistance factor is specified as 0.4 for routine analysis and 0.5 for analysis based on parameters obtained from dynamic field measurements.
3.4 Florida DOT (1997) The resistance factors recently developed by the Florida DOT in a calibration study are given in Table 7. This presentation is clear and presents a good representation of the Florida practice, one of the most modern in the United States. Table 7. Florida DOT (1 997) resistance factors Design methodology Resistance factor SPT97 0.65 PDA 0.65 Wave equation analysis 0.35 Static load testing 0.75 The resistance factor for wave equation analysis seems to be rather small when compared with the value for SPT 97 a static analysis procedure. However, the Florida DOT follows a design process where the static analysis is used only for determination of estimated pile length for bidding. It is interesting that a resistance factor is used in this application. This is the first presentation of such an approach and it merits firther study. If data is collected on the error in the length estimate on each job it would be possible to adjust the resistance factor to arrive at the best possible length prediction. 3.5 Eurocode (I 994)
Table 6. Canadian Bridge Code (1988) resistance factors Type of unit Resistance factors 0.4 Reinforced concrete Cast-in-place concrete 0.4 Expanded-base concrete 0.4 Prestressed concrete 0.4 0.5 Steel H-section Unfilled steel pipe 0.5 Concrete-filled steel pipe 0.4 , Wood 0.4 The third category is geotechnical formula (static analysis) and the particular formula that is used is not given but must be approved. The soil properties used in the formula are specified to be factored by 0.5 for cohesion and 0.8 for friction angle. The capacity resulting from the use of the factored soil properties is then factored by the resistance factors given in Table 6. The resulting capacity shall not be greater than 2.5 EApK or the structural nominal resistance. The quantity E is defined as the modulus
308
This document is lengthy and complex with a large number of descriptive, limiting conditions that are usually not stated quantitatively. The requirements for driven piles will be summarized briefly and should be read with the understanding that only the more quantitative portions of the document are presented. In the Eurocode, both driven piles and drilled shafts are handled in a single section and it will be discussed here in the same way. A nominal strength as used in North America codes is not defined. Rather, the characteristic bearing capacity is defined to be Rc, = SRc, is the measured capacity, and the values where krn of 5 are given in Table 8. The quantity 5 is determined by the number of load tests used to determine the capacity with values given for up to three load tests. The values for 5 are specified without regard to the total number of piles on the job. Thus, a job
with 2,000 piles and one with 100 piles would have the same !-factor for three load tests even though for the large job three tests would be a rather small sample. Only static load tests are mentioned in the code. No values are given for other methods of capacity determination such as dynamic test, wave equation, or dynamic formula. There are no comments regarding the various quality control procedures. The characteristic bearing capacity is further reduced in a second step by component factors to obtain the design bearing resistance according to the relationship (3) where Rbk, and Rsk are the toe and shaft characteristic bearing capacities, respectively, and (pb and cps are the toe and shaft component factors, respectively. The value Rtk is the total characteristic bearing capacity where it cannot be divided between toe and shaft and cpt is the associated component factor on the total capacity. They have been inverted here to more easily compare with the other codes. TABLE 8. Eurocode 6 factors' Number of load 1 2 3 tests Mean of R,,, used 0.67 0.74 0.77 Lowest of &,used 0.67 0.80 0.91 'The values of E contained in the Eurocode are actually the inversi of those presented here. They were inverted to more easily compare with other codes. The values for ( p b , cps, and cp, are given in Table 9. The problem with multiplicative resistance factors is illustrated with this code. If the highest factors and the lowest factors are combined this gives a range of resistance factors ranging from 0.70 to 0.52. Table 9. Eurocode cp values cp h Component factors 0.77 Driven piles 0.63 Bored piles 0.70 CFA piles
(Pq
cp t
0.77 0.77 0.77
0.77 0.67 0.71
The resistance factors contained in Eurocode were developed by calibration to current practice. No probabilistic calibration was performed. 3.6 Danish Code of Practicefor Foundation Engineering (I 985) The Danish Code discusses the use of static analysis and makes recommendations for the establishment
309
of soil properties from subsurface investigation information. It also recognizes the Danish Formula for dynamic analysis of pile capacity based on driving resistance and the use of static load tests. It divides piles into two safety classes with different resistance factors. The following is quoted from the Danish Code of Practice: The coefficient cp given for piles and ground anchors and used for static design (performed with characteristic strength parameters) as well as for assessment of the driving resistance, only apply to the load bearing capacity determined from shear strength tests on soil samples. In cases where the load-bearing capacity is established by test loading, cp is defined for piles actually subjected to test loading, while different values are given for the other (non-tested) piles. The resistance factors are given in Table 10. Table 10. Danish Code (1 985) resistance factors Safety Class Normal High Without test 0.50 0.45 With test 0.63 0.57 Actually tested piles 0.71 0.65 The resistance factors contained in the Danish Code were selected to match current practice. 3.7 Australian Standard (I 995) The Australian Standard for the design and installation of piles is concise and specific. All aspects of design are covered in addition to the axial strength requirements which are discussed here. The resistance factors, given here in Tables 11 and 12 are complete. No information is available as to the methods used to generate the factors. It is interesting that resistance factors are specified to be within a range with suggestions contained in Table 12 for selecting the appropriate value within the range. Structural design specifications have traditionally given nominal strength values that were not-toexceed quantities and lower values than those specified were used, as appropriate, in the judgement of the engineer. This approach would imply that the lower number in the given range is unnecessary. However, the availability of the guidance given in Table 12 makes the range of values useful to the designer. Some of the factors of Table 12 will not be relevant to all of the capacity verification methods contained in Table 11. No recommendations are given regarding the use of measurements made at the end of driving (EOD) versus beginning of re-strike (BOR). The specification does not deal specifically with changes in the resistance factor with increased numbers of tests.
Table 11. Australian Standard resistance factors Method of strength Range of values assessment 0.70 - 0.90 Static test to failure 0.70 - 0.90 Static proof test PDA test to failure with sig0.65 - 0.85 nal matching PDA test to failure no signal 0 50 - 0.70 matching PDA proof test with signal 0.65 - 0.85 matching PDA proof testing, no signal 0.50 - 0.70 matching 0.45 - 0.65 Static analysis using CPT data Static analysis using SPT 0.40 - 0.55 data in sand Static analysis using labo0.45 - 0.55 ratory data for cohesive soils 0.45 - 0.55 Wave equation Driving formulae for piles 0.50 - 0.65 in rock Driving formulae for piles 0.45 - 0.55 in sand Proprietary displacement piles, using well established 0.50 - 0.65 in-house formulae
Driven Bearing Piles (1999) including both ASD and LRFD versions. This code is for building design in the United States and it can be used with either the ACI or AISC structural design specifications. Therefore, there are two different sets of resistance factors matching the different load factors of the two codes. The resistance factors for the ACI loads were arbitrarily selected for inclusion in Table 13. Table 13. PDCA resistance factors Method of strength Range of values determination Static load test'32 0.75 - 0.90 0.70 - 0.80 PDA test3 0.60 Wave equation Dynamic formulas 0.43 0.43 Static analysis - Clay only 'A range of percentages tested from 0.5 - 10 percent is specified 2Static tests can be replaced by dynamic tests at a rate of 4 to 1. 3A range of percentages tested from 2-10 percent is specified. This code is quite specific regarding resistance factors for specified percentages of piles tested. Due to limited space resistance factors are given here as a range in some cases. In the code specific limiting resistance factors are given for particular percentages of piles tested. This code also contains a Commentary that offers additional details and recommendations.
Table 12. Guide for selecting of resistance factors Upper end of range Lower end of range Limited site Comprehensive site Investigation Investigation Simple method of Calculation
Sophisticated design method
Average geotechnical properties
Conservative geotechnical properties
4 SUMMARY AND CONCLUSIONS
Published design parameter correlations
Site-specific design parameter correlations
Limited control
Carefbl control
Less than 3% piles dynamically tested
15% piles dynamically tested
Less than 1% piles statically tested
3% piles statically tested
The values given in the various codes can be compared by assuming a specific live load/dead load ratio and then determining the associated factor of safety for the loads and resistance factors specified. A live load/dead load ratio of 0.3 was selected. Then associated safety factors were determined and are summarized in Table 14. For the case of the use of a static load test for capacity determination a large range of safety factors are specified. The Florida, Ontario, Danish, Australia, and PDCA codes use values ranging from 1.70 to 2.08. In the case of the Australian and the PDCA the range is controlled by the quality control used. The Canadian Bridge code and the Eurocode use considerably larger values. When PDA testing is used the PDCA and Australian codes are very similar in a range of about 2.0 to 1.8. Florida, Ontario, and Canadian Bridge Codes
The specification also includes some resistance factors for the structural limit state for concrete and timber piles. It is the most complete pile design specification of all those reviewed. 3.8 PDCA LRFD Code (I 999) The Pile Driving Contractors Association (PDCA) has adopted recommended Design Specifications for
310
Table 14 Comparison of safety factors for the different codes, L/D = 0.3 Capacity Wave PDA Static Determination Equation Load method Test Florida 3.90 2.37 1.81 Ontario 3.12 2.50 2.08 Canadian Bridge 3.22 2.50 2.50 Eurocode 2.68 Danish 1.70 Australia 2.91-2.38 2.01-1.54 1.87-1.46 PDCA (ACI) 2.43 2.08-1.82 1.94-1.62 use larger values in the range of about 2.5. Eurocode and Danish does not mention the use of dynamic testing. The use of wave equation is covered by all except the Eurocode and Danish codes. Here the values are somewhat more scattered ranging from about 2.4 to 3.9. There are three clearly written design codes available today, Florida, Australia, and PDCA. In each of these cases the conditions for the use of various capacity determination and quality control procedures are clearly linked to values of resistance factors. They are based on modern driven pile design procedures. The other codes all contain some degree of vagueness in the definition of resistance factor use. In some cases the code does not recognize modern driven pile design practice. Methods are available for determining load and resistance factors rationally by probabilistic analysis. Studies should be performed using the many available data bases. It is absolutely necessary that the selection of resistance factors recognize modern driven pile design practice. REFERENCES American Association of State Highway and Transportation Officials, AASHTO LRFD bridge design specrfications. Is‘ ed., 1997 Interim, AASHTO. Washington, D.C., USA. American Association of State Highway and Tranportation Officials, 1994. AASHTO LRFD bridge design specrfications. 1’‘ ed. Washington, D.C., USA. American Concrete Institute, 1963. “Building code requirements for reinforced concrete.” ACI 3 1863. Detroit, Michigan, USA. Barker, R. M., J. M. Duncan, K. B. Rojiani, P. S.K. Ooi, C. K. Tan, & S. G. Kim. 1991. NCHRP re-
port 343: manuals for the design of bridge foundations. Transportation Research Board, National Research Council. Washington, D. C., USA. Canadian Standards Association. 1988. “Design of highway bridges.” CANICSA-5 6-8 8. Rexdale, Ontario, Canada. Cornell, C. A. 1969. “A probability-based structural code.” ACI Journal. December, pp. 974-985. Danish Geotechnical Institute. 1985. “Code of practice for foundation engineering.” DGI Bulle tin 36, 1 Maglebjergvej. DK-2800 Lyngby, Den mark. Dennis, N. D. & R. E. Olson. 1983. “Axial capacity of steel piles in sand.” Proceedings of the Conference on Geotechnical Practice in OfSshore Engineering. ed. SI. Wright. ASCE, pp. 389-402. Ellingwood, B., T. V. Galambos, J. G. MacGregor & C. A. Cornell. 1980. “Development of a probability based load criterion for American National Standard A58 building code requirements for minimum design loads in buildings and other structures.” National Bureau of Standards. Washington, D. C., USA. European Committee for Standardization. 1994. “Eurocode 7: Geotechnical design - Part 1: General rules,” Central Secretariat, Brussels. Florida Department of Transportation. 1997. “1 997 load and resistance factors for bridge design, foundations.” September. Goble, G. G., F. Moses, & R. Snyder. 1980. “Pile design and installation specification based on the load factor concept .” Transportation Research Record 749, Transportation Research Board, National Research Council. Washington, D. C, USA. Goble, G. G. & J. Berger. 1994. “Soil resistance factors for LRFD of driven piles,” Proceedings, International Conference on Design and Construction of Deep Foundations, Vol. 11. Federal Highway Administration, Washington, D.C., USA. Goble, G.G. 1999. “Geotechnical related develop ment and implementation of load and resistance factor design (LRFD) methods,” Synthesis of Highway Practice, National Cooperative Highway Research Program, Transportation Research Board, National Research Council, Washington, D.C., USA. Ontario Ministry of Transportation and Communication. 1992. “Ontario highway bridge design code and commentary,” 3rded. Hansen, J. B. 1966. Code ofpractice for foundation engineering, Bulletin No. 22, Danish Geotechnical Institute, Copenhagen, Denmark. Pennsylvania Department of Transportation, Unpublished Report. “Data base on pile driving records and load test results.” 31 1
Pile Driving Contractors Association. 1999. “Recommended design specifications for driven bearing piles,” St. Louis, Missouri, USA. Rausche, F., G. Thendean, H. Aboumatar, G. E. Likins, & G. G. Goble. 1997. “Determination of pile driveability and capacity from penetration testing,” Final Report, FHWA-RD-96- 179, National Technical Information Service. Springfield, Virginia, USA. Standards Australia. 1995. “Piling - design and in stallation,” 1 The Crescent, Homebush, NSW 2 140, Australia.
312
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Pile testing competitions - A critical review J. M.Amir Foundution Control Institute Limited, Herzlia, Israel
B. H. Fellenius Urkkada TechnologyLimited, Ottawa, Ont., Cunadu
ABSTRACT: A number of pile testing competitions, using low-strain techniques, took place in different
countries between 1987 and 1996 and soon started heated debates by the professional community. Since the outcome of these competitions was obviously contentious, they do deserve a critical review. Such a review may serve as a basis for more productive competitions in the future. These, in turn, will highlight both capabilities and limitations of existing methods and contribute to the advancement of new techniques. The authors describe for each competition the testing scope and program, piles tested, the nature of the defects installed, the participating parties, and the results obtained. They go on to analyze each competition, specifically stressing those items that if done differently could have significantly improved the outcome. Based on the lessons learned fiom these events, the authors present ground rules for fbture pile-testing competitions. In addition, the importance of organizing competitions also in downhole testing applications is strongly recom e n d e d . 1 INTRODUCTION
2 PILE TESTING COMPETITIONS
In spite of the rapid progress in piling techniques (and maybe because of it), defective piles and drilled-shafts are still encountered at many construction sites. Among all the methods designed to test the integrity of bored piles, only two have proved to be of real practical value: The sonic (echo) and the ultrasonic (cross-hole) methods. The sonic method was first applied a quarter of a century ago (Steinbach and Vey 1975). Since then, it has established itself as the leading method for testing the integrity of all kinds of piles. With the advent of handheld computing, sonic testing has become more reliable and at the same time more affordable. The sonic method is based on pressing a sensor against the surface of the pile head while hitting the surface with a hammer. The hammer blow creates a low-strain wave that travels down the pile and is reflected fiom the pile toe, as well as fiom any abrupt change in the pile impedance. The hammer may be either plain or instrumented, and the results may be analyzed and presented in either time or frequency domain. An extensive treatment of the sonic method is presented by Turner (1997). The popularity of the method has brought a proliferation of both instrumentation and testing laboratories. Consequently, it naturally became a subject for competition.
2.1 Objectives In principle, pile-testing competitions should be held with the some or all of the following objectives in mind: 1. Kindling the competitive spirit amongst developers, manufacturers, and users of equipment 2. Establishing the actual (as opposed to claimed) capabilities and limitations of the method 3. Indicating where advances in the state of the x t are required. 4. Serving as milestones to monitor progress in both instrumentation and analysis tools 5 . Providing an opportunity for potential clients to obtain reliable comparative data regarding the performance of available instruments 3 COMPETITION OVERVIEW Following is a brief’ summary of five such competitions, held in Belgium, The Netherlands, and the USA (three of them), respectively. 3.1 Ghent 1987 The frrst known integrity testing competition was
313
held in Ghent, Belgium in 1987 (De Jaeger et al. 1987). The Belgian Society for Soil Mechanics and foundation engineering organized this event. Altogether, the twenty test piles were constructed by four different methods, five piles each: 1. Atlas helicoidal (sawtooth profile) piles, diameter 430/530 mm 2. De Waal precast concrete driven piles, 320x320 mm 3. Socofonda CFA piles, diameter 460 mm 4. Fundex piles, bored with rotated casing and cast in situ , diameter 390 rnm The five testing firms that participated were given the pile diameters, and told that the pile lengths range between 11 m and 16 m. No defects were knowingly produced, and the task in hand was to determine the correct length. The best overall length agreements were obtained in the Socofonda and Fundex piles. The lengths obtained for the Socofonda piles were 93% and 100% of the correct lengths. For the Fundex piles, the spread was between 94% and 102%. Such results are perfectly acceptable. On the other hand, poor results were reported for the precast piles (82% to 120%) and the Atlas piles (101% to 125% for three of the piles, with no results at all for the other two). The testing of the Atlas piles also produced poor results. Although the Atlas piles had the lowest L/D ratio (26 to 30), all testers reported lengths that were too high. The conclusions from this exercise are as follows: 1. The precast driven piles were difficult to test both due to higher shaft resistancc and to having the highest L/D ratio (41 to 53). 2. The testers of the Atlas piles probably neglected the fact that a helicoidal pile exhibits a wave velocity that appears to be much lower than that of a straight-shafted pile (Vyncke and van Nieuwenburg 1987). 3. The CFA piles and the Fundex piles were the easiest to test. This is probably due to their lower L/D ratios (32 to 37) as opposed to the precast piles. 3.2 California 1990 The California test program was carried out in the fiamework of a research project for the Federal Highway Administration (Baker et al. 1993). It took place on two sites: Cupertino, with dry gravelly and sandy soil and San Jose, with clayey soil under groundwater. The piles had a nominal diameter of 915 mm and lengths ofbetween 7.6 m and 18.9 m. ‘l’here were five participants in the program, applying four testing methods: Sonic echo (time domain), transient dynamic response (fiequency domain), cross-hole (ultrasonic) and radioactive (gamma-gamma). All the participants were provided in advance with the lengths and shapes of the piles that they were to test.
3.3 Texas 1990 The Texas test program was a continuation of the FHWA project (Baker et al. 1993). Altogether, nine bored piles were constructed, seven of which had known irregularities. All piles had a nominal diameter of 915 mm, with lengths varying between 11 m and 24 m (L/D ratios between 12 and 26). The irregularities were of different character and magnitude. Four of the piles had a single planned nccking at a dcpth of betwcen 3 and 18 m, the reduction in cross section being between 12 and 50 percent. Three other piles had both increased and decreased cross sections at various depths. In addition to the planned defects, some unplanned defects occurred and were recorded during construction. Five testing firms took part in this competition. The testing firms received fbll information regarding the subsurface conditions, as well as the lengths of the two reference piles. No m h e r data about the existence of defects was divulged. The tests were conducted in two stages: In stage one, only surface (sonic) methods were used. In stage two, the contestants were allowed to lower testing equipment into access tubes which were prepared beforehand. The results of the Texas program may be summarized as follows: 1. At depths smaller than 7 m below the top of the pile head, 80 percent or more of the testing firms managed to identify all important defects in the cross section. 2. The success rate dropped to 60 percent at a depth of 9 m. 3. All participants failed when the defect (necking) was located at a depth of 18 m. 4. Even at shallow depths, participants failed when the reduction in cross section area was merely 12 percent. 5 . In general, an enlarged cross-section was more difficult to fmd than a reduced one. 6. The participants had difficulty in determining the length of the pile when there was a major necking at mid-length, or when a defect existed near the toe of a long pile. 7. As expected, no participant could identify a “soft bottom” condition. 3.4 Delj? 1992 The Delft competition took place in conjunction with the Fourth International Conference on the Application of Stress-Wave Theory to Piles (Smits 1996). Twelve laboratories, employing six different types of instruments, participated in the event. The testing objects in this case were rather uncommon: All of the ten test piles were made fiom precast concrete and installed in the following way: First,
314
closed-end steel tubes were driven to a predetermined depth. Then, a thin bed of sand was placed at the bottom of the tubes and the precast piles were lowered into the empty tubes and placed on the sand-bed. The space around the piles was then filled with a bentonite-cement mixture supposed to represent the local soil stiffness. The test piles had a nominal cross-section 250mm square. Two piles were straight shafied, with respective lengths of 17 m and 18 m. Of the rest, six piles were produced with the cross-section along a given length either enlarged to 300 mm. square or reduced to 200 mm square, or both. The two remaining piles had a sawed notch, 10 mm. thick and occupying one half of the cross section. Similar notches were also produced in two of the piles with enlarged section. The testing circumstances were also noteworthy, in two important respects: First, the participants were not allowed to approach the piles, and the notary public’s clerk was mobilized to hit the piles with the hammer. Second, the participants were given beforehand the exact shapes of all the piles, and their task was to decided which of these shapes best fits each of the reflectograms they obtained. All the participants managed to achieve was a correct fit for between three and seven piles, with a mean success rate of 44%. The scores for the individual piles varied between zero (straight shaft, L = 18m) to 100% (straight shafi, L= 17m).
7% of the tests confirmed the integrity. In the anomalous piles, 82 % of the defects were found. These rates improved somewhat in the fmal reports, to 25% and 83%, respectively. The success rate for the defective piles may seem impressive, but a deeper look into the matter is far less encouraging: Of all the reported defects, only 36% managed to fit the depth of the defects within +20 percent. Of these, only 37% provided the size of the defects within -t20 percentage points. On the other hand, the participants reported on the average 1.3 “phantom”, i.e. nonexistent, defects per pile. None of the contestants, or the testing instruments used, demonstrated a markedly outstanding performance. 4 EVALUATION 4.1 Ghent 1987 The organizers declared that the piles were between 11 m and 16 m long. Thus, a tester who would have reported a uniform length of 13.5 m (the mean of the above limits) for all the piles would be accurate to within f.10% in 90% of the piles. While highlighting the influence of construction method on testability, the Ghent competition totally neglected the most important purpose of sonic testing -- finding defects! Publication of the test results was very complete, and included a special seminar where the results were presented and discussed.
3.5 Texas 1996 This competition (Samman and O’Neill 1997) took place on the campus of the University of Houston, Texas. Altogether, twenty-two piles were tested. Eleven of the piles had a diameter of 460 mm and were bored to a depth of 4.6 m. The other eleven piles were 760 mm in diameter and 7 m long in the ground. Some of the piles were constructed with polymer slurry while the rest were cast in the dry. Six of the piles were regular piles, while sixteen piles had planned built-in defects. These defects were produced from 25 mm thick soft rubber mats, laid horizontally. The mats were placed at different depths, but not more than one per pile. Each occupied between 10 and 50 percent of the total cross section of the pile. Eight laboratories took part: Two were from government agencies, five were commercial and one academic. The contestants were asked to report for each pile whether it is sound or defective, and, in the latter case, specify the depth and severity of the defects. The participants were to submit two reports: A preliminary report on the same day, and a final report within five days. The results, as can be expected, were far from satisfactory: In the piles intended to be sound, only
4.2 California 1990 Since the organizers of the California testing program gave the participants full details regarding the planned defects, success rates have no meaning. It is therefore questionable whether the California test may qualify as a competition and therefore the case is not pertinent to the present paper. (The results are of course interesting in other contexts). 4.3 Texas 1990
While in California, the participants were given information beyond that normally provided to testers, in Texas they got too little. When a pile testing firm is invited to a construction site, it is customary and necessary to provide it with all relevant information, such as soil data, pile construction records and piling logs with the asmade length and details of any irregular events that may have happened. Testing under the “Texas rules“, with no a-priori knowledge of the pile length, is therefore the exception and detracts fi-om the effectiveness of the results. In other respects, the planning and execution of the Texas tests was very effective. The tasks were 315
well graded from easy through difficult to impossible. Thus, the performance of contemporary systems was well defined. This competition proved convincingly that the sonic method is a viable technique for investigating pile integrity. It showed that sonic equipment is able to identifj most important defects where they matter most, that is in the upper part of the pile. On the other hand, it demonstrated that the sonic method is unable to distinguish features that are relatively small or located deep down the pile.
4.4 Deyt 1992 As expected, the Delft 1992 event triggered a lengthy debate in Ground Engineering magazine (Stain 1993, 1993a, Ellway 1993). The main criticism was aimed at the following points: 1. Most routine sonic testing is done on cast in situ piles, with an inherent variability of both concrete quality and shaft resistance soil friction and a rough top surface. Precast piles in an artificial “soil” with smooth tops thus cannot do not represent real-world life conditions. 2. The unusually high L/D ratio (72) is generally considered to be beyond normal testing limits. 3. Most of the anomalies, and especially the saw cuts, were outside the theoretical performance envelope of the sonic method. 4. Actual testing was performed by inexperienced people, not familiar with fine points of the test. In all important respects, the Delft competition did little to advance the state of the art, and in fact was a large backward step from the Ghent affair. With the whole setup being detached from the real testing world, it had only reinforced the (erroneous) belief that sonic testing is not to be taken seriously, being based on little more than guesswork. 4.5 Texas 1996
In view of the poor results obtained in Houston 1996, the organizers declared that sonic testing “may not be reliable enough to be regarded as a standalone measure of the assurance of drilled shafts”. Could it be that not the sonic method was to blame, but the organization of the competition? In principle, the Houston competition had the correct ingredients to simulate a realistic testing assignment. The main factor that detracted fiom the success of this competition lay in the nature of the “defects”: To be applicable, the sonic method utilizes a wavelength that is large in comparison with the pile diameter. A defect with a vertical dimension of 25 mm is therefore well beyond the capability of the sonic method unless it occupies all (or almost all) of the cross section of the pile. Since the defects in Houston occupied at most one half of the total pile
area, it took a lot of good luck to discover any of them. The fact that some defects were placed as close as 300 mm to the top only made things worse. Moreover, a rubber sheet, such as that used to model cracks in the piles, has a low stiffness when unstressed. In contrast, a rubber sheet stressed by the weight of the concrete has a considerable larger stiffness that does not deviate enough from the stiffness of the concrete. It would be have been very hard to detect a reflection from the rubber sheet. Since most competitors were keen to find defects, and the nature of the defects made them practically undetectable, the competitors found defects even in perfectly good piles. This is the main explanation for the participants discovering defects also in the good piles (plain coin tossing would do markedly better!). 5 DOWNHOLE TESTING COMPETITIONS Admittedly, the sonic method has a few basic flaws: First, the wavelength used is about 3 m, which provides rather poor resolution and second, both input (hammer blow) and output (accelerometer signal) are remote from potential defects. To overcome these drawbacks, the industry has developed instrumentation that is lowered into the pile through-access tubes+ Historically, access tubes were first used for testing piles with radioactive isotopes. This method is now fast disappearing due to its limited range, environmental limitations, and regulatory requirements. It was largely replaced by ultrasonic instrumentation, using wavelengths in the range of 50 mm to 100 mm. Modern ultrasonic equipment (Amir and Amir 1998) combines long range (-3m) with high resolution. With a suitable setup, it can also perform tomographical imaging and produce two-dimensional vertical sections. Another technique, still experimental (Samman and O’Neill 1997), utilizes clear plastic tubes and a downhole video camera. In view of their obvious advantages, downhole testing of piles has become the preferred method in certain sectors such as bridges and high rise buildings. The time is ripe to organize suitably designed competitions which would greatly benefit the piling industry. 6 RULES FOR FUTURE COMPETITIONS To be effective, competitions must satisfy certain minimum criteria. Based on the analysis of five such competitions, The following rules are therefore suggested: 1. The test program should be based on sound theoretical foundations - participants must not be asked to perform the impossible.
316
2. The tests should be carried out on real piles, 3.
4.
5.
6. 7. 8.
9.
REFERENCES
conventionally constructed in real soil. The piles should have different lengths and length-to-diameter ratios. As a rule, organizers should create no more than one anomaly per pile. Anomalies should be carefully designed and constructed to resemble, as far as possible, anomalies that are actually encountered in practice. This includes soil pockets and zones of weak, honeycombed concrete. The anomalies should be of different magnitudes, with an importance ranging between minor irregularities to complete discontinuities. Tests should be carried out by experienced personnel, familiar with the testing systems Participants should be provided with normal testing conditions. Pile heads should consist of reasonably goodquality concrete. Testers who desire to improve the surface must be given an opportunity to do
Amir, E.l and Amir J.M (1998) Recent Advances In Ultrasonic Pile Testing, Proc. 3rd Intl Geotechnical Seminar On Deep Foundation On Bored And Auger Piles, Ghent ; 18 1-185 Baker, C.N., Parikh, G., Briaud, J.L., Drumwright, E.E. and Mensah, F. (1993): Drilled Shafts for Bridge Foundations, FHWA, McLean, Virginia De Jaeger, J., Carpentier, R. and v.d. Broeck, M. (1987): Integrity Tests, Ch. IV in Seminar on Pile Dynamic Testing (Integrity And Bearing Capacity), Brussels Ellway, K. (1993): Objectives Of Competition Are Unclear, Ground Engineering, June; 8 Samman, M. M. and O’Neill, M. W. (1997): An Exercise In Seismic Testing Of Drilled Shafts For Structural Defects, ADSC Foundation Drilling, Dec. - Jan; 1 1- 17 Smits, M. Th. J. H. (1966): Ch. 3 - Pile Integrity Tests, Application of Stress Wave Theory to Piles: Test Results, Balkema, Rotterdam;25-54 Stain, R (1993): Test’s Integrity Is Questionable, Ground Engineering JanuaqdFebruary, p. 7 Stain, R (1 993a): Competition Was Not Applicable, Ground Engineering, April; 15 Steinbach, J. and Vey, E. (1975): Caisson Valuation By StressWave Propagation Method, .J. Geotech. Div. ASCE, Vol 101 No. GT4, April ; 361-378 Turner, M. J (1993): Integrity Test Usefulness Is Not The Issue, Ability is, Ground Engineering, July/August pp. 2728 Turner, M. J. (1997): Integrity Testing In Piling Practice, CIRIA Report 144, London Vyncke, J. and van Nieuwenburg, D.(1987): Ch. I1 - Theorie Van De Dynamische Proeven: A. Integriteit, Seminar On Pile Dynamic Testing Integrity and Bearing Capacity, Brussels. Samman, M.M. and McNeill, M.W. (1997): Fiber-optic inspection of drilled shafts, Foundation Drilling, Sept.-Oct; 16-19
--. qc)
10. Participants should be provided with sufficient soil data (borehole logs) and pile data in the manner and to the extent usually provided to testers on actual construction projects. This includes the as-made lengths and any special events observed during construction. 11.Participants should not get any data regarding the special features installed in the piles. 12.In addition to the piles specially prepared for testing, the competitors must be given an opportunity, where possible, to test “ordinary” control piles at the same site. 13.The integrity of the piles should be investigated also by conventional methods, such as coring and pile extraction, in order to provide reference to both the integrity of the piles and the success of the integrity testing competition. 14.The competitions setup and program should be reviewed and sanctioned by a reputable international body, such as APTLY, the Alliance of Pile Testing Laboratories. 7 SUMMARY
Pile testing competitions represent a major technical and financial effort for organizers and competitors alike. To profit from this investment, these competitions should be planned very carefully. The experience accumulated from pile testing competitions in the past can serve as a good basis for planning successful competitions in the future. Such competitions should be open to all available testing methods, both commercial and experimental. Such events should be coordinated with APTLY and published in full in a technical journal or conference that is readily accessible to the general piling community. 31 7
This Page Intentionally Left Blank
Application of StreSs-Wave Theory to Piles, Niyama & Beim (eds) @ 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
The need for quality assurance in the dynamic pile testing industry J. F?Seidel Monash Universily, Melbourne, Vic.,Australia
ABSTRACT: The theme of this Stresswave 2000 Conference “Quality Assurance on Land and Offshore Piling” is one that is timely and worthy of exploration. This discussion paper has taken a broad view of many aspects of quality assurance as they apply to engineering projects. The role of dynamic pile testing in the quality assurance process is highlighted. It is concluded that as dynamic pile testing is a key aspect of quality assurance in many foundation contracts, it is essential that dynamic testing also be subject to the same rigor of quality assurance. It is asserted that the current ad hoc training process is unsatisfactory. The paper argues for an industry standard for training and evaluation of competency. Several models are suggested. This should be associated with a system of formal review for testers deemed to have only basic skill levels. A scheme for prequalification for testing contracts, which applies conditions to testers with basic skill levels, is also suggested.
1 INTRODUCTION Quality assurance has become a catch-cry in many countries around the world. Like motherhood, the pursuit of quality cannot be criticized. However, unlike motherhood, the meaning of the term ”quality assurance’’ is open to interpretation. Clients can also try to effect quality assurance on their projects in several ways. In this section, a number of different methods and applications of quality assurance will be discussed in principle and with reference to both civil engineering projects and dynamic pile testing.
1.1 Qualily Assurance Systems In some countries, the term Quality Assurance (QA) denotes a system of checking, reporting and documentation which that has become an integral part of the contractual process. The purpose of the QA system is to set in place a process which will ensure the final built quality of the product. Implementation of the QA system is usually the responsibility of the entity (contractor, specialist subcontractor etc.) performing the work, There is no direct supervision of the contract by the client or his representative. Any aspect of the work which does not conform with the specification is the subject of a Non-Conformance Report, NCR (a particular terminology used - equivalent terminologies are possi-
ble). Submission of an NCR flags a problem which requires explicit consideration and rectification before consequent work can proceed. All work or rectification must ultimately meet the contractual and specification requirements. Although ensuring the final construction quality by this process is a laudable goal, the reality is that quality assurance techniques do no more than address consistency of processes and adherence to target outcomes. Quality Assurance as a system is well suited to processes such as manufacturing, which are repetitive, devoid of subjective decision and where the product can be tested or measured to ensure compliance with specification. Mass production of consumers products is an ideal application. By inference, QA techniques are not well suited to processes which are one-off, involve subjective decisions and experience, and where the end-product is not always available for direct measurement. Most civil engineering projects would fall into the latter description. Foundation engineering, in particular, is a field in which the knowledge and experience of each member of the team - designer, construction engineer, foreman and crew are critical to successful construction. Each project, and even each pile within a project brings new challenges because of the uniqueness of the ground and groundwater conditions. For instance, the pile driver must be responsive to changes in pile blow count, and must be 319
aware of conditions that will potentially damage the pile during installation. The driller must prevent collapse of the shaft, and must be sensitive to the drilling equipment to ensure that a suitable founding layer has been attained. These are all aspects of the process which are based on knowledge, experience, subjectivity, good practice, and even intuition, which cannot be defined in a contract specification, or measured for compliance. and are therefore outside the ambit of a QA System. This is not to say that QA Systems do not have a role, or should not be an element of the engineering process. Rather, it is important that the limitations of these systems be understood and acknowledged in the formulation of a more global approach. What must be avoided is a belief that by virtue of implementation of a QA System, that the end product is necessarily of high quality. In the author’s experience, there is a real danger that the application of such techniques can actually reduce the quality of construction. The QA System requires such a plethora of documentation that time which should be spent “at the coal face” ensuring real quality is spent filling in forms to document compliance. Dynamic pile testing can, and often is an integral part of QA systems for driven piles, usually in the aspect of confirmatory testing, which will be discussed subsequently. It is also possible and desirable, that the wider capabilities of dynamic pile testing be utilized in setting parameters or guidelines for the installation process, i.e. as an integral part of the construction control process. However, this use is generally underutilized, as it is not as easy to prescribe this type of application in the System. It is also noted that dynamic pile testing is also applied to only a limited percentage of the piles installed on a contract (usually 15% or less). An inference must be made that the quality of the test sample is similar to the quality of the remaining piles installed.
1.2 Qiiality Asszirance
on self-regulation and the professional integrity of each player in the process, are open to criticism on this account. The potential disadvantages of this approach are : e it is an additional cost to the project; the liability for any problem may be partially transferred from the contractor to the supervising body; the effectiveness of this process is dependent on the knowledge and experience of the particular superintendent. Particularly for foundation engineering, where many proprietary systems are used, and where the skills of foundation construction are very specialized, the ability, knowledge and experience of the person providing oversight will generally be inferior to that of the specialist foreman and crew undertaking the work. In this case, the supervision is not an effective approach for quality assurance. In the particular case of dynamic pile testing, it is unlikely that a construction supervisor would be able to effectively oversee the testing process.
1.3 Quality Assurance - professional review Quality assurance can also be applied to the engineering design and construction processes. This can be in the form of either a design check, review or independent design. The independent design would normally be performed by an external organization, but checks and reviews could be undertaken both inhouse or by an independent external engineer. The use of professional reviews is common for larger or more complex projects. It is implied in this process that the design or construction records be transparent and verifiable, and that independent reviewers with the necessary skills to interpret this information are available. Further comments on this aspect with regard to dynamic pile testing will be made later.
1.4 Quality Asszirance - Preqzicdifcation
- sirpervision
A more traditional approach to ensuring a quality outcome for construction projects has been by means of independent supervision of the construction process. Typically, the client employs a person knowledgeable in construction techniques to oversee the construction in either a part-time or full-time capacity. Duties typically include taking spot measurements of critical dimensions. and ensuring compliance with the specification and with good practice. This method is still in wide use in some countries, and with particular clients. The primary benefit of this approach is that the quality of the construction process and the completed product is assessed by a person independent of the process itself, and whose exclusive role is to check for quality. QA systems. relying as they do
On larger and more complex projects, it is also common for contractors or consultants to demonstrate that they have the requisite skills and experience to successfully complete a project or provide advice. Without prequalification, the organization cannot even bid to provide services. Prequalification might typically require the organization to provide details of previous relevant projects completed; available resources; experience and qualifications of key field and office personnel and references. Prequalification is adopted by some statutory bodies with regard to dynamic pile testing organizations. In the absence of any effective alternative, the process is typically based on whether the organization has provided testing services for a reasonable length of time, or on the advice of independent refe320
rees. To .the knowledge of the author, this approach is not widespread for dynamic pile testing. It is also applied to organizations rather than individual testers.
load factors and strength reduction factors under the Limit State approach. Nevertheless, there is an important implied assumption that the capacity estimate provided is the best, unbiased and ‘correct’ estimate available within the constraints of that technique.
1.5 Quality Asstiva~ce- Direct Testing
A key element of quality assurance in civil engineering projects is the testing of constituents, individual elements or in rare cases, large or complete systems. With specific reference to foundations, the quality of concrete is routinely tested by standard compression testing, or chemical testing. Weld quality is also evaluated by inspection and ultrasonic testing. There are either absolute or statistical standards with which the tests must comply. Such routine testing of constituents fits well within Quality Assurance Systems or more traditional contractual arrangements, as compliance or non-compliance is easily demonstrated. The testing of individual piles is also commonly stipulated as part of a project specification. In this case, a variety of techniques can be used. For driven piles, the final pile set and (optionally) temporary compression are measured for every pile to obtain a measure of axial pile capacity using a driving formula such as Janbu, Hiley, ENR or Gates. Although these methods are applied to the entire population of piles, they are known to be simplistic estimates of capacity, and hence subject to high levels of uncertainty. In many countries, for all but the smallest projects, more sophisticated, and hence more reliable testing is performed in some combination of static load testing, dynamic pile testing, StatnamicB pile testing, Osterberg testing and various pile integrity testing techniques. The particular testing strategy developed for each site is a function of economics, available technologies, pile type, site conditions and historical precedent. Testing of piles as individual components of the final system also fits well within the framework of a Quality Assurance system in which quantitative outcomes can be measured against the required specification. Engineers generally recognize that despite conducting a targeted pile testing program, some uncertainty still exists for the following reasons: e Only a fraction of the contract piles are tested using high-level techniques. The capacity of untested piles must be extrapolated or otherwise inferred; a Every testing method provides an estimate of the (axial) pile capacity at the time of testing, and each method has an associated uncertainty. These uncertainties are accommodated by specification of ultimate capacities which incorporate an appropriate factor of safety or through prescribed
2 QUALITY ASSURANCE -THE TESTING PROCESS As noted, direct testing is a key element of quality assurance in many civil engineering projects. It is a particularly common technique in foundation engineering due to the uncertainties introduced by the natural stratigraphy which is not known perfectly inadvance of any foundation contract. As verification is a key element of the overall construction process, it would be ilIogical if the testing were not itself subject to the principles of quality assurance. This section considers quality assurance of the testing process in the particular context of static and dynamic pile load testing.
2.1 Static Load Test@ Standards and Specifications which set minimum requirements for compliance exist as part of the framework of quality assurance for testing. For static load testing, ASTM D 1 143-81(1 994)el (ASTM, 1994) is the U.S. National Code of Practice which sets standards for the equipment, calibration, procedures and records required for testing piles under axial static compressive loads. Equivalent Codes of Practice for static pile load testing are mandated in many other countries. If tests are performed in accordance with the appropriate standard, it is generally assumed that the load-settlement response is a true representation of the pile load-settlement response. This may not, however, be entirely correct. Fellenius (1 984) describes the large errors which can be introduced by using a manometer attached to the hydraulic jack which is simultaneously a load application device and a load measurement device. It is more appropriate to use a separate load cell so that the load estimate is true and unbiased. AS2 159 (Standards Australia, 1995) mandates the use of a load cell for static pile load testing. The complexities of interaction effects between the test-pile and reaction pile, anchors or reaction weights are usually ignored, and will not be dealt with here, other than to note that both the inferred capacity and settlement characteristics can be affected. Engineering analysis is required to correct these physical effects. It should also be noted that there are many alternative procedures for static pile load testing. Standard methods are known variously as Maintained 321
Load Tests, Quick Maintained Load Test and Constant Rate of Penetration Test (other names are used). Within these broad categories, an infinite number of specific test regimes are possible. Due to the different loading paths. any pile subjected to the various tests will exhibit a different load-settlement response. The significance of the different responses, the separation of elastic, plastic and creep components, and the extrapolation to service behavior is a matter for engineering analysis and interpretation. Furthermore, Fellenius (1 980) notes that the interpretation of ultimate capacity from a static loadmovement curve is not unique. Application of the many constructions proposed (e.g. Davisson Offset Limit , Brinch-Hanson, Chin-Kondner) result in significantly different estimates of ultimate capacity. It can be appreciated, therefore, that correct interpretation of the simple static load test is more complex than it would first appear. Analysis of the test data would generally be performed by a specialist geotechnical engineer using accepted methods that are in the public domain. In general. however, the assumptions made in the analysis of the load test data are transparent (it is, after all only a correction which is applied to the measured response), and verifiable. The fraternity of local geotechnical engineers is usually large enough to enable a professional review to be made, if required. Furthermore, as noted previously, in the majority of cases, the static load test response is taken to be a true representation of the pile load-settlement response. For contractual purposes, it is often necessary only to ensure compliance with the specification of the peak applied load and the settlement at one or more defined loads. Further analysis is not undertaken, and the test outcome is accessible and open to direct and immediate interpretation by both structural and geotechnical engineers.
2.2 Dynamic Pile Load Testing Just as for static load testing, Standards exist which prescribe the requirements for dynamic pile testing methods. ASTM D4945-96 (ASTM, 1996) sets standards for the equipment, calibration, procedures and records required for testing piles using dynamic impacts. AS2 159 (Standards Australia, 1995) stipulates requirements for the use and interpretation of dynamic pile testing. The Institution of Civil Engineers (ICE, 1996) have published specifications m d good practice guidelines on inter alia dynamic pile testing. Other similar documents with national authority exist, although fewer than would exist for static load testing. The interpretation of dynamic load tests has both similarities and differences to the interpretation of static load tests. They are similar to the extent that
correct interpretation requires specialist geotechnical knowledge. However, the following important differences exist: The direct output of a dynamic test is not a load-settlement response, but usually pilehead strain-time and acceleration-time responses; * The measured test response is a dynamic response, and the static behavior which is to be determined must be extracted from the test using either simplistic or more complex analytical or numerical techniques. The fact that the direct test outcome bears no resemblance to the load-settlement response is significant. This means that most structural or geotechnical engineers are unable to interpret the test result. To this extent the results are not transparent, and the technique is therefore considered “black box” technology - sometimes with the attendant negative connotations. Interpretation of these pile-head time records is a specialized technique, which is generally known or understood by the small number of practitioners who are providing dynamic pile testing services. This is not to say that information on how to interpret dynamic pile testing records is not well published and available in the public domain. However, the reality is that the technique is so specialized that those not directly involved simply “leave it to the experts”. Local professional review is also not generally available, as the only potential reviewers are likely to be testing or construction competitors, and commercial and professional sensitivities about releasing data have even resulted in claims of intellectual property over test records. This results in a problem with verification of results. Dynamic pile testing suffers from the problem that if the process is not typically transparent or verifiable, the client is not in a position to independently assess the skill, understanding and knowledge of the tester. He also cannot assess whether the estimate which has been made is actually the best estimate, given the constraints of the technology. It appears to be a universal experience, judged by the author’s personal communications with colleagues around the world, that not all practitioners providing dynamic pile testing services have adequate skills. The author is aware of cases where dynamic pile testers have given gross errors in advice due either to poor data quality which is undetected or ignored, or due to misinterpretation or incorrect analysis of the test records. For obvious reasons, these cases are not detailed here. The author’s experiences are not unique. These errors in advice (if detected by the client) can not only affect the client’s confidence in the practitioner, but also their view of the reliability of the dynamic testing method in general. This has an
322
unfortunate and undeserved flow-on effect to all practitioners. This is to say nothing about our responsibility as professionals to provide professional advice, and our moral obligation to ensure the integrity and safety of the structure and people who could be affected by a collapse. We must collectively address the challenge of quality assurance within the dynamic pile testing industry world-wide.
3 QUALITY ASSURANCE IN THE DYNAMIC PILE TESTING INDUSTRY The previous section has addressed quality assurance of the testing process in the context of the foundation industry and with regard to dynamic pile load testing in particular. The following section discusses some possible strategies to ensuring that dynamic pile testing services are provided in accordance with the principles of quality assurance. This discussion is placed in the context of current practice.
3.1 Training A fundamental requirement for ensuring compliance with the principles of quality assurance is that the individual providers of advice are suitably competent. Four stages have been identified in the general development of competence. These can be applied to the specific case of dynamic pile testing: 0 Stage 1 : Consciotisly incompetent. The person is not able to perform a task in a competent manner, and is aware of their inability. This person is not a danger, because he/she will generally be prudent enough not to provide advice. e Stage 2 : Unconsciously incompetent. The person is able to perform tasks at a basic level, but is unaware of what they don’t know, and the implications of their advice. This person is a danger, because he/she will provide advice without the necessary skills to assess whether this is provided on a sound basis. e Stage 3 : Consciotrsly competent. This person has reached a stage where they have achieved a basic to advanced understanding. They are also aware of what they don’t know and the possible implications of their lack of knowledge. In general, this person will provide advice within the limits of their knowledge, and seek assistance in areas outside their competency. They would only be dangerous when they do not seek appropriate advice outside the limit of their expertise.
* Stage 4 : Unconsciously competent. This person has become an expert; has an intimate knowledge of the subject area which they can apply without effort and can apply their knowledge to areas beyond their direct experience. In developing strategies for quality assurance in the dynamic pile testing industry, all four stages of competency must be catered for and addressed. Systems should be in place both to identify the stage of competency, and to prescribe an appropriate level of autonomy or independence. It will be seen from the previous descriptions that the Unconsciously Incompetent person - unaware of their own limitations is the person that requires the greatest attention. The key to competency is knowledge. Knowledge is typically acquired by one of three methods by formal or informal training, education, mentoring or reading; through experience; and finally by making mistakes. All are powerful methods of learning. Obviously, in a contractual environment, the last method is undesirable, and should be avoided by one of the first two methods. Experience, unfortunately can only be acquired over time. Training is the traditional way to quickly develop competency. There are no current standards or guidelines for training within the industry, and no standards or certification which can demonstrate competency. Because of the highly specialized nature of this field, the subject matter is not generally covered in an undergraduate engineering degree. Because of the limited interest and again because of the scarcity of people qualified to teach in this area, there are no formal post-graduate courses available in pile dynamics (to the author’s knowledge). There is, therefore, no formal qualification or certification which can be obtained which indicates competency in this field. There is no requirement to either hold a degree or equivalent undergraduate qualification in Civil Engineering, or in Engineering in general. This is not to say that such a qualification is necessary, although such a degree provides obvious background to some of the principles employed in dynamic pile testing. Guided experience, self-help and training can provide an equivalent level of competency over time. An individual or organization new to dynamic pile testing could expect one of the following training experiences: * An on-site training program of 1 to 3 days, generally provided in association with purchase of new equipment. Training is conducted by the equipment manufacturer, or a designated agent. This training would only be an introduction to use of the equipment and underlying theory. At the conclusion of this training, the user would be at either Stage 1 or Stage 2 competency in the above model
323
An informal training program conducted onthe-job and in-house by a tester from the same organization. The trainer in this case may have been given the original training by the equipment manufacturer, or may have been given similar informal in-house training. In this way, 2'Id or 3rdgeneration training is quite common. There are no controls on the quality of such training. It would typically concentrate more on the particular types of projects encountered, and practical aspects of testing, and would give less broad overview, principles and theory than the manufacturer's training program. Manuals will generally be made available to supplement the training. The quality of the training will only be as good as the competency of the trainer. Any misconceptions of the 2'Id or 3rd generation trainer will be passed on to the new trainee. This.current ad hoc method of training is not an ideal, and is not consistent with a quality assurance approach. In order to comply with quality assurance principles in training, it would be necessary to provide a consistent minimum standard of training to all new practitioners. If the training is to be industry-wide, it would be desirable to provide this through an independent industry organization or educational institution. The practicalities of providing uniform training to a dispersed world-wide market would tend to suggest the need for the material to be delivered in distance education mode. Concentrated workshops could be an alternative model. 0
3.2 Testing Conipetency The purpose of training should be to establish basic competency in the providers of testing advice so that at the completion of the training period they are moving into Stage3 - Consciously Competent. Of course, this stage will not be fully reached without a period of field experience to reinforce the principles learned during training. With further experience, discussion with peers, attendance at industry seminars and conferences, there would be an expectation that the tester would progress to advanced and eventually expert status. Under a quality assurance philosophy. there would be a need not only to provide and undertake training, but to assess competency both of those that undertake training, and those that are already providing dynamic pile testing services e It needs to be established that testers have achieved and maintain a basic level of competency, and 0 Those testers moving from Stage 2 to Stage 3 (hence from a basic level to an advanced level) should receive an appropriate level of
review from a tester with either advanced or expert status.
3.2.1 Evaltiation of competency Assessment of competency could be formulated in different ways. Some of these proposals may be difficult to develop in practice: By a mentoring scheme in which assessors (accredited by an independent industry group) would evaluate the competency of an applicant over a period of time. Both field and analysis skills could be assessed.; By an independent review panel (comprised of acknowledged industry experts) that could assess a submission from an applicant based on examples of the applicant's work; a list of projects completed; details of training undertaken and referee reports. An interview process could also be part of such a scheme; By standard examination. This could cover both data acquisition and data interpretation skills, but could not assess practical skills on site. A multiple-choice format would give a most objective assessment of capability Any method used to assess competency should be able to effectively distinguish three levels of competency - basic, advanced and expert. All approaches should be capable of giving feedback to applicants on areas of weakness so that targeted training and improvement is encouraged. As skills in dynamic pile testing need to be reinforced by regular practice, it may be necessary to instigate a system of regular review and re-appraisal, particularly for those with only basic skills, and for those who may test on an infrequent basis. 3.3 Review Having established a system of assessing competency, it is a logical consequence to ensure that those practitioners assessed to have at most basic competency be required to obtain review from others with either advanced or expert status. The ideal situation would be for such partnering to be undertaken in-house with someone with higher competency. This person would be required to "sign off' on all testing and analysis undertaken by their junior. Where a person of higher qualification was not available within the organization, arrangements would be required for review by any eligible reviewer. The arrangements for such review, including commercial and legal aspects, would be a matter for joint agreement. It is anticipated that standing arrangements would be made with a particular reviewer to streamline the process. This would lead to an effective mentoring arrangement and encourage transfer of knowledge.
3.4 Preqzral~fic~rtion
The principle of proportionality; pile impedance; The relationship between mass density, modulus and wavespeed; Material properties; Allowable stresses; Principles of one-dimensional wave mechanics; Computation of pile wavespeed in easy and hard driving conditions; Recognition of easy, moderate and hard driving conditions; Identification of high compression stress levels at the pile head and at the pile toe; Estimation of pile toe stress; Identification of high tension stresses before and after the 3L/c time; Identification of damage intensity and location; Recognition of a broken pile, pile joints and changes in section; Recognition of response from end-bearing piles, and piles with small and large shaft resistance; Estimation of end bearing, shaft resistance, and shaft resistance distribution; Recognizing high and low cushion stiffness; Recognizing bending and poor hammer-pile alignment; Understanding the effect of resistance, or impedance changes close to the transducers; Principles of valid data adjustment; Pile compression and tension capacity; Estimating Case damping factor; Basis of the Case method of capacity determination, and factors influencing capacity estimates; The implications and conditions for pile unloading; Pile set-up and relaxation; Mobilization and under-mobilization of capacity; Hammer performance and transfer efficiency;
As noted earlier, prequalification is a quality assurance technique which is used for contracts requiring high levels of expertise or resources. Some statutory authorities already require prequalification for providers of dynamic pile testing services. It is suggested that this system could be more widely adopted, and that acceptance be based on the levels of competency assessed by the evaluation process. The requirements for review, as suggested in the previous section could be formalized in the conditions for prequalification. It would be important that providers not be excluded from providing services, but rather that their provision of services should be accepted conditional on demonstration of effective and timely review arrangements.
4 BASIC SKILLS The continuing development of dynamic pile testing equipment has enabled the operator to have a vast array of information available in real-time. Algorithms have been developed which provide critical feedback to the operator at critical stages of testing alerting the operator to potential hazards (such as development of damage, excessive stresses, excessive bending etc.) Many calculations which were previously undertaken manually are now automated, freeing the operator to concentrate on more critical observations. Although these developments are positive, there is a danger that the operator will lose the ability to critically evaluate the results which are provided by the equipment. No algorithm is flawless, and no equipment, however smart the software, can replicate the abilities of an expert to critically evaluate the pile responses and make appropriate judgements and decisions. Without a detailed understanding of the basis for the computations, the limitations of these computations, and the implications of changes in parameters or assumptions cannot be known. If the operator is to remain in effective control of the testing process, and not become a slave to the equipment, it is imperative that the operator develop skills to make the necessary independent critical judgements. Only by developing these skills can he progress to advanced and then expert level. Following is a tentative list of the types of skills necessary for effective dynamic pile testing. Many of these relate to recognition of conditions from the forceivelocity and upward/downward wave responses: o Recognition of valid and invalid data in a variety of scenarios; 8 Field measures required to rectify poor data;
5 CONCLUSIONS This paper argues the need for dynamic pile testing to be subjected to a rigorous industry-wide quality assurance scheme, compatible with the foundation industry which the testing serves. The scheme should be one which delivers real, fundamental quality, and not one which merely produces documentation to satisfy administrative requirements for an audit trail. Ideas for a possible comprehensive model have been suggested. The challenge to adopt such an approach should be taken up by all players in the field
325
of dynamic testing including practitioners, industry groups and clients. Without adopting the principles of quality assurance, the reputation of the industry as a whole will continue to suffer at the expense of alternative testing methodologies.
REFERENCES ASTM ( 1994) ASTM D 1 143-8 I ( 1 994)e 1 Standard Test Method for Piles Under Static Axial Compressive Load. American Society of Testing and Materials, West Conshohocken, PA ASTM (1996) ASTM D4945-96 Standard Test Method for High-Strain Dynamic Testing of Piles. American Society of Testing and Materials, West Conshohocken, PA Fellenius, B.H. (1980) The analysis of results froin routine pile load tests. Ground Engineering, September, 1980 : 19-3 1. ICE (1996). Specification for Piling and Embedded Retaining Walls. Institution of Civil Engineers. Thomas Telford Publishing, London ISBN 0 7277 2566 1 Standards Australia (1995) AS2159-1995 Piling - Design and installation. Standards Australia, Sydney. 0-7262-9884-0
326
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
High strain dynamic pile testing, equipment and practice G. Likins Pile Dynamics Incorporated, Cleveland, Ohio, USA
E Rausche Goble Rausche Likins and Associates Incorporated, Cleveland, Ohio, USA
G.G.Goble Goble Rausche Likins and Associates Incorporated, Boulder, Colo., USA
ABSTRACT: High strain PDA testing has become common for verification of capacity of both driven and bored piles. It also investigates other aspects of the installation of driven piles such as hammer efficiency and driving stresses. The high benefit of dynamic test results for the relatively low costs of the tests has resulted in widespread use. An overview of typical practice in the United States is given including a review of codes which guides the practical applications. The equipment hardware and soRware required to meet the current needs and the vision of the fbture is detailed, and the expertise required is discussed.
1 INTRODUCTION
As technology changes, the construction industry adapts. Construction of increasingly larger structures requires increased ultimate loads. Piles therefore become larger and longer. To install these piles, the pile driving equipment has changed dramatically from simple drop hammers to air, diesel or hydraulically powered hammers with greatly increased energy rating. Increased complexity and increasing loads, coupled with a need to maintain economy, requires better quality control and/or lower safety factors. A century ago, dynamic formula, now widely considered as unreliable, was the only available means of assessing the pile capacity from a “measurement” of set per blow. Test piles were subjected to static tests for final proof. In the 1950’s, the wave equation method of analysis using digital computers was developed and gave more dependable results since it was based on more accurate hammer and pile models, although the soil model was empirically developed. Personal computers have made wave equation analysis a readily available tool for estimating capacity and driveability. Wave equation analysis requires assumptions about hammer system performance and soil behavior models. Unfortunately, these unknowns can cause considerable variation in results when for example the hammer efficiency from a poorly maintained
327
hammer is greatly different than assumed. Fortunately, we can now simply measure the hammer performance and driving stresses as a result of research beginning in 1964 at Case Western Reserve University in Cleveland Ohio under the direction of Dr. G.G. Goble (Goble et al. 1975, Goble et al. 1980). The initial goal of the Case research project was to measure the pile capacity using the pile hammer as the loading device. Capacity was evaluated by both closed form solutions and discrete numerical analysis and correlated with hundreds of static tests to refine the procedures and establish databases to assure the necessary reliability. In addition to estimating capacity, the equipment developed over many years into a complete inspection of hammer performance, driving stresses and pile integrity and is known today as a “Pile Driving Analyzer@”,or simply the ‘‘PDA”. 2 CURRENT PROCEDURES FOR MEASUREMENT
The Case research project resulted in both closed form “Case Method” solutions (for capacity, energy transfer, driving stresses and pile integrity) and rigorous numerical modeling CAPWAP’ software. Both required measurement of force and velocity of the pile. These parameters are routinely obtained by measuring strain and acceleration with the PDA using reusable bolt-on sensors for both strain and
Figure 1. Windows PDA processing program. acceleration quickly attached to any pile size and pile type in any weather condition. Sensors for underwater pile testing applications have proven reliable (Harnar et al. 1996). Recent advances include “smart sensors” which remember their calibration information and transmit it to the PDA. As an alternative to measuring strain and converting to force, top transducers used in the early research project which were sized to the pile dimensions and placed between the hammer and pile to transmit and measure force are now being seriously considered again. The signals are processed and data acquired by a PDA. Modern PDA systems are PC based with large mass digital storage and high resolution graphic screens. Data is acquired and analyzed by the user friendly Windows PDA-W program as shown in Figure 1. The PDA is traditionally operated by a trained engineer who travels to the job site with the equipment, prepares the pile, and attaches the sensors to the pile. In recent times, pile preparation and especially sensor attachment are often done by the pile driving crew. The time required to drill the holes and attach the sensors is often only 5 to 15 minutes per pile tested (substantially less effort than a static load test). The engineer first evaluates the data for quality. The PDA program checks for loose connections, unstable results, various ratios including the known theoretical proportionality between force and velocity at impact. If problems are detected, a warning is given, and corrective action can be taken. Once the data quality is satisfactory, the engineer interprets the data on-site using the PDA. He gives advice or opinions and answers questions that prompted the testing request. Following the on-site test, the engineer may further analyze the PDA data with CAPWAP, a signal matching computer algorithm which extracts the soil model from the measurements and provides a simulated static load test result.
Being on-site gives the engineer a feel for the project. He sees how the piles drive, the condition of the hammer, and the care that the crew gives to the installation process. However, this on-site process requires a considerable cost including hours of travel time and large travel expenses. The engineer usually arrives prior to the first test pile being driven. Unfortunately, this is often a time which is at best a guess due to weather, pile delivery, and assembly of the driving system. Further, the contractor’s efforts to drive the first pile may be delayed due to a variety of reasons or equipment problems which are usually resolved during the test pile program and prior to production pile driving. In summary, during the test pile program considerable time is spent simply waiting. The active time for the PDA engineer on-site is often small.
Figure 2. PAL on site (left) with cell phone (right)
3 NEW METHOD FOR DYNAMIC PILE
TESTING A new version Pile Driving Analyzer (PDA) called PAL has many powefil features to revolutionize dynamic pile testing. This smaller unit has a touchscreen for user friendly data entry and a self contained rechargeable battery sufficient for a full day’s operation. Data is stored on a removable memory card. This PDA can be connected by the engineer on-site directly to a laptop running the Windows PDA program to provide full on-site PDA processing capability. However, modern technology now allows for a more cost effective approach to dynamic pile testing. The PAL on-site can be remotely operated through cell phone technology by an engineer in the office, see Figure 2. The sensors are attached to the pile by the crew on-site or a trained technician. The pile sensors are connected to the PAL, which dials the cell phone and connects to the office PDA engineer.
From this point on, the PDA engineer operating the PDA-W program on his office PC controls the remote PAL to collect the data and send it digitally to the office. The office engineer sees the data from the on-site remote PAL in real time as the test is in progress. Site observations are communicated with the office engineer with a cell phone or by the PAL’S “message communicator” (which sends short messages either preprogrammed or user constructed). In turn the office engineer interprets the PAL data and communicates his advice to the sit: pr.ctic.!!y jnstantaneously. Why test in this remote mode? Simply stated, there are significant cost savings. In many cases, the cost to the owner of the engineer’s travel time to site and the travel expenses approaches the cost of the test itself. Since the pile crew can attach the sensors to the pile, the PDA engineer could operate the PDA from his office and save this time and travel cost. With reduced costs, the project owner can perform more testing for the same total cost. The PDA engineer does not waste time waiting on-site for the test. But perhaps more importantly, the data analysis can begin immediately after data collection resulting in earlier availability of final results to speed up the construction and decision process. The report can be issued sooner. On large projects, the PDA engineer may still travel to the site for the initial test piles. However, subsequent production pile quality assurance testing can be performed remotely. This results in lower testing costs and easier scheduling of routine periodic tests. Thus, more tests can be performed at lower costs resulting in improved quality assurance for the project. 4 DYNAMIC TESTING USES AND GOALS Obviously, whether testing on-site or remotely, the dynamic test must be performed with the correct goals to achieve the maximum benefit. Objectives of high strain dynamic pile monitoring during installation include evaluation of the energy transfer and driving stresses to assure the pile can be safely driven to the desired depth without damage. Capacity may be of interest during driving to establish the bearing layer or establish the driving criteria. However, for long term capacity evaluation, restrike data is usually required due to soil strength changes with time. During driving, the PDA calculates energy transfer from the integral of the product of force times velocity. The energy transferred is then compared with the hammer’s rated energy. When low hammer performance is indicated, the hammer can be serviced (cushion, compressor and hoses, air valves, fuel pumps, or piston rings inspected or corrected, etc.) or the cause of low performance determined and corrected so that hammer performance and thus 329
pile installation productivity is improved. Maximum driving stresses are investigated to reduce the likelihood of pile damage. The PDA gives the direct compressive maximum at the sensor location. Since up to four strain sensors are attached to the pile and monitored separately, bending and local contact stresses can be assessed and hammer pile alignment improved. The maximum tension stress below the sensors for concrete piles is computed from the pile top measurements. If tension is excessive, a lower hammer stroke, or increased pile cushion thickness may be required. The maximum compression stress is estimated at the pile bottom for all piles driven to hard bearing layers. These stresses need little interpretation other than comparison with stress limits imposed by code or specification. The PDA can investigate the pile shaft for damage. Evaluation of pile bearing capacity is more complicated because the soil strength is often altered by the installation process. The capacity during driving is often less than the long term pile capacity particularly for piles driven in fine grained soils (clays, silts and even fine sands) due to excess positive pore pressures generated during driving which reduce the effective stresses. As these pore pressures dissipate after driving ends, the pile shaft resistance increases. Capacity reduction during driving is also caused by lateral pile motions which create an oversized hole; with time, the overburden pressures reach equilibrium on the pile perimeter and increase shaft resistance. This phenomena of capacity gain with time is called soil setup. Therefore, dynamic testing during restrike tests after a sufficient wait period usually yield a better indication of long term pile capacity than a test at the end of pile driving. The wait time required is longer as the soil grains become finer. Although less common, relaxation (capacity reduction with time) has been observed. Relaxation can be a serious problem for piles driven into weathered shale, and may take several days to h l l y develop. These losses can be caused by exposing the shale to water, due to fracturing of the rock or by heave from driving adjacent piles. Pile capacity estimates based upon initial driving can significantly over predict long term pile capacity. Therefore, piles driven into shale should be tested after a minimum one week wait either statically or dynamically (with particular emphasis on the first blows). Relaxation has been observed for displacement piles driven into dense saturated silts or fine sands due to a negative pore pressure effect at the pile toe. Restrike tests after a few days are usually sufficient, with emphasis on early “high energy” blows. With sufficient experience on a site or in some limited geographical area, the engineer may eventually reduce the amount of restrike tests and just apply a reduction factor to the end of drive capacity to estimate the final resistance available.
Dynamic load testing indicates the activated or mobilized pile capacity at the time of testing. At very high blow counts (above about 10 blows per inch, or less than 2.5 mm set per blow), dynamic test methods tend to produce lower bound capacity estimates as not all resistance (particularly at and near the toe) is fully activated. This can occur when piles are driven to rehsal or in cases where setup increases are large and the hammer is then not capable of moving the pile during testing. Several solutions to overcome this under prediction dilemma for refbsal conditions are available depending on the site conditions and availability of equipment. One solution is to apply a few blows at higher energy. The higher energy can be generated from a higher stroke, or a larger hammer or a big drop weight with more energy resulting in a smaller blow count (less than 10 blows per inch set; greater than 2.5 mm set per blow), and thus mobilize the full capacity. Another solution for closed end steel pipe piles at refusal is to fill the pile with concrete and then test as a composite section. The now stiffer pile will have a higher applied force and thus overcome more resistance. Another method is to add the end of drive end bearing to the restrike shaft resistance by superposition to estimate the service condition total load (Hussein 2000); this should only be done if the restrike is at rehsal conditions and where no toe relaxation is possible. Alternately, if setup (and relaxation) can be estimated, the pile need not be driven to the full required ultimate capacity but rather only to a usually reduced “target” capacity or to refbsal. A CAPWAP analysis usually is made to confirm the PDA field capacity result. The soil is modeled similar to wave equation methods. The wave transmission is modeled by the method of characteristics (CAPWAP Manual 1999). The hammer model is replaced by the measured force and velocity as a boundary condition. Since these measurements are redundant, the soil model can be iteratively investigated. If the measured velocity is input into the CAPWAP analysis model, then the force required to hold the system in dynamic equilibrium can be computed and subsequently compared with the measured force (usually the wave down is input and the wave up is computed and compared). The soil model is adjusted either automatically by the program or by the engineer until the computed and measured forces (or wave up) are in agreement. The final soil model (distribution and dynamic and static parameters) then describes the soil behavior during the hammer impact. CAPWAP has been proven to have good agreement with static load test results (Likins et al. 1996). The CAPWAP pile and soil models can be subjected to a simulated static loading to produce a load movement curve comparable to static test. CAPWAP also provides resistance distribution
results where negative friction (downdrag) may be a concern. Numerous factors are usually considered in pile foundation design. Some of these considerations include additional pile loading from downdrag or negative skin friction, soil setup and relaxation effects, cyclic loading performance, lateral and uplift loading requirements, effective stress changes (due to changes in water table, excavations, fills or other changes in overburden), settlement from underlying weak layers and pile group effects. These factors merit consideration when considering the interpretation of dynamic testing results. The foundation engineer should determine if any of these considerations apply to his design. 5 USAPRACTICE
The USA is interesting since practically every possible pile-hammer-soil combination can be encountered. Most PDA testing in the USA has been on driven piles with the testing engineer on-site with the PDA. A considerable experience has been accumulated. Early tests from the research project at Case included extensive correlations with static load tests. Private consultants began in 1972 to apply the method in their own private sector projects, either testing themselves or specifling testing in project documents (often resulting in the contractor hiring a testing firm to provide the service). Based on favorable experience and the results of the research project, highway agencies specified dynamic testing. With time and the support of the Federal Highway Administration for this testing, more highway agencies have included dynamic testing in their practice, either by acquiring the equipment or by hiring outside consultants. As a result of this varied exposure, many contractors see a benefit and now hire testing firms to assist them when piling problems arise, or they often propose dynamic testing as an alternate when static testing is specified. For small projects with only a few piles, a couple piles per structure are tested. Since the time to install all piles is relatively short, often the piles are tested during driving or with relative short wait times during restrike. For medium sized projects, the first production piles often serve as dynamic test piles and are distributed over the site to check site variability. Usually some restrikes are included in the test program. In a growing number of cases, longer wait times before restriking are being employed to take more advantage of the ususal strength gains with time. For larger projects, the amount of static testing is generally reduced after establishing a correlation of dynamically and statically tested piles, and then supplemented by additional dynamic tests to increase the total
330
percentage of piles tested. This improves the overall quality assurance while reducing testing costs. Because of the large amount of dynamic testing performed, various codes and specifications are now in place for USA application. In 1986, the D4945 consensus standard was adopted by ASTM (American Society of Testing and Materials) for “High Strain Dynamic Testing of Piles.” Testing in the USA follows these guidelines (ASTM D4945-96 is most recent revision). Beginning in the 19803, the Federal Highway Administration offered a demonstration project for dynamic pile testing. They currently provide workshops for State highway agencies which includes dynamic methods as an important part of their recommendations. Many State highway agencies have developed their own specifications for use. Florida DOT has probably the most extensive code and PDA is an integral part of their quality control procedures. AASHTO has general specification T298 (similar to but improved version of ASTM D4945). In additional they recognize dynamic testing with its own reliability and benefit when assigning safety factors to foundations in the “AASHTO Standard Specifications for Highway Bridges”. In general, higher levels of testing result in a lower safety factor. The US Army Corps of Engineers included dynamic testing in their 1993 manual on “Design of Pile Foundations”. The American Society of Civil
Figure 3. Drop weight test on augercast pile. 331
Engineers (ASCE) developed a consensus standard for pile testing entitled “Standard Guidelines for the Design and Installation of Pile Foundations”(ASCE 20-96) which includes dynamic pile testing. The private sector work generally has followed the local highway department practice and the three regional building codes. A recent joint effort of these three regional code authorities resulted in the International Building Code (IBC) with application in April 2000. IBC allows ASTM D4945 testing to evaluate static capacity. Before IBC adoption, many consultants employed dynamic testing on their private projects for either capacity evaluation or monitoring hammer performance and driving stresses Two USA organizations (Deep poundations Institute and Pile Driving Contractors Association) have endorsed PDA testing. DFI produced a consensus document “Inspector’s Manual for Driven Pile Foundations (Second Edition 1997)” with positive comments on dynamic testing. PDCA has promoted a model code (dated 1999) in both working stress and LRFD versions which allows substitution of static tests by dynamic tests aRer a proven correlation with considerable reductions in safety factor and thus cost savings for increased percentages of piles tested. It is conservatively estimated that every year several thousands of piling projects have dynamic pile testing on driven piles. While occasional PDA dynamic testing for drilled piles in North America has been made starting in 1974, this procedure is now routinely applied on drilled shafts and augercast piles in many countries in Asia, Europe and South America using drop weights (Rausche et al. 1985, Hussein et a., 1996). Figure 3 shows high strain acceleration and strain sensors attached to an augercast pile. The pile was extended above the ground surface with a thin liner. The lower section of the liner was then removed and the sensors attached to the resulting smooth concrete using anchors in the same manner they are attached to driven concrete piles. The pile top surface is usually flat and relatively smooth and only needs some minimal plywood cushion to distribute the impact over the entire top surface. A steel plate is then placed above the plywood as a striker plate for the impact weight, If reinforcement protrudes from the pile top, the pile can be built up above the reinforcement and then removed after the test. To perform the high strain test, the drilled shaft or augercast pile is then subjected to an impact of an impacting weight. In most cases a simple drop weight is preferred. Figure 3 shows a two ton drop weight (four H piles welded together) being positioned to test an augercast pile. Other drop weight designs include solid steel cylinders, concrete filled steel pipes or heavily reinforced concrete blocks. As a general guide, the weight should be at least 1 to 1.5% of the desired ultimate capacity to be
economy. In some countries the ultimate pile capacity has been greatly increased as a result of more dynamic testing; for example, in Sweden the allowable loads for the same identical piles have approximately doubled as a result of codes requiring higher percentages of piles to be tested by PDA. There are numerous country specific codes detailing application of dynamic PDA testing (Beim et al. 1998). Monitoring hammer performance for offshore oil platform installation is regularly specified by the oil companies; if driving is greatly different than expected, then capacity is further evaluated.
proven (Hussein, 1996) to assure load activation at reasonable stresses. Larger existing weights can be used provided the weight and shaft diameters remain about comparable. Regardless of size, shape or composition, the drop weight is generally guided to an axial impact by a short set of leads as in Figure 3, and is raised by cable and dropped by releasing the drum brake. An alternative and preferable drop method involves raising and securing the weight and then completely releasing it for a true free drop (e.g. releasing hydraulic jaws such as used for vibratory hammers used to grab steel piles, or by tripping a simple mechanical release). The test for a drilled shaft usually consists of a few separate impacts. A low drop height is first applied to assess signal quality and alignment of the weight with the shaft. After each impact, the net permanent displacement or “set per blow” is carefdly measured to evaluate full capacity activation. Compressive stresses are compared with the concrete strength. Alignment adjustments are made if necessary and a second higher drop height is applied. The test continues with increasing drop heights until either the set per blow exceeds a value sufficient to insure the full capacity activation, or until the indicated capacity is above the required ultimate capacity, or until the stresses become too large and the risk of pile damage is then too high. Most tests are completed in less than five impacts. If the pile top has been built up to accommodate the dynamic test, the extra top section is removed to facilitate completing the foundation. The measured pile top strain and velocity data are analyzed by CAPWAP to independently check the total capacity mobilized for each blow. A CAPWAP analysis can be performed in a short time on site after each impact to determine if the set per blow is low so that the full capacity has not yet been activated and another larger impact is required. Upon completing the CAPWAP analysis, a simulated static load test is obtained. 6 WORLDWIDE PRACTICE
PDA use outside the USA revolves around established local practice. In many locations, testing drilled shafts as described above is the primary application. In some locations, testing of driven piles is predominate and follows procedures common in the USA. Many contractors drive precast segmental regularly reinforced concrete piles. Often these contractors obtain the equipment and perform the testing themselves. In many cases a “design-build’ process is common and the contractor is then encouraged to find better foundation solutions and is fully responsible for the foundation installation; in such cases many have found great benefit in dynamic testing to assure quality and
7 CONCLUSIONS Dynamic pile testing with the PDA with subsequent CAPWAP analysis has become a routine practice for engineers and contractors worldwide. The methods have been applied to driven piles during driving to monitor the hammer and driving stresses. Because pile capacity is a function of time due to changes in soil strength due to effects of the driving process, long term capacity is usually evaluated during restrike several days after initial installation. In many parts of the world, dynamic pile testing methods have been successfully applied to drilled shafts and augercast piles by applying an impact of a large drop weight. To activate the full soil resistance (and thus correlate best with a static test failure load) for either driven or drilled piles, the energy input must be sufficiently large to produce a 2.5 mm set per blow or more. Numerous codes and specifications now direct the proper application of dynamic testing. New technology has been introduced to allow the engineer in the office to remotely monitor dynamic pile tests on site.
REFERENCES Beim, J., Gravare, C.J., Klingmuller, O., Li, D.Q., & Rausche, F. 1998. Standardization and codification of dynamic pile testing, a worldwide review. Proceedings, seventh international confirewe on piling and deep .fo2tndation.r Deep Foundations Institute. Vienna, Austria. CAPWAP manual 1999. Goble Rausche Likins and Associates, Inc. Cleveland Ohio USA. Goble, G., Likins, G.: & Rausche, F. 1975. Bearing capacity of piles from dynamic measurements, final report. Case Western Reserve University, Cleveland Ohio USA. Goble, G.; Rausche, F.; & Likins, G. 1980. The analysis of pile driving - a state of the art. Proceedings, first international conference on the application of stresswave theory to piles, Stockholm, Sweden. Harnar, N.& Likins, G. 1996. Underwater dynamic Testing Experience. Proceedings, Fijih International Conference on the Application of Stress-Wave Theory to Piles, Orlando, Florida, USA.
332
Hussein, M., Likins, G., and Rausche, F. 1996. Selection of a hammer for high-strain dynamic testing of cast-in-place shafts.. Proceedings, F@li International Confirence on the Application ofStre.s.7- W m e Theory to Piles, Orlando, Florida, USA. Hussein, M., Sharp, M.R., & Knight, W.F. 2000. Superposition for evaluating pile capacity. Proceedings, sixth international conjkrence on the application of stress-wme theory to piles, Sao Paulo, Brazil. Likins, G.: Rausche, F.: Thendean, G.>& Svinkin, M., 1996. CAPWAP correlation studies. Proceedings, Jisth international conference on the application of stresswave theory to piles, Orlando, Florida, USA. Rausche: F., Goble. G. 8r Likins, G. 1985. Dynamic determination of pile capacity. ASCE joiirnal of geotechnical engineering, Vol. 11 1: March 1985.
Other Referenced Standards: ASTM D4945-96. Standard test method for high strain testing of piles. American Society of Civil Engineers, 1997. Standard guidelines for the design and installation of pile foundations, ASCE 20-96. US Anny Corps of Engineer~~1993.Design of pile foundations, Engineering manual (EM 1 10-2-2906). U.S. Department of Transportation, Federal Highway Administration, 1996. Design and construction of driven pile foundations. Publication No. FHWA-HI-96-033, two volumes. American Association of State Highway and Transportation Officials, "AASHTO", 1993. Standard method of test for high strain dynamic testing of piles. AASHTO Designation T 298-93. American Association of State Highway and Transportation Officials "AASHTO", 1996. Standard specifications for highway bridges, (Sixteenth edition). Deep Foundations Institute, 1997. Inspector's manual for driven pile foundations, (Second Edition). Pile Driving Contractor's Association (PDCA), 1999. Design specifications for driven bearing piles. Canadian Foundation Engineering Manual, 3rd Edition, 1992. International Building Codc, 2000 Section 1807.2.8.3.
333
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) (CJ 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
Dynamic load testing and Statnamic load testing for acceptance and design of driven piles in Japan T. Matsumoto - Department of Civil Engineering, Kcrnazawu University,Japan K. Fujita - Department of Civil Engineering, Science University of Tokyo, Noda, Japan 0.Kusakabe -Department of Civil Engineering, Tokyo Institute of Technology, Japun M.Okahara - Public WorksResearch Institute, Miizistry of Coizstruction, Tsukuba, J ~ p t z N. Kawabata -Nippon Steel Corporation, Tokyo,Japan S.Nishimura - Fugro Geoscience Company Limited, Tokyo, Japan ABSTRACT: This paper presents the applications of dynamic pile load testing and the Statnamic test in Japan, in view of the acceptance and the design of driven piles. This paper covers 1) a historical review of research activities concerning pile load tests, 2) definitions of the terms “dynamic load test” and “rapid load test”, 3) aspects of dynamic load testing and 4) interpretation of the Statiiamic test. Emphasis is placed on the use of soil investigations and soil tests in parallel with the dynamic load test and the Statnamic test to obtain a more reliable load-displacement curve from the tests. Also discussed is tlie importance of conducting redriving tests to estimate the load-displacement curve after ‘set-up’, the increase in the pile capacity with elapsed time after pile driving, for piles driven in saturated soils.
This paper is coniposed of 4 parts: 1 ) a historical review of pile load test research in Japan, 2) definitions of dynamic and rapid loading, 3) aspects of dynamic load testing, and 4) interpretation of the Statnaniic test.
I INTRODUCTION Intensive research activities on the dynamic load test (DLT) have been conducted during the past twoe two decades. and research activities on the Statnaniic test (STN) have been conducted during the past decade in Japan. These research activities are accompanied by a change in foundation design in Japan from tlie conventional allowable stress design to the limit states design and the performance based design. In the new design concepts, accurate estimation of loaddisplacement curve, as well as quality assurance of pile foundations, will become issues of vital importance (Kusakabe 1998). Actually, research activities on the dynamic load test and the Statnamic test in Japan have been focused on the estimation of the static load-displacement curve for a pile from tlie test signals. This paper presents the application of the dynaniic pile load test and the Statnamic test in Japan, in view of the acceptance and the design of driven piles. Emphasis is placed on the use of soil investigations and soil tests in parallel with the dynamic load test and the Statnaniic test signals to obtain a more reliable load-displacement curve froin the tests. Also discussed is the importance of conducting re-driving tests to estimate the load-displacement curve after ‘set-up’, the increase in tlie pile capacity with elapsed time after pile driving, for piles driven in saturated soils.
2 HISTORICAL REVIEW OF RESEARCH ACTIVITIES CONCERNING PILE LOAD TESTS IN JAPAN
Principal research activities in the area of foundation engineering in Japan are listed in Table 1. Research activity on dynamic load testing was initiated in 198 I by the Japanese Geotechnical Society (JGS). The Committee on Estimation of Bearing Capacity of Steel Pipe Piles from Dynamic Load Test was formed in the Japanese Association for Steel Pipe Piles in 199 1 led by Prof. Kazuma Uto of Toltai University. They conducted dynamic load tests on steel pipe piles having various diameters and lengths driven in various soils with various types of hammers, to compare the bearing capacity and the load-displacement curves derived from the dynamic load tests with the results of the static load tests of the piles during 1991 to 1995. They also conducted their own test program in 1993, in which static, dynamic and Statnaniic tests on steel pipe piles driven in sand were conducted to investigate the ‘setup’ phenomena and the performances of the various
335
test methods. The results of these research activities are to be presented in two papers at this Conference. The first Statiiamic tests in Japan were performed in 1992 on a cast-in-situ concrete pile (Chosokabe et al. 1993) and on a steel pipe pile (Matsumoto et al. 1994). In 1993, a private joint research group (Research Group on Rapid Pile Load Test Methods) was formed led by Prof. Osamu Kusakabe of Tokyo Institute of Technology, aiming to compile the existing knowledge about rapid pile load tests and examine basic characteristics and applicability of the tests, as well as the scientific interpretations of the Statnamic test results. The Research Group coiiducted their own research program, in which several factors which might influence the Statiiamic test results were investigated in the field, such as type of pile, the installation method, the loading history, the curing period and the soil conditions (Kusakabe & Matsuiiioto 1995; Kato et al. 1998). Based on the accumulated research results and the continuing efforts of the Research Group, JGS
agreed to form the Research Committee on Rapid Load Test Methods in 1996 to critically review the features of rapid load tests, including the Statnamic test, in more general teims, and to prepare a draft document of testing standards for rapid load tests. In 1998, the Committee on Standardization for Vertical Load Test of Piles was formed in JGS for the revision of the current standards for vertical load tests of piles, with the scheduled enforcement of the revised standards in the year 2000. The revised standards will iiicludes dynamic load tests and rapid load tests, as well as various types of static load tests. The Research Group and the Canadian Embassy in Japan co-organized the second International Statiiamic Seminar in Tokyo in 1998. More than 40 articles were submitted to the Seminar from 11 countries, showing the large increase in the use of rapid load test in the world. A technical session on the Global Srctndardization of Statnamic Testing was held during the Seminar, aiming to have a global equivalency of standards for rapid load tests.
Table 1 , Research activities for pile foundations by Japanese Geotechnical Society (JGS) and other institutes in Japan. Year Institute Event Committee 011Standardization of Lateral Load Test of Piles. JGS (JSSMFE) 1980 - 1982 Research Committee on Bearinrr, Capacity of Piles Installed by Construction Methods with JGS (JSSMFE) 1980 - 1982 Low Vibration and Low Noise. Chairman ofTechnica1 Corninittee on f’enctrability and Drivability of Piles of ISSMFF JGS (JSSMFE) 198 I - 1985 (Prof. K. Fu,iita) JGS (JSSMFE) Sub-Committee on Wave Propagation Theory. 1982 - 1989 National Symposium on f’enetrability and Drivability of Piles, Tokyo. 1984 JGS (JSSMFE) Committee on Standardization of Tension Load Test of Piles. 1 986 - 1987 JGS (JSSMFE) Chairman of Technical Committee on Pile Driving of ISSMFE. JGS (JSSMFE) 1986 - 1989
; 1987 - 1991 1987 1989
JGS (JSSMFE) JGS (JSSMFE) JGS (JSSMFE)
199 1
JGS (JSSMFE)
Committee 011 Standardization of Vertical Load Test of Piles. 32nd JCS Symposium oill~oundationsSupported by Friction Piles. National Symposium on Pile Drivability and Application of Stress-Wave Theory to Piles, Tokvo. National Symposium on Vertical Load Test of Piles and Decision of Bearing Capacity.
Committee on Estimation of Bearing Capacity of Steel Pipe Piles from Dynamic Load Test. First Statnamic test on cast-in-situ concrete pile in Japan. First Statnamic test on steel pipe pile i n Japan. Research Committee on Limit State Design of Foundation Structures. 1992 - 1995 Comparative study on static, dynamic and Statnamic load tests on steel .pipe 1993 . .piles driven in sand at Masaki test yard of’ Sumitorno Metal Industry. Research Group on Research Group 011Rapid Pile Load Test Methods. 1993 - 2000 Rapid Pile Load Test Methods Research Committee on Rapid Pile Load Test Methods. JGS 1996 - 2000 Sub-Committee 011Seismic Design of Pile Foundations. 1996 - present Architectural Institute of Japan Research Committee on I’resent and Future of Japanese Foundation Design and Soil 1997 - present JGS Investigation in view of International Equivalency. Committee on Standardization for Vertical Load Test of Piles. 1998 - present JGS 1998. 10 Research Group on 2nd International Statnamic Seminar, Tokyo. Rapid Pile Load Test Methods Publication of Standard for Vertical Load Test of Piles. 2000 JGS (Impact load test (dynamic load test) and rapid load test are included in this standard.) Japanese Society for Soil Mechanics and Foundation Eng. (JSSMFE) was renamed Japanese Geotechnical Society (JCS) in 1997. JASPP: Japanese Association for Steel Pipe Piles 1991 - 1995 1992
JASPP Takenaka Corp. Kanazawa Univ. JGS JASPP
336
Figure 1. Classification of pile load test inethods in t e r m of relative loading duration (after JGS Research Committee on Rapid Pile Load Test Methods 1998).
Above. we have reviewed the history of research concerning pile load testing in Japan. In the next section, we turn to the theoretical definitions of dynamic and rapid load testing used in Japan. 3 DEFINITIONS OF DYNAMIC LOAD TEST AND RAPID LOAD TEST When the dynamic load test and the Statnamic test are used to derive a static load-displacement curve for a pile, allowance for dynamic effects should be made in the interpretation of the test signals. The difference between the dynamic theory and the static theory is whether inertial forces are considered or not, in the sense of physics. However, in the field of geotechnical engineering, 'dynamic effects' also include such things as (1) the strain rate dependency of soil strength and deformation characteristics. including creep, (2) the viscous effect on interface strength between the pile and the soil, arid (3) changcs in the drainage condition of the soil, in a broad sense. It does not seem that current pile technology and geotechnical engineering fully cover such 'dynamic effects' observed in dynamic load testing, Statnainic testing, and even in static load testing. The degree of influence of each dynamic effect depends on the penetration rate, the acceleration of the pile, and the loading duration. Several interpretation methods are available for the dynamic load test and the Statnainic test, such as applications of finite element method (FEM), onedimensional stress-wave theory and single rigid inass modeling of the pile. Of course, these interpretation methods have limitations in their application in practice, accordiiig to the purpose and the loading condition, as well as the soil conditions. Therefore. it may be necessary to classify pile load test methods
in terms of loading and soil conditions, in order to select the appropriate interpretation method for each pile load test. The Research Committee on Rapid Pile Load Test Methods (1 998) proposed a practical classification of pile load test inethods shown in Figure 1 in terms of relative duration, t,., based on the works of Holeymaii (1993), Middendorp & Rielefeld ( I 995), Karkee et al. (1997), Nishimura et al. (1998) and Matsurnoto (1998). The relative duration, t,., is an index for the loading duration relative to the traveling time for a stress-wave up and down in the pile, and is defined as follows (Karkee et al. 1997): I,. = T 1 ( 2 L l c )
in which T is the loading duration as defined in Figure 2, L is the pile length and c is the bar wave velocity of the pile. The relative duration, I,, is identical to the relative wave length. A, defined as (IHoleyman 1992) :
Each bar in Figure 1 indicates the range of relative duration for which the corresponding dynamic effect is not negligible. The range of stress wave phenomena in the pile which are negligible was determined from the works of Middendorp & Bielefeld (1 995) and Nishimura et al. (1998). They carried out simulations of onedimensional stress-wave propagation in piles and found the relative duration beyond which wave propagation phenoinena are negligible. The range of the influence of the generation of excess pore pressure is based on Holeynian (1 992) which suggested the relative wave length, A, necessary for 90% of excess pore pressure to dissipate for various 337
loading, for concrete and steel piles. Typical loading dlirations of the Statnamic test fall in a range from approximately 100 to 200 ms. Therefore, the Statnamic test can be regarded as a rapid load test, in most cases. In section 4 below, dynamic load testing in Japan will be discussed and in section 5, an example of rapid load testing, the Statnamic test, will be taken up.
soil types. The range of viscous damping (the influence of the relative velocity between the pile
4 DYNAMIC LOAD TESTING IN JAPAN 0
30
60
90
120
150
I n this section, test equipment for gathering dynamic test data, methods used for interpretation of this data, and the use of the results of the interpretation for acceptance and design of piles will be discussed.
Time (ms) Figure 2. Definition of loading duration. Table 2. Typical loading duration for the boundary between dynamic and rapid loading, and for the boundary between rapid loading and static loading. Pile type Bar wave velocity, c Typical range of pile length, L Loading duration for t,. = 5 L,oading duration for t,. = 500
Concrete pile 4000 ni/s 10 - 50 n1 25 - 125 ins 2.5 - 12.5 s
4.1 Test equipment
Steel pile 5 120 m/s 10 - 100 111 20 - 200 ins 2 - 20 s
and the soil on the increase in shaft resistance) was roughly estimated based on the proposal by Randolph & Deelts ( I 992). 'rhe boundary between dynaniic loading and rapid loading (Statnamic loading) may be regarded as equivalent to the boundary between the loading durations where the stress wave propagation phenomena in a pile are negligible, and those where they are not. This Research Committee on Rapid Pile Load Test Methods regards 1, = 5 as the relative loading duration representing the boundary betwxn dynamic loading and rapid loading, based on the works by Middendorp et al. (1 995) and Nishimura et al. ( 1 998). It should be mentioned that even if the wave propagation phenomena in a pile are negligible f o r I , 2 5, the inertial force of the pile as a concentrated mass can not be neglected. The boundary between rapid load testing and static load testing can be regarded as equivalent to the boundary between the loading durations where the soil resistance dependent on the inertial force of the soil and the rate dependency of the soil strength can be virtually neglected, and those where they can 1701. The Research Committee regards t, = 500 as the relative loading duration representing the boundary between rapid loading and static loading, based on the proposal by Middendorp & Bielefeld ( I 995). As a consequence, rapid loading is defined as a load test having the relative loading, I,, of 50 5 I, I 500. Table 2 shows the typical loading duration for the boundary between dynamic and rapid loading. and [or the boundary between rapid and static
Dynamic load test equipment commercially available in Japan are the PDA system developed by GRL Corporation and the FPDS system developed by TNO that are widely used all over the world. In these dynamic monitoring systems, changes with time of accelerations and strains near the pile head are measured, and they are converted to the time variations of the velocity, v, and the force, F. It is not uncommon that the pile is mounted with strain gages to measure the pile forces in a static load test. When the dynamic load test is conducted on such a pile. the variations with time of the strains measured at two levels of the pile near the top are utilized to obtain the amplitudes of downward traveling and upward traveling stress-waves at the pile head (for example, Hayashi et al. 1994) from the two-point strain measurement (Lundberg 1984; Matsuinoto et al. 1992).
338
4.2 Interprettrtiori nzet1iod.s qf. tJynmTic load test J ignnls The measured dynamic signals are interpreted using the CASE method (Goble et al. 1975) and/or the wave matching analysis. It is seldom that the CASE method alone is used to estimate the bearing capacity of a pile, since the ./,-factor used in the CASE method, which allows for the dynamic resistance. is regarded to be too empirical, and the set-up phenomenon is not considered at all if the pile driving test is conducted during initial driving only. The result from the CASE inethod is usually calibrated with the result of the static load test of the pile, and it is used for driving control of the other piles at the site of the tested pile. The wave matching analysis of the re-driving test is regarded as more reliable and informative than the CASE method, because the load-displacement curve
as well as the bearing capacity are estimated through the wave matching analysis and the influence of the set-up phenomenon is thought to be reflected in the estimated load-displacement curve. Foundation engineers in Japan often use commercial computer programs such as CAPWAPC and TNOWAVE in which the soil model proposed by Smith (1960), shown in Figure 3, is usually employed, although extensions of the Smith model are available in the programs. Criticisms of the commercial programs remain, which are summarized as fbllows: 1) the soil model proposed by Smith is too empirical, 2) the influence of soil inside an open-ended pipe pile on wave propagation in the pile is not accounted for, and 3) the difference in the drain condition of the soil between pile driving and static loading is not taken into account.
KWAVE, developed in Japan (Matsumoto & Takei 199l), is a computer program which takes into account the wave propagation in the soil inside thc pipe pile and the influence of the internal shaft resistance on the wave propagation in the pile. The pile-soil system adopted in KWAVE is shown in Figure 4. This pile-soil modeling follows Randolph (1 987). In the KWAVE program, the improved soil models for the shaft and base of the pile proposed by Iiaiidolph & Simons (1986) and Deeks (1992) respectively have been incorporated. Figure 5 shows the shaft model and the base model respectively. One of the rational soil models is that, from the soil test data such as the shear modulus, Poisson's ratio and the density, we can determine the values for the elastic spring and the radiation damping constant of the shaft model, as well as the elastic spring, the dashpot and the soil mass of the base model. A change in drainage condition between pilc driving and static loading may be approximatelj taken into account by using different Poisson's ratiox for pile driving and static loading (Matsunioto et al. 1997, Hayashi et al. 2000).
4.3 U,c of cIj~nai?iic loud test ~icc'eplunce~ i n dcles ign of piles
i ~ s i / I t . s for-
tlic
Fujita & Kusakabe (1988) have pointed out the following four unsolved problems in estimating the static bearing capacity from the dynamic load test signals: 1) the effect of the amount of elapsed time between the end of initial pile driving and static load test on the difference between the bearing capacities estimated from the dynamic load test and the static load test, 3) the effect of the method of the static load test on the results of the static load test, 339
3) the effect of the method used for determining the bearing capacity from the measured loaddisplacement curve on the determined bearing capacity, and 4) the effect of the behavior of pore water pressure during and after pile driving on the load displacement curve. The problems 1) and 4) have been vital issues in the use of the dyiiamic load test in Japan, because 63% of the dynamic load tests on steel pipe piles in Japan have been conducted offshore (Wakiya et al. 2000). Research concerning the problems 1) and 4) are discussed below. It is widely thought that the set-up phenomenon is largely related to the generation and dissipation of excess pore pressures around the pile during and after the pile driving, which result in the changes in the effective stresses in the surrounding soils and acting on the pile shaft. In the authors' knowledge. the first theoretical attempts to estimate excess pore pressures around a driven pile were made by Nishida (1963) and by Ladanyi (1963). In Japan, field measurements of pore pressures around driven piles wcre done by Koizumi et al. (1967), Wakiya et al. ( 1 992) and Matsumoto et al. (1 995) for clay soils and by JASPP in 1993 for a sandy ground (Shibata et al. 2000). These works clearly showed that the set-up plienomenon is largely controlled by the generation and dissipation of excess pore pressures around the pile during and after the pile driving. Figure 6 shows the increase in the static pile capacity with elapsed time after the end of initial pile driving (JASPP 1995). An open-ended steel pipc pile. 35 in in length, 610 niin in outer diameter and 12 mm in wall thickness, was driven with a diesel hammer in a clayey ground. The static pile capacity, R,, was derived from the wave matching analysis (CAPWAPC analysis) of the dynamic load test signals. The set-up ratio, defined as the ratio of the bearing capacity at re-driving test to the bearing capacity at the end of initial driving, attained a value of 5 about 10 days after the end of initial driving, and the set-up ratio leveled off after that time. Hayashi et al. (2000) also showed a similar result for steel pipe piles driven in a mudstone ground. It inay be still difficult to predict the increase in the pile capacity with elapsed time after the end of initial pile driving, although some theoretical solutions have been proposed for the set-up phenomenon (for example, Randolph & Wroth 1979). One reason for this may well be that haininer performance is also a factor for the set-up ratio, in addition to the pile configuration and the soil conditions. If a high performance haininer is used. the time required for the pile driving process is
340
Figure 6. Increase in static bearing capacity of a steel pipe pile driven in a clayey ground (JASPP 1995).
reduced, compared to the case in which a low performance hammer is used. That is to say, the niagnitudes of accumulated excess pore pressures generated at the end of initial pile driving depend on the performance of the hammer used, resulting in different set-up ratios for the same pile driven in the same ground with different hammers. There are several papers which report that the bearing capacity of a pile derived froin the dynamic load test for a pile is comparable with the bearing capacity obtained from the static load test, if the dynamic load test and the static load test are conducted a sufficient time interval after the end of initial pile driving.
Figure 7. Comparison of end and shaft capacities obtained from static load test with those estimated from dynamic Load testing and derived from various pile design codes for steel pipe pile driven in mudstone ground (Michi et al. 1996).
Figure 7 shows a comparison of toe and shaft capacities, Qrl and Q.$,obtained from the static load test, tlie dynamic load test and various pile design codes, for a steel pipe pile ( l l n i in length, 0.8m in diameter, 12mm in wall thickness) driven in a mudstone ground (Michi et al. 1996). The dynamic load test and tlie static load test were conducted 65 hours and 29 days after tlie end of initial pile driving, respectively. Figure 8 shows similar comparison for a steel pipe pile driven in a sandy ground (Shibata et al. 2000). It can be seen from Figures 7 and 8 that the bearing capacity and the Qp/Q5 ratio derived from the dynamic load test are the most accurate, compared to the empirical pile design formulas available in Japan and other countries. There are cases where conducting re-driving tests is difficult due to practical considerations. In such cases, calibration of the results of tlie initial dynamic load test on a pile with the static load test at the site is required to roughly estimate from the initial dynamic load tests the bearing capacity of tlie other piles after the completion of the set-up phenomena.
JHA: Japan Road Associatio~l JAC: Architectural Institute of Japaii JR: Japan Railway JPH: Japan Port and I-larbor Association Hiley ( I ) : Hiley's formula with use of potential hammer energy Hiley (2): Hiley's formula with use of transfered hammer energy Figure 8. Comparison of end and shaft capacities obtained from static load test with those estimated from dynamic Load testing and derived from various pile design codes for steel pipe pile driven in sandy ground (Shibata et al. 2000).
5 STATNAMIC LOAD TESTING IN JAPAN Here, we will discuss the interpretation of the Statnaiiiic test signals to estimate the static loaddisplacement curve of a pile The Statnamic tests which have been conducted in Japan from 1992 are summarized by Kato et al. (1998) and Nishimura et al. (2000). Kato et al. ( 1 998) compared tlie pile head force-displacement curves. FA/,,vs M', measured in the Statnamic tests on 34 1
various types of piles in various soil conditions with various construction methods with the loaddisplacement curves, P vs w, of the piles obtained from the static load tests. Let the discrepancy between Fsl,zvs w and P vs w be called the 'dynamic effect' in the Statnaniic test, here. Based oil these comparisons, they showed that the dynamic effect depends on 1) the acceleration of tlie pile, 2) the against penetration rate of the pile, 3) the level of FIo2 the static yield load, PJ,, 4) the proportion of shaft and toe resistances, 5 ) the stiffness of tlie pile shaft. 6) the soil type, and 7) the drainage condition. They surnniarized the 'dynamic effect' in tlie Statnainic test qualitatively as follows: (1) the dynamic effect is small when F,,,, is below P,,, (3)the dynamic effect is larger for the toe resistance than for the shaft resistance, (3) the dynamic effect is larger when the ground is saturated, especially for the toe resistance, and (4) the dynamic effect becomes larger as tlie pile penetration rate increases. And, they concluded that interpretation of tlie Statnaniic test signals is required to correctly estimate the static load-displacement curve from the Statnanii c test . Kato et al. (1 999) studied various interpretation methods of the Statnamic test signals, including the Unloading Point method, one-dimensional stresswave propagation analysis (wave matching analysis), FEM, and systematized their applicability in practice. The interpretation method mostly used in Japan is the Unloading Point method (Kusakabe & Matsumoto 1995), based on the single mass rnodeling of. the pile proposed by Middendoi-p et al. (1992). A weak point of this method is that any change in the drainage conditions between tlie Statnamic and static loadings is not taken into account. Therefore, the static load-displacement curve derived from the Unloading Point method is thought to closely correspond with the loaddisplacement curve obtained from the quick load maintained load test. FEM analysis is regarded as a promising interpretation method for the Statnaniic test, because elastic-(visco)plastic soil behavior and complicated boundary conditions are incorporated in the analysis. Although there still is room to improve tlie existing interpretation methods of the Statnamic test, the ineasurements of additional parameters such as axial forces down the pile shaft and the movement (or acceleration) of the pile toe are also useful to make the interpretation of the Statnamic test more re1i ab le.
6 CONCLUDING REMARKS It seeins that the Statnamic test is increasingly being viewed as an alternative to the conventional static
load test in Japan. It is not uncommon that the dynamic load test results are calibrated with the Statnamic test results. However, a large part of the Statnamic tests as well as the dynamic load tests are used only for the acceptance of the piles. It is undeniable that pile designers in Japan stick to pile design codes excessively. The fact that standards for the dynamic load test and the Statnamic test (rapid load test) have not been available in Japan also is an obstacle to wide uses of the new pile load test methods. Another aspect to be considered is pile design codes. There is no reduction in the safety factor in response to an increase in the number of load tests beyond 1. Furthermore, pile design is performed based on the bearing capacity alone, since vertical deformation of pile foundations is not specified explicitly in the design codes. However, this situation is currently improving. Actually, the JGS standards for vertical pile load tests are to be revised this year to include standards for rapid load testing and dynamic load testing. Furthermore, the Research Committee on Present and Future of Japanese Foundation Design and Soil Investigation in view of International Equivalency was formed in JGS in 1997, aiming to change foundation design codes from the allowable stress design to the framework of the limit states design and/or the performance based design. Quick, cheap and accurate pile load tests such as the dynamic load test and the Statnamic test are expected to play important roles in the trend of' changing foundation codes in which accurate estimation of the load-displacement curve as well as quality control of pile foundations are vital issues. REFERENCES Chosokabe, M., Yamashita, K., Kakurai, M., Fukuhara, T. and Yamada, T., 1993. A Statiiainic loading test applied for a cast-in-situ concrete pile. Proc. Annual Meeting of' Architectural Institute of ,Japan: 176 1- 1762 (in Japanese). Decks, A.J. & Randolph, M.F., 1995. A simple model for inelastic footing response to transient loading. /tit.
342
Proc. of 3rd Int. Cot$ of Application ofStress- W a v e Theory to Piles, The Hague, Netherlands: 195-2 15. JASPP, 1995. Methods for prediction of bearing capacity of steel pipe piles by dynamic load testing. Report of Committee on Methods for Prediction of Bearing Capacity (in Japanese). Kato, K., Kusakabe, 0. & Matsumoto, T., 1998. Characteristics of rapid load test appeared in various piles. Jozw. qf Geotech. Eng., JSCE, No.589/111-42: 155-166. Kato, K., Horikoshi, K., Matsumoto, T. & Kusakabe, O., 1999. Interpretation methods of Statnainic load test results. Jour. oj'Geotech. Eng., JSCE, No.624iIII-47: 267-282. Karkee, M., Horiguchi, T. & Kishida, H., 1997. Static and dynamic tests for evaluation of the vertical load bearing capacity of piles. Proc. Deep Foundufions Institute 22titl Annual Member 's Coi$erence, Toronto, Canada: 199-2 14. Kusakabe, 0. & Matsumoto, T., 1995. Statnamic tests of Shonan test program with review of signal interpretation, Proc. First Int. Statnaniic Seminar, Vancouver: 1 13- 122. Kusakabe, O., 1998. Changing foundation design code and the role of Statnainic test, Proc. 2nd Int. Statnainic Seminar, Tokyo, Japan. Balkeina (to be published in 2000). Koizuini, Y. and Ito, K., 1967. Field tests with regard to pile driving and bearing capacity of piled foundations. Soils and Founckrtion, JSSMFE, V01.7, No.3: 30-53. Ladanyi, B., 1963. Expansion of a cavity in a saturated clay medium. ./our. Soil Mech. Found. Div., ASCE, vol.89, No.4: 127- 16 1. Lundberg, B., 1984. On the separation of overlapping elastic waves from measured data, Proc. 2nd Int. Conj. on Application qf Sfres.s-M/uve Theory to Piles, Stockholm: 44 1-452. Matsuinoto, T. & Takei, M., 1991. Effects of soil plug on behaviour of driven pipe piles, Soils and Foundutions, V01.3 I , N0.2114-34. Matsuinoto, T., Sekiguchi, H., Yoshida, H., & Kita, K., 1992. Significance of Two-Point Strain Measurements in SPT. Soils and Foundations, Vo1.32, No.2: 67-82. Matsumoto, T., Tsuzuki, M. & Michi, Y . , 1994. Comparative study of static loading test and Statnaniic on a steel pipe pile driven in a soft rock. Proc. Int. Con( and Exhibition 017 Piling mid Deep Foundations, Bruge, Belgium: 5.3.1-5.3.7. Matsuinoto, T., Michi, Y. & Hirano, T., 1995. Perforinance of axially loaded steel pipe piles driven in soft rock. ./our. of' Geofech. Etig., ASCE, Vol. 12 1, No.4: 305-3 15. Matsumoto, T., Michi, Y. & Hayashi, M. (1997): Reliability of dynamic load testing on steel pipe piles in soft rock, Proc. 14th ICSMFE, Hamburg, Vo1.2, pp. 1 185-1 188. Michi, Y., Matsumoto, T. & Futatsuka, Y., 1996. Reliability of dynamic load testing compared with soil parameters: A case study on foundation piles of Noetsu bridge. Proc. 5th hit. Coiif'. 011 Application of Stress-Wave Theory to Piles, Orlando, USA: 465 - 479. Middendorp, P., Bermingham, P. & Kuiper, B., 1992. Statnamic loading testing of foundation piles. Proc. Of'dt/? bit. Conf of Applicufion of Stress- W a v e Theory to Piles, The Hague, Netherlands: 58 1-588. Middendorp, P. & Bielefeld, M.W., 1995. Statnaiiiic load testing and the influence of stress wave phenomena. Proc. of 1st Int. Statnariiic Seminar, Vancouver, Canada: 207-220. Nishida, Y., 1963. A basic calculation on the failure zone and the initial pore pressure around a driven pile in clay. Proc. 2nd Asian Conf: Soil Mech. Found. Engrg., Vol. I , JSSMFE, Tokyo: 2 17-2 19. Nishiniura, S., Matsumoto, T., Kusakabe, O., Nishiumi, T., Yoshizawa, Y., 2000. Case studies of Statnamic load testing in Japan. Proc. 5th Int. Cot$ on the Applicu/ioti q/' the S/i.e.s.s- Wave Theoiy to Piles (to be published). Randolph, M.F., 1987. Modeliiig of the soil plug response during pile driving. Proc. 8th S. E. Asian Ceotechiiicul Cot$, Bangkok, Vo1.2: 6.1-6.14. Randolph, M.F. & Simons, H.A., 1986. An improved soil model for one-di~nensionalpile driving analysis. Proc. 3rd Iiit. Con$ on Nunz. Meth. in Ojfihore Piling Nantes: I - 17. Randolph, M.F. and Deeks, A.J., 1992. Dynamic and static soil models for axial pile response. Proc. of3rd Int. Cot$ 017 Application qfStress- W u v e Theory to Piles,Hague: 3- 14.
Research Committee on Rapid Load Test Methods (formed in Japanese Geotechnical Society), 1998. Research activities toward the standardization of rapid pile load test in Japan. Statnamic loading test '98, Proc. 2nd Int. Statnamic Seminar, Tokyo, Japan. Balkema (to published). Shibata, A., Kawabata, N., Wakiya, Y., Yoshizawa, Y., Hayashi, M. & Matsumoto, T., 2000. Comparative study of static, dynamic and Statnamic load tests of steel pipe piles driven in sand. Proc. 5th Int. Conz on the Application of the Stress- Wave Theory to Piles (to be published). Wakiya, T., Hashimoto, O., Fukuwaka, M., Oki, T., Shinomiya, H. & Ozeki, F., 1992. Ability of dynamic testing and evaluation of bearing capacity recovery from excess pore poressure measured in the field. Proc. of 4th Int. Conj of Application of Stress-Wave Theory to Piles, The Hague, Netherlands: 665 - 670. Wakiya, Y., K.Nishiumi, M.Hayashi, A.Shibata & S.Nishimura 2000. The case studies of dynamic load test in Japan. Proc. 5th Int. Conf on the Application of the Stress-Wuve Theoty to Piles, Sao Paulo, Brazil (to be published).
343
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Ethics and money. Are they compatible? H-Goldernberg& J. J.Goldemberg Geotecnica Cientec, Buenos Aires, Argentina
ABSTRACT: From the point of view of a testing house ... Is the final objective of the construction industry the site itself or QA? New technologies appear and are offered to the market as the salvation- where in some cases they try to substitute the engineer’s eyes and knowledge- to assist the designer to verify if prediction matches performance. But is he exchanging money for a useh1 service or for a tick in the QA box? More piles are been tested each year all over the world. But do we know more about them or are we just collecting tons of data? Testing houses are getting more popular while their specialists write papers in every available publication. Are these engineers enough educated to interpret results and diagnose? What is the role of ethics, when the consultant is under pressure from every part involved and he is being paid by one of them?
A modern tendency is to supervise and assign a qualification to the evolution of the building site through the Quality Control, better known as QC. But what does it actually mean? Is it, maybe, the final objective of the site itself or the evolution of the quality of works done in order to be checked against the projected idea? Nowadays, when the fiemy of the regional economies requires low costs and time limits practically impossible to comply with, the QC concept generates a dichotomy with the demands of the market. Where this market, thanks to advertising and marketing, offers their products and services enwrapped by the I S 0 9000 halo; conditioning the building industry (producer of the container for these products or services) to join up to this trend. It is not the authors intention to disregard the necessity of quality in sites, but it is to call the attention saying that the QC, as well as the Quality Assurance (QA), requires indispensables items such as: an adequate project, organisation, methodology, aptness, responsibility and honesty by those involved in it. At first sight it seems that these ingredients have opposite directions to low costs and short duration. And us, in charged of the QA of deep foundations are not unaware of this phenomenon. In the cold world these words describe, the mercantilist conception of life prunes a new branch of what is human each morning and of our engineering judgement. In the last decades we have witnessed the evolution of science, technology, and knowledge.. .and
345
the ethics to the world of the ‘ant ethics’, where everything is the same and nothing is better. This unsolved equation, where the independent variables such us time, price and quality have been taken, gradually, to the distortion of the QA concept together with the proper behaviour of the Engineers, to obtain as a result a form where the QA box is ticked. That is to say interchanging money for a simple signed paper instead of a useful service. Notable contradiction since more piles are been tested each year all over the world. New technologies are being developed while the old ones are being updated. Various are offered to us as the panacea, the magic solution, the unique answer we were waiting for. The ‘technology’ that does without the engineering knowledge, feeling and judgement. This creates the proper conditions to take us far away from the purpose we were hired, which is to measure and inform about the state of the foundations. But even worse, it misrepresent the Engineer ideal and it highlights how degraded the human ethics is...as we all know that there are no magic solutions in our professional scope. But do we really know better our foundations or are we just collecting data? Are we contributing with vital information that will allow the Engineer clarify the real behaviour of the pile-soil interaction? Can we show the relation between prediction and performance? It is evident that the increasing tendencies all over the world to start leaving aside the prediction- as the last verdict summon- for the real measurement of the
behaviour, brought the increase in the testing houses with it. But does this greater demand to know the reality grew in the same way as the knowledge of those responsible summiting the verdict? Because technology did and what is most, it even superseded the expectations. Nowadays it is possible to find all kinds of equipment to perform even the most unthinkable measurements; any colour, size and origin. The evolution of science and technology has invaded our lives in such a way, that it is difficult to imagine our daily routines with out them. And Engineering cannot run away from this trend neither can our specialty . We blindly believe that technology will solve everything. It seems all sense of proportion has been lost, for, even if technology can help with interesting tools to the diagnosis, the main engine is still the engineering knowledge based on experience. Even though Engineering will hardly survive without technology, it is dangerous to think that only it will save the future of our profession. If we favour the electronic connections and chips over the brain connections, we are doomed to failure as professionals and as homo sapiens. We are destined to become animals capable of performing basic and repetitive operations (the donkey can blow a flute, but that does not make it a musician) Some Engineers show a reverential posture as regards machines which they judge mistakenly, very difficult to use. They think that their employees, who use them with odd familiarity, show as much intellectual capacity as the researchers that developed the Stress Wave Theory. The fact of being able to sit in fiont of a keyboard turns the employee into a genius to his boss’ eyes, in a gifted being if the one supervising is the company’s chairman.. .if continuing to increase generation gap. In general terms, it is believed that a technically complex environment endows a person of a great intellectual capacity; it is as stating that whoever sits in fiont of a TV. set will be endowed with knowledge in electronics. It seems that relying on the mere presence of electronic means turns out to be enough stimulation for the intellectual capacity of the engineers. The simple fact of selecting and observing a screen is a week substitute of the real mental activity. The users end up learning how to press buttons and getting rewards, as laboratory animals, instead of a real comprehension of the cause-effect relation. If we keep on going on this dangerous path, one day we will find an exquisite brochure showing an equipment which will seduce us with its ergonometric design and low price; and it will have two little lights: the green will indicate an undamaged pile while the red one damaged! It is necessary to admit that the Engineer’s work is different fiom the repetitive tasks the collar employees or the industrial workers do, which has been
346
mechanized in order to increase their production. In our case, not only, each event in the site can be physically different, but we also have to analyse as separate entities the boundary conditions that it is subjected. But should the manufacturers take all the blame? Of course not. A good part of it resides in the transfer of the knowledge and not of the technology. The purpose of the ones involved in higher education should be: ‘teach carpentry and not hammer’. We need to teach the ‘why’ and not the ‘how’ of this speciality. Tools come and go. Teaching engineers giving priority to tools limits their knowledge towards them, and therefore their future. Anybody who knows how to read and write and has developed a reasonable abstraction capacity and logical reasoning can learn a computer programme in any moment of his life. That is why, it is important that the education should exploit people’s capacities and skills, the ones which he would be able to comply with the challenges and knowledge that are constantly presented in real life. The focusing on technology- in the computers, for example- will lead us to have obsolete graduates. The question is not, teaching instrumental problems of a technology that develops speedily, but teaching the Engineers how to think. Giving them the ability of thought built up on a concrete knowledge basis, that will allow them to react intelligently before the unceasing changes, even the technological ones, that will appear inevitably in the decades to come. The role of the educational institutions is not that of producing tailor made first class workers, but human beings capable of thinking and being creative, that is what the most lucid businessmen are asking for and that the politicians, blinded by the ‘high technology’, do not have in mind. There are no doubts that everybody needs to learn how to use a computer and to be able to manage the informatics field with certain ability, which it is getting easier than imagined every day. We should always remember that computers are tools and not the aim itself. They are not capable of replacing the intellectual skills and knowledge the Engineer should learn in order to be able to incorporate him to the world economy. In some cases it is being taught as if the study of deep foundations could be understood through the usage of some recipes or simple operations in order to manipulate the equipment, obtaining results on the LCD which the Engineer prefers not to argue because: “I got it fiom the equipment”. Having as a result prefabricated answers that don’t belong to him. What happened with the power of knowing true fiom false? And the answer is in the relationship teacher-student and in its content.. .in the desire of teaching and learning.
What is the use of learning? The question is not new. Almost 2300 years ago, a young geometry student asked Euclid: ‘What will I get learning these things’. The master called his slave and said: ‘Give him some coins. It seems he needs to get something in order to learn’. Together with the world‘s educational problem, in which our speciality is not exonerated, it is proclaimed that no previous knowledge is required therefore it is being taught that knowledge is not hierarchical, that is to say, you can start constructing a building at any level, without foundations. By eliminating the existence of a certain sequence and continuity in education, we are stating that sequence and continuity have nothing to do with the thought itself. We find ourselves those who state that in order to interpret foundation dynamics no deep knowledge in geotechnics is needed. In other words, it is the same to test a pile which is imbedded in the soil that an isolated column! Euclid’s student is now the one sitting at the schools desks! The obsession of earning one’s living is the easiest way to waist it and it is not worthwhile. The present educational system searches for the utility and presumes the death of knowledge. Some of today’s teaching methods are usage instructions to create robots operated by ignorants with ideological lumps. The real knowledge is not on the Internet. but in the brain. Large budgets are assigned to be invested on hardware and the updating of software. Little or no attention is paid to humanware, that is to say the users. And this is the last member of the trilogy Manufacturer-Educator-User we will make reference. The user, before becoming one, must be a student. In other words, he must be eager to learn and willing to be taught (The authors do not advice to follow Euclid’s student’s example) and as a Chinese proverb says: “The masters open the door, you must enter by yourself ’. ‘Some months ago I thought that there was a certain culture which was not given to us through spontaneous experiences, as it is not given as a gift the acquisition of a language or being able to play an instrument. Learning is hard and requires discipline, training and effort”, Luis Landero (Spanish writer) Learning is a job itself, a complex task which demands sacrifices. Learning assumes that each person will take up an individual effort in order to modify himself After more than 14 years, in the business, we can only be sure in believing that learning requires work, discipline, responsibility, and commitment fiom the student as well as fiom the teacher. There are no shortcuts for a quality education and the reward is not a shallow exaltation, but
347
a deep and long lasting satisfaction that may come months or years later. The students do not know that they do not know. This ‘happy unconsciousness’ shows that the system has not even been able to show sobre the antagonism between true and false, knowledge and the lack of it. This lack of information will generate, in part of the new generations, a failure experience due to the contradiction between the high expectations and the insufficient knowledge. It can have an unthinkable impact on society as well, because of the presence of a group of Engineers whose demands will be incoherent compared with their skills. All the community involved in the pile area must keep an eye on this week balance between Manufacturer-Educator-User( equipment-kno wledge-use). Non of these parts can afford the cost of becoming the week link of this chain, otherwise it will break. Each of us has the obligation of being better every day and the noble mission of helping the other two links meeting their commitment of being ethic and having a proper behaviour. If this simple statement can be performed, our job will last becoming a market necessity. And as in every modern market, transactions involves money, it is demonstrated the axiom that ethics and money are compatible.
This Page Intentionally Left Blank
h strain dynamic testing of driven and cast in situ piles narnic testing of large piles
This Page Intentionally Left Blank
ApplicationofStress-Wave Theory to piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
Keynote lecture: Pile acceptance criteria for large diameter and cast in situ .piles Robert E Stevens Fugro-McClellund Marine Geosciences Incorporated, Houston, Tex., USA
ABSTRACT: Case histories are presented to illustrate the use of our pile acceptance criteria for large diameter piles driven in very dense sand, hard clay, and rock. A fourth case history illustrates the use of combined CAPWAP analyses to interpret the results of redrive tests in clay. 1. INTRODUCTION
cost effective way to determine the static pile capacity or develop new methods to compute pile capacity. A total of 26 pullout tests were performed at eight offshore sites on 20- to 30-in.diameter open-ended pipe piles driven to penetrations of 36 to 72 ft. Tensile capacities were 2.5 to 6 times greater than the static pile capacity computed using the API method. Load tests were then performed on two 59-ft long, 24-in.-diameter open-ended pipe piles driven in the very dense sands at Ras Tanajib. A new procedure was developed to compute pile capacity in very dense sands. Helfrich, Wiltsie, Cox, and Al-Shafei (1985) and Al-Shafei, Cox, and Helfrich (1994) describe these load tests in more detail. The ultimate axial capacity is computed using the static method of analysis. The lower bound capacity is computed using the API RP 2A-WSD (1993) method, and is usually that presented in the soil boring report. The upper bound capacity may be computed using the Ras Tanajib method for very dense sands, an upper bound shear strength profile for hard clays, and rock end bearing except for weak rock of poor quality. Additional details are given by Wiltsie, Stevens, and Vines (1984) and Stevens and AI-Shafei (1996).
It is essential that pile and hammer instrumentation be used to evaluate pile acceptance for large diameter driven piles, especially should refusal occur or if easier driving than predicted is encountered. Pile instrumentation can also be used to evaluate acceptance of cast in situ piles. Pile acceptance at refusal is based on: the static pile capacity computations performed using soil parameters obtained fiom a high quality geotechnical investigation, the soil resistance to driving obtained from a CAPWAP analysis, the soil resistance to driving obtained from the Case-Goble formulation, and the soil resistance to driving determined from wave equation results obtained using the measured driving system performance data and the field blow count. ESTIMATE DESIGN PENETRATIONS Design penetrations are estimated from the lower bound static pile capacity curve computed for soil conditions determined from a high quality site specific geotechnical investigation. Due to limitations in the pile capacity procedures, length effects, and soil variability, the computed capacity may be in error. Piles driven to shallow penetrations in the Safaniya Field of the Arabian Gulf, for example, had soil resistances to driving that were as much as 10 times greater than the computed static pile capacity. When discrepancies of this magnitude are observed, pile load tests may be a
3.
EVALUATE PILE DRIVABILITY
We recommend evaluating pile drivability using procedures recommended by Stevens, Wiltsie, and Turton (1982). Driving system parameters are selected from our data base established from over 300 platform installations worldwide. The data base includes hainmer efficiency for 40 different 35 1
pile driving hammers, cushion stiffness and cushion coefficient of restitution for 18 different cushion materials, and system efficiency for 100 different hammer-cushion configurations. Wave equation analyses are performed for a range in driving system performance parameters determined from the mean and standard deviation of values in our data base. Minimum blow counts are computed for the lower bound coring case. If our evaluation of pile drivability indicates premature pile refusal or excessive driving stresses, the hammer and/or wall thickness that permits the piles to be driven to design penetration is determined. 4.
MONITOR PILE DRIVING
Our pile monitoring system records force and velocity signals using strain transducers and accelerometers attached on opposite sides of the pile near the top. The impact stress and velocity, the energy transmitted to the pile (or system efficiency), and the soil resistance to driving are computed for every hammer blow. We estimate the soil resistance to driving using the Case-Goble formulation assuming a damping coefficient that is a function of soil type. The impedance or TNO methods. as discussed by van Foeken, Daniels, and Middendorp (1996), can also be used. Estimates of the hammer efficiency and the cushion stiffness and cushion coefficient of restitution are obtained for selected hammer blows using force-time characteristics obtained in a preinstallation parametric study. We recommend using controlled hard driving, as proposed by Stevens, Wiltsie, and Middlebrooks (1984) whenever the soil boring indicates the presence of rock layers or cemented zones. Driving is continued if the end bearing is decreasing or erratic, and the driving stresses at or just above the driving shoe are acceptable. Driving is terminated if the driving stresses are excessive, or if refusal is reached and the end bearing is constant or increasing with depth. Our hammer monitoring system measures the velocity of the ram just prior to impact using Hall sensors and a small magnet attached to the ram. Hammer efficiency is determined for every hammer blow from the ram impact velocity. We generally recommend that only ram velocity be monitored when easy driving is expected in order to evaluate pile acceptance should blow counts lower than predicted occur. A signal matching program, such as the CAse Pile Wave Analysis Program (CAPWAP)
developed by Rausche (1970), or TNOWAVE, discussed by Courage and Bielefeld (1992), is used to estimate soil quake and damping parameters, and the distribution of the soil resistance to driving along the length and at the toe of the pile. The pile is divided into continuous segments and calculations are made using a traveling wave algorithm. Either the measured pile top force or velocity is used as a boundary condition, and the complementary quantity is computed and compared with the measured quantity. The set of soil parameters is varied until a best match between measured and computed pile top force or velocity is obtained. 5. PILE ACCEPTANCE Piles with adequate lateral capacity rehsing above design penetration are accepted if: the required tensile and compressive pile capacities are obtained from the upper bound static pile capacity curves. o the required tensile and compressive capacities are confirmed by the soil resistance to driving obtained from a CAPWAP analysis, and 8 the required compressive capacity is confirmed by the soil resistance to driving obtained from the Case-Goble formulation and wave equation analyses performed using the measured driving system performance data (i.e., the hammer efficiency, cushion stiffness, and cushion coefficient of restitution) and the field blow count. Piles can also be accepted if the required tensile and compressive piles capacities are obtained from a re-evaluation of pile capacity. The soil resistance to driving profile obtained from the Case-Goble bearing capacity is used to modify the elevation and thickness of soil strata. When pipe piles are driven open-ended, relative movement between the pile and soil occurs both on the outside and inside of the pile wall. Skin friction is, therefore, mobilized on both of these surfaces. The end bearing area in this case is equal to the cross-sectional area of steel at the toe. When a pile plugs, the soil plug moves with the pile during driving. This friction is mobilized only on the outer wall, and the end bearing area is then the gross end area of the pile. The pile initially plugs after each delay. Because the distribution of the shaft resistance on the inside and outside of the pile cannot be evaluated, we recommend waiting a few minutes and restriking the pile. 352
Pile acceptance at refusal is simplified when the pile driving hammers selected are large enough to overcome the long-term static pile capacity, and the CAPWAP and Case-Goble soil resistance to driving can be considered a lower bound because more resistance is mobilized at larger pile displacements. Typically, the pile driving hammer selected for a particular installation may be large enough to drive the piles to design penetration, but not large enough to overcome the long-term static capacity. In clay, the skin fiiction during driving is generally much smaller than that mobilized under static loading because large excess pore pressures are generated during continuous driving. CAPWAP analyses can be used to estimate the distribution of the soil resistance to driving along the length and at the toe of a pile during continuous driving and after a set-up period. By combining these results, it is possible to proof test a pile without the expense of mobilizing a larger hammer to the site. High strain dynamic testing can also be used to evaluate the capacity of cast in situ piles (also known as drilled shafts or bored piles). Hussein, Likins, and Rausch (1996) made recommendations concerning the hammer weight, drop height, and cushion thickness. The hammer weight should be at least equal to 1.5 percent of the expected capacity. The hammer drop height should be approximately 8.5 percent of the pile length, with a minimum value of 2 m. The thickness of the plywood cushion is determined from: t
=
L2/2D
where: t = cushion thickness, mm; L = pile length, m; and D = pile diameter, m. The minimum cushion thickness is 100 mm, and the cushion thickness should be increased 150 mm when the pile length exceeds 30 m.
6. CASE HISTORIES Case histories are presented to illustrate the use of our pile acceptance criteria for very dense sand, hard clay, and rock. A fourth case history illustrates the use of combined CAPWAP analyses. 6.1 In our fwst case history, piles were driven to design penetrations of 35 to 84 m using Menck MRBS 2500-S, 3000, 4600, and 5000 air/steam hammers, and a Menck MHU 1000 hydraulic hammer. Cone penetrometer tests indicated a predominantly dense to very dense silty sand
profile with clay seams and layers. Drivability studies indicated that a Menck MHU 1000 hammer could drive the 48-in.-diameter piles to 84-m penetration if the wall thickness was increased to 2.00 inch. Blow counts with the Menck MHU 1000 hammer ranged fiom 25 to 40 blows per 0.25 m at 60-m penetration, fiom 35 to 65 blows per 0.25 m at 70-rn penetration, and fiom 45 to 85 blows per 0.25 m at 80-m penetration. With the exception of Pile C-3, all piles were driven to design penetration. Pile C-3 was driven to refusal at 56.8-m penetration, 1.2 m above design penetration. Pile C-3 was accepted because the required compressive capacity was obtained fiom the upper bound static pile capacity curve computed using Ras Tanajib pile design parameters, and the required compressive capacity was confirmed by the CAPWAP analyses, the Case-Goble bearing capacity, and the hindcast resistance to driving obtained fiom wave equation analyses. The required compressive design load for Pile C-3 was 26.1 MN. At fmal penetration, soil resistance to driving determined by CaseGoble (Fig. I), CAPWAP (Fig. 2), and wave equation analyses were 28.7, 31.8, and 33.4 MN, respectively. As shown in Fig. 3, the Case-Goble bearing capacity was greater than the lower bound compressive pile capacity. 6.2 In our second case history, piles were driven to rehsal in a hard silty clay stratum between penetrations of 118.9 and 134.3 ft with a Vulcan 560 hammer. Piles F-5 and G-6 were driven without interruption to penetrations of 120 and 134 ft, respectively. Pile G-5 had delays of 22 and 40 minutes that had little effect on the driving. Pile F-6 had a delay of 3.1 hours that almost resulted in pile refusal, as shown in Fig. 4. The restart blow count for Pile F-6 was 730 bpf at 99-ft penetration. At this depth, the required compressive capacity of 4523 kips was not obtained fkom the upper bound static pile capacity curve. The pile was driven to 118.9-R penetration, which was deeper than the required The pile was design penetration of I16 ft. accepted because the minimum blow count was obtained. Although not a requirement for pile acceptance, the required compressive capacity was confrmed by the CAPWAP capacity of 5878 kips (Fig. 5), the Case-Goble bearing capacity of 5200 kips, and the soil resistance to driving backcalculated from wave equation analyses of 6050 kips. 6.3 In our third case history, piles were driven to practical refusal in a strong gypsum 353
Penetration m
Blow Count b10.25m
53.50 53.75 54.00 54.25 54.50 54.75 55.00 55.25 55.50 55.75 58.00 58.25 58.50 58.75 58.80
78 82 72 74 62 72 81 77 79 81 88 81 105 313 5601.05
Average Standard Deviation
Hammer Efficiency %
Cushion Stiffness M Nlm
Coefficient of
Restitution
Transmitted Energy kN-m
System Efficiency %
Resistance to Driving MN
Impact Stress MPa
Impact Velocity mlsec
Reflected Stress %
Maximum Stress MPa
38 41 43 44 44 44 44 44 43 43 42 43 44 42 42
28.6 29.9 29.6 29.7 29.7 29.3 29.1 29.1 28.8 28.4 28.1 28.2 28.8 29.3 28.7
144.0 156.6 152.7 151.8 149.0 146.0 144.3 144.1 141.4 139.0 137.3 137.3 136.2 131.4 129.8
3.52 3.85 3.84 3.84 3.79 3.73 3.69 3.68 3.62 3.56 3.53 3.52 3.52 3.44 3.39
27.1 24.7 23.9 24.1 25.3 25.3 24.5 23.9 24.5 24.0 24.4 23.4 25.0 29.0 31.O
174.9 185.2 182.8 183.2 182.3 179.7 178.2 177.6 175.5 172.8 171.8 172.3 173.6 174.6 172.6
43 (1)
29.0 (0.5)
142.9 (7.3)
3.6 (0.1)
25.3 (2.1)
177.1 (4.4)
--
--
--
75 75 75 75 75 75
I5,026 14,186 14,518 11,646 11,156 10,876
0.60 0.60 0.61 0.61 0.60 0.60
339 363 379 385 389 385 386 390 381 379 374 380 385 374 374
75 (0)
12,901 (1,708)
0.60 (0.01)
378 (12)
Figure 1. Log of Driving System Performance (Pile C-3, Menck MRBS 5000 Hammer)
354
Blow No. 2787
Final CAPWAPC Capacity: Ru 7154.1,
Soil Segment NO.
I 2 3 4 5 6
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Skin 6436.7,
n
Ru kips
UP kips
Down kips
Unit Resistance with Respect to Depth Area kipslft kipslf2
8.1 14.7 21.3 27.9 34.5 41.1 47.7 54.3 60.9 67.5 74.1 80.7 87.3 93.9 100.5 107.1 113.7 120.4 127.0 133.6 140.2 146.8 153.4 160.0 166.6 173.2 179.8 186.4
88.6 48.0 5.2 4.8 35.6 65.6 85.2 88.8 77.5 62.3 47.7 41 .I 63.2 127.4 206.6 243.2 217.8 194.9 266.8 452.5 666.0 763.6 662.0 463.8 362.6 384.1 384.3 327.2
7154.1 7065.5 7017.5 7012.2 7007.5 6971.8 6906.2 6821.O 6732.2 6654.7 6592.4 6544.7 6503.6 6440.4 6313.0 6106.4 5863.2 5645.4 5450.5 5183.6 4731. I 4065.1 3301.5 2639.5 2175.7 1813.1 1429.0 1044.7 7174
88.6 136.6 141.8 146.6 182.3 247.8 333.1 421.8 499.4 561.7 609.3 650.5 713.7 841.I 1047.7 1290.9 1508.7 1703.6 1970.5 2423.0 3089.0 3852.6 4514.6 4978.4 5341.O 5725.1 6109.7 6436.7
13.42 7.26 .79 .72 5.40 9.93 12.90 13.44 11.74 9.43 7.22 6.23 9.57 19.29 31.28 36.83 32.98 29.51 40.40 68.51 100.84 115.61 100.23 70.22 54.90 58.15 58.19 49.55
1.07 .58 .06 .06 .43 .79 1.03 1.07 .93 .75 .57 .50 .76 1.53 2.49 2.93 2.62 2.35 3.21 5.45 8.02 9.20 7.97 5.59 4.37 4.63 4.63 3.94
,053 ,053 .053 .053 ,053 ,053 ,053 ,053 .053 ,053 ,053 .053 .053 ,053 ,053 ,053 .053 .053 ,053 ,053 ,053 .053 ,053 .053 ,053 ,053 .053 ,053
.100 ,100 .I 0 0 ,100 ,100 .100 ,100 ,100 ,100 .I 0 0 .I 0 0 ,100 .I 0 0 ,100 ,100 ,100 .100 .100 .100 .I 00 .100
34.53
2.77
.053
,100
57.07
.351
. I 54
Depth Below Gages ft
Depth Below Grade
95.8 102.4 109.0 115.6 122.2 128.8 135.4 142.0 148.6 155.2 161.8 168.4 175.0 181.6 188.2 194.8 201.4 208.1 214.7 221.3 227.9 234.5 241 . I 247.7 254.3 260.9 267.5 274.1
Suffl of Ru
Toe 717.4 kips
229.9
Average Skin Values
717.4
Toe
Smith Damping Sift
Soil Model Parameters/Extensions
Skin
Toe
Case Damping Unloading Level Soil Plug Weight
,663 0
,488
(% of Ru) (kip9
Quake inch
.i00 .i00 ,100 ,100 ,100 .I 00 .I 00
2.66
Figure 2a. Results of CAPWAP Analyses (Pile C-3)
exceeded, the required tensile capacities were obtained from the lower bound static pile capacity curve, the required compressive capacities were obtained from the upper bound static pile capacity curve computed using rock end bearing, the required tensile and compressive capacities were confirmed by CAPWAP analyses, and the
stratum between penetrations of 58.3 and 62.2 A with a Vulcan 560 hammer. Driving was terminated because stresses exceeded 85 percent of yield in the Grade B steel above the driving shoe. The piles were accepted because the required penetration for adequate lateral capacity was 355
Figure 2b. Results of CAPWAP Analyses (Pile C-3) 356
Figure 3. Soil Resistance to Driving (Pile C-3) 6.4 Our fourth case history is for a series of redrive tests performed with a hammer that is clearly not large enough to mobilize the full soil resistance. The soil resistance mobilized during a series of redrive tests performed on a 1.6-mdiameter steel pipe pile driven to a penetration of 26 m in a very silty clay is presented in Fig. 8. The four lower bound soil resistance profiles are for continuous driving. The pile was redriven by applying only two consecutive hammer blows with a PMJ-400 hydraulic hammer after delays of 6 minutes, 15 minutes, 33 minutes, 2 hours, and 66 hours. The resistance generally increases with time, but the resistance after a 33-minute delay appears to be slightly smaller than the resistance
required compressive capacities were confirmed by the Case-Goble bearing capacities and soil resistances to driving obtained fiom wave equation analyses performed using measured driving system performance data and the field blow count. The driving system performance data for Pile B-2, shown on Fig. 6, indicates a Case-Goble bearing capacity of less than 1460 kips to 28-ft penetration, and less than 2140 kips to 60-ft penetration. The required compressive capacity of 3390 kips was obtained at 62-fi penetration, and was confirmed by the CAPWAP capacity of 4823 kips (Fig. 7), the Case-Goble bearing capacity of 5072 kips, and the hindcast capacity of 6050 kips.
357
ksi
Impact Velocity Wsec
Reflected Stress %
Maximum Stress ksi
4340 4650
19.1 22.4
10.6 12.7
42 62
28.8 30.0
60 59 59 60 56
461 0 4530 4500 4470 4460
22.5 22.3 22.4 22.8 22.3
12.8 12.7 12.8 13.0 12.7
36 39 38 34 38
30.1 29.7 29.8 29.6 29.3
188 195 186 187 186
60 62 60 60 60
4560 4590 4740 4800 4750
23.4 24.1 23.4 24.3 24.3
13.3 13.6 13.6 13.8 13.8
35 34 40 39 39
30.1 30.8 31.3 31.4 31.5
0.86 0.86 0.89 0.78 0.79
196 198 206 177 167
63 63 66 57 53
4580 4450 4460 4470 4780
24.6 23.7 22.8 22.5 23.6
13.9 13.5 12.9 13.1 13.9
35 39 42 45 47
31.3 31 .O 30.0 30.2 32.4
380,600 367,900 384,900 367,900
0.78 0.75 0.79 0.75
168 174 173 166
54 56 55 53
5070 5030 5260 5200
24.3 24.7 25.3 25.1
14.1 14.1 14.4 14.2
51 48 48 45
33.5 33.1 34.0 33.7
316,200 (107,600)
0.82 (0.05)
182 (11)
58
42
( 4)
( 7)
31 .O ( 1.5)
System Efficiency %
Resistance to Driving ki~s
Impact Stress
Restitution
Transmitted Energy ft-kips
66,700 162,100
0.70 0.81
162 184
52 59
76 73 73 73 67
204,400 209,200 209,200 213,900 216,300
0.83 0.85 0.85 0.87 0.88
186 183 184 186 174
236 207 200 193 182
71 74 80 72 75
320,700 423,100 384,900 423,100 406.100
0.89 0.88 0.79 0.88 0.84
111 112 113 114 115
158 172 200 195 188
77 77 78 77 71
414,600 414,600 427,300 258,500 384,900
116 117 118 118.9
204 188 222 294
73 79 73 74 75
Blow
Hammer Efficiency %
Cushion Stiffness kipshnch
Coefficient
730 335
78 77
101 102 103 104 105
270 260 256 249 352
106 107 108 109 110
Penetration ft -
Count
99 100
bf
Mean Standard Deviation
( 3)
of
Figure 4. Log of Driving System Performance (Pile F-6, Vulcan 560 Hammer)
358
Blow No. 170 Final CAPWAPC Capacity: Ru 5877.5,
Soil Segment
Depth Below Gages
Depth Below Grade
NO.
ft
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
205.4 212.0 21 8.6 225.2 231.9 238.5 245.1 251.7 258.4 265.0 271.6 278.2 284.9 291.5 298.1 304.7 31 1.4 318.0
Skin 5761.9,
Sum of Ru
ft
Ru kips
UP kips
Down kips
6.4 13.0 19.6 26.2 32.9 39.5 46.1 52.7 59.4 66.0 72.6 79.2 85.9 92.5 99.1 105.7 112.4 119.0
13.4 15.9 23.1 35.2 60.0 108.6 172.6 220.2 224.9 21 0.6 249.9 388.6 583.6 740.7 806.5 772.2 639.3 496.6
5877.5 5864.1 5848.1 5825.0 5789.8 5729.8 5621.2 5448.6 5228.4 5003.5 4792.9 4543.0 41 54.5 3570.9 2830.2 2023.7 1251.5 61 2.2 115.6
13.4 29.3 52.5 87.7 147.7 256.3 428.9 649.1 874.0 1084.6 1334.4 1723.0 2306.6 3047.2 3853.8 4626.0. 5265.3 5761.9
Average Skin Values Toe
320.1
Toe 115.6 kips
Unit Resistance with Respect to Area Depth kipslf2 kipslft
Smith Damping slft
Quake inch
2.02 2.40 3.49 5.31 9.06 16.40 26.05 33.24 33.95 31.79 37.71 58.66 88.08 111.80 121.74 116.56 96.50 74.96
.I6 .I9 .28 .42 .72 1.30 2.07 2.64 2.70 2.53 3.00 4.67 7.01 8.89 9.68 9.27 7.68 5.96
.068 .068 ,068 ,068 ,068 .068 .068 ,068 ,068 ,068 ,068 ,068 ,068 ,068 ,068 ,068 .068 .068
,080 ,080 ,080 .080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080 ,080
48.42
3.84
,068
,080
65.46
,247
,065
115.6
Soil Model Parameters/Extensions
Skin
Toe
Case Damping Unloading Level Soil Plug Weight
1.004 36
.073 1.72
Figure 5a. Results of CAPWAP Analysis (Pile F-6)
analysis is about 4400 kips (19.6 MN), as shown in Fig. 9. In the standard CAPWAP analysis, the resistance mobilized was only 3305 kips (14.7 MN). To help put this in perspective, Fig. 10 shows that the resistance mobilized by a Vulcan 060 hammer is about 3300 kips (14.7 MN), and the resistance mobilized by a Vulcan 560 hammer is about 4100 kips (18.2 MN), assuming the same soil and pile parameters. Comparing the three hammers shown in Fig. 10, we see that by mobilizing a hammer having 50 percent more rated energy results in only a 10 percent increase in the maximum soil resistance overcome, and mobilizing a hammer having 150 percent more rated energy results in a 37 percent increase in the maximum resistance overcome.
mobilized after a 15-minute delay, and the resistance after a 66-hour delay is less than the resistance mobilized after a 2-hour delay. The maximum resistance overcome by the pile driving hammer is about 3300 kips (14.7 MN). The hammer is too small to mobilize the hll soil resistance during the redrive tests. This is shown very clearly for the redrive test performed after the 66-hour delay. Almost no soil resistance was mobilized over the bottom quarter of the pile. In our combined CAP WAP anaIyses, we have assumed that the soil resistance mobilized during continuous driving is a lower bound, i.e., the soil resistance mobilized on a particular pile segment is assumed to be the larger of the actual resistance mobilized or the resistance mobilized during continuous driving. The soil resistance mobilized after a 66-hour delay in the combined CAPWAP 359
Figure 5b. Results of CAPWAP Analyses (Pile F-6) 360
Blow Count
mf
Hammer Efficiency %
Cushion Stiffness kipslinch
Coefficient of Restitution
Transmitted Energy ft-kips
System Efficiency %
Resistance to Driving kips
Impact Stress J s -i
Impact Velocity Wsec
Reflected Stress %
Maximum Stress ksi
1
51
114,000
1.oo
152
49
980
22.4
14.7
-42
22.4
16 17 18 19 20
44 48 52 58 54
89,700 92,500 89,100 84,900 89,700
0.96 1.oo 0.95 0.89 0.96
125 145 147 7 52 154
40 46 47 49 49
990 1177 1170 7190 1150
20.1 21.6 21.9 21.0 22.4
12.7 13.0 13.4 12.0 13.9
-41 -39 -33 -36 -37
20.1 22.7 22.3 22.4 22.6
21 22 23 24 25
50 56 71 56 72
114,000 114,000 108,100 103,900 97.600
1.oo 1.oo 0.94 1.oo 0.93
162 175 198 179 200
52 56 63 57 64
1170 930 1235 1185 1267
23.3 24.2 24.5 24.4 24.4
14.1 15.1 13.5 14.8 13.8
-45 -50 -45 47 -40
23.6 24.2 24.6 24.6 24.7
Penetration R 15
26 27 28 29 30
10 10 10 11 14
78 79 84 82 78
94,000 92,200 97,300 90,400 94,900
0.89 0.87 0.83 0.85 0.90
204 204 206 208 208
65 65 66 67 67
1370 1372 1458 1515 1613
24.8 24.5 24.8 25.0 25.1
12.8 12.8 12.8 12.9 13.0
-42 43 -43 -40 -38
24.9 24.9 24.8 25.2 25.4
31 32 33 34 35
15 14 13 13 12
78 79 80 77 78
94,900 104,200 103,200 96,700 105.200
0.90 0.90 0.89 0.92 0.91
209 21 1 21 1 212 212
67 68 68 68 68
1623 1563 1548 1560 1615
25.2 25.4 25.3 25.4 25.4
13.1 13.1 13.2 13.1 13.2
-37 -39 -40 -41 -40
25.4 25.5 25.4 25.5 25.5
36 37 38 39 40
13 12 12 14 12
79 77 75 79 78
95,800 97,600 99,400 95,800 87,000
0.91 0.93 0.95 0.91 0.92
213 213 21 1 214 213
68 68 68 68 68
1685 1646 1616 1779 1823
25.5 25.5 25.2 25.4 25.3
13.2 13.2 13.2 13.2 13.2
-40 -39 -36 -34 -30
25.7 25.7 25.6 25.5 25.5
Figure Sa. Log of Driving System Performance (Pile B-2, Vulcan 560 Hammer)
361
Impact Stress
O/O
Resistance to Driving ki ps
Coefficient of Restitution
Transmitted Energy ft-ki ps-
System Efficiency
ksi
Impact Velocity Wsec
Reflected Stress %
Maximum Stress ksi
hf
%
Cushion Stiffness kiwiinch
41 42 43 44 45
17 14 16 16 17
79 79 78 78 80
78,600 79,200 87,700 87,000 78,000
0.91 0.92 0.93 0.92 0.90
213 215 214 212 213
68 69 68 68 68
1986 1829 1763 1894 2139
25.2 25.4 25.3 25.3 25.4
13.2 13.3 13.3 13.2 13.2
-26 -28 -30 -28 -24
25.3 25.5 25.6 25.4 25.6
46 47 48 49 50
20 20 20 23 20
80 80 80 79 79
85,600 77,500 84,900 85,600 85,600
0.90 0.89 0.89 0.90 0.90
212 21 1 213 213 212
68 68 68 68 68
2133 2094 2077 2055 1927
25.2 25.1 25.1 25.2 25.3
13.1 13.0 13.1 13.2 13.1
-25 -25 -22 -18 -2 1
25.4 25.4 25.3 25.5 25.6
51 52 53 54 55
18 15 15 17 14
79 80 80 82 81
78,000 78,000 86,300 76,900 104,200
0.90 0.90 0.91 0.88 0.90
212 214 216 217 216
68 68 69 69 69
1851 1868 1850 1781 1727
25.3 25.4 25.4 25.7 25.9
13.1 13.2 13.2 13.2 13.3
-27 -26 -26 -26 -30
25.6 25.6 25.6 25.9 26.0
56 57 58 59 60
15 16 15 16 17
81 82 82 80 81
104,200 103,200 103,200 105,200 86,300
0.90 0.89 0.89 0.91 0.91
216 217 215 215 217
69 69 69 69 70
1758 1837 1841 1816 201 1
25.9 25.9 25.9 25.8 25.7
13.3 13.3 13.3 13.3 13.3
-3 1 -31 -29 -3 1 -24
26.1 26.1 26.1 25.9 25.9
61 62 62.2
27 200 708
79 78 81
95,800 61,600 45,700
0.91 0.79 0.70
212 183 169
68 58 54
3253 5125 5072
25.3 22.3 21 .o
13.2 11.4 10.6
4 62 69
29.4 34.3 32.1
74
91,900 (13,000)
0.91 (0.05)
200
64
1794
( 24)
( 8)
( 796)
24.6 ( 1.5)
13.2 ( 0.7)
-29 ( 22)
25.4 ( 2.2)
Penetration ft
Blow Count
Mean Standard Deviation
Hammer Efficiency
(11)
Figure 6b. Log of Driving System Performance (Pile B-2, Vulcan 560 Hammer)
362
Blow No. 739 Final CAPWAPC Capacity: Ru 4823.2,
Soil Segment 0.
1 2 3 4 5 6 7 8 9
Skin 3676.8,
Toe 1146.3 kips
Depth Below Gages ft
Depth Below Grade ft
Ru kips
UP kips
Down kips
Unit Resistance with Respect to Depth Area kipslft kipslfa!
167.5 174.0 180.6 187.2 193.7 200.3 206.9 21 3.4 220.0
9.5 16.0 22.6 29.2 35.7 42.3 48.9 55.4 62.0
99.8 143.9 223.9 353.6 522.3 646.6 655.0 567.0 464.8
4823.2 4723.4 4579.5 4355.7 4002.7 3479.8 2833.2 21 78.2 1611.2 1146.3
99.8 243.6 467.5 821 .I 1343.4 1990.0 2645.0 3212.0 3676.8
15.19 21.91 34.09 53.85 79.53 98.46 99.74 86.34 70.78 59.30
Sum of Ru
408.5
Average Skin Values Toe
1146.3
Smith Damping Slft
Quake inch
I .38 1.99 3.10 4.90 7.23 8.95 9.07 7.85 6.43
.I26 .I26 .I26 .I26 ,126 ,126 .I26 .I26 .I26
,194 .I 94 ,194 .I 94 .I94 .I 94 ,183 ,164 ,148
5.66
.I26
.I 82
745.82
.301
,148
Soil Model Parameters/Extensions
Skin
Toe
Case Damping Unloading Level Soil Plug Weight
1.358 0
1.015
(% of Ru)
6.47
WPS)
Figure 7a. Results of CAPWAP Analyses (Pile 9-2)
7. CONCLUSIONS Pile instrumentation records pile top force and velocity. from which CAPWAP and Case-Goble soil resistances are obtained. Soil resistances are also back-calculated f?om wave equation analyses performed using the measured hammer efficiency, cushion stiffness, and cushion coefficient of restitution. Hammer instrumentation measures ram impact velocity, from which hammer efficiency is obtained. At refusal, we recommend accepting piles when the required tensile and compressive pile capacities are obtained from the upper bound static pile capacity curves, the required tensile and compressive capacities are confirmed by CAP WAP analyses, and the required compressive capacity is confirmed by the Case-Goble bearing capacity and the soil resistance to driving obtained fiom wave equation analyses performed using the measured driving system performance data and the field blow count. We also recommend accepting piles when the re-evaluated pile capacity indicates that the
required tensile and compressive pile capacities are obtained. The soil resistance to driving profile obtained f?om the Case-Goble bearing capacity is used to modify the elevation and thickness of soil strata. Pile acceptance can also be based on combined CAP WAP analyses. The soil resistance mobilized on a particular pile segment is assumed to be the larger of the actual resistance mobilized. or the resistance mobilized during continuous driving. When piles are driven to design penetration, we recommend accepting piles as long as the minimum blow count is obtained. Minimum blow counts are computed for the lower bound coring case and a range of hammer efficiencies. Only ram velocity is required to be monitored to confirm hammer efficiencies.
363
Figure 7b. Results of CAPWAP Analyses (Piles B-2) 364
Figure 8. Soil Resistance Determined from CAPWAP Analyses 365
Soil 1000
Resistancel Kips 2000 3000 4000
Figure 9. Soil Resistance Determined from Combined CAPWAP Analyses
366
5000
Figure 10. Driving Resistance-Blow Count Curves American Petroleum Institute (1 993), Recommended Practice for Planning, Designing, and Constructing Fixed Offshore Platforms, API Ihmnmended Practice 2A-WSD (w 2A), 20th Ed., API, Washington, D.C.
REFERENCES Al-Shafei, K.A., Cox, W.R., and Helfi-ich, S.C. (1994), “Pile Load Tests in Dense Sand: Analysis of Static Test Results,’’ Proceedings, 26th Offshore Technology Conference, Houston, Vol. I , pp. 83-103. 367
Courage, W.M.G. and Bielefeld, M. W. (1992), "TNO automatic signal matching," Proceedings of the Fourth International Conference on the Application of Stress-Wave Theory to Piles, The Hague, The Netherlands, September 2 1-24, 1992. Helfrich, S.C., Wiltsie, E.A., Cox, W.R., and AI-Shafei, K.A. (1985), "Pile Load Tests in Dense Sand: Planning, Instrumentation, and Results," Proceedings, 7th Offshore Technology Conference, Vol. 1, pp. 55-64. Hussein, M., Likins, G., and Rausche, F. (1996), "Selection of a Hammer for High-Strain Dynamic Testing of Cast-in-Place Shafts," Fifth International Conference on the Application of Stress-Wave Theory to Piles, Orlando, Florida, September 1 1- 13. Rausche, F. (1970), "Soil Response fi-om Dynamic Analysis and Measurements on Piles," Ph.D. Dissertation, Division of Solid Mechanics, Structures, and Mechanical Design, Case Western Reserve University, Cleveland, Ohio, 320 p. Somehsa, P. and Stevens, R.F. (1989), "Pile Driving Measurements for Assessment of Hammer Performance and Pile Adequacy," Symposium on Underground Excavations in Soils and Rocks," Asian Institute of Technology. Stevens, R.F., Wiltsie, E.A., and Turton, T.H. (1982), "Evaluating Pile Drivability for Hard Clay, Very Dense Sands, and Rock," Proceedings, 14th Offshore Technology Conference, Houston, Vol. 1, pp. 465-481. Stevens, R.F., Wiltsie, E.A., and Middlebrooks, J.R. (1 984), "Controlled Hard Driving," Proceedings, 2nd International Conference on the Application of Stress Wave Theory on Piles, Stockholm, pp. 162-169. Stevens, R.F. and AI-Shafei, K.A. (I 996), "The Applicability of the Ras Tanajib Pile Capacity Method of Long Offshore Piles," Proceedings, 28th Offshore Technology Conference, Houston, Vol. 1, pp. 171-180. Stevens, R.F. (ZOOO), "Pile Acceptance Based on Combined CAP WAP Analyses," Sixth International Conference on the Application of Stress-Wave Theory to Piles, Sao Paulo.
van Foeken, R.J., Daniels, B., and Middendorp, P. (1 996), "An improved method for the real time calculation of soil resistance during driving," Fifth International Conference on the Application of Stress-Wave Theory to Piles, Orlando, Florida, September 1 1- 13, 1996. Wiltsie, E.A., Stevens, R.F., and Vines, W.R. (1984), "Pile Installation Acceptance in Strong Soil," Proceedings, 2nd International Conference on the Application of Stress Wave Theory on Piles, pp. 72-78.
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Analyzing the bearing capacity mechanism of large diameter diving casing cast-in-situ concrete piles by using high strain dynamic testing Xi Liang Fujian Academy of Building Research, Fuzhou, People’s Republic of China
ABSTRACT: In this paper, the bearing character of large diameter diving casing cast-in-situ concrete piles by using high strain dynamic testing was discussed, and some problems in construction aspect were analyzed. Advantages of bearing behavior and economy of this pile were indicated, and some suggestions were put forward in the course of construction. Key words: large diameter diving casing cast-in-situ concrete piles, high strain dynamic testing of piles, shafi resistance, toe resistance approximately 20-25mm. Diameter of pile former is mainly of@700mm and its thickness 22mm.The pile hammer is OED that usually used in driven preformed concrete pile. The model is mainly MH72B, D62-22 and K80. The weigh of ram is 60801tN. Its impact force can reach more than 7300kN and most impact energy can exceed 200k.I. On thc top of the pile there is jacketing vibrator equippcci. An exciting force of it is about 300-400kN. Whcn if begin to cast concrete, vibrator is operated. It can 1101 only improve the ability of pulling the former bu1 also make the concrete be vibrated sufficiently . t 1 equipped with pile shoe beneath the pile tip. ’l’he shape of pile shoe is similar to a cone. The rigidity of pile shoe has been reinforced by installing strengthening rib to improve the ability of penetrating hardpan. The diameter of pile shoe is larger than that of former as about 100mm. It make the soil strata around pile skin not rebound entirely when pulling the former, thus reducing resistance of pulling former and ensuring construction successfully. The processes of construction are presented as fo11ows : (1) Placing the pile shoe and pile driver perching. (2) driving the former; (3) examining pile tip with lamp; (4) placing cage of reinforcement and casting concrete; ( 5 ) operating the vibrator and pulling the former; (6) pulling the former to the designed elevation and finishing all records.
1 INTRODUCTION With the development of economy in Fujian province, great changes have taken place in city’s visage. Since great deals of high buildings are constructed, the bearing capacity of piles is required to improve with most economical method to the greatest extent. At present, the main types of piles of high buildings in Fujian province are pre-formed concrete pile or bored piles and belled piles by hand supporting on medium or slightly weathered rock. In the course of engineering construction in Xiamen of Fujian province, a rather novel type of pile-large diameter diving casing cast-in-situ concrete pile has been always encountered. By means of high strain pile testing and analyzing with CAPWAPC program, the author discovered that this pile has better advantage in bearing capacity and economical character comparing with that of other types of piles. This type of pile fits for many areas in Fujian.
2 ‘THE CONSTRUCTION PROCESS OF LARGE DIAMETER DIVING CASING CAST-IN-SITU PILES The outside former diameter of large diameter diving casing cast-in situ piles is about 0 560 700mm, mainly 0 700mm, @ 650nim, @ 6OOn1m and 0 560mm, respectively the thickness of former is
-
369
3 PROJECT EXAMPLE 27 piles in DH garden in Xiamen area are conducted hinge strain dynamic testing. The typical subsurface condition of this project is shown as follows: (1)miscellaneous fill, its thickness is about2.Om, (2)muck, its thickness about 18-23m, (3)clay, its thickness about 2-6m, (4) residual sandy clay, its thickness morn than 20m. Designers initially intended to adopt 0 1OOOmm bored cast-in-situ piles in this project and conducting static testing of two bored piles with a length of 35m. But they attained a capacity of only 3500kN and 4000kN. They considered adopting driven preformed concrete piles afterwards. But because the surface of residual sandy clay is strongly wavy, the length of pile can’t be controlled accurately. Varying pile length results in the cost of construction increasing accordingly. So driven pre-formed pile was refused. Thus for the pile foundation of this project, large diameter diving casing cast-in-situ pile with diameter Q 600mm and CD 700mm was adopted. Pile penetration is about 29-39m. The bearing stratum of piles is residual sandy clay. 8 piles with diameter Q600mm are tested by using high strain dynamic testing except 4 piles come across rock in the pile tip. The ultimate bearing capacity is about 4500-5800kN, the mean bearing capacity is equal to5150kN; 15 piles with diameter 0 700mm are tested. the ultimate bearing capacity is about 4600-6800kN, its mean value is equal to 5873kN. One pile among them is tested with static and dynamic comparison experiment. This pile is carried out high strain dynamic testing in advance and determines its ultimate bearing capacity of 6400kN; and then it is carried out static load testing. Under the condition of load of 6100kN, the settlement of it is equal to36.9mm. The relative error between static testing and dynamic testing is 4.92%.
The geological column around the testing pile, Q-S curve of static load experiment and CAPWAP result are shown in fig. 1, fig2, fig3, respectively.
Fig.2 Q-S curve of static load experiment of testing pile in DH garden
The author has analyzed all the 23 piles with CAPWAP program. It follows from the analyzing result that both the shaft resistance and toe resistance can be brought into a quite high level, so is it even under the condition of very feeble subsurface. These data are not only higher than that of small diameter diving casing cast-in-situ piles or bored cast-in-situ piles, but also higher than that of driven pre-formed piles. The ultimate toe resistance and shaft resistance of this pile and values comparing with other piles arc listed in table 1. (The data with a symbol “*” conic from Technical Code for Biddii7g Pile Fozriw’cifioiis 1. A statistic result shown that the bearing capacity provided by large diameter diving casing cast-in-situ pile with diameter Q 700mm and length 32m, is larger than that of bored cast-in-situ pile with diameter Q 1200mm and the same length or two preformed concrete piles with a section 400mm X 400mm and the same length. It can be seen that this type of pile has remarkable economical benefit. Table 1 . Ultimate shaft resistance and toe resistance of large diameter diving casing cast-in-situ piles in DH garden and values cornoarinn with other tvoes of oiles Ultimate Ultimate shaft resistance(kPa) ~ i ~ ~ ~ l l Resid iial Pile toe resistance aneous Muck Clay sandy clay fill (1) 7150 46 19 85 I03 _ _
(2)’
370
24
14
66
22 13 64 19 11 52 Large diameter diving casing cast-in-situ pile Pre-formed concrete pile Bored cast-in-situ pile Small diameter diving casing cast-in-situ pile
(3)’ (4)’ (I) (2) (3) (4)
5900 1400 4200
82 _ _ 78 63
The statistic data of static testing of 20 large diameter diving casing cast-in-situ piles among 6 projects have been collected. The stress-strain electric experiments of the pile shaft are conducted for 5 piles among these piles, and the ultimate shaft resistance in miscellaneous soil, clay and residual sandy clay are obtained. The mean values of them are respectively 57kPa, 79kPa and IOSkPa.
4 ANALYSIS MECHANISM
OF
BEARING
CAPACITY
The CAPWAP analyses of testing piles above have shown that the shaft resistance and toe resistance of this type of pile are all high. It follows from this result that the mechanism of bearing capacity of this type of pile is as follows: in the respect of toe resistance, because of high penetration ability of this pile equipment which can impact former to hardpan even to weathered rock, the bearing stratum absorbs great energy passing by steel pile shoe and is compacted, and its modulus of deformation increases greatly, which is similar to the driven preformed concrete pile, but its bottom area is far greater than driven pre-formed concrete pile, What is the main reason why the toe resistance of this pile can reach so high value. In the respect of frictional resistance, also similar to driven pre-formed pile, the pile former penetrate into the soii strata by blowing of rani, soil around former is extruded. Soil around pile rebounds and combines close with concrete of pile after casting concrete. But because the interface between pile and soil around driven pre-formed pile is smooth, while the interface between pile and soil around large diameter diving casing cast-in-situ pile is rough, meanwhile large diameter diving casing cast-in-situ pile has not the mudcake problem of bored pile, therefore, the shaft frictional resistance of this pile is larger than that of pre-formed pile under the normal stress. Analyzing result of CAPWAP program of high strain testing indicates that the end bearing of pile is about 35-45% in all load of pile top, and its mean value is 38.9%, which is larger than that of cominoii bored cast-in-situ pile under the same geological condition. The result also shows that this pile is buoyant pile while its lengt~i-diameterratio (LiD) is about 50. Because there are not sediment under the pile tip and mudcake around the pile skin, the shaft frictional resistance and toe resistance exert simultaneously to resist the load of pile top, thus the pile obtain higher bearing capacity. For the bored cast-in-situ pile, because of the influence of sediment, the effect of toe resistance may not bring
into full play even when the shaft frictional resistance has been exerted to an extreme extent.
5 PROBAl3LE EXISTING PROBLEMS After driving the former of large diameter diving casing cast-in-situ piles has been finished, the pile bottom is required to be examined with lamp. Meanwhile, some effective technology measures have been taken. There, some common faults of quality in small diameter diving casing cast-in-situ piles, such as pile tip mixing mud, etc., can be avoided, and the qualified ratio of testing of large diameter pile piles exceeds far that of the sinalf diametcr piles. However, when carrying out testing of a pile in SF garden in Xiamen of Fujian province, the author also discussed a quality problem. The diameter of this pile with a bearing stratum of residual sandy-clay is equal to Q> 600mm, and penetration length is 3 1Am. By means of static load experiment, its ultimate bearing capacity is only 2 150kN, which is only one half of most entrusted load. After sending the pile 970mn1, then high strain dynamic testing is carried out. Its ultimate bearing capacity was calculated only 2620kN either. It follows from the curve that the pile was damaged seriously on the location of 1 I m beneath the pile top. (The geological coiuinti around pile, pile profile and pile model and CAPWAP final results are shown respectivcly i n fig.4, fig.5 and fig.6) This location happened to be the boundary between muck and sandy clay, So w e can infer the reason of causing this problem. Firstly, when the pile penetrate densely into the saturated soft soil during the process of construction, soil strata around the pile is crushed to cause excess pore water pressure. The pressure makes the ground hunch and the pile bear a horizontal pushing force. If the concrete intensity of adjacent piles is still very low, the pile perhaps damaged by the horizontal thrust force. Secondly, at the boundary between hardpan and soft soil, the speed of pulling former is so rapid that it causes the pile mixing muck. In fact, on the other field with a deep muck such as DH garden introduced above, many piles are found some defects existing on the location of mitck by testing. The reason may be the same as that of SF garden. 4 CONCLUSIONS 1. The testing result of large diameter diving casing cast-in-situ pile indicates: Its shaft frictional resistance and toe resistance can resist together the 371
Fig.3 CAPWAP final result of testing pile in DH garden
Figure 4. Geological column around testing pile in SF garden
load of pile top. The values of their shaft resistance and toe resistance are far greater than that of common cast-in-situ piles. This pile has notable economical benefit and is worthy of generalizing to
Figure 5. Profile and model of testing pile in SF garden
other area in Fujian province, especially in the area with thicker residual soil and deeper weathered rock 2. Quality problem of large diameter diving casing cast-in-situ piles has been reduced greatly 372
F i g 6 CAPWAP final result of testing pile in SF garden
comparing with small diameter diving cast-in situ piles. But when driving in the saturated sof’t soil such as deep muck, the order of driving piles and the speed of pulling the former must severely controlled, and the distance between piles is properly widened. 3. High strain dynamic pile testing has playcd a i l important role in engineering inspection. ‘l’hc bearing capacity mechanism of pile can be analyzed with this testing method and the location of pile damage can be determined. It has obvious social and economical benefits. With regard to large diameter diving casing cast-in-situ piles, it should accumulate constantly the data of comparison between static and dynamic testing, especially that of contain the stressstrain electric experiments of the pile shaft.
REFERENCE
373
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Analysis of dynamic load tests on steel rails piles E M.A. Lima, J.C.A.Cintra & N.Aolu University of SCo Puulo, SCo Carlos, Brazil
ABSTRACT: This paper presents an analysis of dynamic load tests performed on steel rails piles (TR-68) at the Experimental Foundation Field of the Department of Geotechnics, Engineering School of S2o Carlos, University of SBo Paulo, Brazil. The piles have segments of 12 and 3m, welded according to NBR 8800/86, comprising a maximum driving length of 27m. The objective of this analysis is:To carry out a wide study of steel track pile bearing capacity; To validate a tool predicts in NBR 6122/96: the rebound. Validate with the control method of driving piles and verify the utilization of the steel rails piles with structural element of foundation. The rebound measurements were performed with paper and pencil in eighteen piles and were analyzed by applying increasing energy level. The energy was applied by the gravity hammer falling of a drop heights of 0.2 - 0.4 - 0.6 - 0.8 - 1.0 - 1.2 and 1.5m. The bearing capacity of the piles, determined by extrapolations of mobilized resistance - displacement curves is compared with two load tests employing the Pile Driving Analyzer (PDA) and with one static load test on a representatives piles.
drop heights of0.2, 0.4, 0.6, 0.8, 1.0, 1.2, and 1.5 m. The elastic (rebound) and permanent (set) displacements were recorded through paper and pencil, placed at the top of the pile. Based on the measured rebounds, the mobilized load was calculated according to UTO et al. (1985) and VELLOSO (1987) propositions.
I INTRODUCTION
Dynamic load tests with increasing energy has been used in Brazil for 10 years and allows the bearing capacity determination of the soilfoundation system from a load- displacement curve defined by a certain number of points. The rebound technique has also been spread in Brazil as a driving control tool of all piles in a construction site. In practice, the measured rebound signal is the wave equation solution. The rebound recording at the top of the pile with pencil and paper can be used to estimate the mobilized load for the corresponding energy level. Assessment of the mobilized load displacement curve for several energy levels allows the ultimate load estimation.
- Displacenieri f parcels C2 arid C3 deierniirinfrori. C2 and C3 displacement parcels are the first and major variables in the load test assessment using the rebound technique. Inadequate assumptions of these variables can lead to misrepresentative mobilized load values Scatter results were obtained for the formerly assumed values C3 = 2.5 mm and C = set. Thus, the displacement parcels C2 and C3 were determined by AOKI (1986) proposition
2 DYNANAMIC TESTS PERFORMED DURING PILE DRIVING Dynamic tests with rebound measurement at the end of the driving of the piles were performed according to CHELLIS (195 1) proposition, improved by AOKI (1989). Increasing energy levels were applied by the 20 kN gravity hammer, with
Satisfactory results were achieved by evaluating C3 with this methodology. C3 mean value was 0.98 mm, with a standard deviation of 0.34 mm, corresponding to a coefficient of variation of 34.7
375
the stabilization is achieved even if the energy keeps increasing;
%. The maximum and minimum obtained values were 0.40 and 2.27, respectively. Figure 1 illustrates
C3 values normal distribution curve.
Fibwe 2. The plot of the inean value of C3 dsplaceinent versus the liartuner potential energy
The set displacements tend to a nearly zero value for the lower energy levels, and increase as the applied energy increases.
Figure 1. Illustrates C3 1,alues nonnal distribution cume
Table 1 presents an statistic analysis for the applied height blows, in an attempt to exemplify C3 variation at each applied energy level. Tabel 1. Statistic analysis Blon. Potenc. Mean height Enerb? C3 (111) (kJ) (111111) 0.2 4.0 0.79 0.4 8.0 0.91 0.6 12.0 0.9') 0.8 16.0 1.03 1.0 20.0 1.03 1.08 1.2 24.0 1.05 1.5 30.0
-13enririg cnpncitj of the piles The bearing capacity of the piles is evaluated from the static mobilized load - dynamic displacement curve For the mobilized load values calculated by VELLOSO (1987) expression, the shape of the mobilized load - displacement curves shows a clear failure in most of the cases, clearly defined by a vertical asymptote However, as UTO et a1 (1985) expression did not allow a clear observation of the ultimate failure, an extrapolation of the mobilized-displacement load was then necessary The extrapolation was performed through VAN DER VEEN (1 953) method Conventional failure load methods by DAVISSON (1972) and the Brazilian Standard NBR-6122/96 were also used Table 2 and 3 show the mean bearing capacity obtained by the above-mentioned methods
for the applied height blows. Stand. Coeff. Mar. Min. Desv. of Var. Value Value (111111)
(%))
(111111)
(111111)
0.28 0.34 0.35 0.33 0.33 0.34 0.34
35.3 37.4 35.4 32.0 32.0 31.5 32.4
1.73 2.08 2.21 2.21 2.21 2.27 2.27
0.45 0.40 0.64 0.73 0.73 0.73 0.79
It can be observed that C3 values are affected by the maximum displacement and the permanent values of each blow. As these displacements depend on the applied energy levels, one can affirm that C3 value is an energy level dependent variable. This statement is better understood by observing Figure 2, which presents a plot of the mean value of C3 displacement as a hnction of the hammer potential energy. The above results substantiate AOKI (1991) statement' the tip quake (C3) is a variable that depends on the applied energy level Dynamic load tests displacement parcel analysis validate the results acquired by NIYAMA & AOKI (1 991): The rebound displacements increase with the applied energy up to a maximum limit. Afterwards,
Table 2 - the niean bearing capacit? obtained b j mobilized load alues calculated b\ VELLOSO (1987) elpression Velloso Veen Nl3R Davissoii L Ma\ (in) Value (kN) (kN) (kN)
\
(kN) Mean 263 Stand Des\ 0 46 Coeff ofVar 1 7
376
1120 160.0 14.3
1267 239.0 18.9
1247 1201 260.0 257.0 20.9 21.4
Table 3 The mean beanng capacih obtained by mobilized load UTO et a1 (1987) expression Uto L Max NBR Da.cisson Veen (in) Value (kN) (kN) (kN) (kN) Mean 26 3 926 1499 874 741 Stand Den 0 4 6 83 173,O 94.0 81.0 10.8 10.9 Coeff of Var 1 7 9.0 11.5
1 alues calculated b\
Tables 4 and 5 present the mobilized load values for each applied energy level. Table 4 The mobilized load values for each level calculated by VELLOSO (1 987) Blon Pot R Stan Coef May height Eiier Mean Den Var Value (m) (kJ) (kN) (nun) ('3)) (nun) 4.0 745 181 24.3 1066 0.2 0.4 8.0 880 218 24.8 1285 0.6 12.0 1000 237 23-7 1489 0.8 16-0 1o-M 19-4 18.6 14% 1.0 20,O 1018 215 20.5 1592 1.2 24.0 1132 258 22.8 1902 1.5 30,0 1056 158 15.0 1112
applied energ?
Table 5. The mobilized load values for each level calculated by UTO et al. (1985). B1o.c.c Pot. R Stan. Coef Max. llclglll Encr. Mca n Dcs! . Var. Valuc (mm) ('%) (mm) (111) (k J? (kN) 0.2 17.6 668 88 1.0 498 16,7 789 0-4 8.0 599 100 15.7 922 0.6 12.0 696 109 13,2 1000 0.8 16.0 755 100 1.0 20.0 793 105 13.2 1033 24.0 873 12.9 1168 113 1.2 30.0 914 10,2 1092 1.5 93
applied energy
Miii Value (min) 338 295 605 786 756 786 800
Min. Valuc (Inm) 3 16 343 5 15 563 587 640 758
Mean Displ (inm) 8.8 11.3 14.0 15.8 17.2 19.3 22.1
Mean Displ. (ill111) 8.8 11.3 14.0 l5,8 17.2 19.3 22,l
Results show that the mobilized load increases as the applied energy increases. Although VELLOSO (1987) method generated higher mobilized load values, the average ultimate load value for the curve achieved by UTO et a1 (1985) method is higher than the average ultimate load corresponding to the curve obtained by VELLOSO (1987) method. This substantiates the major role that the loaddisplacement curve shape plays on the definition of the ultimate failure load. The acquired failure loads follow the same trend of the mobilized load values when DAVISSON (1972) and the NBR-6122/96 methods are employed. This happens because the values by VELLOSO (1987) are higher than the values by UTO et al. (1985). The set displacement does not define the ultimate resistance of the soil-pile system, as shown in Figure 3
377
Figure 3 The plot of set versus the ultimate resistance of the soll-plle system
3 DYNAMIC LOAD TESTS PERFORMED
DURING PILE RESTRIKE AOKI (1 989) proposition on using increasing energies to define the mobilized load-displacement curve was also followed during this phase. The dynamic load test with the PDA was carried out during piles E-23 and E-10 restrike, both considered as representative of the set of piles. Potential energy levels of 4.8, 9.6, 14.4, 19.2, 24.0, 28.8, and 36.6 kJ corresponding, respectively, to hammer fall heights of 0.2, 0.4, 0.6, 0.8, 1.0, 1.2. and 1.5 m were applied in an attempt to compare the results of these tests with the results by the rebound technique. Pile restrike was performed one year and eight months prior the ending of the driving, and the results were analyzed at each energy level by CASE and CAPWAP methods. The rebound and set displacements were also recorded through paper and pencil. Signal trace analysis using the PDA and the rebound technique (concurrently measured) for each applied energy level are exposed in Table 6 and 7. Signal traces measured by the PDA were assessed through CASE and CAPWAP methods, while the rebound signal traces were assessed through UTO et al. (1 985) and VELLOSO (1987) propositions. The dumping constant J, was assumed to be 0.7 for the CASE method analysis. This value is within the range proposed by RAUCHE et al. (1985) for piles with tip embedded in a clayey silt soil.
Table 6 - The results of restrike arial!Led E-10 B l o ~ Energ! Case Capnap Height Potencial (kN) (kN) (111) (kJ) 0.2 4.8 420 444 0.4 0.6 0.8 1.0 1.2 1.5
of E-10 pile Rebound Vclloso (kN) 513
+ 9.6 + 784 14.4 790 1023 1046 1060 1272 1136 19.2 24.0 1120 1299 1316 1170 1344 1316 28.8 36.0 1460 1490 1681 + This blov did not shon n a good signal
with this dynamic load test. This statement is better understood by observing Figure 4.
Rebound
Uto
Tabele 9 - The quake values obtained b) tlie CAPWAP anal! sis Golpc E-23 E- I 0 H Energia Quake C3 Quake C3 Potencial Lateral Ponta Lateral Ponta (in) (kJ) (inin) (inni) (inm) (inin) 0.2 4.8 1.95 1.70 0.98 0.19
(kN) 370 510 647 693 810 810 997
1.36 + 9.6 1.88 14.4 2.59 2,00 1.03 1.23 2.05 0.30 19.2 0.99 1.25 1.22 24.0 0.94 0-60 1.50 28.8 + 36.0 + 2.39 + This blon did riot slion ii a good signal
0.4 0.6 0.8 1.0 1.2 1.5
Table 7 - The results of restrike analyzed of E-23 pile E-23 Blou Energy Case Capwap Rebound Rebound Height Potencial (kN) (kN) Velloso Uto
9.6 630 673 789 14.4 760 600 964 19.2 900 811 1149 1.0 24.0 960 1102 1234 1.2 28.8 990 1110 1418 1.5 36.0 + + 1418 + This blon did not shon 11 a good sigiial
0.4 0.6 0.8
+ 0.35 0.32 0.35 0.40 0.14
512 604 651 768 862 862
The lateral friction and tip resistance values, as well as the obtained quake values through the CAPWAP analysis are presented in tables 8 and 9. Tabele 8 - The lateral friction and tip resistance values obtained by the CAPWAP analysis. Golpe E-10 E-23 H Energia Capwap Atrito Ponta Capwap Atrito Ponta (in) Potencial (kN) (kN) (kN) (kN) (kN) (kN) 0.2 0.3 0,6 0.8 1.0 1.2 1.5
4.8
444
+ 9.6 14,4 1023 19.2 1272 24.0 1299 28.8 1346 1490 36,O + This bloli did not
435
8
418
+ 671 831 173 600 1094 178 811 862 437 1102 795 551 1110 + 911 579 shown a good signal +
368
50
568 440 428 862 896
103 160 383 240 214
+
+
Restrike results show that the structural element ultimate load limits the bearing capacity of the piles. The same conclusion can be observed either by the PDA and the rebound techniques. The mobilized and ultimate loads obtained by VELLOSO (1987) proposition using the rebound technique yielded close results with respect to the load values generated by the dynamic tests with the PDA and analyzed via CASE and CAPWAP methods. However, load values obtained by UTO et al. (1985) proposition did not show good agreement
Figure 4 - the mobilized resistance-displacement curve in tlie pile restrike.
Thus, it is demonstrated that the rebound technique employment in pile driving control,
378
assessed by VELLOSO (1987) proposition, is a highly reliable tool in comparison with the wave equation based methods. The signal trace assessment through the CAPWAP method shows a virtually perfect superposition of the measured and the calculated load values. A prevalence of the lateral friction load is shown by the CAPWAP analysis, even with a substantial part of the tip load. Despite the scatter shown by some points, the lateral resistance increases with the increase of the applied energy level. Further study is needed to understand this dispersion and the absence of continuity on the resistance curves with depth. However, some hypothesis can be drawn to explain this: A CAPWAP model incompatibility, which assesses the mobilized load for an energy level considering it to be the ultimate system load; The soil residual tensions which dissipates at the higher energy blows; The tip resistance also increases with the increase in the applied load. Specifically, the scattering on pile E-23 data regarding the lateral resistance could again be observed. Although the mobilized load estimates showed good results, comparisons between the PDA tip C3 and the rebound technique estimated C3 were poor. Further explanations are needed, but a preliminary hypothesis is that the assumed C3 in the dynamic formulas consider the tip and friction parcels together, while the C3 assessed by the PDA takes into consideration the pile tip parcel only.
Figure 5 - the result of the static load test.
4 STATIC LOAD TEST The static load test was performed on pile E-23, after completion of the dynamic load tests. The result of the test is shown in the Figure 5 Ten load stages of 120 kN were applied, corresponding to 10 ?o' of the estimated ultimate load. The test was terminated at a load of 1200 kN and a maximum settlement of 17 mm, corresponding to 10% of the larger pile dimension. The test was interrupted because the applied load approached the reaction system allowable resistance. The static load test corroborates the trend of the other tests, showing that the ultimate rail structural resistance is probably the variable that limits the set of piles. The pile-soil system response was elastic because the pile tip lies in a very resistant layer. Figure 6 compares the dynamic load test results and the elastic pile compression.
Figure 6 - Comparisons between the dynamic load test results and the elastic pile compression.
Comparisons between static and dynamic load tests can be better evaluated through figure 7. All dynamic test analysis showed appropriate results with respect to the static test results. The dynamic test with rebound measurements and assessed by VELLOSO (1987) proposition yielded the closest results. 5 DRIVING AND RESTRIKE COMPARISONS
Rebounds were measured during the driving and restrike test phases. Figure 8 shows the mobilized load - displacement curves during both phases.
379
The assessment of C3 is the major drawback of the dynamic rebound analysis. However, the method developed by AOKI (1 986) generated excellent results of the mobilized and ultimate loads. The mobilized and ultimate load values determined by VELLOSO ( 1 987) proposition and the rebound technique showed satisfactory results in comparison to the mobilized and ultimate load values obtained by dynamic tests with the PDA, assessed via CASE and CAPWAP. Comparisons with the static load test show results that encourage the employment of this methodology. Rebound measurements are highly affected by the applied energy level, increasing with the applied energy growth up to a maximum limit. Afterwards, stabilization is reached even if the energy keeps increasing. Good results are not obtained with the assumption of a constant C3 value. Because it considers a single point in the load displacement curve, the dynamic load test with the PDA needs some adjusting regarding the used models and the test procedure. C3 values obtained by CAPWAP analysis showed a great scatter, with no logical order with the increasing in the energy. The load-displacement curve ensures more reliable results for the ultimate load evaluation. Conservative analysis can be obtained if the load displacement curve is not used. Time effect on the soil resistance recovery was verified with the rebound technique, during driving and restrike. An increase of 30 and 100 % was observed in the initial resistance. The steel rails piles has a excellent performed when it is request for axial compression.
Figwe 7 - Comparisons between static and dynamic load tests.
7 ACKNOWLEDGEMENT Figure 8 - the mobilized load - displacement curves during the driving and restrike.
The mobilized load - displacement curves show a considerable improvement of the soil resistance with time, in a phenomenon named setup. By analyzing the ultimate load of piles E-10 and E-23, a set up between 30 and 100 % was found. 6 CONCLUSIONS
The authors are grateful to SCAC Fundaqaes e Estrutura Ltda for execution the dynamic load test with PDA. REFERENCES ABNT/MB 3372 (199 1) Estacas - Prova de Carga Esthtica A B N T N R 6 122 (1996) Projeto e ExecuCSo de FundaC6es A B N T N R 8800 (1986) Projeto e Euecu@o de Estruturas de AGOde Edificios (metodo dos estados limtes) A B N T N R 13208 (1991) Estacas - Ensaio de Carregarnento Dmhico AOKI. N (1986) Controle iii sitir da capacidade de carga de estacas pre-fabricadas via repique elastic0 da cravaGdo PublicaCSo da ABMS/NRSP. ABEF E IE/SP, p 48 AOKI. N (1989) A new dynamic load test concept In XI1 INTEFWACIONAL CONFERENCE ON SOIL MECHANICS AND FOUNDATION ENGINEERING. TC ”
The dynamic load test with rebound measurement is a very straightforward and, because it does not require any type of equipment, almost costless technique.
380
Pile Driving, RIO de Janeiro Proceedings for the Discussion Session 11, T 1, p 1-1 AOKI. N (1991) Carga Admissive1 de Estacas Atraves de Ensaio D i n h i c o In I1 SEMINAFUO DE ENGENHARIA DE FUNDACOES ESPECIAIS. S2o Paulo, v 2, p 269-292 LIMA, F M A (1999) ”Analise de pro\ a de carga dinimca e estaca metalica do tipo tnlho” Disserta@o de Mestrado Escola de Engenhana S2o Carlos EESC - USP. 15Op NIYAMA. S , AOKI, N (1991) CorrelaCBo Entre Provas de Carga Dininiica e Estatica no Cainpo Exyenmental da EPUSP/ABEF _In I1 SEMINAFUO DE ENGENHARIA DE FUNDACOES ESPECIAIS. S2o Pado Anais. 1 1, p 285-293 RAUSCHE, F, GOBLE. G G (1985) D!nainic Determination of Pile Capacity Jourtial of Geotechiiical Ei?grneerri~g ASCE, \ 111. n 3. p 367-383 UTO. K, FUYUKI. M, SAKURAI. M (1985) “An EquaQon for the D j nainic Bearing Capacity of a Pile Based on Wave Theon “Proceedings of the Iiiteriiatroiial S~i?lposiiri?ioil Penetmbrlrtv a i d Drivabilitv of Piler San Francisco. v 2 VAN DER WEN. C (1953) “The Bearing Capacity of a Pile” In 3rd INTERNACIONAL CONFERENCE ON SOIL MECHANICS AND FOUNDATION ENGINEERING, ZWch v 2 , p 81-90 VARGAS. M (1990) Provas de Carga em Estacas - Uma apreciaGBo Historica Solos e Rochas Vol 13, Umco, p 3 VELLOSO. D A (1991) “Capacidade de carga por meio do SPT SEFE 11. SBo Paulo, p 293-3 12 VELLOSO. P P C (1987) Fuiida@es - &-lspectosgeoteciiicoc PublicaC2o do Departainento de Engenharia Civil da PUCFU.\ 2.p 300C-300H ‘I
I’
381
This Page Intentionally Left Blank
A ~ ~ l ~ cofaStress-Wave t~o~ Theory to Files, N i y a ~ a& Beim (eds)02000 Baikema, Rotterdam, lSBN 90 5809 1503
A discussion of penetration matching on high strain dynamic pile testing Xu Dingliang Provinc~alTesting Centre o f ~ r a fEnginee~.in~~, ~c Fujian Province, People’s ~
e
~of China u ~ ~
~
c
Wu Shiming Tongji University,Shanghai, People’s Republic of China
Xiao Liping The Inc of Foundution und Construction, Fuzhou, Fujian Province, People’s Republic of China
ABSTRACT: This paper points out the real connotation of penetration matching in High Strain Dynamic Pile Testing. The relationship between some coefficients and CAPWAP result is set forth through the discussion of two actual engineering examples. This paper may be useful to adopt the standard of penetration matching correctly arid to judge the reliability of a CAPWAP result by analyzing the value of some main coefficients. A perfect penetration matching in High Strain Dynamic Pile Testing is always regarded as an important rule to judge a pile’s capacity correctly. The material [I J point out “while using CAPWAP or other similar program to analyze a pile’s capacity the rules must be obeyed, they are: firstly, the accepted coefficients in program should be within a reasonable range; secondly, the matching quality of the curve should reach a certain standard. For concrete piles, the value of MQ should less than 10. Thirdly, the calculated penetration should approximate the tested value”. The state regulations of High Strain Dynamic Pile Testing” also point out: When a matching analyzing coming to an end, the calculated curve should app~oxiniatethe tested curve and the calculated penetration should approxi~nateto tested penetration”. Then, what’s the real connotation of penetration matching? What condition does the standard of penetration matching suit? How will a perfect penetration matching affect tlie control discreteness of analyziiig result? The paper is devoted to discuss these questions from the definitions of penetration and analyzing two examples of actual engineering.
BCTq = 1.O I’ ( Utrn - qav )
(1)
Where Utm is the largest displacernent at a pile’s bottom; qav is the weighted average of elastic-li~it of a pile section (include the pile bottom-section). See Fig. 1. The calculating formula is as follows:: qav = C(qi
*
Rui)/Ru
(2)
Where qi is the elastic-limit of every pile’s section include the pile bottom-section) , the Rui is the static resistance of every pile’s section and RU is the pile’s total static capacity. The value Utm - Qav, or the reciprocal of BCTq the pile bottom’s in forrnula (l), approxi~~ates remains displacement really. See the length AB in Fig. 1. We know the calculated penetration would approximate tested penetration on a pile’s top if the curve matching reached a perfect quality. Therefore, the demand of “perfect curve matching” and “perfect penetration matching?’ means that the penetration on a pile’s top should be equal to the pile bottom’s one and the tested penetration while the CAPWAP analyzing come to an end. It follows that so-called penetration ~ a t c l i i ~ i g 1 The real connotation of p e n ~ ~ a t i o nsuits only integrate piles which are rigid enough. If a pile has a fatal defect in its body or it has a lower matching rigidity body, the penetration on tlie pile’s top is really unequal to the bottom’s one; hence it is According to the section of BACKGROUD in obviously unreasonable to demand perfect Curve material 121, the formula to calculate the times of matching and perfcct penetration matching to be hitting (The Reciprocal of the times of hitting is satisfied at same time. Otherwise the accepted Penetration) in CAPWAP is: mechanical coefficierits of the pile and the soil will contain deviation 383
t
section, and the fissuring under a pile’s bottom etc. There are two examples below showing the possible errors after the standard of a perfect penetration matching and a perfect curve matching has reached. Example 1: At ii worksite were constructed prestressed pipe piles with a diameter of 500mm and a length of 38111. After the piles were completed, we testers used a 6.2t diesel hammer to excite. As different CAPWAP coefficients were adopted, difi’erent results were gotten as 1-A and 1-B: Result 1-A. The analyzed result is as in Table 1, where “Rut” eqrtals the pile’s judged static capacity, “MQ” equals the matching quality value, “Dcw” equals the calculated penetration, and “Dms” equals the tested penetration. ’I’he matching curves is as in Fig 2, where “Fur Msd” equals the measured force curve, “Vel Msd” equals the measured velocity curve, and “For Cpt” equals the calculated force curve.
P i l e bottom’s damping f o r c e
Pile b o t t o m 7
Fig.
The displacement c u r v e of a pile b o t t o m section
2 The possible range of deviation after a perfect penetration matching is reached The PUTose Of adoPtillg the standard for good penetration matching is to control the discreteness of the analyzed results in cooperation with the standard for good curved matching, so as to obtain a better analyzed result. Undoubtedly, the idea is correct. But, besides the static capacity, there are a lot of
Table 1. The anajyzed result 1-A of CAPWAP
coefficients in CAPWAP that may affect the calculated penetration such as radiant soil damping on pile’s circumference and bottom section, the elastic-limit on pile’s circumference and bottom
Rut
MQ
Dcw
Dms
5789.IKN
4*25
2.79MM
2.80MM
Result 1-B. The analyzed result is as in Table 2. The matching curve is as in Fig 3.
384
Table 2. The analyzed result 1-B of CAPWAP. ~
Rut
4089.1KN
MQ
Dcw
Dms
4.13
2.76MM
2.80MM
different results were gotten as in Results 2-A and 2B: Result 2-A. The analyzed result is as in Table 3. The matching curve is as in Fig.4 Table 3. The analyzed result 2-A of CAPWAP.
Example2: At a worksite were constructed prestressed pipe piles with a diameter of 5OOmin and a length of 39.2m. After the piles were completed, we testers used a 5.0t free falling haixnier to excite. As different CAPWAP coefficients were adopted,
Rut
6123.9KN
385
MQ
3.25
Dcw
0.955MM
Dms
1. OOMM
Result 2-B. The analyzed result is as in Table 4. The matching curve is as in Fig 5.
penetration matching has been conducted. In Examplel, compared to the average value, the range of deviation of the judged static capacity is & 17.2%. In Example 2, compared to the average value, the range of deviation of the judged static capacity is +-21.2%. Analyzing the change of the coefficients in CAPWAP in the two examples above, we can find that the bigger the adopted dynamic damping force is, the smaller the pile’s static capacity is. The value of soil elastic-limit reflects the difficulty to penetrate a pile into the soil. The bigger the value of soil elasticlimit is, the smaller the peiietration value is.
Table 4. The analyzed result 2-B of CAPWAP. Rut
MQ
3.67
3975.9KN
Dcw
0.965MM
Dins 1. OOMM
The main coefficients in CAPWA‘E’in the above two examples are as in Table 5. It shows a certain range of deviaticn still exists even though all the matching quality values are less than 5.0 and a good
Fig 5.
I Example
1
2
Result 1-A Result 1-B Result 2-A Result 2-B
The dynamic damping force Damping force of radiant soil the pile’s circum damping force
I Calculated Soil elaStiC-limit
SS=0.25
SK=2.2 MS=16 SK=O MS=O SK=O MS=O
Q2=0.18 QT=0.20 QS=O. 18 QT=0.20 QS=O.l8 QT=0.20
SS=0.67
SK=O MS=O
QS=O. 18 QT=0.20
SS=0.189 SS=O.G9
386
pile’s static capacity Ru
5789.1KN 4089. IKN G 123.9KN
3975.9KN
In order to satisfy the standard of a perfect matching, the radiant soil damping is adopted in Result I-A ( SK=2.2, MS=16--indeed, the radiant soil damping force under this status is less than that under the status of SK=O and MS=O), the damping force on pile’s Circumference is decreased (SS=O. 189) and the static capacity is increased(Ru=5789.1 KN). While in Result 1-B, the coefficients of the radiant soil damping is set to zero (SK=O, MS=O, it means a huge radiznt soil dampkg force is set), the damping force on the pile’s circtimference is increased (SS=0.690) and the static capacity (!
t
R‘A
Fig. 6
decreased (SS=0.25) and a higher calculated static capacity (Ru=6 123.9KN) is accepted. While in Example 2-B, the higher elastic-limits of the pile’s circumference and bottom section are adopted (QS=0.3, QT=0.35), the damping force on the pile’s circumference is increased (SS=0.690) and a lower calculated static capacity (€1 u=3975.9KN) is accepted. If we summarize the relationship among the radiant soil damping force (Rr), the damping force on pile’s circumference and bottom (Rcb) and the pile’s calculated static capacity (Ru), we can draw the relationship as in Fig.6. If we summarize tha relationship among the elastic-limits of the pile’s circumference and bottom section (QS and QT), the calculated static capacity of the pile (Ru) and the value of calculated penetration (Dcw), we can draw the relationship as in Fig.7. According to Fig.6 and Fig.7 we can judge whether a CAPWAP result tends to a safe one or not by checking the main coefficients in CAPWAP.
387
.c Dcw I
QS,
QT
Ru
Fig. 7
3. CONCLUSIONS We can get the four conclusions through the above discussion: A. The real connotation of penetration matching is that the calculated penetration of the pile should approximate the pile bottom’ s one and the tested penetration. B. The standard of the perfect penetration matching does not suit all piles. It suits only the integrate pile that is rigid enough. If a pile has a fatal defzct in it’s body or it has a less rigid body, the standard won’t be suitful. C. It is necessary to adopt the standard of the perfect penetration matching for an integrate pile that is rigid enough, but the analyzing result may be discrete in a wide range. It is necessary to study further the dynamic damping coefficients of soil such as the QS, QT, SS, ST, SK and SM etc. D. Fig.7 and Fig.8 may give a hint for judging the reliability of a CAPWAP result under the current technical level.
REFERENCES [ 11 The “Tentative Regulations of High-Strain Dynamic Pile Testing” published by the State Supervision and Testing Center of Construction Engineering in 1989. [2] “CAP WAP Manual” Goble Rausche Likins and Associates Inc. in 1993
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Experience gained and difficulties in performing dynamic load test in composite piles made with steel rails 6.P. Bernardes Department of Civil Engineering, UNES8 Guuratinguetb, Brazil
C. S.Andreo & C.GonGalves Benaton Pile Foundation Company, Scio Paulo, Brazil
ABSTRACT: The results of dynamic instrumentation in two construction sites that used steel rails as pile foundation are presented in this paper. The first dynamic load tests were executed in piles made by single steel rails of the type TR32 and TR37. In the second group of dynamic tests, the piles were made by a composition of two and three steel rails of the same type TR37. The difficulties in placing the sensors, the effect of hammer blow eccentricity and the influence of the non-uniform welding along the pile length are presented in detail and discussed.
1 INTRODUCTION Piles with unusual cross section as steel rails, represents a difficult task for engineers that carry out dynamic instrumentation. The first problem is how to place the transducers and the accelerometers when the pile cross section does not permit that pairs of sensors to be fixed diametrically opposite or, even for one sensor, to be installed centrally in relation to pile geometry and with the hammer blow. This unusual position of the instruments certainly will interfere in the signals and subsequence interpretation. No difficulties are encountered for placing the pairs of sensors in pipe piles and any deviation in the average signals of force and velocity in the beginning might be related to hammer blow eccentricity that is not difficult to adjust the hammer system. For the case of "H" and "I" steel piles, depending the size of the web and flange, both pair of sensors can be placed in opposite side of the web or diametrically opposite in the flanges (Paraiso 1990). However for this type of piles, force induction is never completely centrally which might result in bending and a great deviation of the signals. One possibility to overcome this problem is to use more pairs of sensors (Bergdahl and Moller 1984) in order to select the best signals for further analysis. More difficulties are encountered when doing dynamic instrumentation in very slender piles and sheet piles (Hartung et al. 1992). They strongly emphases that; "piles with other geometry shapes and consequently other specific characteristics than normal piles, the usual test procedure and evaluation method are only partially adequate". 389
In some regions in Brazil, steel old rails are available in abundance due to the process of renewing the rails road system. This condition make more attractive to use old rails, single and composite, as pile foundation instead of others conventional type of piles. Some requirements for using old steel rails as foundation elements such as the effective cross section area and the yield stress are specified in the Brazilian Standard NBR-6122. Here, the results of dynamic load tests in such piles are presented and discussed. 2 INSTRUMENTATION OF SINGLE RAIL PILE The first experience in performing dynamic instrumentation in single steel rail was during the foundation construction of a residential building in SBo Paulo. The positions of the accelerometers and the strain transducers in the first attempt are presented in Figure la. For this test setup, there was a great deviation in the signals and it was not possible to do further analyses with the average signal. After all the efforts, including the alignment of hammer pile system, the best signal for one pair of sensor is presented in Figure 2, test (a). The second test setup is illustrated in Figure Ib. One piece of rail web was cut and welded in order to place the instruments diametrically opposite. Of course there was an increasing in the pile impedance in that area but, it was possible to evaluate the quality of the signals in each pair of sensors and storage the average signals with little deviation in the beginning, Figure 2, test (b).
Figure 1. Test setups for single steel rail (A T = strain transducer)
=
accelerometer;
Figure 3. Capwapc@signals match. Table 1. Mobilized static soil resistance fiom CAPWAPC@ analyses.
(a> 5,7 500 375 125 14,6 (b) 5,1 645 495 150 13,2 Obs.: EMX = measured energy; R,,t = mobilized total resistance; Rsk = skin resistance; Rt,, = toe resistance; DMX = pile top displacement. Figure 2. Force and velocity traces for different test setups steel rail type TR 32.
The results of the CAF'WAPC@analyses base on wave up matching are presented in Figure 3 for the force (F) signal. The quality match for both cases is very good with "Mqno" equal to 2.17 and 2.46, respectively. However, the mobilized soil resistance was considerable different as presented in Table 1. For test (b), where force (F) and velocity multiplied by pile impedance (ZV) signals are congruent in the beginning, the mobilized soil
390
resistance is 29% higher than test (a) where great deviation occured. In this example, it is clear the necessity of a good control of the test procedure and instrumentation set up to carry out dynamic test in such piles otherwise, their results should be use with caution.
3 INSTRUMENTATION OF COMPOSITE PILE The difficulties and the experience with dynamic instrumentation in composite steel rails were gained
during the foundation construction for a shopping center in the north of Brazil. In order to reduce the number of piles and to increase the bearing capacity, it was used composite piles made by two and three steel rails welded according to Figure 4. The sensors for the first test attempt were placed in the web. However random signals make completely impossible to do any hrther analyses. One of the problems was related to the method used to weld the two rail segments. To avoid bending of the pile by high temperature, the welding was made every other 40 cm along the entire length. Unfortunately, the piles were driven without the welding process complete these 40 cm gaps to guarantee a uniform link along the pile length. It was practically impossible to do any measurement when placing the sensors in the segment that present this gap between the rails. The signals were recorded using the test setup of Figure 4a. For the pile segment above the ground surface the gaps were also welded to minimize the high frequency observed in the first test setup attempt.
Figure 5. (a) Force and Velocity traces for composed two steel rails type TR 37 (b) CapwapcO signal match.
Figure 4. Test setups for composite piles (A T = strain transducer).
=
in the mobilized soil resistance about 52%. However, since no better measurement was possible the results of the CAPWAPC@analyses for both cases, should be use with caution as design criteria.
accelerometer;
However the gaps still exist inside the ground and the best signals measured are presented in Figure 5a. The signals are congruent in the beginning and the influence of the nonhomogeneous welding is clearly seem. The force signal matching fiom CAPWAPC analysis in Figure 5b gave a quality match (“Mqno”) of 4.85. The mobilized soil resistance for plug condition was 672 kN with 50% lateral resistance. The pile geometry made with three steel rails and the tests setup are presented in Figure 4b. Afker many attempts, including placing the sensors in only one rail, the best signals were measured for the setup illustrated in Figure 4b. The stored signals and the force matching are presented in Figure 6. The great deviation in the signals and the gaps along the welding resulted in a “Mqno” equal to 4.29. For this situation the mobilized soil resistance from CAPWAPC analysis (Fig. 6b) was 1020 kN being 40% lateral resistance. Comparing the two pile geometry fi-om Figure 4 there was an increasing
4 CONCLUSIONS Here, the difficulties in placing the transducers and accelerometers in steel rail to carry out dynamic instrumentation were described. For single rail, when placing one pair of sensors in the web and the other in the flange, it is suggested to register the signals separately. For fkther analysis, one must choose the data that present signals congruent in the beginning. Good signals were obtained when using one piece of the web welded in the head rail, in order to place the instrumentation diametrically opposite. For this condition, where little deviation was observed, it was possible to use the average values. For composite rails, the welding must be continuously along the pile to avoid multiple reflection of the signals. For double rails it is possible to place the transducers diametrically opposite in the web. However, for this particular
@
@
391
Hartung, M.; Meier, K. & Rodatz, W. 1992. Dynamic Pile tests on Sheet Piles, Fourth Int. Con$ on the Application of Stress- Wave Theory on Piles, The Netherlands. 259-264. NBR-6122. Standard.
1996.
Brazilian
Pile
Foundation
Bernardes, G.P. 1989. Dynamic and Static Testing of large Model Pile in Sand, Doctor Thesis. Geotechnical Division. The Norwegian Institute of Technology. Trondheim: Norway. Benaton Pile Foundation Company 1998. Dynamic Load test Results, Technical Report 001/98 (in Portuguese) SBo Paulo.
Figure 6. (a) Force and Velocity traces for composed three steel rails type TR 37 (b) Capwapc@signal match.
case, the best signal was for the sensors fixed in the flange. For pile geometry made with three steel rails, welded according to Fig. 4b, the signals should also be registered separately or more pairs of sensors are necessary for hrther analysis. One possibility to eliminate any eccentricity of the ram during itnpact is to use the impact head of the helmet spherically shaped at the top (Bernardes, 1989). REFERENCES Paraiso, S.G. 1990. Steel Piles- Second Part, Technical Note, CTCEWINI, Construction Magazine N.2217, SBo Paulo. 21-24 (in Portuguese). Bergdahl, U. & Moller, B. 1984. Stress Wave Measurement in Slender Steel Piles Driven by Pneumatic Hammers and Drop Hammers, Second Int. Con$ on the Application of Stress- Wave Theory on Piles, Stockholm: Sweeden. 260-272.
392
Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
The application of high strain dynamic pile testing to screwed steel piles J.G.Cannon Independent Geoscience Pty Limited, Melbourne, Vic., Austrulia
ABSTRACT: Screwed steel piles have been in commercial use in Australia for about 3 years and are becom-
ing a popular foundation support for small to medium commercial structures. Although new to Australia this type of pile is described in the old British Foundations code CP4 -1954. In Australia they would previously have been described as micro-piles as they were used for small loadings only but they are now becoming used for more regular pile applications, with serviceability loads increasing to well over 1OOOkN. They comprise a relatively small diameter steel tube with one or more relatively large diameter screw flights near the toe and sometimes at intervals along the shaft. The structural and geotechnical design of these piles is still evolving. This paper will describe several projects where dynamic pile testing has been successfully used to demonstrate the capacity of these piles with a range of sizes, lengths and ground conditions. In one case a static load test was conducted after the dynamic test and provided a close correlation with the dynamic test result in a Grade A prediction. The general methodology used to conduct the successful tests with the PDNCAPWAP system will be described 1 SCREWPILES These piles comprise a steel tube of 325 or 350MPa grade steel with one or more auger screw flights near the toe or at intervals along the shaft. Originally the screw flights were quite “steep” much like a “cfa” auger piling rig, but more recent designs have used a much “flatter” screw and it is much wider than would be used for an auger rig intended for removing soil. Piles for some projects are galvanized for corrosion protection. The piles observed by the Author include a pointed vertical plate at the toe, which prevents any substantial amount of soil entering the steel tube, hence internal fiiction does not need to be considered. Piles have been excavated after installation to inspect disturbance of soils along the shaft. Anecdotal reports fiom independent geotechnical engineers after inspecting excavated anchors in sandy clay suggest that the auger flight does not significantly disturb the surrounding ground mass. This is also suggested by high shaft fiiction shown in dynamic test results as presented in this paper. The piles are screwed into the ground using a hydraulic torque motor attached to earthmoving machinery, normally being an excavator. No specialised purpose-built pile installation machinery is
required and, as commonly available equipment is used, reliability and ease of repair would appear to the Author to be good. Although the equipment that is used is commonly available these piles still appear to be installed by specialised screw piling contractors. We are aware of at least 3 companies that specialise in manufacture and installation of screw piles in Australia. As this pile type is presently used at the low capacity end of the piling “market,” price competition appears to be vigorous. A photo of a typical installation is shown in Figure 1. Piles currently range in size fiom 76mm outside diameter x 3.8mm wall thickness with a screw flight diameter of 350mm and serviceability loads of 501 O O k N to 273mmODx9.3mm wt with a screw flight diameter of l m for serviceability loads of up to a claimed 1800kN. Connections for the pile to the hydraulic motor and splices to other sections of pile currently comprise a bolted external flange butt joint. This has advantages for dynamic pile testing as the horizontal flange stiffens the tube against buckling at the top during the impact of a drop weight. However it is the Author’s view that this has a disadvantage for the behaviour of the pile in that if this flange penetrates below ground level then an open annulus is 393
Figure 1 - Installation of screw pile
bearing. It is accepted that the torque provides a qualitative measure of resistance but the overburden pressure, plasticity and moisture content of the soil and the relative proportion of clay in the soil, particularly at the level of the screw flight, would appear to also affect torque resistance and thus only when these conditions are also considered can the installation torque provide a quantifiable measure of vertical load resistance. With the present “state-ofthe-art” it is considered that the resistance should be assessed using geotechnical calculations, with appropriate uncertainty factors, termed “geotechnical reduction factors” in the current Australian Piling Code (Reference 2). If higher geotechnical reduction factors are to be used (ie lower “factors of safety”) then the Code requires a proportion of piles at each site to be tested in order to verify load resistance. The Author agrees with this requirement. At least one of the contractors has chosen to use dynamic testing by an independent consultant to provide this resistance verification.
created around the pile that reduces lateral support. However, in very soft soils it may be an advantage in reducing negative skin friction. It is understood that at least one company is developing a new pressed fitting, which does not extend beyond the outside diameter of the steel tube. The Author has not dynamically tested a pile with the new joint. Concerns have been raised by some clients that, based on design calculations, the connection between the screw flight and the pile tube is a weak point in the system as almost all pile resistance to vertical load is supposed to be provided by the screw flight. Concerns include the stiffness of the auger flight and the strength of the welded connection to the steel tube. Although these concerns are well founded the Author’s experience is that CAP WAP analyses suggest the shaft does supply considerable resistance through shear fiiction with the soil. The Author has also not observed any failure of the screw flight connection during testing, which is generally to loads of 2 to 2.5 times the serviceability load plus dynamic effects. Nevertheless, it is understood that this concern is also being addressed with development of what has been termed an “interbrace” connection to increase stiffness of the joint and the length of weld at the joint. This should also increase the stiffness of the piling system. The screw pile contractors all attempt to relate installation torque to the load resistance of the pile and adopt torque as the pile acceptance criteria. The Author is not hlly convinced that the installation torque can provide a quantitative measure of vertical load resistance, which is mostly provided by end
2 DYNAMIC PILE TESTING The Author uses the PAK Pile Driving Analyzer from Pile Dynamics Inc. together with the associated CAPWAP signal matching software. The method and current “state-of-the-art” has been described in Goble et a1 ( I 996). A special cable drop hammer and frame was made up by one of the Contractors specifically for dynamic testing of his screw piles. See Figure 2. The drop weights comprise 2tonne and 4tOnr1e solid circular billets of steel of about 350mrn diameter and 2.6m and 5m long respectively. The guide frame allows for a maximum hammer stroke of about 2m. The frame is also supported laterally by 4 guy wires
Figure 2 - Test Rig
394
that are generally tied to adjacent screw piles, being either production piles or temporary anchors specifically for the test. The large weight of the rams compared to the pile weights provides for a long-acting input pulse. This can be seen in the test results presented below, where the resistance also acts for a long time compared to conventional dynamic tests on piles with larger shafts. Testing has so far been conducted on the steel tubes only prior to any concrete infill, where used. Owing to the relatively small shaft diameter compared to the pile capacity it has been generally found that testing is limited by shaft stresses rather than geotechnical capacity of the piles. However this is not always the case.
3 CASE STUDIES
Independent Gaoscience
04-Apr-97
Sydney International Airport, Project: 97037 Pile: SF9 Blow: 1 Data: ReA Foundations Collected: 03-Apr-97 Operator: Jon Cannon CAPWAP(R) Ver. 1996-2 CAPWAP FINAL RESULTS Total W A P Capacity:
482.0; along Shaft
182.0;
at Toe
300.0
W
==J==_PP-=sP_==F=SP--==-~===T=-===~=~==============*~=--=-=-==~~-===========
Soil Dist. Sgmnt
No.
1 2 3 4 5 6 7 8
Depth Below Grade
RU
Below Gage5
Force in Pile at Ru
Sum of Ru
m
m
kN
IrN
kN
482.0 481.9 481.8 481.6 480.7 474.8 393.7 389.2 300.0
.1 .Z .5 1.3 7.2 88.3 92.8 182.0
1.1 2.1 3.2 4.3 5.3 6.4 7.4 8.5
.6 1.6 2.7 3.8 4.8 5.9 6.9 8.0
.1
.1 .2 .9 5.9 81.1 4.5 89.2
Average Skin Values
22.8
Toe
300.0
Unit Resist. Smith Respect to Damping Depth Area Factor kN/m kPa slm
Quake
W.
.13 .08 .22 .81 5.56 76.30 4.24 83.99 22.76
.47 .30 .78 2.91 19.86 272.49 15.14 299.97
.292 .292 .292 .292 .292 .292 .292 .292
mm 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000
76.49
.292
8.000
2439.02
.079
7.000
Soil Model Parmtera/Extensions
Skin
Case Damping Factor ( a of loading quake) unloading Quake ( % of Ru) unloading Level (mm) Resistance Gap (included in Toe Quake) Soil Plua Weioht ( kNl
.900 50 10
Toe .400-Smith Type 100 .500 .15
Figure 4 -Summarv table for SP9
3.1 Sydney International Airport This was the first project where the Author had encountered screw piles and the design appeared to be under early development. Several configurations were tested, being 89x5.5mm tubes with single 350mm screw flights near the toe (working load 1OOkN) and 168x7mm tubes with either two 700mm screw flights (working load 500kN) or with 5 flights (working load 600kN). All piles penetrated the ground about 8m. Ground conditions were moderately dense to dense medium sands.
Figure 3 -Plots for SP9
CAPWAF’ modelling of the piles should not include the screw flight as taking part in the travel of stress waves along the pile ie a significant increase in shaft impedance over a very short distance, of only a few millimetres, at the screw flight is not required. As the flights are only a plate attached to the 395
side of the pile the Author does not accept that it truly is involved in the travel of the stress wave and we almost never model the flights as anything other than supplying a point resistance with low stiffness. Owing to the small shaft diameter conventional (2Fi2V) gauges were placed individually at 90” intervals around the piles. Gauges were bolted to the pile using drilled and tapped holes as with conventional testing of larger steel tube piles. Testing was started with small blows and then larger and larger blows were used until shaft stresses reached or approached the rated yield stress of the steel. As the piles are small diameter and therefore relatively elastic it was found that a cushion was unnecessary between the drop weight and the piles. If the piles had been cut off then an external connection flange was re-attached prior to testing in order to stiffen the top of the pile shaft. The purpose of the testing was proof testing of working piles rather than as pure research tests. In all cases the testing demonstrated more than 2 times the working load. During the testing the piles were “set” as with testing of conventional piles. The fact that a sizeable permanent “set” could be achieved suggested that the testing was mobilising a reasonably high proportion of the available geotechnical resistance. However, the testing was generally stopped owing to high dynamic stresses in the shaft rather than reaching the ultimate geotechnical resistance of the piles with very high “sets.” The piles with 5 flights could not be “set” at all and they were probably significantly over-designed. The Author has not seen this design used again. We CAPWAPed data from all piles. During CAPWAP analysis we used high “quakes” for both
shaft and toe resistance as the shaft resistance was really provided by high level “toes.” The CAPWAP models show resistance concentrated at the flight locations, which seems reasonable. The CAPWAP second toe option was not used to keep compatibility with the multi-flight analyses although this might be just as valid. The CAPWAP analysis was in other respects normal and we would expect similar accuracy to conventional analyses. The results show considerable resistance and owing to the high hammer/pile weight ratio and the elasticity of the system the resistance “stays on” for a long time. This is shown in the data plots and results table for piles SP9 and SP23 in Figures 3, and 4.
3.2 Redcliff Hospital, Brisbane
Independent Geoscienoe
27-nay-98
Redcliff Hospital, Project: 98057 Data: steel Foundations Pile: B-19 Blow: 11 Collected: 27-my-98 Operator: Jon Cannon CAPWAP(R) Ver. 1996-2 M A P FINAL RESULTS
Total CAPWAP Capacity:
1808.9:
along Shaft
1108.5; at Toe
-_===pi
Soil Dist. Depth Sgmnt Below Below
No.
1 2 3 4 5 6
Gages m .9 1.4 1.9 2.4 2.8 3.3
RU
Force in Pile at RU
kN
kN
49.6 49.6 49.6 900.6 59.2
1808.9 1759.3 1709.7 1660.1 759.6 700.4 700.4
Grade
m .6
1.1 1.6 2.1 2.5 3.0
.O
Average Skin Values
184.8
TOe
700.4
700.4
Ir)(
~
sun Unit Resist. Smith W a k e of Ru W. Respect to Daaping
kN
Depth kN/m
Area Factor kPa s/m
mm
49.6 105.20 152.91 99.2 105.20 152.91 148.8 105.20 152.91 1049.3 1910.27 2776.55 1108.5 125.49 182.40 1108.5 .OO .OO
.146 .146 .146 .146 .146 .146
2.616 2.616 2.616 2.616 2.616 2.616
369.50
.146
2.616
569.61 1297.10
soil node1 Parameters/Bxtensions
Skin
Case Damping Factor Unloading Level Soil Plug Weight
.740 20
( % of Ru) ( kN)
.007 25.000 Toe .022-Smith Type .25
Figure 6 - Summary table for B 19
Figure 5 - Plots for B19
Ground conditions comprised completely weathered sandstone. which was essentially very stiff-hard clayey sand or sandy clay with reported undrained shear strength of up to 600-700kPa. This project was much more recent and the piles were much larger and more heavily loaded than previous tests. Piles were 2190DxSmm wt with 2 flights near the toe. The flights were nominally 0.85m diameter but were actually made from square plates 0.76ni across with rounded corners and thus the area would be slightly less than a 0.85m circle. Note that the flight diameters were about 4 times the shaft diameter and it is questionable that they were stiff enough to truly be working at the outer edge. The pile desigrdconstructor is aware of this possibility and is developing a stiffer flight as described earlier. Required working loads were 850 to lOOOkN and it was necessary to demonstrate 2 times the working load. The CAPWAP plots and summary table for pile B19 are shown in Figures 5 and 6. Resistance was again modelled with concentrations at the flights, 396
which is considered reasonable. The fi-iction predicted along the shaft away from the flights is also high at about 15OkPa, which suggests a Meyerhof ~ 0 . 2 5 which , is also considered reasonable. The piles were very short for dynamic testing and it was necessary to use special start-up procedures in order to by-pass CAP WAP default minimum length limits. However, other than the abnormal initialisation a standard CAP WAP modelling procedure was followed. We again CAPWAPed data from all piles. A static load test was conducted on this pile after the dynamic test. The results of the static test were obtained some months after the dynamic testing was reported so the dynamic test was a “Grade A” prediction. Both the static and the dynamic test passed the initial elastic behaviour of the pile and the “break“ in the dynamic test, which uses an elasticplastic model, appears to model the change in slope of the static test with an acceptable accuracy. The
Figure 7 Statiddynamic correlation for B19
correlation is shown in Figure 7. The static load test was conducted in accordance with the Australian Piling Code AS2 159- 1995 to 1.5 times serviceability load with measurement of deflection vs time at various loads. These time dependent measurements have been removed from the static results presented here. The dynamic test mobilised considerably more resistance and the test was again limited only by stresses in the pile shaft.
Independent Geoscience
06-Oct-99
Redlands Hater Aoapital, Project: 200011 Data: Steel Foundations 2t Pile: 109 extension Blow: 1 CAPWAP(R) Ver. 1996-2 Collected: 05-0ct-99 operator: Jon cannon CAPWAP FINAL RESULTS Total CAPWAP Capacity:
425.0; along Shaft
222.0; at Toe
203.0
kN
===C=_==_P=======rJ=F======rFPL=li====IJ====-====
Soil Sgmnt No.
Dist. Below Gages m
1
1.9
2
2.8
3
3.7 4.7 5.6
4 5
3.3 Redlands Muier Hospital, Brisbane This was a recent project and testing had become reasonably procedural. Two pile sizes were tested. The smaller piles were 89x5.5mm with a single auger flight that was basically a 350mm square with chamfered corners. Larger piles 1 14x6mm were also tested. These were also fitted with a single flight that comprised a chamfered square but the chamfers were larger such that the flight was essentially a 450mm diameter circle. Pile penetrations varied from 2.2 to 5ni. Ground conditions comprised stiff clays over stiff to hard gravely clays. The prqject engineer required testing to demonstrate at least 2.5 times the “working” or serviceability load. This demonstrated load varied between 250 and 625kN. We tested 8 piles, which comprised 15% of the total piles. It is the Author’s opinion that with this proportion of testing the required “factor of safety” of 2.5 is unnecessarily high. In accordance with the Australian Piling Code, which is a “partial factor” Code. it would be possible to adopt a geotechnical reduction factor of 0.8. For a typical pile this would suggest an overall “factor of safety” over serviceability of about 1.8.
Depth
Force in Pile at RU kN kN
RU
Below
Grade m 1.6 2.5 3.4 4.4 5.3
425.0 392.5 357.9 311.9 256.2 203.0
32.5 34.6 46.0 55.6 53.2
Average Skin Values
44.4
Toe
203.0
Sum
of RU kN
32.5 67.1 113.1 168.8 222.0
Unit Resist. Smith w. Respect t o Damping Depth Area Factor kN/m kPa s/m 34.85 37.04 49.32 59.61 57.04 41.89
97.35 103.47 137.76 166.51 159.32 132.88 1276.73
Soil Model Parameters/Extenaions
Skin
Case Damping Factor Unloading Level
.500 60
Wainhr
( 8 of
r
nu)
kNI
Quake mm
,185 1.200 .1851.200 .185 1.200 .I85 1.200 .185 1.200 ,185
1.200
.104 39.000 Toe .258-Smith Type .lO
Figure 9 - Summary Table for 109 extension
In this case we did not conduct CAPWAP analysis of data from every test. CAPWAPs were initially conducted on data fiom those piles with a mobilised resistance close to the requirement, this being for about half of the tests. A very close correlation was found between the Case method field estimates and the CAPWAP results, and because the other tests had proved 120 to 170% of the required ultimate resistance using the field “Case Method” hrther analysis was considered unnecessary, especially with regard for the already high “factor of safety.” One of the most highly loaded test piles was tested at 2 different penetrations with the pile being extended by 1.5m between tests. The torque resistance of this pile was quite high during installation at both levels but during testing the measured set was very high (32 and 37mndbl) and the mobilised resistance was also correspondingly low at both penetrations. This is an example of why installation torque does not give a good indication of pile capacity. Eventually this pile was augmented with additional piles. CAPWAP results for this pile are shown in figures 8 and 9. 4 CONCLUSIONS Dynamic testing can be conducted quickly and accurately on this unusual new pile type. Testing can be conducted using similar procedures to dynamic testing of more common pile types with small modifications to allow for the shorter pile length, if necessary. Resistance concentrated at the flight locations provides a realistic model of the pile. The use of testing to verify pile performance should be conducted on every project, as the installation torque does not give a reliable quantitative estimate of pile resistance. Testing on each project is
Figure 8 - Plots for 109 extension
397
already common practice for driven piles where measurement of pile set appears to give a better quantitative indication of pile resistance than the installation torque applied to screw piles. A Grade A dynamic prediction has demonstrated an excellent correlation of the dynamic test results with conventional static load test results.
REFERENCES G Goble + G Likins ( I 996) “On the Application of PDA Dynamic Pile testing” “Proceedings of Fifth International Conference on the Application of Stress Wave Theory to Piles” Orlando, Florida USA. September, Townsend, Hussein, McVay Editors. pp263-273 AS 2 195-1995 “Piling - Design and Installation” Standards Association of Australia
398
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Case study on the application of high strain dynamic pile testing to non-uniform bored piles J.G.Cannon Independent Geoscience Ply Limited, Melbourne, Vic., Australia
ABSTRACT: Dynamic pile testing is used frequently to prove the performance of driven pre-formed piles in Australia. It is either specified by the designer or is offered as an alternative to conventional static testing by piling contractors. However, it’s use on cast in place piles is much less frequent. This paper describes a project at Noosa Heads in Queensland, Australia where an excellent “Grade A ’ correlation has demonstrated that dynamic testing can provide a good prediction of the load vs displacement behaviour of cast in place piles even when the pile shaft is not as designed. The static test pile had the highest over-consumption of concrete at the site, with more than 2 times the design quantity of concrete used during construction. A non-uniform shaft is known to make dynamic testing more difficult and probably less accurate. Nevertheless a very good correlation was obtained between the static test and the dynamic test results for both overall mobilized resistance and the stiffness response of the pile. The static test included measurements that demonstrate potential problems with static testing and these will also be discussed. 1 THESITE
The site is on Hastings St, which is the coinmercial/tourist centre of Noosa Heads in Queensland, Australia. This is located only a few metres behind a popular surf beach and is immediately adjacent to a tidal inlet and small river. A geotechnical investigation of the site had been conducted and 4 borehole logs were provided that described subsoil conditions. The logs suggest that the site is underlain by coastal sand dune or “beach sand” material to a depth of about 10m and this sand is underlain by very stiff to hard clays with SPT results generally 35 but some as low as 18. The sand is loose to moderately dense with standard penetration test results generally in the range 10 to 20 but with some higher and some lower measurements. The SPT results in the sand did not necessarily increase steadily with depth.
2 THEPILES
The foundation contract was let as a design and construct package and the contractor adopted 6OO1nln nominal diameter “cfa” cast in place concrete piles founded at a depth of about loin (ie entirely within the near surface sand with the toe being influenced by the underlying hard clays). These piles were
constructed using a continuous flight auger (cfa) that is drilled into the ground to a pre-selected level or depth and then a cement grout or, as in this case, high slump concrete is injected down the hollow core of the auger as the auger is withdrawn without rotation. The rig used on this project allowed for monitoring of concrete volume and pressure throughout construction of each pile. This permits the pile constructor to assess whether the pile is “consuming” more or less concrete than would be expected for the nominal shaft diameter. When the pile concreting is completed a reinforcing cage is lowered with vibration into the high slump concrete. The piling contractor adopted a high “geotechnical reduction factor” (ie low factor of safety) i n the design and in accordance with the Australian Piling Code (see references) it was necessary to demonstrate the load vs resistance behaviour of the piles. He decided to adopt a single static load test and 4 high strain dynamic tests. One of the dynamic tests was conducted on the static test pile to establish that the dynamic testing would provide a good prediction of a static load test. Many of the piles for this project “consumed” more concrete than would be expected for the nominal shaft diameter but the greatest “overconsumption” was during construction of pile 63 and this was selected for both static and dynamic testing.
399
The over-consumption on this pile was 105%, ie more than 2 times the required volume of concrete was used during construction. The contractor’s equipment provided for measuring pressure and volume throughout construction of the pile so the pile profile could be estimated. The contractors record for pile 68 is shown in Figure 1 . Most of the extra concrete consumption is shown as a cone between 5 and IOm depth. 3 STATIC PILE TESTING
A static load test had been conducted on pile 68 prior to the dynamic pile testing. The contractor was careful to avoid the dynamic testing consultant becoming aware of the static test results before the dynamic test results were reported. The static test was conducted with several cycles in accordance with the Australian Piling Code. Applied load was measured using the jack pressure only. There are shortcomings to this system that are described below. Displacement was measured using 3 dial gauges. A check of pile displacement was also taken using a level survey. Load was applied by jacking against a reaction beam and displacement of this beam was also measured by level survey.
4 DYNAMIC PILE TESTING
The test piles were cast above ground level inside a steel sleeve of about 4mm wall thickness and about the same diameter as the pile for about 2 5 pile diameters above the surrounding ground This was done at the same time as casting the pile or as soon as possible after casting the pile in order to keep the concrete for the extension of the same strength and age as the remainder of the shaft After the concrete had hardened the bottom half of the steel sleeve was removed This leaves a substantial steel collar at the top to reinforce the pile top during impacts of a drop weight and allows the test equipment to be attached at a level where there is a regular smooth surface with no additional impedance that might interfere with strain measurements The location of the test gauges had a diameter close to that of the pile shaft and had similar reinforcing The Author uses the PAK model Pile Driving Analyzer from Pile Dynamics Inc together with the associated CAPWAP signal matching software The method and current “state-of-the-art” has been described in Goble et al (1996) The option to test with 4 strain gauges was not adopted for piles of this size
Figure 1 - Construction record for dynanuc/static test pile
400
The contractor supplied a “Hydroquip” HQ5 hydraulic piling hammer to strike the piles. Some rebuilding of the hammer’s hydraulic valving had been conducted to increase energy transfer efficiency. Highest energy transfer efficiency during this testing was 76%’ which we consider to be high for a hydraulic hammer with a Stonne rain striking a solid concrete pile of this diameter. Testing generally commenced with one or two small blows (0.5m stroke) to ensure haininer alignment was satisfactory and then two or three blows of high energy (1.2m stroke) were applied to gather test data for later analysis with CAPWAP. During CAPWAP analysis the pile was modeled using the construction record but some additional enlargement of the shaft was required near the top. The Author considers the additional pressure created by the shaft extension after the contractors monitoring record was completed justifies this. The model pile volume in the CAPWAP model was very close indeed to the recorded volume of 205% of the nominal design
sonnel that conducted the test claim this was not the case and consider that the jack did not reach the end of it’s travel. However the measurements of the reaction beam, which was also deflecting elastically, show the same behaviour, with almost no deflection during the last load application cycle. This is shown in figure 3. The Author considers that the maximum load applied to the pile did not exceed 1700kN.
5 RESULTS Figure 3 - Hastings St Static Test Reaction Beam
The static results are summarized in Figure 2, below together with the CAPWAP load vs deflection prediction plotted on the same axes.
Figure 2 - Hastings St Load vs Deflcctioii
It can be seen that in the static test when the pile reaches a displacement of slightly less than 151nni the inferred load increases but there is no corresponding deflection of the pile. The Author considers that this is impossible and that there must have been an error in the test measurements The most likely error was that the jack reached the end of its travel or jammed such that although there was an increase in jack pressure and hence inferred load on the pile, in reality load did not increase and the pile consequently did not deflect. The contractor’s per-
As the static test was conducted On this pile prior to the dynamic test it is relevant to plot the dynamic test as an additional cycle to the static test The results of the other tests at this site show a lower “break-point” in the load vs deflection behaviour and the Author considers this to be related to the loading history of the piles The “cfa” constiuction is a non-displacement construction method Owing to the stress relief that occurs durinz construction it would appear that these piles deflect more during initial loading Subsequent loading cycles appear to behave with increased stiffness up to the point of previous niaxiinuin loading This behaviour has also been noted by the Author at other projects with similar piles in sand ground conditions If deflection is a critical acceptance criteria for- this pile type i n sand ground conditions then it inay well be necessary to “preload” the piles by “driving” them after construction If the inaxiinuin applied load during the static load test was 1700kN this correlates well with the “break point’ at about 1800kN in the CAPWAP load vs deflection prediction for this pile The unload/reload stiffness shown i n each of the load cycles of the static test is also of interest as this correlates quite well with the initial loading stiffness shown in the dynamic test results The stiffness of the static test on initial loading during each of the cycles also correlates reasonably well with the dynamic test prediction after the
40 1
“break point.” The Author considers the dynamic test was the first time the pile experienced sufficient deflection to generate a resistance of more than about 1700kN and so after this point the lower initial loading stiffness is valid. The load vs deflection behaviour shown in other dynamic tests at the site were similar, with similar initial loading stiffness and stiffness after the “break point” but they showed a much lower “break point” and a typical example is shown in Figure 4. The Author considers this is because these other piles
Accuracy of the results for bored piles appears similar to driven pre-formed piles provided the pile shaft can be realistically modeled. This requires some knowledge of the shape of the pile shaft. The load-displacement behaviour of “cfa” piles in sand appears to be related to previous load history with initial loading to any level of load being less stiff than reloading. Consequently dynamic testing should be conducted with as few blows as possible if it is necessary to predict initial load stiffness. Designers should be aware of low initial load stiffness. If displacement of this pile type is critical then pre-loading either statically or by “driving” should be considered. 7 REFERENCES G Goble + G Likins (1996) “On the Application of PDA Dynamic Pile testing” “Proceedings of Fifth Inteniational Confereiicc 011 the Application of Stress Wave Theory to Piles” Orlaiido, Florida USA September, Townsend, Hussein, McVay Editors pp263-273
AS2 195-1995 “Piling - Desigii and Installation. Standards Association of Australia. Figure 4 - Coiiiparisoii Pile 68 Static vs Pile 2 18 Dyiiaiiiic
have not experienced high loading and deflection before the test as did the static test pile. However the initial stiffness in these other dynamic tests is still maintained to higher loads than shown in the first cycle of the static test. The Author considers the small blows applied to the pile at the start of each test cause this and owing to the previous loading by the small blows the dynamic test results should be plotted some distance to the right. The pile with more preliminary blows prior to the “test” blow also showed a higher “break point” in the prediction of static load vs deflection, however there were insufficient tests to say that this behaviour has been proven. Further analysis of the test data may provide more information on this behaviour. In particular it may be worthwhile analyzing several blows from the one test to assess the change in “break point.” It is suggested that if it is hoped to avoid this behaviour that the number of small blows applied before the hll test blows should be minimized. This would appear to minimize the “preliminary” loading of the pile and thus provide the best prediction of the first loading deflection behaviour of a cast in place pile. 6 CONCLUSIONS
Dynamic testing appears to be just as valid for bored “cfa” piles as for driven pre-formed piles. 402
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
igh capacity dynamic load tests for bored piles in Sydney shale David J. Klingberg & Phi1 Mackenzie Wagstaff Piling Pty Limited, Brisbane, Qld., Austruliu
ABSTRACT: The installation of bored piles is undergoing some scrutiny with respect to acceptable parameters to be used in the design for shaft resistance and end bearing. A dynamic test pile program was designed for several shallow bored piles socketed into Sydney shale. A twenty (20) tonne drop hammer was used for the dynamic testing. The results were compared to locally accepted design parameters with the test results indicating a significant saving to be made with the adoption of the test pile results. 1 INTRODUCTION The design and installation of bored piles in and around Sydney, Australia is a highly competitive market. The development of the highly refined designs needed for such a competitive environment would prove difficult without the aid of test pile programmes to allow for the confirmation of the design parameters. With the knowledge and confidence that normal construction techniques would provide pile capacities well above the ‘acceptable’ values given from well-used publications for the Sydney area, it was decided to instigate a test pile program to more accurately determine the design parameters to be allowed in the geotechnical design of the bored piles. The test pile program consisted of the installation of three ( 3 ) No. test piles that were subjected to dynamic testing using a twenty (20) tonne, free-fall drop hammer. Two (2) of the piles were constructed with a styrofoam base in order to determine only the shaft resistance component of the pile capacity (ie. there was little or low base resistance due to the presence of the styrofoam at the pile toe). In addition, various construction techniques were used to determine the difference (if any) in the performance of the shaft resistance - one of the test piles was constructed using a grooving tool to roughen the pile-soil interface along the shaft of the pile, whilst the other two shafts remained ungrooved. All test piles were sacrificial piles and were tested to structural ‘failure’ of the pile head under the impact of the testing hammer.
The remainder of this paper presents some of the specifics of the test pile programme, summarises the results of the dynamic testing and CAPWAP@ analyses completed for the test piles and provides a comparison with some published design parameters for bored piles in Sydney shale. 2 SITE DESCRIPTION AND GEOLOGY The site is located in the north western suburbs of Sydney, Australia. Reference to the Sydney 1: 100,000 geological series sheet and fieldwork results indicate that the site is underlain by stiff to hard grey and brown clay over shale and fine sandstone-siltstone laminite. The design values adopted on this project as specified by the geotechnical report are given in Table 1. The test piles were founded in the Class I1 and I11 shales. The values in Table 1 were determined in accordance with the recommendations of Pells et a1 (1978). This is typical of the majority of piling projects within the Sydney region. Table 1. parameters. Rock Strength Verylow Low Medium
403
Geotechnical report Class IV I11 I1
Bearing Pressure (@a) 1,000 2,000 4,000
recommended Skin Friction (@a) 100 200 400
3 TEST PILE PROGRAMME
Table 3. Dynamic testing results summary. Bile Stroke Set TC EMX RMX No. (mm) (mm) (mm) (kNm) (kN) TP1 715 1.0 8.0 62.1 10,960 TP2 320 0.0 6.1 32.1 9,280 TP3 1,070 1.0 10.2 58.7 8,960 Set = pile set in mm/blow TC = temporary compression in mm EMX = max. energy transferred to pile head in kNm RMX = PDA capacity estimate in kN
Test piles TP1, TP2 and TP3 were installed as sacrificial piles for the project. The piles were 500mm diameter (nominal) bored piles with total lengths between 6.3 and 7.1 metres using 50MPa concrete. The installation details and techniques used for each test pile are summarised in Table 2. Table 2. Test pile installation summary. Pile Pen. Pile Grooved Pile WO. (m) Size Shaft Toe YiN Type” (mm) 5.3 500 Y Foam TP1 TP2 5.5 500 N Foam TP3 4.8 500 N Standard * - The pile toe type indicates the method of placement of concrete at the pile toe. ‘Foam’ indicates a 200mm thick styrofoam block was inserted at the pile toe prior to the placement of concrete to provide a pile with negligible toe resistance. ‘Standard’ indicates the concrete was placed in the usual manner at the pile toe to provide a pile with ‘normal’ toe resistance. To enable the three test piles to be subjected to dynamic testing, special heads were cast onto the piles to contain the impact and bursting stresses from the impact of the twenty (20) tonne hammer. The typical dynamic test data is shown in Figure 1. The results of the dynamic testing are summarised in Table 3. PllETEST
PDlPLE DRNNGANALSER
E
R
12500
EMXBlWWm DMX 9 o m
WU2 3430W
,2;
~
30mS
(W
(W
(W
11,846 10,484 6,804
1 11 1,201
11,847 10,495 8,005
Table 5 . CAPWAP’ Unit shaft resistance summary. Depth CAPWAP@Unit Shaft Resistance (m) (Ha) . _ TP 1 TP2 TP3 100 1.o 585 685 1,100 450 2.0 1,150 1,275 825 3.0 1,400 4.0 1,500 1,275 1,175 1,275 2,100 5 .O 2,100
0;RMX 1 1 0 1 0 ~
kN
TP1 TP2 TP3
permanent pile displacements achieved during the dynamic testing were very small and would not have been sufficient to mobilise the ultimate capacity of the pile. The under mobilisation of capacity during a dynamic test is normally related to the end resistance of the pile, but in this instance it may also be applicable to the shaft resistance and that the unit shaft resistance values given in Table 5 may be conservative. All of the test piles were subjected to fiu-ther analysis using the Case pile wave analysis program, CAPWAP@. The results of these analyses are summarised in Table 4. The unit shaft resistance values determined from the CAPWAP@analyses are also summarised in Table 5 .
TP1-RS3 20T DROP M
Table 4. CAPWAP’ analysis summary. Pile Shaft End Total No. Resistance Resistance Resistance
../ \ 4:-, c?
_--__
.*
-,,\,
50 ;“s
,,,,I-‘
4 COMPARISON WITH RECOMMENDED DESIGN PARAMETERS Figure 1. Typical dynamic test data. It should be noted that the data from Table 3 suggests that ultimate pile capacity has not been k l l y mobilised during the dynamic testing. That is, the
The design of bored piles in shale and sandstone in the Sydney region has traditionally been in accordance with the parameters as given by Pells et a1 (1978). The parameters have recently been updated @‘esl et al, 1998) to include ultimate design
404
parameters by a factor of approximately 4.0 to 5.5. The large difference in pile capacities may have very significant cost ramifications for even the smallest piled foundation.
values so that limit state theories can be adopted. These values have been summarised in Table 6. Table 6. Design parameters summary. Ultimate' CApWAp@ Ratio Class Typical Shaft Shaft E Adhesion Adhesion (MPa) (Ha) (Ha) I >2000 1,000 nfa nfa I1
700 to 2000
600 to 1,000
1,275to 2,100
2.0to 3.5
I11
200 to 1200
350 to 600
450 to 1,500
1.3 to 4.3
IV
100 to 500
150
da
nfa
V
50to300
50to 100
nfa
nfa
5 CONCLUSION
# - ultimate shaft adhesion after Pells et al, 1998.
The data from Table 6 clearly shows that the results from the dynamic tests suggest the traditional values of unit shaft adhesion for the Sydney shales are conservative with factors ranging from approximately 1.3 to 4.3. It is, however, noted that the values obtained are in part a result of good construction control and installation technique and that lesser values may be applicable when poor construction control or installation technique are evident. The difference between the dynamic test results and the recommended design parameters can be highlighted with a comparison of design capacities from each method. These have been provided in an approximate and comparative form in Table 7. It should be noted that the base capacities have not been included for the comparison as given in Table 7 as the testing programme was not designed to mobilise ultimate base resistance. The data from table 7 indicates that the shaft resistance values obtained fiom the testing programme exceeded the geotechnical recommended Table 7. Design capacity summary. Depth Geotechnical CAPWAP (m) Report Capacities" Capacities* (W
(W
1 2 3
4 5 Total
155 3 15 3 15 3 15 630 1,730
TP1 400 1,475 1,895 2,290 3,390 9,450
TP2 415 1,840 1,840 2,120 2,135 8,350
TP3 125 560 1,360 1,360 3,400 6,805
405
The test pile programme, using the techniques of dynamic testing methods, has shown that the use of traditional bored pile design parameters for Sydney shale may provide conservative designs by factors ranging from approximately 1.3 to 4.3. The benefits of the test pile programme were highlighted further when compared to the design parameters recommended in the geotechnical site investigation report where the factor ranged from 4.0 to 5.5. These factors were achieved in shaR resistance only as the testing programme was not designed to determine ultimate base resistances. However, it should be noted that the results achieved in this case study were the result of good construction control and installation technique and that lesser values may be applicable when poor construction control and/or installation technique are evident. It is imperative that the designer take these factors into consideration when attempting to employ the indicated values from this study. The case study has also shown that, given good construction techniques, high shaft adhesion values are obtainable that exceed the 'normal' design parameters for the shales found in the Sydney region. However, it should also be noted that it is probable the parameters determined in this case study are lower bound solutions to the ultimate capacities due to the inability of the test hammer to fully mobilise the pile capacity without causing structural distress to the pile top. REFERENCES CAPWAP Manual (1996). Pile Dynamics Inc. Cleveland, Ohio. Pells, P.J.N., Douglas, D.J., Rodway, B., Thorne, C.P. and McMahon, B.J. 1978. Design 'ioadings for foundations on shale and sandstone in the Sydney region. Australian Geornechanics Journal, G8: 31-39. Walker, B.F. and Pells, P.J.N. 1998. The construction of bored piles socketed into shale and sandstone. Mini symposizim on receiif developments ill piling practice in Sydney. Australian Geomechanics Society. Aitpst I998. Presented by Pro$ H.G. Poulos.
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Predicting uplift deflection from dynamic pile testing W.G.Chambers & D. J. Klingberg Wagstaff Piling Pty Limited, Brisbane, Qld., Australia
ABSTRACT: Predicting pile uplift deflection from dynamic load testing on concrete and steel piles is very difficult due to the variable nature of the soils encountered at any site. Two piles, one a 275 mm square reinforced concrete pile and the other a 406 mm diameter (8.8 mm wall thick) steel pipe pile, were tested dynamically with the Pile Driving Analyzer" (PDA) at the end of initial driving and on restrike. These piles were then tested statically to determine pile uplift deflection. For a successfd pile model behaviour response from CAPWAP@analysis, the calculated static deflection from the dynamic test should match the field measured static deflection. This paper reviews techniques applied during CAPWAP@analysis in an attempt to closely match the field measured static deflections in both piles. 1 INTRODUCTION The test piles were to be subjected to static and dynamic load testing to determine load carrying capacity and load deflection characteristics of the piles. Two test piles were installed to determine the uplift capacity of piles driven for a project near Cairns in northern Queensland, Australia. Test pile 1 (TP1) was a 275 mm square reinforced concrete pile, 12.0 metres long with full length reinforcement comprising 4Y28 (i.e. 4 bars of 28 mm nominal diameter steel of 400 MPa yield strength). Test pile 2 (TP2) was a 406 mm outside diameter steel pipe pile (driven open-ended) with wall thickness of 8.8 rnm and a length of 12.8 metres. Pile design methods (Poulos & Davis, 1980) suggest various factors be applied to the shaft resistance values from compression test results for the determination of upliR capacity of piles. It is generally accepted that a reduction factor between 0.5 and 1.0 be used on the shaft resistance for an estimated pile uplift capacity. If a typical value of 0.8 is used on the shaft resistance to determine uplift capacity, a 'Class A' prediction of deflection from CAPWAP@ analysis can be made at various load increments (i.e. deflection predictions completed prior to any static tests being performed).
407
High strain dynamic pile load testing utilising the Pile Driving Analyzer" (PDA) is based on the simplified Case Method for the determination of load carrying capacity (Rausche et al. 1985). Field data obtained is normally analysed by the CAse Pile Wave Analysis Program, CAPWAP@(Rausche et al. 1972). 2 GEOLOGY The site is located some 5 km south of the centre of Cairns in Queensland, Australia. Reference to the Queensland Geology (Scale 1:2,500,000) indicates the site is underlain by Quaternary alluvials and lacustrine deposits. These comprise interbedded layers of sandy clay, sand and clay. The general subsurface condition consists of stiff to hard sandy clay and clay to 17.7 metres. A 1.0 metre thick sand lens is located approximately 11.0 metres below ground level. 3 HIGH STRAIN DYNAMIC TESTS
The piles were subjected to high strain testing using the PDA at the end of initial driving and restrike one day later. The piles were installed using a Banut piling rig operating a six (6.0) tonne hydraulic drop hammer.
3.1 Test Pile I Total Resistance
The PDA test results for TP1 are summarised in Table 1 and the CAPWAP@'analyses results are summarised in Table 2.
As for TP1, if it is assumed that 80% of the shaft resistance as determined from CAPWAP@'analysis was available for uplift capacity, a 'Class A' deflection prediction can be calculated. The predicted pile head deflections are summarised in Table 6.
where: EMX - maximum energy transferred RMX - PDA capacity estimate
Table 6 Summary of Predicted Pile Head Deflections - TP2
From the restrike CAPWAP@analysis, a 'Class A' prediction of deflection was determined assuming 80% of the shaft resistance was available for uplift capacity. The predicted pile head deflections are summarised in Table 3. Table 3 Summary of Predicted Pile Head Deflections - TP1
*It was noted that the CAPWAP@'analysis indicated that TP2 would not attain the maximum static test load if 80 % of the shaR resistance was used as the estimated tension capacity. 4 STATIC LOAD TEST The static load tests were conducted in accordance with the Australian Standard@'Piling Code AS 2159 1995. The tests performed were incremental sustained load tests in uplift. The tests were carried out to a maximum test load of 150% of the Strength Limit State Load, N* (Design Action Effect) which was approximately equivalent to a maximum test load of twice the pile 'working load'.
3.2 Test Pile 2 4.1 Test Pile I The PDA test results for TP2 are summarised in Table 4 and the CAPWAP@'analysis results are summarised in Table 5.
The maximum static test load for TP1 was -951 kN. This test was conducted some three (3) days after initial pile installation and it was anticipated that the effects of pile-soil set-up would influence the comparison of the dynamic and static load test data. The results, together with the 'Class A' predictions, are summarised in Table 7. 4.2 Test Pile 2
where: EMX - maximum energy transferred RMX - PDA capacity estimate
The maximum static test load for TP2 was -702 kN. This test was conducted some five (5) days after 408
Table 7 - Static Load Test Results - TP1
1
A’ Prediction -235
1.36
0.78
I
If the pile-soil set-up effects are ignored, hrther analysis demonstrated that if a reduction in pile modulus of 80 % is applied to the CAPWAP@model, excellent agreement is obtained with the static load uplift test results. A comparison of the revised and ‘Class A’ predictions with the static load test results are presented in Table 9 and shown graphically in Figure 1.
initial pile installation and again it was expected that pile-soil set-up effects would influence the static-dynamic comparison. The results, together with the ‘Class A’ predictions, are summarised in Table 8. able 8 - Static Load Test Results - TP2
-70 -140 -234
I I I
0.12 0.24 0.44
I
I I
0.59 1.20 2.00
reduction in effective pile stiffness due to loss of concrete tensile capacity). The maximum test load corresponded to a reinforcement stress of approximately 385 Mpa, which is well in excess of the stress level that would cause concrete cracking (i.e. approximately 130 Mpa).
I I I
Table 9 - Summary of Static Load ncluding Revised CAPWAP@Deflectio I CAPWAP@analvsis Results Revised “lass A’ Deflection with Load Predicted reduced (kN) Deflection Modulus (mm) (mm) 0.78 2.21 -23 5 1.58 4.48 -476 -571 I 1.89 1 5.53 -666 I 2.21 1 6.66 -761 I 2.54 I 7.87 -856 I 2.88 I 9.19 -951 I 3.25 I 10.66 \
I
rest Results s - TP1 Static Load Deflection (mm)
I
I
~
*It was noted that the CAPWAp@analysis indicated that TP2 would not attain the maximum static test load if 80 % of the shaft resistance was used as the estimated tension capacity.
-
1.36 4.26 5.36 6.46 7.57 8.68 9.92 _i
Test Pile 1
Precast Concrete Pile - 275 rnm. 40 h
E
E
5 COMPARISON OF RESULTS
30
,g
g
5.1 Test Pile I The static load uplift test results and the ‘Class A’ prediction of deflection from the CMWAP@analysis were not in close agreement although the static load test confirmed that pile failure did not occur at the maximum test load. The maximum uplift pile stress was very high and the test results indicated reasonable agreement between the CAPWAP@’ prediction and the static load test up to approximately -200 kN. However, at higher loads the correlation deteriorated. It is postulated that the difference between the predicted and measured deflection was due to the cracking of the concrete section and effective loss of pile modulus (i.e.
ClassA 20
a
1 I0
Static
+ +
Revised
+
al
B
0
0
-236
-476
-571
-666
-761
-856
-951
Uplift Load (kN)
Figure 1 - Chart of Static Load Test Results including ‘Class A’ and Revised CAPWAP@ Deflections - TP1
409
5.2 Test Pile 2 The static load uplift test results show that the ‘Class A’ predictions of deflection from CAPWAP@analysis were conservative. This result was expected due to the subsurface conditions encountered at this site and the timing of the static load test. The static load uplift test was performed some five (5) days after initial pile installation and the dynamic tests were performed at the end of driving and only one (1) day after installation. The additional set-up time between the static load uplift test and the initial driving is likely to account for the difference in the settlement predictions. The static load uplift test for TP2 was carried out to a maximum test load of -702 kN. If it is assumed that 100 % of the shaft resistance was available as tension capacity, the predicted pile head deflections at -562 and -702 kN were 4.80 and 6.11 mm, respectively. A summary of the revised settlement predictions are presented in Table 10 and graphically in Figure 2. Table 10 - Summary of Static Load Test Results
Figure 2 - Chart of Static Load Test Results including ‘Class A’ and Revised CAPWAF@ Deflections - TP2 dynamic testing and CAPWAP@analyses provided a conservative estimate of the pile uplift capacities. That is, the test piles did not fail at the predicted maximum uplift capacities as indicated by the CAPWAP@analyses. This was mostly attributed to the time delay between the dynamic and static load tests and the effects of pile-soil set-up. The results also indicated that where the pile impedance changes were not induced during the static load test, CAPWAP@ analysis provided a conservative prediction of pile head deflection under uplift loads (i.e. steel pipe pile, TP2). However, it was shown that for TP1 (reinforced concrete pile) the effects of cracking under the uplift loads significantly influenced the pile head deflections and that a reduction in pile impedance of 80 % gave excellent agreement with the static load uplift test results. 7 REFERENCES
Poulos, H.G. & Davis, E.H. (1980) Pile Foundation Analysis and Design. John Wiley and Sons. Brisbane.
6 CONCLUSIONS
Test piles 1 and 2 (TPl and TP2) were subjected to a static and dynamic load test programme to determine the uplift and load deflection characteristics for the two piles (i.e. reinforced concrete pile and steel pipe pile). The test programme was also used to determine the reduction factor to be applied to the shaft resistance (as calculated from a dynamic compression test) in estimating the uplift capacity. The static load uplift tests indicated that the use of 41 0
Rausche, F., Goble, G.G. and Likins, G.E. (1985) Dynamic Determination of Pile Capacity. Journal of the Geotechnicai Engineering Division, ASCE, March Vol. 111, No. 3, Pages 367 - 383. Rausche, F., Moses, F. and Goble, G.G. (1972) Soil Resistance Predictions from Pile Dynamics. Journal of the Soil Mechanics and Foundation Division, ASCE, Sept. Vol. 98, No. SM9, Pages 917 - 937.
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Applicability of dynamic load test on a toe improved pile J. Sakimoto & N. Kita Obuyashi Corporation, Tokyo,Japan
S. Nishimura & TTakeda Fugro Geoscience Company Limited, Tokyo,Japan
ABSTRACT: Many dynamic load tests have been carried out to validate the bearing capacity of pile in Japan. However, most of them are performed on the driven piles. The dynamic load test was also performed on the pile having a grouted pile toe by cement in order to improve pile capacity recently. This type of pile was used as the foundation of underground sub-way station structures and platforms In this paper, the author presents an applicability of dynamic load test on the so-called “toe-improved pile” around H-shaped cored steel pile. A result of signal matching analysis using added-mass model for the shaft resistance on the pile clearly shows a reflection of stress waves from the pile toe and suggests that the dynamic load test will be applicable to interpret the bearing capacity of the toe-improved pile. viaduct where the pile driving faced installation difliculties due to less overhead room for the driving work. Under such circumstance, the following pile installation method was adopted. Firstly, the ground was pre-drilled by the auger with diameter of 600mm to the designed pile toe depth of 9m from the ground level. Then cement milk that was designed to provide compression strength of 24Mpa was grouted into the bottom of bore-hole through a grout pipe. Consequently, the H-shaped steel pile with a length of 2.0m was continuously installed into the bore-hole filled by the cement milk. The H-shaped steels were successively connected with the plates, which were welded to the flange and the web, by the bolts for the installation. Total length of the test pile was 10m and the toeimproved length was 2.5m from the pile tip as shown in Figure 1. Table 1 presents a specification of the H-shaped pile.
1 INTRODUCTION
In Japan, many dynamic load tests have been carried out for evaluating the bearing capacity of piles. However, majority of them are performed on the driven piles. Recently many different kinds of pile are introducing to the market. Those piles are aiming to minimize an influence to the adjacent structures particularly in the dense city area. Some of them have an improved pile toe produced by the cement grouting. In this paper, an example of the dynamic load test result on an H-shaped pile that has the improved-toe by the cement grouting is presented. The pile was designed for the underground sub-way station structures and platforms. A dynamic load test was performed on the pile for evaluating the bearing capacity. One of the features in the signal interpretation stage was adoption of the add-mass modeling in the signal matching analysis to derive the bearing capacity of the pile in static. The result of analysis suggests that the added-mass modeling could be able to simulate the behavior of cement grouting part of the toeimproved pile by the dynamic load testing. 2
INSTALLATION METHOD AND PROPERTY
Table 1. Speclficationof the H-shaped pile. Shape
H-350
SOIL
Total pilc length (m) 10.0
(2.0ws,
Young’s
(MPa)
Cross ~ e n s i t y Section (kgin1’) ’4rea
2 . 1 ~ 1 0 ~ 7850
C1rcurn*ircntldi
length
(mA
(In)
171.9
2.054
2.2 Soil property Soil condition of the test site consists of the cemented sandy silt up to 6.0m below GL and the cemented silt bellow it as shown in Figure 1. Both stratums were very stiff and N-value of SPT indicated N=50. However, pre-drilling by the auger
2.1 Iiwtallntioii method of the pile The sub-way station was planned to built underneath of the existing railway station and the railway 41 1
for the shear stress between the grouted cement and its surrounding soil and 12.0MPa for the yield compression stress of the soil below the pile toe. Table 2 shows the calculated design capacities of the pile. The table shows that the total bearing capacity of Case 3 is largest, which is 4.1 1MN, and of Case 1 is smallest which is 1.76MN.
Figure 2. Design patterns of the bearing capacity. Table 2. Design Capacities of the H-shaped pile. Stratum
Stress (MPa)
Circumferential length
Length (m)
Capacity (MN
Disturbed soil Grouted cement
0.001 0.300
2.054 2.054
5.8 2.5
0.012 1.541
Toe
12.00
0.017
0.206
Case I
Figure 1. Soil property and the location of the test pile.
disturbed the soil condition around the pile so that the pile toe was necessary to be improved by cement grouting. Figure 1 presents the soil property and the location of the test pile.
Disturbed soil Grouted cement (Pile vs Conc.) Grouted cement
Case 2 (Conc
3 DESIGN OF THE BEARING CAPACITY
vs Conc )
Toe
Three types of toe were designed for the load test. Variety of the bearing pattern at the pile toe was evaluated in each design. Figure 2 presents the design patterns of the bearing capacity on each pile. In the first case, the shaft fiiction would act on the interface between the grouted cement and the surface of H-steel, and the toe capacity would be generated on a cross-sectional area of the H-shape steel. In the second case, the shaft friction and the toe resistance were considered to be generated on the circumference and on the cross-sectional area of the square formed by the size of H-shape steel, 350mm x 350mm, respectively. In the third case, the shaft fiiction and the toe resistance were considered to be generated on the circumference and the cross-sectional area of the grouted cement body having 600mm in diameter respectively. The figures of design value were assumed to be 0.001MPa for the shaft stress of the disturbed soil, 0.3MPa for the bonding shear stress between the grouted cement and the surface of H-steel, 0.15MPa 41 2
Disturbed soil Grouted cement
Case 3
Toe
0.001
(m2,
Shafi capacity ‘Total capacity 2.054 5.8
1.552 1.759 0.012
0.300
0.7
2.5
0.525
1.000
0.7
2.5
1.750
12.00
‘Slm2, . area
0.123
1.470
Shaft capacity Total camcity 2.054 5.8 1.885 2.5 C.S. area (m2) 0.283
0.7 I2 2.182 0.012 0.707
0.001 0.150 12.00
ShaA capacity Total capacity
3.392 0.719 4.1 I 1
4 DYNAMIC LOAD TEST 4.1 Test equipments A fiee fall steel hammer with the mass of 1.695ton was adopted as the loading equipment. FPDS-5 system was used for the measurement of the dynamic behavior of the test pile during dynamic loading. Figure 3 presents the set up of test equipments. This test was the frst experience of the dynamic load test to the H-shaped steel pile in Japan. The
authors expected that the quality of the signal would be different according to the measuring points due to non-uniform cross section of the H-shaped steel pile. Transducers were therefore set on the flange and the web as shown in Figure 4.
Figure 5 Measured signals (Blow No 1 )
Figure 5 shows the measured signals of Blow No.1, which are time vs. force and time vs. velocity multiplied by impedance. 5 SIGNAL MATCHING ANALYSIS
5 1 Snirth niodel niialysis TNOWAVE program was used for the signal matching analysis to estimate the static resistance from the test result At first, Smith model (Smith 1960) was applied as the soil resistance model, which is the most popular model in dynamic analyses of the driven pile Figure 6 shows the measured and calculated up ward traveling waves from the result of the analysis The values of soil resistance parameters determined by the analysis are presented in Table 3 The total static resistance calculated from the analysis closes with the value of the design Case 1, but the shape of up ward wave shows that the soil resistance was fully mobilized i n this test I n Figure 6, two signals are almost equivalent However the difference between the two signals is observed after the peaks of the signal At the next
4.2 Test yrocedirre and niensirred sig71al
After set-up of the testing system, the dynamic load test was started. The hammer was dropped from the height of 2.0m to the pile head that would generate the rated energy of 33.2kNm. Five blows were carried out with the same rated energy. The settlement of the pile head was about 1.Omm per blow. The energy transferred to the pile was approximetly 1OkNm. The driving resistance was around 3.2MN. Small differences about the quality of the signals were observed due to the difference in measuring locations. The signal measured at the flange had slightly higher quality than measured at the web. This difference was assumed due to the different between flange and web. The thickness of the flange is thicker than the web. There was another worry about the quality of the measured signals due to the influence of the pile connecting plates welded and the bolts. Such mechanical connection might be potential to interrupt proper transmission of the stress wave between the each segment. However the influence of the joint was not appeared on the measured signals.
Figure 6. Result of signal matching analysis using Smith model.
413
Table 3. The result of the analysis including the added mass yield stratum Thickness stress (In) (Mpa) Disturbedsoil 5.3 0.001 Disturbedsoil 0.5 0.310 Grouted cement
Toe
2.5
Quake (mn) 1.0 2.0
Damping Added ( ~ ~ s / m 3constant ) (kg/1n’) 0.001 0.0 0.350 0.0
0.310
2.0
0.350
12.00
2.0
5.00
Static shaft resistance Total Static resistance
289 0.0 1.552 1 759
W, = m g *a (1) where W, = the stress due to the added mass @Urn2), m g= the mass constant (kg/m2), and a = the generated acceleration (m/s2) The value of the mass constant was derived from the magnitude of the added mass, which was calculated by the volume from the size of bored hole, multiplied by the density of the grouted cement Figure 8 shows the measured and the calculated up ward traveling wave signals from the result of the added mass model analysis. Table 3 also presents the adopted values of the mass constant The estimated static resistance did not changed in the new analysis It is presumed that the pile settlement was too small. However Figure 8 shows that the two signals are more similar including the section after the peak of the signals compared to the first analysis (see Figure 6) It is presumed that the motion of added mass, which was the grouted cement actually, might generate the force that pull the pile down to the ground By this phenomenon, a dip of the up ward traveling wave signals appeared As a conclusion, the added mass analysis might be possible to simulate the actual behavior of the grouted cement as a lumped mass
6 CONCLUSIONS
Figure 8 Result of signal inatclung analysis using Added model
11121SS
step it was tried to make the two signals closer at this part of the signal
A dynamic load test was performed on the H-shaped pile which has improved toe section by the cement grouting Following conclusions are drawn from the test results and the consequent matching analyses / The added mass analysis is potential to simulate the actual behavior of toe-improved pile with the cement grouting by the dynamic load test / Minor difference of the qualities of the signal was observed between the signals measured at different locations But, the signal measured at the flange had better quality than that at the web / The influences of the pile connecting plates were not appeared in the measured signals The test result suggests that the dynamic load test can be applicable for the load test of the toeimproved piles However, it is still necessary to accumulate many test data applied on the similar type of piles, and it is also expected that the added mass analysis is possible to provide more reliable estimation of static resistance of the pile
5 2 Added mass model aricr&srs
Added mass model (TNO 1996) was adopted to evaluate the behavior of the grouted cement as a lumped mass Figure 7 shows the pile-soil model adopted the added mass model The inertial stress due to the mass, w.l,is defined as
REFERENCES Smith. E A L. 1960 Pile driring analysis by the rlare equation. ,I Soil .\lecli Found, DIV , 1SW, I ‘0186, t’o SW, 35-61 TNO report - 7’4’0-DLT Dwamrc Lond Testiiig Signal A\latchiiig Lkers .Uaiiunl. 1985 - 1996
414
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Dynamic pile testing practice in Finland H.Jokiniemi, J. Hartikainen & P: Korkeakoski Geotechnical Laboratory, Tampere University of Technology,Finland
ABSTRACT: High Strain dynamic pile testing was introduced in Finland approximately 15 years ago. Today the use of PDA is everyday practice at nearly every job site. The method has been proven to be an especially effective and economical quality control tool for commonly used large diameter steel pipe piles. The Low Strain Pulse Echo Method is also quite popular for impact driven reinforced precast concrete piles in Finland. The benefit of the method is that as many as 300 piles per day can be tested meaning low cost per pile. This paper describes the use of the both methods in Finnish piling practice. research and development. Statnamic has not yet been introduced in Finland. Precast square reinforced concrete piles are the most popular pile types in Finland. The most commonly used cross-sections are 250x250 m2and 300x300 mm’. The concrete strength should be at least 45 or 50 MYa depending on the piling class. Minimum steel-area ratio is 0,6%. Typical design loads for the above mentioned pile sizes are 437,5 to 810 kN. The piles are driven using 30 to 50 kN drop or hydraulic hammers. Design code for driven piles includes instructions for the stroke and required end of driving set. These instructions are based on the method presented by Broms & H e h a n (1972). Their formula is based on the stress wave theory in which the shape of the initial stress wave is assumed and the soil below the pile toe behaves ideal plastic. Typical strokes at the end of driving are 0,25 to 0,50 m and permanent sets 7 to 25 mm per ten blows depending on the hammer type and size, design load and pile length. In spite of the simplicity of the formula it usually gives reasonable criteria for piles driven with drop hammers. However, the piles driven with new hydraulically accelerated Junttan HHK-A-hammers will easily be overdriven. The reason for this is the great efficiency of the latter new hammer type which the formula can not take into account well enough. The use of large diameter steel pipe piles has increased considerably in Finland during the last 15 years. One reason for this was a thorough development project between manufacturers, owners, design engineers, contractors and the university. vary between 1,5 to 7,O MN. Large diameter steel pipe piles are provided nearly always with rock shoes. The
1 INTRODUCTION The soil deposits in Finland were mainly formed during the last glaciation or thereailer as a result of various geological processes. In southern Finland, where the population is the greatest, the thickness of the soil sedimentary deposits is on the order of magnitude of about 10 metres but can also be as much as 60 metres. The bearing layer beneath the soft soil layers is typically very dense glacial till or very hard bedrock. The glacial till usually contains stones and boulders (Heinonen et al. 1997). The above described soil conditions have had remarkable effects on typical Finnish installation methods and pile types. However, it must be admitted that the Finnish concrete piling technology has mainly been adopted from Sweden. As far as large diameter steel pipe piles are concerned, Finland is the pioneer among the Nordic countries. The great efforts in the field of stress wave theory and measurements in Sweden have probably also had an d u e n c e on Finland. Today PDA is everyday practice at nearly every job site and practically the only way to verlfy the bearing capacities of the large diameter steel pipe piles. 2 FINNISH DRIVEN PILING TECHNOLOGY Due to the above described soil conditions most of the piles in Finland are impact driven end bearing piles. Static loading tests are performed very seldom, only a few per year, and these tests are practically serving Typical range of diameters are 356 to 1016 mm and wall thicknesses 10 to 16 mm. Typical design loads 415
rock shoe consists of a dowel and stiffeners which are joined to the thick bottom plate. The aim of the rock shoe is to protect pile base from damages. In some cases the rock shoe can be chiseled into the rock to ensure proper seating. The piles are mostly driven with hydraulic hammers or sometimes using the Frankimethod. The Finnish made Junttan hydraulic piling rigs and hammers arc the most popular equipment among contractors. Due to the large variation in diameters, wall thicknesses and hammers, part of the large diameter steel pipe piles should always be tested dynamically. Large diameter steel pipe piles have turned out to be extremely effective in the construction of railway underpass bridges. The bridge deck is fixed to the piles after constructing it beside the railway track and then placing it on the piles (Hartikainen et al. 1999). In Finland the piling of almost 100 bridges has been carried out during short traffic breaks, often during weekend nights. It is obvious that the only way to verify the bearing capacity is to use PDA. Unfortunately the schedule is often so tight that the PDA-engineer doesn’t have time to make a CAPWAP but use only CASE-method. However, most of the piles are driven to the very dense till or rock where the piles will have plenty of capacity. The lighter hammers may not activate the required ultimate capacity, but the expertise of the PDA-engineer together with careful review of the site investigations are of extreme importance. Small diameter (D=60 to 323 mm) steel pipe piles or driven steel micropiles have recently become more and more popular. They have replaced driven concrete piles at sites where soil layers include boulders, bedrock surface is inclined without supporting layers or where, soil displacement causes problems to nearby structures. Driveability studies using wave equation and in-situ measurements have been demonstrated that it is often possible to drive these piles effectively without excessive driving stresses with the same hammers as used with typical concrete piles. Figure 1 shows the most popular pile driving hammer in Finland, the Junttan HHK-A. The largest model available in Finland nowadays is HHK-9A and the product range of Junttan comprises models fkom HHK-4A up to HHK-18A. All models have a 1,2 m stroke and the number in the hammer type denotes the weight of the ram in metric tons. The ram is slightly accelerated mainly to compensate friction losses during the drop. Therefore the velocity of the ram using full stroke just before impact is 4,9 to 5,O mls. This velocity and the kinetic energy of the ram can be measured with the energy measuring device made by Junttan for each blow. Jokiniemi (1997) analysed the performance of these hammers.
416
Figure 1. Junttan HHK-A hydraulic hammer
3 NATIONAL INSTRUCTIONS CONCERNING DYNAMIC PILE TESTING IN FINLAND The Finnish design codes for driven piles (LPO-87) should be used merely to design precast concrete and wooden piles. These codes were updated in 1987, which means that only a few recommendations are given for dynamic pile testing. It is common practice in Finland to use factor of safety FS=2.0 for dynamic pile testing of concrete piles. The amount of test piles varies depending on the designer and owner. With railway projects 5-10% of the concrete piles are subjected to dynamic pile testing. Up to 30 piles per day can be measured using restrike. However, it is not unusual that similar piles with same design loads used on house building projects are not measured with PDA but are just driven according to the criteria presented in the codes. The practice for low strain testing varies even more because nothing is said in the codes. Again the method is most often used on the railway projects and it is not unusual that every pile is measured. In an 8 hour work shift it is possible to measure more than 300 piles, providing ladders need not be used to reach the pile tops. Large diameter steel pipe piles are normally applied the same factor of safety as for concrete piles although it is possible to use FS= 1,8 on railway projects when the piles are driven to the sound hard rock.
With highway bridge projects normally at least one steel pipe pile per support is measured. On railway underpass bridge projects one pile per track per support should be measured. The CAPWAP-analysis is not required in any Fiimish code. Therefore the choice of the proper damping factor JC in a CASE-method is sometimes at the discretion of the PDA-engineer and his experience. 4 DYNAMIC PILE TESTING SYSTEMS AND THE AMOUNT OF TESTS IN FINLAND Today there are three different companies or organizations in Finland who conduct dynamic pile testing. Two of these companies use PDI’s system and the third uses TNO’s system for high strain testing. For low strain testing there are three different systems: PIT, SIT and PET. The total number of piles measured yearly in Finland is not accurately known by the authors. However, TUT Geotechnical Laboratory is using PDI’s PAK and Piletestxom’s PET in addition to the research also for construction control service at actual job sites. In 1999 more than 600 precast concrete piles and about 300 steel piles were measured by the laboratory using PDA. Low Strain Tests were made for more than 2500 piles. 5 SELECTED EXAMPLES
5.1 Large diameter steel pipe pile and buoyant force during driving As mentioned earlier, evaluation of dynamic pile testing during construction of railway underpass bridges during weekend nights requires sometimes f a t decision making. The existence of large boulders, which can ‘force’ the pile to lean or even prevent the pile from penetrating to the ~ u or required m depth, is checked on every pile position by means of percussion drilling. However, the soil layers can also be very soft and sensitive with high water content causing the closed ended steel pipe pile to ‘float’ during driving. If the pile does not penetrate deep into the hard bearing layer it can even slightly rise &er the weight of the hammer is removed from the top o f the pile. The phenomen makes it difficult to activate the required capacity as the pile may rise several c e n t ~ e t r eduring s the upward stroke of the ram. This “slack” generates a tension wave as shown in the upper curve in Figure 2. Normally under such circumstances the piles are filled with water or loaded several hours with the deadweight of the hanuner after reaching the required depth. 417
Figure 2. Empty steel pipe pile and the same pipe filled with water.
Figure 2 illustrates two dserent stress waves which are measured on the same pile. The diameter of the closed ended pile is 8 13 mm and the wall thickness is 12,5 mm. The upper curve is measured when the pipe is empty and the lower one when filled with water. The impact is caused in both cases using 1 1500 kg drop hammer and a 1,5 to 2,O m stroke. The length ofthe pile is 21,O rn and above the very hard bearing layer are very soft soil layers with a high water content. According to the upper curve the toe resistance is low at 2L/c as the force is low and velocity high. However, later the force increases to quite a high level suggesting that the pile has risen a few centimetres from the very hard bearing layer. By filling the pipe with water the buoyant force acting on the pile is compensated and the pile remains at it’s lowest position. Prediction of the above described behaviour is not easy. On one railway underpass bridge project in Fidand the closed ended steel pipe piles stopped totally on soft clay/silt at a depth of 23 m due to a large buoyant force. After removing the hammer one pile rose slowly as much as 6 m. This can really be called a false stop.
5.2 Small diameter steel pipe pile, friction joint
Smll diameter steel pipe piles installed by different means are superior in underpinning projects. Wave equation anaIysis, such as GRL‘UIEAP, and actual measurements have demonstrated, that it is possible to drive these piles with larger hammers without damages as has earlier been supposed in
Finland. Therefore these piles are nowadays competing more and more with impact driven concrete piles as the contractor can use the same rig and hammer as with concrete piles. The benefit of these piles compared to the square concrete piles is that each pile can also be visually checked after the driving. Earlier these piles were often driven by using light drop hammers (M=250 ..SO0 kg) or pneumatic hammers. Even very light hydraulic hammers, primarly intended as a rock breaking tool, have nowadays been considered. To establish a driving criteria for the light hammers is a very difficult task and the reality and simulation can differ considerably. The signal quality is not often reasonable due to very high acceleration caused by steel to steel impact. Even moderate sized stones can cause a false stop. In Finland, small diameter steel pipe piles are usually jointed using friction splices. Two types are available: external sleeve splice and internal double conic splice. Figure 3 demonstrates one possible problem concerningjointing which can be detected by PDA. The stress waves are measured from 36 m long D=168,3 mm t=lO,O mm closed ended steel pipe pile. The pile consists of three 12 m long elements and the jointing has been made by using internal double conic splices. The pile has been driven to the required depth several days previously using a 3000 kg drop hammer and a 0,20...0,25 m stroke. This stroke causes a downward travelling force in the magnitude of 500 kN which is low compared to the required ultimate capacity of 1200 kN. Restrike tests were started with one low stroke H=0,20 m blow. The two upper curves in Figure 3 represent this blow. The aim of this blow was without breaking the setup to verifL the alignment between the hammer and the pile etc. The locations of the splices can be observed as local small compressive waves. This phenomen is according to the theory that the splice location represents sudden increase of pile impedance. The two lower curves are measured from the next blow using 0,60 m drop height. It is clearly visible that the upper splice is very loose and not tighten properly as a large tension wave reflects fkom upper splice location. However, during later blows the splices gradually tightened. Although the 3000 kg drop hammer is suitable in magnitude for the above described sized piles, this example demonstrates that the internal double conic splices designed for high tension capacity require a few high energy blows to tighten sufficiently. Figure 4 shows determination of the average wave speed for the same pile by using three different systems: PDI’s PAK, Piletest.com’s PET and “NO’S SIT. The pile is 7 m long 300 x 300 mm2 precast reinforced concrete pile driven through a very soft soil.
418
With the PDA this possible problem can be observed and the necessary procedure for the tightening of the splices can be established.
Figure 3 . Loose mechanical splices
5.3 Precast reinforced concrete pile, wave velocify using different systems In Finland the age of the precast reinforced concrete pile should be at least 14 days before it is permitted to be driven. Under certain circumstances, 7 days old concrete piles are allowed to be driven. The typical average wave speeds in Finnish concrete piles are quite low. At the same site different piles can be stored, handled and driven in different ways. This causes different amounts of hairline cracks to the piles and thus different average wave speeds. layer to hard rock or large boulder. The effective length in PDA was 6,5 m. The PDA testing (upper curve) was made by using a 4000 kg drop hammer and 0,3 m drop height. The toe resistance is very high and shaft resistance very
low. Any damage or even small cracks can not be detected. The both low strain testing curves (middle curve using PET and lower curve using SIT) show no signs of cracking but the toe is visible in both curves by using moderate amplification. The average wave speed with every system is approximately the same c=3000...3 100 m/s. These values are however smaller than average wave speeds in Finnish concrete piles. On the other hand, wave speeds of c=4000...4100 m/s are not impossible for Finnish concrete piles. The variation in interpreted wave speeds between different piles at the same site can also part!) be explained by the influence of the pick-up points. It is probable that the pick-up points are structurally weak locations due to early litling of the pile f?om the mould or defective handling during lifting. As the driving stresses may also be excessive at the location of the lower pick-up point, the pile toe can easily be interpreted to wrong location i.e., pick-up point by using high wave speed. This concerns only the low strain method. In Figure 5 is an extreme damage at the point of lift presented.
TUT Geotechical Laboratory during 1993-1997. The analysed data consisted of 326 different piles from 77 different sites. Most of the piles were tested
Figure 5. Extreme damage at the point ofthe lift.
during end of driving. Transferred energy to the pile EMX and maximum displacement of the pile DMX were recorded by the PDA. The permanent set per blow s was recorded manually and the ram stroke estimated as well as possible. The form of the proposed dynamic piling formula is
5.4 Pile driving forrnula,for large diarneler steel pipe piles Viljakainen (1998) analysed dynamic load tests for large diameter steel pipe piles which were made by
P”
OTE .Wr .H =
s-%C
where P,
predicted static capacity
OTE Overall Transfer Efficiency ( E M W E ) PE
Potential energy of the ram
W,
weight of the ram
H
stroke of the ram
s
permanent set per blow
C
elastic deformation of the pile and soil (DMX-s)
The Finnish hydraulic hammers were classified according to the model and typical values of OTE were established. In figure 6.is comparison between Equation (1) and Case-Goble capacity, hydraulic hammers Figure 6 presents a comparison between Equation (1) and Case-Goble capacity. According to the Figure 6, the CO~elationis strong.
Figure 4. Wave speed determination by three different systems 419
REFERENCES Broms, B.B. & Hellman, L. 1972. Methods used in Sweden to evaluate the bearing capacity of end-bearing precast concrete piles. IVA Pdlkomniissionen rapport 34: 27-30. Stockholm. Hartikainen, J. & Koskinen, M. & Suomalainen J. 1999. Effective construction of railway underpass bridges on steel pipe piles. F.B.J. Barends et al. (ed). Geotechnical Engineering for Transportation Infrastructure: I41 -1 46. Rotterdam: Balkema. Heinonen, J. & Hartikainen, J. & Kiiskila, A. 1997. Design of axially loaded piles-Finnish practice. F. De Cock & C. Legrand (ed.), Design of Axially Loaded Piles European Practice: 133-160. Rotterdam: Balkema. Jokiniemi, H. 1997. Egciency analysis of hydraulic Junttan pile driving hammer. Proceedings of the XIVth International Conference on Soil Mechanics and Foundation Engineering: 1077-1080. Rotterdam: Balkema.
Figure 6. Comparison between Equation (1) and CaseGoble capacity, hydraulic hammers
Paikowsky, S.G. & Regan, J.E. & McDonnell, J.J. 1994. A simplified field method for capacity evaluation of driven piles. Report No. FHWA-RD-94-042. Springfield, Virginia.
Equation (1) is actually the same as “true” Hiley formula or Paikowsky’s (1 994) uncorrected Energy Approach prediction. The correlation in Figure 6 may be slightly improved by introducing an additional factor for the soil below pile toe, which should consider the dynamic resistance. It is probable that Equation (1) overestimates the capacity in cohesive soil and therefore the soil type should somehow be considered However, most of the large diameter steel pipe piles in Finland are driven to very dense glacial till or rock. The decision whether the pile has reached the rock is in many cases not hlly reliable without dynamic load testing. Referring alone to the previous, it must be admitted that the profit of the Equation (1) is only to give the magnitude of the capacity. It is not recommended to be used instead of dynamic load testing.
6 CONCLUSIONS The Finnish driven piling technology and dynamic load testing practice is briefly presented. Some unusual examples are also given. ACJSNOWLEDGEMENTS The pile research group at Tampere University of Technology wish to thank Junttan Ltd, Rautaruukki Ltd, Technology Development Center TEKES, VRTrack Ltd, Finnish Road Administration and Finnish piling contractors for long lasting co-operation and financing of research projects.
420
Viljakainen, J. 1998. Steel pipe pile’s dynamic load testing and driving simulation. Master Thesis. Tampere University of Technology. Tampere. (in Finnish).
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
riving behavior of large diameter steel pipe piles Young-Nam Lee Hyundui Engineering and Construction Company Limited, Seoul, Korea
Jong-Sub Lee JungAng Engineering nnd Construction Company Limited, Seoul, Korea
ABSTRACT For the construction of a 4 8 km long Multi-Purpose Jamuna Bridge, 2 or 3 large diameter open-ended steel pipe piles were used for the piers’ foundation A total of 123 piles were driven for 50 piers and 2 test piles from the river bed through the normally-consolidated upper sand layer and rested on top of gravel layer Two types of piles, having 3 15 or 2 50m diameter, were driven to depths of 69 to 74m with the rake of 6 1 Dynamic pile tests were performed on 24 selected piles during pile driving and soil plug length inside the pile was also measured after driving These piles were plugged about 75% of total length of pile driven Piles soon after driving showed a skinfriction dominant pile behavior, that is, 92% of total capacity being developed by skin resistance Quakes values and Smith damping factors were almost constant regardless of pile diameters, reflecting the influence of uniform soil condition at the site
1 INTRODUCTION
and service crossing over the Jamuna River at Sirajganj approximately l00km to the North-West of Dhaka The proposed bridge consists of a total 48 span modules founded on 50 pile groups at each pier location (Young-Nam Lee, and Jong-Sub Lee, 1997) Each span is approximately lOOm long The bridge deck consisted of pre-stressed balanced cantilever units bearing onto concrete piers Each pier location is founded on 1 6 absolute rake 2 and 3 steel pipe pile groups Pile diameters were 3 15m for the 2-pile groups and 2 5m for the 3-pile groups(Figure 2) The bridge was linked to 120m long approach viaducts at West and East Bank
Large diameter steel pipe piles were used in Bangladesh to support Jamuna Multi Purpose Bridge The Bridge was constructed across the Jamuna river, with the surface consisting of normally consolidated loose to medium dense micaceous sands and gravel layer The used piles are either 3 15m or 2 50m diameter steel pipes of variable wall thickness Pile driving was done using a MHU 1700 Hammer(ram weight 102 tonnes) This paper describes the dynamic load test results performed on 24 selected piles during pile driving and soil plug length inside the pile measured after bottom and top section driving in all the driven piles. The Plug length ratio, skin friction distribution during driving, quakes values and Smith damping factors of large diameter steel piles in the Jamuna bridge are suggested.
The Jamuna Bridge was constructed to carry a four lane highway, a 240kV electric power line, a gauge railway, and a high pressure gas pipeline The contractor was a joint venture led by Hyundai Engineering & Construction Co., Ltd. of Korea and Mitsubishi of Japan
2 PROJECT SITE CHARACTERISTICS
The Jamuna river is as wide as 15km during flood season and 5km during dry season. The main stream of Jamuna river shifts each year so river training work is needed.
2.1 /+(?jt?cfdefciih The Jamuna Multi Purpose Bridge(Figure l), Bangladesh, consisted of a 4.8km long combined road, rail
42 1
2 2 Pile details
The piles for the 4 8km Jamuna Bridge are large diameter open ended steel pipe piles There is a total of 121 piles and all are driven on a 1H 6V inclination The Piles are either 3 15m or 2 50 diameter steel pipes of variable wall thickness in the range of 40 to 60mm, and driven in two or three sections The splice is accomplished by use of common offshore “stabbing guides” which are internal pipes of slightly smaller diameter with pointed end to assist in alignment of the two sections, the thickness of the stabbing guides is 25mm and the length of those is ap422
proximately 3m. The piles had the “Shear keys” located neat the pile top and also near the pile bottom(Figure 3 ) . These keys were approximately 20mm thickness, semicircular in shape, located in alternate planes perpendicular to the pile axis, and extended from about 3m below the pile top to about 13m below the pile top. The shear keys at the bottom began one diameter above the pile toe and extended about 13m from the pile toe. Wall thickness for the 3.15m diameter piles was 60mm at the top, reducing 55mm, then 5Omm, then40mm, before terminating at the toe with
Figure 3. Stabbing Guide and Shear Key a 6Omm drive shoe. Wall thickness for the 2.5m diameter piles was 50mm at the top, reducing to 45mm, 40mm, and 60mm at the toe. All piles consisted of two or three sections with total lengths between 38-451n. The second or third sections were equipped with stabbing guide to ensure the pile alignment during stabbing and welding operations. The stabbing guides were 3.4m long.
2 3 Haninier details
Pile driving was done using a MHU 1700 Hammer(ram weight 102 tonnes) All piles extend above the ground and water surface so that the underwater capability of the hammer is not utilized The hammer is nominally rated at 1700 kN-m for underwater operation (1900 above water) The hammer energy readout was monitored This energy is a computed kinetic energy obtained from velocity measurement by two proximity switches located above the impact surface and thus underestimates the kinetic energy at the impact surface due to both gravity and double action of hammer. The pile installation (Figure 4) was performed by Hyundai Heavy Industry using the crane barge “HD 1000” A so-called temporary jig jacket was installed at the piling area supported by four 0 61m diameter temporary anchor piles The pile sections were stabbed through the jig jacket supported by a top guide and bottom guide mounted on the jacket
acteristics are given below (Fugro-Engineers B.V. 1996). CPT results were plotted as shown inFigure 5. 3.1 Upper Sand Layer
These are normally consolidated loose to medium dense micaceous sands. They are fairly uniform and extend to approximately PWD-64 to PWD-68m. Their in-situ density is loose in their upper levels, becoming medium dense depth. Below around PWD50m, they contain occasional fine gravel.
3 SOIL CONDITION
The soils encountered in this project area have been broadly classified into three Units( 1 ) Upper Sand, 2) Gravel, 3) Lower Sand” ). Brief notes on their char423
Based on the CPT data, the excess pore water pressure remained essentially hydrostatic during CPT penetration. Inclusions (lenses) of fine sand and silt at random levels were also indicated. CPT results consistently show a linear increasing tip bearing pressure as a fhction of depth. 3.2 Gr.nvel Lnyer The gravel layer consists of a mixture of both sand and gravel. The gravel sizes range from fine to coarse (up to 6Omm diameter) and occasional cobbles (maximum diameter 80mm) The gravel at approximately PWD-70111 was found in all boreholes and the majority of CPTs The layer is of variable grading, both vertically and laterally The layer often exhibits a “sandwich” layering, consisting of a thick layer of gravel, underlain by sand, and thereafter gravel Occasionally the complete layer is gravel. The gravel layer is approximately 7m thick and dense The excess pore pressure again remained essentially hydrostatic during CPT penetration and no inclusions of silts or clay were found. Below the Upper Sand layer is a sand and gravel layer (where pile tips are founded) with gravel sizes up to 60mm diameter All piles were driven to a depth of 74 to 76m (toe elevation PWD-72m) 3.3
L O M W
Snmi Lnyer
Micaceous sands encountered at elevations varying between approximately PWD-7 1 m and PWD-80m, and exended to depths beyond those investigated (GL-I 20m) Their in-situ density is probably dense
The CPT qc profiles are erratic, probably due to the dense state of the sand, with high negative pore water pressure readings (down to atmospheric) obtained during the CPT pushes. These responses are probably caused by a combination of dilatancy and a lower permeability than the upper sand layer. From the borehole data the sands appear to be uncemented, and old (Pleistocene) in comparison to the overlying Holocene “Upper Sand”. From both the borehole and CPT data, the “Lower Sand” contains inclusions (both layers and lenses) of very dense sandy silt and very stiff to hard clay. The “Lower Sand” is clearly a different layer than the “Upper Sand’, both in terms of its geology (age) and geotechnical (density) characteristics. The detail soil characteristics is summarized in Table 1 4 DYNAMIC BEHAVIOR 4 I P / I ~Lmgrh r < c m
Soil plug measurements were made to investigate soil plug behavior during driving by using a so-called plug follower, a ballast weight placed on top of the soil plug This weight was connected to a wire guided to the outside of the pile shaft to enable the tneasurements of soil plug movements relative to the pile head The plugging can be indicated by the specific recovery ratio and plug length ratio A better indication
Figure 5 . CPT Results 424
Table 1. Soil Characteristics of Jamuna Site Mineralogy Quartz Feldspar wolume Percentage) Mica Finer(%) Size
D1 mm) ()(
D5(1(1111n)
50-65
55-65
35-65
10-20
15-20
10-20
5-15
10
5-15
2-10
0-10
8-20
0.060.20 0.200.40
0.205.00 0.5010.00
0.020.06 0.080.40
Coefficient of Uniformit
of plugging can be provided by the specific (incremental) recovery ratio, which is defined as the incremental change in soil plug lengths with change in pile penetration. The specific recovery ratio is the first derivative of pile penetration depth with respect t o soil plug length (Paikowsky et al, 1989). In Jamuna project , the soil plug length could be checked only two times over driving. The soil plug length was measured after driving of bottom section and after completion of pile driving(bottom and top section) Because it is impossible to calculate the specific recovery ratio, we used the Plug length ratio, which is defined as the length of soil column inside the pile over the total penetration. The Plug length ratio of Jamuna Bridge pile was plotted in Figure 6 Plug length ratio at the end of
Figure 7. Typical Force and Velocity x(EA/C) Traces at a) 3Om, b) 50m, and c) 70m Penetrations for a 3.15m Diameter Pile bottom section driving for the 3.15m diameter piles is about 84.8%, and its value for the 2.50m diameter pile is about 78.5%. Plug length ratio at the end of top section driving of 3.15m diameter piles is about
Figure 6. Plug Length Ratio 425
4.2 Dynamic Pi Ie Testing Results Dynamic pile testing of the 24 selected piles during driving was accomplished with use of Pile Driving Analyzer, and the measured data analysis was performed using CAPWAP(GRL, 1996). Typical dynamic pile testing result was plotted in Figure 7 at the selected depths of 30m, 50m, and 70m. The prediction of the driving resistance was also checked using the simplified Case method(PD1, 1995). The pile capacity calculated by Case method increase with respect to pile penetration depth using case damping Jc 0.5. 4.3 Pile Copacip
Figure 8. Ultimate Capacity at the end of Driving versus Skin Friction Ratio 76.1%, and its value of 2.50m diameter pile is about 72.1%. Final location of the interior soil plug was about PWD-l 5m for 3, 5m piles and PWD-20m for the 2.5m piles, respectively.
Capacity results obtained by the dynamic testing represent those present at time of testing. Generally data presented was for the end of driving. A summary of results is contained in the attached Figure 8. The average capacity at the end of drive of the 3.15m diameter piles is around 70MN, and about 53MN for the 2.5m diameter piles (this capacity is in excellent agreement with the pile diameter). In both pile sizes, shaft friction accounts for about 92% of the total. Unit shaft friction is generally low (less than 100 kPa) down to a depth of about 60m and is developed primarily along the outside skin friction of the pile.
426
Below this level (which roughly corresponds to the top of the shear key location) the unit friction is about 250 to 350 kPa, and probably represents the sum of both inner and outer friction (Figure 9) In addition, the inner soil plug will be to be removed by water jetting and air lifting A concrete plug will then be installed followed by base grouting Thus final inner friction should be discounted entirely, setup on the outside has not been filly achieved at the end of driving, and the soil failure mechanism at ultimate load for the pile toe will be changed due to plug installation and base grouting Thus the capacity of the soil and its dynamic response are only vaguely related to driveability or installation of the piles themselves It is expected that the long term service load these piles will very significantly greater than that presented here, provided that the base grouting is properly performed
one of the most important case study about the one of the largest piles driven in normally consolidated sands. Dynamic pile testing was carried out to investigate dynamic pile behavior and soil plug length was also checked. Piles soon after driving showed a skin-friction dominant pile behavior, that is, 92% of total capacity being developed by skin resistance. Quakes values and Smith damping factors were almost constant regardless of pile diameters. Typically the shaft and toe quakes (QS and QT) were 0.20 to 0.27cm and 0.27 to 0.4cm, respectively. The Smith damping factors for shaft and toe (SS and ST) were about 0.45 to 0.61s/m and 1.24 to 1.69s/m, respectively. 6 REFERENCE
Fugro-Engineers B V ( 1996), " Interpretive Report, Phase I Ground Investigation, Jamuna Bridge Bangladesh, " Report No K-2380/117
4 4 Sorl Pammeler
Dynamic soil parameter results are plotted in Figure 10 Typically the shaft and toe quakes (QS and QT) are 0 20 to 0 27cm and 0 27 to 0 4cm, respectively The Smith damping factors for shafi and toe (SS and ST) aree about 0 45 to 0 61s/m and 1 24 to 1 69s/m, respectively The CAPWAP UN value of unloading was typically 0 4
GRL and Associates, Inc (1996), CAPWAP - CAse Pie Wave Analysis Program, Manual, Cleveland, Ohio Paikowsky, Whitman, and Baligh ( 1989), A New Look at the Phenomenon of Offshore Pile Plugging, " Marine Geotechnology, Vo1.8, pp.2 13233.
5 CONCLUSlON The Jamuna Bridge pro-ject in Bangladesh provided 427
PDI (1995), PDA - Pile Driving Analyzer : PAK User's Manual, Pile Dynamics Inc. I'
I'
Young-Nam Lee, and Jong-Sub Lee (1997), Behavior Prediction and Capacity Estimation of The Large Diameter Piles, Research Report, Hyundai Institute of Construction Technology.
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
ynamic load test of cast in place pile using a free fall hammer S. Niyama, G.C.de Campos & S.Navajas Institute of Technological Research of SGo P a d o State, IPIZ: SGo Paulo, Brazil
S.C.Paraiso & C.M.C.Costa Geomec, Engenheiros Consultores, Belo Horizonte, Brazil
G. E. Barbosa Construtorcr Andrade Gutikrrez S.A., Brazil
ABSTRACT: The use of high strain dynamic test applied to cast in place piles is growing in the foundation engineering practice in Brazil. This paper presents a single case of the use of this method to assess the bearing capacity of large bored pile, part of the foundation of the new Silo Paulo subway line. The use of a special hydraulic self-propulsion hammer delivering high energy, locally developed, allowed the verification of the parameters required by the foundation design. The dynamic test was conducted according to the common procedure, in Brazil, consisting of the application of variable energy. costs. Within this limitation some alternatives of foundation were studied, resulting in the indication of single bored piles or single bored piles connected to the rock mass through root piles executed in their base. Therefore, the test had as main objective, the assessment of the capacity of the single bored piles, particularly their tip resistance, to evaluate the need of the complementary root piles. The option of using high strain dynamic load test, instead of a static load test, was due to economical reasons and also to the time available. Even from the technical point of view, there was a lot of difficulties to perform a static test in the place, close to intense and heavy traffic.
1 INTRODUCTION This paper describes a case of the application of high strain dynamic test applied to a short large diameter bored pile, part of the foundation element of the new superficial subway line in Silo Paulo city. A special hydraulic hammer device was used in this test in order to transfer the impact force to the pile top, delivering sufficient energy to mobilize the total shaft resistance along the pile. The test which was carried out by applying five blows, with increasing energies, made possible the evaluation of the bearing capacity by the CAPWAP method. The Franki type pile was selected as the foundation element for this job, considering the geological and geotechnical conditions, as well as the economical aspects. Nevertheless, enviromental restrictions in part of the job area, where one or two floors old commercial buildings are located including a small medical center with very sensitive equipment - demanded a review of the foundation design. Theoretical study of the vibration and noise level induced during the Franki type piling was done in order to predict the possible effects in those buildings and in medical equipments. The results of the study indicated that dynamic piling should be avoided. The local geological formation, as presented in the next item, limited the length of the foundation element to approximately 11 m, depth of the bedrock. Any use of piles below this depth would demand techniques and equipment with vew high
2 CHARACTERISTICS OF THE SOIL The subsoil in the test area was composed of a stratigraphy in which Quaternary sediments overlay gnaissic residual soils and fractured rock at low depth. The Quaternary sediments from the Silo Paulo basin are very heterogeneous due to their deposition in a sequence of lacustrian regimens. The rocks that underlay the Quaternary and Tertiary sediments are of Upper Proterozoic age, partially constituted by a metassedimentary rock and part of an igneous looking rock. Wheathering of these rocks is also extremely heterogeneous and the residual soils originated from these rocks can show a micaceous 429
Figure 1. Soil profile and schematic of the hammer device sandy silt, as well as a clayey silt matrix depending on their rock of origin. The dynamic testing was done on a production pile. The area was investigated by SPT borings to refusal, one located exactly where the test pile was executed. Figure 1 presents the soil profile from the SPT boring, which presents the sequence of layers as described below. Presence of 1.9m superficial landfill layer, above 2.8 m of soft organic silty clay layer. Below the clay layer there is a fine sand layer with gravels (0,95 m) and, after this, there is a sandy silt residual soil layer. At 11.26 rn depth appears, the top of the bedrock constituted by granite and gnaisse. The water level in this place is located 3.4 m below the land surface. 3 DESCRIPTION OF THE TESTING AND IMPACT DEVICE
The upper 3.0 m of the pile was specially cast in place to allow the support of the hydraulic hammer. This pile was dug using bentonite slurry up to the bedrock depth, then the steel reinforcement cages were introduced and finally the fresh concrete was placed, prepared to have a characteristic compressive strength of 20 MPa. The time interval between casting and testing the pile was of 26 days, so complete cure of the concrete could be assumed. The combination of the vertical and horizontal forces including bending moments, acting on the pile top, resulted in the predicted working load of 2,260 kN for the particular tested pile. The tested pile is one of a group of 8 piles, supporting a single pillar of the elevated way of the metro line. The value of 3,200 d s for the wave velocity was adopted based in the field mesurements. Considering 25 kN/m3 as the mass density of the pile, a value of 26,100 MPa was adopted as the elastic modulus of the pile.
3.I Characteristics of the testedpile The main characteristics of the tested pile are: 1.20 m diameter, total length of 10.50 m, and length below the ground level of only 6.40 m.
3.2 Characteristics of the hammer The device designed to apply high impact forces on the head of the bored pile consists of a free fall self-
430
A plywood cushion thickness of t = L2/2D where t (thickness) is expressed in mm, and L (pile length) and D (pile diameter) are in meters is suggested with a minimum value of 100 mm; an additional 150 mm should be added when pile length exceeds 30 meters. In this particular case, it was determined that a ram of 140 kN, a maximum height free fall of 1.2 m, and a plywood cushion thickness of 150 mm by 1.1 m diameter should be used to comply with the recommendat ions. The diameter of the plywood cushion should be about 80 to, at most, 90 percent of the shaft diameter to be able to centralize the blow. Thickness of steel striker plate to be placed above the top of the pile is suggested to range from 50 mm to 100 mm, and its minimum diameter should be equal to the cushion diameter. The ram is ideally a simple drop weight which can be raised to a variable height. To facilitate the assembly rams are made in segments which can be assembled at the job site. The ram to be used needs to be "guided" to assure that the pile top and ram bottom are perfectly parallel during impact, to assure uniform contact stresses during impact.
propulsion hammer with special features to test piles with high load capacity and variable diameters, varying from 700 mm to 2500 mm. Selection of a proper hammer size is essential for successful high-strain dynamic testing. Hammer weight, drop height, and cushion details must be appropriately chosen so that hammer impact causes sufficient pile movement in order to mobilize the required soil resistance, and to assure that dynamic stresses in the shaft will not impair its structural integrity. The hammer apparatus also was constructed in a way to facilitate mobility around job sites and assure a uniform impact to the head of a pile. The free fall hammer conceived by GEOMEC and in use recently in Brazil consists of the main components, shown in Figures 1 and 2: a steel striker plate, a cylindrical - 2 pieces - steel cage (both of these components being referred to as hammer), a set of round 20 kN steel parts allowing the building of a ram of up to 200 kN with a maximum drop height of 3 m and a hydraulic system to lift the ram with an automatic catcher, which activates and controls the hydraulic jacking equipment. The steel striker plate consists of a round 2.1 m diameter by 100 mm thick steel plate, with four lateral latches that adjust themselves according to the diameter of the pile to be tested. It also functions as a hammer support on the pile top and as a helmet. After positioning and centering the steel striker plate on top of pile to be tested, the assembly of the hammer is started in this sequence: the ram, the cylindrical steel cage, the ram lifter's hydraulic system, the hydraulic jacking equipment, and, finally, the four diametrically opposed telescopic support braces, which maintain the stability of the hammer during the test. Figure 3 shows details of the placement of the hammer on the pile top. To apply each impact, the ram is lifted by the hydraulic system up to the desired height, and the automatic catcher releases the ram on a free fall inside the steel case upon a hydraulic command, hitting the plywood cushion that transmits the high energy to the pile head. The drop height is controlled by a milimetrically graded scale located outside the hammer steel case. To determine the hammer weight, drop height and plywood cushion thickness to be used in this particular test, the bibliographic recommendations of Likins, 1994 and Hussein et al, 1996 were followed: The hammer weight should be at least equal to 1.5% of the anticipated static test mobilized load, The hammer drop height should be approximately 8.5% of pile length, with a minimum value of 2.0 m, 43 1
Between the steel striker plate and the top of the pile, a 20 mm thick by 1.0 m diameter steel wired rubber pad was installed, in order to evenly distribute compressive stresses on the pile head, absorbing the dynamic impact, and preserving the structural integrity of the pile top, above the electronic sensors which will measure data to be associated with the test.
5 blows. The test was interrupted because the quality of the signals was considered satisfactory to make CAPWAP analysis later. For each blow, set measurements were collected from the shaft of the pile, in order to evaluate the pile-soil interaction behavior. For the first blow, 15 mm of permanent set was observed and practically no permanent set was measured in the following blows.
3.3 Test procedure Instrumentation for the dynamic test followed the recommendations of ASTM D-4945/89, AASHTO Designation T 298-93, and ABNT - NBR 13208 (Brazilian Standard). In the present case, considering the 1.2 m diameter of the pile, four strain transducers and four accelerometers were used, located by pairs on each 90" along the pile perimeter. Sensors were installed at 3.46 m from the top and for better adherence of the sensors onto the shaft, its' surface was polished. Field testing was done using the PDA (Pile Driving Analyzer), PAK-95 model, manufactured by Pile Dynamics Inc. The Case Method was used to interpret the data in the field.
Figure 3. Detail for the fixation of the hammer to the pile top The test was done by a sequence of blows with variable drop heights as a dynamic load test with increasing energy. This type of test has been discussed by Aoki, 1989; Niyama & Aoki, 1991; Hussein et al, 1992; Aoki & Mello, 1992, Beim & Aoki, 1996, Paraiso & Mello, 1998, among others. The test was carried out with drop heights of 0.4 m, 0.4 m, 0.8 m, 1.2 m and 1.0 m, a total of 432
4 RESULTS Figure 4 presents the force and velocity x EA/c curves as function of time, for the fifth and last blow, showing a good proportionality. Table 1 presents the results of the maximum Case Resistance (RMX) for each blow applied, for three values of Jc (Case damping factor) and considering the wave velocity of 3,200 m/s. Table 2 presents the main parameter obtained by CAPWAP analysis for the five applied blows. During the analysis it was observed that the best match of the curves was obtained when it was considered a wave velocity of approximately 3,600 m/s, final value adopted for the interpretation of the results. Figure 5 shows the results obtained for the last blow and Figure 6 presents the resistance distribution for all blows. The CASE method indicated values from 5,800 to 8,300 kN for the mobilized resistance, considering Jc = 0.3, value close to that verified later by the CAPWAP analysis. Nevertheless, CAPWAP analysis presented smaller values varying from 5,730 kN to 7,471 kN. This difference could be the consequence of the different values of wave velocity adopted in the field for CASE method and in the office for CAPWAP analysis. The efficiency of the present hammer device varied from 21% to 34% along five applied blows. The CAPWAP analyses show that the maximum RMX was obtained for the blow with maximum energy (blow 4). The skin friction resistance reached the maximum value for the third blow; after this, despite of increasing energy, the skin friction value was smaller, suggesting that the saturation of the available resistance was reached. The toe resistance increased during the sequence of blows, showing that its total mobilization only occurs after the saturation of skin friction resistance. Considering the maximum value of resistance obtained by CAPWAP analysis it was determined
H
CSI
CSX
TSX
EMX
FMX
DMX
(mm) Jc=O,l (kN) (MPa) (kJ) (MPa) (MPa) (m) 0.7 12 5,300 8.9 4.6 3.0 6,500 0.40 12.9 6.3 0.8 19 7,200 4.1 7,400 0.40 9.6 1.1 38 10,900 5.9 10,200 17.2 0.80 9.9 0.7 49 11,200 20.3 7.3 9,700 1.20 0.1 46 11,400 1.oo 23.7 10.0 7.6 8,500 where: H - drop height CSI - maximum tension registered by one of the strain transducers CSX - medium tension stress among four strain transducers TSX - maximum tension stress EMX - maximum energy transferred FMX - maximum compression forces DMX - maximum downward displacement RMX - maximum resistance mobilized during the blow, by CASE method Jc - CASE damping factor
H (m) 0.40 0.40 0.80 1.20 1.oo
Ru (kN) 5,730 6,100 7,000 7,47 1 6,540
RP (kN) 2,000 2,100 2,130 4,195 4,640
Rs (kN) 3,730 4,000 4,870 3,276 1,900
Qs (mm) 2.10 2.10 2.52 1S O 1.20
433
Qt (mm) 2.10 2.60 4.10 4.34 6.40
R M x (W) Jc=0,5 5,200 6,200 7,500 7,900 7,200
Js CASE 0.14 0.10 0.32 0.23 0.10
Jc=0,7 5,000 6,100 7,400 7,600 7,000
Jt CASE 0.32 0.35 0.22 0.08 0.26
that, in relation to predicted working load for the pile (2,260 kN), the safety factor would be greather than 3. However, considering that no permanent set was verified during the application of the blows, it is possible to say that, probably, the total available resistance was not mobilized by the blows.
Figure 6. Distribution of total resistance for the five _ . blows.
AASHTO T 298-93 - American Association of State Highway Officials Standard method of test for highstrain dynamic testing of piles. ABNT NBR 13.208/94 - AssociaqBo Brasileira de T6cnicas - Estacas de Carregamento Dinamico (in Portuguese). Aoki, N. A New Dynamic Load Test Concept. In: ICSMFE, 12, TC Pile Driving, Rio de Janeiro. Proceedings for the Discussion Session 14, v. 1, p. 14, 1989. Aoki, N. and Mello, V.F.B. - 1992 - Dynamic loading test curves. Proc. Fourth International Conference on the Application of Stress-Wave Theory to Piles. Hague, Holland. Beim, J.W., Aoki, n. - 1996 - Dynamic load test method with variable energy. Proc. Fifth International Conference on the Application of Stress-Wave Theory to Piles. Orlando, USA. Hussein, M., Rausche, F. and Likins, G. - 1992 Dynamics of pile driving as a function of ram drop height. Proc. Fourth International Conference on the Application of Stress-Wave Theory to Piles. Hague, Holland. Niyama, S. and Aoki, N.- 199 1 - CorrelaqBo entre provas de carga diniimica e estatica no campo experimental da EPUSP/ABEF. 2'd Seminario de Engenharia de Fundaq8es Especiais e Geotecnia, SEFE, SBo Paulo, Brazil (in Portuguese). Paraiso, S. and Mello, L.G. - 1998 - Variable energy dynamic load test on a 1.0 m diameter CFA pile. BAP 111, November.
5 CONCLUSIONS The present work showed that the use of high strain dynamic testing for large diameter bored is an economically feasible alternative, with advantages upon the conventional static tests. The development of the device designed to apply high impact forces through a free fall self-propulsion hammer made possible the execution of dynamic testing on this kind of pile. The results obtained using this technique are satisfactory and have given important subsidies for the improvement of foundation design criteria.
ACKNOWLEDGEMENTS The authors would like to mention the valuable support offered by Andrade Gutierrez General Constructor for all the help in testing the pile. REFERENCES ASTM D 4945-89 - American Society for Testing and Materials Standard test method for high-strain dynamic testing of piles.
434
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
ynamic testing of large auger pile using free fall loading system in a harbour work S. Niyama, S. Navajas & G.C.de Campos Institute for TechnicalResearch of Siio Paulo Stute, IPT, Siio Paulo, Brazil
ABSTRACT: This paper presents results of high strain dynamic testing carried out on three concrete auger piles, using a special free fall loading system, at Portocel Harbour Terminal enlargement job, in Espirito Santo State. The geometrical characteristics of the piles are: 1.0 m dimeter, 34.71 m, 34.85 m and 33.48 m in length, and maximum compression working load varying from 2,940 kN to 3,293 kN. Previous drivability study was performed for the design of the impact force application device, which was manufactured by the Contractor. Finally, the results obtained by the drivability study were compared with the data obtained during the field tests. 4 m of a superficial soft clay layer, followed by a hard silty clay with presence of sand layers with SPT values between 20 and 40 up to the 25 m depth, limit of the existing borings.
1 INTRODUCTION The dynamic testing of cast in place piles is still relatively little used in Brazil, although a lot of bibliography is now available worldwide. This paper shows an experience with auger piles submitted to dynamic testing, accomplished with success in the Portocel Harbour Terminal amplification job, in Espirito Santo State, located at southeast region of Brazil. The application of high strain testing instead of the three conventional static load tests, initially foreseen in that job, allowed a significant saving of time and also of involved costs. The reaction system in a static load test would demand the execution of anchors in a place with water table varying from 4 to 8 meters, with inherent difficulties and costs. The use of the dynamic method was possible because of the recent revision of the Brazilian Standard for Design and Execution of Foundations - NBR 6122 which has dynamic testing as one of the alternatives accepted for evaluation of the bearing capacity of pile foundations.
3 DRIVEABILITY STUDY A driveability study using the WEAP program was conducted to predict the performance of the special hammer system designed to allow the application of the high strain dynamic load according to the Brazilian Standard NBR 13.208 (1 994). A free fall steel hammer was simulated with 80 kN weight, cylindrical form, and 2.0 m lenght. An efficiency of 95% was adopted and the fall height was varied between 1.0 m and 1.5 m. The hammer cushion was considered as a wood disk with 89 mm thickness, cross section area of 2677 cm2 and elastic modulus of 3,242 MPa. The weight of the steel helmet was considered as 29 kN. For the pile cushion plywood disks were simulated, havin 50 rnm thickness, cross section area of 6,362 cmLi and Elastic Modulus of 207 MPa. The pile length, in that study, was of 31 m, the upper portion of 16 m length with 1.0 m diameter and lower portion of 15 m length with 0.9 m diameter. The concrete elastic modulus was of 40,000 MPa and the specific weight 24.0 kN/m3. Total soil resistances of 3,000 kN, 6,000 kN and 9,000 kN were simulated with different resistance distributions between friction and end
2 SOIL CHARACTERISTICS The Portocel Terminal is located close to the city of Barra do Riacho and of the industry Aracruz, about 80 km to the north of the capital Vitoria in Espirito Santo State. The soil characteristics presents besides the water table varying from 4 m to 8 m depth about
435
Figure 1. Summary of the results of the driveability study for triangular friction resistance distribution.
Hammer =SO kN Hammer =1 .Om
Maximum compression stress (MPa) 13,5
Maximum tension stress (MPa) 625
Permanent Set (m) 3,9
Energy
weight
Bearing capacity (kN) 3000
stroke
6000
13,5
536
175
36,l
bearing. Also, two forms of distribution of the friction resistance were considered: a rectangular form and a triangular one, both beginning at 18.6 m depth down to the pile bottom. Smith damping used for the soil along the shaft was 0.65 s/m, and a quake of 2.5 mm. For the tip resistance, a Smith damping of 0.50 s/m and a quake value of 8 mm were considered. Figure 1 shows a typical result obtained for the case of triangular skin friction resistance distribution representing 80% of the total static resistance. This driveability study presented the results shown on Table 1, for a total bearing capacity equal to the working load, and for twice that value. Those results showed the feasibility of performing the dynamic testing, using the device designed for application of the impact force.
(kJ) 36,l
4 CHARACTERISTICS OF THE PILES The tested pile was installed by driving a steel pipe pile to a hard soil layer and drilling the internal soil plug up to a depth defined by the designers (below the end of the pipe pile) and then inserting the reinforcement cage and placing the concrete. Three test piles have 1,0 m of concrete diameter in the portion with steel pipe casing and 0.90 m diameter in the portion without casing. The relations between the total lengths and drilled lenghts are 34.71 d 18.42 m; 34.85 m43.13 m; and 33.88 d 1 8 . 3 0 m. Additional portion of the piles (from 3.3 m to 3.5 m) were executed above their tops, in order to make possible the use of the impact application device, as well as the sensors. The average resistance of the concrete pile after 436
Figure 2. Schematic of the hammer device Figure 3. Device for impact application
28 days was of approximately 42 MPa. The concrete of the piles was placed more than 30 days before the dynamic tests.
Figure 2 presents the schematic drawing of the device used and in Figure 3 the same device is observed already positioned in the top of the pile by an ordinary crane available at the job site. In this picture the piles located in the proximities can also be observed, with part of the reiforcement cages appearing from the pile heads.
5 IMPACT FORCE APPLICATION DEVICE The impact force applicatjon device was developed and built by GEOBRAS, responsible for the execution of the piles in this job site, under general orientation of IPT - Institute for Technological Research of Silo Paulo, considering the results obtained in the previous drivability study. The device was composed of a steel tube having 1.0m diameter and 8.0 m total length? closed in the bottom by two steel plates of 25mm thickness that also served to support the system upon the pile top and also as the pile helmet. The impact was generated by dropping a cylindrical steel mass of 8,600 kg weight, hoisted by on hydraulic motor fastened in the top of this steel tube, where an automatic de-hook system was designed. The fall height was controlled visually through spaced holes of 10 cm diameter drilled along the steel tube. The hammer cushion consisted of a wood disk 800 mm in diameter and with 100 mm thickness. The pile cushion three plywood disks consisted of 800 mm in diameter and with 25 mm of thickness each one.
6 INSTRUMENTATION The instrumentation consisted of a pair of force transdutores and accelerometers, used in high strain dynamic test. They were fixed on the concrete surface of the pile, through two windows done in diametrically opposite positions on the steel casing surface, located 2,5m from the pile top. It was decided to install the sensors on the concrete surface and not on the steel casing, because the perfect adhesion of both materials could not be verified. 7 TEST DESCRIPTION The high strain dynamic test was conducted adopting the procedure widely used in Brazil, which consists in the application of a series of blows with 437
Pile
1
2
3
d ;2"
hfghts 3'd series series (m> (m> (m> 1.oo 1.oo 1.oo 1.25 1.oo 1.25 1S O 1.oo 1S O 1.75 1.oo 2.00 1.oo 2.00 2.00 1.25 1.oo 1.oo 1S O 1.25 1.25 1.75 1S O 1.50 2.00 1.75 1.75 2.00 2.00 1.oo 1.25 1.25 1S O 1S O 1.75 1.75 2.00 1.75 2.00
increasing drop heights (Aoki, 1989; Niyama and Aoki, 1991). Table 2 shows the different heights used on each pile during the three series of impact applications. In the case of Pile 1, the first two series were used to verify the hammer behavior, allowing also certain accommodation of the cushion materials. The permanent set in all piles measured by
topographical leveling showed a displacement of less than 1.O mm per blow.
8 RESULTS OF DYNAMIC TESTS Figure 4 presents a typical result obtained by the instrumentation for different hammer drop heights, with curves of force and velocity. Because of the increasing drop heights, those results show also growing values of maximum force. The peaks of those curves, however, are not pronounced, showing that the impacts were relatively well softened by the cushion system. In fact, there was the concern not to damage the piles as they were actual production piles, not test piles. The summary of the results obtained for Piles 1, 2 and 3, by the PDA (Pile Driving Analyzer), with bearing capacities estimated by the Case Method, is presented in Tables 3 to 5 . The hammer device presented a low efficiency, from 13% to 26%, computed as the relation between ENTRHU (transferred energy measured by PDA) and potential energy. Pile 1 , the first one tested, presented a lower efficiency due to the significant energy loss at the automatic de-hook system. After the corrections to minimize the loss, a significant efficiency improvement was verified. Table 6 presents a summary of the results of the CAPWAP analyses, where RU is the static
438
Table 3. Results o f t e CASE nethod 1 SERIES I BLOW DMX EMX H(m) (mm> (kN.m) 12 3.5 1st 1.oo 17 4.8 1.25 20 5.4 11 3.3 12 3.1 1.oo 12 3.O 1.oo 12 3.3 1.oo 17 15 3.7 1.oo 12 2.5 10 3rd 1.oo 16 4.2 1.25 19 4.2 1S O 23 5.6 1.75 26 4.5 2.00 5.2 27 2.00 28 5.8 2.00 DMX = maximum ( spiacement EFIC = eficiency (EMX/(weight x drop height)) RMX = maximum resistance
4,200
I
4,700 4,700 5,000 4,500 5,800 6,300
J=0.2 3,800 5,000 5,500 4,100 4,400 4,400 4,600 4,800 4,200 5,200 5,500 6,200 6,800 7,100
RMX J=0.4 3,200
(kN) JzO.6 2,700
4,500 4,900 5,700 6,200 6,400
4,000 4,300 5,200 5,500 5,800
J=0,8 2,200 4,000 4,400 2,700 2,600 3,000 3,100 3,100 3,000 3,500 3,700 4,600 5,000 5,200 5.100
6,300 EMX = maximum energy FMX = maximum force
Table 4. Results o f , e CASE iethod fc pile 2 EFIC. EMX DMX SERIES I BLOW (%) (mm> (kN.m) 20 17 2.8 21 23 3.8 1.25 22 29 4.3 1.5 23 34 5.0 1.75 23 40 4.0 23 20 3.7 23 25 3.9 1.25 26 33 5.9 1.5 22 33 4.9 1.75 23 40 2.0 5.5 22 24 4.2 1.25 3rd 24 31 4.7 26 39 6.3 23 40 5.4
FMX
6,600 7,200 7,900 8,200 8,500 6,800 7,200 8,000 8,000 8,500 7,100 8,100 8,500 8,600
Table 5. lesults of the CASE method for the Dile 3 FMX EFIC. SERIES BLOW DMX EMX (kN) (mm) (kN.m) (%) H(m) 21 1 20 I 5,900 1 I 1st 6,500 22 28 1.5 6.700 19 28 1.75 21 7,800 36 2.0 19 5,300 16 1.o 211d 22 1.25 30 1.5 26 1.75" 33 1.75 32 2.0 defect ve blow ~~
I
+-
439
I
I
6,900 7,800 8,600 8,800 9,200 7,200 7,700 8,500 8,500 9,000 7,500 8,500 8,900 9,100
J=O.O 7,100 8,000 8,300 9,400 6,100
J=0.4 5,700 61400 7,100 7,200 7,300 5,900 6,300 6,600 6,800 7,200 6,200 7,000 7,000 7,200
J=0.2 6,300 7,100 7,800 8,100 8,200 6,500 7,100 7,600 7,800 8,100 6,900 7,700 8,000 8,100
0J=O.O
1 1
1.L-J
J=O.O 4,300 5,400 5,900 4,500 4,900 4,900 5,100 5,300 4,800 5,700 6,100 6,700 7,400 7,700
1
I
1
J=0.2 6,600 7,300 7,600 8,500 5,700
I
I
RMX J=0.4 6,000 6,600 6,900
6,200 6,900 6,500 7,800 7,100
1
J=0.6 5,200 5,900 6,300 6,400 6,400 5,300 5,700 5,800 6,100 6,300 5,600 6,200 6,200 6,400
(kN) J=0.6 1 5,500 6,000 6,200
I
1
5,600 6,300 6,000 7,200 6,300
1
J=0.8 4,600 5,300 5,700 5,600 5,000 5,000 5,300 5,400 5,300
J=0.8 5,000
5,300 5,600 6,200 4,200 5,100 5,600 5,400 6,500 5,600
1
PILE
E13 E47 E53
DROPPING HEIGHT (m) 2.0 2.0 2.0
“SET” (mm) 0a 1 0a 1 0a 1
RU (kN) 7,806 8,195 8,322
RS
(rn)
7,266 5,186 5,578
resistance, Rs the skin resistance and Rp the tip resistance. The CASE dampings and the quakes (Js, Jp, Qs and Q p ) corresponding to the skin and tip resistances are also shown. According to the results of CAPWAP analyses, the bearing capacity of the piles were considered varying between 7800 KN to 8300 kN, which are satisfactory values according to the design requirements.
RP (kN) 539 3,009 2,743
JS
0.915 0.478 0.655
JP
0.022 0.667 0.246
Qs (mm) 2.97 1.63 1.95
QP
(mm)
2.70 2.76 1.47
Niyama, S. & Aoki, N. Correlation of Dynamic and Static Loading Tests in the Experimental Field of University of SBo Paulo. In: SEFE, 2, Silo Paulo. Proceedings, 1991 (in Portuguese).
9 CONCLUSIONS The use of the high strain dynamic test method for evaluation of the bearing capacity of large auger piles was considered fully satisfactory by application of impacts using a special device built for this job site. The use of a previous driveability study was helpful for the design of the hammer device. However, that experience showed that a better evaluation of the effiency to be adopted in the study is necessary. In this study, a 45% global efficiency for the dropping height of 1.0 m was adopted and the measured values in the field were 14%, 23% and 19%, respectively for the three piles. Probably the main loss occurred in the hoisting system and also due to the lack of parallelism between the face of the mass ram and the hammer cushion during the impact. The measured impact velocity was 0.5 lm/s, when the theoretical value was 1.41 m/s for a drop height of 1.O m. ACKNOWLEDGEMENTS The authors wish to thank G E O B R h for the opportunity to perform this work and also IPT for the incentive to prepare paper for this conference. REFERENCES ABNT Brazilian Standard for Design and Execution of Foundations, NBR 6122. Rio de Janeiro, 1996 (in Portuguese). ABNT Piles - Dynamic Loading Test, NBR 13208. Rio de Janeiro, 1994 (in Portuguese). Aoki, N. A New Dynamic Load Test Concept., In: ICSMFE, 12, TC Pile Driving, Rio de Janeiro. Proceedings for the Discussion Session 14,v.1, p. 14, 1989.
440
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Dynamic load test on high capacity pile socketed in basaltic rock Skrgio C.Paraiso, Clhudia Maria C.Costa & Ecidinkia Pinto Soares Geomec,Engenheiros Consultores, Belo Horizonte, Brazil
ABSTRACT: This paper is about the installation of a large diameter pile into weathered basaltic rock and the interpretation of variable energy dynamic load test done with a specific free fall hydraulic hammer. This allows modulation up to 200kN with a maximum drop of 3 meters. A comparison is done with bearing capacity estimations, utilizing empirical methods. This article presents a case of a 1,30 meter diameter high capacity pile socketed in basaltic rock with a working load up to 4400kN. It also consists of the foundation and construction details, geological and geotechnical site characterization, the testing program, variable energy load test, pile instrumentation, tests results, discussions and conclusions.
1 INTRODUCTION The use of the dynamic load test for the evaluation of the load capacity on high load capacity piles with large diameters also serves for the verification of the integrity of the piles. Its employment has been steadily increasing in Brazil with the use of the hydraulic hammer with an available potential energy of up to 600kN.m, and has been especially developed for tests on this specific type of pile. The dynamic load test has became an important tool for the quality control of designed foundations. In this article, the results of the executed dynamic load test are presented along with a description of the equipment used for the test and semi-empiric analyses that were considered in the project phase for the evaluation of the interaction between pile and soil.
the BR-153, on the border between the states of Santa Catarina and Rio Grande do Sul, Brazil. 2.1 Geological framework
The local of the implantation of the Ita hydroelectric dam presents basaltic spill dominions of the Serra Geral formation, with a thickness of approximately 400m, covering the sedimentary rocks (Botucatu sandstone) where the Uruguay River forms a great loop denominated Volta do Uva. The flat, subhorizontal layers of these spills are evident in the hillsides of the river, where descending ruptures occur in the form of steps.
2 GEOLOGICAL AND GEOTECHNICAL SITE CHARACTERIZATION The instrumentation and dynamic test were conducted on pile T6lpillar P5, Figure 1, an integral component of the foundation of the bridge on the Uruguay River, located in the complex of the Ita hydroelectric dam. This bridge has a total extension of 6 3 0 q with a central span of 130% located on
Figure 1 - Cap Block - Pile location. The following are identified as the main geological structures present in the area: DIABASIC DIKE - with an approximate thickness of 20 meters, in a vertical N40W direction, and diagonally cutting the forced bypass tunnels;
44 1
3 BEARING CAPACITY PREDICTIONS
STRUCTURAL ALIGNMENTS - linear, subvertical geological features, where differentiated movements occur along the length; and DESCONTINUITY J - structures present in the bed of the river at an elevation located between 240 and 250 meters, a sub-horizontal zone which is weathered and fractured with a high rate of permeability (spill 9. The area is also characterized geologically by an excessive, thick mantle of soil, saprolito, and weathered rock, with a depth varying from 10,O to 20,O meters in the hillsides, and that could surpass 30,O meters in other localized areas (Antunes Sobrinho, 1999 and Infanti, 1999).
Non-confined simple compression tests were performed on core samples of basalt rock with a 22% 5 R.Q.D. 5100% whose results, presented in Chart 1, were submitted to a probability analysis in the concept of Gauss, allowing a 95% margin of reliability (a characteristic value): Chart 1 - Results of the compression tests. I$ CS Rupture Tension -OC (MPa)
I
2.2 Geotechnical characterization of the test area Keeping in mind that we are dealing with a design of great size, the implantation area was thoroughly investigated, through rotary core borings along with special tests in the laboratory and field. Figure 2 presents the results of rotary boring SM-02, used as a reference for the tested pile. In agreement with this log of boring, the underground stratigraphy is composed of a superficial layer of silty clay, not very sandy, with a medium thickness of 10m and an average SPT of 10 along with basalt boulders -having an RQD (Rock Quality Designation) varying from 0 to 48%. It has a recovery between 38 to 98% - with a medium thickness of 6m, interspersed by a layer of silty clay, not very sandy, and terminating in sound basalt rock with veins of quartz - an RQD and a recovery varying from 91 to 100% - to the limit of the boring.
1
1
2
1
89.0
52.5
3
138.8
4
89.0
5
1
149.7
6
1
65.3
7
1
164.0
8
1
63.6
9
1
105.0
Maximum Value (o~M~J Minimum value (OcMin) Mean value (X) (s) Standard deviator Characteristic value (x) I;, ~ = X - 1 . 6 5 ~
= = = = =
164MPa 52.5MPa lUlMPa 40MPa 35MPa
According to Zhang (1998), the ultimate capacity for a pile embedded in rock can be estimated as such: (QSp
= Qal+
Qp
in that: (QJP = ultimate bearing capacity; Qal = Lateral skin friction; and Qp = point bearing capacity.
a) Skin Friction Broms (1988), and apud Horvath et al. (1983) have proposed the following equation for weak rocks:
z ,
= (0.20 to 0.30).,/0,
ZMax
Figure 2 - Geotechnical characterization.
10.2) to (0.30)
442
=
=
MPa
maximum unit socket shear; and emoirical factor.
To obtain the oc value, the pile embedded in the weathered basaltic rock having an RQD of 0 was considered. According to the local geology and geomechanical classification presented in Chart 2 (Lima, 1983, Bieniawski, 1984 and Barton, 1974), a RQD <25% average was adopted, stated as, cy slOMPa. The following values found for the unitary lateral friction (zh.ia~) and resistance due to the lateral fiiction (Qa$, are presented in Chart 3.
L d=l+0.4LI3.4 bs in that: L, = socket length, and b, = socket diameter.
(3.5)
substituting the values, we have: 3.2 I 3.4 (3.6) I .3 The evaluation of the end bearing capacity, whose results are presented in Chart 4, considers the characteristics of the support rock, signifying a recovery of 73 and a 48% R.Q.D. It has a characteristic resistance of 0, of 35MPa, according to the results of the probability analyses obtained through simple compression rock samples (Chart l), values which are confirmed in Chart 2. d =1+0.4-=1.98
CLASSIFICA TION
RQD 100
Excellent
> 200
90 - 100
Excellent
> 200
75 -90
Good
I00 - 200
50 - 75
Regular
50 - 100
25 -50
Weak
25 - 50
< 25
Very Weak
1-10
Chart 4 - Toe resistance.
According to Canadian Foundations Engineering Manual (1985) and Ladanyi and Roy (197 1):
4,,
=o,.k,d
(3.3)
: 0.13
0.17 0.20 0.13
0.17
0.20
CT~;M~~
164.0 42.21 56.20 64.94 56,140 73,420 86.370
o,M,,
52.5
13.51 17.67 20.79 17,970 23,500 27,650
35.0
9.01
101.0 26.00 34.00 40.00 34,580 45,220 53.200
in that: Q M ~= maximum toe bearing pressure; X
average unconfined compressive strength cr, = of rock core; K ~ , , = empirical factor; and d' = depth factor.
11.78 13.86 11,980 15,670 18,430
It was considered, in the design phase, that the ultimate capacities of skin fiiction and toe resistance would be the following:
Chart 3 - Skin friction.
3270kN 2 to oc = lMPa, 7Max = 0.25MPa 11,980kN s to oc = 35MPa, qMax = 9.01MPa and ksp= 0.13 (Q,Jp 3270kN + 11,980kN = 15,250kN QUi
Qp =
TMax
(ma)
0.25
0.20 0.25 .0.30
0.20
10
0.63
0.95
8230 10,320 12,420
8
0.57
0.71
0.85
7450
9280
11,100
6
0.49
0.61
0.73
6400
7970
9540
0.79
0.30
5
0.45
0.56
0.67
5880
7320
8760
3
0.35
0.43
0.52
4570
5620
6800
2
0.28
0.35
0.42
3660
4570
5490
1
0.20
0.25
0.30
2610
3270
3920
1 8
--I
k,
1 5
I -
It has been verified that the safety factor for a work load of 4400kN is 3,4, and therefore satisfactory. 4 HIGH DYNAMIC LOAD TEST 4.I Pile and Instrumentation Features
The conditions of construction and the test, are represented in Figure 3.
(3.4)
443
Figure 3 - High Dynamic Load Test - Typical setup. The pile was cast in place with the concrete fck=25MPa (characteristic compressive strength) and heavy longitudinal steel reinforcement associated with spiral reinforcement cage. Instrumentation for the Dynamic Load Test (DLT) and the test itself followed the recommendations of ASTM D-4945/89, AASHTO designation T-298/93 and ABNT-NBR 13208/94 (Brazilian Standard). Constructively, the shaft space incorporated between the confining soil and the top of the alluvium of basalt is constituted by a steel casing with a 3/8" thickness and an internal diameter of 1,20m driven with a casing oscillator rig. In order to carry out the test, a supplementary 47cm shaft of concrete was executed above the steel casing, where the conditions of longitudinal reinforcement were maintained according to project, altering the spacing of the spirals to 8,Ocm. The excavation of the rock was accomplished with the use of explosives and a compressed air system, for the purpose of expulsing underground water which originates in the hydraulic load of the Uruguay River. After the excavation in rock, the pile was concreted to a quota of 333.62, corresponding to Figure 3.
The instrumentation proceeded through the installation of four strain transducers and four accelerometers, positioned every 90 degrees, in a manner to minimize the possible effects of bending resulting from the dynamic impact. In the case of the accelerometers, two piezoeletrics and two piezoresistives were used. The opening of four "windows" became necessary in the steel casing, for the installation of the sensors in the concrete nucleus, as represented in Figure 4. Chart 5, which follows, specifies the characteristics of the tests. Chart 5 - Test features. COLUMN = P5 BORED PILE = T6 SOCKET
4
LENGTH
MU.
ROCK
120
130
(tD
3.20
440
_. .
444
TOTAL
(4
*Maximum worklng J
LENGTHS (m)
16.29
BELOW
BELOW
SENSORS
GL
15.01
14.01
4.2 Hammer Description
- The hammer weight should, at least, be equal to
1.5% of the anticipated static test mobilized load;
Dynamic Load Test on cast in place piles with high load capacity have had routine application in the USA, Europe, Asia etc. and in Brazil. The technical literature available gives an account of several cases of dynamic tests on cast in place piles, as discussed by Rausche and Seidel (1984), Balthaus et al. (1985)Townsend et al. (1991), Seidel et al. (1996), Sawai et al. (1996) etc. The option of using the high strain dynamic load test (DLT) instead of the static load test, which is routine in many places, comes from the attractiveness of the economics and the scheduling benefits. The cost of the high strain dynamic test in Brazil is 25% less than the cost of a static load test (SML-slow maintained load), and represents a reduction of 85% in time consumption. The device designed to apply high impact forces on the head of the bored pile consists of the free fall self-propulsion hammer with special features to test piles with high load capacity and variable diameters, varying from 70cm to 250cm. The selection of the proper hammer size is essential for successful high-strain dynamic testing hammer weight, and drop height. The cushion details must be appropriately chosen so that hammer impact causes sufficient pile movement in order to mobilize the required soil resistance, and to assure that dynamic stresses in the shaft will not impair its structural integrity. The hammer apparatus must be constructed in a way to facilitate mobility around job sites and assure a uniform impact to the head of the pile (Hussein et al., 1996). The free fall hammer conceived and in use in Brazil consists of the following main components, shown in Figure 3 : the steel striker plate, the cylindrical - 2 pieces - steel cage (both of these components being referred to as the hammer), the set of round 20kN steel rams allowing modulation up to 200kN in maximum drop of 3rr4 and the hydraulic system to lift the ram by an automatic catcher, which activates and controls the hydraulic jacking equipment. The steel striker plate consists of a round 2.1m diameter by 100mm thick steel plate, with four lateral latches that adjust themselves according to necessities in relation to the diameter of the pile to be tested, and functions as a hammer support on the pile top and as a helmet.
- The hammer drop height should be approximately
8.5% of the pile length, with a minimum value of 2.0m; and - A plywood cushion thickness of t = L2 / 2 0 , where thickness ( t ) is expressed in and pile length ( L ) and pile diameter ( D ) are in meters, a minimum value of lOOmm is suggested, with an additional 15Omm being added when the pile length exceeds 30m. In this particular case, it was postulated that a ram of 120kN, with a maximum free fall height of 2 . 5 ~ and a plywood cushion thickness of 150mm by 1.l m diameter would be used to comply with the recommendations. The diameter of the plywood cushion should be about 80 to, at most, 90 percent of the shaft diameter in order to be able to centralize the blow. The thickness of the steel striker plate to be placed above the top of the pile should range from 50mm to 100its minimum diameter being equal to the diameter of the cushion the ram, ideally, is a simple drop weight which can be raised to a variable height. In order to facilitate, rams can be made in segments which can be assembled at the job site. The ram to be used needs to be “guided” to assure that the pile top and ram bottom are perfectly parallel during impact, in order to assure uniform contact stresses during impact. Between the steel striker plate and the top of the pile a 20mm thick steel wired rubber pad was installed with a 1.0m diameter in order to evenly distribute compressive stresses on the pile head, absorbing the dynamic impact, and preserving the structural integrity of the pile top above the electronic sensors, which measures data to be associated with the test.
4.3 Technical Criteria
To determine the hammer weight, drop height, and plywood cushion thickness to be used in this p&icular test, the bibliographic recommendations of Likins (1994) and Hussein et al. (1996) were followed:
Figure 4 - Details of the windows in the steel casing for the installation of the accelerometers and strain transducers
445
4.4 Case Method and thefield DLT
The test was done by a sequence of blows with variable drop heights as a dynamic load test with increasing energy. References can be seen by Aoki (1989), Niyama and Aoki (1991), Hussein et al. (1992), Aoki and Mello (1992), Mello and Aoki (1993), Beim and Aoki (1996), Mello and Paraiso (1998), among others. It’s important to note that the sequence of blows was done specifically on a concrete shaft inside in steel case. The DLT with increasing energy can be seen as a cyclic test similar to a static load test with increasing load steps where each loading cycle corresponds to an increasing energy impact. The test started with a drop height of 0.5m up to 2.5m with increments of 0.5m. The input data supplied to the PDA-PAK95 equipment for the beginning of the DLT were: e SP (pile specific weight): 24kN/m3 e AR (pile area on the sensors level): 11,309.8cm2 EM (pile dynamic modulus): 3 12tf7cm2 e EA/C (pile impedance): 1,007.31tfXm e WS (wave velocity): 350Ods 0 JC (Case Damping factor): 0.5 (medium value) At the same time precise millimetrical set measurements were collected from the shaft of the pile for each blow. Figure 5 presents the force and velocity x EA/C traces as function of time for each blow, obtained from the strain transducers and accelerometers processed by Case Method and reprocessed by Datpro Program. These traces exhibit good proportionality between force and velocity, and the peaks of force show an increase as the energy from the dynamic impact rises, which is perfectly consistent.
Figure 5 - Force and Velocity traces. developed internally in the steel casing and between the steel casing and clay soil, as following: - drop height; H FMX = the maximum compression force; EMX = the maximum energy transmitted past the transducer; %EMX = efficiency; SET = permanent SET; DMX = the maximum downward displacement at the transducer location; total mobilized resistance; toe resistance; lateral skin friction; lateral quake; toe quake; Damping factor; lateral damping factor; and toe damping factor. 5 CONCLUSIONS
4.5 Cupwup Analysis In the Capwap analysis the pile was initially modeled considering it’s uniform cross section. However, in the case of the pile tested, it was necessary to model impedance variations along the depth into the rock for improving the “match” between the curves of force and velocity measured and calculated. In this particular case, the analytic procedure was extremely necessary. Chart 6 summarizes the results of the Capwap analyses for each blow applied on the top of the pile. Chart 7 shows the typical results from the Capwap Analysis with reference to all five blows. It considers the values of skin friction, in the shaft space, inserted in the alluvium of the basalt,. between quotas 321.00 and 317.80, and toe resistance disregarding the portion of skin friction 446
Having the results presented by Capwap Analysis and Case Method as a reference it is concluded that: 0 The traces of force and velocity x EA/C presented in Figure 6, shows with clarity the increase of the load mobilized with the applied growing energy. 0 The liquid energy transferred to the pile grew with each applied blow, arriving at 29% in the fifth blow, a low value for the hammer in fi-ee fall without the interference of mechanical friction. The reason of the low transfer of energy is in consequence of the great loss of energy with the shock absorber system used. e It is verified that the maximum mobilized resistance of friction and point are manifested respectively when the fourth and fifth blows and confrmed with the obtained values of Qs and Qt according to Chart 6.
Chart 6 - Capwap analysis - general data.
-
BLOW
Rs SKIN up to 321.00
(kN)
Level (kN)
Rs-SKIN between level 321,OO and 3 17.80 (kN)
1840
1661.0
179.0
KS
'I (m)
R p Point Resistance (kN)
4025.0
RU Total ReRU sistance RU (Rs + Rp) (kN) (kN)
FMX (kN)
EMX (kN.m)
4030.0
9.0
% SET DMX EMX (mm) (mm)
~
1
0.50
4204.0
5865.0
15.0
0
2.72
2
1.00
4572
3148.5
1423.5
5101.0
6524.5
9673.0
5750.0
20.0
16.7
0
3.80
3
1.50
6621
361 1.5
3009.5
5180.0
8189.5
11,801.0
9020.0
37.0
20.6
0
4.85
4
2.00
6365
2938.0
3427.0
7050.0
10,477.0
13,415.0 12,460.0
61.0
25.4
0
6.21
5
2.50
4114
1387.0
2727.0
11.129.0
13,856.0
15,243.0 16,420.0
86.0
29.0
0
7.45
Chart 7 - Capwap analysis - socketed in rock. H (m)
RU (kN)
RP (kN)
1
0.50
4204.0
4025.0
2
1.00
6524.5
5101.0
BLOW
Ks
SMITH
Qs (mm)
(mm)
179.0
1.377
1423.5
2.280
(kN)
Qt
J CASE CAPWAP
Js (s/m)
J, ( s h )
1.030
0.640
1.754
0.87
1.750
1.ooo
1.165
0.25
3
1.50
8189.5
5180.0
3009.5
3.002
2.587
0.701
0.876
0.44
4
2.00
10,477.0
7050.0
3427.0
4.158
3.900
0.7 13
0.75 1
0.52
5
2.50
13,856.0
11,129.0
2727.0
4.000
4.682
1.136
0.426
0.41
It is observed that the measured value of zero set in all the blows leads to the consideration that the pile still presented an additional point load to be mobilized, however, the value of the load mobilized in the fifth blow indicated factor of safety larger than 3, in relation to the predicted working load, being opted by the paralyzation of the test. 8 It is considered that the ultimate load mobilized in the dynamic test can be interpreted in the following way: - Ultimate Bearing Capacity (5th blow) = 13,856kN - Ultimate Bearing Capacity, considering the friction resistance (4th blow) added to the point resistance th blow) = 14,556kN The test showed the saturation of the skin friction in the 4th blow and maximum point mobilized in the 5th blow, whose obtained values present a close approximation with those considered in the semirnpiric analyses in the project phase. The dynamic test in piles of high load capacity has become a reliable tool and has reduced costs and shortened evaluation deadlines for the quality of this type of deep foundation, as long as the hammer to be used presents the available potential energy to the load to be mobilized.
6 ACKNOWLEDGEMENTS The authors wish to thank the support given by the Company CBPO-Grupo Odebrecht,and Mello de
Azevedo contractor responsible for the construction of the bridge on the Uruguay River and, especially, to engineer Claudio Menin, project planner and responsible for the structural calculations of the bridge.
7 REFERENCES AASHTO T 298/93 - American Association of State Highway Officials Standard - Method of test for high-strain dynamic testing of piles. ABNT NBR 13208/94 - Associaggo Brasileira de Normas TCcnicas - Estacas Ensaio de Carregamento Dinilmico, in Portuguese. ASTM D4945/89 - American Society for Testing and Materials Standard - Test method for highstrain dynamic testing of piles. Antunes Sobrinho, J. et al. - 1999 - Criterios de projeto na construqgo da barragem da UHE Ita. XXIII Seminario Nacional de Grandes Barragens. In Portuguese, Belo Horizonte, Brazil. Aoki, N. - 1989 - A new dynamic load test concept. Proc. Discussion Session on Pile Drivability, XI1 ICSMFE, Rio de Janeiro. Aoki, N. & Mello, V. F. B. - 1992 - Dynamic loading test curves. Proc. Fourth International Conference on the Application of Stress-Wave Theory to Piles. Hauge, Holland. Balthaus, H. G. et al. - 1985 - Dynamic load test 447
German practice. Proc. 1lth Int. Cod. Soil Mechanics and Foundation Engineering, San Francisco, USA. Barton N., Lien, R. & Lunde, J. - 1974 - Engineering classification of rock masses for the design of tunnel support. Rock Mechanics, 6 (4), 189-236. Beim, J. W. & Aoki, N. - 1996 - Dynamic load test method with variable energy. Proc. Fifth International Conference on the Application of Stress-Wave Theory to Piles, Orlando, USA. Beim, J. W. and Paraiso, S. C.. - 1992 - Dynamic testing of enlarged base Franki piles. Proc. Fourth International Conference on the Application of Stress-Wave Theory to Piles, Hauge Netherlands. Bieniawski, Z. T. - 1984 - Rock mechanics design in mining and tunnelling. Balkema, Rotterdam. Broms, B. B. et al. - 1988 - Bored in residual soil and weathered rocks in Singapore - BAP I Ghent - Belgium. Canadian Foundation Engineering Manual - 1985. Capwap Users Manual - 1996 - Goble Rausche Likins and Associates, Inc. - Cleveland, Ohio, USA. de Mello, L. G. & Paraiso, S. C. - 1998 - Variable energy dynamic load test on a 1.0m diameter CFA pile. Proc. Belgium - BAP 111. de Mello, V. F. B. & Aoki, N. - 1993 - Updating realism on large diameter bored piles. Deep Foundations on Bored and Auger Piles, Ghent, Belgium. Hussein, M. et al. - 1992 - Dynamics of pile driving as a function of ram drop height. Proc. Fourth International Conference on the Application of Stress-Wave Theory to Piles. Hauge, Holland. Hussein, M. et al. - 1996 - Selection of a hammer for high strain dynamic testing of cast in place shafts. Proc. 5'" Int. Cod. on the Application of Stress Wave Theory to Piles, Florida, USA. Infanti, N. et al. - 1999 - Tensees residuais nas obras subtenheas da UHE Ita. XXIII Seminario Nacional de Grandes Barragens, Belo Horizonte, Brazil. Lima, M. J. C. P. A. - 1983 - Prospecqiio Geotecnica do Subsolo - Livros Tecnicos e Cientificos Editora S. A., Rio de Janeiro, Brazil. Niyama, S. & Aoki, N. - 1991 - Correlaqilo entre provas de carga diniimica e estatica no campo experimental da EPUSP/ABEF. In Portuguese, 2rd -Seminario de Engenharia de FundaGdes Especiais e Geotecnia, SEFE, Siio Paulo, Brazil. Rausche, F. & Seidel, J. P. - 1984 - Design and performance of dynamic tests of large diameter drilled shafts. 2"d Int. Conf. on the Application of Stress Wave Theory on Piles, Stockholm. Rausche, F. et al. - 1985 - Dynamic determination of pile capacity. Journal of Geotechnical Engineering, Vol. 111 - Nc 3 - ASCE: 367 - 383.
448
Rausche, F.- 1988 - High strain testing of drilled shafts. Seminar for dynamic testing of cast in situ piles for capacity and integrity - Boulder, CO. Seidel, J. P. et al. - 1996 - Dynamic testing of barretes for a cement silo project. Proc. Fifth International Conference on the Application of Stress-Wave Theory to Piles. Orlando, USA. Townsend, F. et al. - 1991 - Dynamic load testing of drilled shafts - Final Report, Univ. Florida, Gainsville, USA. Zhang, L. & Einstein, H. H. - 1998 - End bearing capacity of drilled shafts in rock. journal of geotechnical and geoenvironmental engineering, vol. 124.
7 SPT measurements and special field monitoring test
This Page Intentionally Left Blank
Application of StreSS-WaVe Theory to Piles, Niyama & Beim {eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Keynote lecture: Frequency characteristics of stress wave and penetration during SPT K. Fujita Department of Civil Engineering, Science University of Tokyo,Noda, Japan
ABSTRACT Recently wave equation analysis has been successfully applied to the SPT to determine the energy transfer ratio for adjusting the N-value, but should not limited to this purpose The SPT N-value has been applied to the design of foundations for more than SO years, however, the correlation between the Nvalue and the design parameters have been estimated crudely based on experience, so that new methods or constitutive equations should be established for determining the design parameters rationally. The frequency characteristics of the stress wave and the behaviour of the penetration of the rod system are introduced, because the relation between them could be keys to hrther studies 1 INTRODUCTION
As the ISSMFE was unable to standardize the method and equipment to be used for the SPT internationally, the SPT Working Party of the ISSMFE Technical Committee on Penetration Testing proposed the “International Reference Test Procedures on SPT”, at ISOPT-I, Orland in 1988. A method using stress wave equation analysis was introduced in Document 4- Calibration Methods for the adjustment of the N-values in the SPT to account for different magnitudes of driving energy, as a means of quality control for the SPT. The method was established by referring to studies made by Uto et a1 (1974), Schmertman et al. (1979) and others In the author’s opinion, the calibration methods are unable to account for a difference in the size of equipment, however, this could be corrected by calculations and the formula to be applied for the adjustment of N-values should be revised. Matsumoto et al. (1992) have applied the method of two-point strain measurements proposed by Lundberg & Henchoz (1977) to the measurement and analysis of a SPT experiment carried out in the field The displacement-time diagram, the transmitted energy-time diagram and stress-time diagram were verified, and the calculated data corresponded well with the measured data Dynamic measurements of the energy transfer ratio have been carried out in practice in USA and other countries with increasing frequency. Aboumatar et al. ( 1 996) have presented a typical dynamic SPT records including the measured force, adjusted velocity and displacement with time computed by
integrating the velocity, which was equal to the observed final set as shown in Figure 1 The wave equation soil constants, i e the quake, damping constant and mass for use i n the wave equation analysis, as well as the static soil resistance obtained from the calculated toe force and motion were determined reasonably well Abou-matar & Goble ( 1997) have contributed a valuable paper which broadly discusses the problems associated with the dynamic measurements and analyses concerned with the SPT They pointed out that changes in the cross section of the drill rod may cause substantial changes in the measured Nvalue and presented method for detecting and avoiding various problems based on onedimensional wave mechanics Fujita & Ohno (2000) introduced studies concerning the application of the stress wave theory to the SPT in Japan, and included a figure that indicates that the SPT hammer strikes the rod more than three times afier the first impact They suggested that analysis of the frequency characteristics of the measured stress wave would provide valuable information for the study of the SPT concerned The term = f i known as “wave propagation speed’, etc is applied in one-dimensional wave mechanics and assumes that the rod length is infinite or the rod area (diameter) is negligibly small The wave propagation speed in the anvil, rod connector or hammer is about 60 to 70% of c So, the analysis should be take into consideration the frequency of the wave
451
h
E E 4 c
5 E
y
-cac 6 d 2
-44-
-
\'elocity.(LA,c) (IrN) --Divplaceinent
P O I C(hN) ~
0 CL
I
I
I
I r)
I
I
-4
(mtn) I
I
40 Time (ms) Figure 1 Dl ii;uiuc SPT ineasureinent records - force, L elocih and displacement (aficr Abou-matar et a1 1096) 0
\
70
30
II
Final penetration
netrdtiori vs time l'enetratiori
.Maximum penetration
I 10
20
30
40
50
GO
'50 I0
Time ( ms)
Figurc 2. Strcss and pcuctration L'S. tiinc diagram. (Fujita 2000).
In this paper, the penetration or vertical movement of the rod, the frequency characteristics of the stress wave and the relation between the rod penetration and the frequency of the stress wave will be introduced, while the study is still proceeding. 2 PENETRATION VS. TIME DIAGRAMS Figure 2 is an example o f a diagram showing records of the penetration of the rod with time and the stress in the rod with time (stress wave) measured near the top of drill rod during one blow of the SPT. The figure shows that the hammer strikes the rod again (second impact) about 14ms after the hammer blow (first impact) and the penetration increases accordingly. The penetration at the time of the second impact is smaller than the maximum values 452
of the penetration which occurs between the first and second impacts due to the rebound of the rod and/or ground before the second impact. In this case, during the completion of the SPT there are at least two impacts of the hammer, and this evidence should be considered in the wave equation analysis. Figure 3 is a diagram summarizing the measured penetrations of the rod as a finction of time during the part of the SPTs at a depth of l l m . The maximum penetration and the final penetration for each blow are around 12 to 13mm and 4 to 7mm, respectively, and the rebounds of the rod and/or ground are between 6mm and 8mm. While the final penetrations (set) per blow which contribute to the N-value are scattered, the maximum penetration per blow is almost identical in all cases. Attention should be paid to this from the stress wave analysis point of view, because the energy
Figure 3 Penetration 1 s time diagram - at a depth of 1 l n i (Fujita 2000)
case shown in Figures 3-4, in ground which consists mostly of sand and gravel layers of alluvial and diluvium strata, and where the water table is 1 10m below the ground surface and the N-value is smaller than 50 Figure 4 shows the relation between the maximum penetration and the fhal penetration of the rod, (a) on a linear scale and (b) on a logarithmic scale measured at various depths during the SPT, without adjustment for the energy transfer ratio The energy transfer ratio varies from blow to blow in this case, mainly between 0 6 and 0 9, with an average of 74 Values of the maximum penetration smaller than 12mm were not observed during the SPT at this site, except for the case when the energy transfer ratio is small When the maximum penetration is about I 3 to 14mm,the final penetration will be less than about 1 Omni When the maximum penetration is larger than about 22mn1, the difference between the maximum penetration and the final penetration is less than about 21nm The maximum penetration occurs after the time (2L/c), in this case 3 FREQUENCY CHARACTERISTICS OF
STRESS WAVES The straidstress with time measured by strain gauges installed near the top of the drill rod as a function of time is called the “stress wave” in this paper, and an example was shown previously in Figure 2 In the wave equation analysis of the SPT, the hammer, guide rod, anvil, drill rod, coupler and sampler are generally assumed to be made of the same material and to have the same cross-section. except for the hammer The downward travelling wave arid upward travelling wave caused bv the hammer blow are propagate in the rod in the downward and upward directions, respectively, at the wave propagation
Figure 4 Relation betncen the iiia\iiiiuiii pcnetratioii and final penclration (Fiijita 2000)
transmitted to the rod is likely to correspond to the masinium penetration, not only to the final penetration So, dynamic measurements of the penetration (vertical displacement) should be made The strain or stress and the penetration of the rod were measured about 2 to 2 5m aboveground, near the top of the drill rod and below the anvil, in the 453
= . i i 111 ~ for the material, based on speed c- = ,/A one-dimensional wave dynamics However, the speed depends on the shape of the members, and decreases when two-dimensional wave dynamics are applied For example, the speed in the hammer or anvil would be about 70% of the one-dimensional wave propagation speed for the material The waves are partially reflected at the ends of the rod system and at the junctions between the different cross sections so that various waves are propagated in the rod following the hammer blow In order to examine the frequency characteristics of the stress waves, the method of Wavelet transformation was chosen, with which it is possible to obtain a thi-ee-dimensioiial diagram giving the frequency, time and power spectra Figure 5 shows two frequency vs power diagrams obtained by Wavelet transformation of stress waves, measured at (a) a depth of3m and (b) a depth of 17m The natural frequencies of the rod systems employed at (a) and (b) are 0 286kHz and J>
t
i I
I
I
0 1 ISkHz, respectively, assuming the wave propagation speed is = Jty. while the predominant frequencies shown in Figure 5 (a), (b) are calculated to be 0 19kHz and 0 074kHz, respectively, which are about 65% of the natural frequencies In Figure 5 , more than four prominent frequencies can be seen including the predominant frequency, these peaks are referenced to as A, B, C, D ..., from low to high frequency The natural frequency is located between the frequencies of the peaks B and C According to the time-power diagram (not included in this paper), the peak B occurs about the time of the second impact between the hammer and the rod As shown in Figure 6, the reciprocal of the time between the first and the second impacts corresponds well with the fiequency at the peak B 4 RELATION BETWEEN POWER AND PENET RATION Figure 7 indicates that the power at the peak A corresponds to the final penetration of a hammer blow of the SPT The frequency at the peak A is estimated to be around 15 to 20Hz, which is probably determined by the duration of the rod movement of 50 to 70 ms, as shown in Figure 3 The unit of power is ((N/mm2)s)2,where the .s is included in the term is the data sampling inteival, 5 s in this case The power is an index provided i n the Wavelet transformation, which is probably related to the penetration and the stress in the rod as well as the dynamic and static soil parameters, however, it has not yet been clearlv identified in the study ofthe SPT
I
Pztietration 71nm
001
01 1 Frecluel1cy (hHz)
10
(b) at a depth of 171n figiirc 5 Pciictration 1 s poucr spcclrum (Fqita 2000)
Figure 6 Reciprocal of the tiinc bctnceii 1 “ and 2’ld 1111pacts plottcd agaiiist fiequcnc! at peak B (Fujita 2000)
454
nicnsiircnieiils on SPT. llriic. .>" Iiii. 'oil/. n i l , l ~ ~ p l ~ c o i i o ~ oj'.sfrf~.~.~-lCnl~f~ ?%lV)i?' if, I'ilC.V, Orlniltl. 16.3- 175.
Figtire 7. Penetration plotted against power st pcak A. (Fujita 2000).
Abou-inarar. H.. X: G. Goblc. 1997. SPT dynamic analysis and tnc:isiircincnIs. .1. o/( iGK . I Fuiita. K. 2001). Strcss vavc i k o n iipplicilion 10 standard pcnctr;itioii ICSI in japan. /+iic. c;" iiii. ( ' o i i t iiiJ .i/~p/icii/ioi7 ~/.S/rc.v.v-~rmv /h<wcl'/oPilc,s, S m fmilo. (to be publislicd) ISSMFE Tccllnical Coininittcc on Pcnctratioii Testing-SPT \\.orking party. 1988. Standard Pcnclration Tcst (SPT). Irircrt~:itionnl rcfcrciicc tcst proccdurc. Iri Dc Kuilcr (cd). /'c~I1~~ir(J/IOIl 7?s/iilg I9HX. 0 r I ~ i I d: ( 1 ).7-26. Rotterdam : B;i Ikcnia. Lundbcrg. B.. X: A. Hcnchoz. 1977. Analysis o f el;islic w i ~ c s froin two-point slmin mcasurcmcnl. .I. of' I*~.qurii~rei~/. .\/<.ch. 17(6): 213-218. Matsumolo. T.. H. Sckiguchi. H. Yoshida. X: K. Kiln. 1992. SigiiBc;incc or two point strain nicasiirciiicnl in SPT. Siii1.s nntl I'iiiiI~tlnikin.s.32(2) : 6 7 4 2 Uto. K.. M. Fuyiiki. H. Kondo. M. Morilwra. (G H. Mmniniira. 197-1. Considcmtion of .\-\.alrie nnd strcngth of ground froiii a vicn point or w v c tlicon. I'roc.. t*hc.~il!\. of' Liiginwriivg, 7bkni [ hiiw-si/,c. (in Japanesc).
5 CONCLUDmG REMARKS
Since Terzaghi & Peck published their book "Soil Mechanics in Engineering Practice", in 1948, the SPT N-value has been widely applied to the practical design of foundations, because correlations are given between N-values and the relative density. consistency, angle of internal friction, bearing capacity of footings and earth pressure etc. although these were estimated crudely based on experience There are probleins involved in their proposals, for example, the etyect of the overburden pressure on the N-value is not accounted for, the energy transfer ratio is not considered, the SPT equipment and procedures are not standardized and there is no instrumentation The SPT N-value has been regarded as an index of the shearing strength of the soil, rather than the shearing strength of the soil, until now. However, dynamic measurements and the analysis by applying the one-dimensional wave equation to the SPT are now available, and the energy transfer ratio and the static resistance of the rod system can be estimated. The time has come to obtain the shearing strength of the soil directly from the SPT by incorporating dynamic measurements and the analyses and to determine the design paratneter rationally. The frequency characteristics of the stress wave and the behaviour of the penetration could be keys for further studies, as well as for establishing constitutive equations. REFERENCES Aboii-matar. H.. F. Rausche. G. Thendean. G. Linkins. Xt G. Goblc. 1996, Wave equation soil constants froin dynamic
455
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
The application of energy conservation Hamilton’s principle to the determination of energy efficiency in SPT tests N.Aoki & J.C.A.Cintra University of S8o Puulo, S8o Carlos, Brazil
ABSTRACT: The Hamilton’s principle is applied to the analysis of variations in the kinetic and potential energies and in the work done by the non-conservative forces, during the dynamic SPT tests. This paper presents a mechanically sound conceptual analysis of the energy transfer in the SPT hammer blow event different fiom that of the energy content of the first compression wave, as proposed in the ASTM D4633-86. 1 INTRODUCTION The equations of the Newtonian impact (de Mello 1971) and more recently the wave equation theory (Abou-matar & Goble 1997) have been traditionally applied to the analysis of energy lost in the SPT tests. The efficiency factor q of energy transfer is defined by the relationship between the measured energy of the first compression 1ongitudin:l wave and the normalized SPT potential energy T = 474 J. This theoretical efficiency would be proportional to the rod length (Schmertmann & Palacios 1979). The energy conservation Hamilton’s principle is more generalized than Newton’s law and can be regarded as the fundamental law of dynamics (Langhaar 1962). According to this principle, during the sampler penetration time interval, the actual applied kinetic energy ENTHRU is transformed into elastic deformation energy of the rod-sampler-soil system and into work done by the non conservative forces (includins soil damping) acting on the sampler. It must be observed that the ENTHRU is actually independent of the rod length although their evaluation is done by the measurement of deformation energy of the rod-sampler-soil system. Indeed the SPT blow efficiency will be related to the kinetic energy that reaches the sampler-soil system. In this instance it is showed that the actual efficiency is inversely proportional to the rod length. 2 HAMILTON‘S PRINCIPLE The variational expression of Hamilton‘s principle (Clough & Penzien 1975) is
where 6 = variation in the time interval (t2 - t,), T = total kinetic energy in the system, V = potential energy in the system, W,, = work done by non conservative forces (including damping). Any reference system of coordinates can be used. The first application of the Hamilton‘s principle for foundation engineering was the analysis of driven piles under increasing energy dynamic loading test (Aoki 1997). 3 EFFICIENCY IN THE SPT TEST 3.1 Traditional definition of SPT efficiency. The 623 N hammer weight impacts the anvil over the rod-sampler-soil system assemblage (Figure 1).
Figure 1. Standard Penetration Test system The impact height h = 0.76 m, the total length of rod is C and the sampler length is s. The anvil is located at the origin of the vertical reference 0-x axis and the top of the sampler is the section A. The kinetic energy of the first compression wave at the top of the of rod section is defined by T, =
457
ta
F(0, t ) . v(0, t ) dt
(2)
where To = maximum energy applied to the rodsampler-soil system (ENTHRU), F(0,t) and v(0,t) = force and velocity at the section 0. The maximum energy transferred to the top of sampler-soil system is expressed by
where T, = energy content of the first compression wave, F(0,t) and v(0,t) = force and velocity at the section 0, and t, = cutoff time at the arrival of the first tension pulse. Figure 2 presents the corresponding force-time diagram in the section x = 0 at the top of the rod.
(5)
where T,\= maximum energy transferred to the top of sampler-soil system, F(e,t) and v(t,t) = force and velocity in the section A, and t, = elapsed time for the compression wave to reach the section A. Figure 4 presents the energy wave in the x = 0 section at the top of rod-sampler-soil system and in the x = /1 section at the top of sampler-soil system. Figure 2. Force-time diagram at the top of the rod. kinetic energy
The efficiency factor q (Schmertmann & Palacios 1979, ASTM 1986 & Belincanta 1998) is
1'
-r
11
=
-+.loo T
(3)
time (t) 7
where T, = eqergy content of the first compression wave, and T = normalized SPT potential energy (474 J) Figure 3 presents a typical shape of the curve relating the efficiency factor and the rod length. 0
20
40
60
Figure 4. Kinetic energ! naves at thc top of the rod aiid at the top of the sampler sections
80 100
Figure 3 Efficient! factor 7 and rod length relationship (Sclimcrtiiiann & Palacios 1979) This figure suggests that the efficiency factor q tends towards zero in the case of short rod length, and towards a maximum for rod length greater than 15 m. This is an important feature that will be discussed thereafter under the light of Hamilton's principle.
The maximum value of kinetic energy at section 0 is T,, and the maximum value of kinetic energy at section A is T, Corresponding to the particular solutions of stress wave equation for these sections. Although the kinetic energy defined by either equation (2) and (4) is referenced to the same section x = 0 (top of rod-sampler-soil system), it is emphasized that the kinetic energy T, considers only the time integration correspondent to the first tension wave while T,,is the maximum applied kinetic energy Considering that the kinetic energy expressed by equation ( 5 ) is responsible for the deformations of sampler-soil system below the point A, it is proposed to redefine the SPT efficiency by 1
where q" = efficiency measured at section A, T,\ = maximum energy transferred to the sampler-soil system, and T"= normalized SPT potential energy (474 J)
3.2 ICedefinition of SP7' effrcrencj..
The maximum energy applied in the top of rodsampler-soil system is the well known ENTHRU T,,
=
f" F 0
(0. t ) v (0. t ) dt
(4)
3 3 Ayplicnfron of HciniiItoti',sp7-1~1c1~3le to /he SPT coiistnrit energy 4mnniic kondrrig test.
Let us consider the application of Hamilton's principle to the time interval (t2 - tl), where tl is the time when the compressive wave arrives the top of rodsampler-soil system, and t2 is the time when all the 458
kinetic energy is dissipated at the end of the hammer impact. Figure 5 presents the displacement - resistance curve in the section x = 0 at the top of rod-samplersoil system.
where pmax = maximum downwards displacement, pe = recovered elastic displacement, and pp= permanent sampler penetration. The recovered elastic displacement is correspondent to the elastic energy of deformation accumulated in the rod-sampler-soil system. Most of this energy is due to the energy accumulated in the rod length t and in the soil beneath the sampler. The permanent sampler penetration is directly proportional to the work done in the rod-sampler-soil system. More efficient is the hammer blow greater will the permanent sampler penetration. This observation shows that the definition of the SPT blow efficiency is best correlated to the work done and not to the applied kinetic energy. Now, let us consider the application of Hamilton's principle to another time interval (t2 - tc), where t, is the time when the compressive wave arrives the top of the sampler-soil system, and t2 is the time when all the kinetic energy is dissipated at the end of the hammer impact. At the moment of the maximum penetration of the sampler into the soil, the kinetic energy TA is transformed into deformation potential energy of the sampler-soil system only. After unloading, this deformation energy is transformed into work done in the permanent penetration of the sampler into the soil and in recovered elastic energy accumulated in the sampler-soil system. The difference (To - TA) in figure 4 can be considered roughly equal to the elastic energy accumulated in the rod length t. When the rod length tends towards zero the kinetic energy TA tends towards To and the efficiency measured at the top of the sampler-soil system becomes a maximum. Considering the expressions (2), ( 5 ) and (6) it can be concluded that in this case the kinetic energy T, becomes zero and the efficiency measured at the top of sampler is actually a maximum and not a minimum as indicated in figure 3. Figure 6 presents :he probable relationship between the efficiency q measured at the top of sampler-soil system and the rod length.
Figure 5. Resistance - displacement curve at the top of rod-sample-soil system. According to figure 5 the maximum kinetic energy reaching the rod-sampler-soil system at the time tl is
where T = total kinetic energy, and To = maximum energy applied to the rod-sampler-soil system (ENTHRU) . The maximum downwards displacement (pmax) corresponding to the maximum mobilized resistance (R), occurs at the point B of the curve, when the velocity becomes equal to zero and the kinetic energy T is transformed in potential deformation energy accumulated in the rod-sampler-soil system, not taking into account others energies losses,
where V = potential deformation energy accumulated in the rod-sampler-soil system, equal to the area OBCO in figure 5. At the end of hammer impact, correspondent to the point D of the curve, the potential deformation energy is transformed in work done by non conservative forces and in elastic recovery energy
0
20
40
60
80 100
0
(9)
5
where V = potential energy in the system, W,, = work done by non conservative forces including damping = area OBDO, and V, = elastic recovery energy = area BCDB. At the time t 2 after the unloading it is possible to record the displacement components due to the elastic and permanent deformations
efficiency (Yo)
Figure 6. Efficiency factor q' and rod length relationship
459
In the case of very short rod the blow efficiency wfi be independent of the rod length. For small rod Of and very low relength and’or in the sistant soil below the sampler, the applied kinetic ene r a To = TA Will be transformed into work done in the sampler-soil system, during the permanent sampler penetration, and pmax =
pp
(1 1)
In this case the elastic soil deformation energy can be neglected and the kinetic energy is almost all transformed into work done during the permanent penetration of the sampler into the soil, resulting
In the case of elastic soils the elastic recovered energy will be proportional and the work done in the permanent penetration inversely proportional to the compressibility of the sampler-soil system. Finally does the failure load of sampler soil system be always reached? Yes, for SPT in the range I < N 5 60, considering that the permanent penetration for one blow is greater than 10% of sampler diameter, to be in accordance with Terzaghi 1942 small pile rupture definition. For N>60 and in the cases where elastic recovered energy predominates N has no meaning as soil parameter.
4 CONCLUSION The SPT efficiency calculated by the expression (6) and the energy conservation Hamilton’s principle showed that: a) this efficiency is inversely proportional to the rod length and, b) although not proposed, the best definition of the SPT blow efficiency would be related to the work done in permanent penetration of the sampler and not to the applied kinetic energy. Finally it can be remembered that the energy transferred by the SPT hammer impact not always mobilizes the ultimate resistance of the soil.
REFERENCES Abou-matar, H. & G.G. Goble 1997. SPT dynamic analysis and measurements. J. Geotech. & Geoenviron. Engrg., ASCE, 123( 10), 92 1-928. Aoki, N. 1997. Evaluation of the ultimate bearing capacity of piles driven with increasing energy (in Portuguese). Doctorate thesis, University of Sao Puulo, Sao Carlos, Brazil. ASTM 1986. Standard test method for stress wave energy measurement for dynamic penetrometer testing systems; D4633-86, West Conshohocken, Pa., 775-778. Belincanta, A. 1998. Evaluation of intervening factors in the penetration resistance of the SPT (in Portuguese). Doctoi’ate thesis, Universify of SuoPuulo, Sao Carlos, Bruzil. Clough, R.W. & J. Penzien 1975. Dynamics of structures. New York: McGraw-Hill.
460
de Mello, V.F.B. 1971. The standard penetration test. Proc. PanameriCan c o n ! soil Mech. Found. Engrg.., 4,San Juan Puerto Rico, June 1971, 1: 1-87. Langhaar, H.L. 1962. EnergV methods in applied mechanics. New York: John wiley. Schmertmann, J.H. & A. Palacios 1979. Energy dynamics of SPT. J. Geotech. Engrg. Div., ASCE, 105(8), 909-926. Terzaghi, K. 1942. Discussion on pile driving formulas. Proc. ASCE, 68(2), 3 1 1-323.
Application of Stress-Wave Theoryto Piles, Njyama & Beim (eds)02000 Balkema,Rotterdam, ISBN 90 5809 150 3
Correlative study of Smith damping coefficient and SPT blow count Robert Y. Liang Civil Engineering Department, University ofAkroiz, Ohio, USA
CT: The accuracy and consequently the usehlness of wave equation analysis in pile driving problem are largely hinged upon accurate input of wave equation soil parameters that are representative of the soil conditions at the site. Since Smith model parameters have been used in most wave equation analysis techniques, emphasis is given to the development of algorithms to determine Smith model parameters from static ultimate soil resistance R, (including shaft resistance f, and toe resistance R,), quake Q and damping J. In this paper, a methodolozy for using SPT values to determine Smith model parameters is developed. The approach used herein to develop usefbl correlations consists of a number of consecutive steps, where the static soil resistance is first determined, soil quake Q evaluated for a variety of soil types, and finally, the damping coefficient J is determined by computer simulation of the standard penetration test with the help of GRLWEAP program. The results obtained indicate the usehlness of the adopted method when compared to existing cases anaiyzed by the CAPWAP. 1. DETERMINATION OF STATIC CAPACITY
2. DETERMINATION OF QUAKE Q FROM SPT
Liang (1999) described a method, and put down two sets of equations to estimate soil resistances from the standard penetration blow count number SPT-N for different pile types and different soils. In the present study, the adjustment factors m,q~ and ~1 are all assumed to be unity for both shaft and toe resistances. Thus, the shaft and toe resistances can be calculated as follows.
2.1 Measured Quake in Piles
a) Shaft Resistance (kPa)
"
110+.V
7I
Sand
J
Y
366.67,V f' =-
Clay ._
+3.5Ne
j
b) Toe Resistance ( H a ) Clay
Sand
Quake is defined as the maximum movement of the pile relative to the soil in order to hlly mobilize ultimate resistance to pile movement (or moving pile). Terzaghi and Peck (1967) indicated that the maximum shaft resistance would be developed in clays at a pile movement equal to about 1% of the pile diameter. Coyle and Reese (1966) indicated that the maximum shaft resistance would be developed in clays with approximately 6.3 mrn (0.25 in.) of pile movement. Reese and O'Neill (1 989) provided generalized results from numerous load tests. For cohesive soils, they indicated that full development of the shaft resistance occurs at a settlement of about 0.5% of the shaft diameter, while the toe resistance requires about 5% of the shaft diameter to be hlly mobilized. Cohesionless soils required settlements of about 1% of the diameter of the shaft for a hll development of shaft resistance, and in excess of 10% of the shaft diameter to hlly mobilize toe resistance.
Ho (1994) measured the t-z curve in compression bored piles in Sinsapore old all-jvium.He graphically
46 1
showed the mobilized curves at pile shaft and toe. He indicated that the shaft resistances increase with increasing applied loads, with no signs of leveling off in the mobilization curves. As for the pile toe, he indicated that the trend sharply increases in a more like curvilinear manner with settlement measured at pile toe. A summary of soil quake values as deduced from the SPT-N, and pile tests is presented in Table 1, below. Table 1 Summary of Measured Quake Values as Obtained from the SPT-N. and Pile Load Tests * Author
model parameters were also obtained via CAPWAP match procedure. The database can firther be broken down into two sets. One set of data contains 208 pile cases (120 individual piles) that have both dynamic testing and subsequent static load test results. The second set of data contains a total of 403 pile cases with only dynamic pile testing results. To investigate the significance of soil type and time of dynamic testing (i.e., end of driving- EOD, and Beginning Of Restrike- BOR), a statistical analysis of the Smith model parameters was conducted on the Paikowsky’s database, particularly the quake and the damping factor. Table 2 lists the means and standard deviations of the Smith model parameters for different soil types.
(1967)
Table 2 Statistical Analysis of the Smith Model Parameters Based
(1966) (1989). cohesive soils. (1989), Soils. Ho (1994)
Sand
iU1 Soils
0.-30
0.30
0.15
0.15
0.15
0.53
0.64
0.61
036
0.36
0.47
0.60
Clay
Mean
Non-coh. 0.25 (0.040.46)
Mean
0.30
030
0.70
2.2 GRL’s Quake Values from CAPWAP Analysis One of the most frequently used methods of determining the Smith model parameters is the backcalculation of those parameters via the CAPWAPprogram (Goble, et al, 1988). Essentially, the measured velocity-time records from PDA constitute the input file to calculate the force-time histories. By adjusting Smith model parameters iteratively, one can obtain a good match between the calculated and measured force-time histories, from which the Smith model parameters can be identified. GRL performed the CAPWAP analysis of 82 piles in the FHWA report (1988). Based on their work and data, an average soil quake value of 0.10-in can be suggested at the pile shaft for all types of soils with a tolerance ranging from 15 to 27%. As for the quake at the pile toe, sandy soils were found to have the largest values and the least variation, that’s to say using an average value of 0.39-in for piles in sand will produce an error of no more than 5% of the actual value. For orher types of soils (clays, silts, gravel with sand, rock. or a mixture, a value 0.22 in is recommended with a tolerance of no more than 25% of the real pile toe quake.
s/m
I
Std Dcv
0.60
0.40
0.53
Mean
0.77
0.57
0.63
Std Dev
0 77
0.67
0.70
Table 3 Statistical Analysis of Smith Model Parameters based on Time of Testing (EOD and BOR).
2.3 Paikowsky’s Quake Value from CAPW-ZP Method
462
j
Table 3 shows the statistical analysis results based on the time of dynamic testing, regardless of soil type Finally, Table 4 shows the combined statistical analysis results, considering both soil type and time of testing. From these analysis results, one may
0.50
A data base consisting of 61 1 piles has been collected by Paikowsky, et aL(1994). In their report, the smith
’
0.50
J,. s/m
0.43 0.13
0.53
observe that the quake does not vary sigdicantly with the soil type and the time of dynamic testing. However, the damping factor seems to be affected mostly by the time of testing, rather than the soil type. Table 4 Statistical Summary for Smith Model ’arame :rs at Different Testink times. Par. summary E.0.D B.0.R Soil
Sand J,, slm
Jp. s/m
Qs,
Qp,
cm
cm
Clay
J,, slm .
Mean
0.53
0.67
Std. Dev.
0.53
0.53
Mean
0.43
0.80
Std. Dev.
0.13
0.90
Me2
0 28
0.30
Std. Dev.
0.13
0.15
Mean
0.61
0.53
Std. Dev.
0.48
0.20
Mean
0.43
0.73
Std. Dev.
0.40
0.53
Figure 1 Shaft quake in SPT test.
2.4 Quake Values from SPT Test Figure 2 Toe quake in SPT test
GRL, in the FHWA report proposed a method to determine the Smith model parameters from the modified SPT test in a manner similar to the PDA testing. The modified SPT procedure requires the measurement of the dynamic force and the velocity near the top of the SPT drill rod during sampler driving, as well as the N-value, and for every test depth the recorded force and velocity signals at the end of driving of the sampler are analyzed and processed via GRLWEAP program, thus the Smith model parameters, €&, Q, and J were back calculated through an integrated iterative parametric optimization process. Two types of analysis were performed: the first utilized the ultimate resistance k,and quake, Q, obtained from the static test while solving for the damping factor J in the CAPWAP analysis, where the ultimate resistance from static pile test was determined from the load-displacement curve based on Davisson’s failure criterion. While the second analysis “dynamic” analysis assumes that all three soil parameters &, Q, J) were unknown, and
they were solved for by matching the calculated force curve. The results of these two analyses are graphically presented in Figures 1 and 2, showing the variation of soil quake at pile shaft and toe, respectively, as a function of pile installation depth, and soil type. Based on Figures 1 and 2, and for further analysis and calculation of Smith type damping coefficient in this paper, soil quake values at the pile shaR and toe will be taken as the averages of the two figures, meaning that the quake values at the shaft and toe will be equal to 0.5 and 0125 in, respectively.
463
3 1 Determining Damping S using GWWErkP Program
The wave equation analysis program- G was used to simulate the SPT results To model the driving process of SPT, the progiam requires data input that feeds and describes the hammer characreristic, the rod and sampler and the soit. The SPT drill rod is a hollow 0.7-in outer diameter tube with a wall thickness of 3/8 in The sampler must have the same imer diameter as that of the rod, a greater outer diameter of 2 0-in, and 2 25eieas, the input data describing the h a m e r properties are summarized in Table 5 The elastic modulus of the drilling rod and the split-spoon sampler is 30,000 ksi (207 -GPa), and the specific weight is 492 pcf (77 7 W/m') The shaft and toe resistances can be cdculated from SPT-N values using Equations 1 and 2, respectively Based on the dimensions of the SPT sampler, the total shaft and toe resistances can be calculated, from which the eotal soil resistance and the ratio of the shaft resistance to the total soil resistance C ~ be R obtained This ratio is the required soil resistance distribution in the Program As indicated earlier in this paper, the soil quake values will be deduced Erorn Figures 1 and 2 that assign 0 050 and 0 125 for soil quake at the pile shaft and pile toe, respectively
P. Project No Name 1 LUC75081
412in
I 175-1128
$12 in
1 6
Pile
Pene.. Date ft 48.56 29-May-96
21-Jan-97
29
Location Toledo
I
Toiedo
8
~
11
175-8.91 1754.61
!I;i;
(612 in ~
$22in (612 in
I75982 I750891
18 ~
30-Jan-97
i ;24-F;-97 ~
52 63
2 I -khy-97 28-X r-98 20-May-97 3-Jun-97
13
LIC1308- 416 in 2.1 17-Jul-97 80 13 GRE-35- W12r53 27. 30 05-.4ug-97 2080
~
,
Toledo Toledo Toledo Toledo Toledo Toledo
Colurnbus Greene
~
I
Stroke
I
30 in
I
762m
1
22
The evaluation of the damping coefficient is achieved through a number of consequent iterations in a parametric optimization process, rasing the developed software. The static resistance, soil quake, along with an initial guess for damping coefffcient constitutes the input dzta to the program, so that the standard penetration number SPT-Nis evaluated for that set of data. This process will be repeated until the obtained SPT-N matches the actual- recorded number 3.2 Results and Discussion
Du~ingthe course of this paper project, the research team has perfomed series of dynamlJc pile tests at 23 ODOT piling project sites. From those field tests, a total of 34 piles- listed in Table 6, have been tested
464
I
7 175-8.39 $12 in 51 28-Jm-97 Toledo -~~~~~
POR-59578 23 I\/IAH8317
Portage
$12 in
33
27-Xpr-98
$11 in
41
17-Jw-98 YoWgsto\Vn
using the Pile Driving halyzer (PDA). Liang (1999) provides a detailed description for those piles, and their records as obtained from the PDA. Tables 7 and 8, report the calculated damping factor 3 together with the C M tested piles, for 60%, and 70% Note that column (I) in Table 8 lists the average pile shaft damping factor 3 for project site as obtained by G will have the same value for all pile shaft segments when operating in the automated CAF" Column (2) lists the shaft-damping factor J calculated via the method introduced in this paper. Usually, the SPT-N value varies with depth, therefore several
esuIts ofDamuinn J Ec = 0.7). Pmj NO.
(0)
I
I I
1 2 3 4 5 6 7 8 9 10 11
1
12 13 14 15
Toe 3 ( s/ft ) Shaft 3 (dft) CMWAP 1Predxcted CAPWM Predicted (3) (4) (5)=(4)xO.J2 (1) (2) 0.113 1.672" 0.268 0.104 0.161 0.095 0.162 0.206 0.175 0.225 I 0.07 I 0.144 I 0.036 10.147 I 0.062 0.152 0.251 0.354 10.502 0.211 0.497 0.209 0.302 0.04 0.22 0.047 0.501 0.210 0.05 0.075 I 0.105 { 0.145 0.227 10.331 1 0.139 I 0.117 I 0.239 I 0.231 I 0.479 I 0.201 0.12 0.155 0.255 10.295 0.124 0.24 0.101 0.129 0.136 0.177 1.278* 0.319 0.134 0.2 0.032 1 0.513 0.185 I 0.059 10.202 I 0.085 I 6.954* I 0.202 I 0.024 10.096 I 0.040 0.101 0.218 0.06 0.28 0.118 0.165 0.095 0.061 0.233 0.098
1
I
I
I
1
1
I I
1
I
I
1
I
1
I
1
3 4 5 6 7 8 9
I
I
1 I
0.07 0.152 0.22 0.05 0.105 0.117 0.12
0.187 0.148 0.307 0.243 I 0.37 1 0.292 10.101 I 0.080 10.1841 0.145 0.297 0.235 0.197 0.156
I I
I
0.036 0.194 0.068 0.354 0.617 0.216 0.04 10.615 I 0.215 I 0.047 10.613 I 0.215 I 0.227 10.397 0.1391 0.231 0.593 0.208 0.255 0.358 0.125
I
I
1
I
I
1
0.239 1 0.134 1 0.133 10.148 I 0.37 I 0.293 I 0.077 I 0.36 I 0.25 19 0.524 0.123 0.027 20 0.116 0.189 0.018 0.212 0.078 0.224 0.251 0.182 21 0.232 0.06 0.141 0.03 22 23 I 0.305 I 0.151 I 0.036 10.318 I A ~ ~0.177 I I 0.181 I 0.111 l0.288 I * Not included in the average value. 17 18
I
1 1
0.062 0.151 0.105 0.089 0.076 0.097 0.134 0.121
I
I
0.123 0.27 0.213 0.052 0.384 0.134 0.239 0.179 0.141 0.133 0.201 0.070 0.37 0.364 0.288 0.077 0.442 0.155 0.524 0.16 0.126 0.027 0.313 0.110 1201 0.116 (0.2471 0.195 1 0.018 10.275 1 0.096 1 21 0.078 0.274 0.216 0.251 0.233 0.082 0.06 0.18 0.142 0.03 0.293 0.103 22 23 0.305 0.188 0.149 0.036 0.396 0.139 0.111 0.358 0.120 AV 0.177 0.227 0.180 * Not included in the average value.
values for shaft damping factor J along the pile shaft can be obtained. The predicted damping values in Table 8 and in the subsequent Tables are all lengthweighted average values along the pile shaft. It can be seen that there is a difference in the shaft damping J between the C N W N results and the predicted values. Nevertheless, their average values are very close. As for project 13, the shaft damping J from WAd) is too big, so the shaft damping factor J of s project is not included in the average v Columns (3) and (4) in Table 8 are the Clap and the predicted values of the toe damping fa ly. Similar to the shaft damping, the values for toe damping shown in column (3) are the average values. As for the toe damping in column (4), if there is more than one pile with different pile toe elevations, then average value is used. Similar to the shaft damping, the values for toe damping factor J from CAPWAP for projects 1 and 11 are too, therefore, they were not included in the calculation of the average values. It can be seen that the difference between the CBBWAP results and the predicted values is rather significant. The author recornmends that the calculated toe damping factor J being modified by a reduction factor of 0.42. The modified toe-damping factor J is shown in column
16 17 18 19
( 5 ) . In this way, the predicted average value (0.121) is close to the C average value (0.1 11). The comparison between C N W M and predicted values are shown in Figures 3 and 4 for shaft and toe damping, respectively. Figure 5 shows the sirmlar comparison for the modified toe damping. Figure 6 presents the correlation between the predicted damping factor J and §PT N values, where the damping values are shown for all SPT N values from all projects. The damping values were too high, so author recommends that the damping J be modified with a modification factor of 0.75. Then the correlations between the predicted damping J and the SPT-W values are those shown in, in which, two approximate fitting curves for clay and sand are plotted. The equations for these two approximate fitting curves are as follows. N N
Clay
0 ,3375 N 62 . 5 + N
Sand
0 .637
J =
465
62 . 5
+
3
F
Efficiency
0'5 I A v e . J 0.4
t
=
= 0.6
0.181
1 //'
i 0.3
L
i..::-,-,
Average J = 0 177
00 00
,
01
I
,
,
I
I
I
N
,
05
02 03 04 CAPWAP shaft J (sift)
Figure 3 CAPWAP Vs Predicted Shaft Damping.
0.6 r-
1
DO
-
1 t
-
/
L
,' Efficiency =0 6
0.5 j-
value
Figure 6 Correlation between modified damping and N values ( E c = 0.6)
06
05
AAer modification (0 79)
F
Efficiencv
,,,\
,'
0.7
,
a-
/
,
.
I
Avenge J 00
0 1
00 r
0 1 [-*
=
0 177
II
__-I____i i_L-I__J
0 2
02
0 4 C.\P\VXP \ h a f t J (sift)
Figure 7 CAPWAP
0 5
06
Vs Predicted Shaft Damping
! '
0 -0 - L 00
02
01
03
04
OS
00
CAPWAPtoe J (sift)
Figure 4 CAPWAP Vs Coefficient
Damping
Oh
1 0.6 4 z"- . < 14
O S L
c-i
04;
5
~
AAer modification (0.42)
? . 0.3
{ &
=
i
s
/'
,'
1
/
07
/
/
,
i
/
/ / /
,/
8
j 0.2
3
Efficiency = 0.6 Ave. J = 0.121
0.4 7 - 1
3
Efficiency
i Average J = 0 358
0.5
t
F-
.=
h
0.1
/'
Mod factor= 0 35
Toe
Piedicted
4
,/'
4-
-
/' /'
.
on
0 1
I)?
n3
0.1
CAPWAP toe J
ns
06
n-
(sift)
Figure 8 CAPWAP Vs Predicted Toe Dampmg. 00
A ^ _ _ _ -
00
01
-&-I-
_I___
02 03 04 05 C.\PW,AP Toe-J (sift)
__I_ A _.
06
07
is 0.6. For a hammer efficiency of 0.7, the correlations between the damping J and the SPT N values are presented in Figure 7. It can be seen that the damping J for Ec = 0.7 is larger than the damping J for Ec = 0.6. This can be explained easily by the fact that the static soil resistance correlations remain
Figure 5 CAPWAP V s Predicted Toc D a m p m i n g Coefficients ( After modification 1
The correlation results discussed above are based on the assumption that the hammer efficiency of SPT 466
the same regardless of the hammer efficiency. As a result, if the hammer efficiency is high, the increased energy is resisted by the dynamic soil resistance indicated by an increased damping J. Table 7 along with Figures 7 and 8 provide the results and correlation between predicted damping for h may be seen that the predicted damping is higher damping. As discussed before, a modification factor of 0.6 is recommended if one uses that Ec = 0.7.
COWCLUSIONS Based upon the field test results, data analysis, and literature review, the author may draw the following conclusions: The equations relating pile static capacity, and damping coefficient- both at pile shaft and toe, are reasonably accepted from the engineering background. Although there were some deviations from the values obtained from the case method, the average values were close, indicating the validity of the approach. The bigger differences in the damping coefficient may be attributed to the fact that, the predicted damping values will include both the errors in the static resistance, and the damping coefficientcommutative As a result, the use of SPT blow count in the prediction of static resistance, and damping coefficient may be considered to be satisfactorily powerful. The SPT provides a good predictive tool for soil quake, only at pile toe.
1. 2. 3
3.
4.
5.
4.
. m., and Reese, L. C.(1966), “Load Transfer for Axially Loaded Piles in Clay”, JSMFD, Asce, Vol. 92, No. S W , pp. 1-26. FHWA (1988). “Manual on Design and Construction of Driven Pile Foundation”, FHNA-DP-66- 1. Goble Raudche and Associates, (1988), CAPWAP Manual, Cleveland, OH. Liang, R. Y. (1999), “Development and Implementation of New Driven Pile Technology”, Final Report submitted to the ODQT, May, 1999, Report no. FHWMQh991008. Paikowsky, S. g., Regan, J. E., and McDonnell, J. J. (1994), “A Simplified Field Method for Capacity Evaluation of Driven Piles”, Publication no. FMWA-IQD-94-042. Reese, L. C., and O’Neill, M. W. (1989), “New Design Method for Drilled Shafts from
467
Common Soil and Rock Tests”, Proc. Of the Congress on Foundation Engineering: Current Principles and Practices, Evanston illinois, June, 1989, Vol. 2, 1026-10389. 7. Terzaghi, K. and Peck, R. B. (1967). “Soil Mechanics in Engineering Practice”, John Wiley and Sons, Inc., New York.
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
tress wave theory application to standar
enetration test i
K. Fujita Deparfmerit o j Civil Engineering,Science UrziversiQoJ’Tokyo,No&, Japan
M.Ohn0 HuzacIpm Technical Reseur-chInstitute, Tsukuba,Japan
ABSTRACT: Studies related to the application of stress wave theory to Standard Penetration Test(SPT) in Japan are introduced together with a report on the discussions we have had on the dynamics and mechanism of the SPT as well as aspects of the issues to be studied. Uto et al. were the first to apply the stress wave theory to the SPT in 1970’s. Matsumoto et al. have applied the method of two-point strain measurements to the SPT, and presented the calculated stresses, penetration and transmitted energy as firnetions of time in diagrams which corresponded reasonably with measured data. Fujita et al. have proposed a fundamental equation for the equilibrium between the driving and resisting energies considering the energy transfer ratio, an adjustment method for N-values, and Fujita’s diagram summarizing the SPT results, etc.. 1 INTRODUCTION
2 STUDIES BY UTO ET AL.
The SPT Working Party of the ISSMFE Technical Committee on Penetration Testing presented the “International Reference Test Procedures on SPT”, at ISOPT-1, Qrland in 1988. In Document 4Calibration Methods, a method using stress wave equation analysis was introduced for the adjustment of the N-values in the SPT to account for different magnitude of driving energy, as a means of quality control for the SPT. Many papers related to the application of stress wave theory to the SPT have been contributed and the measurements of the energy transfer ratio have increasingly been made in actual practice in the field in USA and other countries. However, few high frequency measurements of the movement or penetration per blow of the rod system have been published or reported in connection with the stress wave equation analysis of the SPT. The application of the stress wave equation analysis to the SPT in Japan started after the study of Uto et al. in the 1970‘s. Most of the studies carried out include high quality measurements of the penetration during the SPT in order to establish the mechanics or dynamics of the SPT which are not only for the purpose of quality assurance of N-values but also for attempting to find the shearing strength ox characteristics of the soil. This paper will introduce a state-of-the art report on studies concerned with the application of stress wave equation analysis to the SPT or related matters which have been done in Japan. These are mostly written in Japanese and include unpublished data.
For applying wave equation analysis to the SPT, LJto et al.( 1974) proposed a model using the generalized solution of St. Venant assuming the reflection coefficient to be the boundary condition at the bottom of the rod and that the rod is supported on rigid ground and struck by a hammer which is considered to be a rigid body. A series of experiments were carried out using an oscilloscope and strain gauges etc. for monitoring the stress waves and a high-speed movie camera for recording the movement of the top of the rod. The calculated and measured maximum compressive stresses in the rod were 154 M N m 2 and about 84 MN/m’, respectively.
469
3 STUDIES BY MTSUMOTO ET AL.
Matsumoto et al. (1992) have applied the method of two-point strain measurements proposed by Lundberg & Henchoz (1977) to a series of instrumented SPTs that they carried out. The upper gauge was located on the rod at a point of 225 rnrn below the anvil which had a height of 80 mm, and the distance between the upper and lower gauges was 600 mm. The total length between the top of anvil and the bottom of the sampler was 3.39mYthe hammer was operated by means of the rope and cathead method and the energy transfer ratio was found to be between 0.6 and 0.9. The SPT was carried out based on the Japanese Industry Standard (JIS) from a depth of 1.32m to 1.62111 and the number of blows was 12
(N-12) through clay and sand layers. The monitored stress in the rod was recorded at a sampling rate of 5 ,U s for the wave equation analysis, employing a DC preamplifier with a capability of 250 kJ3z and an A/Dconverter. The displacement of a target on the rod was measured by an electrical device with a sensitivity of 0.01 111111. Matsumoto et al. (1992) have proposed a procedure for handling errors in the measured strains such as noise, using two correlation factors for each of the upper and lower strain gauges in order to find the true strains. According to case studies, if the value of one correlation factor is taken to be unity, the other correlation factor would be in the range between 0.985 and 1.015. Verifications have been made for the displacement - time curves, the stress waveforms and the energy transmitted to the rod with time as shown in Figures 1-3. The calculated data compare well with the measured data so the twopoint strain measurements and the procedure for handling errors in the measured strains can be regarded as very useful tools for the wave equation analysis.
Time(nis)
Calculated, o Measured Figure 1 . Calculated and measured displacement-time diagram at a blow of SPT(after Matsumoto et al.1992).
E
500
0
5
10 15 20 Ti me(ms) --Calculated, - - - - Measured Figure 2.Calculated and measured transmitted energy-time diagram 31 ;I blow of SPT(after Matsumoto et al.1992).
4 SET-UP FOR THE LABORATORY SPT STUDY BY FUJITA ET AL. For the laboratory experiments results to the SPT, Fujita et al. have used the arrangement of testing equipment, instruments etc. shown in Figure 4 and Table1 since 1993. The SPT equipment conforms to the JIS (Japanese Industrial Standard) specifications
- 1001
0
7
5
10 Time(ms) (a) Calculated
10
15
20
1s
20
'I? me( ms) (12) Measured Figure 3. Calculated and measured stress-time diagrams at a blow of SPT (after Matsumoto et al. 1992).
Fi ure 4. Set-up for laboratory SPT. (&er Nagasaki et al. 1996)
470
Table 1. Specification of equipment and instruments. Hammer -w:63.5kg, OD:19.5cm, ID:S.Ocm, h:33cm Anvil -dia:8Scm, height6.5cm Rod -OD:40.5mm, A:5.3cm: W:4.68kg/m -OD:Slmm, ID:35mm, length:685mm Sampler Steel tank -dia:59.4cm, depth:105cm Pressure hag -SO, 100, 150, 200, 300, SOOkPa by air Soils -Toyoura sand etc. Strain gauge -Kyowa, semi-conductor, 1=21nm, GF:-98, 120 d? Amplifier -Kyowa, DC, 200kHz A/D convertor-Sun system supply, ADH-12A Displacement sensor --Izumi, Ldzer type, max.100mm, 40mm at l k H z
and a heavy steel tank, 0.60m in diameter by 1.05m high was built to control the variable soil conditions for the laboratory tests. A pressurized air bag is installed a load representing the overburden pressure between the loading plate fitted to the top of the tank and the surface of the test soil. The length of the guide rod and the distance between the surface of the anvil impacted by the hammer and the bottom of the sampler are 1.50m and 1.80m, respectively, so that the overall length is 3.30m. As the method of two-point strain measurements (TPSM) proposed by Lundberg and Honchez (1977) is employed for the measurements and analysis, the upper and lower sensors are installed on the rod 40cm and 60cm below the top of the anvil, respectively, for measuring the strains and velocity of the stress waves in the rod. Each sensor consists of four semi-conductor strain gauges attached around the circumference of the rod, at intervals of 90’. A laser type displacement sensor, capable of monitoring a 1 kHz wave with an amplitude of 40mm is provided for measuring the movement of each blow of the sampler at a certain point along the connecting rod. Three high frequency DC amplifiers and an A/D converter are employed for measuring the strain and movement of the rod and the monitored signals are digitized and recorded on a floppy disk on a computer at a sampling rate of 5 U , s. The laboratory test ground provided in the tank is made uniform aiming for a relative density either dry or saturated condition, and Toyoura sand is generally used. Overburden pressures of 50, 100,200, 300 or 5OOkPa can be applied by using the pressurized air bag. The tests were carried out according to the provisions given in the SPT procedure of the JIS, in which the hammer fall height is specified to be 750mm (it will be changed to 760mm in 2000), so that the kinetic energy is 467J, not 473J. The hammer is hung by a wire on its vertical axis and is released by cutting the wire, so that when the rod system is held in a proper perpendicular position, the hammer can be dropped in a free fall, i.e. no
Figure 5. Diagrams summarizing a laboratory SPT. (ai‘ter Fujila, 1997).
fiiction effects due to the guide rod. Measurements were made at every hammer blow after the excess pore water pressure had been dissipated in the case of the tests using saturated sands. Analyses using Matsumoto’s program and digitized data are carried out to obtain a stress (strain) time diagram, a penetration - time diagram, a velocity of rod - time diagram, a transmitted energy -time diagram and an energy transfer ratio (e). One of the results of the test and analysis is shown in the set of diagrams in Figure 5 , which shows that a) the maximum compressive stress induced in the rod by a hammer blow is about 200 MN/m2, b) the maximum and residual penetrations are about 3.6mm and 3.4mm, respectively, and c) the maximum velocity of the rod movement is about 9 d s . Figure 5 also indicates that the hammer strikes the rod more than three times after its first impacts. 5 FUNDAMENTAL EQUILIBRIUM EQUATION AND FUJITA’S DIAGRAM The following equation can be used to express the equilibrium between the kinetic energy given by the SPT hammer impact and the resistance due to work done by the sampler, if the kinetic energy is trans47 1
rnitted perfectly to the connecting rod system or sampler.
wh = R,.S where W = weight of the hammer (622."); h = free fall height of the hammer ( 7 6 0 m ) ; R, = dynamic resistance of the sampler; S = penetration per blow (mm); and wh = kinetic energy (473.35). In practice, the energy transmitted to the connecting rod system is smaller than the kinetic energy of the hammer and the energy transfer ratio e is defined as,
Figure 6. Fujita's diagram for representing penctration,energy transfer ratio and dynamic resistance at each blow o f SPT. (after Fujita 1997).
where e = energy transfer ratio; E, = energy transmitted to the connecting rod system or sampler. Fujita (I 997) has proposed a fundamental equilibrium equation (3) applicable to each hammer blow of the SPT as follows.
In equation ( 3 ) , eWh is calculated by the wave equation analysis of the measured stress with time at the two points on the rod, and &5' is the measured penetration per blow, then &Rd corresponds to the dynamic resistance of the sampler. The reason why &,S and &Rd are used in equation (3) will be discussed later. Fujita (1997) has proposed a diagram with orthogonal axes as shown in Figure 6 for representing equation (3), the x and y axes, using the same logarithmic scale are the penetration per blow S and the dynamic resistance R,, respectively, and the energy transfer ratio e are given by lines each with a slope of 45'. Figure 6 is the so called Fujita's diagram and can summarized the data obtained at each blow of the SPT and is useful for the interpretation and quality assessment of the SPT and for the quality assurance of the N-value. Figure 7 shows an example of the results of the SPT laboratory experiments which were carried out on dry Toyoura sand in a test tank prepared with an uniform relative density of about 40%, 60% or 80%. The measured penetration, calculated dynamic resistance of the sampler and the calculated ratio of the energy transferred to the rod system or sampler at each hammer blow can be plotted as one point on the coordinates of the diagram. Figure 7 illustrates that all the data lie along a line for which e is around e=0.8 or e=80%, and the plots for different relative densities are found to be located in their own groups. In the case of the relative density of 40%, for example, all the plots belong to a group corresponding to larger penetrations and
Figure 7. Penetration, energy transfer ratio and dynamic resistaricc at each blow of SPT in terms of rclativc density (Dr). (after Fujita 1997).
smaller dynamic resistances. The range of each group depends on the penetration depth of the sampler which varies from 150 to 450mrn during the SPT. So, the penetrations of the sampler per blow are larger at shallow depths and become smaller when the depth becomes larger, while the dynamics resistances are correspondingly smaller and larger.
6 ADJUSTMENT OF THE N-VALUE USING THE SQUARE ROOT OF THE ENERGY TRANSFER RATIO According to the International Reference Test Procedures, it is recommended that the N-value should be adjusted to the N-value corresponding to the energy transfer ratio of 0.6 or 60% in the reference procedures, if required, assuming the N-value to be 472
inversely proportional to the energy transfer ratio
(4.
Figure 8 given by Nagasaki et al. (1998) shows three sets of measured data with different energy transfer ratio being plotted on a Fujita’s diagram. Each was obtained by a blow of the SPT hammer at the same depth in each test tank containing dry Toyoura sand prepared with the same uniform relative density. The different energy ratios were produced intentionally by changing the free fall height of the hammer. The three plots in Figure 8 are located nearly on a straight line which is orthogonal to the lines of the energy transfer ratio. This means that the magnitude of the penetration is proportional to the square root of the energy transfer ratio e, not proportional to e. As the N-value is defined as the number of blows necessary to achieve a penetration of 300mm, and the N-value has to be inversely proportional to the square root of the energy transfer ratio, i.e. &, the adjustment of the N-value should be made accordingly. Table 2 indicates the comparison of the N-values which are adjusted in terms of the energy transfer
ratios given by Kovacs et al. and others, by the method of the International Reference Test Procedures assuming the N-value is inversely proportional to the energy transfer ratio and the method proposed by Fujita assuming the N-value is inversely proportional to the square root of the energy transfer ratio.
7 RELATION BETWEEN THE ANGLE OF THE HAMMER BLOW AND ENERGY TRANSFER RATIO The energy transfer ratio of the SPT represents the efficiency with which energy is transferred from the kinetic energy of the hammer to the energy transmitted to the rod system, and it is generally understood that its magnitude depends upon the method of hammer release, e.g. larger for the trip method and
(A)
Angled blow hetwecn bottom of hammer and top of anvil
Free fall height of hammer (cm) (11) Free fall height of hammer vs penetration Figure 8. Adjustment of penetration or N -value b y J e . (after Nagasaki et al.1998).
Table 2.Adjustment of N-value by energy transfer ratiocomparison between Int’l reference test procedures and Fujita’s niethod.(after Fujita,l997). Energy transfer ratio
Symbols Adjusted N-value for N-value Int’l ref. Fujita’s method NI00 E?? __ 18 23 Kovacs’s max Nne 20 19 24 K O V ~ C Save ’S Ntj7 27 22 28 Int’l reference 30 23 30 N45 40 27 35 Kovacs’s min Nv 55 31 40 18,3:Assumed N-values, other N-values are caluculated by the respective adjustment methods.
e 1.00 0.92 0.67 0.60 0.45 0.33
Referenccs
(c) Free fall height of hammer vs energy transfer ratio Figure 9. Effects of angled blow to penetration and energy transfer ratio.(after Fujita et a1.1994).
473
smaller for the rope and cathead method. Fujita et al. (1994) have frequently observed that the bottom of the hammer contacts the top of the anvil at an angle during the hammer blow of the SPT, since the rope of the hanging hammer, the centerline of the hammer and the centerline of the rod are usually not in a straight line and usually are not completely vertical. The amount of the deviation is likely to depend on the method of hammer release and workmanship of the operator, and these many affect the magnitude of the energy transfer ratio. To examine the effects of angled contact during the hammer blow of the SPT on the energy transfer ratio by applying the wave equation analysis, a series of laboratory tests simulating the SPT were carried out using a l m long instrumented steel bar with a diameter of 25mm, an anvil with a surface inclined at 0,1,2 or 3 degree and a steel donut type hammer weighing 49N provided with an inside diameter suitable for making contact at an angle. The results show that when the slope of the surface of the an\d is increased, the energy transfer ratio becomes smaller as shown in Figure 9(a), and the penetration of the rod per blow also becomes smaller as shown in Figure 9(b). Figure 9 indicated that the SPT apparatus should be set up accurately in a straight vertical line in order to ensure the free fall of the hammer.
tween hammer blows should be determined properly. 9 INFLUENCE OF PORE WATER AND OVERBURDEN PRESSURE ON THE NVALUE Studies of the influence of water content and overburden pressure on the N-value are interesting issues to be examined. Nagasaki et al. (1997) have contributed the results of a series of laboratory SPTs using dry and saturated Toyoura sands with relative densities of about 40%, 60% and 80% prepared uniformly in a heavy steel tank, employing two types of pressurized air bags for simulating overburden pressures of 100kPa, 300kPa and 5OOkPa. Figure 10 gives the results of the test, and shows that the N-value increases with overburden pressure. The N-value of the dry sand at relative density of 60% is approximately equal to that of saturated sand and the N-values of dry sand at the densities of 40% and 80% are smaller and larger than that of the saturated sand, respectively. On the other hand the Nvalues are larger and smaller respectively, when another type of pressurized air bag is employed for the saturated sand as shown in Figure 11.
8 PROBLEMS RELATED TO EXCESS PORE WATER PRESSURE Since the dynamic resistance of piles at the end of the driving operation had been considered to be the same as the static pile bearing capacity derived from the load vs. settlement diagram using Davission’s method. Fujita & Kusakabe (1988) have pointed out that the number of days that have elapsed since the pile installation is a very important factor for evaluating the static bearing capacity of piles driven below the ground water table. The recent introduction of a set-up ratio, around 2 to 5 in magnitude, could be a solution for the said problem, if the ratio could be determined properly. When the SPT is carried out in ground below the water table where excess pore water pressure is induced by the hammering of the sampler, similar problem may occur, In this case, a smaller N-value would be obtained than if there had been no excess pore water-pressure. The magnitude of the excess pore water pressure depends upon the type of soil, volume of displacement, the interval and number of hammer blows etc., and dispersed with time at a rate proportional to the pressure and permeability of the soil. In rare cases a negative excess pore water pressure may occur locally. The excess pore water pressure affects the magnitude of the N-value and so the interval be-
Figure 11. Effcts of pressure bag to relative density and N-value of saturated sand.(after Nagasaki et al. 1997).
474
Table 3. Comparison of I'\;-valuc between dry and saturated sands at relative dcnsity of 40,60 and 80%.(after Nagasakiet al. 19'97). DJ 40% Dr- 60% D r - 80% Gibbs-l-loltz (1957) Nrr>N,. Nd> N , N(J>N, Y anase (1963) N(s>NLV N ( J > N , Nct>N,br Yaniacla et al. (1004) Ni>N,, N(s=N,. Nis
These results are summarized in Table 3, which also includes other contributions by Gibbs & Holtz (1957), Yanase (1963) and Yamada et al. (1994). According to Table 3, there are two cases in which the N-value of dry sand is larger than that of saturated sand, and three other cases in which the N-value of dry sand is approximately equal to that of saturated sand at relative density of 60%, while it is either larger or smaller at the relative density of 40% and smaller or larger at the density of 80%. Table 3 suggests that either the N-value is not always proportional to the effective stress in the model ground or the N-value of the saturated sand is significantly affected by the type of air pressure bags employed for the laboratory tests. 10 CONCLUSIONS The application of stress wave equation analysis to the SPT is very useful for the quality control of the SPT as well as for the quality assurance of the Nvalue. The laboratory experiments on the SPT and modified tests using various types of uniform model grounds contained in a steel tank which is provided with an air pressure bag to simulate overburden pressure loads are also very advantageous for studying the mechanics and dynamics of the SPT, because it is possible to produce various ground conditions and to have repeatability. However, there have not been enough studies of this kind in Japan. With respect to the International Reference Test Procedures, it is recommended that a method should be introduced to adjust for the length of hammer, because when the length of the hammer is longer, the penetration per blow is larger, according to the unpublished results of a recent modified test. The analysis of the frequency characteristics of the measured stress wave in the laboratory or on an in-situ SPT would provide valuable information in this field, according to the latest study now in progress. Further studies are recommended.
REFERENCES Fujita, K., 0.Kusakabe. 1988. On the evaluation of static bearing capacity. B.H.Fellenius(ed), Proc. 3"' Int. Con$ on Application of Stress-wave Theory to Piles. 525-534. Ottawa: Fujita, K., R.Tanaka, M.Umemura, & T.Une. 1994. Angled blow of hammer against head of piles. Proc. 49'" Annual Con$ of JSCE. 3: 954-955. (in Japanese). Fujita, K. 1997. Referable N-values of unstandardized standard penetration tests. J. of JSCE. 82(12): 24-27. (in Japanese). Fujita, K. 1997. Mechanism of SPT, interpretation and evaluation of N-values. Kisoko. 25( 12): 2-13. Tokyo: Sogodoboku-kenkyusho. (in Japanese). Matsumoto, T., H.Sekiguchi, H.Yoshida, & K.Kita. 1992. Significance of two-point strain measurement in SPT. Soils and Foundations, (32)2: 67-82. Nagasaki, T., K.Fujita, T.Une. 1996. Dynamic resistance of SPT in dry sand subjected to overburden pressure. Proc. 51" Annual Conf of JSCE. 3A: 758-759. (in Japanese). Nagasaki, H., K.Fujita, Y .Imamura, K.Nohara & K.Yamamoto. 1997. N-value, energy transfer ratio and dynamic resistance in saturated Toyoura sands. Proc. 32"d Annual Con$ on Geotechnical Engineerng. JGS. 191-192. (in Japanese). Nagasaki, H., K.Fujita, T.Yoshinaga & H.Yamaguchi. 1998. A rational adjustment of SPT N-value by square root of energy transfer ratio. Proc. 53rdAnnual Con$ of JSCE. 3A: 762-763. (in Japanese). Uto, K., M.Fuyuki, H.Kondo, M.Morihara & H.Matsumura. 1974. Consideration of N-Value and strength of ground from a view-point of wave theory. Proc. Faculty of Engineering, Tokai University. (in Japanese).
475
This Page Intentionally Left Blank
ra
This Page Intentionally Left Blank
Application of Sfress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
ynote lecture: Vibratory driving analysis Alain E. Holeyman Universite‘Catholique de Louvain, Louvain-la-Neuve, Belgium
ABSTRACT : Engineering issues related to vibratory driving of piles and sheet-piles cover many facets including the long-term bearing capacity of the installed pile, its vibratory penetration resistance, the performance of vibrators, the degradation and liquefaction of the soil around the vibrated profile, and vibratory nuisance to the environment. An argument is made that those issues will be adequately tackled once combined into a comprehensive framework of analysis where proper understanding of soil behavior is the key. The present paper however focuses mainly on our current engineering ability to assess vibro-drivability, i.e. predicting the vibratory penetration log of a given pile into a given soil profile using a given vibrator. Testing undertaken to provide insight into the pile-soil-vibrator interaction and its modelling is reviewed. Several available methods to establish the vibratory performance of a pile from its vibratory capacity are discussed. A rational procedure to model the dynamic nonlinear soil structure interaction during pile vibratory driving is discussed in more detail. Degradation of the skin fiction upon cyclic shear stress is evaluated by applying elements of earthquake engineering practice used to assess liquefaction potential. The present ability to assess the vibratory capacity of a pile from the monitoring of its vibratory performance is critically reviewed. Finally, suggestions for hrther research, design and practice are provided.
1 INTRODUCTION
1.1 Scope
The main purpose of this key-note paper is to present the author’s present view on engineering issues relating to the drivability of piles and sheet-piles using vibrators and the inverse problem, i.e. deriving the pile resistance from its vibratory perf-ormance. It is based on a survey of the relevant literature and original research in the area.
means of installing piles and sheet piles in appropriate soil conditions. Major vibrators manufacturers are now located in Germany, France, The Netherlands, USA, the Former Soviet Union and Japan. Although technological developments have been brought to enhance the initial concept and extend commercial application of vibrators, little is mastered by the engineer when it comes to addressing soil related issues. That limitation in the engineering knowledge is viewed by many as an impediment for the vibratory driving technique to enjoy its full potential.
1.2 Historical Development 1.3 Phenomena at P k y
The vibratory driving technique appears to date back to the early 30’s when it was simultaneously developed in the former USSR and in Germany (Rodger and Littlejohn, 1980). The observation by the Russian soil dynamics researcher Pavyluk that soil resistance could be reduced thanks to vibrations led to the industrial use of vibrators to drive piles, according to Barkan, 1960. Extensive research on the effects of vibration on soils was conducted in the 40’s and 50’s by Barkan, while the vibratory driving technique was gaining acceptance as an economical and effective
Three major actors play a role in the mechanics of the vibratory driving process, as illustrated in Fig. 1 : (I) the pile to be driven, (2) the selected vibrator, and ( 3 ) the imposed soil conditions. The pile can be fully described by its material and geometry. The vibrator mechanical behavior can be assessed based on its specifications and operational range, as discussed in Section 2. Soil conditions are usually characterized by means of standard investigation tools such as CPT soundings, borings and laboratory tests.
479
Figure 2. Installation and design process of vibratory driven piles
1.4 Engineering Issues
Figure I . Vibratory Driving : Players and Issues
Those investigation tools are geared towards answering general design questions (mostly static) but are not well suited to characterize soil behavior under pile installation conditions, specially if the piles are vibratory driven. Because it is has been established for more than half a century that soil resistance during vibratory driving (likewise during impact driving) is lower than the long-term bearing capacity, these two resistances should be distinguished. As shown in Fig. 2, one can estimate the vibratory capacity from the long-term bearing capacity by taking soil degradation eEects into account. Conversely, one can estimate the longterm bearing capacity from the vibratory capacity if soil set-up can be accounted for. A hndamental understanding of soil behavior under vibratory loading is required to establish the relationship between the pile vibratory resistance and its long-term bearing capacity. Soil resistance degrades upon cyclic shearing mainly because of fatigue of the soil skeleton in cohesive soils (Vucetic, 1992), and of effective stress reduction in granular soils (Casagrande, 1938). The effective stress can be ultimately reduced to nearly zero, at which point the soil behaves in a fluid-like manner. These phenomena will be reviewed in more detail in Section 3.
Engineering issues related to vibratory driving cover niany facets, as iIlustrated in Figs. 1 and 2. They prompt the following questions : - What is the long-term bearing capacity of the installed pile? We h o w it depends on the pile geometry, on the soil parameters, but also on the vibratory process. - How are the soil’s long-tem strength parameters influenced by the vibratory process? By how much will the soil compact, and what is the magnitude of the potentially induced settlement? .- How are vibrations transmitted to the surrounding soil, and how much potential damage can they cause to neighboring structures? -. Will a given vibrator be able to drive the pile to the required design depth? If so, at what speed? Are there soil types that strongly limit vibratory penetration depth? - Are there ways to assess the vibratory capacity of a pile from the monitoring of its vibratory performance? - Is there a vibratory testing technique and interpretation leading to estimating the long-term pile bearing capacity? One can actually state that all issues will be properly tackled when combined into a comprehensive framework of analysis where proper understanding of soil behavior is the key. As clarified in Fig. 2, the present paper will focus on our current engineering ability to assess vibro-drivability, i.e. to predict the vibratory penetration log of a given pile into a given soil profile using a given vibrator. Testing undertaken to provide insight into the pile-soil-vibrator interaction and its modelling will be reviewed in Sections 4 and 5. We will then focus on some available methods to establish the vibratory performance of a pile from 480
its vibratory capacity (Section 6), and look into the potential to establish the reverse relationship (Section 7). FinalIy, Section 8 provides suggestions for further research, design and practice. Because of space hnitations, the paper does not f ~ c u on s other important engineering issues such as: bearing capacity of vibro-driven piles derived from soil characterization, vibrations transferred to the environment, and equipment specifications. 2 PILES AND VIBRATORY EQUIPMENT
2. I Vibrated Piles Pile types or profiles mostly used in combination with the vibratory driving technique include: - sheet piles installed for temporary shoring, cofferdam and permanent retaining and containing walls, - H-piles vibro-driven as deep foundations or vibrated to help install underground hydraulic barriers, - Tubes to install cast-in-steel-shell (CISS) piles - Precast prestressed concrete piles - Steel profiles to vibro-compact granular soils at depth. Port, harbor, near-shore and offshore projects very oRen take advantage of the vibratory penetration techniques, as the environment lends its self to substantial tolerance of vibratory disturbances. Steel and concrete profiles are generally cylindrical or prismatic, and can be characterized by the following geometrical and mechanical properties : A [m2]:profile section L [m]: profile length x [m]: profile perimeter E [Mpa!: Material Young’s Modulus p [kg/m 1: Material volumic mass
Figure 3 . Mechanical action of a vibrator
The vibratory action imparted to the pile is produced by counter-rotating eccentric masses actuated within an “exciter block”, as shown in Fig. 3a. The centrifugal forces acting as a result of inertial effects on an even number of symetrically moving nmses combine into a sinusoidal vertical force :
F,(t) =
me .a2sin (at ) = F,sin ( a t )
(1)
where F, = maximum centrhgal force of the vibrator p] me = eccentric moment of the vibrator [kg.m] o = angular frequency of the vibrator [rad/s] Alternative quantifications of the angular fiequency are the rotation speed R [rpm] and the fiequency v [Hz], with :
R
=
6Ov=60-(a/h)
(2)
The vibratory action can be therefore assessed once both the eccentric moment and the operating fi-equency are known. That action will be balanced by reactive inertial effects of masses undergoing the imparted vibratory movement and by soil reactions opposing the profile movement. Provided the center of gravity of the rotating masses belongs at all times to the profile neutral axis, the exciter block is assumed to exert a purely longitudinal force onto the profile. The exciter block is connected to the profile via a clamping device and is suspended to a carrier. The suspension device includes a vibration isolator mechanism consisting of a quasi-stationary heavy mass directly suspended to the suspension hook and an intervening spring, generally consisting of elastometer pads. The vibrator can be viewed as a two degrees of freedom system moving in the longitudinal direction (see Fig. 3b) : an exciter block of mass h&b and an isolator block of mass Mlb., sometimes called bias mass. Therefore Mvib = h& f Mlb. Those two masses are interconnected via an isolation spring with constant ki. In addition to the effort generated by that spring, the mass mb is subjected to gravity (g) and
The profile section can be more fully characterized by its shape, and inside and outside perimeters if closed. This allows one to calculate the areas of the profile in longitudinal and transversal contact with the soil, once an embedment depth z [m] is assumed. The mass of the profile M, equals PAL [kg] while the longitudinal wave speed in the profile is given by c = J E i p [Pnls]. Although they may at time play an important role, transversal and flexural properties of the profile will generally be ignored in the analysis that confines itself to the longitudinal behavior of the profile.
2.2 Mechanical action of a vibrator The mechanical action of a vibrator onto a profile consists of two part: a vibratory action and a stationary action.
48 I
the sinusoidal force described by eq. (1) whereas the mass Mib is subjected to gravity and the suspension force T. The net quasi-stationary action on the pile resulting from the carrier operation and vibrator is the weight of the vibrator mass and its clamp Mcl deducted by the suspension force:
2.3 VibratorMovemerit The movement of the vibrated body will depend on its so-called dynamic mass and the soil resistance. Specifications of vibrators often list a “maximum amplitude” S,. That number [generally expressed in mm] corresponds to the total (i.e. double) amplitude of movement for a free hanging vibrator, thus assuming a dynamic mass consisting of the exciter block M,b and the clamping device ml:
Ssp = 2 so
= 2 nze/(Meh
+ Mc,)
It should be noted that the double amplitude does not depend on the operating frequency, as the center of mass of the free mechanical system remains stationary, irrespective of the frequency. The amplitude of the free hanging pile to be vibrated will always be smaller than the specified amplitude, as can be derived from eq. (4a), where the dynamic mass is increased by that of the pile (PAL).
with s = actual (single) amplitude of the dynamic mass. A power for the vibrator is often listed in the specifications. It generally corresponds to the nominal power of the motor actuating the eccentric masses. It does not correspond to standardized operational conditions of the vibrator in action. Power consumption is indeed dependent upon testing conditions. Barkan suggests that under pile vibratory conditions, the power follows a squared velocity law:
W[kW] = M,,
. n . ( s w ) ~=Pt.K
Experimental verification of that law shows the n value to depend on soil type and pile type; a range of 15 to 50 Hz is observed.
O’Neill and Vipulanandan (1 989) provide an expression of the theoretical power required to maintain the vibrating regime of a dynamic mass in the absence of soil reaction but accounting for the presence of the isolating spring and the bias mass Mib. That formula is however of limited practical use as it provides very low estimates of the power. It is the author’s opinion that power limitation of the equipment is neither sufficiently characterized, nor (therefore?) properly accounted for in vibratory driving analyses conducted to date. 2.4 Types of vibrators Two main types of vibrators are commercially available: hydraulic and electrical. In both cases, the motor is housed in the vibrator and powered through a transmission line connected to a separate or carriermounted diesel-hydraulic or diesel-electric power pack (see Fig. 3a). Hydraulic vibrators are lighter than their electrical counterparts, because of the smaller size of the motor. The adjustment of the operating frequency is more readily available on the hydraulic vibrators, which also explain why they are more commonly used. Five types of vibrators can be distinguished based on operating frequency and eccentric moments, as summarized in Table I . It can be noted that initial improvements of the vibratory driving technique targeted the speed of driving, whereas more recent improvements are attempting to mitigate environmental impacts associated with the technique. Noteworthy amongst recent developments is the “variable vibrator”, which can adjust on-the-fly its effective eccentricity by shifting the phase angle between a multiple of 4 masses. The claimed advantage of such an adjustment is to avoid “soil resonance”, a term coined after the observation that vibration levels pass through a peak upon vibration start-up and shut-down. This phenomenon will however be shown later not to be necessarily related to a particular fkequency. Vibrator choice amongst practicionners is generally based on experience and field verification. Rodger and Littlejohn (1980) have summarized that body of experience into a table recommending fi-equency and amplitude parameters for different piles and soil types. Those recommendations are reproduced herein as Table 2.
Frequency range [rpm]
Eccentric moment [kg.ml
“Standard frequency”
1300-1800
High fTequency
2000-2500
Variable eccentricity Excavator accessory
Type
Resonant dnver
Maximum centrifugal force FNJ
Free hanging double amplitude [mm]
up to 230
up to 4,600
up to 30
6 to 45
400 to 2,700
13 to 22
2300
10 to 54
600 to 3300
14 to 17
1800 to 3000
1 to 13
70 to 500
6000
50
20,000 (in theory)
6 to 20 Self destructing
482
1
Cohesive soils All cases wlgh acceleration Low displacement amplitude
Dense cohesionless soils Loosc cohesionless soils Low point resistance Ihgh pomt resistance ] Heavy piles Light piles &gh acceleration Low frequency. Large displacement amplitude IIigh acceleration
Predominant side resistancc
Predominant side resistance. Predominant end resistance
Requires high acceleration for either sheanng or thixotropic transformation
Iiequlres high acceleration Require? lugh displacement aniplitude and low Requires high acceleration for fluidization frequency for maulmum impact to permit elasto- for fluid17ation plastic penetration
Predominant side resistance.
the vibrating profile.
3 SOIL BEHAVIOR UNDER VIBRATORY
LOADING 3.1 Fzn~dumentals
3 ,2 Static and Cyclic Stress-strain Behavior
As the profile undergoes a vibratory vertical motion of amplitude s, it communicates to the lateral neighboring soil shear stresses and shear strains, as sketched in Fig. 3b. It is also forcing normal and potentially convective movement of soil below the pile toe. As those mechanisms govern soil resistance along the shaft and at the toe, the understanding of the shear stresdshear strain relationship, i.e. T (y)? within the soil becomes of paramount importance. That aspect of soil behavior has been more extensively studied within the field of earthquake engineering, leading to the characterization of so-called constitutive relationships, generally on the basis of laboratory testing of soil samples (mainly triaxial testing and simple shear testing). The constitutive relationshipsthat represents the complex large-strain, dynamic and cyclic shear stress-strain strength, behavior of the medium surrounding the vibrating profile require the characterization of the following elements : e Static stress-strain law expressing nonlinear behavior under monotonic loading and hysteresis upon strain reversal, Shear modulus at small strains and ultimate shear strength, Softening and increase of hysteretic damping with increasing strain, a Effect of strain rate on initial shear modulus and ultimate strength, o Degradation of properties resulting fi-om the application of numerous cycles, and last but not least, e Generation of excess pore pressure leading substantial loss of resistance and possibly to liquefaction. The following paragraphs address key components of the constitutive relationships and provide insight on the intrinsic soil behavior in the vicinity of
A typical soil response to uniform cyclic strains with amplitude yc is represented in Fig. 4, which highlights the following hndamental parameters: Gmax : initial (or tangent) shear modulus zc : shear stress mobilized at yc Gs: secant (or equivalent) shear modulus h: hysteretic (or intrinsic) damping ratio;
Q
Q
483
A.= AW I 27cyczc
(5)
with AW = Energy lost during a given cycle. Both G, and h are strain-dependent parameters that need to be described by specific laws within a given cycle. T~~~ is the ultimate shear strength, revealed at large strains. znlaVand G, are shown to decrease with the number of cycles (cyclic degradation). 3.3 Initial Shear modulus and ultimate shear
strength (emu and zmJ
Numerous studies have dealt with the initial shear modulus to be used in earthquake engineering (e.g. Drnevich et al., 1967). Most of them are supported by parameters determined in the laboratory which are generally not available when a vibratory penetration issue arises. However, correlatins with CPT test results have been more recently developed (Seed and De Alba, 1986, Robertson and Wride, 1998) 3.4 Secant Shear Modulus and Hysteretic Damping (Gsand A.)
As can be observed in Fig. 4, G, decreases with the shear strain during the initial monotonic loading. The curve that represents the initial monotonic loading is referred to as the initial "backbone" curve, because it also serves as the basis to generate the
towards the upward curvature of the stresdstrain curve at large cyclic strains. From the point of maximum straining, the unloading curve is described by the following equation, in accordance with Masing's rules 1 and 2 (Masing, 1926):
The energy dissipated within a loop depends for a given soil on the amplitude of the cyclic strain. Empirical data collected in laboratory tests indicates that the damping ratio increases with yc as the soil undergoes higher plastic deformations. Dobry and Vucetic (1987, Vucetic and Dobry, 1991, and Vucetic,1993 and 1994) have suggested a unifying approach to accommodate the influence of the nature of the material characterized by the plasticity index (PI), as indicated in Fig. 5 3.5 Strain Rate Effects
Although it is well known that undrained modulus and shear strength increase with increasing strain rate (P=ay /at ), experimental data generated using different apparatuses and loading conditions lead to different conclusions. Viscosity mechanisms may well provide a suitable framework to understand the strain rate effect observed when comparing fast and slow undrained monotonic stress-strain curves, as well as to explain the roundness of the loop tips during a sinusoidal strain-controlled cyclic test. Evidence would point to the fact that sands and non plastic silts have very small viscosity in that their stressstrain loops exhibit sharp rather than rounded tips (Dobry and Vucetic, 1987). Figure 4. Soil Behavior under Constant Cyclic Shear Strain Amplitude Loading (From Vusetic, 1993; 1994)
family of curves corresponding to unloading and reloading. Kondner's mathematical formulation ( 1963) is frequently employed to describe the initial backbone curve in earthquake engineering. That hyperbolic law is best represented in terms of reduced variables, q , the mobilization ratio and 6, the relative shear :
The mathematical fhctions proposed in the fiterature to represent the nonlinear viscosity also depend on the type of experimental observations. A power law is often adopted :
q = zI z,, = 6 /(6+1)
and
y, =zm /Gmax
Yr is called the reference strain. Two of the three parameters Gmax, y,, and Tmax, are generally derived M~~~ extensive l a b from laboratory ratory surveys by Robertson and Wride (1998) point 484
Figure 5. Soil stiffness degradation resulting from cyclic shear (Vucetic, 1993)
'kin
Laboratory results conducted at constant cyclic strain show that in many soils, the degradation index after N cycles can be approximated by the following relationship as suggested by Idriss et a1 (1 978):
=
with
Tkin = kinetic ultimate
shear strength [ P a l shear strength [Wa] = shear strain rate [s-]] The advantage of that mathematical form is that resistance does not vanish as the strain rate goes towards zero. The power law also requires the strain rate to vary by orders of magnitude to provide tangible increases in both the modulus and the ultimate strength. The J coefficient and n exponent depend on the nature of the soil. Based on pile driving data, have been suggested for plasn=0.2 and J=0.3 tic soils. J should therefore essentially depend on the plasticity of the soil and become quite limited for granular materials.
xsta
= "static" ultimate
A = A'-'
The exponent t, called degradation parameter, depends mainly on the amplitude of the cyclic strain and the nature of the material (PI), as suggested by Dobry and Vucetic (1988) and as indicated in Fig. 6 Vucetic, 1993). It is noteworthy that the degradation parameter assumes a zero value at strains smaller than a cyclic "threshold" shear strain, y,, The threshold strain increases with the plasticity of the soil, as suggested in Fig. 6. 3.7 Soil liquefaction Vibration induced compaction of saturated sands has received attention not only from the earthquake engineering community, but also from vibrocompaction specialists. Recent advances tend to indicate that build up of pore pressures (eventually leading to liquefaction) and volume reduction of cyclically loaded materials are the expression of the same phenomenon, i.e. the irreversible tendency for a particulate arrangement to achieve a denser packing when sheared back and forth. Under drained conditions, the volume reduction is immediate. Under undrained conditions, the tendency for volume reduction is expressed by an increase in the pore water pressure (see Fig. 7), such that the effective stress is reduced to a value that may be close to zero. It is then necessary to wait for the soil to consolidate in order to see the volume reduction take place. The strain driven evaluation of the build up of pore pressure as suggested by Dobry et al. (1 979) is an approach that lends itself to a direct transposition
3.6 Degradation Law When subjected to undrained cyclic loading involving a number N of large strain cycles, the soil structure continuously deteriorates, the pore pressure increases, and the secant shear modulus decreases with N. This process known as cyclic stiffness degradation can be best characterized on the basis of strain controlled tests for the type of loading involved with the vibratory penetration of piles. Typical results of strain-controlled tests are sketched in Fig. 5 , where the degradation is clearly expressed by the decrease of the amplitude of the peak stress mobilized at successive cycles. The quantification of the degradation process calls for the introduction of the degradation index A, defined by:
(9)
Figure 6. Effect of Plasticity Index (PI) on soil degradation (Vucetic, 1993)
(10)
Figure 7. Build up of residual pore pressure in different sands in undrained cyclic strain-controlled tests (Dobry et al., 1982)
485
to the problem of the vibrations induced by a vertically vibrating pile. It also allows one to evaluate potential changes of the void ratio based on a cyclic strain rather than stress history, as supported by laboratory drained tests conducted on sands by Youd (1972). That framework of analysis entranced by the threshold cyclic strain concept embodies in a single model the intrinsic relationship between degradation and pore pressure build-up, with the advantage that it can be applied to general categories of soils (sands to clays) The excess pore pressure generated during cyclic loading has been shown (see Fig. 7) to increase with the shear strain and the number of cycles for a given soil type. The damage parameter K approach (Finn, 1981) can be adopted to evaluate the excess pore pressure 6u resulting from a particular strain history, as characterized by the following equations :
d~ / d o = A / 4 In (1 + rd2 ) *
with Relative Energy Loss given by Eq. 5, and =<ex’ (16) with C = 5 and 5 = length of strain path = 4 N yc, for constant amplitude cycles
4.1 Conceptual model testing Tests have been conducted by several Russian researchers to investigate the “vibro-viscous” resistance of soils. In particular, Barkan (1963) reports on the sphere test, shown in Fig. Sa, where a steel ball is sunk into a vibrated soil vessel with the assistance of a bias force. Penetration speed is shown to obey Stokes sedimentation law (see Fig. 8b), allowing one to determine an equivalent viscosity p . The inverse of that equivalent kinematic viscosity [cm.s/kg] was shown to vary linearly with the relative level of acceleration (a/g), passed a threshold value of approximately 1.4 for a dry sand (see Fig. 8c). The influence of the water content on the “vibroviscosity factor l/p” of a sand vibrated at constant a/g is also shown in Fig. 8d, highlighting the near total loss of vibro-penetrability at optimal water content.
(11) 4.2 Pile-soil interface testing
K
(12)
4 PILE VIBRATORY DRIVING TESTING The above discussion of soil behavior under cyclic loading does not encompass the particular geometry of the profile-soil interface, nor does it consider the continuous penetration of the profile that leads to successive exploration stages into “virgin” soil behavior. That is why a number of experiences have been conducted to reveal soil-structure interaction w i t h a vibratory framework. Based on the ambition and complexity of the tested interface, one can categorize various experiences reported in the literature as conceptual, interface, and both reduced and fullscale testing.
Soil shear strength resisting the pulling out of a vibrating steel plate against a normal stress controlled medium sand (vibratory direct shear box) has been investigated in the early days by Levchinsky and Savtchencko (Barkan, 1963). The friction coefficient (tan @ = do) was shown to decrease with cyclic amplitude and frequency. The ultimate relative reduction of the friction was also shown to increase with the grain size within the investigated range shown in Fig. 9. Fig 9 shows that the sand vibratory friction angle can easily drop to !h to 1/5 of its static value. 4.3 Reduced scale tests Testing of model profiles in soil tanks were initially attempted by Bernhard (1968), Schmid and Hill (1966), continued by Rodger and Littlejohn (1980), Billet and SBert (1 985) and O’Neill et a1 (1990),
and more recently by Viking (1 998) and Holeyman et a1 (1999). The tests were generally conducted using a lightweight vibrator acting on a heavily instrumented profile. Monitoring included strain gauges, accelerometers and displacement transducers. The soil used was exclusively sand (dry, moist, or saturated), placed at a controlled density, and in some cases, confined at a controlled radial stress. Monitoring of the soil response involved accelerometers, total stress and pore pressure cells during installation as well as compactlon and-in situ testing ager instalYation. Insighthl observations relative to the vibratory toe resistance have been reported by Schmid (1966), who identified three regimes, depending on the magnitude of the driving force : a sinusoidal resistance domain, for a driving force lower than the “resistance threshold” an impact domain, when the upward force exceeds the soil uplift resistance; the toe of the pile alternately separates from the soil and tamps it a phase instability domain, when the downward force exceeds the soil compressive resistance. Rodger and Littlejohn (1980) call upon acceleration amplitude to distinguish: - the elastic state (a<0.6g) - the trans-threshold state (0.6gl.5g). Although their views are contradicted by some of Barkan’s observations, these three different states are stated to be confirmed by dynamic direct shear tests performed by others. Results of‘tarik experiments’have’ been reporte‘d’in terms of friction reduction coefficients, potential optimal operation, and have shed some light on hndamental soil behavior. Correlations have been established between penetration speed and parameters induced by the vibrator (amplitude, frequency) and by the soil (grain size, relative density, and lateral stress). Although conclusions of the tests conducted under different conditions do not consistently agree, those experiments generalJy identified that: penetration speed increased when the relative density decreased and the bias mass increased friction was reduced to 30 to 50% of its static value, while a more limited reduction was noted for the toe resistance optimum operation of the hammer required at times that the frequency or eccentric moment be reduced, while energy transfer was of the order of 40% of the hll theoretical power produced by the vibrator a number of observations cannot be explained. Although reduced scale models are of use, they suffer from improper boundary conditions (at the tank limits) that significantly prevent the vibration energy from propagating away from its source.
4.4 Full scale tests Because of inconsistencies in the conclusions derived from reduced scale tests, research has been conducted in several countries based on full-scale tests. Early hll-scale programmes have been conducted by Barkan (1963) and Davisson (1970). Other programmes have been conducted by manufacturers on specific equipment, but lead to a limited diffusion of their conclusions. More recently, collective European programs have provided actual penetration speed, but within soil conditions that cannot be controlled, only characterized. Monitoring nowadays involvies acceleration, strain, pore pressure, penetration speed, making the tested profile a hlly instrumented probe. Such programs have produced results that have not been hlly analyzed (BBRI, 1994, Sipdis, 1997); others are being presently conducted (KTH, 1999) or planned (IREX, 2000). Publication of such research results would be appreciated by the profession. 5 PILE-SOIL-VIBRATOR INTERACTION
MODELS 5.1 Types of models Models that have been suggested by various authors differ in the way they account for mechanical engineering principles. We will review models purely based on (1) force equilibrium, (2) momentum conservation, (3) energy conservation, and (4) integration of the laws of motion. 5.2 Force equilibrium models The force models aim at predicting whether a vibrator can or cannot overcome an estimated soil resistance. They will not provide an estimate of the driving speed. Jonker (1987) and Warrington (1989) have suggested, respectively:
force generated by the vibrator, per eq. (1) inertia forces of dynamic mass, = Mdp . a surcharge force, per eq. ( 3 ) empirical factor of shaft resistance outside pipe pile, soil resistance outside pile shaft, empirical factor of shaft resistance inside pipe pile, soil resistance inside pile shaft, soil resistance at Dile toe. For sheet-piles Tunke; Company recommends to replace x. z [m2] with 2.81 times the sheet-pile width.
487
5.3 Energy based models Energy based models assume the following general form: R.v, =
p, .W, + ( F , +F').v,
(1 4 4
leading to a direct estimate of the penetration speed :
vP=p, W,/ ( R -F, *
-F,)
(14b)
With: R = soil resistance, v,, = average rate of penetration in m/s, W, = theoretical power delivered to the system, Fi = inertia forces of dynamic masses. Davisson's formula (1 970) to estimate the bearing capacity for the Bodine Resonant Driver suggests :
p, =I--v*s,. R/1000
w,
(15)
where s,: is an empirically determined set [mm/cycle] representing all energy losses. Warrington (1989) has coined eq. (14b) as the 'Vibdrive' formula provided a value of 0. I is used for p, and the power Wt is calculated according to his procedure,
5.4 Momentum conservation models Schmid ( I 968) has suggested a formula implying that, for steady-state penetration, the momentum of the total mass of the vibrator. additional bias mass (MJ, and pile accrued by gravity over a vibration cycie be balanced by the soil resistance impulse: (Ads+hd,,,b+M,)g T= 1
6;Rdt=&T,
(164
with T, = contact time between pile toe and soil within a cycle and a = coefficient between 0.5 and 1, generally assumed to be 213. Conversely, the penetration speed follows a linear trend passed the threshold acceleration amln,which becomes a key parameter to successfdly apply the niethod and estimate T, :
5.5.1 Single degree of freedom (SDOF) Simplest models of the vibrator suggest that the dynamic mass be the focus of attention, thereby assuming that the pile behaves as a rigid body. Newton's second law can therefore be applied to the dynamic mass : a=
mem2sin(wl)
hrfIn,
(174
where U ,,
=Meb
+Md+ M ,
( 17b)
Holeyman (1993) has suggested a method that integrates the inertial effects of the excess force. That excess force is defined as the difference between the sinusoidal driving force and the opposing soil resistance. A distinction is made between the skin ftiction, which is reversible (Uplift resistance = Downward resistance) and the toe resistance, which cannot produce uplift resistance. Attention is also paid to the clutch resistance, which is combined with the skin friction. The soil degraded resistance at the toe and along the shaft is estimated from CPT test results where the friction ratio and acceleration ratio are used to assess the severity of degradation. The method involves an iterative procedure to identifL the coexisting acceleration and soil resistance (17b). The driving speed is obtained by intuitively integrating the net downward and upward accelerations over a complete cycle. The method have been verified and liquefaction parameters hrther refined through calibration with full-scale tests (BBIPI, 1994) Gonin (1998) has followed a similar approach that analytically integrates the effects of an excess force, as shown in Fig. 10. The integration is however performed soleby on the toe resistance, while the skin friction influence is accounted for in terms of damping of the driving force. In addition, the wave equation theory is used to estimate the displacement accrued at the toe over the period of net force exceedance.
5.5 Integration of law.^ of'rnotion Comprehensive accounting of the laws of mechanics requires that movement be described at all times from inertial equilibrium conditions. The simplest models involve a single degree of freedom. I-D models already offer more detailed description of some form of wave propagation, whereas 2-D models might provide hture solutions that integrate all types of wave propagation (compression, shear, Rayleigh, etc..).
Figure 10. Integration of excess toe force (after Gonin, 1998)
Figure 11. Resistance mobilization versus displacement for (a) skin (b) toe compression (After Dierssen, 1994)
Dierssen (1 994) has used a numerical integration scheme to closely follow the time dependence of the skin and toe resistances. Figure 11 provides the shape of the resistance mobilization versus displacement for both skin and toe resistance. One can note that separation of the pile from the soil at the toe is explicitely accounted for.
The model allows the constitutive relationships described in Section 3 to be readily deployed. The major advantage of that shear wave propagation model is to closely follow the development of degradation as more cycles are simulated. It can also provide insight into vibration levels in the vicinity of the pile. Both features are illustrated by Fig. 13 which provides the effective particle velocity calculated at several distances away fiom a profile upon vibrator start up. An apparent resonance is indicated, whereas the model does not include a Iongitudinal or radial dimension that could explain the frequency at which the peak vibration is noted : why? Simply because the model most probably reproduces two soilpile interaction vibratory modes: the coupled mode and the uncoupled mode. In the coupled mode (similar to Schmid’s sinusoidal domain), the soil remains in contact with the slowly vibrating profile, and the transfer of energy from the pile to the soil is nearly perfect. As the vibrator linearly accelerates (between 0 and 0.5 seconds), vibration levels tend to increase with the square of time since start up. However, as the soil begins to degrade, its shear modulus decreases and the specific shear impedance reduces, leading to loss in the energy transfer. At that point, the couplmg between soil and pile suffers some slippage, and therefore time lag. After a sufficient number of cycles, the soil has significantly degraded, and has (60 seconds ageing skipped in Fig. 13) entered into liquefaction.
5.5.2 Radial 1-D model Holeyman (1993b) have suggested the use of a radial discrete model to calculate the vertical shear waves propagating away from the pile. The model, shown in Fig. 12, consists in a succession of concentric cylinders with a linearly increasing depth. The equations of movement are integrated for each cylinder based on their dynamic shear equilibrium in the vertical direction, in a manner similar to that used by Smith (1 960) in the longitudinal direction.
Figure 13. Vibration levels and penetration state parameters estimated upon vibrator startup and regime
Figure 12. Radial I-D model (Holeyman, 1993b)
489
At the shear modulus of the soil in contact with the profile is nearly zero, and very little energy can pass through the fluidized surrounding zone. The soil in the vicinity of the profile cannot anymore follow the profile movement, from which it uncouples itsel, resulting in a lower level of vibration. That example demonstrates that apparent resonance of soil vibration may be no more than the transient combination of increased rotation speed and soil degradation. The model can also shed light on “damping” as it clearly separates geometric damping from the energy losses attributable to viscous and hysteretic behavior. 5.5.3 Longitudinal I-D models Few authors have adapted Smith‘s (1960) classic lumped parameters model to represent the longitudinal behavior of a pile subjected to vibratory driving. Gardner (1981) and Chua et al. (1981) have developed a wave-equation computer code where the vibrator is represented by a two-mass system, separated by a sofi spring, while the exitor black is subjected to a sinuso‘idal force (cfi-. eq. (I)); as shown in Fig. 14. The soil behavior is represented by spring-slider-dasplot systems, according to Smith’s early suggestion. Middendorp and J o k e r (1988), as well as Ligterink et al. (1990) used the TNOWAVE computer program to analyze the driveability of offshore vibratory driven pipe piles, based on the methods of characteristics. The authors identify the need for a soil model able to describe the degradation of the soil resistance as a function of the oscillation history, and warn that soil parameters may depend on opera- ting fiequency and pile movement amplitude.
Moulai-Khatir et al. (1994) have developed together with the University of Houston, the so-called VPDA computer program (for Vibratory Pile Driving Analysis) wherein the action of the hammer is replaced by a static surcharge load and a sinuso‘idal load. The soil model was modified from Smith‘s original in that hyperbolic mobilization curves were adapted for the shafi and toe resistance, as shown in Fig. 15. A simple viscous damper was used to model damping along the shaft, while no damping was deemed necessary at the pile toe. It should also be noted that the GRLWEAP program has included in its latest versions (GRL,, 1998) the capacibilty to model vibratory hammers.
6 VIBRQ-DRIVABILITY ANALYSIS Most of the models discussed in the previous sections should be able to provide a reasonnable match of calculations with relevant field observations provided the models parameters are properly calibrated. The use of energy balance methods is discouraged by the author, while force equilibrium methods are of limited use because they do not provide vibratory penetration speed. Momentum based methods m y produce a penetration speed very similar to that obtained through integration of the laws of notion of a rigid body. Finally, wave equations methods should not produce penetration speeds significantly different form those obtained form a rigid body analysis, provided the vibrator speed is lower than the resonant frequency of the pile, which is generally the case.
Figure 15. Resistance mobilization for (a) skin friction (b) toe compression (After Moulai - Khatir et al., 1994)
Figure 14. Longitudinal I-D model
490
Exceptions to that general case include the Bodine Resonant driver and very long piles for offshore applications (LMO m). In the author’s opinion, the most critical parameter to assess in order to produce a resonable prediction of vibro-drivability is the soil resistance to vibratory driving. That is unfortunately where pertinent information and recent consistent experimental data is cruelly missing. The reliability of the predicted vibropenetration log will strongly depend on the degradation parameters adopted to assess the vibratory penetration resistance form the soil investigation results. The author’s experience leads him to use the following crude ultimate degradations coefficients : 0.15 in sand, 0.4 in silt, and 0.65 in clay for skin friction; as well as 0.55 in sand, 0.7 in silt and 0.85 for end bearikg. A more involved assessment of the degradation coefficient has been suggested (Holeyman, 19%) based on CPT test results. In that method, the soil driving resistance is obtained by interpolation between a static value and an ultimately degraded value. The static base (qs) and shaft (zs) resistance profiles derived from Cone Penetration (CPT) tests results, i.e. from the cone resistance qc and local unit skin friction f, (El cone). The ultimately liquefied base (ql) and shaft (z1) unit soil resistances are derived based on an exponential law as expressed below :
rl=r,[(l-1/ A).e-l”;R+1/A]
(1 8b)
where liquefied soil base resistance [Wa] liquefied soil shaft resistance [@a] FR = friction ratio as measured in a CPT test with E l cone (percentage of the mantle friction to the cone resistance, i e FR = 100 f, /qc) A = empirical liquefaction factor expressing the loss of resistance attribuable to liquefaction (A will be higher for saturated and loose sands and is chosen in the range of 4 to 10) The driving base (qd) and shaft (zd) unit resistances are derived from the static and the “liquefied“ soil resistance depending on the vibration amplitude following an exponential law as expressed below 41
=
21
=
qd=(q -ql ).e +qr
At each depth z the vibratory pile driving resistance is calculated : Rbase
=qd
where $2is the pile section, x the pile perimeter and D the pile penetration. 7 BEARING CAPACITY FROM INSTALLATION MQNITOR1ING Because soil resistance degradation is significant during vibratory driving, one should expect it a challenge to estimate the static bearing capacity from the end of penetration vibratory performance of a driven profile. In the impact driving practice, it is recognized that end of driving (EQD) data generally provides a safe estimate of the pile capacity; that is why beginning of restrike (BQR) or “retap” data is strongly advised to the owner who wishes to tap the value of letting the soil set up. If the end of Vibratory driving (EOV) data is used, methods to estimate the static capacity should allow for recovery of soil degradation, as highlighted in Fig. 2. However significant uncertainty should be expected in the process because the inverse of observed degradation coefficients may range between 2 and 10. That is why extreme caution is warranted when applying so-called pile Vibratory driving (PVD), formulae, even more so than already much detracted (impact) pile driving formulae. A limited number of such PVD formulae have been published; however only one has been, to the author’s knowledge been extensively field tested. The “Snipe” formula is therefore the only one that will be discussed in this paper. The formula is a field-based method was developed in the former Soviet Union according to Steffanof and Boshinov (1 977). The following empirical formula is used to predict the static bearing capacity Q,, :
(194
where QU
=
load capacity, in [kN];
w = power used by the vibrator to drive the pile, where qd zd
a
Ft
=
1/p
=
driving base unit resistance driving shaA unit resistance = acceleration ratio (= dg)of the pile, as obtained from Eq. (1 7a) = =
49 1
in [kW] total weight (force) of vibro-hammer and pile, in [kld]; = (M,ib + MJg empirical loss coefficient (in Soviet practice 143 is safety taken to be 5 in cohesionless
drivability. The observed frequency was 38 fi. From the measurement results, one can observe that : - the vibration amplitude at the pile top (0.65 mm zero to peak) is considerably less than the nominal vibration amplitude which is,
soils) reflecting the influence of driving on soil properties. The bearing capacity in of vibro-driven pipe piles has been verified by A monitoring the behavior of the beginning of impa the finished product strike. This of course requires that a specific procedure be enforced a sufficient time aft EOV installation, in order to allow pore pressures to dissipate and accrue soil setup. The monitoring of the installation of vibratory driven piles is not at all as widely spread as for impact driven piles. Recent improvements in the field monitoring devices FaSA., “NO-System, etc) now allows the geotec’nnical engi celeration, stress and energy equivalent interpretation of formula or CAPWAP method, available for nearly 15 years in the impact driven products. Field monitoring provides a tremendous advantage in controlling the effective performance of the vibratory hammer, as illustrated by the following case history reported by Holeyman et al. (1996). That case involved the installation of a 20.6 m long tubular steel pile with a thickness of 9.5 mm and a diameter of 1 m on a site in Kortrijk ( shows the subsoil profile as depicted by a CPT test (cone M4) performed at the site. The water table was encountered at a depth of -1.8 m. The upper twelve meters consist of very soft river deposits; below the underlying sandy layer was found a very stiff tertiary clay layer in which the tube had to be driven. A preliminary calculation using eqs. (1 7) through (20) pointed out that the necessary time to install the pile to a depth o f 2 0 m with a PTC 3 0 W V vibratory hammer was 13% minutes. However, driving met rehsal at a depth of 11 m. The reason for the difficult driving and the difference between the predicted and the observed penetration speed was explained by measurements taken during the actual driving of the pile. The pile vibration amplitude was measured by means of a velocity transducer placed at the pile head and a velocity transducer (protected by a cover) at the pile toe. Figure 17 shows the monitored amplitude of vibration at the pile top and at the pile base upon loss of
me -- -
M
me +M
-
26000kg.mm = 2.3mm (6500 + 4820)kg
- the amplitude at the pile base (0.45 mm zero to peak) is smaller than the amplitude at the pile top (0.65 mm) It would appear that the pile base amplitude (0.45 mm) is not sufficient to allow the pile to penetrate as the stress-strain behaviour for clayey soils is primarily elastic for small amplitudes. ossible explanations for that observation were considered : - An important soil (i.e. clay) mass was sticking to the vibrating pile, leading to a more important vibrating mass, leading to a smaller vibration amplitude. - The vibratory hammer was unable to deliver the required energy, and thus maintain its nominal plitude or frequency. A characteristic of the C variable eccentric hammers is that a lack of power results in a reduction of vibration amplitude (rather than a reduction of frequency (Houze, 1994)). - A smaller amplitude at the pile base was obtained due to the elasticity of the pile. By applying the observed vibration amplitude to the calculation model (Figure 181, a much better correlation between the calculated and the observed penetration time was obtained. The pile was placed at the bottom of an excavation at -2.5 m and penetrated 4.5m under its own weight. As a result, observed and calculated penetration rates are reported starting at level - 7 m. Figure 18 evidences that the difference for the predicted and observed penetration times for the site in Kortrijk was not due to an incorrect estimation of the dynamic soil resistance but due to an incorrect estimation of the vibration amplitude, which happened to be limited by the nominal power of the power pack. A more powefil power pack
Figure 17. Record from the vibration amplitude upon refusal
Figure 16. Subsoil profile site at Kortrijk
492
Billet, P . Siffert. J G . (1989) "Soil-sheet pile interaction in \ ibro-piling" Journal of Ceotechnical Engineering, ASCE. Vol 115, No8. pp 1085-1101 Chua, K M , Gardner. S . Lowery, L L . (1987) "Wave Equation Analysis of a Vibratory Hammer-Dnveii Pile". Proc Oficrore Techm1og-y Coiif. VoI 4. pp 339-345 Davisson. M T . (I 970) "BRD Vibratory driving formula". Foundatioii facts. Vol 1, No1. pp 9-11 Dierssen. Guillermo. ( I994), "Ein Bodenmechanrsches Modell zur Beschreibung des Vibrationsrammens in komngen Bodcn" Doctoral Thesis, University of Karlsnihe. Germany Dobry. R , Ladd. R S . Yokel. F V . Chrang. R M and Fowl!. D (E982) "Prechction of Pore Water Pressure BuiPdup and Liquefaction of Sands Dtrnng Earthquakes by the Cyclic Strain Method" Uafional Bureau of Standards Building Science Series 138, July 1982. 150 pp Dobry. R and Swiger, W F (1979) "Threshold Strain and Cyclic Beliavior of Coliesioiiless SO~IS" ~ r o c 3 1 ~ Specialiv Conference A ~ ~ s t i nTexas. pp 52 1-525, Dobry. R and Vircetrc M (1987) "State-o€-the-Art Report Dvnamic Properties and Response of Soft Clay Deposits" Proceeding\ of the In11 Svmponurn on Geolechnicnl &,a of.S'ofi Soils. Mexico City. Vol 2. pp 51-87 Drnevrch I J P Hall. J R . Jr , and kchart. F E Jr (1967) "Effects o f Amplit~dcof Vibration on the Shear Modulus of Sand Proceediiigs of the Iiifernatroi?al Svmposium on Fave Propagatioii nnd Dvnat~iicPropertier of Earth hfuferialr, Albtrquerque. N M . pp 189-199 Finn. Vd D L (1981) "Liquefaction Potential Developments Since 1976". Proceedings, Intl Conf oii Recent Advaiices in Geofechnical Earthquake Eiigineering and wrl l A ~ i i a ~ ? ~Sifc sLouis, , Missouri. Vol I€, pp 655-681 Garclner SherriIl. (1987) "Analysis of vibratory driven pile" ~ r o cof 2'ld Inl ~ o n Of M ~ e e ~orni~atron, p Luuenibourg, 5-7 May, pp 29-56 Gonin. J (1998) QueIques reflevions sur Ie vibrofonqage. Revue Fraiqai re de Geoiechnique, No 83. 2'"Ie trimestre. pp 35-39 Harchn. B 0 and Black, W L (1968) "Vibration Modulus of Normally Consolidated Clay mrnal of the Soil Alechmnc I aim' Fowm'atio~rsllil n, ASCE. Vol 94. No SM2. Proc Fapeer 5833. pp 353-369 Holevnan. A (1985) "Dynamic non-linear slun friction of piles." Proceedings of the International Synposiurn on Peiiefrdvlifv and Drrvabik fy of Piles, San Francisco. 10 August 1985. Vol 1. pp 173-176 Holeyman. A (1988) "Modeling of Pile Dvnainic Behavior at the Pile Base during Driving," Proceedings of the 3rd Iiiferiiafionnl Confereiice on the Applicnfion of Stress-Ffni e Thenrv fo Piles. Otlataa. May 1988. pp 173-285 Holeyman. A (19932) "WUPERVIBI An analytical modelbased computer prograni to evaluate the penetration speed of vibratory dnven sheet Piles". Research report prepared for BBH. June. 23p Holcymamr, A (1993b) "HYPERVIP3IIa. An detailed numenca! model proposed for Future Computer Implementation to evaluate the penetration speed of vibratory chrrven sheet
Figure 18. Predicted and Observed penetration log at Kortrijk site compared with prehcted log using actual vibratory amplitude
was brought on site and the piles could be vibrated to design depth using the same vibrator.
After reviewing the present state-of-the-art of vibrat o driving, ~ ~ the following suggestions for krther consideration are offered soil mechanics research is needed in the area of large cyclic deformation to better understand and assess the effects of degradation and liquefaction under those extreme conditions, full scale vibratory driving tests, with extensive field monitoring, will be required preferably to reduced scale laboratory tests, which suffer from improper energy dissipation boundary conditions, potential and transferred power of vibrators need to be better defined, as well as modeled for better description of the mechanical behavior of vibratois, peak vibration of the soil surrounding a profile upon vibrator start up does not necessarily imply soil resonance; it can also result from the combination of increasing fi-equency and degrading soil resistance, monitoring of vibrated profiles is recommended with the view to emulate the benefits accrued by a similar practice for driven profiles, and procedures for vibratory loading tests should be developped '
NCES Barkan. D D . (1963). Methodes de vibration clans la construction. Dunod. Fans, 302 p (Plreiich traidatioi? of Originnl 117 1?uwa~i" Vibsometod v StroitelJSlVe. 1960) BBRI (2994) High performance vlbralopv pile h v e r s base on novel electromagnetic actuation systems and Iinproved UiIderstandng of soil dynamics, Progress reports of the BR17I9ElJRAhl research contrucl CTYI-0.561, 19% Ben&arcL R K . (1967), Fluichzation phenoniena ~n soils during vibro-cornpaction and vibro-pile-hvmg and -prrllrng Hanover. NH 1967. 58 pp CIS Alrnzv Cold Xeg1on.s Research and Engineering ~
493
Piles", Research report prepared for BBRI, September, 54p. Holeyman, A. & Legrand C. (1994). Soil Modelling for pile vibratory dnving, International Conference on Design and Construction of Deep Foundations, Vol. 2, pp 1165-1178, Orlando, U.S.A.. 1994. Holeyman, A., Legrand, C., and Van Rompaey, D., (1996). A Method to preQct the driveability of vibratory dnven piles, Proceedings of the 3rd International Conference on the Application of Stress-Wave Theory to Piles, pp 1101-1112, Orlando, U.S.A.. 1996. Howe, C. (1994). HFV Amplitude control vibratory hammers : piling efficiency without the vibration inconvenience, in DFI 94, pp. 2.4.1 to 2.4.10, Proceedings of the FiBh International Conference and Exhibition on Piling aiid Deep Foundations, Bruges, Belgium, 1994. Idnss, I.M., Dobry, R, and Singh. R.D. (1978). "Nonlinear Behavior of Soft Clays during Cyclic Loadmg." J. Geotechnical Engineering Div. ASCE, 104(GT12), pp. 14271447. IREX (1998), VibrofonGage des pieux et palplanches - Etude exploratoire, (by Le Tirant, P., Borel, S., Gonin, H., Guillaume. D. and Longueval, A.) Jonker, G., (1987). "Vibratory Pile Driving Hammers for Oil Installation and Soil Improvement Projects". Proc. of Alineteenth Annual Ofshore Technology Conf:, Dallas. Texas, OTC 5422, pp. 549-560. Kondner, R. L., (1963). "Hyperbolic Stress-Strain Response: Cohesive Soils." .Journal of the Soil Mechanics and Foundations Division, ASCE, Vol. 89; No. SMI, pp. 115-143. Jan. Ligterink, A., van Zandwijk, C., Wddentorp, P., (1990). "Accurate vertical pile installation by using a hydraulic vibratory hammer on the Arboath project". Proc. 2Jnd annual Technology Conf: in Houston, Texas, May 7-10, pp. 315326. Masing, G. ( 1926), "Eigenspannungen und Verfeistigung beim Messing", Proceedings of Second International Congress ofApplied Mechanics~pp. 332-33 5. Mdendorp; P. and Jonker, G. (1988). Prediction of Vibratory Hammer Performance by Stress wave Analysis, Preprint to the 3'd Int. Cone on the Application of Stresswavre Theory to Piles, Ottawa Moulai-Khatir. Re&. ONeill, Michael W., Vipulanandan, C.: (1994). "Program VPDA Wave Equation Analysis for Vibratory Driving of Piles", Report to The U.S.A. Army Corps of Engineers Waterways Experiments Station. Dept. of Civil and Environmental Engineering, UHCE 94-1. Univ. of Houston, Texas, August 1994, 187 pp. Novak, M., Nogami, T., and Aboul-Ella, F. (1978). "Dynamic Soil Reactions for Plane Strain Case", J. Engrg. Mech. Div., ASCE, 104(4), 953-959. NRC (1985). "Liquefaction of Soils During Earthquakes." National Research Council Committee on Earthquake Engneering, Report No. CETS-EE-00 1, Washington, D.C. ONeill Michael W.: Vipulanandan, C., (1989). "Laboratory evaluation of piles installed with vibratory dnvers". National Cooperative Highway Research Program, Report no
316>National Research Council, Washington, DC. Vol. 1 pp. 1-51. ISBN 0-309-04613-0 ONeilI Michael W., Vipulanandan, C., Wong, D., (1990). "Laboratory modelling of vibro-driven piles". Journal of Geotechnical Engineering, ASCE, Vol. 116, No 8: pp. 1190-1209 Roberlson, P.K. and Wride, C.E. (1998). "Evaluating cyclic liquefaction potential using the cone penetration test", Canadian Geotechnical Journal, No. 35, pp. 442-459. Rodger, A.A. and Littlejohn, G.S. (1980). "A study of vibratory h v i n g in granular soils". Geotechnique, Vol. 30, no 3, pp. 269-293. Seed, H.B. and Idnss, I.M. (1970). "Soil Moduli and Damping Factors for Dynamic Response Analyses." Earthquake Engineering Research Center, College of Engineering, University of California, Berkeley, Report No. EERC 70-10. Seed, H.B. and De Alba, P. (1986) YJse of SPT and CPT Tests for Evaluating the Liquefaction Resistance of Sands" Proc. INSITU 86, V A 22 p. Schmid, W.E., (1969). "Driving resistance and bearing capacity of vibrodriven model piles". American Society of Testing and A4aterials Special Techn., Publ. 444, pp. 362375 Schmid, W.E. & Hill, H.T. (1967). "A rational dynamic equation for vibro driven piles in sand". Symp. Dynamic Properties of Earth Materials. New Mexico Universityy:p. 349. Smith E.A.L., (1960). "Pile-dnving analysis by the wave equation". Jourrzal of the Soil A4echanics and Foundatioris Divisions, ASCE, Vol. 86, August 1960. Vilung, K., (1997). "Vibratory dnven pile and sheet piles - a literature survey", Report 3035, Div. of Soil and Rock Mechanics, Rqyal Institute of Technology, Sweden, 75 p. Vucetic, M. and Dobry, R. (1988). "Degradation of Marine Clays Under Cyclic Loading. ASCE Journal of Geotechnical Engineering,Vol. 114, No.2,pp. 133-149. Vucetic, M. and Dobry, R. (1991). "Effect of Soil Plasticity on Cyclic Response." ASCE Journal of Geotechnical Eiigiiieering, Vol. 117, No. 1, pp. 89- 107. Vucetic. M. (1993). "Cyclic Threshold Shear Strains of Sands and Clays", Research Report, UCLA Dept. of Civil Engineering, May 1993. Vucetic, M. (1994). "Cyclic Threshold Shear Strains of Sands and Clays", Paper in print, ASCE Journal of Geotechnical Engineering Wamngton, Don. C.: (1989). "Driveability of Piles by vibration". Deep Foundations Institute 141h Annual members Con$, Baltimore, Maryland, USA pp. 139-154. Westerberg. E.. Massarch, K. Rainer. Eriksson, K., (1995). "Soil resistance during vibratory pile dnving". Proc. to hit. Symposium on Cone Penetration Testing, Linkoping, Sweden, Vol. 3, Report 3.41, pp. 241-250. You4 L.T. (1972). Compaction of sands by repeated shear straining. Journal of the Soil Mechanics and Foundaiioiw Division, 1972, Proc. ASCE, Vol. 98, SM7, pp. 709-725.
494
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) (c> 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
Computation of ground waves due to piling C. L. Ramshaw, A. R. Selby & P. Bettess University cfDurhanz, U K
ABSTRACT: Pre-formed piles are normally installed into the ground using high energy impact or vibratory hammers. Part of the energy is absorbed in advancing the pile, and much of the remaining energy is transmitted into the ground in the form of outgoing vibrational waves. The current study is towards developing a computational procedure to model the outgoing waves, to calibrate the waves against existing site measurements, and then to expand the model by including simple structural forms so as to estimate structure response, including dynamic soil-structure interaction. This paper reports the stage of the work which generates ground vibrations and some calibration against site data. For impact hammers, the analysis has to be tackled in three phases, of hammedpile impact, pileisoil interface effects, and finally the transient outgoing P, S and R waves. For vibro-driving a different procedure is required. A two stage approach makes an estimate of rigid body oscillation of the pile within a localised soil framework, and then the boundary disturbance is applied to an analysis of the wider soil area for computation of the sinusoidal wave disturbances. Two case studies are presented. must first be calibrated against a substantial database of site records of “green-field” vibrations (Uromeihy, 1990), which is the subject of this paper.
1. INTRODUCTION Pre-formed piles, of steel, concrete or timber, are widely used for transmitting building loads from ground surface down through weak soils to more competent soil or rock strata, while interlocking sheet piles are used for temporary or permanent retaining walls. Pile installation is achieved using high energy vibratory or impact hammers. Some energy is lost into the ground as outgoing P, S or R waves. These vibrations may disturb neighbours (Head & Jardine, 1992), and cause cosmetic damage to nearby structures (Wiss, 1967, Selby, 1991, Todd, 1994). Analysis of risk from vibrations has been attempted by empirical estimates of free ground vibrations and arbitrary vibration thresholds for building types (Attewell & Farmer, 1973, Eurocode 3 part 5 , 1996) with some recognition of frequency effects. Structural analysis for vibration enhancement has been attempted in isolation.
The use of finitehfinite elements to generate ground waves has been demonstrated (Ramshaw et al, 1998), using the package ABAQUSO, and a relatively simple energy transfer at the pile-soil interface. This has now been extended, so as to model the energy input from the hammer, the stress waves in the pile, and the subsequent outgoing ground waves. However, because of the complexity of the system, and the need to use small nodal spacings (x
An ongoing study is directed towards generation of pile-soil models leading to representative ground vibrations from driving, to be followed by pile-soilstructure models which will incorporate dynamic soil-structure interaction. However, the procedure
495
conceptual modei includcs lumped rani an3 anvil mcwes separated by a damped cushion which is represented by a spring and dashpot in parallel. These rest on the pile which is modelkd by a dashpot, see Figure 1.
Calibration against real site data is difficult because of the many physical parameters involved. However, good agreement of both vertical and radial surface vibrations, defined by velocity-time traces, has been achieved for a number of site records. These include bearing piles installed by vibrodriver at the second Severn crossing and impact driving on the M66 near Manchester.
1.1 General aims of paper At present, guidance on vibrations resulting from various types of pile driving is somewhat limited. Various standards exist, but these are empirical and in some cases are conflicting. While empirical equations have been derived for estimation for peak surface vibrations, detailed computation of- the ground waves is lacking. This paper presents simplified methods using finite element techniques for predicting surface vibrations resulting from both vibratory and impact pile driving. The methods have been designed to be used as a preliminary design tool for the prediction of ground surface vibrations where site data are sparse. They are not intended to be rigorous geotechnical analyses.
Figure 1. Hammer impact model. in this stage it is sufficient to represent the pile by a dashpot of impedance equal to that of the pile, defined as =-E,, .A,, (1) P
where E, is the elastic modulus of the pile, A, is its cross sectional area, and cp is the axial pile wave velocity. Deeks & Randolph showed that the behaviour of the hammer impact model is governed by three dimensionless parameters: the cushion stiffness, kc", the anvil mass, ma", and the cushion damping, cc", defined as follows:
2. SIMULATION OF IMPACT PILING The procedure developed here breaks the overall problem into three stages, comprising: A hammer impact model to simulate the force imposed onto the pile head A model of the propagation of the impact waves down the pile shaft (soil response modelled by springs and dashpots) A model of the surrounding ground to simulate the outgoing ground waves. There are two advantages of breaking down the computations in this way. Firstly, each parameter or group of pmimeters can be isolated, and rapid parametric studies can be used to ascribe values. Secondly, more efficient computations are possible; a finite element (FE) model of the pile requires very small nodal spacings (x
* k , .m, k , =?
z
(3) c1 =c(
c
z
(4)
where k,is the stiffness of the cushion spring, n1,. is the mass of the hammer ram, m, is the anvil mass and cc is the damping in the cushion. The second stage of the simulation is the propagation of compression waves down the shaft of the pile. This is done using a model, based upon that developed by Deeks (1992), shown in Figure 2. The pile is modelled by 8-noded axisymmetric finite elements, the toe response is modelled by a mass-spring-dashpot model proposed by Deeks & Randolph (199.51, and the shaft/soil boundary is a spring-dashpot system similar to that developed by Decks (19921, for transmitting shear waves independent of frequency.
The hammer impact model used here is that presented by Deeks & Randolph (1993). The 496
The model for the toe resistance was developed by Deeks from that proposed by Wolf (1988). The complex stiffness of the Wolf model was matched to the results of FE analyses for several values of Poisson’s ratio, v, taken across a large range of dimensionless frequencies. The model parameters can be conveniently expressed in dimensionless terms. The frequency of the applied load, 0,is non-dimensionalised with respect to the radius of the pile base R, and the shear wave velocity of the soil, cs. &.R a,s = C,S
where c, = d(G/p). The static spring stiffness, k, is taken as k = -4.GR
(6)
1-v
The non-dimensionalised mass and damping parameters %, al, POand P 1 are defined as: ,
\7
/
\2
Figure 2. Pile model for axial waves and shear transfer.
The third and final stage of the procedure is to impose the displacement-time functions onto a large FE/IE axisymmetric mesh of the surrounding soils, see Figure 3. The pile response is transferred to the ground model by way of an unrestrained “false” pile made up of axisymnietric finite elements.
Further, Deeks proposed the following expressions to accommodate variations in Poisson’s ratio, v, as: a, = 0 . 6 3 - 3 . 6 ~ + 6 ~ ’ , (94
/j, = 1 . 5 8 - 1 0 . 3 ~ + 1 9 ~ ~
(9b)
The response of the pile shaft to the impact wave is modelled using a new frequency independent transmitting boundary for axisymmetric shear waves derived by Deeks (1992). This boundary is equivalent to viscous dashpots with a distributed damping constant of p.cs (identical to a viscous boundary) and a distributed spring constant of G/2rb, where p is the density of the soil, cs is the shear wave velocity in the soil, G is the shear stiffness of the soil and rb is the pile radius. The velocity-time function derived from the hammer impact model is then imposed onto the simplified pile model described above as a force on the pile head (after multiplication by the pile impedance). The output results are the displacement-time functions at the shaft interface nodes, and at the base.
The pile-soil interface is modelled using a surfacebased contact simulation, with slip controlled by a Coulomb friction model with p=O.1 (Mabsout & Tassoulas, 1994). In order to simulate the horizontal stresses on the pile from the soil, the “false” pile is moved laterally into the soil by a predetermined distance. (This avoids the problem of rigid body motion.) Once the horizontal stresses have equilibrated, the vertical displacements computed from the pile model are applied to the pile shaft nodes of the “false” pile and the soil nodes immediately under the pile toe in a dynamic analysis with time steps of 0.001secs. For simplicity, the soil is modelled as an elastic medium as it has been shown that at a short distance from the pile (about one pile radius) most of the energy is propagated in the form of elastic waves (Massarsch 1992). Parametric studies based on the arrival times of ground waves at various distances from the vibration source indicate that the dynamic
497
soil stiffness is much greater than the static stiffness. The non-linear behaviour of soils with very high stiffness at very small strains has often been observed (Matthews et al, 1996). The infinite elements around the periphery of the FE mesh are required to avoid spurious wave reflections. 3. IMPACT PILE DRIVING AT THE M66 A valuable example of site data which include both pile head records and ground surface vibrations was obtained during dynamic testing of cast in situ piling for bridge foundations on the M66 motorway near Manchester.
Figure 3. Axisymmetric finite elemenvinfinite element mesh of the surrounding ground.
1.0
[ A 0
--3
1 0.3
0.0
9
-0.5
d
i
Depth (m) 0 - 10.5
Soil type
E (Pa) 13 x 106
V
-1.9
P (kgh') 1970
Firdstiff 0.35 CLAY Dense 2 4 x 106 0.30 2100 10.5>> SAND A smalI'strain stifffiess oi-150-x lU^-%a was usedto model the outgoing waves.
-1.0
P 4 -2.0
-2.5
I
-3.0 0.
I
I
5.
I
10.
15.
I
I
20.
25.
I 30.
I 35.
I
I
40.
03
TOTAL T I I E ( n )
[X
90.
l O -3
]
Figure 4. Computed and measured (dashed line) displacement-time functions at the pile head.
In considering the impact of the hammer onto the pile head, a parametric study was conducted to match values of k,*, m,* and cc*. The values which gave the closest match for pile head response were kc* = 0.96, ma* = 0.90 and cc* = 0.10. The normal ranges for these parameters are kc*= 1- 1000, m,*=0.1-0.7 and cc* = 0.0-1.0. The values for the spring stiffness and anvil mass are slightly outside these ranges in this case due to the special pile cap which is required for a SIMBAT analysis.
0.4
[do
I
I
6 2
-0.4
I
-0.8
-
I
i3
ge
I
-----___-____-__
0.0---
2
1
=]
-1.2
I II
-
I 11
8
-1.6
-
IJ I' I
The computed displacements and forces at the pile head show good agreement with the field data as shown in Figures 4 and 5.
-2.0 0.00
The force-time function computed from the hammer impact model was then imposed onto the pile head 498
t
, 0
.oz
1 0.04
I 0 .OS
I 0.08
0.10
Taken overall, and recognising the ground variability, the correlations between the equivalent velocity-time traces are generally encouraging. The main discrepancy appears to be due to a P-wave from the pile shaft which may be caused by an eccentric strike. This is not simulated in the model.
for the second stage of the computational procedure in order to determine the displacement-time functions on the pile shaft. In the third stage, the generated displacement-time functions were imposed onto the shaft nodes of the “false” pile and onto the soil nodes under the pile toe in the FE/IE mesh. The resulting radial and vertical ppv’s at horizontal stand-off distances of 5.5m, 10m and 16.5m are compared with the measured vibrations in Figures 6 and 7 respectively.
4 SIMULATION OF VIBRO-DRIVING OF PILES The process of pile installation by vibratory driver is fundamentally different from impact driving, and
Figure 7. Vertical ppv’s at 5Sm, 10m and 16.5rn respectively. (Measured = dashed line).
Figure 6. Radial ppv’s at 5Sm, 10m and 16.5m respectively. (Measured = dashed line)
499
so modelling of the process must also be different from the above procedure for estimating vibrations from impact hammers. In practice, the method is best suited to saturated fine sands, and liquefaction is induced by the vibration of the pile, which then sinks into the ground under the combined weight of driver plus pile. Many site records have shown that ground vibrations resulting from vibro-driving are continuous and sinusoidal in character, in marked contrast to the transient nature of vibrations caused by impact piling. The technique for modelling pile response to cyclic excitation, and outgoing ground waves, is now proposed as a two-stage procedure. In stage one, the objective is to establish a model for rigid body vertical oscillation of the pile in response to the cyclic excitation of the vibro-driver. This is done by the use of rigid axisymmetric elements for the pile shaft, a limited axisymmetric F E / E mesh representing the soil, and a mechanical model for toe reaction based on a spring and a dashpot in parallel, proposed by Lysmer & Richart (1966), see Figure 8. The pile-soil interface comprises a two-surface contact, with a Coulomb friction model. A static computation is made for the soil/shaft normal stresses due to geostatic stress. A dynamic analysis is then conducted to set up steady state response of the rigid pile to the cyclic excitation from the hammer of F(t) = m.e.w2.
The spring and dashpot constants are given by k = -4.GR (1 1)
I-v
kR ~=0.85-
The time-stepping computation generally stabilises into steady state after 1.0-1.5s with time increments of 0.001s. mesh is established, In the second stage, a €%/E similar to that in Figure 3, and the sinusoidal displacements of the shaft nodes and the base, computed in stage one, are imposed in the form of Fourier series. It is essential in an on-going analysis that there is no spurious reflection from an artificial boundary, and this is achieved by the infinite elements around the periphery of the model. Material damping is applied in the form of a Rayleigh damping ratio given by
=m
= undamped frequency of where U,, vibration. A typical damping ratio of 5% has been suggested by Massarsch (1992) for the elastic range of soil deformations.
The output from the computation gives ground surface vibrations as sinusoidal vertical and horizontal particle velocities. These can then be calibrated against site data.
(10)
where m is the total mass at eccentricity e, rotating at w rad/s.
5. VIBRODRIVING AT THE SECOND SEVERN CROSSING A detailed parametric study of vibro-driving simulation has been undertaken using data obtained from the installation of 1050mm diameter casings for one of the approach bridges to the second Severn crossing. Borehole records for the site indicate soft to firm clay to a depth of 13.8m overlying firm to stiff marl. The casings were installed to a depth of 15.5m using a PTC 50H3 vibratory hammer with an eccentric moment of 50m.kg. The resulting vibrations were monitored by geophones placed on the ground surface at distances of 5.7m, 14.5m and 32.9m. The two stage procedure was followed and parameters were adapted to produce a match between estimated surface particle velocities and
Figure 8. Simplified model for vibro-driving simulation. 500
Figure 9. Radial ppv’s at 5.7m, 14.5m and 32.9m respectively. (Measured = dashed line)
Figure 10. Vertical ppv’s at 5.7m, 13.5m and 32.9m respectively. (Measured = dashed line)
the measured site data. The influence of the various parameters were found to be as follows: The distribution of shaft friction with depth affects the proportions of the ppv’s with relation to each other at various stand-off distances. This may be due to the interaction of the ground waves propagated from the shaft and the toe as observed by Attewell et al, 1991. The magnitude of the dashpot constant affects the magnitude of the ppv’s Doubling the spring stiffness only causes a small reduction in the magnitude of the ppv’s
Increasing the material damping from 5% to 8% only causes a small reduction in the ppv’s The computed ppv’s at the ground surface for a casing installed to a depth of 1l m at a frequency of 15.8 Hertz and a penetration speed of about 0.18mds are shown in Figures 9 and 10. These computations assumed a small strain stiffness of 50 x 106 Pa for the soft/firm clay with v = 0.4 and p = 1750 kg/m3. The computed ppv’s show good agreement with both the radial and vertical ppv’s measured on site over a considerable stand-off distance. 501
6. CONCLUSIONS Simplified procedures have been developed for the computation of ground surface vibrations caused by impact driving and vibratory driving of pre-formed piles. The procedures have been shown to be viable by comparisons with site measurements at two locations, provided that appropriate values for certain parameters can be identified. Further pile driving simulations are currently being undertaken for a large number of sites in the U.K. (Uromeihy, 1990) in order to refine the various parameters, and ultimately, to develop general guidance on the most appropriate parameters for each set of circumstances. The procedures do not require detailed knowledge of site conditions and are therefore particularly useful as a preliminary design tool. The methods are now established for a further phase of the work, which will be to add typical building forms to the finitdinfinite element meshes, for estimation of building response with full soilstructure interaction.
ACKNOWLEDGEMENTS. The funding of this work by EPSRC through Grant No. GNLl8679 is gratefully acknowledged.
REFERENCES. Attewell, P.B. & I.W. Farmer 1973. Attenuation of ground vibrations from pile driving. Ground Engineering. 6.4:26-29. Attewell, P.B., A.R. Selby & A. Uromeihy. 1991 Non-monotonical decay of ground surface vibrations caused by pile driving. Earthquake, Blast and Impact ed SECED. London Elsevier:463-48 1 Deeks, A.J. 1992. Numerical analysis of pile driving dynamics. PhD thesis. Uni. of W. Australia Deeks, A.J. & M.F. Randolph 1993. Analytical modelling of hammer impact for pile driving. Int. J. f o r Num. and Analytical Methods in Geomechanics. 17:279-302 Deeks, A.J. & M.F. Randolph 1995. A simple model for inelastic footing response to transient
loading. Int. J. for Num. arzd Analytical Methods in Geomechanics. 19:307-329 Eurocode 3 1996 Design of steel structures, Pt 5 Piling CENflC 250/SC3/PT5 Head, J.M. & F.M. Jardine 1992. Grouizd-borne vibrations arising from piling. TN 142 London: CIRIA Lysmer, J. & F.E.Richart 1966. Dynamic response offootings to vertical loading. J. of the S.M &F Div. Proc. A.S.C.E 92; SM1 65-91 Mabsout, M.E. & J.L. Tassoulas 1994. A finite element model for the simulation of pile driving. Int. J. for Nunz. Methods in Eizg. 37:257-278 Massarsch, K.R. 1992. Static and dynamic soil displacements caused by pile driving. Application of Stress- Wave Theory to Piles, ed Barends, F.B.J. Balkema Rotterdam : 15-24 Matthews, M.C., V.S. Hope & C.R.I. Clayton 1996. The use of surface waves in the determination of ground stiffness profiles. Proc. Instrz. Civ. Erzgrs Geotech. Eizgng, 1 19:84-95. Ramshaw, C.L., A.R. Selby & P. Bettess 1998. Computation of the transmission of waves from pile driving. Ground dynamics and ninn made processes. ed Skipp, B. 0. T.Telford Publ. London : 1 15- 128. Selby A.R. 1991 Ground vibrations caused by pile installation Proc of 4th Int Conf on Piling arzd Deep Foundations. Stresa DFUTespa :497-502 Stain, R.T. 1992. SIMBAT - a dynamic load test for bored piles. Piling: European practice and worldwide trends. T Telford London: 198-205 Todd, A. 1994. Berwick river walls - an unique scheme to safeguard an historic part of Berwick upon Tweed. ICE papers cornpetition. London Uromeihy, A. 1990. Ground vibration measurements with special reference to pile driving. PhD thesis. Uni. of Durham. UK Wolf, J.P. 1988. Soil-Structure-Interaction Analysis in Time Domain. Prentice-Hall, New Jersey. Wiss J.F. 1967. Damage effects of pile driving vibrations. Highways Res B. 155 : 14-20
502
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
The effect of pile impedance on energy transfer to pile and ground vibrations M.R.Svinkin VibraConsult, Cleveland, Ohio, USA
B.C. Roth Munico Consultants, San Francisco, Cali$, USA
W. R. Hannen WJE,Northbrook, Ill., USA
ABSTRACT: Pile impedance effects energy transfer to the pile and surrounding soil in different ways. Increasing pile impedance increases force and decreases velocity at the pile head and might insignificantly affect energy transferred to the pile. Displacement piles affect predominantly ground vibrations in the proximity of pile driving. Effects of displacement piles are more substantial on amplification of ground vibrations than those from piles with pile impedance reduction.
and pile driveability on induced ground vibrations.
1 INTRODUCTION Pile driving operations are powerful sources of construction vibrations which may adversely affect nearby buildings, harmfully affect sensitive electronics and disturb people. Therefore, certain concerns frequently arise about possible vibrations which might be generated from pile driving activities, especially impact pile driving. The effect of pile driving on ground vibrations depends on the energy transferred to a pile, pile type, soil conditions and the distance from the source. Pile driving vibrations can be a problem with all kinds of pile drivers and all types of driven piles (Woods 1997). However, installation of various piles at a construction site generates different ground vibrations at the same distance from driven piles. Substitution of certain pile types for others can be used as partial mitigation of potentially high levels of ground vibrations. The effect of pile type on induced ground vibrations is at least in part attributable to pile impedance. Considerable data have been collected, analyzed and published with respect to the pile impedance effect on force-energy transfer to piles and ground vibrations generated during pile driving, e.g. Parola (1970), Peck et al. (1974), Heckman & Hagerty (1978), Massarsch (1992), Abe & Thendean (1 996), Woods (1997), Svinkin (1999), Brettmann & Cotton (1999) and others. This paper presents further consideration of energy transfer to piles, assessment of the effect of pile type
503
2 FORCE TRANSFER IMPEDANCE
VERSUS
PILE
Pile impedance is a significant factor in the transfer of dynamic longitudinal force into the pile and from the pile into the surrounding soil. The reason for this is because, independently of how much energy may be applied to the pile head, the force that can be transmitted down the pile is limited by the impedance (Peck et al. 1974). Pile impedance characterizes the pile ability to overcome the soil resistance to pile penetration and develop required capacity. Pile impedance, Z, is defined as
Z
EA/c
=
(1)
where E = modulus of elasticity of pile material; A = pile cross-sectional area; c = longitudinal stress wave velocity. The maximum force, F, measured at the pile head during driving can be expressed as (Svinkin 1994) F =
1:
2-W,
(2)
where L = pile length; W, = energy transferred to the pile; remaining parameters are the same as above, According to equation (2), force transmitted to the
Figure 2. Effect of steel pile (a) and concrete pile (b) impedance on ETR and refusal capacity (after Abe & Thendean 1996)
Figure 1. Effect of hammer (a) and pile (b) cushion stiffness on ETR (after Abe & Thendean 1996)
504
pile is proportional to the square root of pile impedance and also depends on pile length, velocity of stress wave propagation in the pile and energy transferred to the pile. Wave equation analysis simulation (Abe and Thendean 1996) confirmed that the mobilized pile capacity increases with increasing pile impedance (Figure 2).
3 VELOCITY IMPEDANCE
TRANSFER
VERSUS
Table 1. Energy transferred to concrete and steel piles
=.
PILE
J
(3)
2-w, z"L
where all parameters are the same as above. Equation (3) is similar to equation (2), but the peak particle velocity of the source is inversely proportional to the square root of the pile impedance.
4 ENERGY IMPEDANCE
TRANSFER
VERSUS
PILE
The energy transferred to the pile head is computed by integrating the product of force and velocity records (Hannigan 1990). According to equations (2) and (3), pile impedance effects force and velocity in opposite ways. Obviously, it can be expected that pile impedance has negligible impact on the energy transferred to the pile. The performance of a pile driving system or hammer performance, is important to the proper execution of pile driving. The measure of hammer performance obtained from dynamic pile monitoring is the ratio of the transferred energy to the rated hammer energy. This ratio is known as the energy transfer efficiency, ETR. Intensive study of hammer performance was made by Abe & Thendean (1996). They summarized statistical results of ETR compiled from Pile Driver Analyzer outcomes at numerous projects and provided wave equation parameter studies. Six combinations were analyzed for single acting air, hydraulic, and diesel hammers on steel and concrete piles. ETR mean and standard deviation for considered six hammer-pile systems are shown in Table 1. It can be seen that average ETR is 20-
Hammer
Type
Type
Concrete
Steel
Svinkin (1996) derived an equation to calculate the peak particle velocity (PPV) at the pile head, V, in advance of pile driving
v
Pile
Energy transfer efficiency Average ETR
Standard Deviation
ETR Ratio
%
%
Air
40.4
12.0
0.79
Hvdraulic
5 1.2
15.9
1.oo
Diesel
24.8
6.8
0.48
Air
50.2
11.5
0.61
Hvdraulic
82.0
15.1
1.oo
Diesel
34.3
34.3
0.42
34% lower for concrete piles than those for steel piles, but ETR ratios between different hammers for concrete piles are similar to those for steel piles. This is an indicator of almost adequate hammer conditions for driving steel and concrete piles but with lower transferred energy to concrete piles on account of low pile cushion stiffness. The obtained effects of hammer and pile cushion stiffness and pile impedance on ETR are displayed in Figures 1 and 2. For steel and concrete piles, the hammer cushion stiffness does not affect ETR in the commonly used stiffness range above 1,000 kN/mm (Figure la). Therefore, Figure 2a demonstrates the reasonable effect of steel pile impedance on ETR without implication of hammer cushion influence. For steel piles in the commonly used impedance range of 100-3000 kNs/m, pile impedance affects negligibly on ETR for the air and hydraulic hammers but decreases ETR about 30 % for the diesel hammers in the impedance range of 1001000 kNs/m. The pile cushion stiffness significantly effects ETR for all hammers at lower stiffness from 100 to 1000 kN/mm (Figure lb). Probably a share of pile cushion stiffness influence was not completely excluded from the effect of concrete pile impedance on ETR in Figure 2b where considerable ETR variation is observed at the pile impedance range of 300-3000 kNs/m. In this impedance range, ETR decreased 20, 30, and 50 % for air, hydraulic, and diesel hammers, respectively. These results are different from those obtained for steel piles though similar results for both pile types in the same pile impedance range could be expected. It seems that the effect of concrete pile impedance on ETR ought to refine from pile cushion influence with normalization of pile cushion stiffness. Thus, steel pile impedance does not actually affect
505
ETR for air and hydraulic hammers and results in 30 % reduction of ETR for diesel hammers. The effect of concrete pile impedance on ETR is not clear so far.
Table 2. Data for tested piles Designation
Description
Impedance kNs/m
Toe Area crn2
Depth rn
TP-1
HP 362~1.71 rnrn, kN/m
910
222
13.41
TP-2
HP 362~1.71 m, kN/m
910
222
14.94
TP-3
HP362~1.71
910
222
15.54
561
994
7.01
~~
5 GROUND VIBRATIONS VERSUS PILE TYPES The effects of some factors on ground vibrations are obvious. For example, increasing the energy transferred to the pile increases ground vibration intensities, and increasing the distance from a driven pile results in reduction of ground vibration intensities. Influence of other factors on ground vibrations may not be so apparent. It is valuable to know how a change of pile type affects ground vibrations during pile installation. Pile impedance should be considered as a factor affecting the intensity of ground vibrations induced by pile driving. Heckman and Hagerty (1978) and Massarsch (1992) pointed out the important effect of the pile impedance on the peak ground velocity and developed a graph for the amplification factor of ground velocity as a function of pile impedance. The lower the pile impedance, the higher the induced ground vibrations. By way of illustration, they showed that a reduction of the pile impedance from 2000 to 500 kNs/m could increase the peak ground velocity by a factor of 8. According to equation ( 3 ) , for the referenced impedance range, the expected amplification of the peak pile velocity and the peak ground velocity can only be 2. In the following case histories velocity 01ground vibrations were measured from driven piles with different impedances and results obtained were used for assessment of the role of pile impedance in amplification of ground vibrations.
mrn. Id\l/rn
TP-4
356 x 13 mm Pipe Pile Shell w/conical DOint
TP-5
356 x 13 I I I I I I Pipe Pile Shell w/conical point
561
994
7.62
TP-6
406 x 406 mrn Precast Prestressed Concrete
1618
1652
7.01
TP-7
406 x406 rnrn Precast Prestressed Concrete
1618
1652
6.01
365 x 4.5 mm Fluted Pile Shell
204
-500
5.18
~
TP-8
5.1 Case 1 Eight different piles were tested within an Advance Test Pile Program performed at a site in Phoenix, Arizona, in connection with the design and construction of the 1-10 West Papago/Inner Loop Freeway (Hannen & Linehan 1984). A pile description is presented in Table 2. A pile and geophone layout is depicted in Figure 3. The soil consisted of about 1.4 m of clayey sand followed by about 2.4 m of sandy clay followed by about 5.6 m of dense clayey sand underlain by about 2.8 m of sandy clay deposited on 12.2 m of sand, gravel and cobble. The water table was not encountered during 506
Figure 3. Layout of TP-1 - TP-8 piles and geophones
soil boring at the site. A MKT DE-70B diesel hammer was employed for initial driving and restrikes. A hammer stroke changed between 1.5 and 2.6 m. The measured PPV of ground vibrations were adjusted to the maximum hammer stroke. Ground vibrations were measured at distances of 9.1, 30.5, 152.4 m from group 1 of TP-1, TP-3, TP-5, TP-7 piles and at distances of 7.0, 28.3,
Figure 4. Maximum PPV of ground vibrations at distances of 9.1, 30.5 and 152.4 m from TP-1, TP-3, TP-5, and TP-7 driven piles
Figure 5 . Maximum PPV of ground vibrations at distances of 7.0, 28.3 and 150.2 m from TP-2, TP-4, TP-6 and TP-8 driven piles
150.2 m from group 2 of TP-2, TP-4, TP-6, TP-8 piles. Maximum PPV of radial and vertical ground vibrations are depicted in Figures 4 and 5 . Radial ground vibrations had higher intensity in comparison with other components of ground vibrations. Transverse horizontal ground vibrations were not considered in the study because their level was 3-5 times smaller than level of radial horizontal ground vibrations. Four different pile types were tested. H-piles were non-displacement piles and the rest of piles were displacement ones. The analysis of soil vibration records, measured at the same distances from each group of piles with different impedances and driven by the same hammer, revealed some interesting observations.
In pile group 2, the effects of TP-4 pipe pile shell and TP-8 fluted pile shell were major for radial and vertical components at all distances from driven piles.
In general, piles being driven as displacement piles generated the greatest ground vibrations, but there effects were varied. In pile group 1, the effect of concrete TP-7 pile was dominant for radial component at distances of 30.5 and 152.4 m and for vertical component at all distances from the source. e
Driving non-displacement piles induced lower ground vibrations, but the effects of TP-1 and TP-3 H-piles were dominant for radial and vertical components at distance of 9.1 and 30.5 m from driven piles at the pile penetration length between 6.4-7.3 m with the pile toe in dense clayey sand. @
Effects of displacement piles were more substantial on amplification of ground vibrations than those from piles with impedance reduction. @
Ground vibrations increased due to pile penetration depth of 6.4-7.3 m and then started to decrease. This effect is related to soil conditions rather than to pile length. Restrikes were made for TP-2 - TP-6 piles. Restrikes did not actually increase the level of ground vibrations. @
Figure 6 . Maximum PPV of ground vibrations at various distances from driven precast pile and H-pile
Figure 7. Maximum PPV of ground vibrations at disranxs of 7.0 and 9.1 rn versus soil resistance to pile driving
5.2 Case 2
6 GROUND VIBRATIONS VERSUS RESISTANCE TO PILE PENETRATION
Pile driving of four piles and measurements of ground vibrations were made at a site in San Francisco, California, to determine the various seismic effect of installation of two pile types on existing buildings (Municon Conschants 1999). The test piles included two 305x305 mrn precast concrete piles uith lzngth of 18.3 111, Z=910 kNs/m, and two HP 3 0 8 ~ 1 . 0 8(mm, kN/m) with length of 21.0 m, 2=577 kNs/m. The soil consisted of about 11.3 m of loose to dense sand followed by about 1.5 m of sandy clay underlain by a bearing layer of dense to very dense silty sand. The water table was at the depth of 6.1 m. For all piles tested, a predrill was made to the depth of 4.6 m. A Delmag D30-32 diesel hammer was used for pile driving. The hammer stroke was 3.2 m. The soil resistance to pile penetration was mostly in the range of 530 bU0.3 m except the end of driving for concrete piles with the soil resistance between 60100 bY0.3 m. Ground vibrations were measured at various distances from driven piles. The maximum PPV of ground vibrations measured during driving of two pile types are shown in Figure 6. Installation of displacement concrete piles generated larger ground vibrations at distances up to 20 m from the piles than those induced from driving H-piles. A PPV ratio of compared ground vibrations was three at a distance of 3 m from driven piles. Even though the impedance of concrete piles was 1.6 times the impedance of H-piles, the seismic effect of displacement pile installation was more significant in the proximity of pile driving.
SOIL
The soil resistance to pile penetration, measured in blow count per 0.3 ni, changes during pile driving. The higher blow count corresponds with the greater soil resistance. The blow count greater than 100 b110.3 m is considered as high blow count. For various penetration resistance, maximum PPV of horizontal ground vibrations at distances of 7.0 and 9.1 m from driven piles are shown in Figure 7 (Case 1). It can be seen that the high blow count resulted in increased ground vibrations at the pile penetration depth between 4.27 and 7.62 m but did not affect ground vibrations at the greater penetration depth. Analogous results were observed at distances of 28.3 and 30.5 m from driven piles. Holfoway et al. (1980) reported similar phenomena that driving to refusal of concrete, steel shell and H piles at depth about '7.0-7.6 m increased radial and vertical particle velocities by factors 1.2-3.9 ar distances of less than 10 m from driven piles.
7 DISCUSSION OF WSULTS 7.1 Pile Impedance The greater the pile impedance, the greater the pile capacity and the greater the dynamic force that can be transferred to the ground. Pile impedance affects the intensity of ground vibrations in two ways at the same time. On the one hand, according to equation (2), increasing pile impedance increases the force transmitted to the pile 508
and suimxmding ground, but on the other hand. according t ~equation ! (31, increasing pile impedance decreases PPV of pile and ground vibrations. An incress of h m m e r energj- magnifies ground vibrations until the pile impedance allows to increase the force transmitted to the pile and the surrounding soil. The impedance affects in opposite ways in both equations, and the pile impedance effect on the intensity of ground vibrations is not o b l h s . Presented case histories revealed that pile impedance cannot he always used as a predictor of intensit? of ground vibrations. In Case 1, precasc concrete piles and pipe pile shells ~ i t closed h ends generated similar ground T ibrduons though the latter had pile impedance 2.9 rimes less. In Case 2, che dri\.ing of precdst concrete piles induced PP\’s of ground \ lbrarions IZCX rhe prics three times more tlian the driving oc the H-piles eke11chough the latter had pile impedance 1.6 times smaller.
7 . 2 Displacernmr piles Piles being driven as displacement piles transfer p a t e r force to the surrounding ground. The more pronounced effects of displacement piles on ground \ ibrations occurs predominantly at distances less t h m 10 m from pile driving.
7 . 3 Pile-soil load transfer The dynamic force transferred to the pile at each hammer blow is then transferred to surrounding ground through the pile toe as a concentrated load and through the pile shaft as a non-uniformly distributed load. During driving. the various soil resistances to pile penetration are developing as the pile penetration depth is increasing. Generally speaking, dynamic loads transferred from the pile shaft and toe to the ground should be increasing with increasing pile penetration depth. However, this does not always occur. In Case 1. the intensity of ground vibrations was independent of the pile penetration depth and depended on soil properties. Maximum PPV of ground vibrations were recorded during driving all piles through dense clayey sand at depth 6.4 to 7.6 ni. Pile-soil load transfer is realized by means of both concentrated loads from the pile toe and distributed loads generated along the pile shaft. However, for many practical purposes, a pile can be deemed as a
point source of vibrations. This assumption might be correct at dismce. D, from driven pile (Svinkin 1996)
Where c, = velocity of surface wave propagation in the ground; remaining parameters are the same as above.
8 CONCLUSIONS File impedance affects force and velocity at the pile head in opposite ways. Therefore it can be expected that pile impedance does not significantly affect energy transferred to the pile. Analysis of published experimental data and wave equation simulation cunfiims that the impedance of steel piles has a negligible effect on ETR for air and hydraulic hammers. Additional research IS needed to clarify the impedance effect of concrete piies on ETR for all hammer modek. An increase of hammer energy magnifies ground vibrations until the pile impedance allows the force transmitted to the pile and the surrounding sod to increase. Since the impedance affects force and velocity in opposite ways at the same time, the p i e impedance effect on the intensity of grourici vibrations is not obvious. The more pronounced effects of displacement piles on ground vibrations occurs predominantly at distances less than 10 m from pile driving. In many situations, a driven pile. can be considered as a point source of vibrations.
ACKNOWLEDGEMENT The writers are thankful to Mr. Anatol Longinow, Project Manager at Wiss, Janney, Elstner Associates, Inc. for valuable comments.
REFERENCES Abe, S. & G. Thendean 1996. Hammer performance evaluation. In F. Townsend, M.Hussein & M. McVay (eds), Proceedings of the Fifth International Conference on the Application of Stress Wave Theory to Piles, Orlando, Florida: 9 12-927.
509
1997. Dynamic efsects of pile installation on adjacent structures. Synthesis
Brettmann,T. & B. Cotton 1999. Case history: estimating ground vibrations caused by pile driving. DFI 24th Annual Members ’ Conference,
Woods R.D.
Report, National Cooperative Highway Research Program NCHRP Synthesis 253, Washington, D. C. : National Academy Press.
Decades of Technology - Advancing into the Future: 161-171. Hannigan, P. J. 1990. Dynamic monitoring and analysis of pile foundation installations. Sparta, New Jersey: DFI. Hannen, W.R. & P.W. Linehan 1984. Vibration monitoring during the I-10 West Papagohner Loop Advance Test Pile Program. WJE Report 830535, WJE Engineers, Northbrook, Illinois. Heckman, W.S. & D.J. Hagerty 1978. Vibrations associated with pile driving. ASCE Journal of the Construction Division, 104(C04): 385-394. Holloway, D.M., Y. Moriwaki, E. Demsly, B.H. Moore, & J.Y. Perez 1980. Field study of pile driving efsects on nearby structures, Special Technical Publication, Minimizing Detrimental Construction Vibrations, Preprint 80-175: 63-100. New York: ASCE. Massarsch, K.R. 1992. Keynote lecture: Static and dynamic soil displacements caused by pile driving. In F. Barends (ed.), Proceedings of the Fourth
International Conference on the Application of Stress- Wave Theory to Piles: 15-24, Rotterdam: Balkema . Municon Consultants 1999. Pile driving vibration energy attenuation survey - Reports, San Francisco, California. Peck, R.B., W.E. Hanson & T.H. Thornburn. 1974. Foundation Engineering, 2nd ed., New York: John Wiley & Sons. Parola, J. F. 1970, Mechanics of impact pile driving, Ph.D. thesis, University of Illinois, Urbana, Illinois. Svinkin, M.R. 1994. Influence of pile parameters on pile driveability. Proceedings of International
Conference on Design and Construction of Deep Foundations, Orlando, Florida,II: 1150-1164: FHWA. Svinkin, M.R. 1996. Velocity-impedance-energy relationships for driven piles. In F. Townsend, M .Hussein & M. McVay (eds), Proceedings of the
Fifth International Conference on the Application of Stress Wave Theory to Piles, Orlando, Florida: 870-890. Svinkin, M.R. 1999. Prediction and calculation of construction vibrations. DFI 24th Annual
Members’ Conference, Decades of Technology Advancing into the Future: 53-69.
510
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
etemination of modulus of subgrade reaction in a pile with a vibrating apparatus M.Hilmi Acar Department of Civil Engineering, University of S e l p k , Konya, Turkey
ABSTRACT: The pile foundations are mostly under the influence of the horizontal forces due to the earthquake, wind and wave forces. The soil properties is considerably different in this loaded medium. It is necessary to know the value of modulus of subgrade reaction which has an important place in the designing and calculation of the pile foundations. This study is aiming to present by the numerical model the suitability of the value of modulus of subgrade reaction, determined experimentally through a vibrating apparatus placed on a free headed concrete filled steel pile.
d‘w EI 7 + KW 10
1 INTRODUCTION
dx
The pile foundation inay be under a horizontal load as well as under an axial load. The pile problems, which are under the effect of the horizontal force, appears frequently at quay structures, petrol production plants and marine pilings. The horizontally loaded pile is closely related with the beam problems, based on the elastic foundation, and is an application to head of pile in the pile-soil system of the external forces and moments. According to the beam theory based on the elastic foundation:
a‘w
In this equation $4) : The horizontally displacement of the pile x : The depth in soil E l : Pile stiffness 12 : Soil reaction. The soil reaction “p” is a function of the pile characteristics and (1.1) the solubility of the equation is depending on the fbnction, chosen for the soil reaction “f. If we accept that there is a direct relation between the soil function “f and displacement of the pile “ w ” , the soil reaction which effect to the unit length of the pile will be found by the following equation: ( 1.2)
In this equation, K represents the ratio of soil reaction on a point of pile to the displacement at the same point. If the equations (1.1) and (1.2) are united:
(1.3)
/F Coefficient of soil reaction, d= Diameter of the pile, the equation number (1.3) is obtained. The “k” value is given in the literature approximately for the soil classes (Bowles 1988). The aim of the study is; first, to determine in testing manner of the “I? value through mounting a vibrating apparatus on the head of the pile, then, through adding the technical terms, taking into account tlie dynamic effects establishment of a numerical model and its evaluation. The new equation obtained is:
Elax‘
~=-K.w
K=k.d
a2w + ,U---+ k-&v + Kw = 0
at2
at
(1.4)
Theoretically, the pile head flexibility, i.e. tlie horizontal displacement curves, which are the counterpart for a unit force, are obtained for differcnt values of coefficient of soil reaction. The “k” value of the curve is the coefficient of soil reaction, with which curves experiinentally obtained flexibility values fits. 2 EXPERIMENTAL STUDY The experimental study has been carried out one of the composite pile of the quay at Clierbourg (France). The composite piles have been formed through driving of steel tube piles of l n i of diameter and then filling them with concrete without reinforcement. The experiment has been carried out by casting after 2 months of the vibrating apparatus on a pile figure 1.
511
of the system as a whole is provided. Before beginning of the experiment all the control units x e checked. Then a sinusoidal fmce is given in horizontal. 1vay to the head of pile in any choseii frequency iyalue. The force, which will be provided by the vibrating apparatus is calculated as below:
f= frequency, Hz Figure 1. Fixing of the apparatus on the pile
(2.1)
T n this state by dividing the force to& term, the effectir-e force is taken into account. The displacement at head of pile, which is the equivalent to this force is found as a proportion of the speed value (V), determined by the speed detector to the “o”. By- division o f this so found value to the applied fori.e is c,alculatedthe flexibility value of head of pile. (2.’) B!. suhsrituting this obtained espcrinientall>fleribilitj- value in the equation number 2.3. wliicli neglects the basic arid dyiamic effects (Briard 1V S ) . the “k“ value m a y be calculated.
For thc corupleteiy mbedded piles S=------I
Figure 2. Interior view ofthe apparatus
2. 1 The Vibrating Apparatus and Its Chmaiwristics
The vibrating apparatus is fomied of two blocks, rotating in reverse direction in same angular velocity around a vertical axis, driven by an electrical engine figure 2. The apparatus is producing a sinusoidnl horizontal force at (cycle/sec.) frequency hnction. equal to the rotating velocity. The amplitude of this force is proportional to the square of the fiequency. With an aim to measure the horizontal displacement at the head of pile, again on the head of the pile a speed detector is adhered in a manner, in which its direction is parallel to the direction of the vibration. The horizontal displacement of the head of pile is given by the measured speed value (V) with the proportion to the wvalue. The mass of the apparatus: 50 kg, Diameter: 510 mm, height: 200 mm, frequency interval 1 - 14 Hz. The engine is Mavilor type M-600-6OOW.
For the p i t i d l y embedded piles S=-
Q
. ,
(2.3)
&?Ela -?
2Ela.’ Here:
1+2utx
= -I
is the characteristic length
LE1 a: the part of the pile, which is 011 the soil level. However the aim o f the study is the determination of the “k” by a numerical model, including also the dynamic effects. 3 NUMERICAL MODEL
The amplitude of the sinusoidal horizontal force F = F ( u ) s i n ut , which is applied to the head of pile, embedded in an elastic soil, is:
F(w) = h W 2 2.2 Fixation of the vibrating apparatus on the pilc head and measuring of the flexibility (Horizontal displacement, counterpart of the unit.foi~e) An aluminum plate of the vibrating apparatus (500 mm x 500 mm x 25 mm) is bound by bolts to the pile head in absolutely non-moving manner. Then the vibrating apparatus is also bound to this plate in a strong manner by bolts. By this way the behavior 512
“h” is a constant of the apparatus. The cross section of the pile is small compared to its length. The pile is subjected to a simple bending under this force. The “x” abscise has been accepted from the base of the pile to its upper part. In the cases when the operational frequency of the vibrating apparatus “ U ” does not fit with the natural frequency of the
r . = E .I
a a
I
+E I b'b
The coefficient of soil reaction value is:
k, = d, .K,
(Nm-* )
(3-7)
In this formulations d , ,i = 1,...n The pile diameter in each J , pile space K , ,i = 1,...n The soil module in each J, pile space If some part of the pile is up of the soil level, K, =o, k, = 0, p , = 0
(3.8)
3.2 The limit conditions
The pile end behavior (PO abscise) has an effect on the ations ions of the examined system. The w(x,t) are the values of the rotating, bending moment and shear force, depending on the displacement e(x,t), M(x,t), T(x,t). At F O abscise, two of these four are always zero. On the embedded pile: w(0,t ) = 0,8(0, t ) = 0 V t . 3.3 :Mosement equations Figure 3. Divided pile spaces
It is taken into account that an embedded in the soil pile is exposed to external load as shown on Figure 3. If the solution is demonstrated by an S vector: General casc:
pile-soil system, a horizontal displacement will occur. This displacement is : w(x,t , U ) = W ( X ; w>sin(wt + $)
(3.2)
The amplitude of the movement of pile head in the next calculations will be found as bellow:
A ( @ )= W(H;(01
awiax=e
(3.9)
(3.3)
3. I Meckanical characteristics ($the soil and the pile
Sinusoidal case:
?'he pile is formed of fixed sectioned elzments. However the soil is a superposition of the homogenous layers Figure 3. The elements [O,H], situated along of the pile: xo = O l x ,
<x, <...<x, l...<x, = H
s ='
i = 1,...y1
(3.4)
d
ai d
X
=
-T
(3.10)
d T / d x = (ki- p iw2)w
(3.5)
i = 1,. ..n The length of each space i p,, = I ,...n ,The mass of the pile in the space J, r, , i = I , ...n The rigidity of the pile in the syacc J, , and due to the fact that the pile section is composite:
I, ,
1
dZ/dx=8 d g i L i x = a / r , ateachJ~interva1
Each pile, situated in the space J,, and the rnechanical character of the soil is constant.
J , = [x,-,,x,],
fi~x),G(x),M(x),- T(x))Sinot
dW
Due to the sufficient energy the term k'been neglected. Limit conditions:
at
has
Limit conditions:
-
w(0) = 0 w(O,t) = 01 . Pile end(3.12) Ptle end (3.1 1) tI(0.t) = 0 j 8(0)=0 I
513
1
M (,H . t ) = O ,
>
T ( H , t ) = Fsinmt]
External loads
By forming of a global transfer matrix the solution of the system is reached through:
(3.13)
S ( H ) = T S(o)
M(H) =0
(3.14)
T ( H )= F
4 EVALUATION OF THE CALCULATIONS AND EXPERIMENT CONCLUSIONS
The coefficient of subgrade reaction value at J, interval.
k I( ” j = kl -p,w’
The experiment has been done on a partially embedded in the soil, having fixed sections composite pile D=l m. The part of the pile, situated in the soil, is J , , and its situated on the soil part is J , .
(3.15)
The equations number (3.10)-(3.14) has to be solved.
The part, situated in the soil: 3.4 Solution oftlie system in the interval [0,H]
I , = 8 . 2 5 m , p I = 3 . 1 x I 0 3 k g n - ’ , r , =5x10’ NI,?’
The vector “S” is separated to two sub-vectors, which provides the limit conditions at the pile end.
The part, situated on the soil:
s’=/
(wl,el,M ’ J ) , S”=’( w 2 , Q 2 , M 2 , - - T 2 ) ,
The “ k , ” value in the calculation has been chosen
The pile end limit conditions; embedded pile -I
w (0) = O,G‘(O) = O,G’(O) = l,-?(O)
-1
w-(o)= O,%’(O)
in a manner, permitting the appearing of the effects of the characteristics of the soil against the pile movements.
= 0, (3.17)
= O,M2(0)= O,-T2(0) = I,
k~
Each solution of the equations number (3.10)(3.14) is a linear combination.
2.5x107~
3x10‘1
5x10’
N117.2
Flexibility - frequency curve:
a’s’ i- a’s2
(3.18) The a’ and a’ constants will be the solution of the system.
= z x l o i j
s = A / F( F = l ) Horizontally displacement-frequency curve:
A = 1000’S 1 4 ~ ’ (3. 8, For the “a”values, which do not make “0” the determinant of this system:
Speed-frequency curve:
’
=10003S/4n’
Acceleration - frequency curve:
r =1 Through the solution of the equations number (3.10) (3.14) the amplitude of movement at the head of pile is calculated as:
-
-
--
A = w(H)= F
-
w2 ( H ) M I ( N )+ W ’(If)%? -,
( H ) (3.20) T - ( H ) M ’ ( H )+ T I ( E . I ) M 2 ( H )
The calculations are made by transfer matrices. At the first stage at each Jl = [ x ~ - ~ , xspace ~ ] the z,elementary transfer matrix is set up,
S. = 2, .Si-l, i = 1,...n is calculated. Then, T = z,?zJI-,
0 0 d / 4 ~ ~
These values are drawn in Figures 4, 5 , 6, 7. The initial slope in the flexibility curve is horizontal, however, this slope in the displacement curve is high. This state of the flexibility curve is demonstrating only that the forces, which are the counterpart of the small frequency values of the used vibrating apparatus may be neglected. It may be said that the speed is constant on the speed curve between 5-10 Hz, and that this fact is depending to the choseii mechanical characteristics. At the change of frequency in value of 10 Hz, the flexibility, displacement, speed and acceleration values are changed in counter way according to the chosen coefficient of soil reaction. This fact shows that how high is the resistance of the soil, so important is the movement of the pile head.
514
FLEXIBILITY LEGEND
v Experitnental values
-
Unstable measunngs Theoretical vzlues
Figure 4. Flexibility curve (k=2500;5000 kdaNm-') Figure 8. Comparison of the theoretical and experimental results.
Figure 5. Displacement curve of the pile head (k=25 00 ;5000 kdaNm-')
I
I
",I1
The flexibility frequency curves, shown at the been drawn using the figure 8 have k=1000,2500,5000 IcdaNnz-' (@2) values. The obtained experimentally flexibility values have been placed on these curves. As it may be seen at the figure 8, the values 2500 kdaNni-2 obtained as a result of the experiment, are shown on the curve. As a consequence the coefficient of soil reaction value of the soil is 2500 ldaNnz-2(tf / m2) As a conclusion, the numerical model and the experiment result fit between 1 Hz and 1,5 I-Iz. The fact that the experiment results between 1,5 Hz and 4,5 Hz are spread is caused due to tlie lack of the resonance at 3,5 Hz. The hai-tnony between the model and the experiment result after 5 Hz obligate to take into account the dynamic effects in the numerical model. Thus it is observed that the iiumerical model which is out of the resonance space is in agreement with the experiment results. 5 CONCLUSION
Figure 6. Speed curve of the pile head (k=2000;2500;3000;5000 kdaNm-')
Figure 7. Acceleration curve of the pile head (k=2000;2500;3000;5000 kdaNm-')
The study is an evaluation of the fitting of tlie coefficient of soil reaction value, obtained experimentally by a vibrating apparatus, mounted at a composite pile head, embedded partially in the soil, with a numerical model. The vibrating apparatus, fixed at the pile head, is applying during the experiment a measurable siiiusoidal horizontal force to the head of pile. The horizontal displacement, which is the equal of this force is measured, too. The horizontal displaccinent flexibility values, which are the equal to the unit force, are easily calculated in place. Using these values, in the calculations, which do not include the dynamic effects, the not-sure coefficient of soil reaction values are obtained. The aim is to determine the actual coefficient of soil reaction value, by evaluation through a model, including the dynamic effects through a more realistic approach of these flexibility values, obtained experimentally. By this way taking into account the dynamic effects in the iiuinerical model the flexibil515
ity values are drawn by giving different coefficient of soil reaction values. The coefficient of soil reaction value of the soil, which is the coefficient of soil reaction value, used in the curve, drawn in the numerical model, with which curve the flexibility values fit, obtained in the experiment. Through a prepared computer program and in the cases \Then the pile is embedded or articulated and at all kind of soil terms the results of the experiments, made by a vibrating apparatus, may be evaluated. Thus a more realistic coefficient of soil reaction value is obtained. REFERENCES Acar. M.11. 1953. Analyse theorique et experimentale du coinporteinent des pieux. These de docteur-ingenieur, Universite Pierre et Mark Curie, Paris. Baguelin, F. & Jezequel J.F. 1972. Etude expirimentale du comporrement des pieux sollicites horizontalement, Annu1e.s I. T.8.1’.P.11‘’ 297, Paris. Bowles, J.E. I9SS. Foundation Analysis and Design, Mc GrawH i 1I, Ne\v-Ynrk. Rriard, A. 1978. Mesure du coefficieiit de raideur horizontale d‘un pie11 ancre dans le sol a I’aide d’un excicateur a Balourd, Contrnt DBTPCICEBTP, Paris. Chtti, L. & PoLtlos I l.G. 1993. Analysis of pile-soil interaction iirider lcrci~al iociding using infinite and finite elements, i 'amp m d Gcorc‘chnics,Oxford. England.
516
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Salkema, Rotterdam, ISBN 905809 150 3
Analysis of crosswise vibration of pile Chen Renpeng, Zhu Bin & Chen Yunming Geotechnicul Engineering Institute, Zhejiang University,Hmgzhou, People's Republic of Chirzu
ABSTRAT: A rational mechanical model of foundation of precast concrete pile on layered eiastic base when driving is established in this paper. Through dynamic anal!isis. the nurneric solurion of forced reaction of the foundation of single pile which eccentric dynamic \erticill load XIS on pile top is provided. And effects of excitation intensity. impact duration of pile lmiimer, diamcter or length of side of pile. eccentricit), of pile hammer, foundation modulus and coefficient of viscosity of soil to the dynamic response of pile are analyzed. The primary analytic method of studying the pile response LI hich intense axial excitation and moment excitation act on. which is very important when studying the vibration of excited precast concrete pile. pile, and thcq haic not farther discuseJ thc dynamic response of tlie pile ~ h c heccenrric dynamic vertical load acts on. 'Thc niomcnt of pile h m n i c r ncting on pile top is an intciisi'c7c impact loxi whc11
1 INTRODUCTION
As for the design and the construction of pile, it is very important to study the dynamic response of single pile on the condition of intense excitation. Based on the theory that the pile-soil system is linear. there are four models which are discrete modcl, continuous medium model, finite element model and boundary element model. For example, the research that aiialyzing the response of pile which axial dynamic load and crosswise dynamic load respective acting on. and the pile is supported OII rigid soil layer (Shisheng LU & Yachao 131 1987). On the other hand, there are two main niodels considering tlie pilesoil system is non-linear, one is the model consisting of discrete springs, damper and frictional element (Matlock, H.: Foo, H.C. LCt Bryant, L.M. 1978), the other mode1 assurries that there is a cirque-column weakencd rrrca and the mechanical character of soil in the area is different from that of soil out the area (Novak, M. & Sheta, M. 1980, Novak. M. & Han, Y.C. 19S8). These models have provided new approaches to study the dynamic characteristic of the pile, however, these researches only emphasized on the single excitation acting on the
Figtire 1. zjlna15rlicai moLjcl
51 7
additional moment OII the pile top, which thereafter causes the damage of pile in probability. With the increase of this kind of project, the research of this problem becomes more important. In this paper, dynamic property of precast pile under the eccentric load is studied, and effects of various factors on dynamic response of pile is analyzed to improve the acquirement on dynamic property of pile when driving eccentrically, which subsequently provided warrant for engineering design and construct.
2 CROSSWISE DISPLACEMENT AND INTERNAL FORCE ANALYSIS OF PILE Based on Winkler model, the soil around pile is substituted by a series of distributed springs an3 dampers. Pile is assumed as equivalent section, elastic, and the pile top is free. After the point of pile enter into bearing stratum, the penetration of pile is so little that we can assume that the point of pile is fixed in the bearing stratum, and the length of pile on the ground is EI,, the length of pile under the ground is N 2 (Figure 1). The moment caused by eccentric impact force acting on pile top is : M ( t ) = N o siiiwt . e (1) in which t =time, N o =maximal value of the impact force, e =:eccentricity of the impact force, 05 wt < 7r ,w = 7r I I , . where t , =duration of the impact force (Figure 2). When eccentric vertical load is exciting pile top, besides dynamic moment and dynamic axial force, there are pile's dynamic inertial force, counter-force and viscosity damping of the soil acting on the pile. Based on assumptions above, deduced by elastic mechanical, the crosswise vibration equation of the pile excited by eccentric vertical dynamic load on the pile top is:
in which E =modulus of elasticity of pile,
EI =bending rigidity of the pile, y =horizontal displacement of the pile, m,]=mass of the pile per unit length, c =coefficient of viscosity of the soil, foundation modulus and coefficient of viscosity of soil, m =constant of foundation modulus which is proportional to deep, b =length of side or diameter of the pile, x = distance from pile top. Numerical method can solve this differential equation, the partial derivatives that the displacement of the pile to x is dispersed by finite-difference method, and the partial derivatives that the displacement of the pile to t is dispersed by Newmark method. The division of the different unit of the pile is shown in Fig 3.
Assuming that i is the number of the unit of time, and j is the number of the unit of pile. After the partial derivatives of equation (1) are replaced with center different equations and Ncwrnark equations, equation (1) can be written
in which,
as:
[KI =
A . Y,,/-2 + B .
Yl,/-l
[KXYI =
0
+ Cl,/ . Y l , ,
-2,
[ K ] is a coefficients matrix with the
1 - 2
0
,
1
0
o...o
0 0 0 0
- 2 1 O ' . . O B C, D E 0 ... 0 A B C , , D E . . . O
A
- r?
(6)
order of n + 1 :
- 1 2
in which , j = 2
bfI
0
0 0 0 0 ...
...
y k =displacement of
the pile unit which the number is 1 at the time unit which the number is k , A = F = - El 11' '
0 0
0 0
0 0
0 0
0 0 ... A 0 0 . ' . 0
0
0
0
0
O O " ' 0
0
0
0
0
0 0
0
0
B C,,,?D E 0 0 1 0 0 1 0 - 1
the column vector of crosswise displacement in center of every pile unit is:
bI
=
{Y,0 YI,,> Y , , 2 . . Y , 7
'
11-2
,
Yl ) I - ,
>
I'
Y,,,,
(7)
the column vector of load in center of every pile unit is:
the program based on the method mentioned above is provided in this paper, which can present the displaceinent response of every pile unit at any time, and so do the shear force and the moment. (there is no algebra function of in and c in equation (4) and ( 5 ) of the pile unit above gro u 11 d ) where h = step length of center different H = ( H=pile length), s = step length of time, n-3
3 EFFECTS OF CROSSWISE VIBRATION OF PILE When pile haininer impacting on pile top, the pile-soil system is vibrated by the excitation of pile hammer to pile, and the vibration is transient and nonperiodic. Commonly, the time of energy translation is short than cycle period of auto-oscillation of the pile, but the duration of response can be several times of cycle period of auto-oscillation of the pile. There are 5 main factors effecting on the crosswise vibration of pile, which are excitation intensity, impact duration of pile hammer, eccentricity of pile hammer, foundation modulus and coefficient of
a , 6 is constants: a 3 0.25(0.5 + 6)' , 6 3 0 . 5 , s, =(j-2).h,
I, =i..r,
hi, = N o .sinat, ; when 1,
when 0
t,,
NI =0 .
Additionally, there are four equations deduced with boundary conditions, thus, as for every time unit, there is one equation group consist of 17 + 1 equations, which can be written
as:
519
viscosity of soil. The effects of them on the crosswise vibration of pile can be analyzed by 5 examples following, where the crosswise section of the precast concrete pile is square, and the parameters H,HI,H , , E are invariable, that is N = 2 4 m , H, = 2 m , H,= 3 2 m , E = 3 x 1 0 7 K P a . It is found that the maximal moments of the pile are all at pile top through computing. and their values are N o . e , so this paper only analyzed the maximal crosswise displacement and maximal she ar . (1) Effects of excitation intensity on crosswise vibration of pile. Assuming that t , = 0.02s, L' = 300 KN . s / m , n7 = 500OKhi / n?-', e = 0.1m . From Figure 4 and Figure 5 , it is found that the maximal displacement and shear are increasing ivith excitation intensity of the pile hammer, the main reason of which is that the moment at pile top is proportional to the excitation intensity of pile hammer when the eccentricity of pile hammer is stated. The niaximal displacement and shear are affected mostly by length of side or diameter of pile. They decrease while length of side or diameter of pile increases. So we can adopt the lighter pile hammer or descend it in order to reduce the additional internal force of the pile caused by the eccentricity of the pile ham mer . (2) Effect of impact duration 01 pile hammer on crosswise vibration of pile. Assuming that N o =5000KjV, c=300KN.slm, e=O.lm, ni = 500OK.A'! m'. From Figure 6 and Figure 7, it can be found that the crosswise vibration of pile is mostly affected by the duration of impact, and the maximal displacement is increasing with duration of impact, but the maximal shear presents decreasing trend while duration of impact increases. The transient vibratioywhose impact duration is very short increases the internal force of the pile, while reduces the capability of pile to endure dynamic load, and therefore destroy the pile. We usually adopt the heavier pile hammer and thicken the buffer cushion to increase the duration of impact when
Figurc 6. Curve of I he maximal displacement changcd with impact duration
driving, accordingly, the internal force of the pile will decrease effectively. But the heavier pile hammer will increasing the excitation intensity, so we should choose the appropriate pile hammer v, hcn dri\iing. (3) Effect of eccentricity of pile hammer on crosswise vibration of pile. Assuming that
520
Figure 11. Curve of the maximal shear changed with foundation modulus
Figure 9. Curve of the maximal shear changed with eccentricity of pile hammer
N o = 5000KN ,c = 3O O K N . s / n? , = 5000KN / W I ' , I , = 0.02s. From Figure 8 and Figure 9, it can be found that the maximal displacement and shear is proportional to the eccentricity of pile hammer when the impact force is invariable. Actually, what controls the crosswise vibration of the pile when driving eccentrically is the shear wave excited by the moment at pile top. Thus it can be seen that we must adopt a series of methods to
control the eccentricity of the pile hammer. For example, enforcing the management during constructing in order to ensure the pile hammer, the pile cap and the pile in one axis; keeping the thickness of the buffer cushion equal. (4) Effect of foundation modulus on crosswise vibration of pile. Assuming that N o = 5 0 0 0 K N , e = O . l m , c = 3 0 0 K N . s / n i , t , =0.02.~. From Figure 10 and Figure 11, it can be found that the crosswise vibration is less affected by the foundation modulus. (5) Effect of coefficient of viscosity of soil on crosswise vibration of pile. Assuming that t , = 0.02~, N o = 5000KN,
m = 5000KN/m4, e = 0 . 1 ~ From . Figure 12 and Figure 13, it can
be found that the crosswise vibration is less affected by the coefficient of viscosity of soil when the part of pile head is on the ground. But when the part of pile head is under the ground, the damping of soil to pile decreases the maximal
521
Figure 13. Curve of the maximal shear changed with coefficient of viscosity of soil
displacement of pile while increases the shear being. Analyzing with example (4), we can conclude that the effect of soil to pile is little when the part of pile head on the ground, and the length of the part of pile on the ground hardly affects the crosswise vibration of pile. This can be seen on the other hand, the effect of soil to pile is obvious only at the part of pile head. 4 ANALYSIS OF ENGINEERING EXAMPLE A certain precast concrete square pile (Figure 14), H = 24n1, H ,= 2m H z= 2 2 m , b = 0.5m, E = 3.5 x 10‘ KPa . adopting the cylinder diesel hammer weighs 4 tons when driving. Assuming that the inipact force is N o =45OOKN, the eccentricity of the impact force is e = 0.08~1,the impact duration of pile hammer is I , =0.02s. The lognitudinal placing of steel bars are 016, and the placing of hooping is (D6@50, the )
Figure 15. Cur-vc ol displaceriient of pile at different time
grade of concrete is C40, the pile is in three soil layers, the parameters of the soil is in Figure 14. This paper plots the curves based on the conclusion of computing at the time of t , / 4 ; I , /2 and 31, /4 (Figure 15-17). From Figure 15-17, it can be found that the maximal displacement and moment of pile are at the pile top, but the maximal shear is 3 meters away from the pile top. The maximal moment is 3 6 0 m . ? I ? , and the maximal shear is lOOKN . The dynamic response in 10 meters away from pile top is rather great, but it is almost zero out of 15 meters away from pile top. Through the further checking computation of pile intensity, it can be found that the pile was destroyed by the eccentric impact force of 4500KN. The maximal axial load that this pile can bear is 4.88 x 10’ KN , but the maximal eccentric load that this pile can bear is only 3 . 8 5 x l O 3 K N , so the vertical bearing capacity of the pile was decreased 522
viscosity of the hard clay is greater than that of the soft clay, so the internal force of the pile is greater when eccentric driving in the hard clay than in the soft clay. 2. The maximal displacement increases while duration of impact developing, but the maximal shear presents decreasing trend while duration of impact increases. For adopting the heavier pile hammer or thickening the buffer cushion will increase the duration of impact when driving, the internal force of the pile can be decreased effectively by these measures. 3. As for the flexible long pile is concerned, the dynamic response of the pile is little affected by the length of pile. The dynamic response of the pile is almost invariable regardless of length of the part of the pile on the ground when the pile head is on the ground, the main reason of which is that the pile is not restricted by the soil and the dynamic response of the part under the pile head is little. Figure 17. Curve of moment of pile at different ime
REFERENCES
approximately 21%. In a word, the impact capacity that the precast concrete pile cai bear is great decreased because the pile hammer is eccentric.
Lu, S.S. & Lin, Y.C. 1987. The calculation and arzabsis of pile foundation. People traffic publishing House: 304-318. Matlock, H. et al. 1978. Simulation of lateral pile behaviour under ear-thquake motion. Proceedings ASCE Specially conference on Eart hqua ke Engin - eering and Soil Dyizamics . Pasadena, Calif: 600-619. Novak, M. & Sheta, M. 1980. Approximate approach to contact effects of piles, Proc. ofDyizanzic response of Pile Fourzdation. Analytical aspects, ASCE, Florida: 53-79. Novak, M. & Han, Y. C. 1988 Impedance of soil layer with boundary zone. The University of Western, Ontario research Report, GEOT: 16-85.
5 CONCLUSIONS In this paper, the dynamic response of precast concrete pile which eccentric dynamic vertical load acts on pile top has studied, and effects of excitation intensity, impact duration of pilc hammer, diameter or length of side of pile, eccentricity of pile hammer, foundation modulus and coefficient of viscosity of soil on the dynamic response of pile has analyzed. Through analyzing in theory and calculation, it can be concluded as follows: 1. The effect of the soil on the crosswise vibration of pile is obvious only when the length of the part of pile on the ground is shorter than 6 meters. When the part of pile head is under the ground, the maximal displacement of pile decreases while the coefficient of viscosity of soil increases, but the maximal shear increases accordingly. Commonly, the coefficient of 523
This Page Intentionally Left Blank
Application of Stress-Wave Theory fo Piles, Niyama SC Beim (eds) C) 2000 Balkema, Rotterdam, ISBN 90 5809 1503
onitorirag and c0ne;rsl of ynamic effects of pile i ~ S ~ a ~ prior l ~ t itoo ~ driving M. R. Svinkin VibmCoiisidt, Clevelaiid, Ohio, USu
A4BSTRACT:Pile driving induced vibrations may be harmful to adjacent and distant structures, sensitive instruments and people. It is important to assess intolerable vibrations before the beginning of construction activities. Guidelines for preconstruction survey are presented. Analysis of existing methods shows potentialities of various approaches to predict ground and structure vibrations. This paper demonstrate the application of the impulse response function prediction method to solve the geotechnical problem of predicting ground and structure vibrations before pile installation. Control of the predicted vibrations is important step in preventing intolerable-vibrations.
1 INTRODUCTION
Staalduinen and Waarts (1992), Svinkin (1992) and others. This paper presents some guidelines for preconstruction survey, prediction and control of soil and structure vibrations generated by pile driving before the start of construction activities at the site.
Pile driving generates elastic waves in soil which may adversely affect surrounding buildings. Their effects range from serious disturbance of working conditions for sensitive devices and people, to visible structural damage. The dynamic effect of pile installation on adjacent and distant structures depends on soil deposits at a site and susceptibility ratings of structures. It is likely that intolerable structure vibrations may be induced in close proximity of the driven piles, but permanent deformation resulting from vibratory densification of loose granular soils may occur at various distances from the source. It is important to assess the dynamic effect before the beginning of pile driving. Therefore various approaches are used to predict soil and structure vibrations. Monitoring construction vibrations have to be started prior to pile driving to provide the safety and serviceability of sound and vulnerable structures and facilities. Specific guidelines for the monitoring and control of construction vibrations in particular generated by pile driving were prepared by Dowding (1996) and Woods (1997). Separate questions of the dynamic effects of pile installation on surrounding structures were considered, e.g. Heckman & Hagerty (1978), Wiss (1981), Lacy & Gould (1985), Broers & Dieterman (1992), Massarsch (1992), van
2 PRECONSTRUCTION SURVEY
A predriving survey is the first step in the control of construction vibrations to ensure safety and serviceability of adjacent structures and/or distant structures with sensitive equipment like hospital facilities and offices with precision instruments. According to pile driving practice, construction vibrations may cause direct damage to structures at a distance about of one pile length from the driven pile. However, there is no single opinion regarding the maximum radius of a preconstruction survey area with buildings surrounding a construction site. Dowding (1996) suggested a radius of 120 m of a construction activities, or out to a distance at which vibrations of 2 mm/s occurs. Woods (1997) considered distances of as much as 400 m to be surveyed to identify settlement damage hazards. Obviously, a radius of the area of preconstruction survey is various and depends on building condition and utilization. The predriving survey includes a few steps of site investigations. 525
Inspect present conditions of surrounding buildings. Perform damage susceptibility study to establish vibration control limits. Measure vibration background at the area under investigation. Assess problems such as cracking of building or foundation failure. First of all a preliminary desk study of a layout of the area for a preconstruction inspection and a site walk-over observation should be made. Existing cracks found in buildings have to be marked. It is necessary to distinguish cosmetic and structural cracks. Most attention should be paid to cracks in the structures themselves. The width of cracks should be measured with a proper ruler. Determining the cause of cracking is important to predict lengthening and dilatation of old cracks under the vibration effect of pile driving. For assessment of the dynamic effect on surrounding structures it is necessary to take into account the thresholds of damage, cracking and perception. Buildings inspected under the contract requirements are commonly classified depending on structure susceptibility to cracking and proximity of structures to pile driving operations. Structure susceptibility is usually related to the threshold of cosmetic cracking (Dowding 1996) and depends on a degree of degradation of the building structural and nonstructural systems. Apparently, it makes sense to broader consider this terminology as susceptibility of the building-soil system depending on degradation of building systems, utilization of buildings and soil conditions. It is important for certain cases. For example, a building, located in the proximity of the driven piles, identified as having low susceptibility and built on liquefiable soils might have substantially larger deformations than a building identified as having high susceptibility but erected on non-liquefiable soils at long distance from the driven piles. Certainly, for some sites only the threshold of cosmetic cracking could be sufficient. Measurement of a vibration background at soil and buildings with sensitive equipment and/or computerized technology should be a part of the preconstruction survey. The vibration measurement might reveal microvibrations and vibrations induced by industrial machinery located nearby. A waveform recorder has to be used for these measurements. Assessment of the dynamic effect of pile installation on surrounding structures can be made on the basis of prediction of soil and structures vibrations from pile driving. @
526
3 VIBRATION ENVIRONMENT Pile driving operations generate different kinds of waves in the ground, but Rayleigh waves have the largest practical interest. Rayleigh waves become predominant over other wave types at distances smaller than a pile length. Impact hammers induce transient ground vibrations with the dominant frequency between 7-50 Hz (Svinkin 1999), and vibratory hammers induce steady-state ground vibrations in the frequency range of 5-40 Hz (Warrington 1992). The intensity of ground vibrations at the surface and the certain depth are comparable. Linehan et al. (1992) studied the effect of pile driving in the proximity of the gas line with a pipe diameter of 1 m. The sheet and H piles were driven in sandy soils with vibratory and diesel hammers, respectively. For both types of driving, particle velocities on the pipe were less than those measured at the surface at a distance of 1.5 m but became similar at distances about 10-12 m from driving. Barkan (1962) analyzed the variation with depth of vertical displacements in loessial clay induced by a pile driver and concluded that at small depths about of 0.2 to 0.5 wavelength changes in vibration displacements are relatively small.
3.1 Deterministic approach Statistical analysis is used for evaluation of the test data from blasting (Dowding 1996). Such an approach for assessment of other construction vibrations should be carefully used because ground vibrations generated by the dynamic forces applied to the soil directly or transferred through driven piles have deterministic nature. The effect of large plastic soil deformations at the contact area under falling mass on ground vibrations was studied with falling mass of 15.0 tonnes at a site where soil deposits were mostly fine sands with natural moisture content (Svinkin 1996a). The drop height was 10 m. Many impacts were performed at the same spot; consequently, large plastic soil deformations occurred at the point of impact. Records of ground vibration displacements at various distances from the place of impact on the ground are depicted in Fig. 1. Comparison was made for records obtained for two equal impacts with different degrees of plastic soil deformations at the contact area. In particular, vibrations were measured at distance of 43 m for the first and ninth impacts, and at a distance of 57 m for
amplitude reduction of Rayleigh waves, generated by an earthquake, between two points at distance rl and r2 from the source as (Golitsin 1912) r = Z rn
r=33 rn
Figure 1. Comparison of two different displacement records of ground vibrations in fine sands for identical impacts on ground by falling mass of 15.0 tonnes
the first and seventeenth impacts. For the first impact, the falling mass dropped on a flat ground surface, but for the seventeenth impact, it dropped into a pit deeper than 1 m. In spite of considerable soil deformations at the contact area, each pair of ground surface vibrations are similar at locations of measurements. The results demonstrate that at any location on the ground, except probably a zone at close proximity to the source, soil vibration displacements measured simultaneously with impact on the ground are stable, have well-defined shapes, and are independent of the intensity of soil deformations at the contact area. The differences between displacement amplitudes measured during various impacts are within the limits of error of the measurement system. This confirms the necessity of using a deterministic perspective for prediction and analysis of construction and industrial vibrations.
3.2 Attenuation of surface waves The waves travel outward from the construction source and attenuate in the results of geometrical spreading and material damping. The following equation is known to calculate the
527
Where A, = amplitude of vibrations at a distance rl from the source, A, = amplitude of vibrations at a distance r, from the source, y = coefficient of attenuation. The term ( ~ J I - , ) ~indicates ,~ the radiation or geometric damping and the term exp[y(r2-r1)] indicates the material damping of wave attenuation between two points. Equation (1) is highly popular in prediction of the peak particle velocity (PPV) of vertical ground vibrations from construction sources, for example, Massarsch (1992), van Staalduinen & Waarts (1992), Gerasch (1998) and others. Nevertheless, there are obstacles for the application of equation (1) to predict maximum PPV of vertical ground vibrations (Svinkin 1999). Test data along the ground surface shows that for various pairs of widely separated points on the ground surface, values of y can differ by more than an order of magnitude and even change sign. Thus, the coefficient, y,acceptable for small distances may be inadequate for long distances. On account of wave refraction and reflection from boundaries of diverse soil layers, an arbitrary arrangement of geophones at a site can yield incoherent results of ground vibration measurements because waveforms measured at arbitrary locations at the site might represent different soil layers. Moreover, an amplitude of ground vibrations is unknown before the beginning of pile driving. So, problems with uncertainty in assignment or determination of the coefficient, y , and unknown referenced amplitude show that the application of equation (1) to predict ground vibrations is questionable. Brettmann & Cotton (1999) suggested other equation to predict the peak particle velocity, v, of ground vibrations v
=
K(E/r)'.' e-7'
where K = soil constant; E = energy transferred to soil; r = distance from source; y is the same as above. The term K(E/r)o.5 represents the geometric damping and there is some uncertainty in determination of constant K. The term exp[-yr] represents the material damping of the ground and there are problems with determination of coefficient
y described above. So, equation (2) has disadvantages similar to the equation (1). Wiss (1981) found the ground velocity-distanceenergy relationship, so-called scaled-distance approach, to calculated the peak ground velocity at surface distance, D, from a source normalized with energy as v
=
k[D/fi]-"
(3)
Where W, = energy of source or rated energy of impact hammer, k = value of velocity at one unit of distance. The value of 'n' yields a slope in a loglog plot between 1 and 2. It was an important finding because a slope of amplitude attenuation for all tested soils was in the narrow range of 1 to 2. It turned out that the scaled-distance approach was very useful in the assessment of construction vibrations, On the basis of the actual range of energy transferred to piles and the range of the measured peak particle pile velocity at the top of steel, concrete and timber piles, the results of Woods and Jedele (1985) were adapted by Svinkin (1996b; 1999) to calculate the peak ground velocity prior to the beginning of pile driving. The graphic relationship between PPV of ground vibrations and scaled distance from driven piles is used. The results obtained from the graphs depends on PPV of the source. The peak particle velocity at the pile head can be calculated in advance as (Svinkin, 1996b) v = 12Cw.
ZL
(4)
where Z = ES/c is pile impedance, E = modulus of elasticity of pile material, S = pile cross-sectional area, c = velocity of wave propagation in pile, W, = energy transferred to the pile. This new development of the scaled-distance approach eliminates the need to know in advance the factor, k, and enhances accuracy of predicted upper limits of ground vibrations before pile installation. The hammer energy is important source property which effects intensity of surface waves, e.g. equations (2) and (3). However the energy of source is not always the dominant factor in determination of the intensity of ground vibrations. Dowding (1996) made comparison of ground vibrations induced by pile driving with a diesel hammer and by Franki pile driving and revealed that the latter developed two times more energy but induced smaller ground vibrations at the same
distances from the sources. Perhaps the cause of this phenomenon was a higher attenuation of surface waves in the loose soil deposits where Franki piles were driven. This observation underlines the significance of soil contribution to the formation of ground vibrations.
4 STRUCTURE RESPONSE 4.1 Predicting structure vibrations The impulse response function prediction method (IRFP) is based on the utilization of the impulse response function technique for predicting complete vibration records on existing soils, buildings and equipment prior to installation of construction and industrial vibration sources (Svinkin, 1996a; 1999). The impulse response function is an output signal of the system based on a single instantaneous impulse input. Impulse response functions are applied in the analysis of any complicated linear dynamic system with unknown internal structure for which its mathematical description is very difficult. The following is a general outline of the method for predicting vibrations at a distance from an impact source. 1. At the place in the field for installation of the impact source, impacts of known magnitude are applied on the ground. The impact can be created using a rigid steel sphere or pear-shaped mass falling from a bridge or mobile crane, or a hammer blow on the tested pile. At the moment of impact on the ground, oscillations are measured and recorded at the points of interest, for example, at the locations of devices sensitive to vibrations. These oscillations are the impulse response functions of the considered system which automatically take into account complicated soil conditions. 2. Various ways are used to determine the dynamic loads on the ground from different vibration sources. For pile driving, dynamic loads are computed by the wave equation analysis. In the case of the operation of machines on foundations, these loads can be found using existing foundation dynamics theories. For dynamic compaction sites, loads from the source are easily calculated with known falling weights and heights. 3. Duhamel's integral (Smith and Domney, 1968) is used to compute predicted vibrations which will arise from operating construction impact source. For each single output point, the considered input - soil medium - output system is a one degree of freedom system and predicted displacements can be
528
written as follows
where I, = impulse force transmitted from machine to foundation, kNs; fI1z = circular natural frequency of vertical vibrations of foundation, rad/s; f n d = circular natural frequency of vertical damped vibrations of foundation, rad/s; CY = effective damping constant, rad/s; \(t-7) = impulse response function at the output point under consideration; 7 = variable of integration. Examples of predicted results are shown in Fig. 2, and 3. Measurement and prediction of vertical and horizontal ground surface displacements were made at diverse distances from the foundation under a sizeable drop hammer with a falling weight of 147.2 kN and a maximum drop height of 30.0 m. The soil at the site consisted of about 1.6 m of loose sand followed by about 6.8 m of medium density sand and 1 m of sandy clay underlain by about 10 m of slightly moist sand. The water table was about 6 m below the ground surface. The Rayleigh wave velocity was 270 m/sec. A layout of the machine foundation, the place of impact on the ground and geophones is displayed in Fig. 2. Predicted and measured vertical and horizontal components of ground surface vibrations at eight locations are shown in Fig. 3. It can be seen that good agreement is matched in time domain vibration records, except for horizontal vibrations at two locations close to the foundation. This can be explained by the different wave paths from the foundation under the operating machine and the place of impact on the ground. The distance between these two sources was 18.7 m. Lack of coincidence of the two dynamic sources slightly affected ground vibrations at a distance from the machine foundation. Agreement of predicted and measured vibration displacements is quite satisfactory. The differences between the peak predicted and measured vibration amplitudes are less than 30 % at distances larger than 43.0 m from the foundation. For some individual points amplitudes actually coincide. Spectrum analysis of predicted and measured time histories revealed that both records have similar frequency domain curves with the same dominant frequency. Moreover, predicted records are slightly dependent on the dynamic parameters of the source (Svinkin, 1999).
Figure 2. Layout of machine foundation, place of impact loads on ground and geophones
5 CONTROL OF VIBRATIONS There are three ways of how ground vibrations from construction sources can effect adjacent and distant structures. First, excitation frequencies from ground vibrations do not match structural natural frequencies. Second, ground vibrations may generate the conditions of resonance in the building. Third, foundation settlements result from relatively small ground vibrations. For all situations, sensitive equipment in buildings will be exposed to vibrations. The tolerable vibration level can be verified before beginning of construction activities at a site. For structures which natural frequencies do not match excitation frequencies of ground vibrations, structural responses would be lower than the ground vibrations. Prediction of the maximum peak particle velocity independent of frequency could be used. However, it is necessary to point out that low structure vibrations cannot prevent high vibrations of equipment installed in the building. Dowding (1996) demonstrated a case where PPV on the ground and the second floor were 2.03 and 1.52 mm, respectively, but the table on the floor had PPV of 10.5 mm. Levin (1969) found that 3-4 cycles of ground vibrations could generate conditions of resonance in the building. Such conditions can be evaluated on the basis of the prediction of complete time domain records. Significant foundation settlements caused by pile driving in vulnerable sands can result from PPV of ground vibrations as low as 2.5 to 5.1 mm/s (Lacy & Gould 1985). Such velocity values are much less than 51 mm/s which is considered as the tolerable safety limit for buildings. Therefore at sites with a predominant granular soil deposits, the maximum peak particle velocity that varies with frequency should be predicted at the area where intensity of vibration is expected higher than 2.5 m / s .
529
Vertical
b
a r = 2 5 rn
0
m
r = 3 3 in 0 m N
r = 4 3 rn
r = 1 3 2 rn 0
r = 2 0 0 rn
0
r=266
m
in
0.2 s
U
Time (s)
Time (s)
Horizontal r = 2 5 rn
0
m
r = 3 3 rn
.-0
r = 4 3 rn
0 r
r=57 m
0 r
r = 2 0 0 rn
0 l-
r=266 m r-4 Lo
0.2 s
U
Time (s)
Time (s)
Figure 3. Vertical and horizontal soil vibrations from operating drop hammer, measured (a) and predicted (b) vibrations
530
6 CONCLUSIONS Pile driving is a wide-spread source of construction vibrations which may harmfully affect surrounding buildings. It is important to accurately predict vibrations of ground, structures, and sensitive devices prior to the beginning of construction activities to avoid the undesirable effect of generated vibrations. The predriving survey should be made to ensure safety and serviceability of adjacent structures and/or distant structures with sensitive equipment like hospital facilities and offices with precision instruments. Empirical equations provide only calculation of a vertical amplitude of ground vibrations and not always with sufficient accuracy. For pile driving, the scaled distance approach with calculated peak particle velocity of the source is probably the most appropriate method for predicting upper limits of the peak particle velocity of ground vibrations. The impulse response function prediction method (IRFP) has obvious advantages in predicting time domain ground and structure vibrations prior to the beginning of construction activities. The IRFP method provides opportunity to control the tolerable vibration level of soil, structure and sensitive equipment in advance. This is an important step in preventing intolerable vibrations from pile installation. REFERENCES Barkan, D.D. 1962. Dynamics of bases and foundations. New York: McGraw Hill Co. Brettmann,T. & B. Cotton 1999. Case history: estimating ground vibrations caused by pile driving. DFI 24th Annual Members ’ Conference, Decades of Technology - Advancing into the Future: 161-171. Broers, F. & H.A. Dieterman 1992. Environmental impact of pile-driving. In F. Barends (ed.), Proceedings of the Fourth International Conference on the Application of Stress-Wave Theory to Piles: 61-68, Rotterdam: Balkema. Dowding, C.H., 1996. Construction Vibrations. Upper Saddle River: Prentice Hall. Gerasch, W.J. 1998. A way to predict vibrations. Seventh International Conference and Exhibition on Piling and Deep Foundations: 3.3.1-3.3.13. Rickmansworth: Westrade Group Ltd. Golitsin B.B., 1912. On dispersion and attenuation of seismic surface waves (in German). Russian
Academy of Science News, Vol. 6, No. 2, Heckman, W.S. & D.J. Hagerty 1978. Vibrations associated with pile driving. Journal of the Construction Division, A X E , 104(C04): 385-394. Lacy, H.S. & J.P. Gould 1985. Settlement from pile driving in sands. Proceedings of ASCE Symposium on Vibration Problems in Geotechnical Engineering, Detroit, Michigan, G. Gazetas and E.T. Selig, Editors, pp. 152-173. Levin, G.E. 1969. Dynamic influence of drop hammer foundations on surrounding structures (in Russian). Proceedings of the Second Conference on Dynamics of Bases and Foundations: 147-152, Moscow: NIIOSP Linehan, P. W., A. Longinow & Dowding, C .H. 1992. Pipeline response to pipe driving and adjacent excavation. Journal of Geotechnical Division, ASCE, 118(2): 300-316. Massarsch, K.R. 1992. Keynote lecture: Static and dynamic soil displacements caused by pile driving. In F. Barends (ed.), Proceedings of the Fourth International Conference on the Application of Stress- Wave Theory to Piles: 15-24. Rotterdam: Balkema. Smith, G.M. & G.L. Downey 1968. Advanced engineering dynamics. Scranton: International Textbook Company. Svinkin, M.R., 1992. Pile driving induced vibrations as a source of industrial seismology. In F. Barends (ed.), Proceedings of the Fourth International Conference on the Application of Stress-Wave Theory to Piles: 167-174. Rotterdam: Balkema. Svinkin, M.R. 1996a. Overcoming soil uncertainty in prediction of construction and industrial vibrations. In C.D. Shackelford, P. Nelson, and M.J.S. Roth (eds.), A X E , Proceedings of Uncertainty in the Geologic Environment: From theory to Practice, Geotechnical Special Publications No. 58. 2: 1178-1194. Svinkin M. R. 1996b. Velocity-impedance-energy relationships for driven piles. In F. Townsend, M. Hussein and M. McVay (eds.), Proceedings of the Fifh International Conference on the Application of Stress-Wave Theory to Piles: 870890. Svinkin, M.R. 1999. Prediction and calculation of construction vibrations. DFI 24th Annual Members ’ Conference, Decades of Technology Advancing into the Future: 53-69. van Staalduinen & P.H.Waarts 1992. Prediction of vibrations due to pile driving. In F. Barends (ed.), Proceedings of the Fourth International Conference on the Application of Stress-Wave
53 1
Theory to Piles: 159-166, Rotterdam: Balkema Warrington, D.C. 1992. Vibratory and impactvibration pile driving equipment. Pile Buck, Inc. , Second October Issue: 2A-28A. Wiss, J.F., 1981. Construction vibrations: State-ofthe-Art. Journal of Geotechnical Engineering, ASCE, Vol. 107, NO. GT2, pp. 167-181. Woods R.D. 1997. Dynamic eflects of pile installation on adjacent structures. Synthesis Report, National Cooperative Highway Research Program NCHRP Synthesis 253 , Washington, D. C. : National Academy Press.
532
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Full-scale field-test study of dynamic soil resistance of vibratory driven sheet piles K.Viking Deparonent of Civil a i d EnvironnzentulEngineering, Royal Institute c.f Technology (KTH),Stockholm, Swederi
ABSTRACT: The vibratory characterization of full-scale sheet piles vibratory driven into the ground has so far not been performed systematically in detail. This paper presents the preliminary results of a series of fullscale field-tests where the driveability has been continuously monitored. In the light of the presented results, the paper briefly discusses how the complexity of driveability prediction can be subdivided into three parts: vibro-, sheet pile-, and soil-related parameters, all of which significantly affect the driveability.
1 INTRODUCTION Modern variable vibratory drivers to drive sheet piles into the ground are extensively used all over the world. It is not always obvious which available driver that should be chosen to insure driveability. Amongst the uncertainties attached to driveability formulas, the key one is the assessment of the soil resistance during the installation phase. Another one is whether the assumption of pile rigidity, which is reasonable in laboratory tests, can be applied to long sheet piles in the field. The third obstacle is the difficulty in estimating the influence of clutch friction on driveability, an unknown parameter, but believed to be of significant importance. This paper briefly discusses these uncertainties amongst other important parameters significantly influencing the driveability, together with some of the preliminary results of a series of well-documented full-scale field tests.
used instrumentation and main preliminary results of the performed tests presented and discussed. 2 MECHANICS OF PENETRATION INITIATED BY VIBRO-DRIVERS A first step toward predicting the driveability of vibratory driven sheet piles is to define driveability, thereafter determine the magnitude of forces acting on the system, and identify the key factors influencing the driveability. In order to do so, it is the author’s opinion that the complexity of driveability prediction of vibratory driven sheet piles can be subdivided into three main parts: vibro-, soil-, and sheet pile related parameters significantly affecting the driveability. Another issue is the fact that driveability estimations must come out of a trade-off between the accuracy of the driveability model and ease of application to real problems.
1.1 Objectives The overall objective of the ongoing research project at KTH is to improve the understanding of the performance of vibro-driven sheet piles and to develop a better capability of predicting the driveability. From the information collected during the literature studies, (Viking 1997), and conducted laboratory tests, (Viking 1998), a driveability monitory system was designed, built, and appropriately equipped in order to dacument the dynamic soil resistance of full-scale vibratory driven sheet piles. The literature studies reviled that publications associated with full-scale test are not generally available. However, in the foll.owing paragraphs are the
2.1 Driveability The driveability of a vibro-driven sheet pile can be defined either in terms of the rate of penetration or by the forces generated and acting on the vibrohheet pile/soil system. The influences of various parameters on the three main parts are dealt with in the subsequent subsections. 2.1.1 Rate ofpenetration Driveability can be visualized in two ways: i.) as the relationship between driving time t and penetration depth z, (Figure la) and ii.) as the plotted relationship between the global rate of penetration v,, and the penetration depth, (Figure lb).
-,I<>,,
-
/
, .....,...... /
.....
tivc to the penetrating toe. The dynamic shaft resistance has been found to be virtually independent of chosen frequency, indicating that the choice of driving frequency is more important in respect to the dynamic toe resistance. It has also not yet been established how the static surcharge force F,, should be determined on relation to the unbalanced force F,, O’Neill et al. (1989). However, it is well known that increasing the bias weight increases the global rate of penetration. The eccentric moment M , determines the amplitude of both the displacement and the unbalanced force, but it has not yet been established how the best choice of static moment should be determined concerning the driveability. Several authors, (e.g. Rodger et al. 1980) have stated that the displacement amplitude should be chosen (high or small) to drive the pile efficiently at the chosen driving frequency but none gives the underlying assumptions or why it should be so. The conducted field tests reviled that the efficiency of the vibro-equipment is another essential parameter. The peak compressive force measured at the sheet pile head during the field test should be theoretically equal to the sum of the applied static surcharge force, maximum unbalanced force and deceleration of the vibrator, expressed according to Equation (2).
/ ,
b)
a)
igure la-b. Two schematic plots of the driveability curve.
The slope of the t-z curve, described in Figure la, represents the global rate of penetration vp. The enlargement of the t-z curve (Figure la) illustrates the principle of the penetrative motion of the sheet pile, i.e. the different up- and downward parts of the displacement amplitude. The pattern of the recorded v,,-z curve can be used as a tool to study the change in dynamic soil resistance. Rao (1993), involved in the extensive research work at University of Texas at Houston, stated that the penetration speed can be subdivided into three states, ranging from slow to high penetration speeds, (Table 1).
<
Table 1. States of penetration speed values vp in [rnrn/s], after Rao (1993). Slow Fairly high High Vp<32
22< v,, <60
60> v,,
2.1.2 Forces Holeyman et al. (1996) defined driveability using forces instead of penetration speed. Driveability is defined as long as the driving force F, generated by the vibro-equipment exceeds the three dynamic resistance forces generated along the shaft R,, toe R, and in the clutch R,, represented by the following expression: Fg, = (Fo+ FL,) 2 X, + R, + R,
Explanations of parameters in Equation (2) are found in Figures 2 and 3. However, the recorded peak forces in the sheet pile head were less than the theoretical value. These results are believed to be related to energy loss: in the hydraulic system, in the connection between sheet pile and clamp, and also in propagation of flexural and torsional energy, not sensed by the sensors that only record axial load. Recommendations concerning choice of the efficiency factor are provided in Moulai-Khatir et al. (1994). The efficiency for field conditions could be estimated in the range of 20 < < < 25 percent according to the authorsSheet pile parameters It is well known that the influence of clutch friction R, on both driveability and the vibrations in the environment is of great importance. At the moment, good research on the influence of the clutch on driveability is being conducted amongst the European sheet pile manufacturers. However, knowledge regarding clutch friction remains limited. Another sheet pile related uncertainty that exists in applying a driveability formula is pile rigidity, which is a reasonable assumption in laboratory tests, can be assumed for sheet piles in field tests. The natural frequency of a longitudinal vibrating rod can be calculated according to Equation (3). When applying Equation (3) on the instrumented sheet pile, with a bar velocity c, = 5,100 m/s, L = 14 m, the first mode (n=l, half wave mode), a longitu-
This definition of driveability is rather simple but at the same time also powerful. However, the key issue is the difficulty in assessing the dynamic soil resistance along the shaft and toe during the installation phase. The forces in Equation (1) acting on the vibro/sheet pile system are schematically described in Figures 2 and 3.
2.2 Vibrator parameters The choice of optimum vibrator parameters concerning driveability has not yet been established. However, choice of driving frequency seems to give rise to two driving states: fast and slow vibratory driving. The two states are mainly due to the fact that the soil beneath the toe moves out of phase rela534
dinal resonance frequency of fi 182 Hz can be calculated.
f,,= il
Cl,
*
__
2L
(72 = 0,1,2,
...)
(3)
The point to be made here is that longitudinal resonance phenomena generally occur at very high frequencies compared to the driving frequencies used in today’s vibratory equipment.
It would be possible to treat the vibro/sheet pile system as a rigid body, if the displacement of the entire sheet pile is in phase with the motion of the vibro-driver. Such an assumption would greatly simplify the problem to a single-degree-of-freedom vibrating system, where the inertia force and driving force are in phase with each other. In laboratory tests by Viking (1998), a rule of thumb was used to treat the system as a rigid body if the criterion t, 2 2, was fulfilled, where t, is the time it takes the unbalanced force F,, to vary from zero to maximum, and t, is the time it takes a stress wave to travel the length equivalent to four times the length L of the pile. A criterion allowing the model pile to be treated as a rigid body if the driving frequency was chosen to be equal to or less that 12.5 percent of the resonant frequency, i.e. f,ls 0.125f,,. In O’Neil1 et al. (1989) it was stated that a fullsized pile could essentially be treated as a rigid body if the driving frequency fll was chosen to be equal to or less that 10 percent of the resonant frequency f,,of the vibro/pile system. However, from the conducted full-scale field tests it was noted that the vibrohheet pile system used did essentially behave as a rigid body when the driving frequency iiwas as high as 25 percent of the resonant frequency of a free vibrating 14 m long steel rod. Viking et al. (2000) related another important sheet pile related parameter to the lateral flexibility of singularly vibro-driven sheet piles. The flexibility is expiained by the fact that the driving force FCienters the sheet pile web at an eccentric distance e from the neutral axis of the sheet pile. The bending moment M = e-F, induced at the sheet pile head results in a lateral sinusoidal motion U , of the sheet pile during the installation phase. The flexibility of the sheet pile was monitored by a laterally mounted acceIerometer at mid-position of the instrumented sheet pile, (see Figure 6). 2.3 Soil parameters The vibratory technique is known to perform best in loose granular soils. It may not work as well in moderately stiff, saturated clays, dense sand or gravel since the dynamic soil resistance docs not decrease as favorably as it does in loose sand during the installation phase. Results from earlier investigations and research by O’Neill et al. (1989), clearly reveal that the three main soil related parameters that significantly affect the driveability are the initial relative density, degree of saturation, and layers with different density. The primary soil mechanisms behind the favorably reduction of penetrative resistance are the local sand liquefaction, the induced cyclic motion of the soil particles during driving, and the relation between initial void ratio and its critical state value. The degree of saturation of sand has been found to be closely related to the recorded global rate of
Figurc 2. Forccs acting on the vibro/soil/shcet pile system.
535
penetration. The rate of penetration has been observed to be higher in saturated sand than in dry sand, but the difference scems to be more pronounced in the case of low displacement pilcs. Recorded pore pressure readings near the pile shaft (in the sand) have been observed to be higher near the pile shaft than those away from the oscillating pile shaft. The pore pressure development during driving is closely related to the driveability. However, studies by O’Neill et al. (1989) could not relate all of the shear strength loss to the pore pressure build-up during driving. This could be explained either by the fact that liquefaction occurs near the oscillating sheet pile, or by shear strength reduction due to the induced cyclic motion of the soil particles during driving, or a combination of both mechanisms. The relative density is the soil parameter that has been reported to have the most significant influence on the global rate of penetration. The rate of penetration has been found to decrease with increasing relative density. The induced volume changes, (densification or dilation) in the soil are believed to give rise to a characteristic shape of the driveability vp-z curve, see Figure lb. The content of fines (greater than 12%) is a neglected soil parameter that usually dictates if the use of vibro-drivers will be successful or not, a fact that has been experienced by several authors, but none gives the underlying exp!anation.
2.4 Pei i elra 1ioii 171CCIINIZi st i t s of
lli c system
The difficulty in applying a driveability formula comes from the fact that a great deal of trade-off must be made between the accuracy of the candidate formula and the practicability of applying it to real problems. This means that the key issue is to access the magnitudes of both the dynamic soil resistance and the friction force R, in the sheet pile clutch. However, the magnitude of the theoretical forcing function of the vibro-driver Fd, i.e. sum of static surcharge force F, and unbalanced force F,, is easier to estimate. It varies sinusoidally in time with driving frequency and eccentric moment according to Figure 3a. The motion of the sheet pile u(t) is described by a downward directed sinusoidal displacement, correlated in time with the driving force Fd, see Figure 3b. The dynamic shaft resistance R, (Figure 3c) varies between positive and negative, correlated with the up- and downward motion of the sheet pile. The dynamic toe resistance R, varies between zero and maximum, also correlated with the motion u(t) of the sheet pile, and reaches it’s peak value at the lower end of the up- and downward penetrative motion u(t) of the sheet pile, see Figure 3d. No theoretical explanation has been presented for how the generated clutch friction R, varies during the penetrative motion. However, it is well known that the dynamic clutch friction R, is of great significance.
Figure 3. Schematic descriptions of the relationship between driving force, penetrative motion, shaft- and toe resistance.
536
3 FIELDTEST The first test site was situated beside the steel-girder bridges over the Fittja Straits at Virby, situated 20 kilometers south-west of Stockholm. The test site was chosen for the driveability study for two main reasons, firstly, due to the relatively homogeneous soil conditions, and secondly, the high probability of keeping the sensors in place. All of the vibro-driven sheet piles were installed in co-operation with the Swedish foundation company Stabilator AB,using a leader mounted ABIvibro, model MRZ 800V, mounted on ABI-leader system TM 14/17L. The driveability tests were executed on a 14 m long instrumented sheet pile described in Figure 6. Three uninstrumented sheet piles was initially driven into the ground with only environmental effects and penetration speed monitored. The instrumented sheet pile was then driven three times by itself and two times into the clutch of the three initially installed sheet piles, Green & Nilsson (2000).
I
Figure 4. Evaluation of the CPT-1 results.
found approximately 2.1 m below ground level, and the soil, which is relatively well graded, varies between silty sand and gravelly sand. The relative density is estimated to lie between 30 < D,. < 50 % from the CPT using relationship from Jamilkowski et al. (1985). The sand is considered to be normally consolidated with respect to its geological history. The mineral composition of the sand shows that it mainly consists of hard minerals, such as quartz and feldspars. The results from the CP1 probing are presented in Figure 4, showing a soil profile that is relatively homogenous with depth.
3. I Soil conditions The soil investigations consisted of soil sampling at six levels, three CPT tests and 21 dynamic probing tests. Moreover, the pore pressure was measured using one piezometer and one open standpipe. Apart from the top two-meter thick clay layer, the soil consists of more than 40 m of loose to medium dense glacial sand. The ground water table was
Figure 5. Schematic of vibro-, soil-, and sheet pile instrumentation.
537
3.2
Itislrritn~iiluliol?
The instrumcntation system employed to record the driveability gcncrated environmental effects consisted of three parts: vibro-, sheet pile-, and soilinstrumentation, (see Figure 6). The instrumentation on the vibro-equipment consisted of three sensors, monitoring the applied static surcharge force (via oil-pressure), penetrative motion of the sheet pile, and the position of eccentric weights (e.g. the varied static moment Me). The instrumentation on the sheet piles consisted of ten Weatstone mounted strain gauge bridges and three accelerometers with position illustrated in Figure 6. The three low-g accelerometers, Analog Devices Inc., were mounted on specially developed printed circuit boards. Prior to the tests, the instruments were calibrated. During the strain gauge calibration, the instrumented sheet pile was loaded at one end by a hydraulic cylinder and held between end plates welded on a 22 m long I-beam. Prior to the driveability tests, the strain gauges were zeroed when the sheet pile was free hanging in the hydraulic clamp before driving it into the ground. The readings from the calibration procedures were compared to standard known calibrations, described in Green & Nilsson (2000). The data acquisition system consisted of a DATtape recorder Sony PC216Ax together with Sony extension unit PCCX32Ax. The recorded dynamic data obtained from the various instruments were later anal yzed, using the acquisition software package PCscan MK IITM,further discussed in Section 3.3.
A
3.3 Methods of analysis of data collected Driving records of the results recorded on the DATrecorder were analyzed using the computer program PCscan MK IITM.Short time windows of the digitized time histories, at any depth of interest, were exported to MS ExcelTMin order to present and graph the different readings against each other. The choice of window size should contain a sufficient number of cycles of the recorded signal, and the sheet pile should penetrate with a uniform velocity in the considered time window. The development of the presented time window of displacement U ( [ ) of the sheet pile is obtained through double integration of the pile head (or near the toe) acceleration records. Then a correction constant c was applied to the recorded acceleration a(t) to ensure that zero average acceleration exist for the considered time window dt = tllt, of interest. The dynamic force-displacement curves were determined from the time histories of forces and double integration of corrected acceleration records. The corrected acceleration a,(t) was first integrated according to Equation (4) in order to obtain the velocity change v(t) in time, (Figure 8a).
538
$ P
Figure 6. Instrumented sheet pile with location of the sensors mounted on the 14 m long PU16 profile, S240 GP.
I,,
V(t) =
p; dt (I)
(4)
1,
The harmonic penetrative motion of the sheet pile, (Figure 8b), is finally obtained by once more integrating the sum of change in velocity, i.e. Equation (4), and the uniform velocity v,,, at the beginning of the time window, I = I , , expressed according Equation (5).
kgm, f' = 41 Hz,in, = 1.670 kg, a,, = 130 m/s' to about = 515 kN. The actual delivered unbalanced force amplitude to the sheet pile head is given by Figure 9a to %-(236+105) = 170 kN, which gives an efficiency of 6 = 100.(170/515) = 33 percent.
---
so
I
U(t) =
j("(+')
v,,)dl
(5)
-100
1,
I
-200 0,65
4 RESULTS
0,70
trme /
fl)
Observation of the time histories of the sheet pilehead and sheet pile-toe forces, acceleration and displacement of full-scale field tests have provided a unique opportunity to develop a better understanding of the penetrative mechanisms earlier addressed in Section 2.4. A few preliminary records of documented force and acceleration time histories of a sheet pile with no clutch friction, at 3.1 m penetration depth are presented in this section in order to discuss and point out some of the significant aspects of the penetrative behavior of vibratory driven sheet piles in clean sand.
-
0,65
14
-- .-
0,70
--__
-
-
0,80
0,75 time [ s 1
Toe acceleration ot sheet pile B1
-200
r-------
-100
-
---
s-=.
_____
0..
loo.
CZ
539
- -
0,so
0,75
1
Iiead acceleration ot sheet pile I31
4.1 Typical time histories The sign convention for the signals is as follows: positive acceleration occurs during the downstroke, positive velocity corresponds to downward movement of the sheet pile, positive force corresponds to compression of the sheet pile. Relationships of both the lateral acceleration a, at together with the axial acceleration of the head a,, and toe a,have been graphed in Figure 7a-c, at 3.1 m penetration depth. From the preliminary results it appears that the phase difference between sheet pile head and toe accelerations is small. The recorded lateral acceleration of the sheet pile, which also appears to be in phase with the axial, shows a lateral displacement amplitude of about fi, = &Jw'= 5 0 / ( 2 . ~ - 4 1 )=~ 1 mm. At first glance, 5, = 1 mm does not appear significant, but it is in fact about 40 % of the axial displacement amplitude. Figure 8a-b graphs the double integration procedure described in Section 3.3, resulting in the penetrative motion u(t) of the sheet pile toe. The time history of the driving force FCiat sheet pile head and the dynamic resistance R, near the toe has been graphed in Figure 9a-b. The theoretically generated unbalanced force amplitude is calculated according to Equation (2), with 77 = 50 %, Me = 12
I
0,65
0,75
0,70
0,so
rime [ s I
c)
I p r e 7a-c. Typical time histories of acccleration signals.
065
~~
0,70
0,75 rbue [ s 1
*,
b)
1
.
Displaccmnt of shcct pile toe U1
0,65
0,70
t i m I sI
0,75
0,m
Igure 8d-b. Intcgratcd velocity and displacement histories.
3.) The sheet pile with head and toe acceleration sensors behaved essentially as a rigid body during the vibratory installation phase. 4.) The description of the penetration mechanisms visualized in Figure 3, shows several similarities with the recorded behavior of full-scale sheet piles under field conditions. 5.) The preliminary results of the ratio of unbalanced sheet pile head force to unbalanced force produced by the vibrator were approximately 33 percent. It should be noted, all sheet piles came directly from the manufacturer, the test site provided favorable soil conditions, and the vibro-equipment was new.
Ilriviiig h r c r ~its1ic.etpile 131
- 1 10
- 60
. < . t
,-i
-10 40 CX)
1.40
1'93 2%
0,70
0.65
(1)
0,75 time [ s ]
0,65
0.75
0,70
0.80
45
. L . 25
+
6 ACKNOWLEDGEMENTS
-15
The writer would like to express his sincere gratitude to The Development Fund of the Swedish Construction Industry (SBUF) together with Stabilator AB for funding the ongoing research project.
-35 Dynamic force in sheet pile webb 731 near the toe
b,
igure 9a-b. Typical time histories of generated forces.
7 REFERENCES 4.3 Force versus displacement
Green, J., Nilsson, C-O., (2000),Drivbarhets- och omgivningspiverkansstudier av spontdrivning med vibroutrustning., MSc. Thesis, Dept. of Civil and Environmental Engineering, Royal Institute of Technology, (in press). Holeyman, Alain E., Legrand. Christian., Van Rompaey, Dirk., (1996), A method to predict the driveability of vibratory driven piles., Proc. of 5th Int. Conf. On the application of Stress-Wave theory to piles., September 11-13, Orlando FL., pp. 1101-1112. Jamilkowski, M., Ladd, C.C., Germaine, J.T., and Lancelotta, R., (1985), New developments in field and laboratory testing of soils., Theme Lecture, Illh Int. Conf. on Soil Mech. & Found. Engng, San Francisco. Moulai-Khatir, Reda., O'Neill, Michael W., Vipulanandan, C., (1994), Program VPDA Wave Equation Analysis for Vibratory Driving of Piles., Report to The U.S.A. Army Corps of Engineers Waterways Experimental Station., UHCE 94-1, Univ. of Houston, Texas, 187 pp. O'Neill Michael W., Vipulanandan, C., (1989), Laboratory evaluation of piles installed with vibratory drivers. National Cooperative Highway Research Program, Report No. 316, National Research Council, Washington, DC. Vol. 1. pp. 151. ISBN 0-309-04613-0. Rao, Pramod M., (1993), Effect of pile geometry and soil saturation in the behavior of non displacement piles installed by vibration., M.Sc. Thesis, Dept. of Civil and Environmental Engineering, University of Houston, Texas. Rodger, A.A., Littlejohn, G.S., (1980), A study of uibrarory driving in granular soils. Geotechnique, Vol. 30, No. 3, pp. 269-293. Viking, K., (1997), Vibratory driven piles and sheet piles -a literature survey., Report 3035, Dept. of Civil and Environmental Engineering, Royal Institute of Technology. Viking, K., (1998), Driveability studies of vibro-driven model piles in non-cohesive soils -laboratory simulations., Licentiate Thesis 2029, Dept. of Civil and Environmental Engineering, Royal Institute of Technology. Viking, K., Green, J., Nilsson, C-O., (2000), Uppmatta markvibrationer av vibroinstallerad spont., 13th Nordic Geotechnical Conf. on Geotechnics and Sustainable Development., Helsinki, June 5 - 7, 2000 (in press).
Further insight into the penetration mechanisms of sheet piles during vibratory driving can be obtained by observing the load transfer relations (RI-Uand RA-ucurves. The load transfer relation of the dynamic toe resistance in the webb at 3.1 m penetration depth has been graphed in Figure 10. Dynamic force in sheet pile webb B1 near the toe 43
3
2 5
5
.
-15
Y
-35 0,11
0.15
0,16
0.15 1'
0,16
0.17
Iml
igure 10. Load transfer curves of sheet pile B1 near the toe.
The negative toe resistance noticed in Figures 9b10, is assumed to be related to inertia effects in combination with a small portion of shaft resistance on the 600 mm long part of the sheet pile below the position of the toe sensors.
5 CONCLUSIONS Based on this study and the range of variables investigated, the following primary conclusions can be advanced: 1.)The presented instrumentation system has proven to successfully register the dynamic soil resistance. 2.) The position of applied sensors made it possible to document field related properties of vibrocapacity and penetrative resistance.
540
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 0 2000 Balkema, Rotterdam, ISBN 90 5809 1503
Accelerations of a driven pile and the surrounding soil Edward L. Hajduk & Samuel G. Paikowsky Geotechnicul Engiiieering Research Laboratop, Deparfment of Civil und Environmental Engineering, University of Massachusetts, Lowell, Mass., USA
Paul Holscher & Frans B. J. Barends GeoDelft, Netherlunds
ABSTRACT The traditional mathematical formulation of pile driving (e g Smith, 1960) assumes that the soil surrounding the penetrating obiect remains stationary while one-dimensional wave propagation takes place in the pile Paikowsky and Chernauskas (1996) examined large data sets of driven piles and suggested that the soil's inertia greatly affects the actual physical phenomenon controlling pile resistance during driving and hence the accuracy of the dynamic methods Due to the fact that soil inertia is not accounted for by most traditional pile driving models, a new, practical model needs to be developed that accounts for this factor in a routine manner The development of a reliable pile-driving model incorporating soil inertia requires an understanding of the relationship between pile penetration and soil motion Holscher (1995) investigated these relationships in a laboratory and field study To enhance the understanding of the problem, the presented research examines the measurements of pile and soil accelerations during the installation of a closed ended steel pipe pile at a bridge reconstruction site in Newbury, Massachusetts The steel pipe pile, part of a test pile cluster investigating time dependent pile capacity in the Boston Area, was instruniented with accelerometers and strain gages at the pile top, middle, and tip Ground accelerometers, piezometers, and total pressure cells were installed in a silty sand layer about 19 5m from the ground surface at various distances from the pile The layout and installation of the pile and soil instrumentation is provided Measurements showing the effect of pile installation as the pile tip passes by the soil instrumentation are shown Analysis of the pile and soil instrumentation measurements showed (i) such measurements are possible and provide insight into the behavior of the soil during pile penetration and (ii) the majority of soil disturbance occurred at or near the pile tip To develop such a model, actual soil acceleration measurements need to be obtained to provide a framework for understanding the soil motion mechanism. Such measurements can also be used to calibrate the models once they have been developed. As part of a joint research effort into this field, the University of Massachusetts Lowell's Geotechnical Engineering Research Laboratory (UML) and GeoDelft (Delft) of Delft, the Netherlands recorded pile and soil accelerations during impact driving of a closed ended steel pipe pile, designated Test Pile #1. Complimentary measurements of pile forces, pore water pressures, and total soil pressures were also recorded. These measurements were taken at the UML Geotechnical Engineering Research Site located in Newbury, Massachusetts. This paper describes the site, the equipment and its installation, and presents measurements of the pile and soil instrumentation as the pile tip passes near the soil instrumentation.
I . BACKGROUND
Typical pile driving modeling used in current engineering practice is based on Smith (1960) and does not take into account movement of the soil during the installation process The soil-pile interaction is merely being presented as stationary forces activated through the pile's displacement Previous research that examined large data sets of driven piles has suggested that soil inertia greatly affects the physical phenomenon that controls the dynamic resistance of the pile. The analysis of these data sets indicated that under high soil inertia conditions (i e low driving resistance of large displacement piles), the performance of wave matching techniques is extremely poor. For updated data analysis. see Paikowsky and Stenersen (2000) The need for a reliable, physically based soil inertia model is therefore evident (Paikowsky and Chernauskas, 1996,).
54 I
2. SITE AND INSTRUMENTATION
Prior to the installation of Test Pile #I, a ground piezometer field, two test piles, and twelve reaction piles had been installed at the Newbury Site. Paikowsky and Hajduk (1999) describe in detail the layout of the site and the various test pile cluster components. Figure 2 shows the layout of the Newbury Site prior to the installation of Test Pile ff 1.
2. I The Newhirry Research Sire The UML Geotechnical Engineering Research (Newbury) Site was located at a bridge reconstruction site along US Route 1 on the Newbury/Newburyport Massachusetts border. An intensive field and laboratory study of the site was conducted to determine the characteristics and engineering parameters of the subsurface (Paikowsky and Chen, 1998). The Newbury Site soil stratigraphy is typical of the conditions found in the Boston area. The general soil profile at the pile testing location (from ground surface downward) consists of the soil strata listed in Table 1. The soil profile is shown in Figure 1.
2.2 Test Pile ff I Test Pile #I was a 32.4cm (12.75in) diameter x 1.3cm (0.5in) wall thickness x 31.4m (103A) long closed ended steel pipe end-bearing pile. The pile was heavily instrumented for monitoring soilstructure interaction over a lengthy time period. Instrumentation installed within the pile relevant to the pile-soil acceleration research included piezoresistive accelerometers and electrical resistance strain gages installed near the pile midpoint and pile tip. These instruments are listed in Table 2 and are shown relative to the soil profile in Figure 1. The piezo-resistive accelerometers were standard dynamic gages assembled and distributed by Pile Dynamics, Inc. (PDI) of Cleveland, Ohio. Only one interior accelerometer was installed at each pile location. The electrical resistance strain gages were Model LWK-06-W250B-350, manufactured by Measurements Group, Raleigh, North Carolina. Two strain gages were installed at each pile location.
Table 1. Newbury site soil profile. Depth Soil Layer (m)
0-2.4 2.4-2.7 2.1-5.4
5.4-1 1.5 11.5-16.4 16.4-19.3 19.3-21.6 2 1.6-23.9 23.9-26.3 26.3-30.5 N.5+
Granular Fill (Cased) Organic silt and Peat O.C. Clay Soft N.C. Clay N.C. Clay Interbedded Silt. Sand. and Clay Silty Sand Interbedded Silt. Sand. and Clay Fine to Mehuni Sand Till Bedrock
Figure 2 . Newbury site layout.
Table 2. Test Pile #1 Instrumentation. Gage Type Gage Designation Piezo-resistive 1-8-APCB-A Accelerometers (APCB) 1-16-APCB-A Electrical Resistance 1-8-ERSG-A&B Strain Gages (ERSG) 1-S-ERSG-A&B
542
Depth from Pile Top 14.5m 30.2m 14.5m 30.2m
A Pile Driving Analyzer (PDA), also manufactured by PDI, recorded measurements of the interior accelerometers. Each accelerometer was recorded at a frequency of IOkHz for a span of 0 1023 seconds A CR9000 measurement and control system, manufactured by Campbell Scientific of Logan, Utah, recorded the interior strain gages at a frequency of lOkHz over a time span of 05 seconds In addition to the instrumentation installed within Test Pile #1, dynamic transducers consisting of piezo-electric accelerometers and strain gages were attached to the pile top during driving These gages were manufactured by Pile Dynamics, Incorporated of Cleveland, Ohio and were recorded at the same frequency as the interior accelerometers with the same Pile Driving Analyzer Test Pile # I was installed with a Delmag D30-32 single ended diesel hammer on May 29, 1997 2.3 Iiiili'imiet7fntroi~ .for Re rp0t7se
Motirtoi.iiig the
SOI/'.I
A variety of transducers were installed around the planned location of TP#l to record pore pressure, total pressure, and ground accelerations within the soil To allow comparison with previous soil acceleration measurements conducted in cohesionless soil by Delft (Holscher, 1995), it was decided to install the ground accelerometers and total pressure cells within the silty sand layer (depth 19 3 - 21 6m) Three vibrating wire ground piezometers were already installed within this layer as part of the test pile cluster at various distances from the planned location of Test Pile # I A summary of the soil instrumentation around Test Pile #1 in the silty sand layer is listed in Table 3 Table 3 Soil iiistruiiieiitahon v ithiii the silt! sand la!er Elev. Depth Distance Gilge Gage Name fromTP#l (m) T] pe (m) (m) r/R -1566 21.24 0.61 3 7 5 Glid PZ-8 Vibrating 2.19 13.53 -14.0s 19.66 Gild PZ-9 Wirc Piciomctcr Grid PZ- 1 0 2.22 1.3.68 -14.75 20.33 Ground
Acccleroiiietcrs Total Prcs-
siirc Cells
GA-1 GA-2 GA-3 TPC- 1 TPC-2 TPC-3
1 40 2 30 2 93 124 1 90 2 97
8.64 14 17 1809 764 11 75 1832
-13.92 -13 76 -13 76 -1407 -1398 -1388
19.50 1934 I934 1965 1956 I946
KC! I =
Distaiicc froin the pile centcr
R
= Pile
radius
The ground piezometers were Model 4500 vibrating wire pressure transducers manufactured by Geokon Incorporated of Lebanon, New Hampshire Each piezometer was connected to Micro-10 Dataloggers, also manufactured by Geokon. which re543
corded the gages at a frequency of 1 reading /30 seconds throughout the pile installation Paikowsky and Hajduk (1999) provide details of the ground piezometers and their installation The ground accelerations in the deep soil were measured using s e r o accelerometers manufactured by Sundstrand Data Control Inc of Redmond, Washington. These instruments have a resonance frequency of 600Hz, and are therefore filtered by an analog 3'd order filter with a -3dB point at 150 Hz The amplification from 0 to lOOHz is 1 Each ground accelerometer (GA) instrument consisted of a cluster of three accelerometers perpendicular to each other placed in one special cone which allowed for measurements along three orthogonal axes The total pressure cells (TPC) in the soil were composed of PR-transducers manufactured by Druck Limited of Leicester, England The signals are amplified by using HBM amplifiers Measurements of the ground accelerometers and total pressure cells were stored after A/D conversion within a portable computer Data from each GA and TPC was sampled at a frequency of I M z Total signal length for each gage was 2 4 sec, including the pre-trigger time of 0 5 sec Due to the crowded nature of the test site, the ground accelerometers and total pressure cells were installed outside of the reaction piles to (i) prevent disturbance to the installed instrumentation and (ii) allow room for a drill rig to install the equipment Figure 2 shows the locations of the casings and gages relative to the site layout Each ground accelerometer set and total pressure cell was installed into the soil at an approximate angle of 10 degrees from the vertical Installation consisted of placing 1Ocm (4in) OD by 13 7m (45fi) long steel casing through the clay layer using standard boring techniques Accurate location of the gages was achieved using inclinometers to measure vertical and horizontal deviation of the casing relative to the planned location of Test Pile # I and detailed surveying measurements The gages were then pushed into place at the end of standard AW drill rods To insure that the rods were centered within the casing, 3 metal guides were welded to the rods at 120' intervals around the rod diameter at 1 5ni (Sft) increments A typical installation cross-section is shown in Figure 3 The interior of each casing was sealed with bentonite slurry to prevent the boreholes from acting as drain holes, which would affect adjacent pore pressure dissipation measurements Final elevations and distances from TP#l for the various soil gages were determined based on the measurements made during installation and the final surveyed position of TP#l after driving and are summarized in Table 3
Figure 3 TJ pica1 ground accelerometer (GA) arid ~olalpresw e cell (TPC) iiistalla~ioncross-~ectioii
3 . MEAS U REMENTS
Analysis of the pile and soil gage data focused on rneasurements taken during passage of the pile tip approximately 3-5m above and below the soil instrumentation locations This corresponded to a range of soil starting at the interbedded layer (depth 16 3 lm) and ending within the fine to medium sand layer (depth 25 58111) 4. PILE RESPONSE
Problems experienced during Test Pile # 1 installation limited pile acceleration measurements at both the pile iniddle (gage 1 -8-APCB-A) and pile tip (gage 1 16-APCB-A). Gage 1 - 16-APCB-A experienced equipment failure at pile tip depth of 18.90m, directly above the silty sand layer Data acquisition system problems with the PDA resulted in a gap in pile middle acceleration measurements between depths 20.42 and 23.48m. Measurernetits of the dynamic gages at the pile top and the interior electrical resistance strain gages were continuous throughout the driving of Test Pile,# 1 near the soil instnmentation. Permanent set for each blow was determined through a detailed analysis of a video taken during the installation of Test Pile # 1 , Figure 4 shows these measurements converted into driving resistance for. pile penetration around the silty sand layer. 544
Typical measurements of the pile forces and accelerations and the corresponding velocities and displacements with time for a single blow are shown in Figure 5. Figure 5 shows the data for Blow 107 (pile tip depth of 18.901-11) while the pile driving resistance was 1 .Oblows per I Ocni (B 1 Ocm). Figure 5 shows the pile tip experiencing significant acceleration with corresponding high velocity and displacement during driving near the soil instr-umentation. The measurements of high acceleration and displacement at the pile tip, coupled with the low driving resistance of the pile through the siltp sand layer as shown in Figwe 4, make this pile ideal for studying the efTects of soil inertia. Examination of the pile measurements during penetration at and near the silty sand layer showed a consistent drop between the maximum pile top, pile middle, and pile tip accelerations. Figure 6 s h o w the maximuin pile middle and tip accelerations normalized to maximum pile top acceleration with depth. The ratio between pile middle and pile top accelerations was 0.85 0.07, while the ratio between pile tip and pile top was 0.58 7t 0.05 (mean Sr 1 s.d.). This consistent range suggests that a reasonable estimation of pile tip acceleration can be made fi-on1the measured values at the pile top. The ratios presented in Figure 6 are not necessarily representative of standard piles.
*
5 . SOIL RESPONSE 5.1 Soil Veiocip Meastiremenfs Acceleration measurements taken by the ground accelerometers (GA) were processed in the following manner:
o
e
e
Signals fiom the three individual accelerometers within each GA were rotated to an axis system with a pure vertical axis. This produced transformed axes labeled as vertical (parallel to the pile), horizontal (perpendicular to the vertical axis of the pile), and tangential (perpendicular to the plane through the pile and transducer location). The transformed axes are shown in Figure 7. The known slope of the transducers, measured during gage installation, allows for this transformation. The signals are rotated around the vertical axis to reduce tangential accelerations. This rotation is kept independent of pile depth. The signals were inteyrated with respect to time.
Figurc 6. Nomializcd maximum pilc accclcratioii measurcinents with depth.
'i
--
I
.-_-
I
I
Vertical
I
I
1
-I
-
Horizontal Axis
PILE Figure 7 Transfomied ground accelerometer axes
A sign convention was established to identify absolute vertical, horizontal and tangential measurements. The sign convention is shown in Figure 7 . Insight into the behavior of the soil around the pile tip during driving was achieved by examining the velocity vs. time records for individual blows as the pile tip passes the ground accelerometers. Figure 8 presents vertical velocities vs. time for several selected blows. The tinie for each blow in Figure 8 was adjusted to allow for the start of the velocity changes to be located at approximately the same time The data of Figure 8 shows that the location of GAI is outside the major deformation of soil rnovement (plastic flow) and therefore the measurements are in the shape of a disturbing wave The upward (neyative) velocities are larger as the pile tip is above the point of measurement. As the pile tip approaches and passes the gage elevation, the absolute peak velocities changes relation with regards t o the sign,
Figure 5. Tl;pical Tcst Pile #1 measurements (blow 107. dcpth 18.9111).
545
with the upward velocities being larger, indicating the soil is moving primarily downward (positive). Figure 9 was prepared based on the data shown in
Figure 8. The maximum and minimum vertical, horizontal, and tangential velocities are presented as fUnCtiOnS Of the Vertical distance from CA1 to the pile's tip and depth. Figure 10 presents the measurement results from GA2 in an identical format. Inspection of Figure 9 and 10 reveals that the maximum and minimum tangential components remain nearly constant over the examined region, which shows that the applied rotation of the axes was satisfactory. Inspection of the variation of the vertical and horizontal velocities of GA1 and GA2 with the pile penetration needs to consider the location of the accelerometer relative to the pile's tip. As GA1 and GA2 are 8.64 and 14.17 radii from the pile's skin, it seems that the start of the major influence can be identified when the pile tip corner is approximately 9.4 radii (1.52m) diagonally from GA1. At this point the vertical component of the velocity is 8.80 m d s e c downward and the horizontal component is 3 .80 m d s e c , approximately equal to the tangential velocity. The velocities at GA2 are insignificant at that point, suggesting that the outside boundary of ,the moving soil in a plane below or normal to the pile tip, is between 9 to 14 radii away from the shaft The peak velocities of GA 1 and GA2 take place when the pile tip passes by about 1 to 2 radii the horizontal plane on which the accelerometers are located. At this point, the maximum soil velocity at GA1 is 41.20 m d s e c (directed downward) and the corresponding outward horizontal velocity is 24.40 m d s e c . 5.2 A ccwacy of the Soil Acceleration Measurements
Figure S. Vertical velocit_v with time measurements for selected blows of CA1
Analysis of the soil acceleration measurements showed that the frequency content observed in the Newbury testing was much higher than the previous
Figure 9. Maximum and minimum measured velocities with depth for GA1.
Figure 10. Maximum and minimum measued velocities with depth for GA2.
546
testing conducted in Del&, Holland (Holscher and Barends, 1996). This observation raised questions concerning the ability of the ground accelerometers to accurately record the soil accelerations generated by the pile penetration. As stated previously, a 3rd order analog filter was used to attenuate frequencies greater than 1OOHz in the ground accelerometers. If the ground acceleration frequencies were greater than the filter frequency, the recorded accelerations and subsequent velocity and displacement integrations would be inaccurate. To determine if the frequencies in the soil were greater than 1 OOHz, the accelerations in the pile were examined. Figure I 1 presents the accelerations at the pile top and tip for blow 93 (vertical distance to GA1 of 1.85m). A Fourier transform analysis was performed on these two acceleration measurements to determine the frequencies present within the waveforms. The results of the Fourier transforms are presented in Figure 1 1. The results presented in Figure 11 show that at the pile top, the acceleration waveform largely consists of a range of frequencies between 1 to 3000Hz. At the pile tip the acceleration wave is primarily comprised of frequencies between 1 to 600Hz. This substantial reduction suggests that the soil attenuates the acceleration wave as it travels down the pile. Although these observations do not reveal if high fiequency signals are present within the ground acceleration measurements, they suggest that a reduction
Figure 12. Pore and total pressure measurements \\4th depth for Gnd PZ-8 and TPC I.
in frequency content would occur within the acceleration wave in the soil mass. If these frequency reductions did occur, then the ground acceleration measurements should have accurately recorded the behavior of the soil mass due to pile penetration. 5.3 Total Sod and Pore Water Pressure Measurements.
The increase of pore water and total soil pressure with respect to pile tip penetration for gages Gnd PZ-8 and TPC- 1 are shown in Figure 12 The data presented in Figure 12 shows that the increases in pore and total pressures in the soil are caused by the passing of the pile tip AAer the tip has passed, any excess pore water and total pressures generated begin to dissipate These measurements reinforce the observations of the ground accelerometer measurements that the soil surrounding the pile is most affected between an approximate range of Im below to 2m above the pile tip (approximately 3 to 6 pile radii) The ground piezometers measurements were affected by (i) disturbance of the vibrating wire within the gages caused by pile penetration, which led to extreme gage readings, (ii) the fast dissipation of excess pore water pressure in the granular material relative to the slow measurement frequency of the vibrating wire gages, and (iii) pauses in pile driving Extreme gage readings were edited from the data appear as gaps in Figure 12 Pile driving operations were stopped at various intervals to remove pile piezometer caps along the pile and account for the multiple readings at various depths The relatively low measurement frequency of the ground piezometers, caused by limitations of the gage type and data acquisition system, did not allow for detailed pore pressure measurements per blow
Figure 1 1. Typical test pile acceleration measurements and corresponding Fourier transform analysis.
547
CONCLUSIONS
REFERENCES
Based on the measurements presented in this paper, the following conclusions are made: Measurements of pile and soil forces, accelerations, pore and total pressures during pile driving are possible and provide insight into the behavior of the soil during pile penetration. Over the range of pile penetration examined, the majority of soil disturbance occurred at or near the pile tip. Changes in total and pore water pressure around a driven pile are caused primarily by movement of the soil around the pile tip. The accuracy of the ground acceleration measurements and corresponding velocity and displacement integrations at the Newbury Site may have been affected by the filtering of high frequencies. To avoid this problem, ground accelerometers in fbture research into soil accelerations around driven piles should be capable of recording data at the same range of frequencies generated within the pile. Examination of the pile measurements showed that the frequency content of the acceleration wave decreased as the wave traveled through the pile. This attenuation suggests that the frequencies within the soil mass are also attenuated in relation to those generated within the pile. Additional, detailed examination of the Newbury Site measurements is suggested in order to assist in planning fbture research to develop a reliable, physically based soil inertia model to incorporate into pile driving formulations.
Holscher, P. 1995. Dynamical response of satiwated and drv soils. Delft: Delft University Press. Holscher, P. and Barends, F.B.J. 1996. In-situ Measurement of Soil-motion near the Toe of a dynamically Loaded Pile. In F.C. Townsend, M. Hussein. & M.C. McVay (4).Proceedings3 ShInternationaI Coq/erence 011 tlie Applicatioii of Stress-Wave Theory to Piles, 11-13 September, 1996: 26-36. Orlando, Florida. Paikowsky. S.G. and Chen. Y.L. 1998. Field and Laboratory Testing of the Physical Characteristics of the Subsurface at the Newbury Bridge Site. Research Report subirtitted to the Massachusetts Highwav Department, January 1998. Boston, Massachusetts. Paikowsky, S.G. and Chernauskas. L. 1996. Soil Inertia and Use of Pseudo Viscous Damping Parameters. In F.C. Townsend, M. Hussein, & M.C. McVay (ed), Proceedings, 5IhInternational Colf i r m c e on the Application qf StressWave Theorv to Piles, 11-13 September, 1996: 203-216. Orlando. Florida. Paikowsky, S.G. and Hajduk. E.L. 1999. Design and Construction of an Instrumented Test Pile Cluster. Research Report submitted to the A fassachusetis Highwqv Department, September 1999. Boston, Massachusetts. Paikowsky, S.G. and Stenersen. K. 2000. The performance of the dynamic methods, their controlling parameters and AASHTO 2001 Deep Foundation Specifications. Proceeditigs, dhInternational Cotfirence on the .Ipplication of Stress-Wave Theory to Piles, 11-13 September, 2000. S I i Pado City. Brazil. Smith. E. 1960. Pile Driving Analysis by the Wave Equation. Journal of Soil Mechanics and Foundations, Aniericaii Society of Civil Engineers: 3 5-6 1. A irgust 1960.
ACKNOWLEDGEMENTS The presented research was a joint venture between the University of Massachusetts Lowell Geotechcal Engineering Research Laboratory, Lowell, Massachusetts and GeoDelR, DelR, the Netherlands as an additional element to an ongoing research project. Funding for this project was provided by the Massachusetts Highway Department (Test Pile #1 and ground piezometer field installation and measurements, site support and maintenance) and GeoDelR (ground accelerometer and total pressure cell installations, measurements, and removal). The authors would like to thank the contributions of Fokke Wijnstra, Tom van der Poel, and Dr. Frans Barends of GeoDelR, Nabil Hourani and John Pettis of the Massachusetts Highway Department, Car1 Ealy and Al DiMillio of the Federal Highway Administration, and Gary Howe of the University of Massachusetts Lowell.
548
9 Statnamic and other similar techniques
This Page Intentionally Left Blank
Application ofStress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Keynote lecture: Statnamic the engineering of art €? Middendorp TNO Psofouiid, Rijswijk, Netherlands
ABSTRACT: In present standard engineering practice there seems to be a big gap between engineering and art. It is the experience of the author that engineering and art can exist as an excellent combination. This will be illustrated by examples from history and the author’s personal experience. One example will be treated extensively: the continuous development of the Statnamic load testing method as a marvelous combination of engineering and art. The start of the Statnamic concept is described as an interaction between an artist and engineers together with developments on the theoretical approaches and technical applications. Further the present Statnamic state of the art will be discussed briefly.
1 INTRODUCTION
ham, the inventor of Statnamic and the nowadays President of the Berminghammer company.
Art was of minor interest to the author when he was starting his engineering study at the Technical University. Within a short time his interest got a strong impulse when he met his present wife, a painter and artist, at that time studying at the Royal Academy of Art in The Hague. After assisting here with some projects he experienced that the “logical” thinking of engineers is in no way superior to the “associative and intuitive” thinking of artist in finding practical solutions, but that both are complimentary and when combined into an artist-engineer as a person or a team, can result in marvels.
During his career the author was impressed by the many creative solutions of engineers all over the
In the 15‘hcentury the “artist-engineer”, was a socially prominent and respected figure, commissioned by powerful and wealthy patrons, well paid and often regarded as one of the brightest ornaments in sovereign courts. The most famous example of course is Leonardo da Vinci (P.Galluzi, 1996). Because of cultural changes and specialization a gap has been generated between engineering and arts and most engineers are not aware nowadays of their artist-engineer forefathers. Still a strong interest of artists for engineering can be observed in modern art, for example Panamerenko (1996). The artist-engineers are still among us and it was the privilege of the author to cooperate for long period with one of them: Patrick Beming-
Figure 1. Leonardo da Vinci. Automatic file-making machine
551
idea of Fellenius at that time was that dynamic pile testing would become independent of the piling contractor, the pile driving rig, and in many instances contractor’s unionized labor, as well make him and others free to perform dynamic load testing after a good and long set up time. This idea was not original because several such devices were already around, but Fellenius just needed a local practical tool.
world. Many engineers are not aware that their creative solutions can be considered art and that they act
At that time Bermingham had just finished a professional education as a sculpturer in London and worked also for the Bermingham pile driving and hammer manufacturing company. From his childhood on Patrick was interested in both engineering and art and supplied, for example, several ideas for improvement to the Berminghammer pile driving hammers. It was not such a strange idea that the Berminghammer Company chose Patrick to come up with a design for a drop hammer. Patrick contacted Fellenius and suggested a pile loading system design from a fully different viewpoint compared to standard engineers. “Why do engineers want to drop the weight, why do they not send it up into the air?” Intuitively he converted Newton’s Law from Figure 2. Early sketch of a Statnamic piston and cylinder arrangement by Patrick Bermingham
Force = Mass times Acceleration To
as “engineer-artist”. In this paper the author wants to illustrate his view by using the Statnamic development as an example in which one “artist-engineer’’ and many “engineer-artist” contributed with creative ideas and solutions.
Mass times Acceleration = Force (Load) This concept needed a few months to evolve and
The Statnamic development will be described by introducing the start of Statnamic together with some milestones, theoretical approaches and technical applications. Further the present research and developments will be mentioned and problems that still have to be solved.
2 THE STATNAMIC CONCEPT Patrick Bermingham (1998) got the first idea about the Statnamic concept in Hamilton in 1985 while watching a static load test with kentledge, when he first thought about utilizing the inertia of the kentledge (Fig. 2). According to Fellenius (1995) the idea for the Statnamic concept was born in 1987 when he asked Patrick Bermingham to design a drop hammer for impacting a pile to perform dynamic load tests. The
Figure 3. First Statnamic device with accelerometer and early catch mechanism. 552
Bermingham made a first prototype (Figure 3) and started experiments by shooting masses upwards in Hamilton, Ontario in April 1988. He determined the feasibility of accelerating a mass upwards from the top of the foundation rather than dropping a mass onto the foundation
ist-engineer” can be observed, especially the gravel catching mechanism based on the reliable gravity of the earth (Fig.4). With the gravel mechanism a gravel container is placed around the reaction mass and the space between gravel container and reaction mass is filled with gravel. During testing four successive stages can be distinguished. In stage 1 the Statnamic device is ready for launching. In stage 2 the reaction mass is launched upwards by highpressure gases. During this stage the pile is loaded and a Statnamic test performed. Because of the momentum the reaction mass will remain moving upward in stage 3 and the gravel will flow under the reaction mass and over the pile head because of gravity. In stage 4 the reaction mass will fall back and will be caught by the gravel inside the container and the impact load will not be transferred to the pile head but to the subsoil. This creative, simple and elegant principle is still applied as one of the methods in absorbing the energy from the falling reaction mass. Bermingham presented his results and ideas to several parties and also to the author at the OTC (Offshore Technology Conference) at Houston in 1988. Based on the combination of his engineering background and experience with ideas of artists, the author immediately recognized the beauty and power of Bermingham’s idea for pile testing applications. Berminghammer and TNO agreed to start a joint development and decided to do the first prototype testing immediately after the Third Stress Wave Conference at Ottawa in 1988. With the help of Fokke Reiding and Matthew Janes they realized that the long duration feature of the load allowed a fully different approach in instrumentation and analysis compared to dynamic load testing.
Figure 4. First Statnamic trial tests. Bermingham also tried out several concepts of catching the reaction mass when falling back from launching and also here creative ideas from the “art-
It was decided to base the load measurement on a calibrated load cell, to make the measurement independent from pile material properties and to measure displacement directly by the use of an electronic theodolite. The electronic theodolite was a rather expensive instrument and a new tool for measuring displacement was developed based on a laser and a laser sensor, which is still in use. So the basis of measurements became load-time signals and displacement-time signals similar to static load testing. 3 FIRST DEVELOPMENTS In May of 1988 the first model tests where performed with instrumentation provided by T’NO. These first two days of testing confirmed the ability of the very small Statnamic device to produce loads of up to 5 tons with duration of up to 30 ms. From
Figure 5 . Successive stages of Statnamic Testing. 553
also proved the simplicity and practicality of the system in the field and the first load test comparisons proved an unexpectedly close agreement with conventional static load tests. The success of this first prototype enabled Berminghammer to manufacture of a full-scale 5MN tester.
this point onwards the direction of Statnamic was upwards. A second model was built and sent to TNO in Holland where the instrumentation would be developed. In the laboratory at TNO, Statnamic tests were performed using a calibrated load cell and a new laser measuring system developed specifically for the Statnamic test. Both measuring systems worked very well the first day and they have remained virtually unchanged to this date. The next step was to build a Statnamic tester, which incorporated the instrumentation and was large enough to test a real pile in the ground. A 0.6 MN tester was built, which was one tenth of full scale, but still able to test small piles driven into real soil. This load-testing device was first used to test piles in the Berminghammer yard in Hamilton, McMaster University, and Ashbridges Bay. Since that time it has performed tests in Europe, Japan, and the United States. The primary objective of this equipment was to prove the durability of the instrumentation in all weather conditions, and to prove the practicality of the equipment in the field. This equipment was also used to make the first comparisons between conventional static load tests and the new load test method. The 0.6 MN device proved that Statnamic testing could be performed in all types of adverse weather including rain and snow. It
Figure 7. Set up of a 5MN Statnamic device with gravel catch mechanism Statnamic was first called Inertial load testing (Bermingham, P., et all., 1989. The author gave the method its present name Statnamic, realizing that the method was positioned between Static load testFrom the very beginning Statnamic was an international development rather than a regional or national development. Testing of driven and cast insitu piles was carried out in Canada, Holland, Germany and the United States during the first two years. At this time all of the testing was conducted with the aim of gaining a better understanding of the behavior of piles subjected to very quick loading cycles. Statnamic and static load tests were conducted side by side as well as on the same pile in as many different soil types as possible. Every effort was made to collect as much data as possible and to avoid making predictions about the static behavior of a foundation until we could collect a wide range of test results. Today many companies and universities are still collecting and expanding this worldwide database. The first two years of research revealed a great deal about pile behavior when subjected to a Statnamic load of 120ms duration. It was observed that in the elastic range there was a very good agreement between static load deflection and Statnamic load deflection, it was observed that in very soft soils and clays it was possible to apply a much larger load
Figure 6. Patrick Bermingharn launching a 0.6MN device 554
than a static load prior to plunging the pile. In stiff non-cohesive soils and rock sockets it was observed that there was very close agreement between Statnamic load test results and static load tests performed along side. It was also observed that the sequence of loading a foundation had a great effect on the perceived similarity of test results and this had to be taken into account. It was also observed that during a typical Statnamic test the pile would reach maximum displacement at some time after peak force had been applied. In other words the pile would continue to move downwards while the applied load at the pile top was decreasing. At the point where the pile reached maximum displacement the velocity of the pile was zero and then the pile would begin to rebound as the load was further decreased. This observation which was present in nearly all test results except very stiff piles on rock, lead to the development of the Unloading Point Method (UPM) by the author (P. Middendorp, 1992). 4 METHODS OF ANALYSIS
From the very start of development there has been a determined effort to make Statnamic a means of measuring rather than a prediction method. This has meant putting a very strong emphasis on using accurate measuring equipment and recording equipment. The measured data will then be more reliable and may then be examined more closely. From the beginning we have been observing the behavior of foundations subjected to very rapid loading with a view to being able to better understand the mechanism of failure during a Statnamic test. In the end it is hoped that Statnamic testing will stand alone as a rapid test with a distinct method of analysis, which will measure the load deflection behavior and determine the factor of safety of the foundation.
The first approach to analyzing failure was to look to the displacement curve and to analyze the rate of change of displacement, or velocity of the pile. Normalizing the load and plotting load vs. velocity was examined in an effort to pinpoint the load at which the velocity begins to increase. This only worked well when the foundation experienced a plunging failure, and it did not work well when the pile was in a cohesive material. Statnamic test results were also evaluated with a simple 2.5mm offset method, which was analogous to the Davisson failure criterion but much more conservative. All three of these methods of determining the point of failure were far too subjective to be of any great value. In January of 1993, while reviewing the results of pile 7 at Texas A&M the author noticed that during the unloading of the test the velocity of the shaft reached zero at a load, which corresponded closely to the ultimate static resistance. The foundation began to rebound as the load was further decreased. PDA users had observed the significance of the point of zero velocity in the 1970's and some attempts were made to make use of it. However, during pile driving the point of zero velocity at the pile head does not correspond to zero velocity anywhere else in the pile unless the pile is very short and rigid. The author's observation provided both a practical means of determining a significant point on the static load displacement curve and also a means of estimating the damping coefficient directly from the test results rather than from a soil boring. This Unloading Point Method (UPM) assumed that the damping was a constant, which was zero when the velocity was zero, and that the pile was behaving as an elastic body, which could be treated as a lumped
Initially no attempt was made to convert the results of Statnamic load testing into quasi-static load test results, because they would loose integrity in the process. What was recognized was that every Statnamic test result was unique and that very small differences in the relative stiffness of two different foundations could be measured accurately. Much like the dynamic resistance of a driven pile, it is very useful even though there is no direct correlation to static resistance. The Statnamic test has been described as applying a controlled strain while monitoring corresponding deflection. When a test is performed, a predetermined load is applied and the resulting deflection is measured.
Figure 8. Stress wave influences as function of wave number N,
555
measure of toe displacement is desirable. All results are based on measured quantities.
mass and a spring. Subsequent research has concentrated on testing these assumptions and determining the limit of their validity. The unloading point method has provided a very simple universal method of analysis and was first published at the 4ti1International Conference on Stress Waves in 1992. The basic principles of the method have been presented in the Appendix.
F =E,E,A, where F; is the measured force at level i, E; is the measured strain at level i (typically an average of all strain gages at level i), E, is the calculated (or assumed) elastic modulus at level i, and A; is the calculated (or assumed) area at level i
The UPM method is based on the assumption that stress wave phenomena can be neglected. The author studied the validity of the method with the stress wave program TNOWAVE (1996) by varying the pile length with constant load duration. To quantify the stress wave influence he assumed a wave num= D/L, in which D = c.T and T the ber constant N,,, duration of the load, c the stress wave velocity and L the pile length. In this way it was possible to indicate when stress wave phenomena could be neglected and when they should be taken into account.
where U ; is the calculated displacement at level i, L; is the length of the pile segment between levels i and i+l
A valuable extension to the UPM method is the "Modified UPM" (M-UPM) by Justason ( 1997). The method simply involves the averaging of the top and toe velocity and acceleration for calculating the inertia and damping. The method can be applied to any length of pile, but becomes more necessary as the pile becomes longer (low N, numbers). The standard UPM method assumes that pile top velocity and pile toe velocity are in the same range. The M-UPM method is particularly useful when the pile top and pile toe velocity are not in the same range (elastic pile, high toe resistance). Averaging the pile top and pile toe velocities and accelerations yields more accurate inertia and damping forces. The method yields the best results when used in conjunction with an embedded toe accelerometer.
-
wherev, is the first derivative with respect to time of the average displacement for the pile segment between levels i and i+l __
dv. a.I dt -
I
-
-
where a, is the first derivative with respect to time of J The Unloading Point method is performed on each pile segment using the following equation:
F - F - , -5'; -civ, =mja,
Prof. Gray Mullins of the University of South Florida made an additional improvement to the MUPM method, the "Segmental Unloading Point" SUPM. This method uses measured strain gage data to separate the pile into "segments" and perform an M-IJPM on each segment. The data for each segment are added together to produce a total "derived static" load-displacement for the top of the pile. The S-UPM can be applied to any pile, so long as the pile has strain gages distributed over the pile shaft. The first application was the Taipei Financial Center in Taiwan - 1999.
where S; is the equivalent static force for the segment between F, and F,.)
S, represents the friction forces on the each pile segment, with the exception of the bottom pile segment, which also has some component of end bearing. in; is the mass of the pile segment between i and i- 1. The cumulative derived static force at each level can be calculated by the following equation:
The S-UPM method is briefly described below. The Segmental Unloading Point Method extends the applicability of M-UPM to long piles. All assumptions of the Unloading Point Method remain valid. The Segmental Method assumes each segment of a pile behaves as a single degree of freedom system. The method requires embedded strain gauge data. A
1=I
where n is the pile level number, and FSTAT,,is the cumulative derived static force at each level.
556
Hydraulic catching systems eliminate the need for gravel and gravel structure since the upward moving reaction masses are caught at the top of their flight by four hydraulic actuators (or rams). These 3.2m stroke rams are activated by four low pressure (100 bar) nitrogen accumulators, which store compressed nitrogen gas over hydraulic oil. As the weight on the rams is released during a test, the compressed nitrogen quickly expands to force hydraulic oil into the rams causing them to chase the reaction masses to the apex of their flight. The hydraulic oil is routed (in series) through one-way valves at the base of each ram, which restricts reverse flow, and thus the downward movement of the masses. Each of the four rams is independent of the others providing redundancy and safety. The masses remain at this position until the user redirects the additional fluid in the rams back into the accumulators. At which time, a subsequent load cycle can be performed.
The derived static load displacement curve can be drawn at each strain gage elevation using FSTAT,~ and U,.
Each of the above variables represents an entire data set measured over time. The S-UPM was first used for the Taipei Financial Center for 80m piles in 1999
5 THE HYDRAULIC CATCH MECHANISM
By transferring the initial weight of the masses to the rams at the onset of the test it is possible to perform Statnamic testing without a pre-load condition. Additionally, hydraulic catching systems have no minimum required jumg-height for the silencerreaction mass assembly, which is a concern for gravel catching systems. By removing this restriction, low load tests can be performed with much greater than 5% reaction mass. Such tests can produce long duration load pulses greater than 0.5 seconds, thus reducing inertial and damping forces for large portions of the test. Although the set-up time for a 4MN gravel or hydraulic catching systems is comparable, multiple cycles can be performed in a matter of minutes when using the latter. Further, the breakdown typically takes less time. In using gravel-catching systems, great care is exercised in the preparation of the ignition circuitry. An inadequate igniter connection could cost a project as much as a day of delay time. This of little concern when using the hydraulic catching system due to the ability to raise the entire stack of reaction masses with the hydraulic rams so as to access the fuel basket.
Figure 9. Statnamic device with hydraulic catch mechanism (4 MN)
The foundation industry not only wanted to perform larger Statnamic tests but also more of them and at a higher frequency. Also in this case Patrick Bermingham and design engineers came up with a creative solution. In 1995, the hydraulic catching mechanism was built to provide a means of testing, without using the conventional gravel container, or gravel. This simple piece of equipment makes it possible to test up to ten individual piles and to perform multiple load-cycles. The catch mechanism provides the luxury of multiple load cycles within a matter of minutes, the ability to inspect the ignition circuit without disassembly, the benefit of single truck mobilization, and its avoids the environmental problems with gravel retrieval with testing over water.
A substantial portion of all Statnamic testing costs stems from the mobilization of equipment. Typically, a 4 MN test requires two tractor-trailers to ship the combined weight of the equipment and reaction masses (27,000 kg total) where only 20,000 kg is permitted per truck in the USA (30,000 kg in Europe). The device can be equipped with two reaction mass options: (1) an entire set of six concretefilled steel masses, which requires two trucks to ship, or (2) a set of six empty, structurally reinforced steel cans. The empty cans option allows single truck mobilization to distant sites with a total shipped mass of 19,000 kg. Once at the site the cans 557
can be filled with sand, gravel, water or any combination to attain the required mass 3MN and 4MN hydraulic catching systems are now in use in the United Kingdom, the USA and the Netherlands for 3MN and 4MN systems. In 2000 a mechanical catch mechanism will be constructed for a 16 MN device.
6 BATTER PILE TESTING Drop hammers and dead weight static load tests are fully dependent on gravity. One of the big advantages of Statnamic is its independence of gravity because generating the load it is based on inertia forces. This means that the test can be performed in any direction: under batter, lateral and even allows to perform a tension test on a pile. In Figure 9 an example of the application of batter pile testing is presented. It is almost impossible to perform static load test in this over water pile testing situation. The flight of the reaction mass is guided by a support beam
7 LATERAL LOAD TESTING Lateral STN testing is becoming popular in the USA and was strongly encouraged by Barry Berkowitz of the FI-IWA. The first lateral test with a large device was at the Salt Lake City, Utah Airport in a research project with Kyle Rollins of Brigham Young University USA and a 14MN device was used. Mike Muchard and Don Robertson of Applied Foundation Testing perfected lateral load testing. They developed a "jig" for holding the piston and a "sled" for holding the silencer and masses.
Figure 11. Lateral load testing preparations for a 7.5MN test. Major pioneering developments have been performed by Dan Brown (1998) of Auburn University USA in the analysis of lateral Statnamic tests. 8 WATER REACTION MASS TESTING A most significant development is the use of water as a reaction mass when testing piles over-water or near-water. By being able to mobilize the inertia of the ocean or a river, very large tests may be performed with testing equipment weighing only 1 % of
A very significant job was for the Mississippi DOT in 1998 to simulate a ship impact of 7.5MN by the company AFT. Loads of up to 10 MN have now been applied in lateral load testing.
Figure 12, Set up of Statnamic water reaction mass testing
Figure 10. Over water Statnamic testing on a batter pile 558
Figure 15. Artist’s impression of a water reaction pile-driving tool. Statnamic testing using water reaction mass was first done by Berminghammer in 1998 in Hamilton Harbour. These tests went to 600kN. The first fieldtesting was performed for the Port of Lake Charles in Louisiana in 1999 by Applied Foundation Testing assisted by Benninghammer. The loads were up to 5 MN.
Figure 13. Water reaction mass containers the test load. The necessary 5% orlO% reaction mass would be provided by water confined within a vessel and submerged below the surface of the water. This weightless reaction mass makes it possible to perform very large tests of longer duration than are practical today. The use of water reaction will also make it possible to drive piles underwater with a tool, which will be virtually weightless.
9 PILE DRIVING WITH WATER REACTION Geert Jonker of IHC Foundation Equipment envisioned the idea to extend the use the water reaction mass into a pile-driving tool. Berminghammer, IHC and TNO are now working together to build an underwater Statnamic hammer, which will consist of a large inertial mass, made of water and a Statnamic tool capable producing multiple loading pulses. This tool will be used to push an anchor pile into the seabed and measure its capacity at the same time. In the coming years, we will see driving small onshore and offshore piles according to the Statnamic principle. 10 EVENTS To share the experience among pile engineers and to improve the Statnamic test technology the first International Statnamic Seminar was held in Vancouver in 1995 with 25 papers and 54 participants. The Japanese research group on Rapid Pile Load Test Methods organized the Second International Statnamic Seminar in 1998 with 48 papers and 132 participants. The third Statnamic seminar is planned in the Netherlands in 2002. In March of 2000, the Japanese Geotechnical Society (1998) published a standard for “Method for Rapid Load Test of Single Piles”
Figure 14. Water reaction mass testing
559
An ASTM standard for Statnamic type pile testing is in progress.
1 1 RESEARCH AND DEVELOPMENTS Companies and universities have accumulated up to date more than 700 case histories throughout the world. The total number of contract Statnamic tests worldwide has exceeded 1000, with testing now occurring at a frequency of more than one each day. The largest volume of testing to date has occurred in Malaysia, with the number of contract tests exceeding 300. Similar numbers of contract tests have now been performed in the United States. The UK and Japan are close behind in their numbers of tests. An important contribution to the development of Statnamic was supplied by the Japanese Geotechnical Society, which established the research group of Rapid Pile Load Test Methods in 1993. Professor Osamu Kusakabe of Tokyo Institute of Technology chaired the group. A strong promoter and initiator of Statnamic in Japan was Makoto Tsuzuki of Fugro Japan. The research group included 30 private institutes and companies as members. The activities of the Research Group aimed at cataloging the existing knowledge about rapid load tests, examining the basic characteristics and the applicability of the test, and producing scientific interpretations of the Statnamic test results. The University of South Florida (1998) conducted over 150 Statnamic tests in conjunction with privately and federally funded test programs. The tests programs have included: (1) axial load tests on piles and shafts in sands, clays, or rock-sockets, (2) lateral load tests on pile groups and shafts, and (3) plate load tests on sands and full-scale spread footings on sands and vibro-compacted soils (stone columns).
The velocity dependent soil behavior can be split up in soil damping phenomena and strain rate dependency. Soil damping phenomena can be derived straightforward from a Statnamic test. Strain rate dependency for fine-grained soils is still subject to study, for example by the University of Sheffield UK (A.F.L. Hyde. Et all, 1998). Well-documented data from pile load test projects is becoming available to support the insight in strain rate effects (Holyman, et al., 2000). 12 CONCLUSIONS The success of Statnamic stems for a significant part from the concepts and ideas generated by “artist-engineers” and “engineer-artists”. According to Brand1 (2000): an excellent engineer requires not only a firm theoretical knowledge but also comprehensive experience as well as engineering feeling and intuition in equal parts. The author would like to add: the ability to be creative and think in unconventional ways. The success of Statnamic can be further explained by the high degree of international cooperation and research, which has brought the technology to the forefront . The Statnamic community originated from the stress wave community and the author is sure that they will remain in close contact. Both have a common interest in the research of dynamic phenomena of soils and the development of tools for the load testing of piles. The incorporation of the ideas of “artist-engineers” and “engineer artists” will guarantee more marvels in the development of pile testing applications and other fields of engineering. 13 REFERENCE:
The application of Statnamic produces excellent results in stiff andor granular soils, although loading rate effects have to be taken into account. The influence of soil viscosity alongside buildup of pore water pressure in fine grained soils requires further development of analysis tools and experience. (E.L.Hajuk et. al. 1998). The soil viscosity shows up in two different ways: Creeping, this means continuing settlements under constant pile load Velocity dependent soil behavior Creeping phenomena cannot be determined with Statnamic, dynamic load testing and in many cases not even with static load testing. 560
Bermingham, P., Janes, M., ”An innovative approach to load testing of high capacity piles”, Proceedings of the International Conference on Piling and Deep Foundations, London, 1989 Middendorp, P. Bermingham, B Kuiper, Statnamic load testing of foundation piles. 4th International Conference on Stress Waves, The Hague, Balkema, 1992 Galluzi, P., Mechanical Marvels, Invention in the age of Leonardo, ISBN 88-09-20959-1, Instituto e Museo di Storia della Scienza, Florence, 1996. Baudson, M., Panamarenko, Paris 1996
Fellenius, B., Welcome from the Chairman, First International Statnamic Seminar, Vancouver, 1995 Middendorp, P., Daniels, B., The Influence of Stress Wave Phenomena during Statnamic Load Testing, 5th International Conference on the Application of Stress-Wave Theory To Piles Orlando, Florida, 1996 Bermingham, P.D., Statnamic the first ten years, Proceedings of the Znd International Statnamic Seminar, Tokyo, 1998 Hajduk, E.L., Paikowsky, S.G., Mullins, G., Lewis C., Ealy, C.D., Hourani. N.M., The behavior of piles in clay during Statnamic and different static load test procedures. Proceedings of the 2"d International Statnamic Seminar, Tokyo, 1998 Mullins, G., Garbin, E.J., Jr., Statnamic testing: University of South Florida Research, Proceedings of the 2"d International Statnamic Seminar, Tokyo, 1998 Brandl, H., Civil and Geotechnical engineering in society - Ethical aspects and future prospects, Proceedings of the First International Conference on Geotechnical Engineering Education and Training, Sinaia, Romania, 2000 Brown, D.A., Statnamic IateraI load response of two deep foundations, Proceedings of the 2"d International Statnamic Seminar, Tokyo, 1998 Hyde A.F.L., Anderson W.E., Robinson S.A., Rate Effects in clay soil and their relevance to Statnamic pile testing, Proceedings of the Znd International Statnamic Seminar, Tokyo, 1998 Justason, M. D.; Janes, M. C.; Middendorp, P.; Mullins, A. G. Statnamic load testing using water as reaction mass, The 6th International Conference on the application of stress wave theory to piles, Sao Paulo, Brazil 2000. Holeyman, A., Maertens, J., Huybrechts, N., Legrand, C., Preparation of an international pile dynamic prediction event. The 6th International Conference on the application of stress wave theory to piles, Sao Paulo, Brazil 2000
561
14 APPENDIX The Unloading Point Method (UPM) Step I) Determination of static resistance from unloading point Assumption: The long duration Statnamic load F,,, allows modelling of the pile as a zoncentrated mass (M) and springs
F,,,, = Statnamic force (measured 'and known) U
=displacement
(measured and known)
v = du\dt = velocity
(known)
a = d'u/dt'=
(known)
F,,,,i = F,
acceleration
+ F,
F,, = static resistance (unknown) F, = C.v = damping force
(unknown)
C = damping factor
(unknown)
F, = M.a
(known)
Eq U i I i bri uiii : F,,,, = F,,,,I+ FA F%!" = F,
+ F. + Fd
F,, = F,,, - C v - M.a At maximum displacement (Unloading Point) v = 0 + U = inax~rnurn,t =tun,,,
F,,,I = F,,,(t,,,,)
. aiini = a(tunIAx)
Static resistance F. is known at uulli
F,(t,,,,,) , uUlilis a static point
diagram
Step 11) Construction of static load-displacement Assumption: The soil is yielding over range Fw,,,,) to FUni So F, = FU,,i
Over this range the following equation is valid F, = F,,, - F,,,i - FJ
with F, = C.v C = (F.,,>- F,,,I- F,) / v Calculate mean damping factor C,,,,. for above range Now static resistance Fu can be calculared at all points
E, = F.m - Cmw, .V- Fd D r w static load-displacement diagram with F,, and
562
U
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
ynote lecture: Three-dimensional finite element analysis of statnarnic load test Tirawat Boonyatee & Makoto Kimura Kyoto University,Japan
Feng Zhang GifLl Universiq,Japan
ABSTRACT In order to stud! the inechaiiisin of pile-soil interaction during statnamic loading. a FEM program is developed Since it n i l 1 also be applied to the statnamic test of piles in lateral direction. batter piles. group pile. and pile raft foundation, the anal! sis S J stem is formulated i n three-dimensional space Elasto-plastic model is used for the ground material Joint element is also used for slip at the pile-soil interface The analysis program uses the direct integration scheine to solve the elasto-plastic anal!sis in time domain Laboratory tests of single piles are analjzed to iiivestigate the validit! of the developed program Firstly. the simulatioiis of static load test are done and compared \i7ith the nieasurcd data from the tests Then. the simulatioiis of piles uiider statnamic are conducted uiider the same grouiid coiidition Finall!. thc comparisons betu cen the static and statnamic responses of piles are discussed 1 INTRODUCTION
At present, the primary approach to investigating pile behavior under statnamic loading is to make a comparison between the results of full-scale statnamic and static load tests. Unfortunately, this calibration method has an inherent problem in its testing consistency, as was addressed by Amir & Amir (1995). Under the uncertainty and complexity of the target ground, the situation whereby any model can be validated by calibrating statnamic load tests against corresponding static load tests can be demonstrated as follows: 1 If a comparison is made of the results of tests on the same pile, the tests have to be done in sequence. Therefore, the quality of the later tests invariably deteriorates due to the induced residual stresses from the former tests. 2 If a comparison is made between the results of tests on different piles, significant samples should be tested in order to achieve some degree of reliability over the ground uncertainty.
lnstead of studying the data from the field tests, the mechanism of the piles during statnamic loading can be thoroughly investigated by the numerical analysis approaches. Applying the fact that the system of interest is axis-symmetry, Matsumoto (1998) applied a 2D-FEM program to analyze the behavior of a single pile under vertical loading. Although a twodimensional analysis is simpler and the solution can be found faster than with a three dimensional analy-
sis, it cannot be applied to a complex system such as a pile-raft foundation or batter piles Moreover, it is obvious for a horizontal loading case that the analysis should be done in three-dimensional space In order to develop a unified tool that can be applied for various types of problems, a threedimensional finite element analysis program called DYSC (Dynamic and Static systems analysis Code) is originally developed No-tension criteria and a simple elasto-plastic model based on the DruckerPrager theory are applied as the yield functions for the model ground An interface layer element is inserted at the interface between the pile and the soil For the pile body, a linear elastic relation is used since the applied force is lower than its yielding point It is thought that before applying the DYSC to the simulation of piles, its fidelity should be checked by a physical model For calibration purposes, a comparison of the analytical results to the laboratory test results is preferred to the results from the field tests The reasons are due to the complexity of the ground and the qualities of the geotechnical parameters required for the analysis In our previous study, a small-scale statnamic loading device (3SLD-Mkl) was used to conduct laboratory tests of piles under vertical loading (Kimura et a1 1998) This device is used to conduct experiments under a controllable environment Consequently, guaranteeing that the series of tests are done under the same conditions is possible Accompanied with the data from those experiments, a comparison 563
between the simulation and the experiment is made and reported in this paper. 2 DETAILS OF ANALYSIS
2.1 Inyzrt yaranieters To determine the necessary parameters used in the simulations of the statnamic load test, trial calculations are made. As shown in Figure 1, an analysis is conducted in the half area of the pile-soil system. Young's modulus of the pile is back calculated from the flexural test results. Poisson's ratios for the pile and the soil material are assumed to be 0.2 and 0.333, respectively. A parametric study has been made to determine the Young's modulus of the model ground. From various trial calculations and comparisons with data from static load tests, a Young's modulus of 4.9 MPa and a frictional angle of 32 are selected as rational quantities for the ground materials. The frictional angle of 32 for soil gives the calculated ultimate load at more or less the same level as the measured data. A Young's modulus of 4.9 MPa controls the shape of the loaddisplacement of the pile before failure. Using these values, a satisfactory approximation for pile response can be obtained. The load-displacement plot from calculations and measured data is shown in Figure 2.
0
1
2
3
3
5
6
D i sp1acein ent (inin) Figure 2. Load-displacement relations from simulation and experiment.
Table 1. Prouerties of materials Sand
Pile
Density (kg/m3)
1467
2150
Young's modulus (MPa)
4.90
5.00s103
Poisson's ratio
0.333
0.200
Friction angle ( )
32
Dilatmcy angle ()
9
The properties of each material determined from an inverse analysis are summarized in Table 1. 2.2 Pile-soil inteiface model
The Mohr-Coulomb yield criterion is used for the interaction between the pile and the soil. It states that failure will take place if the magnitude of shear stress (z) on the failure plane is equal to the value given by the following relationship:
in which 1 1 denotes the absolute value, 0, is the normal stress on the failure plane, and $0 and CO are material constants for the pile-soil interface. In this study, adhesion, or CO, is assumed to be zero. The frictional angle of the pile-soil interface is assumed to be 0.9 times the frictional angle of sand. Equation (1) can be written in the form of yield fbnction F as
F = jzlFigure 1. Analysis mesh.
564
CT,
tan$o
If the material is sheared to the yield surface and the associated flow rule is adopted, the rates of plastic normal strain d&; and shear strain 7' are given by
which implies 0.0
(4) Increments in shear displacement along the plane are accompanied by increments in normal displacement. The dilation of the shear plane will go unbound under yielding. To avoid this unfavorable behavior, the non-associated flow rule is adopted for the pile-soil interface. By introducing dilatancy angle y, the plastic potential function can be written as
In this study, no dilation is assumed for the pile-soil interface, i.e., w-+ 0.
3 SIMUL,ATION ON STATNAMIC LOAD TEST
After necessary parameters were back calculated from the previous section, a dynamic FEM analysis of a pile under Statnamic loading is done and compared with the experiment results. The loading rate dependency of the ground response is represented by a constant damping parameter. In this simulation, velocity and acceleration are calculated from the displacement by the Newmark method. The stiffness matrix (IQ the damping matrix (C), and the mass matrix (A4) are calculated from the following equations:
K = JB'DBdd 1.01
(7)
565
0.2
0.4
0.6
0.8
1.0
1.2
Di sp1acein eiit (m in ) Figure 3. Load-displacement relations from statiiaiiuc simulation and experiment.
where N is a so-called shape function or displacement interpolation function, R denotes a displacement to the strain transformation fbnction, and 11 is a strain to stress-strain transformation function. Parameter ,U is used to represent the damping constant per volume of interested material in the same manner as the density (p) is applied in the mass matrix. Based on trial calculations, a damping constant of 3.9 MN/(m/s)/m3 (,U as defined in Equation (7)) is used as the material constant for statnamic simulations. The estimated load-displacement relation and the test data are shown in Figure 3 . Note that the initial displacement in statnamic load tests does not conform to that in the static load tests. The initial displacement in the statnamic load tests is a little bit smaller than the corresponding value in the Static load tests. This may contribute to the loading rate effect during the equipment installation process. To correct this inconsistency, the load-displacement relation (from the experiment) of the statnamic load test is shifted to the right in order to match the 'true' displacement in the static load tests. 4 COMPARISON BETWEEN COMPUTED STATIC AND STATNAMIC LOADING RESULTS
The distributions of displacement along a pile under static loading are shown by Figure 4. At the peak load, the elastic contraction represented by the difference in settlement between the pile head and the toe is about 0.05 mm. When compared to the overall settlement, this contraction is about 5% of the pile head settlement. The distribution of axial force along the pile is shown in Figure 5 . The application of load is sustained almost totally by friction resistance, with only a small amount of force being transmitted to the pile tip. The ratio of the end bearing resistance to the shaft resistance is about 1 :4.5.
St3t.k
150 N +250 N +400N 450 N
+
s tatnarfl k: - 0 - 150 N - 0 - 250 1\1 - 0 -400N
- H - 450 N
~"'"1'""1"""'"I""'I""'I''"'I'""I'
()
0
02
0 -4
00
08
10
D 1s p lacem en t ( m m) Figure 4 Distnbdion of displaccnient along pile under static loading
0
50 100 150 200 250 300 3% -100
Load (N) Figurc 5. Distribution of axial forcc along pilc undcr slalic loading.
The pile velocitv versus time relation is shown in Figure 6. At the maximum displacement, the velocity of the pile head is equal to zero. At the same time, the velocity of the pile tip is almost identical to that of the pile head. This implies that the pile behaves as a rigid body at the unloading point. Axial force distributions of the pile during statnamic loading as well as a comparison with those of the static load test are shown in Figure 7. At the same applied load level, the axial force distributions of the two cases are almost identical. There are no substantial changes in load sustained by the pile tip. Increases in capacity mainly contributed to the shaft resistance. The effect of a stress wave is not observed in this analysis. 5 CONCLUSIOK
Figurc 6 Vanation
III
When compared to field tests that have the inherent problem of ground uncertainty, the model tests presented here show the possibility for conducting Statnamic load tests under one uniform condition. Only by this approach can the data be interpreted on absolutely the same basis. Based on the results presented here, the following conclusions can be summarized: 1 The finite element analysis (DYSC) shows that the elastic contraction of the pile is relatively small. This value represents about 5% of the total displacement. From the load distribution plot, the applied load is mainly supported by shaft friction. The model pile is thought to be a friction pile.
vclocitj of pilc licad
566
The ratio of shaft resistance to end bearing resistance is about 1:4.5. 2 Velocities of the pile head and the pile tip are almost identical and equal zero at the unloading point. This supports the assumption that a pile moves as a rigid body at the unloading point. 3 The axial force distribution from the statnamic load test is almost similar to that from the static load test. The stress wave effect, as in dynamic load tests, is not observed in the present. This implies that in the statnamic load test, the pile is loaded in the same manner as in the static load test. At present, the horizontal load tests on group piles have been conducted and reported elsewhere (Kimura et al., 1999). The improvement on DYSC in order to simulate the piles under lateral loading is under operating. REFERENCES Amir. J M and Aniir. E I 1995 A Lumped-Parameter Model for Statnanvc Testing ProL of the 1" Iiit 'I ,Stntiianiic & i ~ i i iiar, I i7iicoiiver pp 221-230 Matsuiiioto. T 1998 FEM anal!sis of Statnainic test on openelided steel pipe pile pro^ oftlze T dIiit 'I StatiiamiL SeiiiiI?W, Ibkvo, pre-printed volunie Balkerna Kiniura. M . Boonjatee, T & Yoshida. A 1998 Expenmental stud! of Statnainic load test b! air-pressure based loading ' ~ Seiviiiar, 70Iil0, apparatus ~ r o cof the 2 1 ' ~~ i i t StntmniiiL pre-priiited volutiie Balkenia Kimura. M . Boonlatee, T & Yoshida. A 1999 Evperinicntal Stud! of Lateral Statnaiiuc Load tests on Group Piles I ' i m oftlie J'" Iiit 'I Colif oii Deep Foiiiidatroii I+actice 111 orprating PILEX-ILK '99,Singapore pp 263-27 1
567
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 1503
Lateral statnamic load testing of model piles Makoto Kimura & Tirawat Boonyatee Kyoto Univei-sify,Japan
ABSTRACT: In order to generate the statnamic loading for laboratory tests, a small-scale loading device is developed. It is driven by the compressed air instead of gas from the explosion. By applying our developed device, statnamic lateral load tests of a single pile in sand are performed under free head condition. The diameter and length of model pile are 24 mm and 1225 mm respectively. The test results show that the unloading point method can also gives a good estimation for the lateral load tests of single piles. haved as the short, or rigid, piles. Therefore, an attempt to make the tests on the flexible piles has been made. In this paper, the results of the tests on the single, flexible piles are reported. The procedure of the tests is the same as the previous study. Firstly, the static load tests of single pile are done in order to establish the benchmarks for the pile resistance. Then, the model ground is completely reproduced and the statnamic load tests are suubsequently carried out. The unloading point method, originally used for vertical statnamic loading, is also used to estimate the static response of piles in this study. Although there are some assumptions that take a certain amount of soil, which sticks to the shaR, into consideration, only the mass of the piles is used in this computation.
1 INRODUCTION
Up to present, the potential of the statnamic load test to estimate the bearing capacity of pile foundation is well recognized. However, since there is little understanding of pile-soil interaction during lateral statnamic loading, the rational interpretations for the lateral statnamic load tests have not yet been established. Generally, it has not been unusual to conduct studies based on the data from the field tests, i.e. by comparing the results of statnamic load tests with those of the static load tests on the same piles. One example is the work done by Rollins et al. (1998). With these data, the consistent interpretation is difficult to achieve due to the uncertainty and the complexity of ground. Since consistent systems for interpretation and comparison can be provided under laboratory test conditions, it is thought that the model tests are an indispensable tool for the study of pile-soil interaction during Statnamic loading. In our previous study (Kimura et al. 1998), a small-scale statnamic loading device (3SLD-Mkl) that can perform the statnamic load test under one constant condition has been developed. By applying pre-compressed air instead of gas from an explosion, a loading similar to the statnamic loading can be generated. This loading device was used to perform vertical statnamic load tests. For lateral load tests, a loading equipment named 3 SLD-Mk2 (Small-scale Statnamic Loading Device-Mark 2) is used to apply statnamic loading to the model piles in a lateral direction. Using 3SLD-Mk2, the tests of single piles and group piles have been successklly conducted in the previous work (Kimura et al. 1999). However, it was found that the model piles used in that study be-
2 LOADING APPARATUS 2.1 Static load test In the static loading tests, a mechanical jack is used to load the target piles. By winding the handle of the jack, the piles can be loaded by the displacementcontrolled method. Noted that a steel plate is used to transmit the applied force from the jack to the piles since their elevations are not the same. The applied loads to the piles are measured by a load cell that is attached to the steel plate. The point of loading is 1 cm above the ground surface. For the displacement of piles, the displacement at pile head is measured by a laser displacement gauge. This gauge is installed on an additional stand to isolate any effects from the target system. For the static load tests of a single pile, strain gauges attached at 569
eleven elevations are used to measure the bending moments along the shaft 2.2 Stattmmic load test
A general view of the apparatus is shown by Figure 1. Statnamic loads are generated by the quick delivering of compressed air into the cylinder. The main part of the device, which is the cylinder, can be moved smoothly by attaching it to the slide rails. This slide rail also restrains the moving direction of the device to the horizontal axis. The flow of air is controlled by a magnetic valve that is shut in the normal state and open when an electric current is received. By applying an electric timer, an adjustable pulse of electric current can be given to the magnetic valve. The mechanism of 3SLD-Mk2 can be shown schematically as in Figure 2. The operation of loading by 3SLD-Mk2 can be divided into the following five stages: Initial stage (Figure 2a.). Air loading stage (Figure 2b.). In this stage, precompressed air is allowed to flow into the cylinder. The air piston will start to load the pile head. Acting stage (Figure 2c.). While the piston is pressing on the pile head, the air in the near-pile side of the cylinder will flow out through the silencer. At the same time, the main body of the device will move away from the target pile due to the reaction force from the pile head. Termination of air supply (Figure 2d.). As mentioned before, the magnetic valve will be triggered by the electric timer and the path of air will be shut off. At this stage, the main body of the device will still be moving for a while because of inertia force Termination of test (Figure 2e.). The motion of the main part is gradually terminated due to the friction forces of slide rails.
Figure 1. The statnamic loading apparatus.
Figwre 2. Loading mecllanism of 3SLD-Mk2.
570
Table 1. Propertics of iiiodel piles Thickness (nun)
1.o
LCil~$Il(cm)
122 5
Weight (8)
591 0
Diameter (nun)
24 0
Young's inodulus (MPa)
8 58\10'
Bending stiffness (Niii')
4 09x10'
The steel plates are also used to transmit the force from the piston to the pile heads. The piles are loaded 1 cm above the ground surface. To ensure that the transmission plates contact with the pile heads and no impact force is induced, a small initial load is applied to the pile's head on the beginning. The measured parameters are the applied loads and the displacement of piles. For statnamic cases, these data are recorded every 0.2 ms. 3 TEST CONDITIONS AND TEST PATTERNS
Hollow brass pipes are used as the model piles. The diameter and the thickness of the piles are 24 mm and I mm, respectively. The bending stiffness is determined from bending tests. The properties of the model piles are shown in Table 1. In this study, No.6 silica sand is used for the ground material. The model ground is prepared by vibro-compaction method. At first, the model pile is placed in the center of the soil chamber. After the model pile is set up, the sand is filled into the soil chamber, and then, it is compacted evenly throughout the twenty-four locations by injecting a vibration rod into the ground. Due to the length of vibrator rod is limited, the ground preparation is divided into two steps. The first step is for the lower and the second step is for the upper layer. The thickness of the lower and upper layer are 60 cm and 57.5 cm, respectively. The relative density of both upper and lower layers are about 60%. The overall density of the model ground is 1467 kg/m'. The properties of the sand and prepared model ground are shown in Table 2. In this paper, two types of tests are reported, as shown in Table 3. For each pattern, the tests are
Table 2. Properties of sand and iiiodel ground. Specific weight
2.6.3
Maximum void ratio
1.03
Miiiiniuni void ratio
0.64
Moisture ratio (%)
0.3
D60 (W)
310
Diu (w)
120
Density (kg/m3)
1167
Relative density (%)
59.6
Frictional angle ( )
36
Table 3. Test patterns.
Pattern
Pile type
Loading type
Length (cm)
TS1
Single
Static
122.5
TS2
Single
Static
122.5
TD1
Single
Statnainic
122.5
TD2
Single
Statnamic
122.5
0
1
2
3
4
5
Displacement (mm) Figurc 3. Load-Displacemcnt relations from static load tests.
Flguc 1 Moment dlstnbutlon along thc pilcs.
571
2 0.3 1 '
conducted twice in order to ensure that there is no disparity in the records. Addition tests will be done if noticeable differences are observed.
' ' ' '
I
'
'
, ' '
I
' ' ' ' '
' ' ' ' ','
I
' ' ' '
'
'
'
'
~
'
i
1
AND
4 EXPERIMENT RESULTS INTERPRETATIONS 4 I Static load tests
By winding the handle, the jack will elongate and thrust the piles, therefore, the piles are loaded by the constraint displacement method A loading speed of 0.6 mm/min is applied The piles are loaded up to 135 N, which is approximately the same level as the maximum force measured from the statnamic load test The relation between the load and the displacement are shown in Figure 3. The moment distributions calculated from the strain along the shaft are shown in Figure 4. Since the induced bending moments reduce to negligible proportions below 72 cm from the ground surface, it can be thought that the model piles behave as the flexible piles 4.2 Sfaiiinniicload tests
Figrlrc 6 Vclocit! and accclcrai~o~~ 1 s iinic
43
C bi?iparisonhetit'eeii stntimmic mu' static load test 1.e.slilt 5
When compared to the data from the static load test, the load-displacement plot from the statnamic load test shows a stiffer relation As shown in Figure 7, the load-displacement curves from the static load test and the statnamic load test intersect near the unloading point under an applied load of 125 N
-4pressure of 0.3 MPa is used to generate the Stat-
4 4 Estii~iatioiiby the ~niloadirig poiiit method
namic load. Note that this value has not been computed on a theoretical basis, but has come from trial tests. The load and the displacement of the pile head are shown in Figure 5 . From the tests, a maximum Statnamic load of 140 N and a maximum displacement of about 4.0 mm are observed. As was also observed in the vertical Statnamic load tests, the maximurn displacement occurs a short time after the peak load due to the presence of damping forces. The duration of loading is about 70 ms. Figure 6 shows the history of the velocity and the acceleration of the piles during statnamic loading. The maximum velocity and acceleration are 0.24 m/s and 25 m/s2, respectively
The unloading point method, originally used for vertical statnamic loading, is also used to estimate the static response of piles in this study. The outline for the calculations can be found in the works of Rollins et a1 (1998) In this method, the inertia can be computed by multiplying the lumped mass of system with the acceleration. Although, it is rational to include a certain amount of soil that moves together with the shaft, only the mass of the piles is used for inertia force Computation The predicted static load-displacement relations compared with the results from the static load tests are shown by Figure 8. Although these plots are 150 t
150
'
'
'
'
' ' '
'
'
'
'
'
'
'
'
'
'
'
'
'
'
' '
'
'
'
'
'
'
'
h
100
-20 5 0 -1 h
0
,
.e
a
0
20
.
40
.
.
.
.
I
.
.
.
.
.
I
. . . . .I
60 80 Time (nis)
Fibwre 5. Load and displacement vs time.
.
100
.
.
.
0
.
120
1
2
3
4
Di splaceinent (inin) Figtrrc 7. Comparison of load-displaccmcnt rclation.
572
j
500
125 100
400
5: W
75
300
20
50
n
200
1-3 25
1 00
0 t
.
.
.
0
.
.
l
1
.
,
.
.
.
2
l
.
.
.
,
.
3
l
,
,
,
,
.
l
1
,
,
.
.
.
i
0
5
0
1
Displacement (inin)
7
I
3
3
5
Di splacein en t (mm)
Figure 8. Estimated load-displacement relations
Figurc 9 Dainping cocfficicnt 1 s displaccmcnt
slightly different, the estimated curves show a good estimation of static response of the model piles. Based on the presented results and the previous study on the relatively short piles (Kimura et al. 1999), it can be concluded that the unloading point method also gives a good estimation for the statnamic load tests in lateral direction.
5 CONCLUSIONS
4 5 Daniprig coeficierif of the imlondirig point
niethod
Accompanying the data from the static load tests, damping coefficient C, used in the unloading point method, can be back calculated Firstly, damping force I*;{ is calculated by subtracting the statnamic loading with static resistance and inertia force (F',it and Mn, respectively) as shown in the following relationship
Note that all the three items on the right hand side of above equation can be obtained directly from the experiments. To obtain the damping coefficient, the damping force is, then, divided by pile velocity L: as in the following relation:
In this study, statnamic load tests of flexible piles have been conducted The data collected from the statnamic load tests are interpreted by the unloading pint method and compared with the measured static responses Conclusions from the study can be summarized as follows 1 Load-displacement plots from the SLT and the STN meet at about the unloading point This evidence agrees well with the assumption of the unloading point method (Horvath et a1 , 1993), which states that dynamic forces at an unloading point are very small and can be disregarded 2 By the calculations that only the mass of pile is taken into consideration, the unloading point method can give a good estimation for the tests of flexible piles 3 Based on back-calculated damping parameter (C), it can be concluded that a good estimation of piles can be obtained even C is assumed constant When compared with the prediction on the piles under vertical loading, the unloading point method also shows a good prediction for piles under lateral loading However, it is obvious for a horizontal loading case that the analysis should be done in three-dimensional space Therefore, an attempt to make a versatile analytical tool that can cope with several loading conditions in three-dimensional space has been made (Boonyatee & Kimura 2000) REFERENCES
The variations in calculated C with pile displacements are shown in Figure 9 Since the velocity of piles is close to zero, C changes rapidly at the beginning and around the unloading point. However, it changes gradually and undulates within a small range during the intermediate stage From this evidence, it can be concluded that for the interpretation of piles under lateral stat namic loading, the parameter C can be assumed constant
BoonJatee. T & Kimura. M 2000 Three-dimensional finite element analysis of statnamic load tests Proc dh Iiit 'I Coiif oil tlie .Ipplication of Strew Iiave 'Ilieon to Piles, Sao Paul(,, Brazil Balkenia Horkath, R G . Beniiinghaiii. P & Middciidorp P 1993 The Equilibrium Point Method of Anal! sis of the Statnainic Loadiiig Test \\ ith Supporting Case Histories Proc of the IRfh 1iinunl Ueetirig of the Deep Founciatioii Iiistituie. l'itttburg pp 61-80
573
Kimura. M . Boonlatee. T & Yoshida. A 1998 Expenmental study of statnanic load test by air-pressure based loading apparatus Proc Znd hit 'I Statnamic Seiniiiar, Tokvo, Japaii Balkema Kimura. M . Boonyatee, T & Yoshida, A 1999 Expenmental Study of Lateral Statnartuc Load tests on Group Piles Proc of the Jih Iiit'l Coif on Deep Foundatioi? Practice 111 Corporatiiig PILETiILE: '99, Siiigapore pp 263-27 1 Rollms. K W . Weaver. T J & Peterson, K T 1998 Lateral statnanuc load testing of a pile group Proc 2'ld hit 'I Statiiatiiic Seminar, Tokvo, Japan Balkem
APPENDIX To make a rough comparison between the lateral loading case with the vertical loading case, the results from the model tests of single piles under vertical loading is provided here. The load-displacements from the static and statnamic load tests of a model pile are shown in Figure A l . 1
0 ~ 00
~
02
~
"
~
~
-- _-----'
~
'
'
~
04
"
~
~
~
06
"
"
'
~
'
'
~
'
~
'
~
~
I 2
10
08
~
Displacement (mm ) Flgurc A 1 Load-displaccincnt rclallons of a inodcl pilc undcr 1 crlical loading
Then, the result from statnamic load test is used to estimate the static load-displacement relation by the unloading point method. Note that this computation do not include the mass of soil that may be stick to the shaft during loading. The estimated curve and the relation from the static load test are shown in Figure A2. I
'
"
"
I
"
'
'
~
"
'
'
~
~
"
'
"
'
1
400
300
x
0
200 100
00
02
04
06
08
I 0
12
Displacement (inin) F i p r e A2 Estimated 1o;iddisplacemciit relations
574
"
~
~
'
~
~
~
~
~
~
"
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Application of the Stress Wave method to automatic signal matching and to statnamic predictions G. Esposito, W. M.G.Courage & R. J.van Foeken TNO Building und Construction Research, Department of Structurul Dynamics, Rijswijk, Netherlunds
ABSTRACT: The Statnamic method is an increasingly popular technique to cany out loading tests on cast insitu piles. The method has proved to be a cost-effective alternative to a static loading test. As-sociated to Unloading Point Method (UPM) and to automatic signal matching, the Statnamic testing technique provides an estimation of the static load-settlement curve of foundation piles. For long piles with wave numbers (Nw) less than 12, however, the method becomes less accu-rate because of stress wave phenomena (concrete piles longer than 35 m or steel piles longer than 40 m). In these cases, the signal matching stress wave method can be used to estimate the static load-settlement curve. The TNO Stress-Wave method is a convenient way to carry out automatic matching on Statnamic signals. The use of TNO Stress-Wave method and of auto-matic signal matching for a Statnamic prediction is investigated in this paper. A case study is introduced and the results of a signal matching are presented. Further, the case study is mod-elled in TNO Stress-Wave and the results of the model are compared to the results of the field Statnamic test. the differential equation for longitudinal stress waves in a pile. The first stress wave program was developed by Smith in 1960 and solved the differential equation by means of the fmite difference method (Smith 1960). The TNOWAVE stress wave package was developed in the early 80s and is based on the method of characteristics (Voitus 1974), where the pile is defined and modelled by elements with a characteristic length. At the interfaces of the elements, a soil model is defined. Between each interface, the pile remains frictionless and a wave travels undisturbed as if in a fiee pile. The soil response is a fimction of the relative dis-placement between pile and soil, the pile velocity, and the pile acceleration. At each interface, part of the wave is reflected and part is transmitted, depending on the properties of the elements and soil resistance. For signal matching, the measured load on the pile head is used as a boundary condition at the pile top. Statnamic predictions can also be made with TNOWAVE (Bielefeld 1995), although the he1 combustion must be modelled. The objective of this study is to demonstrate the application of automatic signal matching for a Statnamic prediction. The case study herein presented refers to a bored pile having wave number lesser than 12 where the UPM proves to be not adequate. The same Statnamic test is fbrther modelled by means of the TNO Stress-Wave Method.
1 INTRODUCTION Where a static load test is expensive and time consuming, a Statnamic load test can be much cheaper and faster. Together with a dynamic load test, Statnamic is a cost-effective alternative to a static load test. Dynamic load tests perform well on pre-cast concrete piles and steel piles. On cast-in-situ piles (auger, bored, etc.), the Statnamic method is preferred, because the pile shape is not known and can vary. To match soil parameters, the cross sectional changes are also considered as soil variations (Middendorp 1998). The Unloading Point Method (UPM) is an accepted method of analysis for Statnamic tests to determine static bearing capacity of foundation piles having wave number greater than 12. The wave number, N,, is defined as the ratio of the product of the duration of the blow and the wave velocity respect to the length of the pile. To interpret Statnamic load tests on long piles with wave numbers less than 12, the unloading point method becomes less accurate because of stress wave phenomena (Middendorp 1995). The pile does not move as a rigid mass with a uniform velocity. The movement of the pile head and pile toe is not synchronized and, thus, the pile velocity and acceleration can not be approximated by the observed pile head displacement. There are several programs which can describe the stress wave phenomena in a pile and are all based on 575
2 THEORETICAL BACKGROUND 2.1 Unloading Point Method (UPM) The long duration of loading (N, > 12) allows the pile to be modelled as a rigid mass on which the following forces act: Applied load or Statnamic force, FSatn; Inertia force of the pile mass, Fa(t) = Mplle a(t), where Mplle is the total mass of the pile and a is the actual acceleration; 8 Soil resistance along the shaft of the pile and at the toe of the pile, F,, which is composed of a static resistance, water pore pressure, and damping force from the soil; Static resistance force, FU-SlatlC; Water pore pressure, Fp; Damping force from soil, Fv(t) = C - v(t), where C is the mean damping value and v is the actual velocity. The displacement of the pile top equals the displacement of the pile relative to the soil. For simplicity, the pore pressure resistance is included in the damping. Figure 1 shows the Statnamic loaddisplacement with dynamic components.
Displacement
U
5. The unloading force of the static soil resistance overcomes the opposing forces and pushes the pile upwards. At the end, final settlement can be observed. The static load behaviour can be derived as follows:
or
At each point, the inertia force and the damping force can be calculated, resulting in the following static resistance:
All terms of equation (2) can be determined from the Statnamic loading data, except the damping factor Cmean. In the unloading point method, at maximum displacement the velocity and therefore damping force is zero. The damping factor is calculated fi-om the average of
obtained from the time range between FStat,, maximum value and the unloading point with the condition that v(ti) is greater than 25% of the maximum velocity contributing to the average. With the damping factor, C,,,,, known, the static load curve can be calculated from the Statnamic load curve at any time. A smooth curve can be constructed from the calculated values using appropriate curve fitting.
Load displacement diagram
-
Figure 1 : Statnamic load-displacement diagram
2.2 TNOWA VE - Stress Wave Method
1. The Statnamic reaction mass is placed on the pile top. The load-displacement behaviour is fully elastic. 2. The reaction mass is launched and Statnamic loading starts. The soil resistance is elastic and inertia and damping forces act on the pile. 3. The static soil resistance reaches ultimate strength and yields. Velocity and inertia forces increase progressively. At the end of this area, the maximum Statnamic load is reached. 4. Statnamic load decreases. Due to pile inertia, displacement is increasing. The soil is yielding and pile velocity reduces to zero. The pile displacement reaches a maximum value and damping forces and velocity become zero. The Statnamic load minus the inertia force equals the static soil resistance at time tu-max.
TNO Building and Construction Research developed the stress wave package TNOWAVE to simulate stress wave propagation in piles. The first application was signal matching for dynamic load testing. The method also applies to signal matching for Statnamic load testing. Other applications include sonic integrity testing simulation and signal matching (Middendorp 1988), pile driving prediction for impact hammers and vibratory hammers (Bielefeld 1994 and 1992) and Statnamic simulation (Bielefeld 1995). The stress wave method for wave numbers less than 12 is much more reliable than those with the unloading point method (Middendorp 1995). In the stress wave method, the soil layers and classification and the actual pile shape are defined. Since signal matching with soil parameters is time consuming and requires an experienced engineer, TNO developed an
576
'automatic' matching routine to help the engineer in the matching process Early automatic matching routines in 1984 calculated pile shapes from measured signals (Ypey 1984) and determined the pile shape of several model piles measured in research conditions without soil As soil influences were not considered, however, the accuracy of practical results was v e q low I n 1 989, more advanced research programs began (Staalduinen 1989) I n 1992, a routine for automatically calculating measured signals was developed (Courage 1994)
The routine used in automatic signal matching of soil or pile parameters has a dif-f-'erentbasis now instead of directlg. calculating the pile shapes and soil models, TNOWAVE matches signals as an engineer would, with a "search strategy" to find the best fit between calculated and measured signals Superbision by an engineer, with experience in signal matching, geotechnical Anowledge, and practical experience in foundation engineering, is still 1-equired Automatic signal matching is not replacing the engineer, but offers support to perform matches more easily and faster
The p remi ses for deve1opi ng autoin at i c signa 1 matching in TNOWAVE were A mathematical model \n/hich accurately describes the behaviour of foundation piles in soil layers, with an appropriate paraineterization of the pile properties and pile-soil interaction In TNOWAVE, the mathematical model is based on the method of characteristics, developed by Voitus van Haniine ( 1 974, 1977), Middendorp (1 966), and others * An accurate and efficient computational algorithm to solve the boundary value problem described by the tnathematical model Such an algorithm is available in TNOWAVE and is referred to as the "farward" model. 0 Reliable measurement data of excitation and response of the pile on the site Using digital processing techniques, the Foundation Pile Diagnostic System (FPDS) is capable of obtaining high quality signals
Usually, time signals are obtained froin measurements. revealing information of lower pile parts and soil layers up to the reflection of tlie pile toe with increasing time It is therefore useful that the method considers new data points iteratively, updating the parameter values and their reliabilities To satisfi the above properties, tlie Kalman filtering method was chosen The I
A method was sought that would determine unknown or uncertain parameter values in the model giving good agreement between the forward model and the measured data The method should incorporate n yrrori ltnowledge about probable values and uncertainties using n p~ioricovariance Apart from parameter values, the method should also give covariance using a Bayesian approach (Norton 1986)
Figurc 2 Schematic flou chart for automatic signal matching
577
For Statnamic load test signal niatching, the soil, measured impact signal. and the pile must be modelled The starting values of the soil parameters for each layer are estimated from soil investigation results, such as CPT-values or SPT-values The layer definition can not be changed. but the soil properties (yield stresses, quakes, and damping factors) are updated until a good match is made In dynamic signal matching, the match is done on the upward travelling wave For Statnamic signal matching, it is better to match on the velocity signal at the pile top The distinction between measured and calculated signals are more pronounced i n the velocity signal comparison The measured signal is split into 5 intervals for which each allowable errors can be set (see Figure 3) Also a certain updating scheme is used First the match is concentrated on the shaft soil model and than on the toe soil model Than a second update is made oii both The iteration is finished with a re-run 011 shaft, toe and the combination The best match during the inatcliing process is stored for further use This is in most cases the match at the end of the update scheme The soil model used for Statnamic signal matching is composed of asymmetric springs and linear dampers along the shaft and an asymmetric spring with gap and linear damper at the toe as shown in Figure 4
Figure i Force \crs~istime signal c m be split into riitcnals
Figure 1 Soil models at shali m d toe
Velocity along pile axis 3 100
E 0 75
3 CASE HISTORY STATNAMIC TEST IN TA1Lli A N 3 1 7 N O WA I 7<
0 :0
0 25
Sigrid Mcrkhiirg
A 16 MN test was performed in Taiwan on a 51 ni long bored pile with a diameter of 1 0 in The pile toe is situated in a sand layer of I7 ni On top of this layer is a 20 in thick clay layer The upper soil is composed of sand, silt, and clay layers The duration of the Statnamic load was about 100 nis giving a wave number of about N,,= S and, thus, the direct application of the UPM was not possible The velocity and acceleration observed at the pile head were not representative of the overall velocity and acceleration of the pile as shown in Figure 5 To obtain a static load-displacement curve, a signal match must be made The soil model along the shaft and toe must be chosen in a way such that the calculated velocity signal at the top of the pile matches the measured velocity After a good velocity match has been made. the measured displacement should also match the calculated displacement The results of the signal match are shown below for velocity, displacement, and upward travelling wave at the top of the pile in Figure 6 The soil match data is listed in Table 1 The yield factor (F\2/F,1, see figure 4) for the shaft springs is
0 00
0 25 000
1000
2003
3000
4000
5000
E000
Pile axis [m]
Figure 5 Velocitj along pile axis at three stages
Table 1 : Soil match data Layer Thickness Yield Stress Lml [MPaI
1 2 3 4 5 6 toe
2.5 5 7 20 7 9.5
0.01 0.0 1 0.027 0.0 I 0.03 1 0.098 8.677
Quake [mml
5 5 5 5 5 513 25
Damping C [MNs/m] 0.03 0.048 0.038 0.076 0.06 1 0.071 0.00 1
1 .O(last layer 2.0; toe is 0.0). The soil damper is assumed linear. 578
-MoblllsedLoad Displ Dragram, Pile Head 3
[mml 0.0
8.0
16.0
,
I
18 4
'\
i
0.6
I
.
d'
-0.6
100
00
Calculated and
- - measured
200 Time [ms]
20.0I
_ _ I
nowAMIm
I"I*)_."
[MNI
Figure 7: Static load-displacement diagram
28 20 12
I
4 1
4 100
0
-Calc. and
200
Time [ms]
--measured
Figurc S: Statnamic iiiodcl -08
0 -Calculated
100
and
_ _ measured
200 Time [ms]
3.2 TNOWAVE - Stress Wave Method A Statnamic test is a complex interaction between the Statnamic device, pile, and soil. As with a diesel hammer, the pile-soil s t f i e s s has an influence on the peak load and jump height of the reaction mass. The modelling of the load at the pile top from the burning of the propellant is important to properly simulate a Statnamic test. The loading of the Statnamic device depends on the initial chamber volume, type, and amount of fuel and reaction mass, vent distance, and vent height. As well, the shape and dimensions of the vent are important (Bermingham 1995). The Statnamic loading in TNOWAVE includes the burning of the fuel according to pyrotechnics and depends on the fuel chemistry, burning coefficients, and geometry of the fuel (Bielefeld 1995). The signal match case history is used to make a prediction. The 16MN device is modelled on top of the pile as shown in Figure 8. A fuel loading of 11.2 kg was used with a reaction mass of 0.8 MN. The
(cl Figure 6: Results of signal match for displacement(a), velocity (b), and upward travelling wave at top of pile(c).
The sum of the damping values over the whole shaft and toe give a damping value of 14 MNs/m3. From the UPM, a damping value of 9.6 MNs/m3 is obtained. Both values are related to the toe cross section. From the best soil match, the static loaddisplacement curve can be calculated, by simulating a static load test on this model using only the displacement dependent part of the soil model as shown in Figure 7. The static resistance from the signal match is 10.2 MN. The UPM gives a static resistance of 14 MN. The latter value is not correct and confirms that a signal match is required for wave numbers less than 12.
579
plied. The inertia force calculated from the total pile mass should be considered to predict the proper static resistance. For wave numbers less than 12, only part of the pile mass will contribute to the inertia force. Signal matching with TNOWAVE offers a good method to determine the static capacity of a Statnamic load tested pile when the wave number is less than 12. Statnamic stress wave predictions give a conservative result. More back analysis is required, especially to predict the launch (jump) height.
REFERENCES
Figurc 9: Results of Statnamic prediction for displacement (a) and force at top of pile (b).
following parameters have been used to model the Statnamic device: combustion chamber area 0.2082 m2; initial chamber volume 0.12426 m’; vent diameter 0.15875 m; vent length 0.1016 m; vent distance 0.1524; plenum volume 1.24542 m3; plenum vent area 0.02 194 m. The Statnamic simulation starts with the preloading of the pile by the reaction mass. After 20 ms, he1 combustion in the Statnamic chamber begins. The pressure in the chamber increases rapidly and as a result the mass is accelerated and launched. The resulting reaction on the pile top introduces forces causing stress waves to travel in the pile. Shaft friction and toe resistance will be mobilised. The displacement of the pile will have its influence on the burning fuel. The calculated displacement and force as hnction of time is shown in Figure 9. The predicted jump height of the reaction mass is 4.46 m, a conservative result since friction forces from gravel are not considered. The calculated force and displacement are about 10% to high. A good comparison can be obtained by using 10% less hel, but can not be currently confirmed by other back analysis. 4 CONCLUSIONS
For wave numbers greater than 12, the UPM (unloading point method) can be straightforwardly ap-
Bermingham. P.. Janes, M.. 1989. An innovative approach to load tcsting of high capacity piles. Proceedings of the International Conference on Piling and Deep Foundations. Lol1don. p.409-4 13. Bermingham. P., White. J.. 1995. @roteclinics and the accurate prediction of Statnamic peak laoding and fuel charge size. First In/erna/ional S/atnniiiic Senrinnr. I ancoiives, British Colunrhia, Canada. September 27-10. 1995. Bielefeld. M. W., Middendorp. P.. 1995. Statnamic Simulation. bi’rst International Statnainic Seminar, I ancoutter, Lh-itish Colimbia, Canada. September 27-30. Bielefeld. M.W.. 1991. Prediction of Installation of Sheet Piles using Vibratory Hammers. Fl’jih lnlernational C’onfesence and Exhibition on Piling and Foundations. DFI. 13- 15 June. Belgium. Bielefeld. M.W. and Middendorp. P.. 1992. Improved PiIe Driving PreQction for Impact Hammers and Vibratory Hanuners. Fourth Inrerna~ionalConference on the .4pplication of Stress Wave Theorv t o Piles. The Hague. 21 -24 September. Courage. W.M.G. and Bielefeld. M.W.. 1992. TNOWAVE automatic signal matching. Application of S/re.s.s-J b l w Theory to Piles, F.B.J. Barends (ed.) Balkema. Rotterdam. ISBN 9054100826. Courage, W.M.G. and Foeken, R.J. van, 1992. TNOWAVE Automatic Signal Matching for Dy-namic Load Testing. Application of 5’tre.w-Wave Theoiy to Piles. F.B.J. Barends (ed.) Balkema, Rotterdam, ISBN 9054100826. Foeken, van R.J., Daniels, B. and Middendorp, P., 1996. An improved method for the real time calculation of soil resistance during h v i n g . Application of Stress- Wave Theory to Piles, F.B.J.Barends (ed.) Balkema, Rotterdam. Hendriks, M.A.N., Oomens, C.W.J., Jans, H.W.J, Janssen, J.D. and Kok, J.J., 1990, A Numerical Experimental Approach for the Mechanical Characterization of Composites. In V. Askegaard (ed), Proceedings of the ninth International Conference on Experimental Mechanics, Aaby Tryk. Copenhagen. Horvath R.G., 1990, Statnamic, an accurate and innovative load test method for high capacity deep foundations. Foundation Drilling, Volume XXVIII, No. 11, Jazwinski, A.H. 1970, Stochastic Processes and Filtering Theoty. Academic Press, New York and London. Kalman, R.E., 1961. A new Approach to Linear Filtering and PreQction. Trans. ASME. Koten, H. van, Middendorp, P. & Brederode, P. van.,1980. An analysis of dissipative wave propa-gation in a pile. Intl. Seminar on the Application of Stress-Wave Theory on Piles, Stockholm, 4-5 June. Middendorp, P., Foeken, R.J. van, 1998. When to apply Dynamic Load Testing or Statnamic Load Testing, Second
580
International Statnamic Seminar, Tokyo, Japan, October 28-30. Middendorp, P., 1993. First experiences with statnamic load testing of foundation piles in Europe. 2nd International Geotechnical Seminar “Deep Foundations on Bored and Auger Piles”, Gent, Belgium, 1-4 June. Middendorp, P., Bielefeld, M.W.,1995. Statnamic and the Influence of Stress Wave Phenomena. First International Statnamic Seminar, Vancouver, British Columbia, Canada, September 27-30. Middendorp, P. and Weele, van A.F.,1986. Application of Charateristic Stress Wave Method in Offshore Practice. Proceedings of the third International Conference on the Numerical Methods in Offshore Piling, Nantes. Middendorp, P. and Zandwijk, van C.,1985. Accuracy and Reliability of Dynamic Pile Testing Techniques. Proceedings of the fourth International Conference on the Behaviour of Offshore Structures. pp 601-609. Middendorp, P. and Reidmg F.J.,1988. Determination of Discontinuities in Piles by TNO Integrity Testing and Signal Matching Techniques. Third International Conference on the Application of Stress Wave Theory to Piles. Ottawa, Canada, 25-27 May. Norton, J.P., 1986. An Introduction to Identification, Academic press, new York and London. Ser. D.J.,1960.Basic Engineering, Vol. 82, pp 95-109. Smith, E.A.L. 1960. Pile Driving Analysis by wave equation. Journal of the Soil Mechanics and Foundation Division, ASCE vol 80. SM4. Staalduinen. P.C. van, Bielefeld, M.W. and Middendorp, P., A.,1989. Direct Technique for Discon-tinuities in Piles. The second International Conference on Foundations and Tunnels. London. TNO report, 1985-1996. TNO-LILT Dyncmzic Load Testing Signal Matching, Users Manual. Ypey. L.P., 1984. De deconvolutie rnethode toegepast op het sonisch doortneten van fundatiepalen. Master’s Thesis, Delft University of Technology. The Netherlands (in Dutch). Voitus van Hamme. G.E.J.S.L, 1977. Bepalen van heiweerstanden uit meetresultaten. (In Ducth). Presentation on 13-eedenhriightiag. Voitus van Hamme. G.E.J. S.L, Jansz J.W., Bomer. H., and Arentsen, 1974. D., Hydroblok an improved pile driving analysis. De Ingenieur. Vol. 86. no. 8. pp 344-352.
581
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) Cl 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
A comparative study of static, dynamic and statnamic load tests of steel pipe piles driven in sand A. Shibata - Kubotu Corporation, Ichikawu, Japan N.Kawabata -Nippon Steel Corporation, Tokyo, Japan YWakiya - Kawasaki Steel Corporation, Tokyo,Japan YYoshizawa - Sumitorno Metal Industries Limited, Tokyo,Japan M. Hayashi - NKK Corporation, Kawasaki, Japan T. Matsumoto - Department of Civil Engineering, Kanazawa University,Japan ABSTRACT Comparative static, dynamic and the Statnamic load tests on three steel pipe piles driven in sandy ground were conducted. In this test program, dynamic load tests were performed at various time intervals after the initial driving to investigate the variation of the pile capacity with different hammer driving energies, and the increase in the pile capacity with elapsed time after the initial driving Two different computer programs were employed for the wave matching analyses to estimate the static load-displacement curves for the piles, which were compared with the static load test results The Statnamic test was also performed on one of the test piles that had underwent the static load test. This paper will discuss the uses of the dynamic and the Statnamic load tests as alternatives to the static load test 1 INTRODUCTION
3. To investigate the differences between the results of various computer programs used for the wave matching.
More than 20 comparative cases of the dynamic and the static load tests on driven steel pipe piles were collected by the Japanese Associatioa for Steel Pipe Piles (JASPP) to examine the use of the dynamic load testing to estimate the static load-displacement curve for steel pipe piles. The data were collected from the field tests with various pile configurations, various soil conditions, various driving hammers used, and various rest periods after the initial driving for the dynamic load test (Wakiya et al. 2000). The collected data suggested that the reliability of the static load-displacement curve estimated through the wave matching analysis of the dynamic load test signals depends on hammer energy, rest period before the re-driving test, the computer program used for the wave matching analysis and soil test data available for the site. Therefore, JASPP conducted their own test program of various load tests on three open-ended steel pipe piles in a relatively uniform sandy ground at Hasaki (the test ground of Sumitorno Metal Industries), Japan, in 1993, to evaluate the use of the dynamic load test and the Statnamic load test for piles in sandy soils as an alternative to the static load test. The emphasis was placed on the following goals in this particular test program: 1. To evaluate the influence of different hammer driving energies on the estimated pile capacity. 2. To measure the increase in the bearing capacity of the pile with elapsed time after initial pile driving, (the so-called "set-up" phenomena).
2 TEST DESCRIPTION 2 1 The f e s fsife md f e s fpr1e.r Figure 1 shows the soil profile and the results of site investigations at the test site The Standard Penetration Test (SPT) and the Cone Penetration Test (CPT) were conducted immediately before and 7 weeks after the pile installation The test ground consisted of fine to gravel sands from the ground surface to a depth of 20 m The SPT N-value measured prior to the pile installation was relatively high and uniform at depths greater than 7 m The ground water level existed at a depth of 5 m from the ground surface The variation with depth of the tip resistance, qc, from the CPT before the pile installation seems to be similar to the variation of the SPT N-values The sleeve friction, A, tends to increase linearly with depth to a depth of 9m and level off for larger depths The layout of the test piles, the location of site investigations and a piezometer are shown in Figure 2 PThree test piles, designated piles D for dynamic load testing, S for static load testing and M for soil measurements, were prepared The test pile specifications are listed in Table 1 Pile S and pile D were individually prepared, so that the dynamic and static load tests could be carried out at the same elapsed time after initial pile driving to minimize the 583
Figwe 1. Soil profile and results of site investigations at the test site.
Figure 2 Lavout of the test piles, the location of site inwtigations and a piezometer Table 1 Specifications of test piles Propert) Value L ni 13 0 Length Outer diaineter U inin 100 0 Wall tluckness t,, Illill 12 0 Cross-sectional area ,A ni2 0 017 Young’s modules E MNlin’ 206x10’ Mass densih p t/in’ 7 85 WaL e 1 elocity c m/s 5120 Mass .U ton 1.78 Cross-sectional area- ;1, includes the cross-sectional area of steel channels for protection of strain gages.
M also, so that piles D and M had the same configuration as pile S, although piles D and M were not instrumented with strain gages. The steel channels increased the net cross-sectional area of the test piles to 0.017m2. Pile M was prepared for the purpose of investigating the change in the soil conditions around pile M before and after the pile driving. The CPTs and the SPTs were conducted at a distance of 0.4m from the center of pile M immediately before and 7 weeks after the pile driving. There was little increase in N-values and qc-values, while there was a clear increase in&-values after pile driving to depths deeper than 8m. The excess pore pressures measured by the pre and post CPTs were very small indicating the relatively high permeability of the sand. An electric pore pressure transducer (piezometer) was placed in the ground at a depth of 11.5 m from the ground surface prior to the initial pile driving. The horizontal distance between the center of pile D and the piezometer was 0.4m, which was equal to the outer diameter of the test pile. 2 2 Test sequences and procedures
influence of the loading history for each pile. Pile S was instrumented with strain gages at a total of 6 levels. Steel channels were welded to the outside of the pile shaft for the protection of the strain gages. Note that steel channels were attached to piles I)and
584
A series of dynamic load tests was performed on pile D after 8 different time intervals, 0 min, 5 min, 15 min, 30 min, 1 hour, 3 hours, 20 hours, 6 days and 30 days, after the initial driving. In the dynamic load tests, the hydraulic hammer (HK65), which allows the control of the hammer driving energy, was used. Pore pressures were measured during the
pile driving. The dynamic signals (time variations of accelerations and strains at the pile head) were monitored by a PDA (Pile Driving Analyzer) and an FPDS (Foundation Pile Diagnostic System), which are widely used in the world on a commercial basis. Two different computer programs, CAPWAPC and TNOWAVE, were employed for the wave matching analyses of the dynamic load test signals to estimate the static load-displacement curve for the pile. A series of static load tests were conducted on pile S, separately. The first test was carried out 6 days after the initial piIe driving and the last test was conducted 30 days after the initial pile driving. The measured load-disp~acementcurves were compared with those estimated from the dynamic load tests on pile D to examine the applicability of the dynamic load test to the estimation of the static loaddisplacement curve for a pile. The Statnamic test was also carried out on pile S after the completion of the series of static load tests, 52 days after the initial pile driving (Nishimura et al. 1995).
of the shaft resistance along the top 3 5m of the pile at 6 days after the pile driving may be attributed to the fact that a gap between the pile shaft and the surrounding soil was generated by lateral movements of the pile near the ground surface during the pile driving. The shaft resistance along the top 3.5m of the pile recovered after a rest period of 30 days, although it is difficult to comment on this mechanism. 4 DYNAMIC LOAD TEST R E S a T S 4.1 The resirifsdirring the irzitiai driwig
Figure 5 shows the change in the total resistance, Rt, with penetration depth of the test piles during the initial driving. The total resistance, Rt, was estimated by the Case method (Goble et al. 1975). The total resistance of each test pile increases almost
Load on pile head (MN)
3 STATIC LOAD TEST RESULTS A total of 5 static load tests on pile S were carried out 6, 7 and 30 days after the initial driving as shown in Figure 3. The maintained step load test method was employed for the first and the last (5th) test Each load step was maintained for 15 min. The number of load steps was 5. The quick maintained load test method was employed for the Znd, 3rd and 4th tests, in which the pile head load was increased to the ~ a x i m u mload in about 20 minutes. It can be seen fiom Figure 3 that the curvatures of the loaddisplacement curves after the yield loads obtained from the first and the last tests are similar and they are more moderate than the other tests in which the quick ~aintainedloading method was used. It is also seen from the comparison of the 4th test and the last (5th) test that the yield load in the last test is smaller than that obtained from the 4th test, althoug~ the rest period for the last test is longer than the 4th test. This fact indicates that the influences of the loading rate andior the Ioading duration are not negligible even for sandy ground when it is saturated. Figure 4 shows the distributions of the shaft resistance, 2, obtained from the first static load test and the last static load test. The shaft resistance, 2, was estimated from the measured axial strains of pile S when it penetrated a distance equal to 10 percent of the pile diameter in each test. There is a slight change in the distributions of the shaft resistance after the rest periods of 6 days and 30 days, except for depths shallower than 3.5m. Accordingly, it is thought that set-up has completed within 6 days after the initial pile driving. The loss
Figure 3 Load 4splacenient c m e s obtained from static load tests (pile S).
0
Shaft resistance. z-(kN/m3) 50 100 150 200 250 300 6 daj s after EOID
"I
'
- 4
E
W
I I
Figure 4. Distributions of the shaft resistance obtaincd from the static load tests conducted 6 and 30 days afrer initial driving (for pile S).
585
linearly with the increasing penetration depth and attains about 2 MN at the end of initial driving. Therefore, the bearing characteristics of all the test piles are comparable with each other. 4.2 The ii$'iierrce of hammer driving energy The static pile capacity, R,,estimated from the dynamic load test may vary with the change in the driving energy actually transmitted to the pile To investigate this aspect in detail, the dynamic load tests were carried out on pile D changing the hammer driving energy after different time intervals from the initial pile driving Figure 6 shows the change in the static resistance, R,,estimated from the wave matching analysis of the dynamic load test signals using CAPWAPC, as a function of the nominal hammer driving energy, EH. The measured values of set per blow, S, are also shown in Figure 6 The static resistance and the set per blow increase gradually with increasing hammer driving energy to 75 kNm. The static resistance attains it's peak value of 2 MN when the hammer energy is greater than 80 kNm On the contrary, the set per blow increases steeply for hammer driving energies greater than 75 kNm, indicating that the pile capacity is fully mobilized by hammer driving energies greater than approximately 75 kNm.
It is recommended that the relations (EHvs R,and E I vs ~ 5') as shown in Figure 6 be obtained prior to the actual pile driving, in order to assure that the pile capacity is fully mobilized by the selected driving hammer. If EH versus R, alone is measured, it is difficult to judge that the peak value of R, is the actual pile capacity or it is a limitation due to the insufficient hammer driving energy. The relations shown in Figure 6 are useful for the selection of an appropriate hammer for each site. 4.3 Set-irp phenomerza
Dynamic load tests were performed after 8 different time intervals measured from the end of initial driving, in order to investigate the set-up phenomena of a steel pipe pile (pile D) driven in sandy ground. Figure 7 shows the change in the total resistance, Rt, which was estimated using the Case method, with elapsed time after the pile driving. Two dynamic monitoring systems, PDA and FPDS, were used to record the dynamic signals. The total resistance, Rt, reaches it's peak 60 min after the initial pile driving, and remains almost constant after that time. The peak total resistance is 1.2 times R, measured at the end of initial driving, that is to say, the "set-up ratio" for the total resistance is 1.2 in this case.
Figure 7 Increase 111 total resistance. R,. nit11 elapsed tiine after inihal driving (pile D)
Figure 8 shows the change in excess pore pressures with elapsed time after each blow for the different depths of the pile tip, during the initial driving of pile D Note again that the piezorneter was placed at a depth of 11.5m from the ground level and at a horizontal distance of 0.2rn from the outside of the pile shaft When the pile tip level is above the level of the piezometer, the magnitude of the first peak of the positive excess pore pressure, Au, increases, and the time instant of the peak of Air becomes earlier, as the pile tip approaches the piezometer level. When the pile tip level is below the piezometer level, the value of the peak positive excess pore pressure, Au, decreases as the pile 586
E
-0.1
.-.
z
Depth of pile tip = 10.0 in
2
2
$ a
> Depth of pile tip = 10.5 in
mn
0.1 0.0 -0.1
+I
initial driving of pile D were conducted to estimate the static load-displacement curve for the pile for each time interval after the initial pile driving. Figure 9 shows the measured wave signals of the re-driving tests. Figure 10 shows the results of the wave matching analyses. The wave matching analyses were conducted using CAPWAPC and TNOWAVE programs. The calculated and measured upward traveling forces are compared in Figure 10.
Depth of pile tip = 1 1.0 ni
0.1
6 da\ss after EOID.
U, m
(I)
3 w
Depth of pile tip = 1 1.5 in
1$ 0.1
0.0
-0.1
Depth of pile tip = 12.0 iii O l t - " " * " ' ~ 4
-0 01 000
° 001
002
003
004
~
005
h
-
Figure 8 Changes in elcess pore pressures uith elapsed time dunng each blou at different depths of the pile tip (pile D)
penetration proceeds It is interesting that negative excess pore pressure is generated after the peak positive pressure, although the magnitude of the negative excess pore pressure decreases as the pile penetration proceeds Although the measured excess pore pressures have not been fully interpreted, the influence of the excess pore pressures on the pile penetration resistance may not be negligible even for sandy ground One of the interesting features of the measured pore pressure is the fact that the pore pressure dissipated almost completely within 30 ms As indicated later, the duration of each dynamic load test was about 30ms also These facts show that the excess pore pressure is not accumulated during successive pile driving, resulting in the small set-up ratio of 1 2 for this site The set-up ratio for steel pipe piles driven in clay grounds is in the range of 3 to 5 (Wakiya et a1 1992, Matsumoto et a1 1995) In these cases, the existence of the accumulated excess pore pressures was confirmed It may be thought that the set-up is completed just after the end of each driving for sandy ground
& '5
6 days after EOID 30 days after EOID
-(b)
2 E
Time (s)
4 --
-: 32: 5 8
Ta)
10. - - - - -___.... -
-1:
~
"
"
"
'
20
15 0l 50
g
00
LL
-05 -1 0 -1 5 0
(a)TNOWAVE \\as use 10
20
30
40
30
40
Timc. t (ins)
20 15 h
z
10
E 05
2 00 r - ~-0 5 -1 0
4.4 Wave matching ctimlyses to estmiate the static load-d s p l n c e ~t ~cim'es l Wave matching analyses of the re-driving test signals recorded at 6 days and 30 days after the 587
-1 5 0
10
20
Time, t (ms) Figure 10 Coinpanson betu eeii ineasured and calculated upv ard tral eling forces in pile D (6 daj s after EOID)
calculated distribution of the shaft resistance along the pile shaft does not coincide with the static load test results. The distributions of the shaft resistance, z, estimated from the wave matching analyses using CAPWAPC were similar to Figure 11. In these wave matching analyses, the traditional empirical soil model proposed by Smith (1960) was used. The maximum soil resistance, the spring value and the damping constant for the Smith model were estimated only from the agreement between the calculated and measured signals. These factors may well be the cause of the difference between the estimated and the measured soil resistance as shown in Figure 11. The static load-displacement curves estimated using the wave matching analyses are compared with the load-displacement curves measured in the static load tests in Figure 12. The estimated loaddisplacement curves are comparable with the measured curves. Especially, the estimated initial pile head stiffness is fairly coincident with the measured values. This fact is usefbl in the limit states design and the performance based design of pile foundations, in which the estimation of the loaddeformation relation will be a vital issue (Kusakabe 1998).
Shaft resistance, z (kN/niL) Shaft resistance, z (kN/m’) oO
100
200
300
0
100
200
300
(a) 6 days after EOID (b) 30 days after EOID Figure 11. Distributions of shaft resistance estimated from redriving tests using TNOWAVE. together with static load test results (piles D and Sj.
5 STATNAMIC TEST RESULTS
Figure 12 Load -displacement cunes obtained from static load tests. and denved froin d!naniic load test using CAPWAPC and TNOWAVE (pile Dj
Figure 11 shows the distributions of the shaft resistance, z, estimated from the wave matching analyses using TNOWAVE and measured in the static load tests. Although the total shaft resistance estimated from the wave matching analysis is comparable with the static load test results, the
A loading device having a loading capacity of 8 MN was used in the Statnamic test conducted on pile S 52 days after the initial pile driving The measured variations with time of force, F,[,, displacement, w ,velocity, 11, and acceleration, a, at the pile head during the Statnamic test, are shown in Figure 13 The loading duration is about lOOms The peak ofF,,, is 3 68 MN, the maximum pile head displacement is 37 mm, and the residual displacement is 16 mm which corresponds to 4% of the pile diameter The maximum downward velocity is 1 m/s The maximum downward acceleration is 60 m/s2 while the maximum upward acceleration attains 120 m/s2 The FAln- w curve is shown in Figure 14 The pore pressures were measured during the Statnamic test also The magnitude of the pore pressure was very low, o 0 1MN/m2 in maximum The Unloading Point method analysis (Kusakabe & Matsumoto, 1995) and the wave matching analysis of these Statnamic test signals were conducted to derive a static load-displacement curve for the pile The KWAVE program developed by Matsumoto & Takei (1991) was used for the wave matching analysis Figure 15 compares the derived load-displacement curves and the load-displacement curve obtained from the static load test The curves compare well for practical purposes 588
Pile h e a d load (MN)
U
0
50
150
100
250
200
i
-
m
ro 0
50
150
100
200
73
250
(d
. E
.5i' 1.0
v
-ei-
0.5
"- 0.0
0' -0.5 0
_.
a, >
30
I : a,
-1.0
0
0
50
50
150
100
100
150
250
200
250
200
Time, t(ms)
Figure 13. Statnarnic test signals of pile S.
40 Figure 15 Load 4isplaceineiit cun es obtaiiied froiii the static load test, derir ed from Statnanic load test using Uilloadiiig Point method (ULPM) and derived from the wave matching anal! sis
capacity, Q, from the static load test, but overestimate the end bearing capacity, Qp, and underestimate the total shaft capacity, Qy. It is seen that the dynamic load test predicts well the total pile capacity as well as the proportions of the total shaft capacity and the end capacity.
Pile head force, Fll,,(MN) 0
1
2
3
7 CONCLUSIONS
4
The results of a comparative study of the static load test, the dynamic load test and the Statnamic test on three open-ended steel pipe piles driven in relatively uniform sandy ground have been presented, and the uses of the dynamic load test and the Statnamic load test in sandy ground as an alternative to the static load test has been discussed in this paper. The following conclusions were derived from this study:
40'
"
"
'
The hammer driving energy should be enough, so that the pile capacity is fully mobilized. A method to select an appropriate hammer driving energy was proposed, in which the mobilized static resistance as well as the measured set per blow are utilized to determine the minimum required driving energy (see Figure 6). Excess pore pressures are generated during pile driving even in sandy ground. However, the generated excess pore pressures dissipate within the duration of each driving without accumulation of residual excess pore pressures. Therefore, the 'set-up ratio' is relatively low in this case study. The set-up phenomena was completed within 60 min after the end of the initial driving process.
"
Figure 11. Frl,,versus II'
6 RELIABILITY OF DYNAMIC LOAD TEST
The ultimate bearing capacity of the test pile obtained from the static load test is compared with the bearing capacity derived from the dynamic load test and various pile design codes in Figure 16. OIn Figure 16, Q is the ultimate capacity which is the sum of the ultimate end bearing capacity, Qp, and the ultimate shaft capacity, 0,. Four Japanese design codes, which are based on the SPT N-values, predict well the total pile 589
0.5
0.0
1.0
15
20
2.5
35
3 0
4.5
4 0
Measured b! SLT (6 da! s) Measured by DLT (6 da) s)
I//
JHA S
P
T
l
JAC SPTl
V
1
/
JR SPT JPH SPT CPT method AASHTO S
P
T
l
I
] ~
AASHTO CPT Hilc!
I\
1 L.,'+C),
ith poicniial hamiiicr cncrg!
Hilc! v, 1111 iiicasurcd liainincr cncrg\00
Q +
10
20
15
I *
1 5
I
1
1
1
1
1 1
1
1
3 0
1
1
35
,
,
,
1
,
,
,
,
40
45
Ultimate End and Shaft Capacities (MN) JHA Japan Road Association J A C Arch~tecturalInstitute of Japan JPH Japan Port and Harbor Association JR Japan Railna! Figurc 16 Coiiipanson of end and sliafl capacitics obtained from the static load test n i l h Iliosc cstmialcd from d! naiiiic load test and dcm cd froni 1;irious p ~ l cdcslgn codes
The load-displacement curve derived from the dynamic load test is comparable with the curve obtained from the static load test, if the dynamic load test is performed after the same time intervals as the static load test, measured from the initial pile driving. However, the distribution of the shaft resistance and the toe resistance derived from the dynamic load test did not coincide with the static load test results. Estimation of soil parameters and selection of the soil model used in the wave matching analysis need hrther study. The load-displacement curve derived from the Unloading Point method analysis of the Statnamic test signals is also close to the curve obtained from the static load test. The ultimate bearing capacity and the Qp/Qy ratio derived from the dynamic load test are the most accurate, compared to empirical pile design formulas available in Japan and the other countries for this case study. This study is an activity of the Committee on Methods for Pile Capacity Prediction (formed in the Japanese Association for Steel Pipe Piles from 1991 to 1995 led by Professor Kazuma Uto, Tokai University) and the Research Group on Rapid Pile Load Test Methods (formed in Japan in 1993 led by Professor Osamu Kusakabe, Tokyo Institute of Technology). The members of these groups and Sumitomo Metal Industries Corporation are greatly appreciated for their valuable discussions, suggestions and support for this study.
REFERENCES Goble. G G . G E Jr Liluns. & F Rausche 1975 Bearing Capacih of Piles from D! iiaiiiic Measurements Final Report, Dept of Civil Eng Case nestern Rescne Unn . Cle! eland. Ohio Kusakabe. 0 1998 Changing foundation design code and the role of Statiiaimc test, Statnaiiuc loadmg test '98. Proc b 7 d h r S'rntlinim Seniiiinr, Tok! 0. Japan Balkema (to be published) Kusakabe. 0 & T Matsumoto 1995 Statnamic tests of Shonan test program nith re\ ieir of signal interpretation. Proc 1Y f hit Stntiiainic Seiiiiiinr 113-122 Vancom er Matsuinoto, T & M Takei 1991 Effects of soil plug on beha\ lour of dn\ en steel m e mlcs Soils K- Fouiidatioiis. JSSMFE, 3 l(2): 14-34 Matsumoto, T., Y.Michi, & T.Hirano 1995. Performance of axially loaded steel pipe piles dri\,en in a soft rock. .low: of Geotecli. Eiig., ASCE. V01.121. No.4: 305-315. Nishimura, S. & T.Matsumoto 1995. Wave propagation analysis during Statnamic loading of a steel pipe pile. Proc. 1st Int. Stntnnniic Seminar: 23-3 3 . Vancouver. Smith. E.A.L. 1960. Pile dnving analysis by the nave equation. Jour. Soil iifecli. Foiiiid. Div.. ASCE. Vol.86. No.SM4: 35-61. Wakiya, Y.. 0.Hasluinoto. M.Fukuwaka. T.Oki & H.Sliiiioiniya 1992. Abiliq of dynamic testing and evaluation of bearing capacity recovery from exess pore pressure measured in tlie field. Proc. of 4th Int. Cot$ oii the .4pplicatioii of Stress-Wave Theory to Piles (B.J Barends Eds.): 665-670. The Hague: Balkenia Wakiya, Y., M.Hayashi, T.Katayama & Y.Kobayashi 1994. Experimental inspection on accuracy of pile capacip prediction in dynamic loading test. Proc. 29th Japnnese .Aiinunl Meeting of Soil Mecli. B Found. Eiig.: 1431-1434 (in Japanese). Waluya. Y.. K.Nisliunu, M.Hayashi. A.Shibata & S.Nis1iinura 2000. The case studles of dynamic load test in Japan. Proc. 6th Itit. Coiif: OM the .Application of tlie Stress- W m e Tlieoy to Piles, Sao Paulo. Brazil (to be published). Yoshizawa. K., N.Kawabata. T.Oki & A.Shibata 1994. Vertical loading test of steel pipe pile dnven into sand focusing on set-up characteristics. Proc. 29th .Japanese ilnnual Meeting ofSoil-Wech. B Fhunrl. Eiig.: 1429-1430 (in Japanese)
590
.
'
I
1
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Case studies of statnamic load testing in Japan S. Nishimura - Fugro Geoscience Company Limited, Tokyo,Japan T. Matsumoto - Department of Civil Engineering, Kanazawa Universi~, Japan 0.Kusakabe -Department of Civil Engineering, Tokyo Institute of Technology,Japan K. Nishiumi -Nippon Steel Corporation, Tokyo,Japan Y.Yoshizawa - SuniitonzoMetal Industries Linzited, Tokyo,Japan
ABSTRACT: More than 70 Statnamic tests have been carried out on various types of piles in various soil conditions with various construction methods, since the Statnamic test’s first application in Japan in 1992. This paper reviews the Statnamic tests in Japan, presenting their chronology and statistics. This paper also discusses the a plicability of the Statnamic test to the estimation of the static load-displacement curve for a pile, based on tEe accumulated results of the Statnamic tests in Japan.
1 INTRODUCTION Foundation design in Japan is beginning to change from the conventional allowable stress design to the limit state design, or the performance based design, following the global trend in foundation design. In the new design methods, accurate estimation of the load-deformation curve as well as quality control of pile foundations will become an issue of vital importance (Kusakabe 1998). Hence, quick, cheap and yet accurate pile load test methods have been sought in Japan. The dynamic load test and the Statnamic test may be regarded as two of the newest developments in construction technology in this area. More than 70 Statnamic tests have been carried out on various types of piles in various soil conditions with various construction methods, since Statnamic test’s first application in Japan in 1992. This paper reviews over the Statnamic tests in Japan, presenting their chronology and statistics. This paper also discusses the applicability of the Statnamic test to the estimation of the static loaddisplacement curve for a pile, based on the accumulated results of the Statnamic tests in Japan. 2 STATNAMIC TEST DEVICES AVAILABLE IN JAPAN The first use of the Statnamic loading method (Bermingham & Janes 1989) in Japan was the load test of a cast-in-situ concrete pile in 1992 (Chosokabe et al. 1993). The Statnamic load testing method has attracted the interest of piling engineers
591
in Japan due to the simplicity and speed of the testing procedure. Statnamic test devices of 8, 16, 30 MN loading capacities are currently available in Japan. They can cover a range of loading force from 3 to 30 MN. The time required for pre-test preparation of the Statnamic test is shorter, while time requited for decomposition of the loading device is longer, because gravel is used to catch the launched reaction mass. Introduction of a system for catching the launched mass such as the hydraulic catch mechanisms (Bermingham 1995) would be desirable to shorten the total time required for the Statnamic test. Statnamic loading has been also applied to lateral loading of two steel pipe piles (Tada et al. 1997). 3
CHRONOLOGY AND STATISTICS STATNAMIC TESTS IN JAPAN
OF
Table 1 summarizes the Statnamic test conditions of the total of 42 sites in Japan. Kusakabe (1998) has reviewed statistics concerning these Statnamic tests using the database listed in Table 1. Some statistics are added in this paper. 3.1 Nuniber. of Stalnamic tests and test sites Figure 1 shows the chronological variation of the number of Statnamic tests conducted over the past 7 years, indicating a sharp increase in 1994 and a steady increase of 6 to 8 tests per year after that. A total of 74 Statnamic tests have been carried out at a total of 42 test sites.
Table 1. Achievement of Statnamic tests in Japan Date No (year111011t11)
Pile tyye
Pile Nuin- Num. Pile ber of of test diam.0 length tests piles (in) O(m)
'Ianned
1
92-5
CISC
4
3
1.3
14.5
5.0
2
92-11
Dri1,en SPP
2
3
93-5
CISC
2 2 1
0.8 1.4 1.2
11.0 13.5 13.5
5.0 80 6.0
3
93-6
BoredPHC
1
1
0.8
54.0
5.7
5
93-10
Driven SPP
1
1
0.4
15.0
4.5
6
93-1
DrivenSPP
2
1
1.4
25.5
2.51
Measured
5,0 5.7
Soil Type Sand I Sandy mavelE v Soft rock Reclamation / Sand
6.3
j,7 3.7
Notes Friction cut: 9m DLT SLT. DLT Friction cut: 13.5m
C Silt sandy gravel Friction cut: 39.5m SLT Sand SLT.DLT Mudstone
Ld = 13m
0.11
8 9
94-2 94-3
DrivenPHC BoredPHC CISC DSPP
21
96-4
CISC
22
96-6
Driven SPP
23 24 25 26
96-9 97-2 97-3 97-4
CISC CISC CISC BoredSPP
27
97-5
Driven SPP
28 29 30
97-5 97-5 97-6
Bored SPP CISC DrivenPHC
97-7 97-9 97-10 97-10
DrivenSPP
32 33
CISC BoredPHC
1 1
34
97-12
DrivenPHC
2
35
98-1
Micropile
1
94-1
31
~
5 6 1 1
2 3 1 1
0.3 0.3 1.5 1.4
7.0 7.0 10.0 29.5
0.300.6 0.300.6 13.0 5.0
1 1
1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1
0.8 1.6 1.6 1.6 1.0 1.5 1.0 0.9 1.2 1.2 0.6 1.2 0.6 1.4 1.4
47.5 29.0 30.0 29.0 12.5 9.0 13.0 18.0 19.5 21.5 20.0 13.4 11.0 70.0 36,0
10.0 10.0 0.3 0.3 8.0 13.0 7.0 13.0 15.0 15.0 8.0 16.0 3.5 16 12
1 1
1.2 0.6
--
6.0 28.0 12.0 12,0
13 4.5 5
0.18 1.5 1.5 1.5 1.0 1.2 1.2
21.5 46.6 46.6 46.7 36.5 13.5 14.5
1.9 24.5 24.5 24.5 7 16 16
1.2
8.0
12
1 1 2 1
36
98-3
DrivenSPP
3
37 38 39
98-7 98-8 98-8
DrivenSPP CISC CISC
1 1 1
1 1 1 1 1 1 1
40
98-10
CISC
1
1
41 42
99-2 99-3
CISC BoredPHC BoredPHC(SC) CISC: cast-in-situ concrete pile DLT: dynamic load test
Clay / Sand 12.1
10.4 11.5
Sandy/ Silw Shale Sandstone Shale
Frictionpile
8.0 13.7 6.3136.9
Sandy soil Aeolian rock Soft rock Sandy gravel
Recycled pile
15.4 15.3 16.0 3.7 15.5 12.3 13.0 3.8 5.2 5,0 23.8 27.2 22.8 8.0 15.2 15.3
Lateral STN
Mudstone
DLT, Lateral SLT
Sand Sandygravel Sandstone Sandy gravel Slate Sandstone I Shale Cemented silt
SLT SLT, Lateral SLT DLT DLT SITGSLL 0
Sand / Hard clay
Recycled pile
Fine sand
1 1 1.2 20.8 20 1 1 0.8 19.0 2 1 1 0.6 22.0 2 SPP: steel pipe pile PHC: precast concrete pile SLT: static load test SIT: sonic integrity test
592
SLT
Silt/Sandygravel SLT Mudstone Ld = 131n
SLT. DLT
Gravel sand
DLT
Calcareous rock Sandy gravel Sandy gravel Soft rock I Weathered rock Sandy gravel
DLT
Cemented silt SC: steel-concrete composite pile Ld:embedded length
3 . 2 Test objectives
Figure 2 presents a breakdown by objective of Statnamic tests at the test site. The Statnamic tests have been conducted to estimate the bearing capacity and load-displacement curves of newly constructed piles in 57% of the test sites, and of existing piles in 12% of the test sites. The Statnamic test was employed in 14% of the test sites for the development of new pile construction methods. The remaining were research projects, concerning aspects of the Statnamic test itself such as loading characteristics, dynamic effects, applicability to the estimation of static load-displacement curves and so on.
3.4 Pile length Figure 4 presents a breakdown of the number of sites with respect to pile length and pile type. Large shares of precast concrete piles having a length less than 10m and cast-in-situ concrete piles having a length less than 20 m are seen, whereas most steel pipe piles range in length from 10 to 40 m, 70 m in maximum. The longer lengths of the steel pipe piles can be attributed to the fact that most steel pipe piles are offshore piles, while all the concrete piles are constructed on land.
Figure 4. Breakdown of the nuniber of tests with respect to pile length and pile type.
Figure 2 Breakdown of the site number nith respect to test objective.
3.5 Soil types
3.3 Pile ~ p e s
The breakdown of the number of sites with respect to the kind of pile is given in Figure 3, where steel pipe piles have been tested at 43% of the test sites, cast-in-situ concrete piles at 38%, and precast concrete piles at 17% of the test sites, suggesting a rather even distribution over each kind of pile. There was one site where a micro pile having a diameter of 0.8m was tested. Out of 18 test sites with steel pipe piles, 10 sites were offshore piling projects where execution of a static load test would have been rather difficult. 593
Figure 5 shows the breakdown of the number of tests with respect to the soil type of the bearing stratum and the kind of pile type. Most piles are seated on firm soil layers such as rock, gravel, sand, soft rock or indurate silt. This is because of the fact that pile design codes in Japan recommend that the pile toe be seated on stiff soil layers. The bearing capacity of piles seated on rock tends to be underestimated in the pile design codes, so loading tests are often conducted on piles seated on rock to obtain realistic pile bearing capacities for design.
A breakdown of the site number with respect to planned maximum load is given in Figure 7. Most tests were conducted to confirm whether piles had design ultimate capacities or not. A merit of conducting a load test is that a safety factor of 3 can be reduced to 2.5 if a load test is conducted. Statnamic tests were conducted at only 3 sites to obtain bearing capacities, indicating that pile designers in Japan stick to pile design codes excessively. 3.7 Loading diu-atron Figure 5. Breakdown of the test iiuinber with respect to soil type and pile hpe.
A typical loading pattern of the Statnamic test is shown in Figure 8, with the definition of loading duration, tr, The relative loading duration, 7;, is defined as follows (Karkee et a1 1997).
3 6 Load magnrtirdes
Figure 6 presents a breakdown of the number of tests with respect to maximum test load and pile type Maximum loads are less than 16 MN for most piles in each pile type Note especially that the maximum loads are less than 8 MN for all precast concrete piles, because these piles were installed by means of a pre-auger method and had relatively low bearing capacities However, there is a clear trend of increasing testing load in recent years, an average of 4 88MN in 1996, 9 44MN in 1997, 16MN in 1998 and 13 3MN in 1999
T, = t L l ( 2 L l c ) in which L is the pile length and c is the bar wave velocity The relative loading duration, T,., is a measure for the influence of stress-wave propagation phenomena in the pile during the Statnamic loading. The Research Committee on Rapid Pile Load Test Methods formed in the Japanese Geotechnical Society (JGS) in 1996 regards the rapid pile load test as a test in which stress-wave propagation phenomena are negligible, although other dynamic effects such as the inertial force of the pile as a mass, radiation damping and viscous damping of the soil are not. They proposed T,. = 10 as a boundary between dynamic loading and rapid loading based on the work of Nishimura et al. (1998) (Research Coiiimittee on Rapid Pile Load Test Methods 1998). In the draft of “Method for Rapid Load Test of Single Piles (JGS 1815 - 200X)”, the rapid load test is defined as a load test with 5 < T,. < 500. A load test with 7;.greater than 500 is regarded as a static load test (JGS Committee on Standardization of Pile Loading Tests 1998) As seen from Figure 9, most Statnamic tests excluding 3 tests, had the relative loading duration, Tr, greater than 5 , satisfying the requirement for rapid loading h
z
z-U
CCI
0 -0
a, .-
Q Q
Q
0
30
60
90
Time (ms) Figure 8. Definition of loading duration.
594
120
150
the sites (for example, Terada et al. 1998). Static load tests were conducted mainly for comparative research of static load tests and the Statnamic tests. 4
LOAD-DISPLACEMENT STATNAMIC TESTS
BEHAVIOR
OF
4 1 Load-dispIacemerif curves Load-displacement curves from 5 1 Statnamic tests out of the 72 tests listed in Table 1 are overviewed to grasp the responses of the piles and the ground during the Statnamic tests with a wide range of loading forces The relationships between the pile head force, Fstn, and the pile head displacement, w, are shown in Figures 11, 12 and 13 for the cast-in-situ concrete piles (CISC), the steel pipe piles (SPP) and the precast concrete piles (PHC), respectively In these figures, Fsttl is normalized by the maximum pile ( zw( m is normalized ~~), by the pile head force, I ~ ~ t ~ land diameter, d The F&v curves may be classified into two groups In one group, the F\t,,-w curves have a characteristic 'tear drop' shape in which the pile displacement, w,tends to increase and decrease with increasing and decreasing the pile head force, FqtI1, regardless of the kind of pile In the other curves have shapes far from the group, the Fstn-~i' 'tear drop' shape In the Statnamic tests classified into the latter group, the pile displacement continues to increase after the start of unloading of I i m , although the recovery of the displacement occurs at a later unloading stage of F,t,, Such behavior of the J;stn-wcurve is notable for cast-in-situ concrete piles which have relatively large masses compared to steel pipe piles and precast concrete piles According to single mass modeling of a pile during Statnamic loading (Middendorp et a1 1992), the total soil resistance, Fsoli, is given as I;'. -1soil - 'stn - M a
Figure 9. Breakdown of the test number with respect to relative loading duration and pile type.
The Unloading Point Method (Midendorp et a1 1992, Kusakabe & Matsumoto 1995) is usually employed to estimate a static load-displacement curve for the pile in the case of Tr > 5 . In cases of Tr < 5 , the wave-propagation analysis (wave matching analysis) is usually conducted to estimate the pile capacity and the corresponding static loaddisplacement curve (for example, Ochiai et a1 1996). 3.8 Pcrr.allel test nielhods
The types of load tests other than the Statnamic test conducted at the test site are given in Figure 10 in the form of a breakdown by the number of sites The Statnamic test (STN) was carried out alone at 56% of the test sites, whereas dynamic load test (DLT) and/or static load test (SLT) were conducted in parallel at the remaining sites. At most test sites where the Statnamic test alone was conducted, it was confirmed that the pile had bearing capacity greater than the design load that was derived from pile design codes. On the other hand, Statnamic test results were used for calibration of a dynamic load test at the test sites where both load tests were carried out for quality assessment and driving control of the other piles at
in which A4 is the mass of the pile, and a is the measured acceleration at the pile head. The total soil resistance, Fsoil, is thought to be the sum of the static soil resistance, F,,, and the dynamic soil resistance, Fv, which is dependent on the pile penetration rate. The Fsoil - 1.1, curves of the Statnamic tests are shown in Figures 14, 15 and 16. Note that F'soil is normalized by the yield load of I'soil, FSoil(y1, that is defined as Fsoilat the maximum curvature of Fsoil - w curve. For most piles, the pile head displacement increases and decreases following the increase and decrease in Fsoil. It can be seen that the influence of the correction of the measured pile head force by the pile inertia, Ma, is predominant for the cast-in-situ concrete piles with large residual displacements. The Fsoil - IQ curves are again classifiable into two groups: the group with elastic responses in which the
Figire 10 Breakdonn of site number uith respect to load test methods
595
I”soli-
it’ curves have the characteristic ‘tear drop’ shape, and the group with plastic responses in which plunging loads are clearly detected in the F s o -~ w~ curves. It may be judged that the pile exceeded the ‘static’ yield load in the Statnamic tests classified into the group with plastic responses, although the dynamic soil resistance, F,,, which depends on the pile penetration rate, is not corrected for.
If the classification o f Statnamic test results into the
elastic or the plastic response type is possible from the measured pile head displacement immediately aRer the completion o f the ~tatnamictest, this may be useful An overview of the measured pile head displace~entsis provided below for this purpose The maximum and residual pile head displacements are shown in Figures 17, 18 and 19 for cast-in-situ concrete piles, steel pipe piles and precast concrete piles, respectively. In these figures, the Statnamic tests which were categorized into the group of plastic responses are marked with a star symbol. It can be seen that the Statnamic tests classified into
596
displacenients are small with ~ ! /typically d less than 0 5% As the number of the Statnamic tests on prestressed concrete piles at actual sites is still low (Tests 7-1 to 7-10 were performed by the Research Group on Rapid Pile Load Test Methods at the Shonan test site Kusakabe & Matsumoto 1995), it is difficult to determine the relationship between pile displacement and the classification of the Statnamic test for precast concrete piles at the present stage 3 3 Rehtiorishrp heh,rveri niaxiniuin F,,,,
~?~cixiiniini
I.,o,[ utrd ut?/oudulgYolrlt fi?fW,I T i LP.
Figure 20 presents the relationships between measured maximum loading force, F,t,,c,,,,,), and Definitions maximum total soil resistance, F,o,~(llld\) of f*~111(,,18~)and J*qoll(lnn,) are presented in Figure 23 The maximum total soil resistances, I*col~(,lla~), are almost equal to the maximum loading forces, l~rtn(nln~), for the wide range of I ~ ~ ~ t nIt~means l , l a \that ) the amplitude of f ~ ~ ~ l ~is( transferred l l l n ~ ) to the ground effectively Figure 2 1 presents the relationships between l~,o,~~l,la\~ and the Unloading Point force, fit ~ p which , is defined as I,\ollat the maximum displacement on f*
the plastic response group tend to have larger maximum head displacements as well as larger residual displacements than the Statnamic tests of the elastic response group for all kinds of pile. For the Statnarnic tests of the plastic response group, the amplitudes of the elastic displacement (difference of the maximum disp. and the residual disp.) seem to be similar in each kind of pile. The Statnamic tests on the cast-in-situ concrete piles are clearly divided into two groups: the Statnamic tests with large maximum displacements and the Statnamic tests with very small maximurn displacements. This fact may reflect the soil conditions of the bottom of the boreholes prior to concreting. For the steel pipe piles, 11 piles out of the 19 piles were classified into the plastic response group, although the amplitudes of the residual 597
This paper owes much to the activities of the members of the Research Committee on Rapid Pile Loading Test Methods in JGS, and permission to use the Statnamic test results which was given by the users of the Statnamic test in Japan. The authors wish to express their sincere gratitude to them here. REFERENCES Bcriniiigliain. P & M Jams. 1989 An innoLatii c approach for load tcstlng or high capacltj piles PIW 117r (‘017/ 017 Piling arid Deep Foundaliont London 109-413 Bcnningliaiii P . 1995 Futurc applications of tlic Statnamic method of testing foundations, l’roc 1ct 1/11 Smrnrrtnic Senirnnr, Vaiicom er, Canada 255-258 Chosokabe, M . K Yanasluta, M Kakurai. T F ~ i k u h a& T Yanxida, I993 A Statnanuc loading test applied for a cast-in-situ concrete pile, l’roc .lnniinl .Ueetii7g of lrchiteciurnl I n criiiire of Japnii I76 1-1762 Karkcc. M T Horiguclii & H Kisliida 1997 Static and dynainic icsts for c\dluahoii of tlic \crlical load bcanrig capacit! of pilcs. Proc Deep Foii17tlatio17r I I ~ W U 2nd I~ 1nniial 1Ieniher ’ 5 ( ‘014ere17ce.Toronto. Canada 199-214 Kusakabc. 0 . 1998 Changing roundation dcsign codc and ihc role of Slatnainic tcsi. Statnainic loading test Proc ,?lid hit Sinmnmic Seiniiinr, T o h 0. Japan Balkeina (to be published) Kusakabc. 0 & T Matswiioto. 1995 Statnainic tests of Slionan tcst prograin L\ 1121 rck icw of signal intcrprctation. Proc Fir$/lilt \tntnannc Setmnnr Vancow cr 1 13-1 22 Middcndorp. P P Bcnningham & B Kuipcr 1992 Statiiamic load tcstirig of fot1ndatlon pilcs. P i w J t h 1171 (‘oi7J on /he 1ppliCntiori o/ 5ti-et+Hki e T17eon to Pile5 (B J Baicnds Ed ). Tlic Hague Thc Nctlicrlaiids 581-588 Rot t crdaiii. Ba lkcnia Nisluinura. S . A Shibata, K Yaiiwshita, N K ~ t aN Ogita & M lsluda. 1998 One &ineiisional stress wile simulahon malj sis of Stahiamc load tests, Statnanuc loading test, Pmc 2 ~ dhit Srntmmrc Senrincw T o h 0, Japan Balkeina (to be pnnted) Rcscarch Connnittcc on Rapid Pilc Load Tcst Mcthods. 1998 Rcscarcli Actii itics tovard Standardmtion of Rapid Pilc Load Tcsl Mctliods i n Japan. Statriunic loading tcst l’roc 2nd lnr Sfntnniiiic Seiirrnar. Tolq 0. Japan Balkeiiia (to be published) JGS (Japanese Geotecluucal Socieg ) Conuiiittee on Standardization of Pile Loaduig Tests. I998 Draft of “Method for Rapid Load Test of Single Piles (JGS 1815 2000)”. Statnrunic loadmg test. l’roc 2nd Int S’tntnrnrr~c Senrmar. Tokyo. Japan Balkcma (10 bc publishcd) O c h m H . 0 Kusakabe. K Sum, T Matsuinoto & S Nisluniura. 1996 Statnanuc Tests on Offshore Steel Pipe Piles for Foundations of Access Bridge for New f i t a Kjushu Airport, Proc 4th 1171 Conf on the. fpplimnfion of Sfress-Ifirve Iheorv to Piles (B J Barends Ed ), The Hape. The Netherlands 997- 1014 Rotteddain Balkeina Tada. K . M Ohnishi. Y Cliono. S Nishiniura & S Sakaiiioto. I997 An cffcct of loading Lclocity r a m at latcral loading test for a steel pipe pde. l’roc 321id -lni?iinl Meerrng, Japanese Geotecluucal Societj . Kuinanioto 1601-1602 (in Japanese) Terada, M . R Nakatsuka, K Olmiwa, N Akisato, K Honkoshi & S Nishiinura, 1998 Applicafion of the Statnamc loadng test for large-scale pier construcfion Statnarmc loadmg test. Proc 2nd Int Statnainrc Seminar. Tokyo, Japan Balkema (to be published) ~
______------
pile obtained in the Statnamic test For the Statnamic tests classified into the elastic response group, F,o,l(md,) almost equals FI I P for most cases On the contrary, for the Statnamic tests classified into the plastic response group, t(\ol~(l,,d\)exceeds FITp by 5 to 30% of F1.1 p, indicating that sufficient load was applied to obtain the static pile capacity The difference between F\ol~(,nrl\) and FIIIP may be regarded as the dynamic soil resistance which depends on the pile penetration rate Figure 22 presents the relationships between k~~l,~llla,) and k i ~ pThis figure indicates that the planned maximum force could be determined to be I I to 1 3 times the expected static pile capacity 5 CONCLUDING REMARKS
An overview of the case studies of the Statnamic tests conducted in Japan has been presented in this paper, showing their chronology and statistics. The results of the Statnamic tests also have been reviewed from the view point of whether the ultimate capacity of the piles was obtained from the Statnamic test or not. It was shown that the ultimate pile capacity was obtained in a large number of the Statnamic tests, encouraging the use of the Statnamic test to estimate the design pile capacity. 598
Application of Stress-Wave Theory to Piles, Niyama & 5eim (eds)02000 5alkema, Rotterdam, lS5N 90 5809 1503
Statnarnic and dynamic load tests for large diameter steel pipe piles supported by a thin bearing layer at Nagoya port in Japan Y. Kikuchi Port and Harbour Research Institute, Yokosuku, Japan
S. Nishimura Fugro Geoscience Company Limited, Tokyo,Japan
M.Tat suta Nippon Steel Corporation, Tokyo,Japan
.-\BS'TR,ICT: Nagoya port is located in center of Japan and knows as one of the most active ports. Steel pipe piles were adopted for the development o f the wharf at the west area of Nagoya port. According to the expericncc o f Great Hanshin Earthquake. the seismic safety factors were considered !or the structural design. To save the construction cost, it was planned to install large diameter steel pipe piles into the thin gravelly sand stratum. Series of Statnamic and Dynamic load tests were carried out to estimate the bearing capacity of !he piles in order to establish a site-specific driving thrmula used for driving control of production piles at the site. I'his paper discusses the parameters !'or the design o!. [he toe capacity of piles that had seated on the thin gravelly sand stratum as well as shalt fiiction o f upper alluvial clayey stratum. And interpretations o f Statnaniic and Dynamic tests with its results are described. i
INTRODUCTION
At Nagoya Nishi-5 wharf (- 14m). locating near Nagoya City in Nonshu Island in Japan. the steel pipe piles were adopted for founding the wharf structures. To achieve economical construction. the piles were designed to drive -40.0 m below N.P. levei so that the piles would be seated on the gracclly dense sand stratum in spite of' penetrating into the deeper diluvial sand stratum existing the depth of' -68m. According to soil investigation rcsults. the soil stratification can be classified as shown in Table 1 . A thickness o f the gravelly sand stratum was relatively thin. varying from 4.5 to h.5m. as a bearing stratum against expected toe rcs i stance. To compare the toe resistances of' the t w ) dif'fcrent t y e s of' open-ended pile. which had just seated on the then bearing stratums. Statnaniic and Dynamic tests were c'arried out on the three test piles. One of the pile toe designs was simple open ended and other installed the cross-rib plates at the toe The instrumentation along the pile shaft was arranged in accordance with the soil stratification as shown in Figure I . 'Two test piles, V-1 and V-3. were simple open-ended pile and V-2 had the crossrib plates at the toe. The V-1 and V-2 piles were seated on the gravelly sand stratum having 4.5 ni thickness in order to evaluate increases of bearing capacity o f toe resistance due to the cross-rib plates. -. I he V-3 pile was installed on the stratum hacing 6.5 m thickness t o evaluate the effect of the thickness of bearing stratum to the toe resistance. 599
All tcst piles had a length of50m and diameter of 1.5ni. 'The wall thickness o f the pile was 17mm at the pile head and 15mm at the toe. The main purpose of the test was to confirm the design capacity of the pile as well as to establish the sitespecific pile driving control procedure to be applied for the driving ofother production piles in the site. 'Table I ~ . General_information on _
soil layers. -~
IIIc'iation
Soil
l'liick-
111)
t\ne
nzss
Remarks
Figure 1. Schematic diagram of test condition.
2 ESTIMATING DESIGNED VALUES OF PILE BEARING CAPACITY The bearing capacity of the piles was derived by the following methods with some assumptions.
2.1 Designed vulue of sha$,friction Equation 1 was established to determine cohesion variation from unconfined compression test results in the upper alluvial clay stratum. The equation has been used for calculating the design value of the shaR fiiction, C.
Two cases were considered for the assessment of shaft fiiction act on the pile. Because, soil behaviors might be change after the pile load tests. The upper soil deposit at the construction site had to be removed till the depth of -1 5.0m and the production piles to be driven in this site would be in the same condition including the test piles. Table 2 shows the calculated values of the design shaft fiiction. Table 2. Design value of shaft friction. Thickness of layer (m) Alluvial clay Case ,aver Bearing layer Test pile Production pile
C = 9.0 -t 3.02 (5 100 ) (kN/m2) Where, Z= depth from N.P.=O (m).
600
32.5 (-8.0t0-40.0) 24.0 (-15.0to-40.0)
1.5 (-40.0t0-41.5) I .5 (-4O.Oto-41.5)
Design total shaft friction (KN) 12240 10820
2.2 Designed value of toe resistance In this design, the effects of soil plugging and of the thin bearing stratum were evaluated and taking into the pile toe resistance. Increase of toe resistance due to the soil plugging effect was calculated by the following two methods. (a) The design code for steel jacket wharfs, in Japan
a = 2 X . ( 5 +X) (b) Specifications for highway bridges, in Japan
a = XI5 where a = Coefficient of soil plugging effect; X = a ratio of an embedded length into bearing layer / pile diameter, D (Xg5). Table 3 presents the values of toe resistance calculated f?om the two methods. The method (a) yielded larger value than the method (b). Table 3. Designed pile toe resistance. Method fa) Coefficient 0.333 of plugging effect Bearing capacity 6170 of pile toe (KN)
Table 4. Designed pile toe resistance considering the effect of thin bearing stratum. Thickness of bearing stratum (m) 6.5 4.5
Bearing capacity of pile toe (KN) Method (a) Method (b) 6170 3700 4760 3280
2.4 Effect of the cross-rib plate on pile toe resistance Increase of pile toe resistance due to the effect of the cross-rib plates were evaluated by using Yamahara’s method (Yamahara, 1963). According to this method, when the rib plate length is 1.5m, the resistance was found 1.5 times greater than that for the pile without cross-rib plate 2.5 Comparison of designed and required capacities Table 5 summarizes the designed values of pile capacities calculated by the methods described above. The total bearing capacity is shown in Table 5.
Method fb’l 0.200 3700
2.3 Eflect of thin bearing stratum to pile toe resistance Pile toe resistance decreases when a pile toe was seated on the thin bearing stratum. Toe resistance of the pile in this case was derived by the method, proposed by Hanshin Highway Department in Japan. In this method, the pile toe capacity is defined by the Equation (4). (4) where, q d = Pile toe resistance at the thin bearing stratum; q,= Pile toe resistance when the bearing stratum is continuous below the toe level; qd = Bearing capacity of weak layer existing below the bottom stratum; kf = Effective thickness of bearing stratum; D = Diameter of pile. Table 4 presents the toe resistance taking the effects of the thin bearing stratums of 4.5m and 6.5m thickness into the account. The table suggests that the calculated result for 6.5m thickness did not reduce regardless of applying the Equation 4 in comparison with the result of Table 3.
Table 5. Designed pile capacity. 4.5 m thick Shaft friction for working piles 10820 fKN1 Open end Cross-rib Toe resistance 3280 4760 (KN) Total bearing capacity 14 100 1 5580 (KN) Test pile v-1 v-2
6.5m thick I0820
Open end 6170 16990 v-3
The required bearing capacities as common and seismiclwind vertical loads for the production piles have been given in Table 6 that shows the values of designed capacities had enough allowance to the required capacities based on the designated vertical loads. Table 6. Required bearing capacity. Loading con- Designed Factor dition vertical load of (KN) safety Common 3500 2.5 Seismiclwind 4500 1.5
Required capacitv TKNj 8750 6750
3 LOAD TESTING PROCEDURE
3.1 Load test sequence Statnamic and Dynamic load tests were carried out on the three test piles to verify the bearing capacities and to establish the site-specific driving formula for 60 1
installation control of other production piles by driving. Table 7 presents the specifications of the test piles. Table 8 shows the load test sequence. As shown in Table 8, Dynamic load test was carried out five times, three times and twice on the V-I, V-2 and V-3 piles respectively. Statnamic test was carried out on each test pile 42 to 51 days after the end of driving. Pore water pressure was measured just before and after the load test performed. Table 7. SDecifications of the test tiles. Test pile v-1 v-2 v-3 Layer thickness 4.5m 4.5m 6.5m Pile toe Open Cross-rib Open $1 500mm x t 17/15mm(23/27m) x L5Om Size Steel (Tension strength = 490 N/mm2) Material
Table 8. Load tests seauence. V- 1 v-2 v-3 Test d e Laver thickness 4.5m 4.5m 6.5m Pile toe Open Cross-rib Open Elapsed time after E.O.D. (days) E.O.D. E.O.D. E.O.D. 2 days 1 days 23 3 days 25 days Dynamic load test 4 davs 28 davs Statnamic test 51 days 49 days 42days
3.2 Loud testing equipment To meet the loading condition of the test piles, Statnamic test was carried out by a 30MN loading capacity device manufactured by Berminghammer Foundation Equipment. A laser sensor was used to monitor displacement of the pile top during the loading. The strain gauges and accelerometers (50G) as shown in Figure 1 were used to monitor forces and accelerations of the test piles. Dynamic load test was performed by using IHCS90 hammer with ram weight of 70KN manufactured by IHC that generates impact energy of 90 KNm. However a fiee fall hammer with ram weight of 50KN was used instead of the IHC hammer for re-driving after 23 days, 25 days and 28 days to the V-1, V-2 and V-3 piles respectively. Force and acceleration responses during dynamic load test were monitored by FPDS system, which was developed by TNO.
4 TESTRESULTS 4.1 Dynamic load tests results As example of results from the dynamic load tests, the measured force and velocity multiplied by impedance curves were presented in Figure 2 and the measured and calculated upward traveling force wave curves in a wave matching analysis were presented Figure 3. They were obtained from the test on the V-1 pile at the end of driving. Table 9, 10 and 11 present the results of signal matching analyses of the dynamic load tests on the V-1, V-2 and V-3 piles respectively. Those tables show the following analysis results. At the end of continuous driving, toe resistance of the V-3 pile with the cross-rib plates was greater than that of the V-1 and V-3 piles without the crossrib plates. During elapsed time of 23 to 28 days, the total capacity of those piles had increased about two times of those at the end of initial driving. However
Figure 2. Measured force and velocity (Vl-03, E.O.D.)
Photograph I . 30MN Statnamic loading device.
602
Table 11. Results of signal matching analyses (V-3). V3-03 E.O.D. U.E. I,.I:. U.E. L.E.
2.0 4.0 4.0
2.0 6.0 18.0 18.0
Clayey (9.0m)
U.E. L.E.
6.0 6.0
35.0 35.0
1J.E. L.E. Sand U.E. (2.31~11 L E . Sand Toe
11.0 11.0 10.0 10.0 18992
56.9 56.9 15.0 15.0 23990
Total shaft resistance(KN) Toe resistance (KN) Total capacity (KN)
1044
A6QA
2702 3746
2813 7507
’
Table 9. Results of signal matching analyses (V-1).
.o
Sandy ( I 1 .Om) Clayey (20.0m) Yield stress (kN/m?
Figure 3. Measured and calculated upward force (VI-03, E.O.D.)
V3-R 1 23days
Clayey (4.34
1
V1-03 VI-RI V1-R2 VI-R3 VI-R4 E.O.D. 2 days 3 days 4 days Sandy
lJ.E.
L.E. Clayey U.E. (20.0m) L.E. Clayey (I.E. (9.0m) L.E. Clayey U.I. (4.3m) L.E. Sand U.E. (2.31~11 L.E. Sand Toe ( 1 1.0m)
_ _... ‘leld
stress (kN/m’)
Total shaft resistance(KN) Toe resistance (KN) Total capacity (KN)
2.0 6.0 20.0 20.0 35.0 35.0
toe resistance of those piles were assumed not hlly mobilized by dynamic loading due to the lack of impact energy by the hammer applied.
10.0 18.0 20.0 20.0 55.0 10.0 18.0 20.0 20.0 55.0 8.0 12.0 14.0 14.0 17.0 8.0 12.0 14.0 14.0 17.0 21791 21991 24990 29988 20491
Statnamic load test was performed on the V-1, V-2 and V-3 piles with elapsed time of 51, 49 and 42 days after initial driving respectively. Each Statnamic test result is summarized in Table 12 to 14. Figure 4 to 6 present the load-displacement curves of each test. Shape of the load-displacement signals were similar each other and suggests that the test piles were still in elastic region during Statnamic loading. Measured axial forces transferred to the pile toes were about 5000KN. To estimate static capacities of the piles fiom the Statnamic test results, the unloading point method named ULP method (Middendorp et al. 1992, Kusakabe & Matsumoto 1995) and the signal matching analysis was adopted. The ULP method was performed using two different acceleration values; one was measured at the pile head and another was averaged value along the pile axis. The reason of adopting averaged values was due to the long length of the test pile that had a relative loading duration, proposed by Karkee (I 998j, T,, of about 5. T,=5 is regarded as to be a boundary between dynamic and rapid loading (Japanese Geotechnical Society 2000) such as Statnamic loading. The signal matching analyses was performed. Smith model (Smith 1960) was adopted as the soil resistance model in the analyses. Figure 4 to 6 also present the force-displacement curves obtained from the results of analysis by the above three methods. Table 15
1.0 2.0 5.0 5.0 5.0 5.0
1.0 2.0 5.0 5.0 8.0 8.0
1052 2455 3506
1386 2477 3862
1.0 2.0 5.0 10.0 9.0 9.0
1727 2815 4542
1.0 2.0 5.0 10.0 12.0 12.0
1854 3378 5232
4.2 Statnumic loud tests results
4865 2308 7174
Table1 0. Results of signal matching analyses (V-2). V2-03 E.O.D.
V2-Rl 1 day
V2-R2 25days
U.E.
1.o
2.0
2.0
L.E. Clayey U.E. (20.0m) L.E. Clayey 1J.E. (9.0m) L.E. Clayey U.E. (4.31111 L.E. Sand U.E. ( 2 . 3 4 L.E. Sand Toe
2.0 4.0 4.0 8.0 8.0 8.0 8.0 9.0 9.0 18992
2.0 8.0 8.0 10.0
10.0 25.0 25.0 40.0 40.0 48.0 48.0 16.0 16.0 8397
Total shaft resistance(KN) Toe resistance (KN) Total caoacitv (KN)
1056
1695
3359 4417
442 1 61 15
Sandy (1 1.0rn)
10.0
12.0 12.0 15.0 15.0 24990
5493 1486 6979
603
Table 12 Statnamic load test result (V-1). Initial load (KN) Initial displacement (mm) Max. loading force (KN) Max. pile head displacement (mm) Max. pile head velocity (m/s> Max. pile head acceleration (m/s) Residual pile head displacement (mm) Max. pile toe axial force (KN)
1,653 2.5 23,861 54.5 1.47 48.92 12 4820
Table 14 Statnamic load test result (V-3). Initial load (KN) Initial displacement (mmj Max. loading force (KN) Max. pile head displacement (mm) Max. pile head velocity (m/s) Max. pile head acceleration (m/s) Residual pile head displacement (mm) Max. pile toe axial force (KN)
1.653 2:5 22,841 5 1.O 1.27 40.71 6.0 4995
Figure 6. Load-displacement curve (V-3). Figure 4. Load-displacement curve (V- 1).
Table 13 Statnamic load test result (V-2). Initial load (KN) Initial displacement (mm) Max. loading force (KN) Max. pile head displacement (mm) Max. pile head velocity (m/s> Max. pile head acceleration (m/s) Residual pile head displacement (mm) Max. pile toe axial force (KN)
Table 15 Static capacities calculated from Statnamic test results.
1,653 2.5 27,173 53.9 1.83 59.25 11.0 5273
* Signal matching analysis
Yield
Sandy (11.0m) Clayey (20.0m) Clayey (9.0m)
(kN/m2) Clayey (4.3m) Sand (2.3m)
U.E. L.E. U.E. L.E. U.E. L.E. U.E. L.E. IJ.E. L.E.
Toe Sand Total shaft resistance (KN) Toe resistance (KN) ‘Total capacity (KN)
v- 1
V-2
V-3
20.0 20.0 80.0 80.0 80.0 80.0 100.0 100.0 120.0 120.0 45000
20.0 20.0 80.0 80.0 80.0 80.0 100.0 100.0 150.0 150.0 40000
20.0 20.0 80.0 80.0 80.0 80.0 100.0 100.0 150.0 150.0 40000
15288
15613
15613
5069 20357
7076 22689
5069 20682
v-1
V-3
V-l
21391
24132
20509
20081
23292
19956
* ULP method Total capacity (Pile head Acc.’l (KN) Total capacity (Averaged Acc.) (KN)
Figure 5. Load-displacement curve (V-2).
summarizes the static capacities calculated by the three methods. The total static capacities derived from the results of signal matching analyses were estimated to be 20357KN for the V-1 pile, 22689KN for the V-2 pile and 20682KN for the V-3 pile as shown in
Table 15. Each value was greater than the designated bearing capacity. Figure 7 presents the comparisons between measured and calculated timepile head displacement curves as the results of signal matching analyses ofstatnamic test.
604
Figure 8. Comparison of distributions of shaft friction.
Figure 7. Results from signal matching analyses for Statnamic tests (Force - displacement relationships).
5 DISCUSSION 5.1 Shaft resistance
Figure 8 presents the Statnamic and Dynamic test results refer to shaft resistance. Figure 8 shows the Statnamic test results where shaft resistances are varying with depth. In the Figure 8, the designed values estimated from the unconfined compression strength (qu/2) that is represented by black lines. Black dots indicates the values measured by a fkiction meter, which is able to measure friction stress simply by turning a steel pipe with a diameter of l O O m r n installed into designated depth in the ground. From Figure 8, it is obvious that the distributions of shaft resistances estimated from Statnamic tests were close to the designed values and also the values measured by the friction meter. In the Figure 9, elapsed time-static total shaft resistances obtained from Dynamic and Statnamic tests were shown on the lower diagram. This figure shows that the total ;haft resistances had been
Figure 9. Variety Of pore water pressure and shaft fricti0n:V-1.
605
increased with elapsed time. During elapsed time of 42 to 51 days (1008 to 1270 hours), the total shaft resistances from the Statnamic test was greater than 15 to 20 times of the end of initial driving. From this result, followings can be concluded. The alluvial clayey soil around the pile was fully disturbed during the initial driving, hence the friction value is small. However after the set-up period, large set-up characteristics in the alluvial clayey soil is indicated. The upper diagram in Figure 9 shows recovery of pore pressures around the surface of the V-1 pile with time after the end of initial driving. Pore pressure gages were installed facing to the alluvial clayey at the levels of -32m and -42m from the pile head as shown in Figure 1. The upper diagram in Figure 9 shows that the measured pore water pressures at the both levels had been settling down in 1100 hours (46 days) and dynamic tests and Statnamic test had generated large pore water pressures around the pile. Figure 9 shows obvious correlation between the recovery of the pore pressure and the set up phenomena of the shaft resistances in elapsed time after the end of initial driving. 5.2 Toe resistance Toe resistances obtained fi-om the load tests were compared with the design values estimated. Those values are shown in Table 16. Table 16. Comparison of toe resistances. r - - - Statnamic (KN) Dynamic(KN) Lilbf: (Matching) (Matching) 3378 v -1 5069 442 1 v-2 7076 v-3 5069 2813
Figure 10. Comparison of load-displacement curves.
6 CONCLUSIONS Safety of the bearing capacities for steel pipe piles seated on the thin bearing stratum were confirmed by the result of Statnamic and Dynamic load tests. In addition, the site-specific driving formula to control driving of the production piles was developed based on the test results. The enforcement of the load tests at the wharf in Nagoya port materialized economical and reliable pile installation in such tricky soil conditions. Following interesting result was obtained from the series of Dynamic and Statnamc tests. Shaft resistance of the alluvial clayey stratum shows large recovery in its set-up behavior. Pore water pressures around the pile in the alluvial clayey stratum had increased at the time of Dynamic and Statnamic tests and settled down natural pressure within 1100 hours (46 days). The effects to the toe resistance due to the thin bearing stratum could not be observed from the test results. However, the pile with the cross-rib plates at the pile toe indicates large stiffness than that of without the cross-rib plates in the loaddisplacement curves.
Designed value (KN) 4760 7140 6170
The toe resistance was not fully mobilized in every case; therefore the effect of the thin bearing stratum (gravel with sand; 4.5m to 6.5m thickness) could not identify from the test results. Figure 10 shows the relationship between force and displacement at the head of test piles during Statnamic test. Stiffness of load-displacement curve of the V-2 pile with the cross-rib plates at the pile toe is larger than those of the V-1 and V-3 piles without the cross-rib plates. This certainly suggests that increase in load capacity of pile due to the presence of the cross-rib plates at pile toe. However, the piles were not loaded up to yield values, therefore it is difficult to verify an effect of the cross-rib plates on allowable resistance of the pile toe.
REFERENCES Japanese Geotechnical Society, 1998. Draft of “Method for Rapid Load Test of Single Piles (JGS 1815 - 2000)”, Proc. 2nd Int. Statnamic Seminar, Tokyo, Japan. Balkema (to be published) Karkee, M., T. Horiguchi and H. Kishida, 1997. Static and dynamic tests for evaluation of the vertical load bearing capacity of piles, Proc. 22nd Annual Member’s Conference, Deep Foundations Institute : 199-214. Kusakabe, 0 & T. Matsumoto, 1995. Statnamic tests of Shonan test program with review of signal interpretation, Middendorp, Proc. 1st Int. Statnamic Seminar, Vancouver, Canada, 113-122.
606
Middendorp, P., P. Bermingham & B. Kuiper, 1992. Statnamic load testing of foundation piles, Proc. 4th Int. Conf: on the Application of Stress-Wave Theory to Piles (B. J Barends Ed.), The Hague, The Netherlands: 581-588. Rotterdam, Balkema Smith, E.A.L, 1960. Pile driving analysis by the wave equation, J. Soil Mech. Found, Div., ASCE, Vo1.86, No.shd4, pp.35-61. Yamahara, H., 1963. Plugging effects and bearing mechanism for steel piles, Summaries of technical papers, Vol, 96, 97, Architectural Institute of Japan, (in Japanese).
607
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Statnarnic load testing using water as reaction mass M. D. Justason - Berminghammer Foundation Equipment, Hamilton, Ont., Canada M.C. Janes -Deacon Industrial Design, Vancouver,B. C., Canada P, Middendorp - TNO Profound, Deljt, Netherlands A.G. Mullins - University of South Florida, Tampa, Flu., USA
ABSTRACT: The following paper describes experiments performed in Hamilton, Canada, using water as reaction mass for Statnamic testing. These trial tests represented the first such tests performed in the world. These tests were performed in Lake Ontario, on a 323 mm diameter pipe pile driven in 8 in of water. The tests were performed using a 0.6 MN Statnamic device and hydraulic catching mechanism with a specially designed hanging structure from which a submerged steel container was used to contain a mass of water. The goal of the testing was to investigate the potential for using water to replace concrete and steel for use as reaction mass for Statnamic tests. This paper also describes Statnamic test results froin the first two Statnamic testing contracts for the Port of Lake Charles, Lake Charles, Louisiana, USA. These jobs were performed by Applied Foundation Testing, Inc. of Green Cove Springs, Florida, USA. These tests were performed on square concrete piles, with test loads ranging from 4.0 MN to 5.0 MN. This paper also summarizes the theoretical research conducted by Middendorp and Courage in 1995, which influenced the design of the actual water reaction mass assemblies for the experimental work of 1998. preceding theoretical work. This paper also describes the test results for the first contract Statnamic tests performed using water as reaction mass in 1999.
1 INTRODUCTION Statnamic load testing is frequently performed on foundations for bridge piers and port structures. Since these locations are over-water, a support structure is required for the Statnamic apparatus. The typical support structure consists of four temporary piles and a steel platform. Although this setup is not difficult, it does require extra time and cost to install. Alternatively, it is possible to construct a platform supported only by the test pile using a collar or sleeve that slides over the test pile. This method has been popular in Japan and has some advantages and disadvantages compared to the four temporary support piles. In both of the methods described above, the Statnamic reaction masses, base frame and gravel container must be mobilized to the jobsite, in addition to materials for the support structure. The idea of using the water surrounding the foundation to replace the conventional use of concrete and/or steel as reaction mass has been discussed for several years. The advantages of this idea would be: mass would not need to be mobilized to the jobsite; a support frame would not be required; and the time for setup and tear-down could be decreased. This paper describes the research and experiments performed in 1998 surrounding the use of water as reaction mass for Statnamic testing, as well as the
2 RESEARCH
2.1 Theory In a typical Statnamic test, a mass (usually concrete or steel) is used as reaction for an upward thrust produced by expanding gases within the Statnamic device. The masses are placed on top of the Statnamic device and are typically accelerated at about 20 times the acceleration of gravity in the upward direction. The result is a downward force on the foundation of 20 times the weight of the reaction masses. Although the applied force on the foundation cannot be sustained (force durations of lOOms are typical), the magnitude of the force is large in relation to the amount of mass needed. Even though the amount of mass required to perform a Statnamic test is small in relation to the mass required for a static load test, there is still cost involved in the mobilization of this mass. It had been proposed in the early 1990's, that for foundations situated in a marine environment it should be possible to use the readily available quantity of water to 609
was able to act on the mass of water as well as the mass of the steel container. As expected. the jumping height of the mass was greatly reduced. Unfortunately, this configuration would place greater physical demands on the catching mechanism. The third model presented the most attractive solution. In this model the top of the steel container was left open, and valves or ‘trap-doors‘ were introduced to the bottom of the container. This cdnfiguration allowed for the same mass of water to be ‘contained’ as in the first two models. thus providing equal inertial reaction for the Statnamic loading event. During the upward acceleration of the Statnamic device the valves in the container bottom remained closed. During the deceleration phase of the event the contained moving mass of water was allowed to flow through the container rather than ‘lift’ the container along with it. Modest jumping heights were observed in this container configuration, even when the container remaining completely submerged throughout the event. This was the most favorable model for the design of an over-water catching mechanism. A schematic of this third model is shown in Figure 1 . Figure I also contains the definitions of the variables used in the following equations. From equilibrium:
provide the mass needed to perform a Statnamic test, thus eliminating the mobilization cost of the concrete or steel masses. Middendorp and Courage of ” N O Building and Construction Research performed one of the first theoretical studies of water as Statnamic reaction mass in 1995. In this analysis three configurations of the underwater container were examined. The first mathematical model described a completely closed steel container, filled with water and submerged. The steel container was assumed to remain entirely below the water surface throughout the test. After the initial upward acceleration of the Statnamic test, it was calculated that the continued upward movement of the container would be very large due to the upward momentum of such a large volume of water. The deceleration force of gravity could only act on the mass of the steel container and not on the mass of water, which was weightless while submerged, and moving upward very quickly. The main advantage of this first model was that a relatively small amount of mass would need to be ’caught‘ by whatever type of catching mechanism was to be devised. The disadvantage was that an unrealistically large jumping height of the mass would need to be accommodated.
FSTN= Fa + F,, + F,
+G
Equation 1 can be rewritten as: d’x d’x . sin a t = Mw. -+ M,. -+ ... Fb,AX dt’ dt’
c, c,
As soon as the Statnamic force becomes zero, the steel containers will lag behind the accelerated water mass since they are decelerated by gravity. The valve at the bottom of the containers wi!l open automatically and the moving water will flow upwards through the container. In this condition the equation becomes:
drag coefficient water in container drag coefficient containers A, = cross section water A, = cross section containers FA = inertia force = M.a = =
M = M,+ M, Mw= mass water Mc = mass steel containers G = MF*g - V c . p g = gravity
Additional theoretical work was performed by Baddour in 1998, just prior to the construction of the first prototype underwater reaction mass container. The main issues addressed were: the additional mobilized mass of water ‘above’ the container, drag forces above and below the container, and drag forces on the sides of the container. Calculations generally agreed with the previous work by Middendorp and Courage. However, the work by Baddour focused more on the forces that would govern the structural design of the steel containers.
Figure 1 Force Diagram of Middendorp and Courage mathematical model of water reaction mass
The second mathematical niodel examined the simple case where the container described in the first model was allowed to rise above the surface of the water. In this case the deceleration force of gravity 610
Most of the remaining questions centered on the physical concerns of assembling such a test. The only remaining task was to undertake a physical experiment.
pile. The top of the container was constructed with a flat top with six 250 mm holes. The container was fabricated with a flat bottom with six 250 mm holes. The purpose of the bottom holes was to allow the container to fill with water as it was lowered over the test pile. Each 250 mm bottom hole was also equipped with a hinged door that opened in the upward direction. These doors were designed to remain closed during the upward acceleration of the container to prevent the water from simply flowing through in the downward direction. The doors did allow water to flow through the container in the upward direction after the completion of a test, to prevent the creation of a low-pressure area at the bottom of the container due to the upward momentum of the contained water and to act as the valves proposed by the Middendorp and Courage model. A photo of the submerged reaction mass assembly is shown in Figure 4.
2.2 Experiment
In the spring of 1998, a series of 12 Statnamic load tests were performed in Hamilton, Canada. These tests were conducted using a 0.6 MN Statnamic device and a prototype submerged reaction mass container. 2.2.1 Apparatus Testing was conducted using a 0.6 MN Statnamic device. A hydraulic catching mechanism was also used to facilitate repetitive testing. The testing apparatus, along with the fabricated reaction mass assembly is shown in Figure 2.
Figure 3 View of test setup clearly showing the support platform spanning the corner of the dock.
Figure 2 Test apparatus consisting of 0.6 MN Statnamic device, hydraulic catching mechanism, platform, and underwater reaction mass assembly (shown above water).
The test pile was located in a corner of the dock structure at the Berminghammer Facility, thus it was possible to span the corner with long beams to provide a temporary platform for the Statnamic apparatus. A view of the test setup can be seen in Figure 3. The fabricated steel reaction mass assembly consisted of a 1.07 m diameter steel casing 2.13 m in length. On the inside of the casing, another steel pipe was used to centralize the assembly over the 323 mm diameter pipe pile that was used as the test
Figure 4 View of submerged reaction mass assembly
High tensile strength anchor rods provided the load transfer from the Statnamic device to the top 61 1
For discussion purposes, the results of three typical loading cycles will be discussed.
and bottom of the reaction mass assembly. Two sets of three rods were used. The first set was 8 m long while the second set were 4 m long. The two sets were coupled together to provide the required length to perform one series of tests with the top of the reaction mass assembly 3.3 m below the water surface (masses stayed below the water surface during testing), and another set of tests with only 0.3 m of submergence (masses jumped out of the water during testing). The test pile was a 20 m long closed end steel pipe pile 323 mm in diameter with a wall thickness of 8.0 mm. The pile was installed using an MIST V16 vibratory hammer with a peak driving force of 1.1 MN. The soil at the test location consisted of 7.8 m of water, overlying soft marine silt sediments to a depth of 33 m. The purpose of the foundation was to provide a consistent elastic load deflection behavior for all tests with the loading capacity of the 0.6 MN Statnam i c device. After the first seven load tests using the above reaction mass configuration, a 250 mm wide flange was welded to the outer edge of the water container. This flange (added to the bottom outer edge), essentially doubled the plan area of the reaction mass assembly. This was done to investigate the effects of drag on the behavior of the underwater reaction mass assembly. The total reaction mass for this testing was 3350 kg, with 1490 kg (44%) provided by the Statnamic cylinder, silencer and reaction mass assembly. The remaining 1860 kg (56%) of the reaction mass was provided by the contained water.
Series 1, Test 4 - Results The first test to be discussed was performed during the first series of tests prior to the addition of the ‘drag flange’. The peak Statnamic load was 480 kN, with a measured upward acceleration of the reaction mass assembly of 10 g. Typical upward accelerations for the same test, performed with a typical amount of concrete reaction mass would have been approximately 15 g. The drag of the water provided a reduction in the upward acceleration of approximately 50%. On land, with conventional reaction masses, the jumping height of the masses would have been approximately 2.3 m, while the measured jumping height of the water reaction masses was only 1.7 m a reduction of 30%. The pressure transducers mounted on the reaction mass assembly suggested peak drag forces on the underside of the reaction mass assembly of 35 kN. The peak drag force on the top of the assembly was 58 kN. These forces occurred at the same time as the peak Statnamic load. Together the drag forces accounted for 93 kN, or about 19% of the applied load. Significant changes in the lateral pressure on the container were not observed suggesting minimal drag along the sides of the reaction mass assembly. 2.2.3.1
Series 2, Test 10 - Results This series of tests was performed with the additional drag flange welded to the bottom of the reaction mass assembly. The peak Statnamic load for this test was 496 kN. The load was very near to the previously discussed test, however the 3% increase in the peak force was achieved without increasing the amount of Statnamic fuel. This was likely achieved by a combination of the increased mass of the drag flange ( 1 20 kg), as well as the increased drag at the peak load. The peak upward acceleration of the reaction mass assembly was approximately 9 g rather than 10 g. And the jumping height of the masses was only 1.4 m, down from the previous 1.7 m, a reduction of 18%. The pressure transducers indicated similar pressures to the previous testing, both on the top and bottom of the reaction mass assembly. The increased area on the bottom of the container translated into an estimated peak drag force of 85 kN. The addition of the drag flange increased the total drag resistance to approximately 140 kN or 28% of the peak Statnamic load. 2.2.3.2
2.2.2 lnstrume ntation Accelerometers were placed on the reaction mass flange of the Statnamic device. These measurements were used to determine the acceleration of the under water reaction mass assembly. The jumping height of the Statnamic device was measured after each load cycle. Pressure transducers were mounted in various locations, both inside and outside the water reaction mass assembly. These measurements were made to help investigate the behavior of the water both inside and outside the steel container. In addition to the above special instrumentation, the typical Statnamic load cell, laser displacement sensor, and pile accelerometer was also used. 2.2.3 Typical Results For all of the loading cycles performed in this experiment, the test pile was not displaced beyond its elastic range, providing an essentially rigid test foundation. Peak displacements of 7-8 mm were typical for the test pile, with net displacements of zero.
2.2.3.3 Series 2, Test I ? - Results Test 12 provided data on one of the highest peak loads of all the tests performed. This test was also
612
performed with the drag flange included on the reaction mass assembly. The peak Statnamic load for this test was 573 kN, very close to the device capacity of 600 kN. For an equivalent land test using the same mass of concrete or steel, the peak upward acceleration would have been approximately 17 g. The peak measured acceleration was 11.5 g, a reduction of 32%. Pressure transducer measurements implied a peak drag force beneath the container of I I5 kN and peak drag forces on top of the container of 40 kN. In this case the drag forces accounted for a total resistive force equal to 27% of the applied load. Compared to Test 10, the drag forces provided a slightly lower percentage to the peak applied force, however the magnitude of the forces did increase with the increased applied load. This suggested that the drag forces provided additional resistance ‘for free’ under additional applied load. In other words, the total ‘effective’ reaction mass was increased without increasing the actual mass.
Figure 5 event
2.2.4 Discussion From the data presented here, it was observed that the drag forces on the top and bottom of the reaction mass assembly both contributed to the overall resistance provided to the Statnamic loading event. More generally, it was observed that the combination of water mass and drag forces provided an effective reaction mass. The valves proposed by Middendorp and Courage appeared to function as intended, although an additional series of tests were also performed with the valves welded shut. These tests showed only small increases to the jumping height of the masses. This was not readily explained by the theory. For tests with the valves open and with the valves shut, there was a visible plume of upwardly moving water that appeared at the water surface about 2 seconds after the Statnamic test (see Figure 5). Tests were also performed in which the submerged reaction mass assembly was permitted to exit the water surface. In these tests, the reduction in the jumping height was significant and, although spectacular, the tests it did produce an increased strain on the catching mechanism. Use of the hydraulic catching mechanism provided a convenient means of performing this research, however, the contribution of the hydraulic cylinders to the jumping height of the reaction mass was not quantified. One small test was performed without the use of the catching cylinders simply to observe the rate of decent of the masses afier the test. This test was not repeated due to the rapid rate of decent that was observed and due to the possible risk of damage to the equipment. This highlighted the need for a catching device.
Plume of water visible shortly after each loading
More detailed results and analysis of this series of tests are provided in Janes (1 998).
3 FULL SCALE TESTING Given the positive results of the small-scale testing in 1998, it was decided that the water reaction mass technology and theory was sound enough for use in the commercial sector. The American testing company, Applied Foundation Testing, Inc. in conjunction with Berminghammer Foundation Equipment of Canada collaborated on a project performed for the Lake Charles Harbor and Terminal District. Two testing contracts were undertaken for the Harbor. The first contract was performed in May of 1999, and involved the testing of two 600 mm square concrete piles. The test load for these piles was 4.0 MN. The second contract for the Harbor was performed in June of 1999 and involved the testing of three 750 mm square concrete piles and three 600 mm square concrete piles. The test load for these piles was 5.0 MN. 3.1 Apparatus Figure 6 shows an elevation schematic of the apparatus used for testing in Lake Charles. For this project it was not possible to use a supporting platform as for the experiments performed in 1998. Without the supporting platform, it was necessary to devise a catching frame that was solely supported by the test pile. The resulting design consisted of two 5 m steel truss towers, connected at the top and
613
Figure 7 Statnamic equipment being placed on a test pile Figure 7 also demonstrates the means by which the system was moved from one test pile to the next. The convenience of a large crane allowed the entire system to be moved from pile to pile without disassembly. On the second testing contract this allowed for as many as three piles to be tested in one day with the possibility for more Unlike the prototype equipment of 1998, the reaction mass container was not equipped with ‘valves‘ in the bottom of the steel casings. The trap-door type valves of the MiddendorplCourage model were replaced by two manually operated gate valves. The superior watertight sealing of these valves allowed the reaction mass containers to actually float for several hours while the above-water equipment was assembled on the test pile. Allowing the containers to float made the logistics of the test setup much more straightforward. The gate valves were opened just before the each test to allow the containers to fill with water and to sink to their starting elevation below the water. During testing the gate valves were left open. After the testing, the open gate valves allowed the water to drain from the containers as they were extracted from the water. In comparison to the 1998 testing, the Lake Charles under-water equipment contained a larger percentage of actual contained water mass. The
Figure 6 Elevation view of the Lake Charles testing apparatus
truss towers, connected at the top and bottom, with enough space in between for the Statnamic device. The catching mechanism was a mechanical latch system running on vertically mounted racks of ‘teeth’. Due to the reduced quantity of mass that was actually ‘caught’, it was believed that a simple mechanical system could provide the required catching capacity. This was in contrast to the hydraulic catching systems that are normally required for land based tests in which the entire reaction mass must be ‘caught’. The water container consisted of six, 1.2 m diameter steel casings with closed bottoms. The length of the containers was 2.0 m. The structural support for the casings was provided by two, 1 m deep steel I-beams, mounted below the casings. The upward Statnamic force was transferred to the containers through high-tensile rods connecting the I-beams with the Statnamic device, similar to the 1998 experimental testing. A photo of the system is shown in Figure 7, in which the entire reaction mass assembly js visible.
614
mass of water was 13,600 kg, while the mass of the containers and associated structures was only 9,400 kg. However, including the mass of the Statnamic silencer, high-tensile rods, and reaction beams mounted on the Statnamic device, the total nonwater mass was very close to the mass of the contained water. In general, the testing system was expected to behave much like the 1998 prototype. Perhaps the largest uncertainty in this test program was the performance of the new ‘mechanical’ catching mechanism. 3.2 Test Results Unlike the 1998 experimental testing, the main focus during the Lake Charles testing was the actual movement of the pile, rather than the movement of the reaction masses. Unfortunately, data was not collected on the upward acceleration of the reaction mass assembl>i. From the test data the force-time curves give the most information about the behavior of the testing apparatus. Jumping heights were recorded, but are not presented here. Figure 8, shows the force-time curve for one of the first 4MN tests on the 600 mm piles. Figure 9. shows a typical force-time curve for a conventional Statnamic testing using concrete and steel masses. The similarity between the two graphs shows the effectiveness of the water reaction mass assembly. Closer examination of Figure 8 shows a loading event of just over 100 ms duration. Figure 9 is at least 10 ms shorter in duration. This was the expected result since the land test was performed with a total reaction mass of 22,000 kg, whereas the water test used closer to 25,000 kg. Although the results were not surprising they do offer direct evidence of the effectiveness of the contained water as reaction mass for the Statnamic test. Also of ncte was the successful operation of the mechanical catching mechanism. As the Statnamic silencer reached the high-point of its jump it was noted that the subsequent downward movement was abrupt - similar to the 1998 experimental testing
Figure 9 Load vs Time using concrete and steel as reaction
when the catching cylinders were deactivated. As a consequence, the descending silencer and reaction masses produced a sharp impact load on the catching frame. Several of the steel teeth as well as the mechanical latches showed noticeable wear. The catching device functioned well nonetheless, and proved the idea of a mechanical catching mechanism to be sound. 4 SUMMARY AND CONCLUSIONS
The main conclusion drawn from the work described in this paper was that water could be used successfully as reaction mass for Statnamic testing. The theoretical work by Middendorp and Courage was useful in the design of the water containers. In particular, the use of a valve at the base of the water containers was highlighted in their work as an important element. In the 1998 experimental testing the mass of the steel containers was high in relation to the mass of contained water, hence the Middendorp and Courage valves had only a moderate effect. The containers used in the Lake Charles contract testing omitted the valves entirely. As the percentage of contained water increases in relation to the overall reaction mass (this is likely for larger capacity tests), it is theorized that the valves will become increasingly important. The experimental work of 1998 was the first testing of its kind, and proved the general concept of using water as reaction mass. Pressure transducers contained in the submerged reaction masses indicated that the drag forces of the container provided an additional resistance to the Statnamic event of 3030%. The moderate jumping heights observed in this testing also demonstrated the usefulness of drag forces. The final proof of the concept of using water as reaction mass was the contract testing in Lake Charles, Louisiana. This testing proved that a lowcapacity mechanical catching mechanism could be used to catch the water reaction masses. It also
Figure 8 Load vs Time usin$ water as reaction mass
615
proved that such a test could be assembled freestanding on the test pile, without the need for a support frame or special support structure. The success of the 1998 experiments as well as the Lake Charles testing point toward the future use of water as reaction mass for Statnamic testing in marine environments.
REFERENCES Applied Foundation Testing, Inc. Phase I1 - Report of Statnamic and Dynamic Load Testing, New Ship Berth at Contraband Bayou - Contract “A“. Prepared for The Lake Charles Harbor and Terminal District. c/o CBK Soils Engineering. Inc.. September 14, 1999. Applied Foundation Testing. Inc. Phase I - Report of Static, Statnarnic and Dynamic Load Testing, New Ship Berth at Contraband Bayou - Transit Shed. Prepared for The Lake Charles Harbor and Terminal District, c/o CBK Soils Engineering. Inc.. May 28. 1999. Baddour. R., 1998. Calculations modeling the reaction mass of hydraulic dynamic effect: Report fiom the tiniversigi of Western Ontario, Depr. of Civil Engineering, London, Ontario, Canada, I998 Janes. M.C. 1998. Statnamic testing using submerged reaction mass, Second International Starnamic Seminar; preliminary proceedings, Tokyo, October 29-3 I 1998. Tokyo: Japanese Geotechnical Society. Jonker. G. 1999. Use of a ballistic hammer for installation of anchor piles for off-shore structures. Conrinuous advances in Moorings Ce Anchors; proceedings, Aberdeen, 26-2 7 M ~1999. J London: IBC UK Conferences Limited. Middendorp. P.. and W.M. Courage, “STN above-water testing with undenvater containers“, Reporr ,from TNO Building and Construction Research, Delfi, Netherlands, 1995.
616
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Introducing statnamic load testing in Europe: Case studies in the Netherlands G.J. J.van Ginneken & l? Middendorp irTV0 Profound, Delft, Netherlands
ABSTRACT: A short introduction is given of the principles of the statnamic (STN) pile load testing. Examples are given of the applications of STN in The Netherlands and some European initiatives. Practical experiences with the use of the 4MN STN device with catching mechanism are described. Concluding economic considerations will be given when applying STN in practice.
1 INTRODUCTION
The duration and loading rate of a STN test can be controlled with the volume of the burning chamber and the shape of the cylinder and piston, the amount and type of fuel and the amount of reaction mass launched. As a result, the load can be introduced more gradually and for a much longer dura-
Statnamic started in the early nineties to satisfy construction Industry’s demand for cost effective and accurate means of testing high capacity foundations. Berminghammer Foundation Equipment of Canada joined with TNO Building and Construction Research of The Netherlands to develop STN. At this moment over one thousand STN load tests have been performed in Canada, United States, Japan, Malaysia, United Kingdom, The Netherlands and many other countries in the world. The equipment available at present can perform STN load tests starting from 0.1 MN up to 30 MN. Design and construction of even higher capacity STN devices (60 MN or more) can be expected in the coming years. The principle of STN is based on the launching of a reaction mass from the pile head. Launching takes place by generating high pressures in a cylinder, caused by the burning of a special fuel. As a reaction on the launching the pile is gently pushed into the soil. The load exerted on the pile head is measured by means of a load cell. The displacement of the pile head is registered by means of a special developed laser sensor. Load cell and laser sensor are integrated components of the STN loading device. No instrumentation has to be installed on the pile shaft. The required reaction mass is 5% of the force generated. The high capacity STN devices are not restricted to test single piles, but also allow the testing of pile groups and structural elements such as bridge piers and spread footings. Because the principle of STN is based on the acceleration of masses, piles can be tested in any direction, also horizontal and under batter.
Fig. 1. Force versus time diagrams
617
tests indeed showed the same or similar results as static tests.
tion compared to dynamic load testing. As an example the force time diagrams of dynamic load test, a STN load test and a static load test have been presented in Fig. I . The long duration of the STN loading keeps the pile under constant pressure and tension stresses cannot develop. The central location of the STN loading device on the pile top guaranties an axial introduction of the load onto the pile. The long duration of the STN loading causes for all pile levels a similar displacement behaviour as can be observed with static load testing. This justifies a simple modelling of pile and soil, in which stress wave phenomena do not have to be taken into account. The pile is considered to be a mass on which the STN force, the inertia force and soil resistance are acting (Fig. 2).
During the first years the desire to use STN for higher loads quickly rose. Statnamic devices to measure up to 30 MN were developed and are still used in larger projects. From the beginning the economic gains in these larger projects were evident. After first setting ground in Asia in mega construction projects STN is now quickly developing a wide acceptance in the USA as well. Testing is steadily growing, having performed up to present some 300 STN tests in the USA alone. The way STN was introduced in the USA was identical to its acceptance in the Far East. First comparative studies were carried out between static and STN load tests to confirm the reliability and accuracy of STN before the transportation authorities adopted the method. Several case histories in the Netherlands and Germany during the early nineties showed satisfactory and good agreement between static load behaviour from STN results and static load test results. It was concluded that: During STN load testing, pile behaviour can be modelled as a mass on which inertia forces and soil resistance are acting. This allows the simple calculation of the static load behaviour. Statnamic load testing can be performed on piles situated in soils with a strong dynamic response. The point of unloading (maximum displacement) in the STN load displacement diagram allows the direct calculation of the maximum static soil resistance during testing. Strain rate sensitivity has to be included for clayley soils.
Fig. 2. Forces acting on the pile during STN loading 2 HISTORY OF STATNAMIC DEVELOPMENTS
Worldwide during the last decade over one thousand statnamic tests have been carried out. The first developments of STN started in Canada, USA and the Far East where alternative ways of pile testing were introduced to accommodate for the high demand for pile testing in strong developing economies. There was much to gain, especially in the larger construction works where pile testing appeared to be cumbersome and expensive. Numerous comparisons between static and STN tests on driven and cast-in-situ pile have been made to convince authorities that STN proves to be good alternative for static load tests. Acceptance was acquired when authorities had seen that the results obtained from the STN
Europe has been somewhat reluctant to accept STN as the alternative testing method. Reasons for this are the existing building codes based on a traditionally strong preference for static load testing for well defined and safe cases. Moreover, for driven piles over the years a market for dynamic load testing had been developed and is now well established. The incentive to adopt alternative methods has not been large because existing regulations did and in many case still do not mention STN as an acceptable method for pile load testing and the market has the perception that STN testing is expensive. The question whether to perform pile testing or not in many cases depends on the funds available and the judgement of the experts on site. Money appears to 618
be dominant over quality assurance and as a result often a lot of construction works go without regular testing of foundation piles. The reason is that static load tests are considered too expensive, time consuming and therefore cumbersome. Dynamic load testing is not always recommended because of its limitations.
model for the strain rate sensitivity of clayley soils under STN loading conditions. Demonstrations have been trend-setting in adopting this new technology into the market. The method used by TNO Profound to introduce this new concept is to organise seminars and training courses for users and potential users to acquaint themselves with this new technology. This introduction involves a theoretical part in which the geo-technical background of STN is being explained and a practical part in which the actual testing is demonstrated.
For STN to be attractive applications are to be found in the lower pile loads testing range up to 2.53.5 MN and competition with established low cost static and dynamic testing methods is strong. In addition, regulations in most European countries are not yet adjusted in considering STN as an acceptable means for pile Ioad testing replacing or in conjunction with static load tests.
Collaboration exists with the leading Universities and Research Institutes in the countries involved securing adequate technical support, acceptance and being able to open steps towards introduction of STN into the national regulations. This means training, educating and continuously introducing technicians into the world of STN.
3 INTRODUCING STATNAMIC IN THE NETHERLANDS AND EUROPE
New countries covered are Spain and Turkey, where they are very interested to get STN introduced. Preparations are made for a first STN demonstration project in Barcelona. In Poland and Hungary progress is made to get STN into the regulations and perform static-STN demonstrations in 2000. Infrastructural EU supported investments for the preaccession countries are the driving force for the introduction of STN.
In response to the increased interest in statnamic in Europe TNO has decided to set up TNO Profound (Professional Foundation Diagnostics) responsible to introduce STN further in Europe. In combination with the knowledge base in foundation technology present at TNO Building and Construction Research a strong basis has made to make STN a success following the examples in the Far East and the USA. 3.1 Stntnarnic load testing with a 4MN device From its start in March 1999, TNO Profound has a 4 MN STN device at its disposal at a central location in The Netherlands for testing on the European continent. This has opened up the market for STN in Europe. Steps are taken to position a second 16 MN device in the Central European market with the aim to have a STN testing facility available on local markets to conduct testing services at commercially attractive prices. In 1999 TNO Profound has been engaged in several research projects in Europe in which the results of Static and Dynamic Load Testing have been compared with STN test results on the same and/or similar piles. Projects have been carried out with Jacbo (The Netherlands), WTCB (Belgium), the Technical University of Budapest (Hungary). Preparations are being made to initiate similar projects in Germany, although Germany has a strong tradition in static load testing the willingness to apply STN in Germany as well is rising now STN tests become economically more attractive. TNO Building and Construction Research is participating in a research project with the University of Sheffield to develop a
3.2 Major considerations for the introduction of Statnanzic on the European continent Regulations The tradition in most European countries is to conduct static load tests on specially made test piles. Loads to be applied range between 1.5 - 2.0 times the design load. These static load tests require a lot of preparation work and piles are to be carefully selected and prepared for testing. This is a costly and cumbersome activity. For this reason static load tests are virtually abandoned in The Netherlands for proof testing and are becoming less popular in other countries counties like Belgium as well. Time and construction costs are increasingly under pressure squeezing cumbersome testing practices to the bare minimum. In Europe most countries the use of alternative load testing methods is not foreseen in the national regulations. This hampers the wider introduction and acceptance of other innovative and more convenient methods. Mainly in Germany dynamic load testing has gained interest because of its convenience to use. A pile driving hammer or drop hammer can be used. Piles to be tested can be selected randomly. The same counts for STN where the reaction mass re-
619
quired is only 5% of maximum load to be applied. Construction engineers of local organisations may decide to accept DLT or STN when judged sufficiently suitable for its purpose. Their application depends on the availability of a testing capacity and is mostly used on a case-by-case basis, where convenience, time pressure and economics play a major role. For cast-in-situ piles STN is the preferred method (P.Middendorp, 2000). TNO Profound is promoting to consider also alternative testing methods like STN into the national regulations of EU countries. This is believed to be an essential prerequisite to get STN widely accepted in Europe. STN demonstrations with the involvement of the academia and regulation authorities are required to get decision makers acquainted with the method.
Statrzanzic as the only feasible alternative for testiiig
The major applications of STN in The Netherlands have been cases where the alternative testing methods where not seriously considered because of technical limitations (DLT on cast-in-situ piles would not generate satisfactory results), time pressure or undesired excessive costs. STN was not requested to replace routine static load testing, but was considered to identify the load bearing capacity of existing piles for re-use or in situations of doubt about the performance of disturbed soil profiles where conventional calculation and testing methods could not be applied. Groundwater seepage has been one of the major causes. STN would also be ideal in situations where a lot of piles are to be tested in a short time frame. For such larger projects STN is still lacking sufficient EU based reference projects.
Ecorzoinic considerations The price per pile for static testing varies considerably depending on the load to be applied and the specific site situation in which the tests are performed. When more piles are to be tested economies of scale cannot be made since each pile needs to be individually prepared. When one compares static load testing with STN, the testing costs of STN are usually 30-60% lower. The real economic gain is not so much in the testing costs as such, but in the considerable time saving when applying STN. For STN the preparation time in minimal, no reaction piles are required, piles for testing can be selected randomly and testing results are immediately available. This allows for instant decisions on additional testing requirements. For calibration purposes of STN it is always recommended to conduct one comparison test with a static load test in unknown soil conditions.
The price of STN and dynamic load testing is largely influenced by the costs of mobilizing the testing equipment. When the testing equipment is used for testing of one pile only the price remains relatively high. When more piles are tested on one project site the price per pile drops considerably. When comparing STN with DLT, pile preparation for STN is easier, but more time is required for the installation of the testing device. Four to 6 piles can be tested a day with a STN device with catching mechanism. Transportation costs in Europe are considerably higher than in the USA and allowable transportation weights per truck are maximized. Therefore in Europe it is preferred to move empty containers to be filled on the site with load material, whereas in the USA it’s easier to transport concrete/steel reaction masses. 4 CASE STUDIES OF STATNAMIC IN THE NETHERLANDS
4.1 Statizamic pile testing in Rotterdam Project Data Location : Rotterdam Harbour : Wilton-Feijenoord Time : June 1999 Client : Van Hattum & Blankevoort Testing : Prefabricated piles 420x420 4 piles, 23 m long (new) 2 piles, 24 m long (existing) Load applied : 2 -4 MN Project Description At one of the piers in the Rotterdam harbour a large construction crane was present for the construction of offshore vessels. It was intended to extend the moving range of the crane to an old reconstructed pier by using the existing foundation piles. It was questioned whether the existing piles under the pier would be suitable to bear this additional load. TNO Profound was requested to apply maximum loads in the order of 3.5 MN to check the load bearing capacity of the piles. Project Characteristics The piles had to be tested over water under difficult site conditions, limited availability of crane capacity and considerable time pressure (sea vessel to be loaded). static load testing and DLT not feasible because of shoreline conditions and accuracy. The 4 MN STN device was moved from pile to pile in one
620
Fig. 4. STN Testing at Eindhoven 4.2 Stutnamic pile testing in Eindhoven
Project Data Location : Eindhoven Railway Station : Large Shopping Complex Time : October 1999 Client : Inpijn-Blokpoel Consultants : Jacbo Avegaar piles (cast-in-situ) Testing 15 piles, 2 1 m long (new) Load applied : 2,5 - 4 MN
Fig.3 .4MN STN Device in Rotterdam. piece. Careful and accurate installation of the STN device above water appeared a time consuming process. Two piledday were tested with one- two testing cycles per pile.
Project Description A large shopping complex was to be built. Construction works were suffering delays and trouble with the foundations would negatively affect the timely progress of works. The piles were located in a 6-7 meter deep construction pit and testing had to be carried out on selected piles. After placing the foundation piles considerable groundwater seepage had occurred and maintained to grow during the construction works. The seepage was stopped through injecting the soil, but the soil stability around the piles was believed to be disturbed. Doubts arose whether the piles under the main structure would perform according the calculations based on soil investigation results and their capacity had to be determined.
Results The following loads were applied to the piles: Max load Displacement (in mm) (static) in kN (total) (permanent) Pile 1: 4190 Pile 2: 3980 Pile 3: 3960 Pile 4 (step 1): 2750 Pile 4 (step 2): 4340
20,s 19,5 18,s 12,l 17,6
395 377 530 2,o 1,5
Project Characteristics First a reference test was carried out on a pile installed in an area which had not been affected by seepage. The results of the STN test on the reference pile were compared with the test results of the seepage affected piles. It was also indicated that the amount of piles to be tested would depend on the testing results. The piles selected for testing were indicated from the ones most severely affected in concentric circles going outwards towards less seepage affected piles.
Conclusion The conclusion was that the existing piles performed above expected calculated bearing load and could be safely used for the pier extension. After the first tests the client wanted the tests to go further up to failure to determine ultimate load capacity. The tests showed that the performance of existing piles may well exceed calculated capacities.
621
Results The aim was to minimize the time spent on testing. During the first tests it appeared that cyclic testing would have a positive influence on the performance of the piles in working conditions. It was decided to conduct 2-3 cyclic loads per pile on the piles most seriously affected. The following loads were applied to selected piles: Max load Displacement in mm (static) in kN (total) (permanent) Pile 1 : 2220 Pile 2: 2130 Pile 3: 3460 Pile 4 (step 1):2450 Pile 4 (step 2): 31 10
11,l
8,1
27,9 9,1 18,5
790 4,l 21,5 2,4 3s
Project Characteristics This project is typical for a common situation in the Netherlands where the standard procedure is not to conduct static load tests on a regular basis, but to rely on the results of CPT tests to approve piling jobs. In case of poor CPT results construction engineers are only inclined to approve the foundations when the piling contractor can prove that the piles installed perform within the limitations. TNO Profound was requested to test and provide an independent judgement on the load bearing capacity of the Omega piles. Cyclic STN tests on a reference pile and on a questionable pile were compared to assess the difference between the two and to judge the acceptability of the tested piles. On both piles 4 cyclic load tests were carried out up to failure capacity to fully satisfy the demands of the construction engineer.
Conclusion It was concluded that the piles proved to be just within the margin of acceptance. Signal matching with TNOSTN revealed that the soil skin friction in the upper 4 m layer was reduced to only 35% to of the total due to the effect of seepage. The total pile capacity was seriously affected by this, but the lower soil layers compensated for the loss occurred and would still have sufficient capacity to provide the necessary support for these pile foundations.
4.3 Statnamic pile testiizg in Utrecht Project Data Location : Utrecht : Jaarbeurs Fair Complex : December 1999 Time Client : Jacbo/HBG : Jacbo Omega piles (Auger) Testing 2 piles, 18 m long (new) Load applied : 2,5 - 3,5 MN
Fig 5. 4MN Statnamic device in Utrecht Results The aim was to minimize the costs and therefore the amount of time spent on testing. The results were satisfactory and the foundation piles were accepted by the construction engineer. The following loads were applied to the piles:
Project Description At the Jaarbeurs fair grounds of the Municipality of Utrecht, new high rise office buildings were planned for construction. For the Omega pile type it was a routine procedure to compare the CPT results derived from tests before and after the pile construction works. These values are then being compared to check for a reduction of CPT values caused by the installation of the piles. In one of the a corners of the building site the CPT results after piling showed values lower than the set limits. STN tests were carried out to verify that the capacity and settlements of the piles would be within the allowable limits.
Max load
Displacement in mm (static) in kN (total) (permanent)
pile 1 pile 2 pile 1 pile 2 pile 1 pile 2 step 1 step 2 step 3 step4
622
1690 2170 2190 3110
1850 16,3 19,2 2570 16,O 15,5 2730 9,9 36,l 3140 18,5 38,5
12,O
13,O 6,O 9,O
10,O 12,O 28,O 31,O
Conclusion Statnamic proved to be a simple and effective way to test the performance of cast-in-situ foundation piles in cases where CPT tests would give doubtful results. STN testing confirmed that in some cases the interpretation of CPT tests alone could lead to questionable conclusions. With STN an effective tool was provided to verify independently the actual performance of the pile tested.
5 CONCLUDING REMARKS When considering the STN testing worldwide it can be concluded that nowadays every other day a STN load test is carried out somewhere in the world. USA, Asia and Japan are taking the lead. Europe is a region where STN has a large potential to grow. TNO Profound has actively started to promote STN in Europe and aims to motivate centres of expertise in European countries to initiate STN activities. STN can replace static load testing, however, when deriving design rules, static load testing will remain the preferred testing method.
6 REFERENCES Bermingham P., Janes M., 1989, An innovative approach to load testing of high capacity piles, Proceedings of the International Conference on Piling and Deep Foundations, London, p.409-413. Middendorp, P., Bermingham P., Kuiper B., 1992, Statnamic load testing of foundation piles. 4th International Conference on Stress Waves, The Hague, Balkema Middendorp, P, 1993, First Experiences with Statnamic Load Testing of Foundation piles in Europe, Proceedings 2nd International geotechnical seminar on Deep Foundations on Bored and Auger Piles, Gent, p .265-272, Balkema Brown, D.A., 1994, Evaluation of Static Capacity of Deep Foundations from Statnamic Testing. Geotechnical Testing Journal, Vol 17, No.4, American Society for Testing and Materials Matsumoto, T., Tsuzuki, M., 1994, Statnamic Tests on Steel Pipe Piles Driven in a Soft Rock. International Conference on Design and Construction of Deep Foundations, Orlando, U.S. Federal Highway Administration Middendorp, P, Bielefeld, M.W., 1995, Statnamic Load Testing and the Influence of Stress Wave Phenomena, First International Statnamic Seminar, Vancouver 623
Middendorp, P., Foeken, R.J. van, 1998, When to Apply Dynamic Load Testing and Statnamic Testing, 2nd Statnamic Seminar, Tokyo Middendorp, P, Ginneken, G.J.J. van, Foeken, R.J. van, 2000, The advantages and disadvantages of D namic Load Testing and Statnamic Load Testing. 6' International Conference on the Application of Stress Wave Theory to Piles, Sao Paulo, Brazil
x
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 750 3
The advantages and disadvantages of dynamic load testing and statnamic load testing l? Middendorp & G.J. J.van Ginneken TNO Profound, Delft, Netherlands
R. J.van Foeken TNO Building and Construction Research, Department of Structural Dynamics, Rijswijk, Netherlands
ABSTRACT: Pile capacity testing by high strain dynamic loading methods is widely applied because of its economy and efficiency compared to static load testing methods (SLT). Frequently applied dynamic loading methods are dynamic load testing (DLT) and statnamic testing (STN). The paper will deal with the very often raised question in practice: What are the advantages and disadvantages and when to apply DLT and STN when pile type and soil conditions are known. Special attention is given to DLT on cast in situ piles, and complicating factors like limited knowledge of concrete material properties and pile shape. The suitability of DLT and STN is discussed for cast in situ piles and driven pre-cast piles by the evaluation of reliability, economy, mobilization of capacity and the chance on pile damage.
1 INTRODUCTION
Pile capacity testing by high strain dynamic loading methods is widely applied because of its economy and efficiency compared to static load testing methods (SLT). The most popular dynamic loading methods are dynamic load testing (DLT) by an impact hammer and statnamic testing (STN) by launching a reaction mass from the pile head. DLT introduces a short duration shock pulse into the pile. STN generates a relative long duration push load onto the pile head. Extensive descriptions of load testing methods and comparisons are published by Holeyman (1992) and Karkee et a1 (1997). However these papers do not deal with the very often raised question from practice: What are the advantages and disadvantages and when to apply DLT and STN when pile type and soil conditions are known. The answer to this question will be treated in the next paragraphs . Special attention is given to DLT on cast in situ piles, because the calculation of the pile load is based on signals from strain transducers mounted on the pile shaft. So for DLT the pile load calculation depends strongly on pile material and cross section properties and factors complicating the analysis like limited knowledge of concrete material properties and pile shape are discussed.
Fig. 1 statnamic test on a cast in situ pile
625
-
-
economy mobilization of capacity chance on pile damage
2 THE APPLICATION OF DLT AND STN ON CAST IN SITU PILES For cast in situ piles both DLT and STN are performed a certain period after pile production, to allow the piles to reach the required compressive strength to withstand the test loads. For DLT strain and acceleration transducers are mounted on the pile shaft near the pile head. The load displacement behavior is calculated by signal matching. For STN the load displacement behavior is calculated in most cases by the Unloading Point Method (UPM), however signal matching techniques are also applied.
2.1 Accuracy in load measurement for STN With STN the load is accurately measured by a calibrated load cell placed on the pile head. The measured load is not dependent on the pile properties. The load measurement error is less than 0.1% of the maximum range of the load cell. 2.2 Material properties and accuracy in load measurement for DLT
Fig, 2. Dynamic load test on a cast in situ pile.
With dynamic load testing strain transducers are mounted on the shaft near the pile head. The load (F) on the pile head is calculated by multiplying the measured strain (E) with the modulus
Fig. 3. Statnamic piston with built in load cell and laser displacement sensor placed on a cast in situ pile
Finally the suitability of both DLT and STN will be evaluated by taking into account the following points: -
-
Fig. 4. Strain transducer mounted on the shaft of a cast in situ pile
accuracy of the load measurements reliability
626
of elasticity (E) of the concrete and the pile cross section (A). F=E. A.&
(1)
The accurate determination of the properties E and A for bored piles is difficult in many cases. To calculate the force from the measured strain in a pile during DLT we need to know the cross section and the modulus of elasticity of the concrete at the measuring level. For piles with homogeneous material the stress wave velocity (c) is used to calculate the E-modulus with E = c2. p c = 2L/T
(3)
Knowing or estimating the stress wave velocity c we can calculate the pile load at the measuring level with the formula
F= c’.p.A.
(4)
So the derived stress wave velocity has a strong influence on the value of the load measured in the pile. An error in the measured load will result in an error for the pile capacity prediction.
The stress wave velocity is calculated from the time (T) it takes for a stress wave to travel over the pile length (L) from the pile head to the pile toe and back to the pile head (Fig. 5). For this method it is required that the reflection coming from the pile toe is clearly visible in the signals. In Fig. 6 the force and velocity times impedance signals of two dynamic load test are presented. The first case shows a clear toe reflection and the stress wave velocity can be calculated accurately. If the toe reflection is not visible one has to estimate the toe reflection time. However an error in the estimated toe reflection time (T) and stress wave velocity (c) will result in a considerable error in the calculation of the E-modulus. For example a 5% error in the stress wave velocity will result in a 10% error for the E-modulus and corresponding load in the pile. Another option in this case is to rely on an estimate for the E-modulus from the pile material properties. Making an estimate on the E-modulus is difficult because it is not a constant value but depends on the age and the quality of the concrete (Franklin, 1971)(Fig. 7), the loading rate (Sparks et all, 1973 ), and even the temperature of the concrete (Abbasi 1990). For example, for static load testing the modulus of elasticity for concrete is in the range of 28 GPa to 32 GPa while for dynamic load testing it is in the range of 32 GPa to 52 GPa
Fig.5. Calculation of stress wave velocity c from toe reflection
627
Fig. 6 Well visible and no visible toe reflection.
Fig. 7 Relations between dynamic modulus of elasticity and age for concretes made with various aggregates
Another complicatjng factor in determing the stress wave velocity c tor cast in situ piles is the fact that the concrete is not homogeneous. The concrete quality will vary over the cross section and over the pile axis. The concrete in contact with the soil will be of lesser quality than the concrete in the center of the pile and the shaft area that has been in contact with the soil might be the location where the strain transducers are mounted. The concrete quality difference over the pile length is caused by the pouring procedure and the difference in concrete pressure
628
during construction. The quality of the concrete near the toe will in general be better than the quality of the concrete near the pile head. This also means that the stress wave velocity will vary with the pile length. So the stress wave velocity calculated with c=2L/T is a mean value for the whole pile. The modulus of elasticity calculated from the toe reflection represents a mean value for the pile and there can be a considerable difference with the modulus of elasticity at measuring level.
urement results. During STN the load duration is long enough that all pile parts move in the same velocity range. Under these conditions the pile can be considered to act as one mass with a pile rigidity behavior similar to static load testing (Middendorp 1995 ) For this reason pile behavior during STN is closer to static load testing than DLT.
2.5 Economy
Fig 8. Cast in situ pile with bulb.
2.3 Injluence of pile cross section variations on DLT capacity prediction To predict capacity from DLT results, signal matching techniques are the most frequent applied methods, (TNOWAVE, CAPWAPTM).Based on a wave equation computer program calculated signals are matched with measured signals by adjusting the computer soil model and pile model in an iterative way. When signals match it is assumed that the computer soil model represents the real soil behavior and the static pile capacity is calculated from it. Pile discontinuities like necking, bulbs, and material changes introduce stress wave reflections, which can influence the calculated signals strongly. Reflections from bulbs yield an almost similar wave equation result as a local stiff soil layer and a necking similarly results as a local soft layer. When pile discontinuities are not properly taken into account, either a proper match cannot be obtained or the capacity prediction will not be reliable. Soil properties can be confused with pile discontinuities.
For DLT on cast in situ piles a drop hammer with a guiding system has to be mobilized. The required ram mass is as rule of thumb 2% of the maximum load that has to be applied. A crane is required to move the drop hammer over the building site. The pile head has to be prepared to prevent damage from impact loading. An epoxy or grout cement is used to smooth the pile head surface to prevent stress concentrations during impact loading. The location of the transducers has to be at least 2 pile diameters from the pile head. When the pile head is located at ground level this requires an extension of the pile head for a similar length or the excavation of the pile head. For small capacity piles multiple piles can be tested in one day. For loads above lOMN the testing rate is normally in the range of two piles per day. For STN a loading device with a reaction mass catching system has to be mobilized. The required reaction mass is as rule of thumb 5% of the maximum load that has to be applied. Local available material can be used as reaction mass to reduce transport costs. A crane or a crawler system is required to move the STN device over the building site. For loads up to 4MN a STN device with a hydraulic catch mechanism can be applied. For higher loads STN requires a gravel catch system. Testing can take between 0.5 and 2 days per pile depending on the pile capacity. However for piles with a capacity less than 4 MN, a loading device with hydraulic catch mechanism can be applied and the number of piles tested in one day are in the same range as with DLT. STN can be even more efficient when the loading device and hydraulic catch mechanism are placed on crawlers. An epoxy or grout cement is used to smoothen the pile head surface to prevent stress concentrations during push loading. 2.6 Chance of pile damage
2.4 Reliability for testing on cast in situ piles With DLT the load on the pile head is introduced by an impacting ram. When the ram is not properly guided and hits the pile in an eccentric way, bending stresses will occur and result in excessive compression and/or tension stresses which can damage the pile. Most cast in situ piles need considerably more displacement to mobilize the ultimate capacity than driven piles. This softer response will easily gener-
Because of the many unknowns that have to be solved to perform a proper DLT signal matching analysis on cast in situ piles, there is considerable chance of errors in pile capacity predictions. The load measurement for STN is similar as for static load testing and unknown pile properties of cast in situ piles will not influence the load meas629
For STN the load displacement behavior is calculated in most cases by the Unloading Point Method, however signal matching techniques are also applied. 3.1 Accuracy With DLT on pre-cast driven piles, the load in the pile is measured by strain transducers mounted on the pile shaft. Pre-cast piles are considered to be of homogenous material and with the method described in section 2.2 and based on the determination of the stress wave the E-modulus can be determined accurately. The toe reflection will be visible at several stages of driving and the stress wave velocity can be determined easily. Only when the pile head is heavily reinforced will the E-modulus at the pile head be different from the E-modulus calculated by the stress wave velocity. With STN the load is accurately measured by a calibrated load cell placed on the pile head. The measured load is not dependent on the pile properties. The load measurement error is less than 0.1% of the maximum range of the load cell. Fig 9. Dynamic load test on a pre-cast pile 3.2 Reliability ate tension waves. Cast in situ piles are not designed to withstand high tension stresses. As soon as allowable tension stress levels are reached the impact energy has to be reduced to prevent pile damage. As a result, DLT has to be stopped at a stage where full capacity has not yet been mobilized. With STN the duration of the loading is long enough to keep the pile is under constant compression and tension stresses will not occur. To prevent bending stresses the piston of the statnamic device is installed exactly on or near the center of the pile head cross section. The launching of the reaction mass, and the resulting push load starts from the center of the pile. 3 THE APPLICATION OF DLT AND STN ON PRECAST DRIVEN PILES For pre-cast driven piles both DLT and STN are performed after a setup period after pile installation. This allows the soil to recover from driving induced disturbances like pore water pressure. In most cases the soil will regain strength during the setup period. For DLT strain and acceleration transducers are mounted on the pile shaft near the pile head. The load displacement behavior is calculated by signal matching. 630
The capacity of driven piles is mobilized at relative small displacements. Both DLT and STN are performed after a set up period. For DLT the pile load displacement behavior is calculated by a signal matching technique (CAPWAPTM,TNOWAVE) in most cases. For STN the pile load displacement behavior is determined by a direct method, the Unloading Point Method (UPM) and in some cases by signal matching. 3.3 Economy DLT has the advantage that the pile driving hammer used for pile installation can also be used for redriving the piles after a set-up period. However, when the pile driving hammer has to be used for constant production, an additional pile driving hammer or drop hammer has to be mobilized. When the mobilization of the full pile capacity is requested, the production hammer might not be sufficient to mobilize pile capacity after the set up period and an additional heavier hammer has to be mobilized. For STN the same economical conditions are applicable as mentioned in paragraph 2.5 (Ginneken, van G.J.J., 2000)
Table 1. Preferences for DLT or STN with respect to economy for driven piles Driven piles Preferred Soil set up I DLT I STN dium medium to high
**
*****
STN
4 CONCLUSIONS For bored concrete piles, auger piles and caissons the dynamic load testing method has some disadvantages and is less suitable and statnamic load testing is the preferred method. The most important reasons for the preference of statnamic load testing in the case of cast in situ piles are:
1. Accuracy in load measurement STN is not dependent on pile material and cross section properties 2. No influence of cross sectional variations STN results are not influenced by cross sectional variations over the pile length 3. No tension during compressive testing STN long duration loading will keep pile under constant compressive pressure 4. Concentric loading Easy placement of STN loading device in center of the pile 5. Pile and soil response closer to static With STN the pile moves as one unit, similar to static load tests. Stress wave phenomena can be neglected resulting in a simple method of analysis
3.4 Mobilization of capacity Set up phenomena can increase the soil resistance considerably. The pile driving hammer used for pile installation might not be able to mobilize the full pile capacity in such a case. Another reason that capacity can not be mobilized with DLT is that the load cannot be increased because compression or tension stresses becoming too high. To mobilize the pile capacity a STN device will be sent to the building site with at least a corresponding loading capacity. Only when the piles are over-designed will the full bearing capacity not be mobilized.
For driven piles both DLT and STN methods can be applied reliably and each has its advantages and disadvantages. A big economic advantage for DLT can be the use of the production rig for testing. A big advantage for STN is the fact that maximum available energy can be used to mobilize capacity and that that testing does not have to be stopped when tensional stresses become too high like with DLT.
3.5 Chance of pile damage For DLT there are some cases with a chance of pile damage. In the case of low friction and a soft toe response tension waves will be generated during DLT. When the maximum allowable tension stresses are reached the load on the pile cannot be increased because this will generate higher tension stresses and the pile will experience damage. In the case of a pile with a hard toe response, for example pile toe on rock, the compressive stresses at the pile toe can theoretically be two times higher than the maximum compression stress at the pile head. This is caused by the superposition of compression stress waves at the pile toe. So, if during DLT the compression stresses at the pile head are higher than half the compressive strength of the pile material, collapse of the pile material at the pile toe will occur. In this case piles can only be tested up to half the compressive strength of the pile material, which may not correspond with the capacity of the pile.
5 REFERENCES Abbasi, A.F., Al-Tayyib, 1990. Effect of hot weather on pulse velocity and modulus of elasticity of concrete. Materials and Structures, 1990, 23, ~~334-340 Ginneken, van G.J.J., 2000, Introduction Statnamic Load Testing in Europe, Case Studies in the Netherlands. Proceedings of the Sixth International Conference on the Application of Stress-wave Theory to Piles, Sao Paulo. Holeyman, A.E., 1992. Keynote Lecture: Technology of Pile Dynamic Testing. Proceedings of the Fourth International Conference on the Application of Stress-wave Theory to Piles, The Hague, F.B.J. Barends, Editor, A.A. Balkema Publishers, ppl95215.
For STN the pile is kept under constant compression and tension waves are suppressed. Superposition of compression waves at the pile will not occur. As with SLT piles can be tested near to the compressive strength of the shaft.
Franklin, R.E., King, T.M.J. 1971, Relations between compressive and indirect-tensile strength of
631
concrete, Road Research Laboratory, RRL Report LR 412 Madan B. Karkee, Takashi Horiguchi, Hideaki Kishida. Static and Dynamic Tests for Evaluation of the Vertical Load Bearing Capacity of Piles. , 22nd DFI Annual Member's Conference, Toronto, Canada, 1997, pp199-214 Sparks, P.R., Menzies, J.B., 1973. The effect of rate of loading upon the static and fatique strength of plain concrete in compression. Magazine of Concrete Research, Vol25/ 1973, No. 83, pp 73-80.
632
10 Case histories, pile set-up and correlations between test methods Prediction reliability
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Keynote lecture: Improving the reliability of pile bearing capacity prediction by the dynamic increasing energy test (DIET) Nelson Aoki University of SLio Paulo, SLio Curlos, Brazil
ABSTRACT: The dynamic increasing energy test (DIET) allows the measurement of the dynamic resistance displacement curve and the prediction of the static resistance - displacement curve of the pile-soil system. The variations of energy and work done are evaluated by Hamilton’s principle of energy conservation, and this evaluation allows the prediction of the dynamic failure load of the system. The usefulness of this approach is confirmed by the good correlation between the results of the DIET and of the static load test. static approaches and the definition of the ultimate resistance, are of foremost importance. This paper will deals with: (a) the ultimate resistance definition in the conventional static load test and in the constant energy dynamic test, accounting for the energy in the system; (b) the Hamilton’s principle of energy conservation (Clough & Penzien 1975); (c) the analysis of the dynamic resistance displacement curve of the dynamic increasing energy test and (d) some case histories. The static pile bearing capacity, in the traditional dynamic loading test, is predicted from the analysis of one blow, among several blows of same energy. This approach can be called a constant energy test. In that case, depending on the magnitude of the observed permanent displacement (set), the Smith’s model static resistance (Smith 1960) could satisfy either the lower or the upper bound theorems of the limit analysis (Chen & Liu 1990). The dynamic increasing energy test (DIET) involves the measurement of the total resistance to penetration or simply dynamic resistance, the displacement and the kinetic energy applied to the system at increasing energy hammer impacts (Aoki 1989b). The result is a dynamic resistance - displacement curve, from which the corresponding static resistance - displacement curve (Smith’s model) is generated. The dynamic resistance - displacement curve is then examined and the variations of energy and work done are analyzed by using the Hamilton‘s principle of energy conservation. Under the action of the failure load the system displaces as a rigid body. Considering the similarity of the failure under static and dynamic loading conditions, since both events involve energy and work done variations, it is possible to evaluate the ultimate
1 INTRODUCTION The main purpose of a dynamic pile test is to predict the static response and it is essential that any transfer model can be applied to both static and dynamic loading conditions (Randolph & Deeks 1992). This is fimdamental for reliable predictions of behavior under static and dynamic conditions (Aoki 1989a). The evaluation of the static resistance by the application of the stress wave theory to one blow of hammer is questionable and some discussion points are: (a) the choice of one blow among several impacts of nearly the same energy; (b) the effects of the elapsed time, from the end of driving to the start of the test; (c) the magnitude of deformations around the pile toe and shaft; (d) the comparison of the dynamic penetration with the static movement; (e) the methodology of the static load test and the evaluation of the pile bearing capacity (Fujita & Kusakabe 1988). The need for the standardization of the pile - soil interaction model, the establishment of soil parameters, the statistical treatment of discrepancies between the model and reality and derivation of design values for the load bearing capacity, which can be used within the modern Ioad and resistance partial factors design approach, are essential for the total foundation quality assurance (Vrouwenvelder & Dieterman 1992). The average resistance and the standard deviation allow verifying the actual characteristic load of the piling and the quality assurance depends on how and who takes the decision to stop the driving in the field. In contrast, the safety in the allowable load philosophy is simply a matter of personal past experience. In any case, the reliability of the pile bearing capacity prediction, by dynamic or 635
bearing capacity with the dynamic increasing energy test (Aoki 1997). The DIET has been largely used in Brazil as an alternative to the traditional static load test and the constant energy dynamic test (Aoki & Mello 1992, Beim & Aoki 1996). The static and dynamic results compare reasonably well (Aoki 1989c, Niyama & Aoki 1991). The DIET can also be used to analyze time effects on pile behavior “set up” after short time intervals. 2 RESISTANCE - DISPLACEMENT CURVE AND THE ULTIMATE PILE RESISTANCE The goal of a loading test is the determination of the ultimate resistance or bearing capacity of an isolated foundation pile. The isolated foundation pile is a system constituted by the structural pile element and the soil layers along the shaft and under the pile toe. This system’s response to the action of an axial load applied to the pile head is the displacement of this point, which is a complex function of many independent variables. Among them, the most important are: (a) time history of the applied force; (b) geometry of the pile and the soil layers; (c) physical properties of the pile and the soils; (d) peculiarities of pile installation that change material’s properties; and (e) elapsed time after the pile installation. During the static loading the applied force and the vertical movement are measured at each loading stage. In this case the applied force is equal to the mobilized resistance and the response of the pile soil system to the static loading is the well-known load - settlement curve. The slow maintained loading (SML), the quick maintained loading (QML), the constant rate of penetration (CRP) and the cyclic loading are examples of static test procedures (Fellenius 1980, Whitaker & Cooke 1961). Nowadays, the definition of the static pile bearing capacity or static ultimate resistance, which is the maximum reaction capacity of the pile - soil system, is a matter of discussion in the pile foundation engineering practice. Nevertheless, it is recognized that the shape of the load - settlement curve should be included in this definition (Reese 1972, Fellenius 1980) and this can be done by the consideration of the deformation energy of the pile - soil system. For a given curve there are many interpretations, each one leading to a particular ultimate resistance. The absolute or relative displacement values and the relationship between the elastic and permanent components of the displacement after the unloading or the time dependent behavior are some of the criteria for the establishment of the static bearing capacity (Vesic’ 1975, de Beer 1988). The absolute displacement equal to 10% of the pile dimension is a typical definition of the ultimate resistance (Terzaghi 1942).
636
The upper bound and lower bound theorems are used to determine the static failure load from energy transformations and equilibrium of forces considerations in the theoretical soil mechanics limit analysis (Chen & Liu 1990). By the upper bound theorem, the load is not less than that of the actual failure load, determined by equating the external rate of work to the internal rate of energy dissipation in an assumed velocity field which satisfies: (a) the velocity boundary conditions and (b) the strain and velocity compatibility conditions. Note that the stress distribution in the system does not need to be in equilibrium. Under this limit load the system loses its capability to store more elastic recoverable energy of deformation and starts to displace as a rigid body. By the lower bound theorem, the stress field must satisfy the stress boundary conditions and must nowhere violate the materials yield condition. The most popular dynamic loading test is the impact of a hammer on the cushion - pile - soil system. In this case the pile bearing capacity can be predicted by the old pile dynamic formulae or by the application of the stress wave theory to the hammer impact (Whitaker & Bullen 1981, Goble et al. 1975). The statnamic test has been developed as an alternative dynamic test (Middendorp 1993). During the hammer impact the time histories of deformation and acceleration are recorded at the instrumented section and the velocities and displacements are calculated by integrating the measured accelerations. The force history is obtained from the deformation and the applied kinetic energy is evaluated by integrating the force and the displacement. The dynamic resistance or total resistance to the pile penetration is evaluated from the analysis of forces and velocities. The correspondent static resistance is evaluated by using the classical Smith‘s model (Smith 1960) or the derived model of CASE and CAPWAP methods (Goble et al. 1981, Rausche et al. 1985, Goble & Likins 1996). The ultimate resistance from one single hammer impact is defined by the Davisson‘s limit load criterion (Davisson 1972). By this criterion the ultimate resistance is defined by the elastic toe displacement equal to (0.15 + B/120) inches, where B is the toe dimension. Finally it can not be forgotten that the soil resistance changes with the time and neither the static nor the dynamic loading can predict the future behavior of the pile - soil system, from the measured values at the time of the test. It can be concluded that the definition of the ultimate resistance is not clear in both static and dynamic loading tests. 2.1 Static force - displacement curve response. The loading is called static when the force (Q) is reached in stages of infinitesimal increments (dQ) and each stage is of infinite time duration. Thus, the
frequency of static loading is equal to zero. However, in practice, there is no static loading once the increment of force (AQ) and the time duration (At) have finite values. Figure 1 presents the static force - displacement curve, resulting from the force increment (AQ) and displacement increment (Ap) and the consequent increment of the deformation energy (AV).
if the zero load condition is reached in the inverse path described for the static loading. In the actual example of a building construction, the unloading path does not exist while the load-unloading cycle is observed only for the live load. The potential energy of deformation (V) stored in the system after the static loading can be elastic recoverable or permanent, depending on the rheological characteristics of the materials. This fact can be only known after the system unloading, when the potential energy of deformation (V) is transformed into recovered elastic energy (V,) plus the work done in the system (W,,) in such a way that
v = v, w,,
(5)
f
Figure 1. Static force - displacement curve.
where V = stored potential energy of deformation, V, = recovered elastic energy, and W,, = work done by the non-conservative forces. The reliability of the ultimate resistance predicted from the static load - settlement or the resistance displacement curves is discussed thereafter.
The rate of internal energy of deformation corresponding to the force increment (AQ) is the power
2.2 Shape of the static load - settlement curve.
g=AV I At (1) where 9= power or rate of the internal energy of deformation increment, AV = increment of energy deformation, and At = increment of time loading. The average power of the actual static or dynamic loading is variable from some watts to various kW. Under the action of the static load (Q), the displacement of the pile head (point D) is expressed by
P = Pe + Pp (2) where p = pile head displacement, pe = elastic displacement component, and pp = permanent displacement component. The nature of the deformations can not be known at the end of the loading stage. The potential energy of deformation (V) at the end of the loading is the hatched area ODFO. The complementary energy (Langhaar 1962) can be calculated by the difference
0
Vc= Q . P - V (3) where V, = complementary energy (area OGDO), Q = applied force, p = total displacement of the pile head, and V= potential energy of deformation. The static equilibrium of forces is expressed by R=Q
Figure 2 presents the typical static load -settlement curve in the case where the ultimate resistance is well defined by the final shape of the curve. This curve correlates the settlement (p) to the load (Q) where it is implicit that the static reaction (R) is always equal to the static load (Q). The failure is characterized when the pile head displaces from point D to point A under the action of the constant limit load (Q,), equal to the ultimate system resistance (Ru). QU
Figure 2. Load - Settlement curve: defined ultimate resistance.
When the limit load (Q = Q,) corresponding to the point D of the loading curve is applied, the limit total displacement is (p = pu), the limit potential energy of deformation is (V = V,), and the limit complementary energy is equal to (V, = V,,). In the case where all materials are perfectly rigidplastic, the unloading path is the horizontal line DF. In that case the deformation energy is completely transformed into work done by the resistant forces
(4)
where R = total mobilized resistant forces (resistance or reaction), and Q = applied force (load or action). According to expression (4), the system response under static loading can be indifferently expressed by the relationship between the force - displacement curve or the resistance - displacement curve. The system response to unloading is called static 637
(V = Wn,). There is no elastic recovery (Pe = 0) and the total settlement OB is permanent (p = pp). For perfectly elastic materials the loading path is the curve OD and the unloading path is DO resulting in a minimum work done (Wnc= 0) and (V = Ve). In actual cases some elastic and plastic deformations take place and the unloading path is DE. The recovered elastic energy (Ve) is represented by the area DFED and the area ODE0 is the work done (W,,) in the system. For any point on the vertical asymptote line between points D and A, all the additional energy of deformation is fully transformed into work done, if the elastic recovery energy becomes constant after point D (DE line being parallel to AC line). It is not possible to know beforehand if this defined failure load is equal to the upper or to the lower bound of the failure load, unless more data concerning the external rate of work and the internal rate of energy dissipation could be known. Figure 3 presents the typical shape of loadsettlement curve corresponding to the case where the resistance reaches a maximum equal to Q m a x at point F and thereafter it drops up to a constant residual ultimate resistance, after point D. The limit load is again the well defined value Qu on the vertical asymptote DA to the load-settlement curve.
where Q = applied force, Q, = physical ultimate resistance, and p = settlement. Expression (6) can also be written lim Q = Q ,
(7)
V+m
where V potential energy of deformation. In this case the load - settlement curve can be described by the well-known expression (Veen 1953) =I
where Q = applied force, Qu= physical ultimate resistance, a = curve shape coefficient, and p = settlement. This expression can also be rewritten as
(9)
The potential energy of deformation is equal to (10) J
L
where V = stored potential energy of deformation, Qu= physical ultimate resistance, p = settlement, and a = curve shape coefficient.
0
Q"
0
--P
load
+-'
)
-s5 2
C
ID
-----
B
D
I
Figure 3. Load - settlement curve: defined residual ultimate resistance.
Figure 4. Load - settlement curve: physical ultimate resistance.
This shape of curve is observed in the cases of floating piles where, after the peak of resistance (Qmax) the system resistance drops up to the defined residual ultimate resistance (QL,).The unloading path is dependent on the pile-soil system characteristics as explained in the case of figure 2. Figure 4 presents the case where the ultimate resistance is not defined. The system displacement increases with increasing loading and the vertical asymptote to the curve appears to be at the infinity. In this particular case the ultimate resistance can be taken as the physical ultimate resistance defined by the condition
Considering expression (3), the complementary energy can be estimated by the expression
where V, = complementary energy of deformation, Q, = physical ultimate resistance, p = settlement, and CI = curve shape coefficient. The limit complementary energy corresponding to the physical ultimate resistance would be lim V,= V,,
(12)
P+W
lim Q = Q,
where V,,
P+a
638
= limit
complementary energy.
When the curve is expressed by Veen's expression (8), the limit complementary energy becomes constant and its value is expressed by
v,,
= -Q
the structural pile material strength. This is expressed by the following condition: (14)
QSQu
"
o!
Therefore the physical ultimate resistance can be defined by the limit complementary energy of expression (13). This criteria has the advantage of involving both the curve shape and the ultimate resistance in the definition of the failure behavior of the pile - soil system (Aoki 1997). Figure 5 presents the case where the OD path of the curve is similar to that presented in figures 1 to 4. The special feature in this case is that the settlement increases linearly with increasing loading up to the maximum load test value Qmax,without the ultimate resistance characterization.
where Q = applied force, Q, = ultimate soil resistance, and R, = structural pile strength. In some cases, where the structural pile strength is reached before the ultimate soil resistance has been mobilized, the system fails by the weaker structural link and the following condition prevails: Q u > Re
(15)
where Q, = ultimate soil resistance, and Re = structural pile strength. This is the case where the pile toe is placed either on very hard and rigid soil material or in very long floating piles. The analysis of the limit maximum elastic potential energy of deformation of the pile - soil system requires the prediction of the maximum elastic displacement of the pile head. The elastic displacement component can be predicted by the expression (Vesic' 1975)
where pe = elastic displacement of the pile head, pep = displacement component due to the elastic shortening of the structural pile shaft, and pes= displacement component due to the elastic deformations of the soil layers under the pile toe. The elastic displacement component due to the elastic deformations of soil layers under the pile toe (elastic pile toe displacement) is equal to
A
Figure 5 . Load - settlement curve: structural ultimate resistance
From point D up to point A, the measured displacement increases linearly under increasing loading level and the corresponding potential energy of deformation increases linearly with loading. This is the case where all the materials of the pile - soil system present strain hardening behavior. As explained previously the unloading path is dependent on the rheological properties of pile and soil materials. In this case both deformation energy (V) and complementary energy (V,) increase linearly with the increasing applied load up to a limit load corresponding to the ultimate resistance of the weakest link in the pile - soil system. In this analysis the following variables must be considered: (a) the ultimate soil resistance, (b) the structural pile strength and, (c) the limit displacement from where the maximum elastic potential energy of deformation is reached and the system starts to behave as a rigid body. Although it is not remarked, in all the cases presented in figures 2 to 4, the applied force (Q) was previously assumed to be always smaller than the soil material ultimate resistance and, on the other hand, the soil ultimate resistance was smaller than
Pes = Pes,p + Pes.f
where pes.p = component due to the action of the load transferred to the soil by the pile toe, and pes.f= component due to the action of the load transferred to the soil by the pile shaft. When the load transferred to the soil by the pile shaft is equal to zero, the elastic displacement of the pile head corresponding to the compression axial load (Q) can be predicted by e , = Q
~
-------+*
~B
i s p]
(18)
where pe = elastic pile head displacement, Q = applied force, L = length of pile; E, = elastic modulus of the pile material, B = diameter of pile toe, and E, = representative elastic modulus of the soil layers under pile toe. Considering expression (1 4) for materials of perfectly elastic - plastic behavior, the limit elastic pile head displacement corresponding to the load equaI to the pile toe ultimate soil resistance will be: 639
where Peu,s = limit elastic pile displacement, and Qu = pile toe ultimate soil resistance. Considering expression (1 5 ) for materials of perfectly elastic - plastic behavior, the limit elastic pile head displacement just before the structural failure of the pile will be: /
until point D, where the limit elastic displacement (Peu,s) is reached under the action of the load (Q,) corresponding to the ultimate resistance of the system. Under the action of this constant load, the pile head displaces from point D to any point A on the vertical line. Observe that, in this loading path, all the deformations are permanent. The potential energy of deformation (V) stored in the system at point A is expressed by
\
where V = potential energy of deformation, Q, = ideal ultimate resistance, p = total displacement, plv = permanent displacement, and Peu,s = limit elastic pile displacement. The unloading path is line AC. The limit recovered elastic energy of deformation is expressed by
where Peu,p = limit elastic pile displacement, just before the structural failure of the pile. The limit elastic energy of deformation, when the ultimate soil resistance is smaller than the structural pile strength, can be evaluated by
where Ves,m, = limit elastic energy of deformation, when the soil is the weakest link in the pile - soil system. The limit elastic energy of deformation, when the structural pile strength is smaller than the ultimate soil resistance, can be evaluated by
where V,, = limit recovered elastic energy (area ABCA). The work done in the system is expressed by
where W,, = work done in the system (area ODACO). It could be interpreted that any plastic deformation in the linear loading path OD was 'concentrated' in the final DA path of the curve. The soil response is not always instantaneous and it takes a more or less variable time, to fully mobilize all the resistance in the system. Most soils change their properties after the execution by freezing (soil setup) or relaxing (soil relaxation). In this case it is important to observe that the load-settlement curve represents the system behavior only for the period of time when the test is under progress. The comparison between two pile load tests is reliable if their results represent the behavior under the same time test conditions. The elapsed time is an independent variable that can affect the results of both static and dynamic loading tests.
where Vee.m, = maximum elastic energy of deformation, when the pile is the weakest link in the pile soil system. The elastic energy of deformation of expression (21) is a component of the limit potential energy of deformation (V,) described in figure 2. Figure 6 presents the idealized load -settlement curve in the case of a pile - soil system constituted by elastic perfectly plastic materials. This model is currently used to describe the actual curves presented in figures 2 to 5. QU
load
2.3 Shape of the dynamic resistance - displacement curve. The loading is called dynamic when the applied force (Q) varies with time (Clough & Penzien 1975). The impact of a hammer on the pile-soil system is a typical example of a dynamic loading. By analogy, with the resistance to the penetration of piles under static load, it can be assumed that the dynamic resistance - penetration curve is similar to the traditional static load - settlement curve represented in figures 2 and 3. The resistance against
Figure 6. Idealized load - settlement curve, for elastic perfectly plastic materials.
In this case the ultimate resistance is equal to the load (Q,) in the vertical line DA of the ideal curve. Along the loading path OD the system response is linear and the energy of deformation is reversible 640
penetration under a single blow of the hammer changes with increasing penetration. However the expression ’dynamic resistance against penetration’ has no definite meaning, unless this term is applied to the final resistance represented by the abscissa (QJ of the vertical asymptote to the penetration curve (Terzaghi 1943). This observation shows the importance of the shape of the resistance - displacement curve and its link with the ultimate resistance in the static loading test. Terzaghi’s final resistance (QJ of the vertical asymptote to the penetration curve is the physical ultimate resistance of expression (6). According to Terzaghi, the question is: is the kinetic energy of the hammer blow enough to mobilize the ultimate resistance of pile- soil system? Figure 7 presents the idealized dynamic toe resistance - pile toe displacement curve corresponding to the idealized load - settlement curve for elastic perfectly plastic materials of figure (6).
o
toe resistance
Rp
a,
E cd
_1
4 .-U a, 0 +*
2 c .a
A
Figure 7. Idealized dynamic toe resistance - pile toe displacement curve for elastic perfectly plastic materials. This ideal curve (Smith 1960) correlates the dynamic toe displacement (Dp) and the total toe resistance (Rpt).The total resistance is usually considered to have two components: the ’static’ one, which is dependent on the displacement and the ’dynamic’ one, which is dependent on the velocity. For the sake of simplicity the total dynamic toe resistance will be referred simply as dynamic toe resistance. In this case the dynamic toe resistance is related to the static toe resistance by the expression
1
dynamic resistance
:
RI
> ? \ ? \ ?
Figure 8. Dynamic resistance - pile head displacement curve, in the case of a single hammer impact. The shape of the corresponding static loading curve is under discussion as pointed out by Terzaghi (1943). At the end of the dynamic loading (point A of the curve) the maximum pile head displacement is (D) and the mobilised dynamic resistance is (RJ.
n
c R~~ [I + J m - ~ ( m , t ) ]
n
R = R, + E R f m
I\,
Rpt = Rp (1 + Js . Vp,rnax) (26) where Rpt= dynamic toe resistance, R, = maximum static toe resistance, J, = Smith‘s toe damping, and v , , =~ maximum ~ velocity at the pile toe. The Smith’s model for the resistance at a given section (m) of the pile shaft is similar to figure 7 and the difference comes from the fact that the shaft can displace downwards and upwards. The total (dynamic + static) shaft resistance or simply dynamic shaft resistance is equal to ft =
where Rt = dynamic resistance, R,, = dynamic toe resistance, and Rf, = dynamic shaft resistance. From expressions (26), (27) and (28) it is possible to conclude that the corresponding static resistance is
where R = static resistance (Smith‘s model), R, = static toe resistance component, and Rfm = static shaft resistance component. Besides the simplified rheological behavior, a fundamental feature of Smith’s model is that the displacement under static loading is assumed to be equal to the displacement under dynamic loading. In the case of a single hammer impact, the loading time history starts at time (tl) when the force begins to be applied on the pile head. When the pile head displacement is maximum, the maximum kinetic energy (T) is the well-known ENTHRU value (Goble et al. 1975). The resulting total dynamic resistance (static + dynamic) or simply dynamic resistance - pile head displacement curve is presented in figure 8.
Rpt
4-
B
where R,-, = dynamic shaft resistance, n = number of sections, Rf ,m = maximum local shaft static resistance, J, = local Smith’s damping factor, and v(m,t) = velocity in the shaft section (m) at time (t). The total dynamic resistance (static + dynamic) or simply dynamic resistance is not mobilized at the same time at all pile sections and is computed by the following expression:
(27 1
1
64 1
At this point the resulting kinetic energy (T) was stored as potential energy of deformation (V) in the pile - soil system. The potential energy of deformation (V) and the maximum kinetic energy (T) are represented by the area OABO. The measurements of the maxiqum total (dynamic + static) resistance (RJ, the maximum dynamic displacement (D) and the maximum kinetic energy (T) can be done, for example, by monitoring the hammer impact with PDATMsystem. The dynamic loading cycle is unique and it is not possible to separate the loading and unloading stages as in the static loading case. The applied forces and the corresponding mobilized resistances vary during the impact time and only the maximum values at each pile section are of importance. The dynamic unloading path starts soon after the maximum displacement (D) is reached and ends at the time (t2) when the final pile penetration is equal to the permanent displacement or set (S). This unloading path is represented by curve AC in figure 8. The area ABCA is the potential recoverable elastic energy of deformation (V,) and the area OACO is the work done (W,,,) by the non-conservative forces. The maximum pile head displacement under dynamic loading measured by either the old cardboard and pencil system or the double integration of measured acceleration is equal to D=K+S
(30)
where D = maximum pile head displacement under dynamic loading, K = rebound or elastic displacement, and S = set or permanent displacement. The elastic displacement component is equal to
where C2 = elastic shortening of the pile, and C3 = elastic displacement due to deformations of the soil layers under the pile toe (pile toe quake). In figure 8 the maximum displacement (D) is equal to segment OB, the elastic displacement (K) is equal to segment BC and the set (S) is equal to segment OC. The set (S) is usually measured by the pencil and cardboard technique. Although fundamental, there is not much interest in measuring the actual dynamic resistance - pile head displacement curve from one hammer impact. The ideal shape indicated in figure 7 for the toe displacement is based on the ideal static response of figure 6. The actual static curves of figures 2 to 5 are being simulated by the ideal curve of figure 6. The CAPWAPB analysis simulates the shape of the static load - settlement curve based on the stress wave theory applied to one single hammer impact measurements, by zeroing the velocity in figure 8. In this curve the toe ultimate resistance is defined by Davisson's limit load criteria. 642
Despite such difficulties, the reliability of the pile bearing capacity prediction can be improved by the dynamic increasing energy testing (DIET), based on Hamilton's principle of energy conservation which will be discussed thereafter.
3 HAMILTON'S PRINCIPLE OF ENERGY CONSERVATION. The ultimate pile resistance has been predicted by the dynamic pile driving formulas and by the application of the stress wave theory to model the hammer impact. In both cases the repeatability of the displacements under impacts of same energy shows that the dynamic loading is a reliable procedure. The fundamental problem to evaluate the energy lost in the pile driving system was clarified when the ENTHRU concept was introduced in practice (Goble et al.1975). From this time, the actual kinetic energy available in the pile - soil system has been currently measured during the pile driving, allowing the measurement of the hammer impact efficiency. Apart from this use, not much has been made to predict the static resistance from the direct energy measurement (Broms & Choo 1988). This keynote lecture brings back the old Hamilton's principle of energy conservation that has not been yet mentioned in the traditional pile dynamic literature (Aoki 1997). It is a valuable tool to better understand the energy transformations in the pile soil system, subjected to the action of a dynamic loading, in the case of the ultimate resistance analysis. Hamilton's principle of energy conservation states that the variation of the kinetic and potential energy, plus the variation of the work done by the non-conservative forces acting in any material system, including the work done by damping and any arbitrary external forces at any time interval tl to t2, is equal to zero. In variational form, Hamilton's principle of energy conservation is expressed by ft':
S (T - V ) dt
+ S,:S
( Wnc ) dt = 0
(32)
where T = system total kinetic energy, V = system potential energy, including the energy of deformation and the potential energy of any external conservative forces, W,, = work done by the nonconservative forces acting in the time interval, including the damping forces and any arbitrary external loading, and 6 = variation during the time interval (t2 - tl). According to Hamilton's principle, during the dynamic loading corresponding to one impact of the hammer, it follows that: (a) at the time (t = tl), the kinetic energy applied to the pile - soil system is equal to (T); (b) at the time when the displacement reaches the maximum value (D), all kinetic energy is
transformed into potential energy of deformation (V=T); and (c) at the time (t = t2) after the unloading, this energy of deformation (V) is partially recovered as elastic energy of deformation (Ve) and partially transformed into work done (Wnc)by the non - conservative forces acting in the pile - soil system. In the case of a static loading, the kinetic energy T is equal to zero and the Hamilton's principle of energy conservation is reduced to the well-known principle of minimum potential energy 6(-V) + 6(W",) = 0
(33)
In this case it is said that the variation of the potential energy plus the variation of the work done by the non-conservative static forces is equal to zero. It follows that, after the system unloading, the energy of deformation (V) is partially recovered as elastic energy of deformation (Ve) and partially transformed in work done (W,,) in the system, as expressed by equations ( 5 ) and (33).
The average soil data are presented in table 1. Table 1 : AveraEe soil laver characteristics. Layer depth Soil description 0 to 1.9 1.9 to 3.3 3.3 to 10.3
Brown medium silty clay fill, N,sI17.=5/30 cm. Brown medium organic clay, N,s/,7,= 5/30 cm. Pink and yellow sandy silt residual soil; veins of fine gravel, medium, Ns/.l'7.=10/30 cm. 10.3 to 16.1 Yellow and grey weathered sandy very dense rock, N.ypr= 43/30 cm Soil under the Very dense sand resulting of weathered 37/30 cm. pile toe gneiss, N.~/>.I.=
The test results are presented in table 2. Table 2: Dynamic increasing energy test results. D S T=V T=V H Rt (field) (PDA) (PDA) (field) (PDA) (curve) (m) (kN) (m) (In) (kJ) (kJ)
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.5 3.0
4 DYNAMIC INCREASING ENERGY TEST The static ultimate resistance can be obtained either in a single cycle of static loading or in a single hammer blow impact as discussed in chapter 2.2. The increasing energy impacts has been used to analyze the behavior of model piles (Koten at al. 1988, Bernardes 1989) or to define some parameters of the traditional pile driving formulas (Whitaker & Bullen 1981). In the test procedure called dynamic increasing energy test (DIET) the pile - soil system is subjected to the action of several increasing energy blows and each blow is instrumented and interpreted, for example, by the PDATMsystem. The difference is that the resulting set of dynamic resistances and displacements are analyzed as a whole by plotting each couple of data from the same origin, although they have been obtained in a cyclic loading (Aoki 1997).
850 1010 1140 1330 1530 1730 1930 2010 2060 2080 2110 2180
0.0035 0.0051 0.0054 0.0072 0.0082 0.0111 0.0132 0.0147 0.0158 0.0160 0.0178 0.0203
0 0 0 0 0 0 0.002 0.003 0.004 0.005 0.006 0.009
1.9 3.0 3.9 5.8 7.3 11.8 15.9 18.6 21.1 21.8 24.7 29.9
1.5 3.0 3.3 5.5 7.0 11.5 15.5 18.5 20.7 21.2 24.9 30.3
For the increasing hammer impacts of height (H), the total measured resistance (RJ, the maximum displacement (D) and the set (S) measured in the field according to the pencil and cardboard method (Chellis 1951) are presented in table 2. The last two columns present the kinetic energy T (PDA) measured by the PDA system and the kinetic energy T(curve) calculated from the dynamic resistance displacement curve using the single expression
4.1 Dynamic resistance - displacement curve. A typical example of dynamic resistance - displacement curve determination carried out with the PDATMsystem is presented and discussed in this item. This is the case of a pre-cast reinforced concrete pile of external diameter of 42 cm, wall thickness of 8 cm and 12 m long. The length below the sensors was 10.9 m and the driven length was 10.1 m. The average structural strength of the pile was 3000 kN. The age of the pile at the time of the test was 6 days. The impacts were applied with a 30 kN free-fall hammer. The shock absorbing system consisted of a steel helmet of 1 kN of weight, a hardwood cushion of 30 cm of height with 6 cm thick plywood cushion.
where T(curve) = kinetic energy calculated from the measured curve, n = number of points, Rt,(i)/ R t,(i-l) = dynamic resistances at blows (i) and (i-1), and D(i) / D(i-1)= displacements at blows (i) and (i-1). Figure 9 presents the dynamic resistance - displacement curve obtained in this dynamic increasing energy test. It is assumed that the initial conditions of any blow do not change with the increasing applied energy. This is the reason why the points of the curve are plotted from the same origin, although they have been obtained in a cyclic load test. The reliability of this procedure can be verified by the comparison of the last two columns of table 2, where the kinetic energy calculated from the curve
643
0 D (displacement)
S (permanent set)
A
K (elastic rebound)
Figure 9. Dynamic resistance - displacement curve plot is in good agreement with the direct measurement with the PDA system, within the measurements accuracy. If the initial condition of the pile-soil system has changed after each increasing loading stage, the measured potential energy of deformation would not fit the calculated value from the dynamic resistance - displacement curve. The maximum displacement at point A corresponding to the last applied blow of 3.0 m of height was 20.3 mm and the mobilized dynamic resistance was 2 180 kN. At this point the stored potential energy deformation was equal to 29.9 kJ. After unloading, the permanent displacement was 9.0 mm. The area ABCA is equal to the recovered elastic energy of deformation and the area OACO is equal to the work done by the non - conservative forces including the damping forces. According to the Hamilton’s principle of energy conservation, the following equation can be applied at the end of any impact of increasing energy: T = V = V,
-t
W,,
(35)
where V, = recoverable elastic energy of energy of deformation, and W,, = work done by the non - conservative forces. Assuming that the unloading path AC is linear, the elastic energy of deformation will be V,- R,(W2) where Rt = dynamic resistance, and K elastic displacement. In that case the work done is
(36) =
rebound or
W,CZR~(W~)-T (37) where Wn, = work done by non - conservative forces. According to the measurements the wave takes 16 millisecond after the impact to reach point A, resulting an average test power equal to 1868 kW. 644
Figure 9 also shows the Permanent (set) and elastic displacement (rebound) components of the displacement at each increasing energy impact. The last point where the set is equal to zero is point E on the curve. In the next impact (H = 1.4m) the system starts to present permanent deformations ( S O ) . It can be observed that, after this impact, the elastic rebound presents variations, but remains nearly constant while the set increases under increasing applied energy up to point A where S=9 mm. The occurrence of the first permanent deformation fulfills the statement of the lower bound theorem for the determination of the system static failure load. Therefore, it can be inferred that 1930 kN is the dynamic resistance that could be taken as the dynamic lower bound of the system failure load (H = 1.40m). At point F of the dynamic resistance - displacement curve (H = 1.8 m), the displacement be,’*ins to increase linearly with the applied energy. This change in the system response involves energy and work done variations under increasing loading, which will be analyzed under the light of Hamilton’s principle of energy conservation. The subsequent search for the ultimate resistance in the DIET procedure is a similar problem to the prediction of the failure load in the static loading. Considering that the energy and the work done under increasing energy rates can be measured, it is natural to analyze their variations in order to check the existence of a dynamic upper bound of the failure load similar to the case of the static loading. For static loading, the upper bound theorem determines the limit load by equating the rate of internal energy dissipation to the rate of the work done by the external forces in the system. This theorem states that the variation of the internal potential energy is equal to the variation of the work done by the external forces, if the rates of variation are referred to the same time interval. The load and the analysis are called “limit” because it is not possible to apply a load greater than this limit value. This limit is equal to the ultimate resistance of pile-soil system. In the dynamic increasing energy test it is possible to measure the variation of the internal potential energy and the work done at any dynamic loading level. Similarly, the limit dynamic resistance can be determined by the upper bound theorem, as follows: (a) choose an energy of deformation (VU) and suppose that it fulfills the limit upper bound theorem; (b) calculate the variation of the energy of deformation for the impacts of energy greater than the value under analysis by the expression AV=V-Vu
( 3 8)
where AV = energy of deformation variation, V = current energy of deformation, and VU = limit energy of deformation under analysis.
(c) calculate the work done variation by AW Wnc - Wnc,u (39) where AW= work done variation correspondent to AV, Wnc= current work done, and Wnc.u= work done at the limit energy of deformation under analysis. The failure dynamic resistance is reached when the energy of deformation variation is equal to the work done variation at the same time interval. This condition can be expressed as AV - AW = 0
(40)
Table 3 presents the values of the potential energy of deformation, elastic recovered energy and the work done in the present example. Table 3. H (field) (m) 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.5 3.0
Elastic recovered e n e r m and work done. K Rt V=T Wnc Ve (PDA) (kN)
850 1010 1140 1330 1530 1730 1930 2010 2060 2080 2110 2180
(D-S)
(m) 0.0035 0.0051 0.0054 0.0072 0.0082 0.0111 0.0112 0.0117 0.0118 0.0110 0.0118 0.0113
(PDA) (kJ)
(PDA) (kJ)
(PDA) (kJ)
1,9 3,O 3,9 5,s 7,3 11,s 15,9 18,6 21,l 21,s 24,7 29,9
0.412 0.424 0.822 1.012 1.027 2.198 5.092 6.841 8.946 10.360 12.251 17.583
1.488 2.576 3.078 4.788 6.273 9.602 10.808 11.759 12.154 11.440 12.449 12.317
The corresponding upper bound limit dynamic resistance of system is equal to 2060 kN. Theoretically, the elastic rebound would be constant from point F, where the system becomes theoretically rigid. Nevertheless the average elastic rebound appears to become nearly constant before this point (see figure 9). This fact can be explained by the residual stresses stored in the system. The elastic deformations that take place between points E and F are not recovered as elastic rebounds after unloading and remain in the system as residual stresses, being incorporated into the permanent displacement S after each blow. After point F, the displacement increases linearly with increasing kinetic energy satisfying equation (40) condition. This is also true for any value exceeding the limit energy 21.1 kJ, as the case where the potential energy is equal to 24.7 kJ. Therefore, the application of Hamilton’s principle of energy conservation to the dynamic resistance displacement curve allows the prediction of the lower and upper bound of the dynamic limit resistance, from the dynamic resistance - displacement curve, in a similar way the limit analysis defines the static failure load by the lower and upper bound theorems. 4.2 Dynamic post -.failure behavior When the failure load is smaller than the structural pile strength, the applied kinetic energy can be greater than the limit energy corresponding to the upper bound of the ultimate resistance, configuring the post - failure behavior of the system (Aoki & Cintra 1997). In this case the kinetic energy (AT) in excess to value (T,) of the upper bound limit is equal to the excess of the work done (AW) by the forces. Observe that the maximum applied energy can not be greater than the limit elastic energy of deformation presented in equation (22). Table 4 shows the values of the excess of kinetic energy, the recovered elastic energy and the excess of work done by the non-conservative forces. considering that the limit dynamic resistance corresponds to the blow of height equal to 1.8 m.
Figure 10 presents the results of this research in graphical form. The difference between AV and AW, represented by the shaded area, diminishes until the value VU= 21.1 kJ has been reached. For this value the shaded area is equal to zero, fulfilling the dynamic upper bound condition.
Table 4a. Post dvnamic failure behavior Darameters V=T R S Wnc V, AT=AV PDA
PDA
field
(PDA)
(I’DA)
(kJ) 21.1 21.8 24.7 29.9
(W
(m) 0.004 0.005 0.006 0.009
(kJ) 8.946 10.360 12.251 17.583
(kJ) 12.154 11.440 12.449 12.317
1790 1800 1790 1810
(kJ) 0 0.7 3.6 8.8
AW (kJ) 0 1.41 3.30 8.64
Figure 11 presents the curve of the excess of energy and excess of work done, considering the limit values VU = 21.1 kJ and Wnc,u = 8.946 kJ, corresponding to the point F of the figure 9.
Figure 10. Research on the limit dynamic resistance
645
AW (kJ) 17
the pile toe. Moreover, it is considered that the maximum toe displacement under dynamic loading is equal to the toe displacement under static loading. Under these basic considerations the static load - top movement curve obtained for last 3.0 m height blow, by the conventional CAPWAPB analysis simulation, is presented in figure 12. Point G on the curve corresponds to the maximum simulated static movement. An alternative to this conventional CAPWAPB simulation is the direct transformation of the dynamic resistance - displacement curve into a static resistance - displacement curve. In this case the static resistance is predicted from the dynamic resistance by the application of expression (29) for each increasing energy impact of the DIET. Figure 13 presents the Smith's static resistance dynamic displacement curve corresponding to the dynamic resistance - displacement curve of figure 9.
AV (kJ)
Figure 1 1. Dynamic loading: post failure behavior.
As showed by figures 10 and 11, in the post failure the excess of energy is equal to the excess of work done in the system. From the practical point of view, this means that, after the dynamic failure load, all the excess of kinetic energy is totally transformed into excess of work done in the system.
4.3 Static resistance - movenzent curve p o m dynamic tests. The displacement of the pile head under static loading can be predicted by equations (2) and (16). It is remarkable that the elastic pile toe displacement expressed by equation (17) is dependent on the load transferred to the soil by the pile toe and the pile However the Smith's mddel toe (figure 7) is dependent only on the pile toe resistance, not taking into account the effect of the load transferred by the pile shaft to the soil layers under
Figure 13. Slnith's model static resistance - displacelne11t curve from (DIET) dynamic increasing energy test.
The static parameters calculated from the data of table 2 are presented in table 4. Table 4b. Static parameters resulting from - the DIET. H R D V, W, V,,, Vs,c (field) (PDA) (PDA) (PDA) (Chellis) (PDA) (PDA)
(m)
(kN) 810 950 1070 1240 1430 1610 1730 1770 1790 1800 1790 1810
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.5 3.0
Figure 12. CAPWAP'S static load - top movement curve for H= 3.0 in blow of the dynamic increasing energy test.
646
Ohs.:
'S
'
(m) 0.0035 0.0051 0.0054 0.0072 0.0082 0.0111 0.0132 0.0147 0.0158 0.0160 0.0178 0.0203
(kJ) 1.418 2.826 3.129 5.208 6.543 10.951 14.458 17.083 19.041 19.400 22.631 27.131
(kJ) 1.418 2.422 2.889 4.464 5.863 8.935 9.688 10.354 10.561 9.900 10.561 10.226
(kJ) 0.000 0.404 0.240 0.744 0.680 2.016 4.770 6.729 8.480 9.500 12.070 16.905
denotes parameters ullder static loading.
(kJ) 1.417 2.019 2.649 3.720 5.183 6.920 8.378 8.936 9.241 9.400 9.23 1 9.612
The analysis of figure 13 and the data in table 4 show that the static resistance R is practically constant from point D to point A of the curve. In this path the corresponding static resistance varies between 1790 and 1810 kN with an average value of 1800 kN. Considering that the shape of this curve fits well the static load - settlement curve of figure 2, it can be concluded that the static ultimate resistance was mobilized at point D and its value is nearly equal to 1800 kN. Table 4 shows that, for the hammer fall height of 1.8 m, the ultimate static resistance is R, = 1800 kN, the limit displacement is p,=15.8 mm, the limit potential energy is VS,, = 19.04 kJ and the limit complementary energy is the nearly constant value equal to VC,,z 9.4 kJ. Point D of figures 13 and 2 corresponds to point F of figure 9. The comparison between the CAPWAP’s static load - top movement curve and the Smith’s static resistance - displacement curve is presented in figure 12. Point G corresponds to the CAPWAP’s simulation of the maximum static movement of the pile head for the last 3 m height blow. Point A corresponds to the last point of the Smith’s static resistance - displacement curve from the DIET. The dynamic increasing energy test resulting from this example is summarized in table 5. Table 5. Dynamic increasing energy test (DIET) results Smith’s static Observation H Dynamic (m) resistance (kN) resistance (kN) 1.40 1930 1730 lower bound 1.80 2060 z 1 800 upper bound
In this example it can be concluded that: (a) the lower bound of the dynamic failure load is equal to 1930 kN; (b) the upper bound of the dynamic failure load is equal to 2060 kN; (c) the upper bound of the static failure load is equal to 1800 kN; and (d) the lower bound of the static failure load is equal to 1730 kN (static load of Smith’s model). Observe that the dynamic resistance continues to increase after the dynamic resistance upper bound has been reached, but the same is not true for the static collapse load. The comparison between the results obtained in the traditional static and dynamic loading tests procedures with the results of the dynamic increasing energy test is presented in table 6. This table do not intend to suggest that the conventional static load test is less reliable than the dynamic load test, either conventional or with increasing energy. It was presented only to show that the dynamic increasing energy test (DIET) provides more information on the behavior of the pile - soil system than the other conventional static and dynamic loading test procedures. The interpretation of the static load test is relatively simple, as opposed to the dynamic tests which 647
Table 6: Static and dynamic loading Measured response Static of the pile - soil loading system Shape of the curve available Energy & work done not available Failure lower bound ? failure upper bound ?
test procedures. Dynamic Dynamic constant increasing energy energy available available available available ? available ? available
require a very sophisticated interpretation. Such interpretation can only be correctly performed by highly qualified professionals. 5 CASE HISTORIES OF DIET APPLICATIONS. 5.1 Correlation between static (SML) tests and the dynamic increasing energy test (DIET) results. Brazilian experience has showed that the Smith’s static resistance - displacement curve predicted from the dynamic resistance - displacement curve measured in the dynamic increasing energy test and performed on prefabricated reinforced concrete piles has presented a good correlation with the load - settlement curve executed in the same pile, in the traditional slow maintained static (SML) load test. Figure 14 presents the correlation between the static resistance - displacement curve from DIET performed at the end of the driving, and the static SML load - settlement curve obtained 25 days later. In this example the concrete pile of 11.5m of length, diameter of 26 cm and 6 cm of wall thickness was driven with a hammer of 22.4 kN of weight, in 9 m thick soft clay layer overlying a very hard silty clay at the pile base, in the tertiary Taubat6 Formation in Sgo Paulo State. The maximum mobilized load was 600 kN in the static and 870 kN in the dynamic test.
Figure 14. Correlation DIET and SML test curves.
Figure 15 presents the correlation between the static resistance - displacement curve from DIET performed one day after the driving and the load settlement curve obtained in the static SML test peformed 45 days later. In this case the concrete pile of diameter of 50 cm, 10 cm of wall thickness and 16.2 m of length, was driven with a 50 kN hammer
kN weight hammer. The sedimentary soil layers are constituted by 12 m of soft clayey silt soft clay layer, overlying 4.0 m of hard micaceous sandy silt and 12.0m of very hard micaceous clayey silt, in the Tertiary Itambe Formation, Parana State. The maximum static load was 2600 and the maximum mobilized Smith's static resistance was 2520 kN. Figure 17 presents the correlation between SML static test and DIET static results performed on the concrete pile of 50 cm of diameter and 9 cm of wall width, 24.0 m of length and hammer weight of 55 kN, driven on shoreline soil formation, through 22.8 m thickness very soft sedimentary silty clay, underlying a 8.7 m thickness of compact silty sand (23
Figure 15. Correlation DIET and SML test curves. weight, in varied sandy soft to medium sedimentary soil, overlying the dense silty sand (20
648
Figure 17. Correlation DIET and SML test curves Figure 18 presents the results of SML and DIET tests performed in three concrete piles, of 80 cm of external diameter, wall width of 15 cm and 48 m of length. These piles were driven in homogeneous marine lacustrine sedimentary soil formation, with 30 m thickness soft silty clay (N=2) layer, overlaying 10 m of medium over consolidated clay (N=5) and 12 m of dense to very dense (20
mum mobilized static resistance in the DIET was 62 YOof the structural pile strength corresponding to the maximum displacement of 3 YOof the pile diameter.
EOD =
Figure 18. Correlation (4) DIET‘S and (2) SML test curves on large diameter piles with the same length Piles I and I1 were subjected only to the dynamic increasing energy test. The set of the points shows that the curve trends are in good agreement with the static results of pile A. The shape of the curves are typical of piles working by the point. The general comparison between the DIET and SML test curves (figures 14 to 18), shows that the load - settlement curve of the conventional SML static loading and the Smith‘s static resistance - displacement curve of the dynamic increasing energy test show a reasonable agreement for engineering purpose. It can be concluded that the dynamic increasing energy test (DIET), as currently used in Brazilian foundation practice, improves the reliability of the pile bearing capacity prediction. 5.2 Setup measured in increasing energy test In coastal sedimentary soft clay formations it is well established that the soil setup takes place very quickly. But how quick is the setup effect? The elapsed time between the end of the driving and the start of the test can be hours or days, depending on the pile test assemblage time. An advantage of the dynamic increasing energy test is that the loading can be performed at time intervals as short as desired. Figure 19 presents the results of a typical investigation of the soil setup in the case of a concrete pile, of 26.6 m of length, external diameter of 42 cm and 8 cm of wall thickness, driven with a free fall hammer of 42 kN of weight, in the sedimentary marine coastal of the Baixada Santista soil Formation in Siio Paulo State. 649
40’
t = 80’
+ t = 10’
t = 20’
o t = 13days
Figure 19. Initial setup in very soft clay layers. The soil was constituted by 5 m thickness soft clayey sand layer (N=2) and 9 m of very soft marine sandy silty clay (N=l), overlaying 12 m thickness soft to medium fine clayey sand (1
reinforced concrete piles under dynamic loading conditions.
REFERENCES Aoki, N. 1989a. Prediction of the Behavior of Vertical Driven Pile under Static and Dynamic Conditions. Drivability of Piles (2): 55-61; Proc.Disciission Session 14, Int. Con$ Soil Mech. and Fozind. Eng. (12), Rio de Janeiro, August I9S9. Japan : Japanese Soc. Soil Mech. and Found. Eng. Aoki, N. 1989b. A new dynamic load test concept. DrivabiliOJ of Piles (I): 1-4; Proc.Discussion Session I4, Int. Con$ Soil Mech. and Foui7d. Eng. (IZ), Rio de Janeiro, August 1989. Japan : Japanese Soc. Soil Mech. and Found. Eng. Aoki, N. 1 9 8 9 ~Discussion . on Driveability of Piles. Proc. Int. Con$ Soil Mech. Found. Eng. (I2), Rio de Janeiro, Discussion Session I j : ( 5 ) : 2977-2978. Rotterdam: Balkema. Aoki, N. & de Mello, V.F.B. 1992. Dynamic loading test curves. In Barends (ed.), Proc. Iift. Coif on Application of Stress- Wave Theory to Piles (4); 525-530. Rotterdam: Balkema. Aoki, N. 1997. Evaluation of the ultimate bearing capacity of piles driven with increasing energy (in Portuguese). Doctor of Engineering thesis, School of Engineering of Sao Carl o ~ University , of Sao Paiilo: 1 10. Aoki, N. & Cintra, J.C.A. 1997. New interpretations of the dynamic loading curves for driven piles based on the energy approach. In M. Almeida (ed.), Recent Developments in Soil and Pavement Mechanics: 467-472. Rotterdam: Balkema. Beer, E.E. 1988. Different behavior of bored and driven piles. In W.F.van Impe (ed.), Proc. Int. Geotechnical Seminar: Deep Foundations on Bored and Auger Piles: 47-82. Rotterdam: Balkema. Beim, J.W. & Aoki, N. 1996. Dynamic load test method with variable energy. In F.C. Townsend & M.Hussein & McVay, Proc. Int. Con$ on Application of Stress-Wave Theory to Piles (5): 274-28 1. Orlando. Bernardes, G.P. 1989. Dynamic and static testing of large model piles in sand. Doctor of Engineering thesis, Department of Civil Engineering, Norwegian Imtitute of Technology, Trondheim. Broms, B.B. & Choo, L.P. 1988. A simple pile driving formula based on stress-wave measurements. Proc. Int. Con$ on Application of Stress-Wave Theoiy to Piles (3): 59 1-600. Ottawa: Canada. ChelIis, R.D. 1951. Pile foundations. New York: McGraw-Hill. Chen, W.F. & Liu, X.L. 1990. Limit analysis in soil mechanics. Developments in Geotechnical Engineering (52). Amsterdam: Elsevier. Clough, R.W. & J. Penzien 1975. Dynamics ofstrzxtures. New York: McGraw-Hill. Davisson, M.T. 1972. High Capacity Piles. Proc. of Lectitre Series. Innovations it7 Foundation Construction. Soil Mech. and Found. Div. Illinois Section. ASCE and Dept. of Civil Eng. Illinois Inst. of Technology. Chicago. Fellenius, B.T. 1980. The analysis of results from routine pile load tests. Ground Engineering v. I3 n. 6, September: 19-3 1. Fujita, K. & Kusakabe, 0. 1988. On the evaluation of static bearing capacity. Proc. Int. Conf on Application of StressWave Theory to Piles (3): 525-534. Ottawa: Canada. Goble, G.G.; Likins Jr., G.E. & Rausche, F. 1975. Bearing capacity of piles from dynamic measurements. Final Report. Depart. of Civil Eng. Case Western Reserve University, Cleveland Goble, G.G.; Rausche, F. & Likins Jr., G.E. 198 1. The analysis of pile driving - A state-of-the-art. In H. Bredenberg (ed.),
650
Proc. Int. Seminar on Application of Stress-Wave Theory on Piles: I3 I - 16 1. Rotterdam: Balkema. Goble, G. & Likins, G. 1996. On the application of PDA dynamic pile testing. In F.C. Townsend & M.Hussein & McVay, Proc. Int. Conf on Application of Stress-Wave Theory to Piles (5): 263-273. Orlando. Koten, H.; Sandbergen, R. & Starink, T.W. 1988. Model test on static and dynamic bearing capacity of piles. Proc. Int. Con$ on Application of Stress-Wave Theory to Piles (3): 368-376. Ottawa: Canada. Langhaar, H.L. 1962. Energy methods in applied mechai7ics. New York: John Wiley. Middendorp, P. 1993. First experiences with statnamic load testing of foundation piles in Europe. In W.F.van Impe (ed.), Proc. of the 2'"' Int. Geotechnical Seminar on Deep Foundations on Bored and Auger Piles: 265- 272. Rotterdam: Balkema. Niyama, S. & Aoki, N. 1991.Correlation between dynamic and static loading tests in Experimental Research of EPUSPCampus. (in Portuguese). Proc. Seininario Engenharia de FundaGGes Especiais - SEFE 11, SE0 Paiilo (1): 285-293. Poskit, T.J. 1992. Keynote lecture: Problems of reconciling stress wave measurements with theory. In B.J. Barends (ed.), Proc. lilt. Conf on Application of Stress- Wave Theoty to Piles (4): 495-507. Rotterdam: Balkema Randolph M.F. & Deeks A.J. 1992. Keynote lecture: Dynamic and static soil models for axial response. In B.J. Barends (ed.), Proc. Int. Conf on Application of Stress- Wave Theoty to Piles (4): 3- 14. Rotterdam: Balkema. Rausche, F.; Goble, G.G. & Likins Jr., G.E. 1985. Dynamic determination of pile capacity. Journal of Geotechnical Engineering ASCE: v.111 n.3: 367-383. Reese, L.C. 1972. Axial capacity. Proc. Specialty Conference on Petformance on Earth and Earth-supported Structures. Purdue Ut7iversity - ASCE, June.v.111: 20 1207. Smith, E.A.L. 1960. Pile driving analysis by the wave equation. Journal of Soil Mech. and Found Division, ASCE, v.86, SM4: 35-61. Terzaghi, K. 1942. Discussions. Pile-driving formula. Progress report of the Committee on the bearing value of pile foundations. Discussion. Proc. American Society of Civil Engineers, v.68 n.2: 31 1-323. Terzaghi, K. 1943. Theoretical soil mechanics: 137- 143. New York: John Wiley. Vesic', A.S. 1975. Principles of pile foundation design. Soil Mechanics Series No. 38. Duke University, School of Engineering. Vrouwenvelder, A. & Dieterman, H.A. 1992. General report 4: Reliability of predictions. In B.J. Barends (ed.), Proc. Int. Conf: on Application of Stress-Wave Theory to Piles (4): 509- 5 17. Rotterdam: Balkema. Veen, C. 1953. The bearing capacity of a pile. Proc. Int. Conf on Soil Mech. and Found. Eng. (3). ( 2 ) :84-90. Zurich. Weele, A.F. 1993. Quality assesment foundation piles after installation. In W.J.van Impe (ed.), Proc. Int. Geotechnical Seminar: Deep Foundations on Bored and Auger Piles (2) BAP 11: 459-467. Rotterdam: Balkema. Whitaker, T. & Bullen, F.R. 198 1. Pile testing and pile-driving formulae. In F.E. Young (ed.), Piles and Foundations 12 1-134. London: Thomas Telford. Whitaker, T. & Cooke, R.W. 1961. A new approach to pile testing. Proc. Int. Con$ on Soil Mech. and Found,. Eng. (4). (2): 17 1-1 76. Paris: Dunod.
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Correlation analyses of dynamic and static loading tests for nine piles Y.M.Zheng, J.M.Zheng & B.Chen Shunde Supervision Stutionfor Quality ancl Sufety of Construction Engineering
ABSTRACT: Can the accuracy of dynamic pile testing meet the demands of engineering? Engineers are always concerned with this problem as well as fall into arguments on it. In this paper, the correlation results of dynamic and static loading tests for nine prestressed-concrete pipe piles are studied. The two testing methods are compared in three aspects including the capacity, the settlement ratio and the load-movement curve. In addition, causes of diversity are discussed. As a tool for the acceptance test, the dynamic testing method is evaluated based on comparisons and analyses. dynamic tests, correlation can be studied in three aspects. However, for piles neither tested to failure statically nor full capacity mobilized dynamically, only latter two aspects, the settlement and the loadmovement curve, can be studied.
1 INTRODUCTION The dynamic method mentioned in this paper is the hi h strain testing in which the Pile Driving Analyze?" manufactured by PDI is employed and measured curves are matched with the CAPWAPC"' program Dynamic testing is performed in accordance with Specification for High Strain Dynamic Testing of Piles (JGJ 106-97 1997), which is a generic standard of the People's Republic of China The static loading test is carried out by the maintained load test in which reaction load system is used The standards of the static loading test are Code for Building Foundation Design (GBJ 7-89 1989) and Technical Code for Building Pile Foundations (JGJ 94-94 1994), which are a national standard and a generic standard of China respectively As for the theory of the high strain testing and the detail specifications of the static loading test, please refer to the standards mentioned above and Application of PDA dynamic-testing method to bored piles (Zheng & Ma 1996) Dynamic tests were performed on nine test piles first, then the pile predicted capacities were used as the basis of static load increments, so that the error due to rough increments might be reduced Pile installation parameters, soil conditions and test results are illustrated with tables and figures in the paper Correlation between the dynamic and static tests is investigated by two quantitative analyses and one qualitative analysis They are the bearing capacity of single pile, the settlement under the ultimate load (Sd in dynamic test, Ss in static test) and the shape of the load-movement curve For piles tested to failure in static tests as well as mobilized sufficiently in
2 PILE INSTALLATION AND SOLL CONDITIONS The nine test piles are chosen from five sites in Shunde district, Guangdong province, China Four of them were installed by silent piling in two sites, the others were driven with a diesel hammer. One is 0500mm, four are 0400mm and others are 0300mm The shortest pile is about 7m long, the longest one is 38m, the others are from 20m to 30m, (Details see Table 1) The boring log of the hole nearest to one test pile is referred as the soil conditions of this pile Among them, since pile #81 of Meidi Shore Garden is far from any boring hole, its penetration depth is less than the bearing stratum depth according to its nearest hole, no boring data can be used for analysis Thus, corresponding stratum description of only eight tested piles is given in Figure 1 (The groundwater levels of all sites are less than 1 m below the ground surface) 3 TESTING RESULTS AND ANALYSES
The correlation results of dynamic and static loading tests for nine test piles are shown in Table 2. Figure 2( 1)-2(9) present the comparisons of the static loadmovement curves and dynamic simulated ones for 65
I
kpth
Pile #33 of
Pile #76 of
Pile #85 of
Pile #27 of
Pile#113 of
Pile #27 of
Pile #429 of
Pile #709 ol
(m)
Gaozan
Gaozaii
Gaozan
Gigui
Yujing
Meid
Huarun
Huarun
fill
1 2
plain fil
plain f i U
vlain fill
1 5
silty clay i f h i d plastic)
3 fine sand (loose
fine sand (loose
fine sand (loose)
plain fil muckv cla.
Sand
silt (little-dense)
dredge
fine sanc (lower p:
6
7
mucky silty
8
salid (loose
mucky silty sand (loose)
1llUCky s1lty sand (loose)
LS siltv
sand
siltv sanc
muck
trongweathcrec
coarsesill(
siltstonc silw clar
fine silt
9
sand(1oose)
10 11 12 13
11
fine sail (loos
fine saiil
fine saic
(loose
(loose'
gravel sand
p v e l sand
15 16
sandv clav
17 sandy c l q
18 muck
19
strong \v&m
20 21 22
IilIlCk?' Silt!
sand (loose
muck! silt! sand (loose
muclq Silt! salid (loose1
23
21 silt?. c1a.i
silo claj
mi1g \\eatllcIx
muck
(fluid dastic
(fluid plastic
ilt iiiterlaycr
fluid plastic
29
muck? silt?
muck! silt!
30
sand (loose
25
silt!, cla!
26
(fluid plastic
27
mucky silt!
28
salid (loose sand (loose)
fine sand
31
32
puddtiig stoiir
medium-densit!, )
puddtng stoiic
33
oarse g r x el
34 7 -
13
36
mllg \wltllc
37
inudstoiic
38
39 40
Figure 1. Conespoiidtng stratuin descriptioii of eight test piles
652
Table 1 Installation Parameters of tlie iiiiic prestressed pipc piles Project Pile Pile EmDate of In- Design Installation Name No Dia bednient stallation Bearing Method (mm) (in) Cap& ity (kN)
Gaomn Market Bigui Garden Y uj i ng Garden Meidi Shore Garden Huaruii
Fiz
Rani Weight Drop Total Blou Count Set of the last / Ma\iiiiuni Height B l o ~ of the last 30 blom s PllC CoIn(cm I 10 pressing Force
blows)
W)
300 300 300 400
29.0 3 1.0 31.0 7.2
Mar. 25.99 Mar. 3 1.99 Apr 3. 99 Apr. 9. 99
800 800 800 1200
1700 1700 1700 Dicscl hanimer 65
113 400
24.5
Apr. 4. 99
1200
Static press
400
38.2
Jun. 3. 99
1400
Diesel hammer 62
2.0
1506
81 400 429 400
29.0 24 8
Jun. 3 1. 99 May 23, 99
1400 1500
62
Diesel hanuner 50
6.2 2.0
396 1369
709 300
19.6
May 23. 99
1000
50
1.5
848
33 76 85 27
27
silent piling
---1.1. 1.1. 1.1
2300 113 (0.2111) 308 219
1.5. 1.5. 1.5
1.2. 1.2. 1.1 1.2. 1.2. 1.1
(0.8111)
Table 2 Correlation results of dyiatnic and static loadmg tests for nine prestressed-concrete pipe piles Project Pile Pile EnibedDate of Design Dq iiaiiuc Testing Results Static Name No Dia nieiit Installation Bearing Bearing Rd (111111) (111) Capaeitj, Beaniig Shaft Re- Toe Re- Capacit! (kN) CapaciQ sistaiice sistarice Rs Rd (kN) OtN) (kN) Gao7a1i 33 300 2 9 0 Mar 25.99 800 1863 1341 522 098 1900 Market 76 300 31 0 Mar 31.99 800 1881 1417 464 1900 0 99 85 300 31 0 Apr 3,99 800 2188 756 1432 1980 111 r306O ---Bigui 27 ** 400 7 2 Apr 9 99 1200 4007 431 3576 Garden Yujiiig 113 400 2 4 5 Apr 4.99 1200 2051 655 1396 2160 095 Garden Meidi 27 ** 400 38 2 Jun 4. 99 1400 4710 4110 600 2 5100 ---Shore Garden 81 ** 400 2 9 0 Jun 31.99 1400 4453 2619 1834 24700 ---Huarun 429 **300 24 8 Maj 23. 99 1500 4734 3605 1129 2 5225 ---Dope Plant 709 ""300 19 6 May 2.3, 99 1000 2943 1990 953 2 3000 ---* St is the settlcineiit under the static failure load. Sd is the settleinent under tlie dynanuc bearing capacity or the largest testing load. Ss IS tlie settlenient mider the static beaniig capacity or the largest testing load Neither the static loadtiig test nor the dynaniic testing reached the ultmate bearing capacity
is
(lw
each pile. Due to limited space, only two CAPWAP matching results are given in Figure 3 and Figyre 4. Comparing both the capacities and the loadmovement curves, we can see that: 1. The dynamic result of each test pile is a little less than its static value, except pile #85 of Gaozan Market. 2. For piles mainly depending on the friction resistance, such as: three piles of Gaozan Market and pile #113 of Yujing Garden, which were installed by silent piling, failure loads were obtained in static tests and soil resistance sufficiently mobilized in dynamic tests, the capacity errors between the static and dynamic are not more than 11%, settlement ratios(Sd/Ss) =l, the load-movement curves of two
653
256 (0 6111)
1.5. 1.5. 1.5
Ss*
sf
St"
Sd*
(111111)
(mm) (mm)
ss
6535 3 3 6 3 1 3 63 30 32 2 3 7 2 7224 35 3 4 1 8 ---- 21 1 1 6 4
107 0 87 083 129
5207 22 1 21 9
101
----
43 7 3 9 9
1 10
-------
364 306 2 6 4 21 2
119 124
----
25 1
195 123
methods are highly close with the second knee points almost identical. 3. For pile #27 of Bigui Garden, two piles of Meidi Shore Garden and two of Huarun Dope Plant, which were driven with a diesel hammer, the main resistance is at or near the pile toe and the static loading test curves are belong to "slight slope" without steeply downward sections. The total settlement couldn't achieve the failure criterion according to Code, therefore the ultimate bearing capacity couldn't be determined. Sd is a little more than Ss for this type, the ratios are from 1.1 to 1.3. As to the load-movement curves of two methods, three of five piles are analogous. While pile #429 and pile #709 are much the same as others at the
beginning and separate obviously from the middle part, and that the load-movement curves from dynamic tests are below the static ones. We have given Some preliminary analyses to above correlation results: Transient impact load is applied to mobilize the soil resistance around pile in the dynamic testing, 654
which is different from the static loading test. The pile capacity is gained by slowly loading in static tests. When the dynamic test is being performed, impact is applied on pile top for many times with a heavy hammer. In general, the upper resistance around the pile shaft is mainly mobilized by former blows. While the lower shaft resistance and end-
Figure 4. CAF'WAF' results of Pile #429 of Huarun
bearing capacity act gradually under later blows, at that time, pile has possibly lost small part of top friction resistance. This is the reason why the dynamic results are sometimes conservative. Besides, according to transfer mechanism of vertical load for pile foundations, adequate mobilization of shaft and toe resistance needs a certain displacement of the pile, which is related to the pile diameter 655
and soil around the pile. Commonly mobilization of shaft resistance requires 5-1 Omm displacement and 10-20mm is needed for sandy soil while for mobilization of toe resistance, even more displacement may be required. The displacement may come to 20-30mm in the dynamic testing. Such a displacement is enough for friction piles to be mobilized adequately. That is
why three piles of Gaozan Market and pile #113 of Yujing Garden have good correlation Nevertheless if lower soil layers have good conditions or the pile toe is built-in firmly, it will be difficult to have all resistance mobilized For example, in Huarun site, the middle soil layer around the pile is gravel sand and the pile toe is built in intensely weathered rock The penetration of the end driving was only 1 2-1 5cmllOblows (normally 2 0-5 OcdlOblows) The dynamic testing was performed one week after driving Thus, it should be expected that soil resistance of piles is hard to be hlly mobilized in this site It should be explained that although the dynamic results are more conservative than the static ones for the two piles of Huarun See Table 2, #429 is a 400mm dia pile with a design capacity of 1400kN, and the soil resistance mobilized in dynamic testing is up to 4734kN, which is more than three times of the design value, #709, 300mm dia , design capacity is lOOOkN, the dynamic capacity is 2943kN, near three times more than the design value Both of results are more than twice of the design vertical load capacity of pipe pile defined in Technical Specification for Prestressed Concrete Pipe Pile Foundations (DBJ/T 15-22-98 1998), which is depend on the material strength of pile shaft It can be calculated by following formu1a
R1, = 0.3 (fee - o p c ) A
(1)
Where R, = the design vertical load capacity of the pile shaft; f,, = the compression strength of the centrihgal concrete pipe pile; oPc= the effective prestressing force of the concrete pile; A = the section area of the pipe pile. This capacity for 400mm dia. pile is about 4200kN and for 3OOmm dia. pile is about 2400kN Hence, it is insignificant for design engineers to increase impact force in order to gain more resistance in dynamic testing, In addition, the ratios between the static and dynamic settlements of pile top are from 1.O to 1.1 under the design load. Therefore, for practice, the difference between the results of the dynamic testing and static loading test are not much at all. CONCLUSION 1. If the soil resistance has an adequate mobilization, results of the dynamic testing and the static loading test are quite close for most of prestressed pipe piles. That is to say, the measurement accuracy of dynamic testing method may be ensured when the correct testing procedure is carried out. Therefore, in general, dynamic pile testing can be used for the acceptance of piles instead of the traditional static loading test.
However, for the short (less than 1Om) or extra long (more than 40m) piles where the resistance according to the safety requirement of design is hard to be mobilized, the static test is recommended. For pipe pile foundations of the first grade building, in order to ensure safety, it is suggested that testing piles should be driven in the design stage and dynamic testing should be performed during the driving and restriking. And then, static loading test should be carried out. After the comparison of two testing results, dynamic method may serve as an acceptance test for working piles. Today in Shunde district, a great number of pipe pile foundations and square pile foundations are tested and accepted by the dynamic testing (More than seven thousand prestressed pipe piles have been tested in recent seven years). Considerable testing cost has been saved from this. As for the advantage in enlarging the testing quantity, ensuring safety and quickening the tempo of construction, it is much more than economical. 2. In Shunde, the design capacity of the prestressed pipe piles is always on the low side. For 300mm dia. piles, design capacity is less than lOOOkN, 400mm dia. less than 1500kN and 500mm dia. less than 2000kN. The measured capacities are close to or even more than three times of the design values (such as the five of nine piles in the cases chosen in this paper) The annual amount of the working pipe piles is up to tens of thousands, it will cause enormous waste and losses. We suggest that the design vertical load capacity defined in Technical Specification for Prestressed Concrete Pipe Pile Foundations, which is depend on the material strength of pile shaft, should be applied for designing, if the pile length and soil conditions are similar to the Huarun and Meidi site. In this way, both safety and economy would be achieved. We are gratehl to Prof. De-qing Li for his suggestion on this paper. REFERENCES DBJ/T 15-22-98 1998. Teclinical Specjfication for Prestressed Coiicrete Pipe Pile Fouidations: 8. A standard of Guangdong Province. GBJ 7-89 1989. Code Sc)r Building Foundafions Desig17: 104105. A national standard of the People’s Rcpublic of Cluna. Publishing House of Construction Industry of China. JGJ 94-91 1993. Technical Code for Buililing Pile Fouiidations:130-136. A generic standard o f the People’s Republic o f China. Publishiiig House of Construction Industry o f China. JGJ 106-97 1997. Specification for High Strain Dvnaniic Testi17g of Piles. A generic standard of the People’s Republic of China. Publishing House of Construction Industry of China. 2%ig Y.M. & C. Ma 1996. Application of PDA dynamnic&sting method to bored piles. Proceedings of PDMPIT Users: 83-94. ,
656
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkerna, Rotterdam, ISBN 90 5809 150 3
Back-analyses of steel pile driving records for quality assurance B. R. Danziger Department of Civil Engineering, Fluminense Federal University,Nitero'i,Brazil
J. S.Ferreira Coupe, Rio de Janeiro Federal University,Brazil
ABSTRACT: Steel piles are driven, in many cases, deeper than the SPT borings. To improve confidence in bearing capacity estimation for piles with penetration exceeding boring depth, steel pile driving records were back-analysed by the wave equation and the Danish formula. A comparison of the results by both methods is presented. Only those data including piles driven to penetrations shorter than SPT boring depth and located no farther than 5 metres from the SPT were selected. A good correlation between the driving resistance evaluated by the wave equation and the Danish formula was found. The authors then assured the confidence of the Danish formula for steel piles with characteristics within the range of the data base. 1 INTRODUCTION
Regardless of all the criticism involved in pile driving formulas, they are still used in many construction sites as a simple method available to control driving operation in the field. Although it is well known that the wave equation model can more faithfblly describe pile behaviour during driving, the driving formula is easier and less time consuming for practical applications. Furthermore, pile driving formula calculation can be extended to the whole piling, providing an efficient tool for quality assurance of the foundation job. Pile driving formulas are based on the conservation of momentum and energy (Chellis 1961). The Brazilian experience incorporated the Danish formula for application to steel piles on land. They are used mainly as a means to assure uniformity control in construction for a given pile job. Smith (1960) introduced the wave equation method that solves the problem of wave transmission along pile length, modelling pile and soil behaviour during driving in a more adequate form. The wave equation allows the introduction of the real boundary conditions provided by the embedded soil. The present paper analyses the driving data from steel piles using both the Danish driving formula and the wave equation. In spite of the numerous driving records available, nearly 10% of the records were
657
selected for the analyses, as only this small part contained complete and confident data set, including an SPT boring not farther than 5 metres away from the pile and extending below pile tip. Most discarded driving records were related to piles driven deeper than the SPT borings. The excessive driving length is mainly due to the high impedance of steel piles and the close range of pile set specifications. The soil nature around the pile length below SPT boring, in those cases, is unknown, posing difficulties for the determination of pile bearing capacity. 2 THE DANISH FORMULA
The Danish formula (DGI 1978) is written as:
where
Q,, - driving resistance 7 - efficiency factor
G - hammer weight
H - hammer drop L - pile length A - pile section area E - modulus of elasticity of pile s - pile penetration SO - elastic pile displacement The DGI (1978) bulletin presents some recommendations concerning mean value of pile length to be considered for piles shorter than 20 times the width of the pile. All the piles of the data base are longer than that limit, so the actual pile length was considered in the analysis. It is also mentioned in the DGI (1978) bulletin that in case of no direct measurements, q should be assumed as 0.7. In the present analysis a q value of 0.7 was adopted, as it also corresponds to the usual procedure in the application of the Danish formula in Brazil. The driving resistance will be evaluated in this paper for the two records usually considered in Brazil: (1) the blow counts for the last 50 cm penetration and (2) the pile set at final penetration, related to the 10 last successive blows. For the whole data base, both records (1) and (2) were taken with a drop hammer falling from one metre height. Table 1 presents the data base containing the complete set of information related to the 19 driving records, including SPT borings extending below pile toe. The columns include, from left to right, the job number, a description of a simplified soil profile, the pile designation and type, the pile section area, the pile length, the hammer weight, the blow counts for the last 50 cm penetration (record number (1)) and the pile penetration for 10 blows (record number (2)). 3 THE WAVE EQUATION ANALYSES
Two wave equation programs were used to estimate pile resistance during driving for the analyses described in this section: the Weap program (Goble et a1 1980) and the Dinexp program (Costa et a1 1988). Details concerning the use of the programs are discussed elsewhere, GRLWEAP (1988), Costa et a1 (1 988) and Danziger (1 99 1). The data needed to run the program were available in the pile driving bulletin, except soil nature and the soil resistance distribution along pile shaft and toe. The soil nature was obtained by means of the SPT boring profile close to the pile. The estimation of soil resistance during driving is necessary to make a proper choice of toe resistance percentage and side resistance distribution along pile shaft.
The driving of steel piles on land is usually performed with the hammer striking directly the top of the pile. Since no cap was used, and the Weap program can not be run without a cap, a solution was found by modelling the pile 0,lO m shorter and considering the pile top as the cap. The same procedure was introduced in the Dinexp program. The data preparation and analysis of each pile followed the sequential procedure: i) Establishment of soil profile (layering and estimation of soil parameters). ii) Estimation of soil resistance distribution during driving for the embedded pile length. iii) Establishment of driving data (hammer, pile). iv) Performance of driving analysis. v) Drawing the driving curve. vi) Introduction of the blow counts from the pile driving bulletin in the driving curve. vii) Determination of the driving resistance. The estimation of soil resistance distribution during driving (step ii above) has been performed according to Semple and Gemeinhardt (1981) and Stevens et all (1982). The soil damping at the pile point (Jp) and at the pile shaft (Js) introduced in the wave equation analyses were (Aas, 1993): Sand layers: J, = 0.1 s/m J, = 0.5 s/m Clay layers: J, = 0.65 s/m J, = 0.15 s/m The quake values were assumed as 2.54 mm (Smith 1960). The driving resistance was obtained from the driving curve, both for the blow count related to the last 50 cm penetration of continuous driving and also for the 10 successive blows. The soil resistance during driving calculated by the Danish formula and the Wave equation (Weap and Dinexp program) are indicated for the two records in Table 2. 4 STATISTICAL ANALYSES
Two statistical analyses were carried out. A one population analysis which included the whole data base and a two population analysis, in which the data were separated according to soil type mostly present along the pile length. Unfortunately, the data base included predominantly clayey soils. Piles embedded mainly in sandy soils included only five cases, which were not sufficient for a statistical analysis.
Table 1. Pile Data Base. Job Number 1
1
2
3
3
3
3
3
3
1
3
3
3
Simplified Soil Profile Depth (m) Type of soil 0-9.2 9.2-14.0 14.0-19.0 0-9.2 9.2-14.0 14.0-19.0 0-10 10-20 20-25 0-8 8-10 10-18 18-30 0-3 3-6 6-18 18-30 0-3 3 -6 6-18 18-30 0-3 3 -6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30
sandy clay silty sand sandy silt sandy clay silty sand sandy silt silty sand sandy clay clayey sand clayey sand sandy clay clayey sand sand! cla! sandy cla! silh sand sandy silt sandy clay sand! clay silh sand sandy silt sandy claj sand! clay silty sand sandy silt sand! cla! sand! cla! silt! sand sand! silt sand! cla! sand! cla: silt? sand sand! silt sand! cla! sand! cla\ silo sand sandy silt sand! claj sandy cla! silo sand sand! silt sand! claj sand! clay silh sand sand! silt sand! cla! sand! cla! silty sand sand! silt sandy clay
Pile
Section
~ ~ ~ gHaininer t h Blow Set 10 Weight counts (1) blows (2) (m) (kN) (I 5Ocin) (inm)
Designation
Type
(cm’)
B- Bloc0 2
I 12” 5 1/-
77 3
14.50
91
192
77 3
14.50
91
194
96 2
22.00
12 2
204
181 8
26.65
13 1
80
184 8
29.50
43 1
93
18
181 8
28.60
13 1
62
25
184 8
29.18
43 1
208
25
181 8
29.19
13 1
156
181 8
29.70
13 1
175
181 8
25.00
13 1
71
181 8
29.25
13 1
156
23
120u120u11 inm
181 8
29.00
13 1
85
25
420\120\11 inin
18-18
29.50
13 1
85
25
B- Bloc0 1
I 1277 5
M
0
w
10
I1 10” 15/8”
16 F3 1
120u120sll
0
imn 16 D12
36
D13
r
120s120sll iniii L
120u120\11 inin
16
U
D9
420u120ull inin
16 D11
L1 120u120ull
inm 16 D 10
120\420\11 llllll
16 D5
16 E1
120u120.;ll mni -1 12Ox120ull 111111
16 E5
16 E6
659
Table 1. Pile Data Base (cont.). Job Number
Simplified Pile Soil Profile Depth (m) Type of soil Designation Type 3 0-3 46 sandy clay c 3-6 E3 silty sand 420u420x11 6-18 sandy silt mm 18-30 sandy clay 46 7 0-3 sandy clay U E2 3-6 silty sand 120u420ull 6-18 sandy silt inm 18-30 sandy clay P I1 4 0-8 clayey silt 8-10 130L 10'' 4 518" silty clay 10-13 sand 1 0- 12 P I1 clayey silt 7A 12-17 10" 4 518" silty clay 17-18 clayey sand 5 0- 1 P I silty sand 12" 6 9/32" 1-1 11B sandy clay 1-27 silty clay I 0-10 P 6 sandy clay 10" 4 518" 10-12 M20 sand 12-18 sandv silt (1) blow counts for the last 50 cm penetration. (2) the pile penetration for 10 successive blows.
Section
Length
Hammer Blow Set 10 Weig11t counts (1) blows (2)
(cm')
(in)
(kN)
(I 50cm)
(mm)
184.8
30.50
13.1
80
20
184.8
3 1.OO
43.1
72
25
96.2
12.00
12.2
191
96.2
11.00
12.2
150
78.0
18.50
7.5
250
18.1
14.70
6.8
219
11
Table 2. Soil Resistance during Driving. Danish Formula and Wave Equation Programs (Weap and Dinexp). Job Simplified Pile SRD (kN) SRD (kN) Weap SRD (kN) Dinesp Number Soil Profile Type Danish Formula 1
1
2
3
3
1
7
Depth (in)
Type of soil
0-9.2 9.2-11.0 11.0-19.0 0-9.2 9.2-11.0 14.0-19.0 0-10
I sand! cla! 12" 5 'A'* silh sand sandy silt I sandy cla! 12'' 5 %" silo sand sandy silt I1 silh sand 10" 4 518" sand) cla! clale) sand r L! cla!e! sand sandy clay 420s420\rll mm cla~e! sand sand! c l a ~ sand! cla) s i l ~sand 420\;120xll min sand! silt sand! c l a ~ sand! cla! 420s120sll s i l sand ~ I11111 sand! silt sandy clay
10-20 20-25 0-8 8-10 10-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30
(1)
(2)
(1)
802.9
1194.3
813.1
952.0
804.9
1191.3
838.5
955.5
921.3
1065.0
1290.5
1559.2
1858.0
2750.0
1836.5
2477 4
1613.6
2263.1
1756.0
2590.0
1876.4
2112.0
1570.4
2201.8
1867.7
2585.0
1627.0
2313.0
1321.0
2014.8
1510.0
2338.5
660
(2)
(1)
(2)
Table 2. Soil Resistance during Driving. Danish Formula and Wave Equation Programs (Weap and Dinexp) (cont.). Job Number
Simplified Soil Profile Depth (in) Type of soil
Pile TYPe
sandy clay silty sand 120.1120sll sandy silt inm sandy clay sandy clay 3 silty sand 120.1120sll sandy silt inin sandy clay 1 sandy clay 3 silty sand 12Ox12Osll sandy silt mm sandy clay sandy clay 3 0 silty sand 120x420.111 sandy silt niin sandy clay sandy clay 3 C silty sand 120s120sll sandy silt inm sandy clay sandy clay 3 0 silty sand 120s420ull sandy silt inin sandy clay sandy clay 3 C silty sand 120s120sll sandy silt inin sandy clay 3 sandy clay silty sand 120.1120sll sandy silt nim sandy clay 3 sandy clay i silty sand 420.1420sl1 sandy silt inin sandy clay I1 clayey silt 1 10”1518’silty clay sand I1 1 clayey silt 10“4 518” silty clay clayep sand I 5 silty sand 12”6 9/32’’ sandy clay 1-4 1-27 silty clay I 6 0-10 sandy clay 10’’4 518” 10-12 sand sandy silt 12-18 (1) blow counts for the last 50 cm penetration. (2) the pile penetration for 10 successive blows.
3
1
0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-3 3-6 6-18 18-30 0-8 8-10 10-13 0-12 12-17 17-18 0-1
SRD (kN) Danish Formula (1) (2)
SRD (kN) Weap (1)
(2)
SRD (kN) Dinesp
(1)
(2)
2311.9
2291.3
2294.3
2011.8
2366.0
2140.0
2172.0
2377.1
1923.5
2191.0
2170.5
2360.0
2217.6
2312.8
1935.0
2037.7
2335.5
2175.0
1785.1
2513.3
1179.4
2265.5
1721.6
2755.0
2176.3
2327.1
1975.3
2129.5
2270.0
2199.0
1828.8
2300.0
1577.3
2079.3
1750.0
2110.0
1818.7
2281.2
1587.0
2086.2
1810.0
2156.0
1760.1
2340.9
1506.8
2111.9
1719.5
2550.0
1683.9
2238.6
1128.1
2013.4
1611.0
2370.0
11 15.4
1391 7
1100.0
1411.5
1245.0
1633.0
1017.6
1091.8
1200.0
1209.6
11 19.0
1170.0
705.2
827.4
602.3
756.9
790.0
1300.0
583.0
681.4
459.5
556.0
918.0
1305.0
I
66 1
4 1 A one poyzrlatron atialysrs Values from Table 2 suggest that a strong correlation can be found between the driving resistance evaluated by the Danish formula and by the wave equation Such correlation should pass through the origin A fbnction of the type y = B x (SRDIVE= B SRDD,tnlsh) was searched, with the coefficient B and the correlation coefficient given by Bussab (1 988) SRDWEis the soil resistance during driving obtained by the wave equation analysis and S R D D a n l s h is that evaluated by the Danish formula Two cases were analysed one related to the data obtained at the end of continuous driving, “the last 50 cm penetration record’, labelled as (1) in Tables 1 and 2, and the other obtained after complete driving, “1 0 successive blows penetration record”, labelled as (2) in the same Tables For each case, two fbnctions were established one from the results of Weap program, SRD\ijq, and the other including the results from Dinexp program, S R h n s , , , (Table 3)
Figure 1. Correlation between soil resistance during driving (in kN) from Danish (1) formula and Wave Equation Dinexp (I). 4.2 A two yoyu Iatiori analysis
Table 3. Soil Resistance during Driving. Statistical Results. One Population. X Variable Y Variable B Correlation Coefficient Coefficient Weap (1) 0.90 0.9946 Danish (1) Danish (1) Dinexp ( 1 ) 1.03 0.9897 Danish (2) Weap (2) 0.92 0,9964 Danish (2) Dinexp (2) 1.10 0.9829 Weap (1) Dinexp (1) I.I 5 0.9958 Weap (2) Dinexp (2) 1.21 0.9917
A good correlation was found, with a correlation coefficient around 0.99 for all correlations, considering both the driving records for the last 50 cm penetration (1) and also for the additional 10 successive blows (2). Similar results were found in the comparison between S R D D a ~ ~and i s ~ ~SRDW (Danish formula and wave equation analysis) and in the comparison of the wave equation performed by two distinct programs (Weap and Dinexp). Figure 1 illustrates one of the linear fknctions passing through the origin. It can be seen that the individual points are really very close to the curve. The surprisingly good correlations between the Danish formula and the wave equation assures the confidence of the Danish formula in evaluating the soil resistance during driving for the steel piles of the data base.
The data from Tables 1 and 2 were now separated into two groups: the first group including the piles embedded in predominantly clayey soil and the second group those piles embedded in sand. In view of the very few cases included in the second group, only the first group was analysed. The results are shown in Table 4.
~~
Table 4. Soil Resistance during Driving. Statistical Results. Piles predominantly in Clays. X Variable Y Variable B Correlation Coefficient Coefficient Weap (1) 0 89 0.9975 Danish (1) Danish (1) Dinexp (1) 1.OO 0 9990 Danish (2) Weap (2) 0.72 0 9840 Danish (2) Dinexp (2) 1.07 0 9988 0 9987 Weap (1) Dinexp (1) 1.13 Weap (2) Dinexp (2) 1.18 0.9982 The same good correlation was found, with a correlation coefficient even higher now. 5 TIME EFFECT
As soon as the penetration for the additional 10 successive blows occur after the pile is driven to its final depth, that is, after the end of continuous driving, the comparison of the estimates of driving results from both records can express a time effect. It should be emphasised, however, that the time lag separating both records is very short in practice,
662
lasting just some minutes. Redriving the pile days later would probably show much stronger effect. The analyses are only related to piles in clay. Piles in sand included very few cases for which time effect is not expected to occur. The results are indicated in Table 5. Table 5 . Soil Resistance during Driving. Time effect from end of continuous driving and additional 10 sucessive blows. X Variable Y Variable B Correlation Coefficient Coefficient Danish (1) Danish (2) 1.20 0.9925 Weap (1) Weap (2) 1.03 0.9897 Dinexp (1) Dinexp (2) 0.92 0.9964 According to what was expected, time effect was not significant within the time lag separating both records. 6 CONCLUSIONS
The paper analyses the driving data from steel piles using both the Danish driving formula and the wave equation. Good correlations were found between soil resistance during driving determined by the Danish formula and two different wave equation programs. High correlation coefficients were obtained when the data base was grouped as one population and also when the data was separated according to soil nature mostly present along pile shaft. Similar results were found when comparing both the Danish formula and the wave equation solution and two distinct wave equation programs. Such results assure the confidence of the Danish driving formula in evaluating the soil resistance during driving for steel piles with characteristics similar to those of the data base. The conclusion above is particularly usehl when the pile penetration is deeper than the SPT boring, which represents a very common situation in the engineering practice in Brazil. For those cases, the driving formula may be the only tool available to verify piling performance. ACKNOWLEDGEhENTS The authors are gratehl to Dr. Alvaro Maia da Costa, from PETROBRAS (the Brazilian State Oil Company), for allowing the analyses with Dinexp Program.
663
REFERENCES Aas, P M 1993 Personal communication, from Norwegian Geotechnical Institute Experience Bussab, W 1988 Analise de F'nridncia e de RegressLio. SLio Paid0 Atual Editora LTDA Chellis, R D 1961 Pile Foundatioris. McGraw-Hill Book Company New York Second Edition Costa, A M , L F R Moreira, N F F Ebecken, A L G A Coutinho, L Landau & J L D Alves 1988 Recent Application of Computer Methods for Drivability Analysis of Offshore Piles in Brazil Proceedings of the Iriteriiatronal Conference on Computer Modellitig r n Oceari Str1~tici.e.665-672 Veneza, Italy Danziger, B R 1991 Dynaniic arm'ysis of pile drising D Sc Thesis COPPE, Federal University of Rio de Janeiro Brazil (in Portuguese) DGI 1978 Code of Practice for Fouridation E~gi~eeririgBulletin I T' 32 The Danish Geotecnical Institute Goble, G G , F Rausche & G E Linkins 1980 The analysis of pile driving - A state of the art report Proceedings of the 2nd Iriteniatiorial Corference on the Applicatrori of Stress Wave Theory or?Piles, pp 131-161 Stockholm GRLWEAP 1988 Mamal for Wave Epntiori Aiialysrs of Pile Driving Goble Rausche Likins and Associates, Inc Semple, R M & P Gemeinhardt 1981 Stress History Approach to Analysis of Soil Resistance to Pile Driving Proceedrrigs of the 131h Offshore Technology Coiiference. Vol 1, 465-48 1 Houston, Texas, USA Smith, E A L 1960 Pile Driving Analysis by the Wave equation. ,Joimal of the Sorl Mecharircs arid Foinidatroris Divisioti ASCE Vol 127, part I, 1145-1193 Stevens, R , A Wiltsie & T H Turton 1982 Evaluating pile drivability for hard clay, very dense sand and rock Proceedirigs of the I 4 rh Ofshore Techriology Coifererice Paper No 4205, Vol 1, pp 465-481 Houston, Texas, USA
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Evaluation of pile set-up from penetration per blow Gary Axelsson Skunsku Teknik AB, Royal Institute of Technology,Stockholm, Sweden
Staffan Hintze Skunsku Teknik AB, Stockholm, Sweden
ABSTRACT: Case histories, published in recent years, show that the long-term increase in bearing capacity of driven friction piles in non-cohesive soil is often large. The present paper describes two cost-effective methods, in which the penetration (set) per blow is used to evaluate the relative set-up. The reliability of the methods was assessed in a comprehensive field-test involving both concrete piles and steel rods, driven in glacial sand. Dynamic loading tests were performed at different points in time, between 0-216 days after end of driving. Both an increase in bearing capacity and a decrease in set per blow were observed over time. The set-up was roughly 40% per log cycle of time for both the piles and the rods. Both methods showed that it was possible to determine the relative increase in bearing capacity with very good accuracy from the set per blow. 1 INTRODUCTION The long-term increase in bearing capacity (i.e. setup) of driven friction piles in non-cohesive soil is often large. Published case histories, summarised by Chow et al. (1998), suggest that the capacity is, on average, doubled over a period of 1- 100 days. A few studies (Chow et al., 1998 and Axelsson, 2000) also indicate that the set-up process continues for several months and possibly even years. Furthermore, the set-up is approximately linear with the logarithm of time. However, capacities older than one week are very seldom used in practice. The reasons for this are often lack of available waiting time, and also that it is relatively costly to perform loading tests at several points in time. The present paper describes two cost-effective methods to evaluate the relative set-up at a specific site by simply measuring the penetration (set) per blow at different points in time. The methods were assessed in a comprehensive field-test involving both concrete piles and steel rods, driven in glacial sand. The methods are believed to be very useful for an early estimation of the pile set-up at a site and for production control purposes.
set-up. In particular, it is the relative increase in bearing capacity that is the main item of interest, not the absolute capacity. Hence, the aim was to see if a general relationship exists between the decrease in set, due to pile set-up, and the related increase in capacity. 2.2 Method I : Empirical con-elation The authors’ have noticed from a number of piling projects, where piles were observed to have exhibited set-up, that the measured capacity (0) as a function of set per blow (S), can be approximated by the following type of relationship: Q=C7.S-b
where a and b are constants. It was firther noted that calculated results from various dynamic formulas and WEAP analysis can be also approximated with very good accuracy by this equation. The empirical method, derived from Equation (l), is based on the following simple equation, which relates the relative decrease in set per blow (SySJ to a corresponding relative increase in bearing capacity (WQ1):
2 PROPOSED METHODS 2.1 Principal coricept The main idea with this study was to see if the penetration per blow could be used to estimate pile 665
To establish the constant, b, in the equation the set per blow and capacity have to be determined at
two points in time. Further, in order to estimate the set-up at a third point in time only the set is required. Alternatively, b could be chosen from wellestablished correlations for a certain combination of pile-hammer-soil conditions. Advisedly, the method should be limited to data from piles where capacity has been hlly mobilised. A permanent penetration greater than 3-5 mm is often an indication that the ultimate capacity has been reached. However, using data from pile testing with small sets will in any case produce conservative predictions, or in other words low b-values. The equation could also be applied when the maximum transferred energy (E,,,,) varies between blows. In this case the penetration per blow is normalised with respect to the transferred energy by simply dividing S by E,,,ax,i.e:
distribution of shafthoe capacity and increased damping) 4. A second WEAP analysis is performed. The penetration, 5'2, and the corresponding capacity, 0 2 , is then evaluated at another point in time. / Q I ,gives the relative increase in ca5 . Finally, (I? pacity Table 1. The effect of time .on the dynamic soil parameters (typical trends). Shaft Damping Quake Increases Constant over time
I
Toe Damping No trend found
Quake Decreases over time
3 TEST SITE AND TESTING PROCEDURE
The test site is situated on the Fittja straits, 20 kilometres south-west of Stockholm. Extensive field testing has previously been undertaken at the site on instrumented concrete piles and steel rods, the main aim being to investigate the nature and the mechanisms of the set-up phenomenon. The results of these tests are presented in Axelsson (1998a,b,c), Axelsson (2000) and Axelsson & Westin (2000).
Results from WEAP analyses, for various pile types and soil models, show that for a certain capacity, the ratio, S/Emax, is almost constant when the set is large (i.e. when the capacity is clearly mobilised). 2.3 Method 2: back-cnlczrlntion iising WEilP
In the second method WEAP-analysis is used to back-calculate the capacity from pile penetration data. The soil model is either based on dynamic testing and C APWAP-analysis, or on experience from similar piling conditions. The method makes it possible to take into account both different transferred energy and time-dependent changes in the dynamic soil parameters. Several studies by Svinkin (1997) and others indicate that the most significant long-term change in the soil parameters is an increase in shaft damping over time. Axelsson (2000) also showed that set-up in non-cohesive soils is closely related to the shaft, and suggests that the set-up can, for practical purposes, be assumed to take place entirely along the shaft. In Table 1 are some typical trends summarised regarding time-dependent changes in the dynamic soil parameters. They are based on a review of the relative limited number of published cases that is available on the subject (Hunt & Baker, 1988, Svinkin 1997, Castelli & Hussein, 1998), as well as results from the dynamic testing in this study. The following procedure is suggested for estimating the set-up from pile penetration data: Asoil model related to the first testing occasion is created. A WEAP analysis is performed, and the capacity, 01, corresponding to the measured penetration, 5'1,is evaluated. Appropriate changes in the soil model are made to take into account the effect of set-up (e.g. re-
3 1 Soil conditroris
Figure 1 shows a plan of the test site with the location of the soil investigation, piles and rods The soil investigation consisted of soil sampling at six levels, CPT and dynamic probing. Furthermore, the pore pressures were measured using one piezometer and one oper, stand-pipe The soil consists of more than 40 m of loose to medium dense glacial sand. The results from the CPT, presented in Table 2, show a soil that is relatively homogenous with respect to the penetration
Figure 1. Plan of (lie test site.
666
resistance. The groundwater table lies approximately 2.0 m below ground level, and the soil being relatively well-graded varies between a silty sand and a gravelly sand. The relative density (Dr) is estimated from the CPT to lie between 35-50 % using a relationship by Jamiolkowski et al. (1985). The sand is considered to be normally consolidated with respect to its geological history. A mineralogical inspection of the sand showed that it mainly consists of hard minerals, such as quartz and feldspars. Table 2. Evaluation of CPT results
Depth [ml 0-2 2-1 1-7 7-13 13-19 19-21 21-23 23-
Soil type Clay Silty sand Sand Silty sand Sand Silt? sand Sand Sand
9c [MPa)
2-3 3-6 3-5 5 -9 3 -8 7-9
3.2 Irisfallatiori of piles and rods Figure 1 shows the location of the piles and rods. The rods (Hl-H21) were installed with the Swedish standard equipment (type “Borros”) used for performing dynamic penetration tests (dynamic probing). The equipment is fairly widespread outside Scandinavia and is classified between a DPH and a DPSH. The weight of the hammer is 63.5 kg and the drop height is 50 cm (E,,, = 63.5.g.0.5 = 311 Nm), with an efficiency of approximately 80%. The rod is a 32 mm diameter solid steel rod, normally equipped with an enlarged 45 mm diameter cone shaped toe to minimise shaft friction. However, to resemble the behaviour of normal pile driving, the dynamic probing in this study was performed without an enlarged toe. The rods were driven to a depth of 19.1 m below ground. Three 235 mm-square concrete piles (Pile A-C) were installed after completing the dynamic testing of the rods. The piles were dynamic load tested for set-up determination. In addition, two of the piles were instrumented with earth pressure cells on the shaft for measuring the long-term stress relaxation The piles were installed under easy driving with a 4tonne hydraulic free-fall hammer (ECH) and a drop height of 20 cm. Piles A and B were driven to a depth of 19.1 m and, after concluding the test program, Pile C was driven to the same depth in between the two piles. A piezometer, placed at a depth of 7.1 m and at a distance of one meter from Pile A, indicated that the excess pore pressure induced from the pile driving dissipated within a few minutes after the passage of the toe.
3.3 DyFianiic f e s f i q Dynamic testing was performed on the three piles and the 21 rods at different points in time from the end of driving, up to a maximum of 72, 143 and 2 16 days respectively for the piles, and up to a maximum of 69 days for some of the rods. The tests were performed using a Pile Driving Analyzer (PDA), together with two strain gauges and two accelerometers attached close to the top of the pile (or rod). To determine the static capacity and its approximate distribution along the shaft and on the toe, CAPWAP analyses were performed on the measured signals for seven of the rods and all the piles. In addition, the static capacity was also calculated for all the tests using the Case Method (RMAX) with damping factors of Jc=0.7 for the piles and Jc= 0.3 for the rods. These values were chosen as they produced capacities that best agreed with the CAPWAP results The drop height was held constant at 0.5 m for all the dynamic testing on the rods, while the drop height for the piles was gradually increased over time, from 0.2 m to 0.8 m, in order to mobilise the ultimate capacity. The penetration aRer each blow was measured by simply marking the pilehod between blows with a sharp object. The accuracy was estimated to be approximately a half-millimetre. 4 RESULTS AND ANALYSES 4.1 Bearing capacity versiis time
Figure 2 shows the relative increase in bearing capacity (Q.’Ql) versus the logarithm of time, where (21 is the capacity one day after end of driving (EOD). As can be seen, the capacity increased roughly linear with the logarithm of time. In particular, the increase in bearing capacity was observed to be on average approximately 40% per log cycle of time for both the piles and the rods. Moreo-
Figure 2. Increase in bearing capacity of piles and rods.
667
ver, the CAPWAP analyses indicated that the toe capacity was roughly constant over time for both the piles and rods, 335 kN (6.1 MPa) and 4 kN (5.3 MPa) respectively, signifying that the set-up took place along the shaft. Although the scatter was fairly large, the CAPWAP analyses also showed that there was a tendency for the pile toe damping and the rod shaft damping to increase over time. Further, there was also a tendency for the toe quake to decrease over time. More detailed information of these tests is provided in Axelsson (1998a,b). 4.2 Appraisal of the empirical method Figure 3 shows the bearing capacity versus set per blow for the rods. Furthermore, Equation (1) is bestfit to the data using regression analysis (least-square method). As can be seen good correlation (R2 = 0.88) is obtained for a wide range of set values (between 1 - 16 mm). However, it is questionable if the three highest capacities were hlly mobilised. If they are omitted, b will increase from 0.64 to 0.71. Depicted in Figure 4 is the bearing capacity (Case and CAPWAP) versus the normalised set per blow for the piles. The evaluated capacity, the transferred energy and the set per blow are also presented in Table 3. In Figure 4, Equation (1) is best-fit to the data using regression analysis (least-square method). As can be seen, very good correlation (R2 = 0.93 and 0.95) was obtained for the relatively large range of capacities and energies. Further, the CAPWAP regression curve is not as steep as the Case regression curve, providing an indication that an increase in soil damping took place over time. It is interesting to note that the constant, b, which determines the curvature of the relationship, is approximately the same for the piles and rods, 0.68 (CAPWAP) and 0.71 respectively.
Figure 4. Bearing capacity as a fhction of normalised set for the piles. Table 3. Results from the dynamic testing of the piles.
Pile A
Time after EOD EOD 1 hour I day 6days 37days 143days
Pile B
EOD 40min. 1 day 6 days 37 days 143days 216days
Pile C
1 dav 72 dais
Capacity (kN)
I
Case 481 630 740 1285 1505 1533 460 585 830 1445 1644 1668 1728 81 1 1436
I CAPWAPl 560 703 890 1247 1354 1441 529 678 1006 1402 1677 1710 1774 744 1441
Set per blow (mm)
-7 6 7 4 4 3.5 7 5 5 4 --
3 4 6.5 6
E,,
I
(Wm)
8.7 10.5 15.6 17.4 23.1 19.9 9.0 7.3 16.1 16.8 18.9 20.3 20.8 16.7 23.8
4.3 Appraisal of the WEAP based method
As with the empirical method, the WEAP based method was also compared with the results from the dynamic testing on the piles and the rods. Table 4 shows the dynamic soil parameters that were used in the WEAP analyses for the piles and the rods. These are the generally-used values (“average values”) and recommended according to the WEAP manual by Goble et al. (1998). In the WEAP analyses the toe capacity was assumed to be constant, as indicated by the CAPWAP analyses, while the shaft capacity was assigned as a variable (and increased over time). An analysis was also performed in order to examine the effect of increasing shaft damping (J.) over time. In this case the damping was gradually increased, from 0.10 s/m for the capacities at
Figure 3. Bearing capacity as a function of set for the rods.
668
EOD, up to 0.30 s/m for the capacities after a twoweek set-up period. Normalised set, S / E,, Capacity h = 0.2 I h = 0.3 I h = 0.4 I h = 0.5 I h = 0.6 I h = 0.7 m m m 91: 2:1 I 1.19 1.17 I 1.15 500kN 600 kN 1.08 1.10 1.07 1.04 1.01 0.99 700 kN 0.86 0.94 0.94 0.92 0.90 0.88 800kN 0.64 0.78 0.82 0.82 0.81 0.79 900 kN (0.45) 0.64 0.71 0.72 0.72 0.72 1000 kN (0.30) 0.52 0.61 0.64 0.65 0.65 i i o o k ~ (0.17j (0.39) 0.5 0.55 0.57 0.58 (0.19) (0.32) 0.40 0.44 0.46 1300 kN 1500kN (0.05) (0.17) (0.26) 0.31 0.36 (0.05) (0.14) 0.21 1700kN 0.25
'
Table 4. Soil model used in the WEAP analyses
Soil parameters Smith damping,
*
Shaft 0.16 s/m, and 0.10-0.30 slm
Quake
2.5 mm
Capacity
Variable
For square piles: d
= 2b I
Toe 0.5 dm d' I 120, (However, > l m ) Constant (Rod: 4 kN,
dz
The results from the WEAP analysis are presented in Figure 5 , together with the measured data for the rods. As can be seen, very good correlation is obtained between measured and calculated capacity for the whole range of data. However, the best correlation was achieved in the case with increasing J, over time, as can also be seen by comparison with the regression results in Figure 3. The results from the WEAP analyses for the piles are presented in Table 5 as S/El,,,, for the different drop heights. In the analyses, EInmwas varied between approximately 6.5 - 23 kNm for drop heights between 0.2-0.7 m. The values where the set per blow was less than approximately four millimetres for which it could be assumed that the ultimate capacity had not been completely mobilised - are displayed in parentheses. Presented in bold are the values that should be compared with the measured values, as the transferred energies are similar. The results show that the normalised set is approximately constant for a fixed capacity, under the condition that the capacity is hlly mobilised. Moreover, the results also clearly support the normalisation, S/El,,, made in Equation ( 3 ) . The results from the WEAP analyses are presented in Figure 6, together with the measured values. Further, a regression analysis was performed on the WEAP results using Equation (l), i.e. the empirical method. As can be seen, the regression curve gives a very good fit (X2= 0.96) to the WEAP data. Although the WEAP analyses could not match the measured data very well, it can be seen that the relative change in capacity as a hnction of set, i.e. the curvature, is fittingly matched, which is the main interest here. This is also revealed by the constant, b, determined by regression analysis, being almost of the same value, 0.68 and 0.72 for the measured data and the WEAP data respectively.
669
4.4 Comparison between the methods In the previous sections the empirical method and the WEAP-based method were compared against measured field data. Since both methods showed very good correlation with measured data, it could be instructive to compare the methods with one another. In Figure 5 and 6 it is shown that Equation (1) can be best-fit to the WEN analyses with very good Correlation. A sensitivity analysis was performed to see if this good correlation could be also achieved for a wide spectrum of soil models. In this case 14 WEAP analyses were performed for the same pile and hammer as above. The drop height was chosen to 0.5 m, the pile length was 20 m, and the shaft distribution chosen as constant with depth.
The results from the sensitivity study are presented in Table 6 in terms of the correlation factor, RZ,and the constants, a and b, that produced the best curve-fit using Equation (1). The range for the set per blow that was used in the regression analyses was between roughly 5 and 25 millimetres. The most interesting result fi-om these analyses was that the variation in the constant b was relatively small, ranging fi-om 0.73 to 0.87, where low values were linked to small damping factors, and correspondingly, high values were linked to high damping factors. On the other hand, the toe quake was shown to have very little influence on b. Some of the results are also illustrated in Figure 7. The influence of pile length was also examined. It was found that by shortening the pile length, h will increase. For example, for a pile length of 5 my the ST2 model (see Table 6) produced a b of 0.84 instead 0.79, for the same range of set. Furthermore, the correlation was also improved, resulting in RZ = 0.99. A similar sensitivity study was also performed for the rods. For a wide variety of soil models, the results produced values of h within approximately 0.80 to 0.85, for a range of set per blow of roughly 2 - 20 mm. It was also noted that a decrease in the shaft quake to 1 m produced slightly higher values of b. The correlation factor was as good as 0.99 for all the regression curves. 4.5 Factors injluencing the constant b
Figure 7. Some results from the sensitivity study on the piles.
As was revealed when applying the empirical equation to both measured data and WEAP results, the constant b showed a tendency to vary within a relatively narrow range of values, approximately 0.60.9. Before this can be applied in practice it has to be examined under a variety of different conditions.
Table 6. Results from the sensitivity study using WEAP. Shaft capacity
Quake (mm)
("/I
Shaft
0 0 0 0 50 50 50 50 50
2.5 2.5 2.5 2.5 2.5
50 50 50 100 100
2.5 2.5 2.5 2.5 2.5
Damping (mls)
Toe dl120 d/120 2dl120 2d/120 d/120 d120 dl120 d120 2dl120
Shaft
2dl120 2dl120 2d/120
0.16 0.65 0.65 0.16 0.65
0.16 0.16 0.65 0.65 0.16
Toe 0.25 0.5 0.25 0.5 0.25 0.5 0.25 0.5 0.25 0.5 0.25 0.5
a
b
R2
Set
Ident.
5098 5190 4536 4673 5551 5159 5080 4714 5034
0.77 0.86 0.74 0.85 0.77 0.79 0.82 0.83 0.75
0.97 0.98 0.98 0.98 0.97 0.97 0.97 0.97 0.97
5.2-25 5.5-25 5.3-24 5.6-24 5.7-26 5.6-23 5.6-23 5.4-24 5.3-26
5176 4740 4694 5538 5207
0.81 0.81 0.84 0.73 0.87
0.98 0.98 0.98 0.96 0.97
5.3-26 5.8-23 5.6-24 5.7-23 5.6-24
T1 T2 T3 T4 ST 1 ST2 ST3 ST4 ST5 ST6 ST7 ST8
670
s1 s2
Nevertheless, it would be of great benefit to be able to obtain a rough estimate of the relative set-up at a site by just striking the piles at different points in time. According to Equation (2), a halving of the set, for instance, would indicate a relative increase in capacity of 63 % for b = 0.7. If b was instead 0.6, the increase in capacity would be 52%, the difference only being 11 % in the estimated set-up. This example demonstrates that the variability in the relative set-up as a function of set is small, if b lies within the mentioned range. The authors would, however, like to emphasise that the two methods presented in this paper are by no means intended to replace pile testing for set-up prediction. Their main purpose is to provide an estimate of the set-up in, for instance, an early testing stage, or during production control. If we consider the following dynamic formula (the Energy Approach) derived from the energy equation by Paikowsky & Chernauskas (1992), and and D,,,, are measured quantities: where E,:,,,a,
It can be seen that if D,,,,, N" S, i.e. the total quake (pile and soil) is very small in relation to the set, the following expression is derived:
s
(5)
In fact the expression is a conventional dynamic formula (initially presented by Sanders in 1851) in its simplest form, thus disregarding the energy losses It also very similar to Equation (1) Nevertheless, it implies that the maximum value b can acquire is 1 0 Hence, high values of b should be expected for small soil quakes and large permanent sets, and further, for short piles with high stiffness (impedance) This was also shown in the WEAP sensitivity study However, a value less than 0 9, but probably in most cases less than 0 8, is more realistic, as near to perfect conditions (a ideal plastic soil and an infinitely rigid pile) are not very likely to be encountered in the field Changes in soil damping over time will also have an influence on h, in as much as an increase in soil damping over time will produce a smaller value, than if the damping was constant It is also possible to reason that for a normal hammer to pile impedance ratio and for non-extreme soil conditions, a value of h greater than about 0 6 is assumed reasonable Again under the circumstances that the ultimate capacity is mobilised At the moment, however, site specific determination of the constant h, by dynamic load testing and set measurements, is highly recommended. at least until further calibration of the constant b has been performed for a variety of field cases
5 CONCLUSIONS
Both methods, the empirical and the WEAP based, showed that it is possible to evaluate the relative increase in bearing capacity with very good accuracy from pile penetration data. The particular advantage with the empirical method is its simplicity, making it ideal for production control purposes. Furthermore, it was noted that the constant, b, can be expected to range between a relatively narrow span, approximately 0.6-0.8, which increases the value of the method even more. The WEAP based method, however, is suggested for use in more detailed evaluations of the relative set-up, as it is possible to account for timedependent changes in the soil parameters. The results also showed that there is very good correlation between the empirical equation and the WEAP analyses for a variety of possible soil models. An interesting observation was that the choice of dynamic soil parameters had little influence on the b-value. ACKNOWLEDGEMENTS The research was funded by the Development Fund of the Swedish Construction Industry, the Swedish Council for Building Research, Skanska AB. Furthermore, important contributions were made by the Commission on Pile Research, Stabilator AB, the Swedish National Railway Administration and the Swedish National Highway Administration. This is acknowledged with gratitude. REFERENCES A\elsson G (1998a). "Long-term set-up of dn\ en piles in noncohesn e soils e\ aluated froin dpnaiiuc tests on penetration rods" Proc I '' hit Co~f on Site Chnracterizntioii, .Itlantn Axelsson G (1998b). "Long-term set-up of dn\ en piles in noncohesive soils". Liceritiafe thesis, Kovnl Itistirute of Technology, Stochliolni Ayelsson G (1998c)."Long-tenn increase in shaft capacih of dmen piles in sand" Proc JthInr Coif on Case Histories 117 Geotech Eiigng , St Louis, l\fissouri, paper I 25 Alelsson G (2000). "Set-up of dmen piles in sand - The effect of constrained dilatancj on the sliafi belia~iourdunng loading". Proc 1171 Coiif oii Geotechnical R- Geological Eiig , GeoEng2000, .\feIboume, -4ustrnlin, (in press) Alelsson G & Westin A (2000). "Torque tests on dmen rods for predichon of pile set-up ' , Proc Iiit Colif, Geo-Dewer ,7000, Denver, Colorado, (IM press) Castclli R J & Hussein M (1 998). "Pile foundation construction for the Buckinan Bridge. Jacksoni ille. Florida". Proc 4th Itit Cotif 0 1 1 Cave Ifirtorier 117 GeotecJi Eiig17g Si I,OUIJ,.I fi s souri',paper 11 01 Chov F C Jardine R J . Naro! J F and Brucj F (1998) "Effects of time on the capacih of pipe piles in dense inaniie sand". -J GeotecJ?Engug , ASCE 1'01 121(3),p p 251-261
.
671
Goble G.G.. Rausche F.. L k n s G. & Associates Inc. (1998), “GRLWEAF’ manual“, P’ersion 1998-1. Hunt S.W. & Baker C.N. (1988), ”Use of stress-wave nieasurements to evaluate piles in high set-up conditions” Proc. 3”’ Iiit. Coi$ App. Stress-wave Theory to Piles, Ottawa, p p 689- 705. Jamiolkowsky M.. Ladd C.C.. Gennaine J.T. and Lancelotta R. { 1983, ”New developments in field and laboratory testing of soils” Theme Lecture, 1 lthInt. Coic O M Soil Alech. B Found. Eiigiig, San Francisco. Paikowsky S.G. & Chernauskas L.R. (1992),” Energy Approach for capacity evaluation of driven piles”, Proc. drh Int. Conf App. Stress-wave Theoty to Piles, The Hague, The iVetherlai7ds,p p 595-601. Sanders J. (185 I), “Rule for calculating the weight that can be safely trusted upon a pile which is driven for the foundation of a heavy structure”. J. Franklin Inst. XYII, p.304. Svinkin. M.R. (1997). ”Soil damping in wave equation analysis of pile capacity”, Proc. 5”’hit. Con$ .4pp. Stress-wave Theor?; to Piles, Orlanclo, Florida, p p 128-143.
672
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Comparative analysis of dynamic and static test of foundation pile Zhou Guoran & Wu Jiaduo Shanghai Harbor Engineering Design and Research Institute, People’s Republic of China
ABSTRACT: Since the non-ultimate bearing capacity of pile dynamic test can’t compare with the result of static test, this paper presents a comparison method between the P-S curves of static tests and the P-S curves of CAPWAPC matching based on experimental data obtained fiom 8 piles. At the same time, this paper also presents an assessment way of the quantitative analysis for the CAPWAPC result. 1 INTRODUCTION
2 COMPARATIVE ANALYSIS
Designer, contractor, supervisor and constructor pay close attention to the determination of the bearing capacity of the pile in pile foundation engineering. Early of this century, the mechanism of the traveling of the stress wave along the impacted rod had been recognized. Smith(1960) presented the numerical solution of partial differential equations for traveling of the stress wave along a one-dimensional rod: and the solution was called wave equation method. Beginning in 1964, Dr. G.Goble(1975), professor of Case Western Reserve University of U.S.A, studied the closed form solution of wave equation and presented Case and CAPWAP analytical methods from measured strain and acceleration of the pile top. These methods have been confirmed all over world. It is a cheap, fast and reliable method for the determination of the bearing capacity of a pile. In some countries, this pile dynamic test method has already become a standard for replacing or supplementing the static load test. Many cases(Stephen S.M. Cheng, Shaheen A. Ahman, 1988) showed there existed little discrepancy between the results of dynamic and static loading test, but the consistency of the two load tests have to be verified if the dynamic loading test is to be accepted widely. In the past, the comparative analysis between dynamic and static test was performed at the ultimate state. However, it isn’t always possible that both dynamic and static test reach the ultimate state simultaneously. This paper presents other comparative analysis method between the P-S curves of static tests and the P-S curves of CAPWAPC based on experimental data obtained from 8 piles, and explains this scientific method.
2.1 CAPWAPC Method CAPWAPC[7] method is an analytic method that calculates the bearing capacity of pile by measured force and acceleration with wave equation. Stress wave is computed fiom the force or velocity measurements. Improved Smith elastic-plastic model was adopted for soil. Soil resistance was modeled as a function of the moving pile, and in two components, namely: one is an elastic-plastic force of static state, the other is a viscous damping force of dynamic state. To start a CAPWAPC matching, a certain curve chosen fiom measured axial force of pile shaft or velocity with time change during impact of the hammer is taken as the initial input of matching. Then the parameters (soil resistance, Quake and damping) are adjusted continuously until computed and measured signals agree. After various parameters of pile and soil are confirmed, CAPWAPC calculates the axial forces and the settlement values of each cross sections of pile and determines the axial force and the settlement of the pile top, and the loading - displacement curve is drawn. This method is called the simulated static loading test. 2.2 Static Load Test Static load test is a testing method where a static action force is applied at pile top, and the settlement (displacement) is measured at pile top. Generally the action force is applied at pile top in increments. Test stopping criterion is that the displacement of pile top can’t stop or reach a certain limit value during a certain stage loading. 673
For reasons such as lack of sufficient reaction load, limit of hydraulic jack, limiting strength of reaction beams, the ultimate bearing capacity of the pile can't be obtained during test. Static loading test was carried out according to "Specifications For Port Engineering Technique" ((1987) JTJ222-87) of the Ministry of Communications of the People's Republic of China( 1987).
When settlements are same, sd~~]=&r-] and Sdr=SSr, the absolute error of maximum loading between both P-S curves is:
E of matching P-S curve relating to P-S curve of static is:
2.3 Contrast between Dynamic Test & Static Test The comparative analysis is needed between the results of dynamic and static load test. In the past, the results of both tests could not be compared if the bearing capacities were not hlly mobilized. Similarly the past comparative method was not so comprehensive although the bearing capacities were in ultimate state. Settlement is a very important data in design. Even though both bearing capacities are at ultimate state for dynamic and static test, and settlement can't be the same, the comparative result between dynamic and static test is difficult to evaluate scientifically. This paper presents a new comparative method that considers the comparison of both bearing capacity and settlement. See figure 1. Figure 1. Calculation method of the relative error of two curves
E reflects the accuracy and reliability of CAPWAPC matching. E provides an assessment way of the quantitative analysis for reliability of CAPWAPC matching.
3 THECASES The results of CAPWAPC matching and static load test of 8 piles had been collected in this paper. AAer submitting the results of CAPWAPC matching of 8 piles to supervisor and owner, then static load test was performed for 8 piles. The size of 8 piles were PHC(Pre-stressed Reinforced Concrete Spun Pile) pipe pile, steel pipe pile and square concrete pile. Testing location is at Marco and Shanghai in P.R. of China. The details are shown on table 1. Table 1. Size and soecification of testing tiles and olaces Type Spec. Length Toe Level Testing Name A-6 T-2 B-3 B-3a B-3c B-3b B-2 E-6
S is the P-S curve of static load test. D is the P-S curve of CAPWAPC matching. When both AP and AS reach an acceptable range simultaneously, the comparative result will describe that CPAWAPC matching is right. E is assumed to be the relative error of the bearing capacity obtained fiom CAPWAPC matching relating to the bearing capacity obtained fiom static test. There is: E=(IAd-Asl)/AsX 100%
where
is the enclosure area of D curve, As is the enclosure area of S curve.
(1)
Ad
6 is defined as the absolute error of both P-S curves. Between the loading of stage i and stage i+l, the area of their absolute error is following:
674
PHC Steel pipe PHC SC PHC PHC PHC Steel pipe
(mm) 0800tl10 0914t16 0800t110 600x600 0800t110 0800tl10 0800t110 0900t19
(m) 50.50 50.50 48.50 48.50 43.50 41.50 45.50 57.87
(m) -44.43 -44.49 -37.71 -38.68 -35.00 -32.59 -38.26 -51.97
place Macao Macao Macao Macao Macao Macao Macao Shanghai
only according to the previous comparative method. The results of both dynamic and static load test of 8 piles are as follow: Table Name
A-6 T-2 B-3 B-3a B-3c B-3b B-2 E-6
Results of d iamic and st ic test Static CAP WAPC Redriving capacity (kN) 7248.4 8303.2 7269.6 800 1.7 6729.6 5448.6 6703.4 12784.7
set (mm/blow) 1 .o 0. I 0.6 1.o 1 .o 2.0 0.5 0.4
capacity (kNI 7326 10567 7890 8453 6199 5636 6199 12768.1
REFERENCES
Soil state for static test Non-ultimate Ultimate Non-ultimate Ultimate Non-ultimate Ultimate Non-ultimate Non-ultimate
The P-S curve obtained &om CApWApC matching is drawn on the p-s curve obtained from static loading test (see figure2-9), then the relative error of maximum loading values of both curves at the same settlement is calculated, respectively. The details are shown in table 3.
comparison of P-S curve
(“/.I A-6
16.19
B-3a B-3c B-3b
29.16 6.89 1.42 12.47 13.35
Goble, G.G. , Likins, G. E.1975. Bearing Capacity of Pile from Dynamic Measurements: Final Report for Ohio Department of Transport, Department of Civil Engineering, Case Western Reserve University. Hussein, M.& Likens, G. 1991. Static Pile Capacity by Dynamic Methods, First Geotechnical Engineering Conference, Cairo University. Hussein, M., Rauche, F., 1991, Bearing Capacity of Deep foundation from Dynamic Measurements and Static Test ... Ten Correlation Cases, Malaysia. Ministry of Communications of the People’s Republic of China, 1987. Specifications For Port Engineering Technique” (( 1987) JTJ222-87) Pile Dynamic, Inc. 1992. Pile Driving Manual. Smith, E. A. L.1960. Pile Driving Analysis by the Wave Equation: Proceedings of American Society of Civil Engineering, vol. 86, No. SM4, pp35-61 Stephen S.M. Cheng, Shaheen A. Ahman, 1988. Dynamic Testing Versus Static Load Tests: Five Case Histories: Proceedings of Second International Conference on Case Histories in Geotechnical Engineering.
error between P. curves The relative erThe error beror between tween max loadings at ultimate cathe same setpacities tlement(%) (”/I 15.36 Can’t compare 0.62 Can’t compare 2.09 Can’t compare 32.72 Can’t compare 2.70 3.32 1.05 Can’t compare 3.88 Can’t compare 4.76 Can’t compare
4 CONCLUSIONS 4.1 This paper presents a comparative method between dynamic and static test. This comparative method may be carried out, no matter whether the bearing capacities of dynamic and static tests reach the ultimate state. This comparative method provides a scientific assessment of the quantitative analysis for CAPWAPC matching. 4.2 There exists a good correlation between the results of dynamic and static test. The discrepancy between static load test and CAPWAPC matching has been determined. 4.3 To mobilize the soil resistance hlly, the set per blow during dynamic test should not less than 2 mm. 676
Applicationof Stress-Wave Theory to Piles, Niyama & Beim (eds}02000 Balkema, Rotterdam, ISBN 90 5809 1503
Dynamic load test and elastic rebound analysis for estimation of the bearing capacity of piles in residual soil P.J. R.de Albuquerque Departamento de Estruturas e Fundapjes, Escola Politkcnica da U S 8 S6o Paulo, Brazil
D.de Carvalho Departamento de Construg6esRurais, Universidade Estadual de Campinas, Brazil
ABSTRACT: In this paper we will present the final bearing capacity results from a static load test and the control by the elastic rebound. These tests have been carried out by driving 0.18m diameter and 14.0m long precast concrete pile, at the Unicamp Campus, in a place where the subsoil is formed of residual diabasic soil, common in Campinas and other regions of the country. These results are compared to the ones obtained afterwards through an instrumental static load testy carried out on the same pile. The results presented refer to of two more dynamic load tests, which were performed close to the pile, where the static load test was performed and also the results of the field tests performed on the site. region, basic intrusive rocks occur from Serra Geral (diabase) formation. The local subsoil is characterized by two types of soil: silt-sandy clay (0 to 6m) and clay-sandy silt (6 to 18m). Water level was reached only at 18.0m. The soil of the first layer has collapsible behavior (Monacci 1995). The average results of the field tests are shown in Table 1.
1 INTRODUCTION In order to know the bearing capacity of a pile, empirical prevision methods are commonly used, since load tests are often economically unfeasible. New bearing verification methods have been developed to reduce costs. Among them is the dynamic load test, already known in the technical area, often providing point and lateral load values very close to the real ones. There is also the Reboundmeter System, developed by IPT/SP (Machado 1995), to check bearing capacity in a simple and inexpensive. Three precast concrete piles were driven (L=14.Om and $=0.18m) in a subsoil formed by residual diabasic soil. The results obtained through the dynamic load test were analyzed through Case, CAPWAPC and IPTCase methods. Specifically for the Reboundmeter (Machado 1995), the Uto et al. (1985) method and Chellis (1951) method were used, modified by Velloso (1987). The load test was the slow type (SML) and counted on the reading of deformations along the shaft, which comes from previously installed instrumentation, thus obtaining loads absorbed by the point and shaft. 2 SUBSOIL The tests have were conducted in an experimental area at Unicamp Campus, in CarnpinadSP. In the
3 PILE
The piles are 14.0m long, the first segment is 6.0m long, and the second is 8.0m long, connected through a metal ring bound by a weld cord. The cross section is circular, with a 0.18m nominal diameter. The concrete used has fck=35 MPa and the
Table 1 Alerage results of field tests Parainctcr 1'I La! er NVT 30 9c
2'ld L ~crJ 73 2473 9 kPa 214 1 kPa 152 60 0% 30 3% 15 5 kN/in3 58 7kPa 22"
920 8 P d 44.2 kPa e 172 n 63 1% IF 23 8% 13 6 kN/in3 Y C 27 7 kPa 6 70" Remark NspT=nuInber of blows of SPT. ql=polnt resistance (CPT). f,=skin frichon (CPT). e=r old rabo. n=porosio. fc
15 =moisture
content. y=specific natural n eight. c=cohesion
(total tension). $=friction angle (total tension)
677
steel has fyk=1500 MPa. In instrumentation CA-50 steel bars (L=0.60m and +=12.5mm) were used, in which strain-gages were installed, bound in complete bridge. These bars were installed at depths of 0.60m, 5.Om, 10.0m and 14.0m, inside an iron sheath left inside the pile while it was being made. Cement slurry was injected, so elements could be connected (Albuquerque 1996). Driving system used is a free-falling hammer, with a ram of 1,650 kg of mass, falling from a height of approximately 0.5m. 4. IPT-CASE SYSTEM Developed by IPT ( S ~ OPaul0 State Institute of Technological Research) this is a computer-based data acquisition and analysis system to be used to monitor dynamic pile driving. It digitizes, analyses and records strength and speed signals produced by the instruments (acceleration meters, specific deformation transducers) placed in a section near the pile head while it is being driven (Figure 1). By analyzing these signals, several important values are obtained in sifu, such as an estimation of pile bearing by the Case Method, maximum strength and speed, maximum transferred energy, maximum and final displacement, and evaluation of pile structure integrity. The purpose of developing this system is to supply the technical area with a domestic and potentially less expensive alternative to PDA
5. STATIC LOAD TEST
The directions established by NBR 12131/91 were followed, adopting the slow load (SML). The loads were made in consecutive stages, with load increases of 40kN, until the load in which the settlements indicated that the rupture of the connection pile-soil was reached. The unloading was made in
consecutive stages, with load reductions equal to 25% of the total load achieved in the test. The soil was excavated from the head of the pile up to the depth of 0.60m7 keeping this area as reference section to determine its Elasticity Modulus.
6. DYNAMIC LOAD TEST AND
REBOUNDMETER SYSTEM A basic set of instruments and other commonly used equipment (force transducers, accelerometer etc) were used (PDA) for dynamic load tests. These tests allowed checking integrity and bearing capacity of the piles (Albuquerque 1996). While the piles were being driven, measurements of elastic rebound took place were performed using the Reboundmeter System developed by IPT, which was in testing period (Machado 1995) The Reboundmeter System is a technique to assess soil mobilized resistance as a pile is being driven, through the elastic rebound value observed. This technique is an alternative to other methods such as conventional load test and dynamic driving monitoring. The value of elastic rebound (K) is obtained by measuring the displacement caused by the hammer blow. The displacement curve along time is measured for a section near the pile head and, using measured values of maximum (M x F) and final (S) displacements of the section observed, elastic rebound (K) is calculated. This value expresses the sum of numbers of maximum elastic compression of the pile shaft (KO) and of soil below the pile point
(W.
The Reboundmeter developed by IPT records and processes signals coming from a sensor. These signals describe the displacement of an instrumented section of the pile along time, as a result of the driving hammer blows. The system has a software that controls data acquisition and makes analyses, and a measuring device (Figure 2) comprising a displacement sensor, a mechanical device and a PC interface card (electronic circuit). The sensor is provided with a small rubber wheel When it rotates, this wheel transmits the rotating movement to the sensor, which converts this movement into displacement signals. The sensor is mounted on a mechanical device, which keeps the wheel pressed against the pile wall by means of a spring to ensure adherence between both surfaces. The displacement of this pile section, as a result of hammer blows, causes a corresponding rotation on the wheel, thus allowing the sensor to record its displacement curve along time 678
Table 5. Results obtained from the static load tcst Pilc QiZi (kN) Q,,li(W 2 219 43
Results obtained from the tests conducted, namely: dynamic load test (analysis: CAPWAPC, Case and IPT-Case) (Table 2), analysis of elastic rebound (Reboundmeter System) (Table 4) and static load test (Table 5 and Figures 3 and 4) are shown below.
9
3
(W
262
According to the results obtained from the static load test, it was possible to calculate the values of a and K in this particular case, for the first and second layers. These parameters are used in the Aoki & Velloso (1975) formula to predict the ultimate loads of the piles (Table 6).
7. RESULTS
Table 2. Results obtained froin the static load tests Pile Analysis QI"(kN) QPU ( W 1 172 36 2 CAPWAPC 20 1 15 3 178 62 1 2 Case
Ql,
Table 6. Values for a and K calculated henceforth the results froin the static load test. Layer a K First (0 to 6in) 3.56% 389 kPa 8.58% 356 kPa Second (6 to lam)
(kN) 208 216 240 213" 248"
Qil
7<+*
,522
1 2
200" IPT-Case 243w 3 224* *Numbers obtaincd from thc a\ cragc of thc last bloM s apphcd to the pile J=O 4 was selected for the analysis by Case because it represents the average value obtained by CAPWAPC. in thc order of 0 2 for the meakenmg of tlie edge. added 0 2 (according to the TPT (1991) that indicates it for tlie cases m which RMX values are used)
The dynamic tests were interrupted before the lOmm/lOblows, usual for this kind of pile, had been reached, since the piles had pre-determined lengths (Table 3). Table 3. Results of the end rate of penetration, obtained from the dnving of the three piles. Pilc 1 Pilc 2 Pilc 3 3 3 1 d 1 0 blows 1lOitun/IO blows 1251m/10 blows
Table 4. Results for the ultimate load. obtained froin the Reboundmeter for the piles 1 and 3 Pile Method Qii (W 1 Uto* 346 3 319 1 Chellis-Velloso 186 3 175 *used for piles with point in sand.
8. ANALYZING THE RESULTS
The values for the ultimate load (Q,,), obtained from the analysis CAPWAPC (208, 216 and 240kN), Case (213, 248 and 255kN) and IPT-Case (200, 243
679
and 224kN), from piles 1, 2 and 3, respectively, were analyzed individually for each pile. The results obtained from the three analysis for pile 1 were very close. For piles 2 and 3 the results were a bit farther than for pile 1. It was verified that the values of QI, obtained by the Rebound Meter are very different for the two piles analyzed (1 and 3) among the analyzed formulas (Table 3). Comparing these results with the ones obtained from the static load test, it is possible to observe, therefore, that the Uto et al. (1985) formula offered superior values, around 330kN. Regarding the Chellis (195 1) method, modified by Velloso (1985), this has presented values of a lower level, around 180kN. Nevertheless, in this case the Chellis (195 1)Velloso (1985) formula must be used, because the Uto et al. (1985) formula is valid for piles with point in compact sand. Based on the data shown in Table 2, it is possible to observe that, related to point load, skin friction and ultimate load, the results obtained from static and dynamic load tests were very close. However the same cannot be said about the methods obtained through the rebound, because they present large variation.
9. CONCLUSIONS
The dynamic load test through the CAPWAPC, Case and IPT-Case analyses, has shown reliability as a means to determine pile side and point loads. The value of ultimate load obtained from the dynamic load test was placed 10% below the one obtained from the static load test. Although the Rebound Meter System did not indicate results comparable to the ones from analysis CAPWAPC, Case and IPT-Case, shows great potential for utilization, mostly due to its simplicity and low cost. However, an improvement in its calculation formulas is required. It is important to emphasize that the Uto et al. (1985) formula is only valid for piles in sand. REFERENCES AssociagBo Brasileira de Normas Tecnicas 1991 MB - 3472191 Esfacas - Provas de Carga Estbtica. Rio de Janeiro, AssociagBo Brasileira de Normas Tecnicas, 4p. Rio de Janeiro: A.B.N.T.
680
Albuquerque, P J R 1996 A~idhse do Contportanwito de Estaca Pre Moldadu de Peqrreiio Dicintetro, Insfruvneizfnda em Solo Residual de Diabasio da Regifio de Canipmas DissertagBo de Mestrado 154p Campinas Feagri-Unicamp Aoki, N & Alonso, U R 1989 Correlation Between Different Evaluation Procedures of Static and Dynamic Load Tests and Rebound In XI1 I C S M F E, Rio de Janeiro Proceedrngs...,v 2, p 1 1 15- 1 1 16 Aoki, N & Velloso, D A 1975 Urn Metodo Aproximado para a Estimativa da Capacidade Cotigresso de Carga de Estacas In Parianterrcaiio de Mec&ica dos Solos e Eqqenharm de Fimda@s, V, Buenos Aires, Arms. .., p 367-376 Aoki, N 1986 Co’ontroleIrA’itu da Capacidode de C’arga de Estacas Pre-Fabricadas L’la Repiqire Elbstico da Crava@o SZo Paulo ABMS, ABEF e IE-SP Aoki, N 1989 A New Dynamic Load Test Concept In XI1 I C S M F E, Discussion Session 14, Technical Committee on Pile Driving, Drivability of Piles, 1989, Rio de Janeiro Proceedings..., v i, p 1-4 Aoki, N 1991 Carga Admissive1 de Estacas Atraves de Ensaios Dinimicos In Seminario de Engenharia de Fundagaes Especiais I1 - SEFE 11, SBo Paulo, Anais ..., v 2, p 269-292, SP Aoki, N , Alonso, U R & Trindade, 0 A 1990 Aplicaqiio de Registrador de Deslocamento Diniimico (RDD) na Avaliagiio da Carga Mobilizada em Estacas Cravadas In Simposio sobre InstrumentagBo Geotecnica de Campo SINGE0 90, Rio de Janeiro Anais..., p 45-51 Chellis, R D 1951 Pile FourzdatronA. TheoryDesrgti-Pratrce New York McGraw-Hill Book Company lnc IPT - Instituto de Pesquisa Tecnologica do Estado de Siio Paulo SJA 1994 Ensnios de Carregamento Dinbniico iia Faculdade de ErzgeiTharra Agricola da Uriicantp, em Estacas Pre Moldadas de Concreto Relatorio no 32 075 Siio Paulo IPTDCC - AMSFEICC Machado, J R A 1995 A Avaha@o da Capacidode de Carga de Estacas, com Base 110 Repique Elastico Medrdo no Final da Crnvap?o DissertagBo de Mestrado 265p SBo Paulo EPUSP Monacci, M D 1995 Estudo da Colapsibilrdade de unz Solo do ramp0 Expertnwztal da Facirldade de Eigenharia Agricola - Urircamp.
Dissertaqiio de Mestrado. 114p. Campinas: Feagri-Unicamp. Uto, K., Fuyuki, M., Sakurai, M. 1985. An Equation for the Dynamic Bearing Capacity of a Pile Based on Wave Theory. In: International Symposium on Penetrability and Drivibility of Piles. San Francisco-USA. Proceedirigs..., v. 2, p. 95-100. Velloso, P. P. 1987. Fiinda@es - Aspectos Geotecnicos. 5a ed., NA 01/82. Rio de Janeiro: DECPUC -RJ.
681
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Strain dynamic testing on pressure-grouted piles Liu Xi-An & Zhang Yao-Nian Fujian Academy of Building Research, Fuzhou, People's Republic of China
ABSTRACT Pressure grouting can greatly increase bearing capacity of bored pile and improve shaft integrity. The grouting effectiveness is evident by performing high strain and low strain dynamic tests. Results of some actual cases are presented in the paper.
1
INSTRUCTION
The bearing capacity can be increased by injection of, cement mix with bored pile a water/cement ratio of 1/1 to 0.5/1 at the bearing end. The grouting pressure is normally 2.5MPa to SMpa, and in some cases can be as high as 10 MPa. In order to control the solidify time and grouting area, some admixture is added to the ceineiit mix. Both the end bearing capacity and shaft resistance can be increased significantly after grouting treatment. The total pile bearing capacity can be increased by more than 30 percent in most cases, and the largest increment observed was 200 percent. The grouting can be used to treat the pile faults such as pile end slime and shaft concrete segregation. In such cases, high-pressure water is used to clean the faulty parts first. Then the grouting of cement mix is performed. The treatment can lead to satisfactory result for moderate extent of faults. The load transfer behavior of grouted pile can be determined by high strain dynamic test and by further analysis of CAPWAP. The result of grouting on pile faults can be understood using low strain dynamic test, which identifies the pile faults and their change due to grouting.
In the first stage, both Ta and Tb Mere l m in diameter and 58m in length. The bearing stratum of the piles was medium weathered granite. The results of static loading tests led to the conclusion that the bearing capacity of Ta and Tb were 8000kN and 9000kN with corresponding displacement of 12.73mm and 14.71mm, respectively. In the second stage, five piles of 800mm in diameter and of length between 35.0m and 36.lm, which were treated by pressure grouting. The bearing stratum of the piles was gravel miued with clayey soils. All of five piles were loaded 10 90001
2 HIGH STRAIN DYNAMIC TEST A case of pile foundation project in Fuzhou, China is introduced. The building is 30 storeys with two basements. The height of the building is 99.8m above the ground level. The total building area is 47000in2. Large diameter bored piles were used for the foundation. The static loading tests of the piles were carried out in two stages.
Fig.1 .The static test load-settlement curves.
Fig.3 The distribution of axial forces measured by static tests for T4. The total load of pile top was 9000kN, and the total toe resistance was 2080kN, and the total shaft resistance was 6920kN.
Fig.2 The distribution of axial forces measured by static tests for T2. The total load of pile top was 9000kN, and the total toe resistance was 2310kN, and the total shaft resistance was 6690kN.
CAPWAP FINAL RESULTS Total
CAPWAP
Soil
Depth
Capacity:
12175.5; along
Shaft
2056.2; at
Toe
3119.4
IcN
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
"%E E$: 1 2 3 4
m 2 . 0
6 7
4.0 6.0 8.0 10.0 12.0 14.0
9 10 11 12 13 14 15 16 17
18.1 20.1 22.1 24.1 26.1 28.1 30.1 32.1 34.1
5 R -
1G
0
Depth
Ru
EES:m
kN
2.0 4.0 6 . 0 8.0 10.0 12.0 14.0 16.0 ~18.1 20.1 22.1 24.1 26.1 28.1 30.1 32.1 34.1 ~
~ v e r a g eSkin Values Toe
37.4 150.2 348.0 397-8 417.6 412.0 321.8 454.4 578.8 633 - 7 647.3 679.1 705 - 3 781 - 8 787 - 5 835 - 2 868 - 1
Force i n Pile at R u ICN
Sum
of
kN
12175.5 12138.1 11987.9 11639.9 11242.1 10824.5 10412.5 10020.6 9636.3 9057 4 8423 8 7776 - 4 7097.3 6392.0 5610.2 4822.7 3987 - 5 3119 - 4
~
Ru
37.4 187.6 535.6 933.4 1351.0 1763.1 2084.9 2539.3 3118.1 3751 8 4399 1 5078 - 2 5783 - 6 6565 - 3 7352.8 8188.1 9056.2 ~
~
Unit Resist. Smith w. Respect to D a m p i r i y D e p t h Area Factor kN/m kN/m2 s/m 18.65 74.88 173.51 198.29 208.21 205.41 160.45 226.51 288.56 315.91 322.71 338.56 351.64 389.73 392 .59 416.40 432.78 265.58
532.7
6.98 28.01 64.91 74.18 77.89 76.85 60.03 84.74 107.95 118 - 1 9 120.73 126 66 131.55 145 - 8 0 146.87 155.78 161.91
-
99.36 5482.17
3 119.4
Fig.4 The CAPWAP final results of T2.
684
.547 -547 -547 -547 -547
<Xiake
mm 3 . 0 0 0
3 . 0 0 0 -3.000 3.000 3 . 0 0 0
.547 -547 .547 -547
2 . 8 6 6 2.550 :-!.355 149 1 970 1 818 1 684 1 561 L 445 1 206 1 094 1 000
-
547
1
87f3
.721
I
200
.547 .S47 -547 .547 .547 .547 .547 .547
CAPWAP
FINAL
RESULTS
CAPWAP Capacity: 11838.7; along _ . ._.- -. -. -.- ._ ._.- ._.- -. _. _ . _ . ._._._. _. = = _ _ - -_- _ - -_= = = _ _ _ _ _ _ _ _
Total
Soil Sgmnt NO.
Depth
Depth
Below
Gages
Below Grade
m
m
Ru IcN
Force i n Pile at Ru kN
1x838 7 1 8 - 8 11819.9 75.0 11744.9 248.6 11496.3 400.7 11095.6 3 3 3 9 1 0 7 6 1- 7 3 9 5 . 4 10366.3 391.8 9974.5 480.4 9494.1 626.4 8867.6 750.4 8117.2 781-7 7335.5 813.0 6522.6 844.2 5678.3 876 1 4.802 - 2 829.9 3972.3 938.0 3034.3
Average
3.0 5.0 7.0 9.0 11.0 13.0 15.0 17.0 19.0 21.0 23.0 25.0 27.0 29.0 31.0 33.0
3.0 5 - 0 7 - 0 9.0 11.0 13.0 15.0 17.0 19.0 21.0 23.0 25.0 27.0 29.0 31.0 33.0
Sum
Skin Values
IcN
-
Toe
~
~
~
550
3034
-
Ru
18 a 93.8 342 - 4 743.1 1077.0 1472.4 1864.2 2344 6 2971.1 3721 5 4503.2 5316.2 6160 - 4 7036.5 7866.4 a804 4
~
-
3 3
at
8804.4;
Toe
3034.3
I ~ N
-===___-__________________:===_
of
~
1 2 3 4 5 6 7 8 9 1 0 11 12 13 14 15 16
Shaft
Unit Resist. Smith w . Respect t o Damping FactOL' Depth Area kN/m lcN/m2 s /m
Quake
9.38 37.52 124.29 200.36 166.96 197.70 195.88 240.22 313.22 375.21 390.85 406.48 422.12 438.06 414.94 469.02
3.50 14.00 46-38 74.76 62.30 73.77 73.09 89.63 116.87 140.01 145.84 x.51.67 157.51 163.45 154.83 175.03.
2 -500 2 -500
-57-7
2.500 2 . 5 0 0 2 -500 2 .so0 2.500 2.500 2 .so0 2 . 5 0 0
-577
2.500
266.80
102.66
-5.77
2
5323.32
-815
1
-577 -577 -57-7 -577 -577 -577
-577 -577 -577 -577 -577
-577 -577 -577
mrn
a .so0 2.500 a -500 2.500 2 -500
.so0 .so0
Fig.5 The CAPWAP final results of T4 The fitting results of CAPWAP were shown as Fig. 4 and Fig. 5, and reasonably closed to the these of static loading tests. The shaft resistance near the pile end was comparatively high, which characterized the positive effect of pressure grouting. It was noted that the toe resistance analyzed by CAPWAP were larger than that by static loading tests. This was simply due to the fact that in the static loading tests the piles were not loaded to the ultimate state, so the toe resistance capacity was not mobilized completely. 3
comparative results of low strain dynamic test before and after pressure grouting are as follows. 3.1 The dynamic test
for
shaft.
LOW STRAIN DYNAMIC TEST Fig.6 The curve of low strain dynamic tests for pile L74-4 indicated the break at the location of 5.5m from pile top (shown in first curve). The defect was verified by boring sample test with the break at the location of 5.6m to 6.8m, which was the mixture of gravel and slurry mortar. The second curve showed the improvement of pressure grouting.
Pressure grouting not only can strengthen the bearing stratum but also improve the integrity of pile. Using PIT tester produced by PDI Corporation, the locations of the defects, the end slime of pile or the strength of bearing stratum can be determined reasonably accurate. The results of the tester clearly indicated the effectiveness of pressure grouting. The
685
Fig.7 The PITWAP final results of pile L74-4 indicated the break at the location of 5.5m from pile top (shown in first curve). The second curve showed the improvement of pressure grouting.
Fig.8 The curve of low strain dynamic tests for pile L80-1 indicated the severe segregation at the location of 2.5m from pile top (shown in first curve). The defect was verified by boring sample test at the location of 2.5m with the concrete clipped with clayey slurry mortar. The clayey slurry mortar was substituted by pressure grouting shown as the second curve.
The comparative curves of piles before and after pressure grouting can be seen as follows.
Fig.9 The curve of low strain dynamic tests for pile L2 1-2 indicated that the pile was set in soft soil (shown in first curve). The situation was verified by boring sample test that the end of pile was placed in the residual clay of 7.33m thickness. not in the medium weathered granite required by the design. Pressure grouting (shown as the second curve) obviously strengthened the bearing stratum.
3.2 The dynamic test curves for pile end. The pressure grouting not onIy apply to the soil surround the pile end but also to the soils along the pile shaft, which combines the pile and the soil as a whole mass. The velocity curves of grouted piles collected by PIT tester show the improvement around pile shaft and at the end of pile.
Fig.10 The curve of low strain dynamic tests for pile L43-4 indicated that the bearing stratum had low strength obviously (shown in first curve). The situation was verified by boring sample test that the end of pile was placed in the residual clay with
686
4. CONCLUSIONS
thickness of 8cm and with thickness of 32cm strongly weathered granite to the medium weathered granite. Pressure grouting (shown as the second curve) obviously strengthened the bearing stratum.
The pressure grouting to pile foundation is a new technology. It enhances the axial bearing capacity of a pile significantly. The effectiveness of pressure grouting can be analyzed by the results of static loading test, the combination of high strain dynamic test and low strain dynamic test. REFERENCES Xi-an Liu, 1998, “The Pile Testing in F L I Z ~ O U International Airport”, Proceedings of the THIRD INTERNATIONAL GEOTECHNICAL SEMINAR ON DEEP FOUNDATIONS ON BORED AND AUGER PILES, Ghent, Balkema, Rotterdam, pp28 1-284.
Fig.11 The curve of low strain dynamic tests for pile L49-1 indicated that the bearing stratum had low strength obviously (shown in first curve). The situation was verified by boring sample test that the end of pile was placed in the residual clay with thickless of 9.2gm instead of medium granite according to design. Pressure grouting (sl1own as the second curve) obviously strengthened the bearing stratum.
si-an Liu, Ping-Hui Zhuallg and Do11g-bo Zhong, 1996, “Distinguish Rock-Socketed Status of pile Toe ’Irain Testing”, Proceedillgs of the FIFTH INTERNATIONAL CONFERENCE ON THE APPLICATION OF STRESS-WAVE -rmow TO PILES, Orlando, Florida, pp.607-611.
Fig.12 The curve of low strain dynamic tests for pile L49-4 indicated that the bearing stratum had low strength obviously (shown in first curve). The situation was verified by boring sample test that the thickness of slime of pile end and residual clay was 39cm. Pressure grouting (shown as the second curve) obviously strengthened the bearing stratum. All of the above cases were chosen from a pile foundation of Changle international airport in Fuzhou, China. The piles are 800mm in diameter and socketed into medium weathered bedrock 1.5m or 0.5m into slightly weathered bedrock. The working bearing capacity of the pile is designed to be 4MN. The low strain dynamic tests were carved out most defects of the piles. Core sampling was conducted for some moderate and all the severely defective piles. Pressure grouting was applies to these defective piles before pile cap construction. The total settlement of the building observed to this date has been not more than lOmm until now.
687
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Assessment of the interface between dynamic and rapid loading tests Madan B. Karkee Department of Architecture and Environment System, Akita Prefectural University,Honjo, Japan
Yoshihiro Sugimura Department of Architecture and Building Science, Tohoku University,Sendai, Japan
Takashi Horiguchi GEOTOP Corporation, Tokyo,Japan
ABSTRACT: This paper is concerned with the investigation and identification of indicators representing the characteristics of rapid and dynamic loading tests. Attempt is made to evaluate the range of values of such indicators from actual loading test observations so as to depict the interface between rapid and dynamic tests. To provide certain consistency, the data from the rapid and dynamic tests utilized in the investigation were low pass filtered to remove components in the frequency range higher than 1OOHz. Three of the parameters that may be obtained directly from the measurements during tests are found to be effective for representation of the interface between the two types of test. The concept of normalized wave length is introduced and utilized for the purpose of investigation. Comparison of the load transfer characteristics in static, rapid and dynamic tests indicate relatively closer resemblance of rapid loading test to static behavior. In contrast the dynamic test tends to be distinctly different. In addition, the extent of resemblance between rapid and static loading is adequately reflected in the three simple parameters proposed. The results of this research are expected to find useh1 practical application in the implementation of rapid loading test in practice. 1 INTRODUCTION The rationality of and the continuing need for relying on loading test results while taking critical design decisions concerning pile foundations is evident from the recent studies devoted to investigation of test results (e.g. Bruno and Randolph 1999, Danziger et al. 1999 etc.). It is usual to employ different types and categories of loading test methods (Karkee et al. 1997) in practice. Irrespective of the method employed, evaluation of the static load bearing behavior of the pile is the desired outcome in practice. When the loading itself is not static, the measured response requires fbrther analytical processing to estimate the static response. Consequently, the choice of loading method generally involves a tradeoff between higher cost of direct measurement by static test and a measure of ambiguity that may be unavoidable in the indirect estimation through dynamic or rapid loading tests. The obvious alternative to static test used to be the dynamic test, which was originally developed as an extension of the pile driving operation (Smith 1960). With the predominance of installation methods other than driving (Karkee 1999), the dynamic test is not limited to driven piles alone any more. In addition, methods of load application intended to lie in a domain intermediate between static and dynamic, such as the Statnamic test
(Middendorp et a1 1993) and pseudostatic test (Schellingerhout & Revoorte 1996), have also been developed These methods are categorized as rapid loading tests (ASTM, 1999) The availability of the different categories of loading test provides for prudent selection of a method suitable for a given situation However, it also means proper distinction between such categories based on the differences involved Considering the close resemblance, the distinction between dynamic and rapid loading tests is particularly important Disregard of the stress wave phenomenon in case of rapid loading test while estimating the static load transfer behavior of the pile may be justified based on such distinctions In other words, clarification of the interface between dynamic and rapid loading tests is crucial for usefbl implementation of rapid loading test in practice The rapid loading considered here is the Statnamic test Considering the similarity in the nature of load application, it is not surprising that some degree of vagueness exists regarding the interfaces between dynamic and rapid loading test methods Identification of clearly defined indicators of how the rapid loading condition differs from that of dynamic loading condition may be usefully utilized for defining such interface One such indicator proposed is the relative duration of loading, T, (Karkee & Kishida, 1999) The validity of T, is further investigated in this paper based on the nature 689
of frequency contents in rapid and dynamic loading tests on bored precast piles. The concept of the normalized wave length is introduced and the applicable range of the dimensionless frequency a0 is also investigated. The extent of the dynamic effect in load transfer through shaft resistance is compared based on the data from static, dynamic and Statnamic tests. The results may be usefbl in defining the basis for disregarding the stress wave phenomenon in rapid loading tests.
Other details of the different indicators are discussed subsequently in the respective sections that follow, where attempt is also made to evaluate their suitability in representing the interface between dynamic and rapid loading tests. 3 TEST SITES AND STATIC LOAD TRANSFER
Three types of loading test results, including static, rapid (Statnamic) and dynamic, are utilized. Both the static and dynamic tests were conducted at the site shown in Figure 2, where the ground condition, steel pipe pile profile and its instrumentation details are depicted. Circled numbers indicate sections along the pile where strain is measured. The shear wave velocity VS profile in Figure 2 is estimated from the correlation with standard penetration test N-values (Japan Road Association, 1990) given by:
2 INDICATORS OF THE INERFACE It is evident that the interface between dynamic and rapid loading tests has to be evaluated depending on how the pile body and the pile soil interface respond to the respective loading methods. Basically, the nature of loading method employed is reflected in the dynamic characteristics of the resulting load and velocity time histories recorded during the test. Prospective indicators considered for investigation are also derived from the time histories, and include: (a) relative duration of loading T,, (b) average rate of loading re, (c) normalized wave length ;1’2L~,(d) dimensionless frequency an, and (e) time lag between load and movement peaks t,,,.Here, 3, is the wave length and LP is the pile length. Discussions are based on actual observations from loading tests.
vs
={
80N1I3 (Sand) 1 0 0 ~ (clay) ~ 1 ~
The relation of the static load POapplied at the pile head and the strain measurement at level 0 is given in Figure 3. Approximately a straight line relationship may be noted except at very high strain. The Young’s modulus of the steel pipe material ES is found to be about 2 . 1 5 105 ~ MN/m2 based on the straight line approximate shown in Figure 3. The value of Es is quite similar to what may be expected for steel, indicating validity of the observations from strain measurement.
Fibwe 1. Definition of the duration of loadmg and the rise time
Here, T,. and re depend on the values of the duration of loading t d and the rise time t e respectively, which are defined as illustrated in Figure 1. The idea of the factor a (cl.0) for estimation of t d was introduced by Karkee and Kishida (1999). In practice, the value of a may be considered to lie in a range such as 0.7 to 0.9. The relations for T,. and re are given by:
T )*
y
=Ctd 2Lp
(1)
Figure 2. Static and dynamic loading test on a steel pipe pile
=-aF
(2)
3.1 Load transfer relation IH static loading
te
where, c is the stress wave velocity of the Pile material and aF is the load increment in time fe.
Strain was measured at three levels as shown in Figure 2, and the corresponding shaft resistance z, and local vertical movement w, relations for the 690
segments ill-0and 0-0are given in Figure 4. The z,-w, load transfer relation is approximated by bilinear representation shown by the solid lines in Figure 4. Similarly, the load transfer hnction relating the pile toe resistance PP and pile toe movement w p is shown to be approximated by a hyperbolic curve in Figure 5 . It may be noted that the maximum pile toe movement in the static loading test is relatively small. Consequently, the mobilized toe resistance is not substantial and the data points in Figure 5 seem to lie in a straight line. Thus the hyperbolic curve approximation also serves to extrapolate the load movement behavior at the pile toe. In contrast, the pile shaft load transfer seems to be h l l y mobilized in Figure 4.
3.2 Cotfifirnmtionof the load transfer.relntioris The load transfer relations of Figures 4 and 5 appear to be quite approximate in nature To confirm the adequacy of such approximation, load transfer analysis (Hirayama, 1990) was carried out to compute the vertical movements at different load levels at the pile head. As may be noted in Figure 2 that the vertical movements were directly measured at the head (WO), midpoint of pile (wm) and pile toe 69 1
(wp).The corresponding movements from the load transfer analysis are compared with the measured values in Figure 6. Fairly good agreement may be noted. The value of ES obtained from Figure 3 was utilized in the analysis.
3.3 Rnpid (Statnnmic) loading test sites The rapid loading test data utilized are from the Shonan test site in Japan (Karkee et al. 1995). A number of tests were carried out in 1995 on the 7m long bored PHC pile shown in Figure 7. Details of the steel pipe pile installed by driving are shown in Figure 8. Statnamic test was conducted on this pile in 1995. Both the piles were at the same site, as may be noted from the ground condition. Again, the VS profiles in Figures 7 and 8 are obtained from Equation 3. Altogether, the results of 3 repeat tests on the pile in Figure 7, and one test on the pile in Figure 8, are utilized in this investigation.
time history to compute the frequency content. To provide a certain consistency in the comparison between different tests, components with frequency higher than lOOHz were removed in all the data by low pass filter. 4.1 Norninlized wnve length The Fourier transforms of the load and velocity time history recorded near the pile head provide the nature of the frequency components involved with rapid and dynamic tests. Since the wave length A of these components is more reflective of the nature of stress wave transmission along the pile length, it is logical to express /z in terms of c as.
where, f is the frequency. Nature of transmission of the component of a certain wave length ;1 along the pile length is basically determined by how large 3, is in relation to the pile length L p . Thus it would be useful to normalize /z by L p . Since 2Lp already appears in Equation 1 in consideration of the importance of the time taken for the stress wave to travel down and up the pile, it is logical to define the normalized wave length as A 2Lp. The Fourier spectra components can also be normalized by the peak spectra. Figure 9 shows the plot of the normalized velocity spectra S I - S , in~ ~ terms ~~~ of~ /z 2Lp for the dynamic and Statnamic tests described above. It may be noted that the spectral velocity components of dynamic and Statnamic tests tend to form distinct groups, with normalized wave length of about 10 to 20 forming as the interface
3 4 I)L’tmniic tesi oil piles and site conditions One of the dynamic test data utilized in the analysis is from the steel pipe pile in Figure 2. Three other test data from 39m long bored PHC piles at a site in Tokyo area, where the F
Figure 9. Nomiallzed spectra and nornialized nave length
It may be noted in Figure 9 that most of the dominant components in the Statnamic test have the normalized wave length of 20 to 200. In comparison, 692
the dominant components in the dynamic test have normalized wave lengths in the range of 0.5 to 10, indicating a fairly clear distinction.
value of the shear wave velocity VSat the upper part of the site where the tests were conducted, the value of a0 can be easily computed fr0m.f:
4 2 Relative dirratiori arid average rate of loading Figure 9 clearly shows that a certain range of normalized wave lengths are involved in the velocity time histories from the Statnamic as well as the dynamic tests. The normalized wave length A,,, of the peak component S I ~ can, , , ~ however, ~ be determined easily from Figure 9. Here, A,,,is the normalized wave length corresponding to the peak spectral velocity, and is referred as the relative wave length
Figure 11. Dimensionless frequency content in loading tests
The spectra of the load time history SP normalized by the peak spectra Sp,,, is plotted with respect to no in Figure 11. Again a distinct difference in the range of no may be noted between dynamic and Statnamic loading tests. Effectively, apO.2 for Statnamic tests and a0<:3.Ofor dynamic tests seems Figure 10. Relative duration & average rate of loading to be a reasonable range based on the loading test data under consideration, indicating a difference of Figure 10 shows the variation of T,- computed with alO.75 in Figure 1 with the values of PL,,, the order of 10 in the applicable range. obtained from Figure 9. The data points for T, denoted by open marks tend to fall on a straight line, 4.4 Time lag betweer?load arid movement indicating that T,. and AL,,, tend to have a proportional One of the effects of the using dynamic and rapid relationship. It is apparent from the tendency that the load application in loading tests is the development relative duration of loading T,. serves as a suitable of the lag time between load and movement. That is, indicator of whether a loading test falls into rapid or the movement peak occurs after a certain time after dynamic category. This is advantageous because the load peak. This is illustrated in Figure 12 for unlike L,,,T,. can be determined quite easily from the typical dynamic and Statnamic loading cases. time history record. The average rate of loading re is also computed based on the value of rise time te with a=O.75 in Figure 1, and is plotted in Figure 10. Compared to the very high re (>800MN/s) in dynamic tests, its value is quite small (7 to 28MN/s) in Statnamic tests, indicating the difference of an order of magnitude 4.3 ~imerisionlessJi.eqz~eiicy
Another important parameter indicative of the nature of load application in rapid and dynamic loading tests is the dimensionless frequency a0 given as: wrQ 27r f r o a0 =-=-
vs
Figure 12. Illustration of time lag between load and movement
vs
It may be noted that the movement in dynamic tests is obtained from the acceleration by integrating twice and the accuracy may not be very good. The
where, r0 is the outer radius of the pile and o is the circular frequency. Considering the average 693
movement is, however, directly measured in Statnamic tests. The values of the time lag t,,. between load and movement are given in Table 1 along with other relevant parameters. The difference in between dynamic and Statnamic tests is not evident. That is, t,,. is not found to be usefbl for distinction between rapid dynamic loading conditions based on the loading test data considered in this study. Further study would be needed for more conclusive outcome.
from actual loading test This is because the soil shear modulus G degrades at higher strain level Conversely, for k,=I. 7kiVm' nini in Equation 7, the effective yalue of G works out ,to be about I IO5kN ~ i i -compared to 2 756JkN ni- at low strain level (a difference of about 25 times) Similarly, for the section 6-1of the test pile in Figure 2, where L's is again about 1 3 0 d s on the average, the effective value of the soil ,shear modulus G ,works our to be 6500kN/ni- (for k,=IO.OkNm- nini in Figure 4). Low strain shear modulus in this case is about 4 times the effective value, compared to 25 times in section 17-5 as mentioned above. Both the damping coefficient cs and the soil stiffness k, are derived directly from the effective shear modulus G as given by Equation 7 Consequently, the selection of a suitable value of shear modulus, applicable to a given loading test, dynamic or rapid, becomes a critical decision in the analysis of the load transfer behavior
5 LOAD TRANSFER RELATIONS Another important aspect to be considered in investigating the different loading conditions is the difference in load transfer fbnctions. The load transfer fbnction under loading conditions in which the velocity is not negligible has been investigated by Randolph (1990) and Danziger et al. (1999). The relation for unit shaft resistance zhas been given in terms of the soil stiffness k,sand damping coefficient cs as: Z = k,Tv+C,-
mu
(6)
fit
k,s = 2'9G and c , 2nro
=
JG'P
=Vsp
Figure 13. Load transfer in shaft resistance in dynamic loading
(7)
where G is the soil shear modulus, ro is the pile radius and p is mass density of soil. In what follows, some investigations are made concerning the difference that may be expected in the behavior of load transfer by shaft resistance. 5.1 Load transfer in dynamic and static tests
As mentioned above, static as well as dynamic loading tests were conducted on the steel pipe pile shown in Figure 2. The unit shaft resistance obtained from the static test is given in Figure 4, where the bilinear load transfer relation is found to represent well the observed behavior. It is also seen that k, is about l.7kN/mZ/ntmat the upper part of the ground (section E-Cl) where VSis about I30mis in Figure 2. Considering G = ~ ( V Sat) ~low strain level? the value of k, works out to be about 42.5kN/m-/mmfrom Equation 7, much larger than the value obtained 694
Assuming that the velocity measured near the pile head in the dynamic test on the pile in Figure 2 acts directly to mobilize the shaft resistance at the upper part of the pile, the unit shaft resistance may be obtained from Equation 6. For this, the values of k,y and c , ~ based on the value of G estimated from the static test can be utilized. The shaft resistance thus obtained is compared with that of the static loading case in Figure 7. Large difference in the shaft resistance behavior between dynamic and static loading, in terms of the initial slope as well as in terms of the magnitude, may be noted. In Figure 4 the static shaft resistance tends to increase up to a movement of more than 20mm. In contrast, the maximum movement affected in the dynamic test is only about 8mm. However, there appears to have been a clear permanent movement of the shaft at the end of the dynamic loading cycle in Figure 13, in spite of the relatively small maximum movement. The result has very important implication, because the maximum shaft movement at which a permanent movement may be expected in
a dynamic test may not be coincident with the movement at which the shaft resistance is fully mobilized in static test. This aspect needs to be adequately reflected in defining the soil model in the analysis for dynamic test.
Similarly, Figure 15 compares the trend of load transfer relation in Statnamic and static tests for the test on steel pipe pile in Figure 8 (test no. 4 in Table 1). The two load transfer relations are noted to be even more similar to one other when compared to those in Figure 14.
5.2 Load trau.fer in Statnaniic tests Following the interrelation between k, and cs noted above in Equation 7, it may be noted that the dynamic and static load transfer functions are basically determined by the effective value of G applicable to a given loading test situation. The movement w at the pile head is directly measured in the Statnamic tests depicted in Figures 7 and 8. The velocity response dvldt is obtained from it by differentiation. Now, knowing that the low strain shear modulus close to the surface a; the test site works out to be about 36697kN/ni-, it may be reckoned from the static test discussed above that the effective value of G would be a certain fraction of this value. Once this fraction is determined, the initial slope of the static load transfer function as well as the Statnamic load transfer fbnction may be defined by Equation 7. Assuming the fraction, noted above for the static loading test depicted in Figure 2, to be applicable here as well, the load transfer relations for the Statnamic tests (no. 1 and 4 in Table 1) may be obtained as shown in Figures 14 and 15.
Figure 15. Sl& resistance in Statnamic test on Steel pipe pile
It would be interesting to see if the indicators in Table 1 reflect the better resemblance of Statnamic test with the static load transfer trend in Figure 15. Table 1 shows that test no. 1 (Figure 14) has smaller values of A,,, (25.5) and T,. (12.96) and higher rate of loading re (13.3MWs), compared to the (38.6), T,. (26.48) and I’, corresponding values of A,,, (9.1MN/) for the test no. 4 in Table 1 (Figure 15). This provides further verification that the indicators, proposed here for representation of the interface between dynamic and rapid loading tests, logically depict the trend towards resemblance or otherwise to the static loading. That is, the three simple parameters may be suitably utilized to see if a loading condition is closer to conventional dynamic test or to the rapid loading test with relatively closer resemblance to the static loading condition. 6 CONCLUSIONS
Figure 11. Shaft resistance in Statnarnic test on PHC pile
Figure 14 gives the relative comparison of the initial slope of load transfer function in static test with that in the Statnamic test on the pile shown in Figure 7 (test no. 1 in Table 1). It can be noted that the difference in magnitude as well as in the trend of the initial slope between the Statnamic and static tests is much smaller when compared to that between dynamic and static test shown in Figure 13. This is a clear indication that the load transfer in Statnamic test tends to be much closer to the static trend when compared to the conventional dynamic test. 695
Attempt has been made in this paper to propose suitable indicators representing the interface between rapid (Statnamic) and dynamic loading tests. The range of the values of such indicators are evaluated based on the actual loading test observations. To provide certain consistency in the comparison between several Statnamic and dynamic tests, low pass filter is used to remove components in the frequency range higher than 1OOHz. The concept of normalized wave length is introduced and utilized in the investigation. The load transfer characteristics in static, rapid and dynamic tests have been compared and investigated. Following conclusions are drawn from the study:
The relative duration of loading Tr tends to have a proportional relation with the normalized wave length A,,, at the peak spectral velocity component. Alrj is referred to as the relative wave length in this paper. 11. Three of the indicators investigated, namely the relative duration of loading T., the average rate of loading re, and the relative wave length A,,j, are found to be well suited to depict the distinction between dynamic and rapid loading tests. Certain range of values of these parameters may be defined to serve as the interface between dynamic and rapid loading tests. iii. Based on the loading test data analyzed in this study, the applicable range of the dimensionless frequency a0 is noted to be less than about 0.2 for rapid loading tests. In comparison, dynamic tests seem to involve a0 of up to 3, and may be more. 11'. The pile shaft load transfer relation in static loading test can be represented by bilinear approximation quite satisfactorily, as evidenced by the comparison between measured and computed load movement behavior at different points along the pile length. 1). The load transfer by shaft resistance in case of rapid loading test tends to be relatively closer to the static behavior when compared to the case of a dynamic test. VI. Different extent of resemblance in load transfer behavior between rapid and static tests can be expected in practice, indicating the need to look at the individual cases in detail before deciding to adopt a certain method of analysis for estimating the static load bearing behavior. 1'11. The extent of resemblance between rapid and static loading tests in relation to the load transfer knction is adequately reflected in the values of the indicators proposed in this paper. Closer resemblance with static loading may be expected in case of rapid loading test with larger values of relative duration of loading Tr, and the relative wave length A,,,, and a smaller value of the average rate of loading re. i.
ACKNOWLEDGEMENT The Statnamic data utilized here constitute part of the series of tests in 1994/95 organized by the 'Research Group on Rapid Load Test Methods'. of which the first author was a member. The authors wish to acknowledge the contribution of all the members of the research group to the research initiative that has helped substantially in developing better insight about the rapid loading test method.
696
REFERENCES ASTM 1999. Standard test method for piles wider rapid axial cotnpressive load (Designation: D 0000-99) Bruno. D. & M. F. Randolf 1999. Dynamic and static load testing of model piles dnven into dense sand. J. Geotecli. 61. Geoenv. Eizgrg., ASCE. 125(11): 988-998. Danziger. B. R.. A. M. Costa, F. R. Lopes & M. P. Pacheco 1999. Back analysis of offshore pile dnving with an improved soil model. Geotechnique, 49(6): 777-799. Hirayama, H. 1990. Load-settlement analysis for bored piles using liperbolic transfer functions; Soils & Foundations. VO1.30. No.1: 55-64. Japan Road Association 1990. Specifications for highway bridges, Part IV, Earthquake resistant design (in Japanese). Karkee, M. B. & H. Kishida 1999. Dynamic, intermediate and static loading tests and application of test results in design practice. Keynote paper: Proc. 3Ih Int. Col$ Deep Foundation Practice incorporating PILET.4LK '99. Singapore, 29-30 July 1999: 13-26. Karkee, M. B. 1999. Developments in low noise and low vibration methods of pile installation in Japan. Proc. 1Ith Asian Regional Col$ Soil Mechanics & Geotechnical Engineering, Seoul, Korea, Vol. I1 Karkee, M. B., H. Horiguclii & H. Kislida 1999. Limit state formulation for the vertical resistance of bored PHC nodular piles based on field load test results. Proc. I I " .4siaii Regional CO@ Soil Mechanics & Geotechnical Engineering, Seoul, Korea, 1999: 237-240. Karkee. M. B., H. Horiguchi & H. Kishida 1997. Static and dynamic tests for evaluation of vertical bearing capacity of piles. Deep Foundation Institute, 2?ld Aiinual Metriber's Conference, Toronto, Canada. 1997: 199-214. Karkee. M. B., 1. Kojima & Y. Slunoda 1995. Influence of different test conditions on the results of Statnaniic load test at Shonan test site. Proc. I"' Int. Statnaiiiic Seminar, P'aiicouver, Canada. 1995:137-147. Matsunioto. T. & S. Nisliiniura 1995. Wave propagation plienomena in Statnamic test of a steel pipe pile. Proc. 5'" Int. Conf. Application of Stress-wave Theor?: to Piles. Orlando. Florida. 1996: 1015-1030 Middendorp. P.. P. Bermingham and B. Kuiper 1993. Statanic testing of foundation pile. Proc. -!Ih Int. Con$ Application of Stress-wave Theoty to Piles, the Hague, 1993: 585-588. Randolph. M. F. 1990. Analysis of the dynamic pile driving. Developments in soil mechanics IV: .-lnlxmcet/ geotechnical analvsis. Elsevier Applied Science Publishers. Poulos. H. G. 1998. Pile testing - from the designer's point of view. Preliminary Proc. 2"" Int. Statnamic Seminar. Tokyo. Schellingerhout, A. & E. Revoorte 1996. Pseudo static pile load tester. Proc. 5'" Int. Conf. Application of Stress-wave Theon- to Piles, Orlando. Florida. 1996: 103 1-1037. Smith. E. A. L. 1960. Pile driving analysis by the wave equation. J. SoilMecli. Found. Div., ..ISCE, 86(1): 35-61.
Application of Stress-Wave Theory to Piles, Niyama & Beirn (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Dynamic load testing on 102 steel pipe piles for bridge foundations on mudstone M. Hayashi Applied Research Center, NKK Corporation, Kawasaki, Japan
T Matsumoto Department of Civil Engineering, Kaizazawa Universig,Japan
M. Suzuki Izurni Research Institute, Shirnizu Corporation, Tokyo,Japan
ABSTRACT: A total of 102 steel pipe piles were driven into diatomaceous mudstone on Noto Peninsula, Japan, for the foundations of a highway bridge. In this pile construction, dynamic load testing was carried out on all of the 102 piles. The results of the dynamic load testing showed that the bearing capacity of all the piles increased rapidly with elapsed time after the end of pile driving, and exceeded the design ultimate bearing capacity one day after the end of pile driving. The sufficient number of dynamic load tests to be conducted at a site to determine the distribution of pile capacity is also discussed in this paper. A non-parametric bootstrap method was employed with the use of the measured bearing capacity of the piles derived 6om the dynamic load tests. It was found that the execution of 10 or more dynamic load tests gives us the distribution of pile capacity that has a acceptable probability of failure which may be obtained from the dynamic load tests on all the piles for this particular site.
1 INTRODUCTION
The sufficient number of dynamic load tests to be conducted at the site to determine the distribution of pile capacity is also discussed in this paper. A nonparametric bootstrap method was employed with the use of the bearing capacities of the piles estimated from the dynamic load tests.
The construction of a new highway route was completed in 1998 on Noto Peninsula in Japan. Steel pipe piles were used for the foundations of the highway bridge named Noetsu bridge No.3. The construction site consists of a thick deposit of a soft rock called diatomaceous mudstone underlain by a sand rock. A total of 102 steel pipe piles were driven into the diatomaceous mudstone. Pile driving of foundation piles for an abutment (AI) and a pier (PI) was carried out in 1995 (Matsumoto et a1 1997). Another abutment (A2) and three other piers (P2, P3 and P4) were constructed in 1996. For these pile constructions, dynamic load tests were conducted on each of the 102 piles at pile installation (initial driving) and after a rest period from the end of initial pile driving (re-driving) in order to estimate the static load-displacement curve for the piles after the completion set-up phenomena. Rational soil models that have been proposed by Randolph & Simons (1986) and Deeks (1992) were used in the signal matching analysis of each dynamic load test. The soil model parameters except the shaft resistance and the toe resistance were estimated 6om the soil test results. The shaft resistance and the toe resistance alone were estimated through signal matching analysis. The validity of this wave matching analysis procedure was examined through the comparison of the results of the dynamic load test and the static load test on a test steel pipe pile.
2 SITE CONDITIONS Figure 1 shows the soil profile along the axis of Noetsu Bridge No.3. The construction site is characterized as a thick deposit of diatomaceous mudstone. The dynamic load testing and the static load testing of a steel pipe pile were carried out in 1991 near the construction site of the bridge (Figure 2), to obtain the design parameters for the steel pipe pile to be driven in the diatomaceous mudstone for the bridge (Matsumoto et a1 1995). Figure 3 shows the soil test results at the construction site and the test site. The variations with depth of the natural water content, w,,,the density, p, the unconfined compression strength, q,, and the secant modulus, E50, are shown in Figure 3. It can be seen that the construction site ground is relatively uniform in plane and in depth. Figure 4 shows the distributions of the SPT Nvalues and the shear wave velocities, Y,, measured by the seismic cone penetration tests. The shear wave velocities which are indicated by the dashed lines were measured at the test site, pier PI and abutment A2.
697
Figure 4. SPT N-values and shear wave velocities measured by seismic cone penetration tests at the construction site of Noetsu Bridge No.3 and the test site.
698
3 TESTPILING 3.1 Test pile
The test open-ended steel pipe pile was driven at the test site with a diesel hammer having a ram weight of 40.2 kN. Table 1 lists the specifications of the test pile. The test pile was instrumented with foil strain gages at a total of 10 levels in order to measure axial forces during the static load test. Steel channels were welded inside the pile for protection of the strain gages, which increased the net cross-sectional area, A , of the pile to 0.041 m2. The increase in the cross-sectional area was taken into account in the wave matching analysis by increasing the wall thickness (by decreasing the inner radius) of the pile. For dynamic load tests, two sets of a strain gage and an accelerometer were mounted 180' apart at a distance of 1 m from the pile head. The dynamic load tests were carried out at the end of initial driving (EOID) and 66 hours after the end of initial driving, and the static load test was performed 26 days after the re-driving test. Table 1. Specifications of the test L Length Ld Embedded length t, Wall thickness 43 Outer diameter 4 Inner diameter A Cross-sectional area Young's modulus E Bar wave velocity c
pile. (m> (m> (mm> (mm> (mm> (m') (MN/m') (m/s>
11.0 8.3 12.1 800.0 775.8 0.04 1 2.06 X IO5 5120
3.2 Procedure of the wave matching analysis Wave matching analyses of the dynamic load test signals obtained at the end of initial driving and the re-driving were conducted using a computer program KWAVE to estimate the static responses of the test pile. The computer program KWAVE (Matsumoto & Takei 1991) is based on the characteristic solutions of the wave equation and the pile-soil system illustrated in Figure 5. The spring value of the internal soil (soil plug) is estimated fiom the one-dimensional modulus, Eo, since the radial strain of the internal soil may be assumed to be negligible:
E,
=-
6), the slider represents the maximum shaft resistance, f , the spring represents the elastic deformation of the surrounding soil, k,, the dashpot in the upper potion represents the viscous damping,
2(1- v)G 1-2v
cv,and the dashpot in the lower potion represents the radiation damping, cr, respectively. According to Novak et a1 (1978), the spring value, k, and the radiation damping, er, of the outer shaft resistance were estimated as follows:
where G and v are the shear modulus and Poisson's ratio of the soil. Rational soil models that have been proposed by Randolph & Simons (1986) and by Deeks (1992) were used for the shaft and the base resistance in the wave matching analysis. In the shaft model (Figure
699
= &G= pV, =G/VS
(3) Although the shaft model shown in Figure 6 was used also for the internal shaft resistance, the radiation damping was set to be zero since the inward radiation does not occur in the soil plug. In the base model (Figure 7), the slider represents the maximum base (toe) resistance, q b , the spring represents the elastic deformation of the soil below the pile base, kb, the dashpot expresses the damping, cb, and the mass is the lumped soil mass, Mb, respectively. The soil parameters, kb, c b and Mb are given as follows for soil plug base and for the annular pile base (Deeks 1992): C,
respectively, and n is the porosity (n = 0.25 for the mudstone). The soil parameters except for the maximum shaft resistance, f, and the maximum base resistance, q b , were determined by means of Equations (2) through (9) with v = ves.The values off and q b alone were established by the wave matching. The static loaddisplacement curve was calculated using the identified f and q b , and the drained value of the base under static spring, kb, and the shaft spring, loading. The ks(stat,c) was estimated according to the following equations (Randolph 1991):
for soil plug base;
cb
3.3 Comparison of the dynamic and static load tests
3.2pV5 3.2G =n(1- v) n(1- v>y5
M b =2d?p-
The validity of the above procedure of the wave matching analysis was examined through the comparison of the results of the dynamic load test and the static load test on the test pile.
0.1 - v 4 1-v
for annular pile base; 8G kb = ~ ( 1 v- ) ( d , + d;) 3.2pVs __- 3.2G "b =n(1- v) n(1- v)V, 3 M b = 2(d; - di )p-
O.l-V4 1-v
where p is the soil density. The mudstone is hlly saturated (degree of saturation S = 0.995) and has a low permeability. Hence, the undrained condition is thought to be maintained during pile driving. In order to take account of this situation, the equivalent Poisson's ratio, v,,, which has been proposed by Verruijt (1969), was used, instead of the drained Poisson's ratio, v, of the mudstone ( v = 0.15), to estimate E0 and the base model parameters such as kb, cb and Mb:
where
K,,
= K,
+ Kf / n
(1 1)
where K,, Kf, K,., and K, are the bulk modulus of the soil skeleton, the pore fluid, the saturated water (K, = 2000 MN/m2) and the air (K, = 2000 kN/m2), 700
Figure 9. The results of the save matching analysis the redriving test.
Figure 8 shows the results of the wave matching analysis of the driving test performed at the end of initial driving. The upper figure (a) shows force vs time and the lower figure (b) shows velocity vs time. Similarly, the results of the wave matching analysis of the re-driving test are shown in Figure 9. The maximum base resistance, q b , was identified to be 1500 kN/m2 for both the initial and the re-driving tests. Figure 10 shows the distributions of the shaft resistance, J estimated from the initial driving and the re-driving tests and the shaft resistance measured in the static load test. Comparison of the distributions of ffrom the two driving tests clearly ...... End of initial driving (EOID) -Redriving (66 hours after EOID) Static Load Test (29 days after EOID)
Shaft resistance, f (kN/rn*) 0 1
:
:
'
100 "
200
300E.L.1.44m
I
Load on pile head, P (MN)
h
. g - 0
s u v
2
1
3
4
5
3 G @
5
10
0E
15
.-U)
20
a
U
25
30 _.
a
35
\
i -Static
',
(66 hours after EOID) Load Test I -
} , , : , ! , ! , , , I (29 days after EOID)
Figure 1 1. Load-displacement curves estimated from initial driving and re-driving tests, together with static load test result.
w
showed that about 80% of the ultimate capacity is the shaft capacity, which is comparable with the static load test result. 4 ACTUAL PILINGS 4.1 Pile specifications and test sequence A total of 32 piles were driven at abutment AI and pier PI in March, 1995. Note that 2 piles out of the 20 piles at abutment AI and 2 piles out of the 12 piles at pier PI had been driven 5 months before this pile construction for drivability assessment. A total of 70 piles were driven at abutment A2, piers P2, P3 and Pq in April, 1996. The geometrical and mechanical properties of the piles are listed in Table 2. Two sets of a strain gage and an accelerometer were mounted 180' apart at a distance of 1 m from the pile head for the dynamic load tests. In the pile construction work in 1995, 28 initial driving tests and 33 re-driving tests were conducted within 4 days without a delay of the usual pace of the pile construction work. The initial driving tests were performed on all the piles except the four piles which had been driven prior to this construction work.
-8
Figure 10. Distributions of shaft resistance estimated from initial driving and re-driving tests, together with static load test results.
indicates that a large set-up of the shaft resistance occurred during the rest period of 66 hours. The results from the re-driving test are comparable with the static load test results. The load-displacement curve derived from the wave matching analysis of the re-driving test is fairly comparable with the static load test result (Figure 11). The result of the wave matching analysis
Table 2. Properties and design ultimate bearing capacities of the foundation piles of Noetsu Bridge No.3. Design ultimate bearing Pile properties Abutments, Number capacities Qu(kN) Piers of piles Ordinary Seismic Embedded Outer diameter Length (m) length L,(m) do (mm> condition condition 9.0 AI 20 3161 8.0 600 3706 (upper 5.0 +lower 4.0) Pi 12 1598 2942 10.0 9.0 600 p2
p3
16
3787
3604
20
1585
2556
-~
1200 2624 18 2050 328 1 (for all piles) Young's modulus E = 2.06 p4
16
A2
X
Wall thickness 4\ (mm) upper 12 I lower 9 9 11.5 upper 12 I (upper 4.5+lower 7.0) 11.0 600 lower 9 upper 14 I 16.5 (upper 5.5+lower I 1.o) 16.0 600 lower 9 upper 12 I 14.5 14.0 600 lower 9 (upper 4.5+1ower 10.0) 8.5 8.0 600 9 10' MN/m2 , Density p = 7.86 ton/m3 , Bar wave velocity c = 5120 m/s
70 1
Figure 12. An example of wave matching for the re-driving test of a pile at abutment A,.
Several piles underwent multiple re-driving tests. In the pile construction work in 1996, 70 initial driving tests (one test for each pile) and 116 re-driving tests (several piles underwent multiple re-driving tests) were conducted within 13 days. 4.2 Wave matching results The procedure of the wave matching analysis described in 3.2 was used for the wave matching analysis of the dynamic load tests. Figure 12 shows an example of wave matching of a
pile at abutment AI. The estimated loaddisplacement curves of all the piles at abutments AI, A2 and at piers PI through P4 are shown in Figure 13. Although rest periods of the piles between the end of initial driving and the re-driving are different, the ultimate pile capacity at re-driving is quite a bit larger than at the end of initial driving. The estimated shaft capacity amounted to about 80% of the estimated ultimate capacity in all the piles. The increase in the estimated static resistance, Qu, of the piles at abutments AI, A2 and piers PI through P4 as a function of elapsed time after the end of initial driving are shown in Figure 14. It is seen that the set-up ceases at about 100 minutes for abutment AI, at 200 minutes for pier PI, at 1000 minutes for abutment A2 and pier Pz. The estimated static resistance after these time intervals exceeds the required ultimate capacity. However, for the pile at pier P3 and pier Pq, it was judged that the ultimate capacity was not attained, because the set per blow was very small, less than 1 mm. The frequency distributions of the estimated ultimate static resistance, Qu, of the piles obtained after the elapsed time of 100 minutes in abutment AI, 200 minutes in pier PI, and 1000 minutes in abutment A2 and pier P2 are shown in Figure 15, with the statistical properties such as the average and the coefficient of variance, COV. All the frequency distributions seem to take the form of normal distribution.
Figure 13. Load-displacement curves derived via wave matching for all piles at each abutment and pier.
702
5 SUFFICIENT NUMBER OF DYNAMIC LOAD TESTS Let us discuss the sufficient number of load tests to be conducted at a site to determine the distribution of pile capacity. The design value, Qd, is assumed to be given by
in which Qk is the characteristic value and YQ is the partial safety factor. If the bearing capacity of the piles takes the form of normal distribution as shown in Figure 16, Q d and Q k are expressed as (Murakami et a1 1988):
in which ,qis the average of Q, VQis the coefficient of variance and k is a factor that depends on the required safety level. If we take the probability of
failure to be 5% for the characteristic value and 0.1% for the design value, the value of k is calculated as 1.64 for the characteristic value and 3.0 for the design value. Using the results for pier P:! and abutment A2 shown in Figure 15, the relationships between the number of load tests and the design and characteristic values are examined, applying the non-parametric bootstrap method. In this method, the number of tests, n, was hypothetically varied f?om 2 to the total number of piles (16 for pier PI and 18 for abutment A2). For each test number, n, n data were selected from the measured bearing capacities in random fashion. Then, the values of ,UQ and VQ were calculated from the selected data to obtain the values of O d and Q k by means of Equation (14). The pair of such obtained is called 'bootstrap sample'. A total of 1000 pairs of the bootstrap samples were obtained for each test number, from which the values for the average and the standard deviation of Q d and Q k were calculated for each test number, n.
Figure 15. Frequency distributions of estimated static resistance, Qu,of the piles at abutments A, and A', and piers PI and P2.
703
Figure 18. Variations of average and standard deviation of Qd and Qkwith of number of load tests at Abutment A?.
Figure 17. Variations of average and standard deviation of Qd and Qkwith of number of load tests at Pier P?.
The variation of Qd and Qkwith the test number, n, are shown in Figure 17 and in Figure 18, together with the range of the standard deviation. It can be seen that the average value converges when the number of tests reaches 3 or 5 , whereas the standard deviation tends to converge when the number of tests reaches 10. These results suggest that the execution of 10 or more load tests give us the design pile capacity that has an acceptable probability of failure which may be obtained from the load tests on all piles for this particular site.
analysis procedure based on the rational soil model to derive the load-displacement curve for a pile was demonstrated in this paper. In the proposed wave matching analysis procedure, soil investigation and soil test data are effectively utilized to estimate the soil parameters except the shaft resistance and the toe resistance, and change in drain conditions between dynamic and static loading are taken into account . The results of dynamic load tests showed: I . The bearing capacity of all the piles exceeded the design ultimate bearing capacity. 2. Set-up ceased within a short period after the end of initial driving, 100 minutes to 1000 minutes, depending on pile length. Further, the sufficient number of load tests to be conducted at a site to determine the distribution of pile capacity was discussed from the view point of a probabilistic design method, based on the results of the dynamic load tests. It was found that the execution of 10 or more dynamic load tests gives us the design pile capacity that has an acceptable probability of failure which may be obtained from the dynamic load tests on all the piles for this particular site.
6 CONCLUSIONS REFERENCES The results of the wave propagation analyses of the dynamic load tests of a total of 102 steel pipe piles for the foundations of Noetsu Bridge No.3 have been presented in this paper. Rational soil models (Randolph & Simons 1986; Deeks 1992) were used in the wave matching analyses. A wave matching
Matsumoto, T., Y . Michi & M. Hayashi 1997. Reliability of dynamic load testing on steel pipe piles in soft rock. PTOC. 14th ICSMFE, Hamburg, v01.2 : 1185-1 188. Randolph, M. F. & H.A. Simons 1986. An improved soil model for one-dimensional pile driving analysis. Proc. 3rd Int. Con$ on Num. Methods in Offshore Piling 1- 17.
704
Deeks, A.J. 1992. Numerical analysis of pile driving dynamics. PhD Thesis, The University of Western Australia. Matsumoto, T., Y. Michi & T. Hirano 1995. Performance of axially loaded steel pipe piles driven in soft rock. Jour. of Geotech. Eng., ASCE, 121(4) : 305-315. Matsumoto, T. & M. Takei 1991. Effects of soil plug on behavior of driven pipe piles. Soils h Foundations, JSSMFE, VOI.3 1 . No.2 : 14-34. Randolph, M.F. & A.J. Deeks 1992. Dynamic and static soil models for axial pile response. Proc. 4th int. Conf: on the Appl. of Stress- Wave Theory to Piles: 3- 14. Novak, M., T. Nogami,& F. Aboul-Ella 1978. Dynamic soil reactions for plane strain case. J. Mech. Eng. Div., ASCE, 104(EM4): 953-959. Verruijt, A. 1969. Elastic storage of aquifers.R.J.M. De Wiesl (ed.), Flow through Porous Media, New York : Academic. Randolph, M.F. 1991. Analysis of the dynamic pile driving, Chapter in Developments in Soil Mechanics - IF Advanced Geotechnical Advances, P.K. Banerjee & R. Butterfield (ed.): Elsevier Applied Science.
705
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Behavior of short CFA piles in an overconsolidated clay based on static and dynamic load tests A.C. M. Kormann & €! R.Chamecki - Structures and Materials Laboratory, LAME, Federal University of Paranci, Curitiba, Brazil
L. RUSSO Net0 - PonthiJcial Catholic Universig of PararicilIn Situ Geotecnia, Curitiba, Brazil L.Antoniutti Net0 -In Situ Geotecnia, Curitiba, Brazil G.€! Bernardes -Department of Civil Engineering, UNESR Campus of Guaratinguetci, Brazil
ABSTRACT: A comprehensive investigation of the behavior of two short continuous flight auger (CFA) piles tested at the experimentation site of the Federal University of Paranh is presented. The piles were submitted both to static and dynamic load tests. After the tests, a visual shaft inspection and complementary SPT borings were performed. Samples of the pile material were tested in order to evaluate the concrete specific weight. The analysis shows the importance of the correct assessment of the pile concrete properties. For a proper comparison between static and dynamic test results performed at the same drilled shaft, the preloading effects should be accounted. Cast-in-place piles may develop increasing point resistance from large displacements resulted from accumulative blows. The CAPWAP static simulations, for blows with energy increasing, are compared with the results of the static load tests. A remarkable agreement was found in the loaddisplacement curves obtained from both test methods.
1 INTRODUCTION The use of dynamic load tests in both precast and bored piles is increasing in Brazil. In some major jobs where continuous ilight auger (CFA) piles were used, dynamic load tests have been carried out for foundation control. There is a number of papers in the literature covering the dynamic high strain testing of drilled shafts (e.g. Seitz 1984, Seidel & Rausche 1984, Likins & Hussein 1995, Hussein et al. 1996, Liu et al. 1996, de Mello & Paraiso 1998). However, the current references are mainly related to the performance of the tests. Pile geometry, concrete properties and soil inertia effects require a proper evaluation for a correct assessment of the dynamic behavior of bored piles. Comparisons between static and dynamic tests at the same pile should account for the preloading effects mainly the accumulative pile displacement. In order to contribute to a better understanding of high strain dynamic testing of drilled shafts, this paper describes the results of the research about the behavior of CFA piles which is in progress at the experimentation site of Federal University of ParanB - UFPR. This is part of the studies about the foundation behavior in the Guabirotuba Geological formation at ParanB State, south of Brazil. The information here presented are also discussed by Kormann et al. (2000).
707
2 THE GEOTECHNICAL PROPERTIES OF THE EXPERIMENTATION SITE OF UFPR AND THE CFA PILES CHARACTERISTICS The Geotechnical Experimentation Site of the UFPR is located at the “Centro Polit6cnico” Campus in Curitiba, Brazil (Chamecki et al. 1998). The natural soil belongs to the tertiary Guabirotuba Geological Formation. Overconsolidated silty clays and clayey silts, with high plasticity, are commonly present soils in this scdimcntary formation. Polished and shiny surfaces are commonly seen in the clayey soil mass. These slickensides follow a pattern of difficult identification. Well defined tectonic structures are also prcsent. Lenses of sands, rich in feldspar, frequently occur inside the clayey mass. Some conglomerate and carbonate deposits may occur at specific sites. A detailed study about the geology of the Guabirotuba Formation is presented by Salamuni (1998) and the geotechnical aspects of these soils are discussed by Kormann (1999). Figure 1 presents representative data of the geotechnical profile and the pattern of the two CFA piles at the time of dynamic testing. At the site of the CFA piles, the subsoil of the Geotechnical Experimentation Site consists basically of high consistency gray and brown silty clay. Lenses of sandy silt and silty sand are also present. The subsoil shows high resistance to the SPT blow count, with N values over 20 blows since the first meter. The CPT test shows stiff clay subsoil, with
Figure 1. Geotechnical profile data and the pattern of the piles at the time of dynamic load tests.
point resistance as high as 12 MPa (Fig. lb). The friction ratio decreases with depth, changing from 4 9% more superficially down to about 3 % close to the test refusal at 6 m of penetration. A high degree of saturation occurs as result of the water present at the discontinuities of the soil mass. The CFA piles were installed in May/1997. Table 1 presents the main features of the two piles and some information obtained by a Turacord system during the installation. Table 1. Pile geometry and Properties Nominal diameter (m) Installed pile length (m) Average injection pressure at the auger head (kPa) Average auger withdrawal velocity (cm/s)
installation data. CFA- 1 0.35 6.0 100
4.7
1. Employing some extrapolation methods (c. g. Van der Veen), it was suggested the ultimate loads of 1006 kN and 1473 kN for the CFA-1 and the CFA-2 piles respectively. 2. The extrapolated values are near the maximum load applied during the load tests. These loads were limited by the structural integrity of the pile caps. It is important to note that the applicability of the extrapolation methods for bored piles have some limitations, especially when the maximum displacements applied during the static tests were insufficient. 3. Although the lengths of the two tested piles differed only by one meter a markedly distinct behavior was observed in their load-displacement behavior. The CFA-2 pile presents more stiff and higher bearing capacity. 4. Such unexpected behavior does not seem to be explained only by subsoil characteristics. The grout pressure might have affected the performance of the piles. In order to improve the understanding about this distinct behavior of the two piles further investigation was planned, which includes: 1. The test of the piles with low strain tests (PIT). 2. The reinforcement of the pile tops and the performing of high strain dynamic tests. 3. A visual shaft inspection of the piles and the surrounding soil. 4. The collection of samples of pile material in order to determine the concrete specific weight.
CFA-2 0.35 7.0 > 200
5.6
3 STATIC LOAD TESTS Two slow maintained load (SML) tests were carried out in Aug-Sep/98 according the Brazilian code NBR-12131. The penetrations were 6.0 and 7.0 m for the CFA-1 pile and the CFA-2 pile respectively. The measured load-displacement curves are presented in Figure 2. The tests procedures, instrumentation setups and interpretation are described by Kormann et al. (1999). The following points were outlined by these authors: 708
Figure 2. Static load tests results.
5 . The performing of additional SPT borings just beside the piles. These investigations and the results of all the available data are presented and discussed in the following items. 4 LOW STRAIN TESTS Since the penetration of the piles was well controlled, the low strain tests allowed to calculate the wave speed, resulting 3550 m/s and 3800 m / s for the CFA-1 and the CFA-2 piles respectively. These wave velocities are somewhat reduced. Although the concrete used in both piles was essentially the same, the higher wave speed of the CFA-2 pile might be associated to a material having a slightly better performance (Kormann et al., 1999).
5 HIGH STRAIN DYNAMIC TESTING In order to perform the dynamic testing the pile tops needed to be reinforced. The reinforcement was carried out using longitudinal steel bars and a cylindrical steel ring mounted at the top of the piles (Fig. la). It is important to note that the reinforcement reduced the pile embedded length by approximately one meter which resulted in some loss of shaft friction near the top of the piles. The dynamic load tests were conducted in Dez/99. For a proper selection of the hammer to be used for dynamic testing of cast-in-place piles, Hussein et al. (1996) suggest a ram weight about 1.5% of the pile static resistance to be verified. A 30 kN free fall hammer was selected for the high strain dynamic tests. One cushion made by wood plate was used to reduce the impact maximum stress. Two piezoeletric accelerometers and two strain transducers were attached to the piles. A Pile Driving Analyzer, PAL model, was employed (PDI, 1999). In order to mobilize as much as possible the soil resistance, the tests employed increasing hammer energies (e.g. Aoki 1989, Bernardes 1989, Beim & Aoki 1996). The drop height was increased in 20 cm increments until the compressive stresses
caused failure of the pile material. The set and rebound of the pile top were recorded for all blows. Table 2 summarizes the dynamic load test data for the CFA-1 and the CFA-2 piles. The first pile tested was the CFA-2 pile. At the beginning of the test the new cushion with low stiffness and low coefficient of restitution generated a low driving system efficiency - a reduced energy (EMX) was transferred. The pile was initially stroked six times with 40 cm fall height until the set became perceptible. During the test, the driving system efficiency increased with the succession of the blows. An efficiency ranging from 12.2 to 38.6% was observed. The dynamic load test comprised seven blows with drop heights ranging from 60 to 170 cm. The maximum tension stresses below sensors (TSX) were small for all blows. The maximum compressive stress (CSX) of 28.5 MPa generated by the seventh blow caused the pile top material to fail just below the steel reinforcement. The CFA-1 pile received six blows with drop heights comprised between 40 and 140 cm. The driving system efficiency increased from 1 1.7 to 32.1% during the test. The tension stresses were slightly higher than i n the CFA-2 pile, but as expected the successive increase in compressive stresses caused the failure of the pile top immediately below the steel ring. The compressive stresses that controlled the end of the test (15.0 MPa at the gages) were lower than the CFA-2 ones. It can be related to bending effects due to eccentric blows or concrete properties. Figure 3 shows the records of force and velocity multiplied by the impedance for the two piles. For the CFA-1 pile, a wave speed of 3000 m/s is suitable. The CFA-2 pile shows a wave speed of 3250 d s . These findings are in full agreement with the wave speeds measured in the PIT tests, which due to the level of deformation usually present higher values. The ratio between the wave speed from both tests for the CFA-2 pile to the CFA-1 one was the same regardless the type of test (1.07- 1.OS). Due to the reduced length of the piles, the width of the pulse impact is considerably higher than 2L/c. Despite that, it will be demonstrated that good results can be achieved. The traces of the force and the velocity of the CFA-1 pile suggest a high skin friction (Fig. 3a). The succession of the increasing energy blows causes the force signal to be close to the velocity signal after 4L/c. It can be related to increased pile point displacements. A considerable pile set (9,4 mm) was measured at the last blow (Table 2). It is expected a high mobilization of the soil resistance. The first record of the CFA-2 pile (Fig. 3b) shows an increase in the velocity just after 2L/c. It is related to a hammer interference. The subsequent records do not present such feature.
709
Table 2. Dynamic CFA- 1 Blow Drop number height (cm) 1 40 2 60 3 80 4 100 5 120 6 140
load test data.
EMX CSX (kNm) (MPa)
TSX (MPa)
Set (mm)
Rebound (mm)
1.4 3.4 5.1 8.2 11.8 13.5
0.4 0.4 0.3 0.5 0.1 0.7
0.1 1.8 3.0 5.2 8.0 9.4
1.9 2.2 2.2 2.3 2.0 2.1
9.4 11.2 12.5 13.8 15.1 15.0
CFA-2 Blow Drop number height (cin) 1 60 2 80 3 100 4 120 5 140 6 160
EMX CSX TSX (kNm) (MPa) (MPa)
Set (mm)
Rebound (mm)
2.2 5.0 8.2 12.2 14.9 18.1
0.5 0.6 1.4 2.2 2.8 4.3
1.6 2.9 3.8 5.1 4.2 3.7
14.1 19.0 21.5 23.9 25.1 27.4
0.0 0.0 0.0 0.5 0.3 0.0
BN - Blow number H - Hammer drop height
0 - Pile ( ~ u c ) 1500 00
BN=l
2.70
H=40cm
4 1s
-7
-7 3 40 rns
+]
1500.00
L
_.--
B N = 2 H=60cm
I
/
\ .-
4
L /
3 20 rns 3 20 rns
I
I
'1
-\ I
'
,
\ .//
A
2
0 5m5 - 20 5m5
Figure 3. Records of force and velocity multiplied by the impedance for CFA-1 (a) and CFA-2 (b) piles.
710
The force signal is considerably higher than the velocity record. It can be due to compressive upward waves that are related to a high skin friction. The comparison between the measured sets and rebounds of the two piles also reflects the distinct behaviors observed both in the static load tests and in the dynamic records. Despite the higher energies applied to the CFA-2 pile, its sets are lower than a half of the CFA-1 ones. The measured rebounds of the CFA-2 pile were as high as 5.5 mm, while in the CFA- 1 pile the maximum rebounds were 2.3 mm. All blows presented in Figure 3 were submitted to CAPWAP analysis. The results will be presented in the item 9.
6 VISUAL SHAFT INSPECTION After the dynamic tests, the soil around the piles was excavated approximately 3.2 m below the ground level. The visual inspection of the soil allows the following remarks to be made: 1. As expected, the silty clay around the piles has polished and shiny surfaces closely spaced. Slickensides following a preferential orientation also were found. The clayey soil around the two piles seems to be similar. 2. A hard stratum of sandy silt is present from the ground level to 1.3 m deep at the CFA-2 pile (see Fig. la). It is important to note that skin friction of this layer acted only on the static test, since the soil around the pile top was removed for the dynamic testing. The following comments can be outlined about the inspection of the piles geometry and material: 1. The surface of the two piles was very smooth and regular. There was no sudden changes in the pile diameter which was expected for this type of pile construction in stiff clayey soils. 2. The actual diameter of the two piles is lower than the nominal diameter (35 cm). The measured diameter ranged between 33 and 34 cm. 3. The reports of the installation data of the two piles indicated excessive amount of concrete used 47-59% more volume than the nominal volumes of the piles. The direct measurements of the diameter of the piles show that there is no reason for the high concrete overconsumption and a better control during pile installation is required. 4. The CFA-1 pile shows some small soil intrusions along the shaft. They are present at the external pile surface and also approximately over to 4 cm inside the concrete mass. The CFA-2 pile does not exhibit soil intrusions. It may explain the lower wave speed (3000 d s ) measured in the CFA-1 pile.
711
7 CONCRETE TESTING In order to evaluate the concrete density, a total of 17 samples of the material of the two piles were collected just below the steel reinforcement and at the bottom of the excavation. Table 3 presents the mean results of the tests. It is important to note that the concrete specific wei ht of the piles - in the range of 19.13-19.52 kN/m - is low. It is due to the high porosity of the material. These findings are in agreement with the low wave speeds measured. It can be noted that despite the low concrete density and probably the low resistance the piles supported well the loads both of the static and the dynamic tests.
5
Table 3. Average pile concrete properties. Pile Concrete specific Voluine of voids / Toweiglit (W/m’) tal volume (9%) CFA- 1 19.52 25.07 CFA-2 19.13 25.56
It should be noted that the concrete specific weight is currently used in wave equation analysis to calculate the dynamic elastic modulus of the pile. It is a common practice to adopt to the specific weight the value of approximately 24 kN/m3. However, if the cast-in-place shaft exhibits a reduced specific weight the analysis would produce overestimated results. In order to illustrate the influence of the concrete specific weight, Figure 4 shows a simplified analysis for the CFA-2 pile. The pile capacity was evaluated with the Case Method. It is observed that for a specific weight range of 20-24 kN/m3 the pile capacity can change about 20%.
Figure 4. The influence of the concrete specific weight in the pile capacity evaluation.
8 ADDITTIONAL SPT BORING Since the excavations around the piles were not sufficient to explain the distinct behavior between the two piles, complementary SPT borings were performed just beside the CFA-1 and the CFA-2 piles. The distance between the borings and the piles was approximately 50 cm. Obviously, the excavation around the piles implies in a stress relief that
Table 4. Complementary SPT boring data. CFA- 1 CFA-2 Depth Blows130 cni Blows/30 cm Blows130 cm Blows130 cm below (first and (second and (first and (second and third 15 cm) second 15 third 15 cm) ground second 15 level cm) NSI’T cm) NSPT
(m) 4.0 5.0 6.0 7.0
16 21 28 30129cni
27 30128cm 30123cm
*
16 22 19 17
27 30 1 2 5 cni 30 19
changes the conditions of the soil. In effect, even the pile installation changes the natural state of stresses. However, since the conditions of the ground around the two piles were similar at the time of the complementary SPT boring, a qualitative comparison can be made among them. Table 4 presents the results. The following remarks can be made: 1. The complementary SPT borings data are in agreement with the SPTT-1 information. 2. The new borings revealed a lens of silty sand just above the point of the CFA-2 pile and 60 cm below the CFA-1 pile. These thin layer is approximately 40 cm thick and it probably appears in the soil profile as presented in Figure 1 a. 3. The N s p ~values along the shaft of the CFA-2 pile are not higher than the CFA-1 ones. At the point of the CFA-2 pile the N s p ~values are lower than the CFA-1 ones. 4. On the basis of the known limitations of the SPT testing, the blow count suggests that the higher resistance of the CFA-2 pile could not be explained only by a stronger soil around the shaft or at the pile point. The hypothesis of the influence of the grout pressure in the performance of the pile cannot be disregarded. Also, since the sand lens intercepts the CFA-2 pile just above the tip, a distinct response of this soil to the installation procedure - such as an slight enlargement - might affect the pile end bearing. 9 CAPWAP ANALISYS: COMPARISONS WITH STATIC LOAD TESTS The records of both the CFA-1 and CFA-2 piles were submitted to conventional CAPWAP analysis. In order to perform a suitable analysis eight pile and soil segments were employed. A concrete specific weight of 20.0 kN/m3 was adopted. The main results are presented in Table 5. The skin quake values of the two piles lies in a narrow range (0.97- 1.622 mm) and they do not depend on the energy level. On the other hand, the toe quakes exhibit a trend to increase with the energy transferred to the pile. It can be related to a growing mobilization of the toe resistance. High toe quakes
were found, in special for the CFA-1 pile. For the two piles the ratio between the toe quake and the measured set is approximately constant - a representative range is 1.O- 1.7. The Case Method damping JC values of Table 5 does not exhibit a trend and should be disregarded. The shaft and toe Smith damping values of the two piles showed dependent on the energy level. At the first blows the toe damping was too high. It decreases with the increasing of the drop height. Such behavior is an agreement with the findings of Aoki & de Mello (1992). These authors show that the damping and quake of the Smith model are not constant soil parameters, but are dependent on the energy level. The use of the radiation damping - that may account for the soil motion associated with the pile movcment in drilled shafts - did not lead to improved results. It seems in agreement with the smooth surface of the piles revealed by the visual inspection and the stiff characteristics of the soil. Rausche et al. (1996) discusses the applicability of the Multiple Blow Analysis (MBA). They recommend such analysis for variable energy blows and provide an example for a drilled shaft. However, for the piles here studied such analysis did not seem suitable. The MBA analysis keeps constant the end bearing. That is not in agreement with a expected increased mobilization of the toe resistance due increasing hammer energies. In this study, the Residual Stress Analysis (RSA) option also did not improve the results. The mobilized capacities of the two piles are compatible with the results of the static load tests. At the end of the dynamic tests, for the CFA-1 and the CFA-2 piles respectively, the maximum mobilized capacities were 10.7% and 48.1% higher than the static capacities reported by Korrnann et al. (1999). The high CAPWAP resistance of the CFA-2 pile suggests that the limited settlements of the CFA-2 pile in the static tests may have conducted to an underestimation of the extrapolated capacity. It is important to note that the CFA-1 and the CFA-2 piles were statically loaded before the dynamic test. These preloading leads to a set of residual stresses both in the pile and in the soil (e.g Massad 1992, Maiorano et al. 1996). The comparison between static and dynamic load tests performed on the same pile should account for the residual stresses. The set of residual stresses does not change the ultimate load. However, the form of static loading curves can be markedly affected. The residual loads generated by a cycle of static or dynamic loading increase the pile head stiffness. In addition, as discussed by Dkcourt (1998), the meaning of the physical failure cannot be applied to bored piles. Cast-in-place piles may develop increasing point resistance over to large displacements.
712
Table 5 . Main CAPWAP results. CFA- 1 1 Blow number Mean Shaft Quake (mm) 1.298 0.764 Toe Quake (mm) 1.298 Shaft Damping (s/m) 1.419 Toe Damping (dm) 0.36 Jc 1.02 MQno 62 1 RS (kN) 24 8 RT (kN) RU (kN) 869 RS - Skin mobilized resistance
CFA-2 6 1 2 4 3 2 3 4 5 0.996 1.001 1.001 1.001 1.000 1.361 0.9 7 1.445 1.526 2.682 4.080 6.413 8.557 11.05 0.757 1.004 2.109 3.422 1.183 0.713 0.533 0.39 0.512 0.324 0.208 0.091 0.099 4.936 2.526 2.055 1.301 2.682 1.411 0.677 0.663 0.13 0.68 0.7 1 0.42 0.41 0.07 0.00 0.18 1.oo 1 .oo 1.28 0.86 0.93 1.32 1.37 2.34 1.96 2.21 1.58 800 1296 1592 1570 1470 590 620 719 760 314 148 167 280 300 300 180 296 310 886 900 1019 1060 1114 1444 1759 1750 1780 RT - Toe mobilized resistance RU- Total mobilized resistance
If the soil resistance degradation effects are reduced, it is expected that the succession of static or dynamic cycles of loading will lead to increased pile capacities. For instance, Liu et al. (1996) points out that when the dynamic test is performed after the static loading, in average a 10% increased pile capacity can be found. On the basis of the discussion above, it seems suitable to compare static and dynamic tests performed at the same pile as a succession of cycles of loading. In order to perform such analysis, the permanent displacement at the end of a cycle of loading can be added to the next one (e.g. Seitz 1984, Niyama & Aoki 1991). Figure 5 shows the static load tests and the CAPWAP static simulations superimposed as cyclic loading. In order to plot the calculated curves, the measured sets were employed. The range of 0.928I . 1 15 is representative for the ratio among the measured sets and the CAPWAP ones. The unloading of the blows number 4 to 7 of the CFA-2 pile was adjusted as a straight line between the CAPWAP maximum displacement (DMX) and the measured set. The simulations plotted in Figure 5 shows a remarkable agreement with the static load tests. The CAPWAP inferred static behavior follows within reasonable limits the trend of the static load tests. For the CFA-I pile simulations, it can be observed that the succession of blows increases the end bearing mobilization and reduces the toe stiffness. The gradient of the last straight portion of the loading cycles increases with the succession of the blows. In effect, the ratio RT / Toe Quake decreases when the drop height increases. It can be related to a nonlinear stress-strain behavior of the soil. Despite the lower displacements, the same trend can be observed in the CFA-2 pile data.
10 CONCLUSIONS At the current stage of the research here presented, the following conclusions can be outlined: 1. The visual shaft inspection of the two CFA piles revealed a very smooth and regular surface. It is consistent with the stiff and hard clayey soil 713
5 1.595 3.573 0.138 1.371 0.84 1.14 1574 300 1874
6 1.622 5.866 0.1 17 0.974 0.63 3.26 1776 309 2085
7 1.000 5.028 0.1 17 0.702 0.59 3.12 1876 306 2182
around the piles. The actual diameter of the two piles, ranging between 33 and 34 cm, is lower than the nominal diameter (35 a n ) . It shows that there is no reason for the high concrete overconsumption reported by the installation monitoring system. The CFA-I pile shows some small soil intrusions over to 4 cm inside the concrete mass. The wave speed of the CFA-1 pile (3000 m/s) was lower than the wave speed of the CFA-2 pile (3250 m/s). 2. The concrete testing of the CFA-1 and the CFA-2 piles showed a reduced specific weight (19.52-19.13 kN/m3) and a high porosity. Despite that, the structural behavior of the piles was satisfactory both in the static and dynamic load tests. However, it should be noted that the concrete specific weight is currently employed in wave equation analysis to calculate the dynamic elastic modulus of the pile. The use of an overestimated concrete specific weight would produce an overestimated computation of the cast-in-place pile capacity. In addition, the low specific weight suggests that the effects in the wave equation results of a possible nonlinear stress-strain relationship of the pile material should be investigated. 3. The distinct behavior observed between the two piles in the static load tests was confirmed by the dynamic load tests. Despite their similarity, the pile-soil interaction of the two piles is markedly distinct. The CAPWAP analysis suggest that the skin friction of the CFA-2 pile is higher than the CFA-1 one. On the basis of the current soil investigation performed, the higher capacity of the CFA-2 pile docs not scem to be explained only by a stronger soil around the shaft or at the pile point. The hypothesis of the influence of the grout pressure on the pile-soil interaction cannot be disregarded. Also, since a sand lens intercepts the CFA-2 pilc just above the tip, a distinct response of this soil to the installation procedure might affect the pile end bearing. 4. For the two piles, the maximum mobilized capacities in the dynamic tests ( 1 1 14 and 21 82 kN) are in reasonable agreement with the results of the static load tests (1006 and 1473 kN). The higher CAPWAP capacities are due to the higher displacements achieved in the dynamic tests. A comparison between static and dynamic load tests performed on the same pile should account for the preloading effects. Cast-in-place piles may develop increasing
Figure 5. Static load tests and CAPWAP static simulations for the CFA-1 (a) and the CFA-2 (b) piles.
point resistance over to large displacements. If the soil resistance degradation is reduced, it is expected that the succession of static or dynamic cycles of loading will lead to increased pile capacities. 5 . In order to compare the results of the static and dynamic load tests, the permanent displacement at the end of a cycle of loading should be added to the next one. As shown in Figure 5 , the CAPWAP static simulations shows a remarkable agreement with the trend of the static load tests. 6. The results here presented demonstrate the potential of the high strain dynamic testing as a valuable tool for the assessment of the load-settlement behavior of drilled shafts. ACKNOWLEDGMENTS The authors would like to express their gratitude to Eng. Alexandre Chwist (Estacas Premold), Eng. Janies Barossi (Sondar) and Norberto E. Calliari for their support in performing the tests here reported. REFERENCES Aoki, N. 1989. A new dynamic load test concept. Proc. Discussion Session 14, TC Pile DiYvirig, X I 1 lilt. Coi$ Soil Mech. Fouiidcition Erig., Rio de Janeiro: 1-4.
714
Aoki, N. & V.F.B. de Mello 1992. Dynamic loading test curves. Proc. 4th Itit. Cot$ Applic. of Stress- Wuve Theory to Piles: 525-530. Beim, J. & N.Aoki 1996. Dynamic load test method with variable energy. Proc. 5th lilt. Conf: Applic. of Stress- Wave Tlieoiy to Piles: 274-28 1. Bernardes, G. P. 1989. Dyiiciiiiic mid stcitic testbig of large model piles in sciiid. D.Sc. Thesis, Norwegian Institute of Technology, Trondheim. Chamecki, P.R., A.C.M. Kormann, N.A. Nascimento & A.S. Dyniinski 1998. Sitio experimental de geotecnia da UFPR objetivos e dados preliminares. Proc. of X I COBRAMSEG, ABMS, Bmsilia, Brazil: 8 19-826. de Mello, L.G. & S. Paraiso 1998. Variable energy dynamic load test on 1.0 m diameter CFA pile. BAP 111, Belgiuii: 32 1-334. Dkcourt, L. 1998. Ruptura de fundaG6es e coeficientes de segur a n p a luz do conceit0 de rigidez. Proc. o f X I COBRAMSEC, ABMS, Brcisilia, Brazil: 1599-1606. Hussein, M., G. Likins & F. Rausche 1996. Selection of a hammer for high-strain dynamic testing of cast-in-placeshafts. Proc. 5th 1nt. Coi$ Applic. of Stress- Wave Tlzeory to Piles: 759-772. Kormann, A.C.M. 1999. Comportamento de argilas rijas: aspectos geotkcnicos da Forma@io Guabirotuba. A m i s du M e w Redoiidci Ccimcteristiceis Geote'cniccis du Fosrii~ip7o Gu(ibirotubLi,ABMS/UFPR, Curitiba, Briizil: 1 19- 128. Korniann, A.C.M., P.R. Chamecki, N.A. Nascimento & A.S. Dyminski 1999. Load tests on continuous flight auger piles in the Guabirotuba Formation. Proc. X I Punciiiiericciti Cot$ on Soil Mecli. iirid Geotech. Eiig. : 1537- 1544. Kormann, A.C.M., P.R. Chamecki, L. Russo Neto, L. Antoniutti Neto & G.P. Bernardes 2000. Estacas hklice continua em argila sobreadensada: comportamento em provas de carga est6ticas e dinbmicas. Proc. SEFE l V , S3o Paulo, Brazil. Likins, G.E. & M.H. Hussein 1995. High strain dynamic testing of drilled shafts and cast-in-place piles. Deep Foulid. liist., 20th Aiiriuul Menibera Cot$ Meet., Cliurlestoii. Liu, C., Q. Lin & F. Shi 1996. Determining the bearing capacity of large-diameter bored cast-in-situ piles by high strain dynamic pile-testing. Proc. 5th lrit. Coi$ Applic. of StressWuve Tlieoiy to Piles: 797-804. Maiorano, R.M.S., C. Viggiani & M. Randolph 1996. Residual stress system arising from different methods of pile installation. Proc. 5th lilt. Coilf: Applic, of'Stress-Wuve Theory to Piles: 5 18-528. Massad, F. 1992. Sobre a interpretaq5,o de provas de carga em estacas, considerando as cargas residuais na ponta e a revers3o do atrito lateral. Park I: solos relativamente homogEneos. Solos e Roclzas 15 (2): 103-115. Niyania, S. & N. Aoki 1991. Correla@o entre provas dc carga dinbmica e esthtica no campo experimental da EPUSP/ABEF. 2'. Seni. de Erig. de FuridugGes EspeciLiis, SCO Ptiiilo, B t - ~ i ~ i285-293. l: PDI 1999. Pile Dt-iviiig Aiidyzer, PAL Model, Users Mariual. Cleveland: Pile Dynamics. Rausche, F., B. Richardson & G. Likins 1996. Multiple blow CAPWAP analysis of pile dynamic records. PI-oc. 5th bit. Cor!f:Applic. of Stress-Wcive Theor? to Piles: 435-446. Salamuni, E. 1998. Tectdriicci dci bucia Sediiiientar de Curitibu (PI?). P1i.D. Thesis, Instituto de GeociEncias e CiEncias Exatas - UNESP. Rio Claro. Brazil. Seidel, J. & F. Rausche 1984. Correlation of static and dynamic pile tests on large diameter drilled shafts. Proc. 2nd Iiit. Conf;Applic. of Stress- Wcive Theory to Piles: 3 13-3 18. Seitz, J.M. 1984. Dynamic testing of bored piles i n noncohesive soils. Proc. 211d lrit. Cot$ Applic. of Stress- Wuve Tlieorj to Piles: 201 -209.
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Static and dynamic testing of the 'Campile' - A displacement, cast-in-situ pile David J. Klingberg & Phi1 Mackenzie Wagstaff Piling Pty Limited, Brisbane, Qld, Australia
ABSTRACT: Throughout the world, developments are continually being made in the foundation industry to overcome construction and environmental difficulties. The need for a low noise and vibration method for installing piles led to the development of such a system called the 'Campile'. The 'Campile' was developed in Australia by Wagstaff Piling as a means of installing piles on sites that are sensitive to noise and vibration, whilst maintaining the advantages of a displacement pile. This paper outlines the results of a dynamic and static test program conducted on a 'Campile', with the static load test taken to failure. The test results indicate a good correlation between the two test methods. the probe is replaced by concrete or grout as the probe is removed from the ground. The finished pile may be reinforced either partially or full-length, depending on the design requirements. The Campiles installed for this project were at a developmental stage and were constructed as a straight-shafted, nominal 450mm diameter pile. Further advances have since been made with the design and installation of the Campile which allows for other sizes and profiles to be constructed, including a 'screw-shaped' pile which provides increased shaft resistance and greater overall pile capacity.
1 INTRODUCTION
Within the foundation construction industry, environmental considerations with respect to noise and vibration levels have become of more concern in recent years. Modern piling contractors continually look for new ways to provide cost-effective and technically sound solutions to these environmentally challenging projects. The Campile was developed by Wagstaff Piling, Australia as the result of such a need. The Campile is a displacement, cast-in-situ pile that is installed under low noise and vibration conditions. As a displacement pile, it combines the advantages of non-removal of soil, but maintains the low noise and vibration of traditional non-displacement, drilled piling methods. This paper presents the outline of the test pile programme completed on a Campile, summarises the results of the 'Class A' dynamic test data and provides a comparison of these results with the static test data. Some recommendations are also included in this paper with respect to methods to be adopted for wave equation analysis for displacement, cast-in-situ piles.
3 SITE GEOLOGY
The project for the installation of the Campiles was located at an industrial site near Sydney, Australia. The site was underlain by loose to medium dense sand and silty sand layers to depths of approximately 4 to 5 metres over a dense sand layer to 11 metres. The target penetration for the Campiles ranged from 5.5 to 7.5 metres where the base resistance was expected to be high in the dense sand layer.
2 DESCRIPTION OF THE CAMPILE
4 TEST PILE PROGRAMME
The Campile is a proprietary pile of Wagstaff Piling and is constructed by screwing a specially designed probe into the ground, which displaces the soil as the probe penetrates the ground. The void created by
4. I Dynamic test results
Test pile TP was installed as a sacrificial pile for the project. The pile was a 450mm diameter (nominal)
715
Campile with a total length of 7.0 metres and a drilled depth of 5.8 metres. The Campile was constructed with grout using a 40MPa design strength. Test cubes were crushed prior to the dynamic test being performed and indicated a strength of approximately 32MPa at 7 days. The dynamic test was conducted when the pile was 8 days old. To enable the test piles to be subjected to dynamic and static testing, a special 40 MPa concrete head was cast onto the pile to contain the impact and bursting stresses from the impact of the test hammer. The test hammer was a five (5) tonne drop hammer, with drop heights up to 13OOmm used during the testing. The dynamic testing and CAPWAP@analysis were completed as ‘Class A’ predictions (ie. without knowledge of the static test results). The typical dynamic test data is shown in Figure 1, where the difference in impedance between the pile head and pile shaft (as expected) is clearly seen. The results of the dynamic testing are summarised in Table 1, with the results of the CAPWAP@ analysis included in Table 2. PILETEST
Table 2. ‘Class A’ CAPWAP@analysis summary. Pile ShaR End Total No. Resistance Resistance Resistance
TP
Lrz
50 m s
...... .................. ....... ..... .........
0 ..................
F
(W 1,797
Table 3. ISL compression test results summary. Load Pile head Creep Creep (kN) Deflection Correction Corrected Deflection (mm) (mm) (mm) 225 0.52 0.00 0.52 450 1.63 0.00 1.63 675 3.06 0.03 3.03 900 4.75 0.20 4.55 1,125 8.15 0.27 7.88 1,350 17.70 5.89 11.81 1,575 25.60 6.41 19.19 1,800 50.00 10.27 39.73
51 DROP W E R
3.2mS
(W
1,427
The test pile was subjected to an incremental sustained load (ISL) compression test in accordance with the requirements of the Australian Standard@ Piling Code - AS2159 (1995). The load test was conducted to a load level where plunging failure of the pile occurred. The results of the ISL load test have been summarised in Table 3 and shown graphically in Figure 2. As dynamic load tests cannot predict any time dependent deflections and to enable a realistic comparison to be made, all deflections from the ISL test have also been corrected for creep. The creep corrected values are given in Table 3.
PD(PLEDRMV,UULIZER
RMX 1890 kN EMX 43.30 kKm DMX 22.8mn WU2 100kN LE 6.10111
370
4.2 Static test results
TP-RST
.....................
(W
1‘
4.3 Comparison of static and dynamic test results
kN
~
l ..,. p..
.. ... .
.
.x . l I
The results of the static and dynamic tests have been plotted in Figure 3. It can be seen that the CAPWN@ analysis provided a conservative load-deflection behaviour compared to the static test data. Afker the results of the static test were known and compared to the ‘Class A’ predictions, the CAPWAP@analyses were re-analysed to provide a predicted deflection behaviour that more closely modelled the static behaviour. The results of the re-analysis are given in Table 4. It is noted that the revised CAPWAP@analysis did not alter the ultimate capacity of the pile, but adopted a different distribution of capacity to achieve the ‘better match’ to the ISL test results. The comparison between the revised CAPWAP@analysis and the static test data is given in Figure 4 and Table 5 . The revised CAPWAP@ analysis showed excellent agreement with the ISL test results.
............................................................ & ; s
.. ,. I
I
I
Figure 1. Dynamic test data for test pile TP Table 1. Pile No. TP
Dynamic testing results summary. EMX RMX TC Set Stroke (mrn) (mm) (mm) (kNm) (kN) 6.8 43.3 1,890 1,300 16.0
Set = pile set in mm/blow TC = temporary compression in mm EMX = max. energy transferred to pile head in kNm RMX = PDA capacity estimate in kN 716
Figure 4. ‘Revised’ CAPWAP@ and compression test results.
ISL
4.4 Discussion on results
Figure 3 Comparison of ISL compression test and ‘Class A’ dynamic test results. Table 4. ‘Revised’ CAPWAP@analysis summary. Pile Shaft End Total No. Resistance Resistance Resistance TP
(W
(lm
776
1,021
Table 5. Revised CAPWAP@and ISL test results summary. Load ‘Class A’ Revised (kN) CAPWAP@ CAPWAP@ Deflection Deflection (mm) (mm) 225 1.78 0.48 450 3.55 0.96 675 6.34 1.50 900 10.59 4.39 1,125 14.89 10.67 16.95 19.20 1,350 1,575 26.23 23.22 43.85 43.79 1,800
(W 1,797
compression ISL Creep Corrected Deflection (mm) 0.52 1.63 3.03 4.55 7.88 11.81 19.19 39.73
The re-analysis of the dynamic test data and the corresponding better match with the ISL test results indicated that the shaft resistance on the pile was significantly higher than initially obtained from the ‘Class A’ CAPWAP@ analysis. The revised CAPWAP@showed approximately twice the shaR resistance of the ‘Class A’ analysis. The re-analysis was achieved with significantly lower skin damping and quake parameters, with only minor changes to the toe damping and quake parameters. It should be noted that shaft radiation damping parameters were used in both the ‘Class A’ and ‘revised’ CAPWAP@analyses (CAPWAP@Manual, 1996). Without the use of this modelling tool, the predicted ultimate capacity from the CAPWAF@ analyses may have been significantly lower than the static test value (approx. 15% to 20% lower). A comparison of the shaft resistance values from the ‘Class A’ and ‘revised’ CAPWAP@analyses was also made with standard pile design calculations using the borehole information for the site (Poulos and Davis, 1980). Three separate design calculations were performed on the borehole data: (a) bored pile design; (b) grout-injected pile design; and (c) driveddisplacement pile design. A summary of these design calculations is given in Table 6. It can be seen from Table 6 that the ‘Class A’ values from CAPWAP@correlated with bored pile design parameters, whereas the revised CAPWAP@ values were slightly higher than those from a driveddisplacement pile design. This confirmed that the Campile was performing in the manner in which it had been designed - that is, as a displacement pile and not as a non-displacement pile.
717
Table 6. Design calculations summary. Depth Stratum Bored Pile (m) Design (kN) 2.5 L/MD Sand 107 3.2 Loose Sand 29 4.5 MD Sand 73 5.8 Dense Sand 199 Total 408
GIP Design (kN) 193 52 132 22 1 598
Driven Design (kN) 214 58 147 23 8 657
CAPWM@
(W
‘Class A’ 126 45 61 138 370
‘Revised’ 253 84 185 254 776
5 CONCLUSION
REFERENCES
The Campile is a proprietary pile type developed by Wagstaff Piling, Australia as a low noise, low vibration, cast-in-situ, displacement pile that produces limited spoil. The Campile was developed as part of an ongoing research programme to provide piling solutions for environmentally sensitive sites, particularly inner city developments where noise and vibration considerations are paramount. The test pile programme for the Campile installed for this case study showed that dynamic testing methods could be used to predict the behaviour of the Campile under static loading conditions. The ‘Class A’ CAPWAP@analysis provided a conservative prediction of the pile behaviour. A ‘revised’ CAPWAP@ completed after the static results were known showed excellent agreement between the dynamic and static methods. It was noted that the revised analysis provided a significantly larger shaft resistance than the initial ‘Class A‘ analysis and adopted significantly lower shaft damping and quake parameters. Comparison of the shaft parameters obtained from the CAPWAP@analyses with calculations from the borehole information for the site indicated that the Campile was performing as a displacement pile. That is, the shaft parameters obtained from the test programme on the Campile were similar to those expected for a driven, displacement pile. It was also noted that the Campile installed for this case study was a development stage of the pile type and was installed as a straight-shafted pile. Further advances in the pile installation process have since been made with the Campile which allows for various sizes and profiles to be constructed, including a ‘screw-shaped’ pile which provides increased shaft resistance and greater overall pile capacity. During the installation of the Campile, torque measurements were also recorded but have not been included in this paper. However, it is noted that the torque readings show good correlation with the CPT data collected for the site.
Australian Standard Piling Code. AS 159 (1995). Standards Australia. CAPWAP@ Manual (1996). Pile Dynamics Inc. Cleveland, Ohio. 1980. Pile Poulos, H.G. and Davis, E.H. foundation analysis and design. John Wiley and Sons.
71a
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
LT the final word? Correlation between DLT and SLT H.Goldemberg & J. J.Goldemberg Geotecnica Cientec, Buenos A i m , Argentina
ABSTRACT: Due to the increase in service loads of foundations, a world wide trend is being followed by geotechnical engineers and designers to measure the real behaviour of the pile-soil interaction. Historically, this was defined by the Load-Displacement Curve, obtained from a Static Load Test, and the concept of bearing capacity.. .if those involved in the project arrived to an agreement on its definition. Technology developed and the concept of Dynamic Load Test (DLT) arrived, bringing with it a much economic and faster way of measuring the response of piles to imposed loads. But is it this what the geotechnical engineer requires? A comparison between DLT and SLT is analysed throughout the eyes of a geotechnical engineer and not from the perspective of a testing house. Different types of piles were tested with both methods, seeking for correlation, side effects and installation influence in results, aiming to know the real performance vs. predictions. 1 INTRODUCTION In the first half of the ‘SOS, it was introduced in Argentina the pile testing speciality based on the Stress Wave Theory; at the beginning throughout the Sonic Integrity Test (SIT), as a parameter for control quality of foundations and, afterwards, with the Dynamic Load Test (DLT) in order to measure the pile-soil behaviour. But up to then, how was determined the bearing capacity of a piled foundation? The answer is simple, in the same way than in the rest of the world, that is, with a Static Load Test (SLT). As we all know, in spite the familiarity SLT has within the engineering community.. .that, if we first agree on what type of SLT, due that the ASTM D 1143-81 Standard has five different methodologies.. .it is time and money consuming. During those years, where Quality Assurance (QA) started to play an important role in the Argentinean construction market, as well as the increment in service loads, leaded to an increase in the demand for verifying piling works. As well, a reactivation of the local economy, the availability of new technology in construction methods and tighter schedules for finishing site works, allowed the DLT to be introduced as an alternative to the cumbersome static test.
719
It is the objective of this Paper to present correlations between Dynamic and Static Load Tests performed on the same pile, including all the information that the Geotechnical Consultant had at the moment of analysing the results and highlighting how it was arrived to those conclusions. It is the moment to mention that in all cases the Dynamic Load Tests were performed using the FPDS-3 equipment, developed by TNO Building & Construction Research (the Netherlands), while the SLTs were conducted under ASTM Standard. When the Static test was carried out, the transfer load structure assembled consisted in a beam and reaction piles anchored well below the influence area of the test pile. 2 CASE HISTORIES. 2.1 Buildings, The site was a group of buildings constructed in the City of Buenos Aires, Argentina. The ground conditions are described in the following geotechnical profile (Figure l), and the pile type was of substitution, reinforced concrete and casted in-situ under bentonite mud (pile characteristics are presented on Table 1).
Table 1. Pile characteristics. Pile Bentonitemud Type I J J I1 111
J
Diameter (m) 0.35 0.40 0.50
‘slow method’ and then 7% of the foundation population (29 piles) were going to be tested with the Dynamic Load Test (the tests information is presented on Table 2) Conclusions: The results of the homologated tests are presented on Figure 2, where it is clearly seen the good correlation between DLT and SLT. The quality of this match depends dramatically on the quality of the geothecnical investigation, the knowledge of the mechanical properties of the concrete and the concrete consumption, in order to check variations in the cross section of the piles.
Length (m) 14.0 15.0 15.0
2.2 Industrial.facility. The site consisted in the expansion of an industrial facility in the South Area of Gran Buenos Aires, where 250 precast reinforced concrete piles were going to be driven with a diesel hammer. Pile data is available on Table 3 . In order to determine the pile length in different areas of the building site, Pile Driving Analysis (PDA) was performed, monitoring Driving Resistances, Enthru Energy, Stresses, etc. The 100% of the pile population was tested with SIT searching for cracked piles, 10% of the foundation (20 piles) was tested dynamically and only 1 ‘slow’ SLT was performed to have a correlation between both methods. The summary of these investigations is presented on Table 4.
Figure 1. Geotechnical profile.
Table 2. Test programme* Test Pile Quantity Service Max. Test Maximum Type Type Load Load Settlement (MN) (MN) (mm) DLT I 10 0.6 1.4 2.6 I1 10 0.7 1.9 3 .O I11 9 0.9 2.4 3.9 SLT I 1 0.6 1.2 2.2 I1 1 0.7 1.4 2.1 Ill 1 0.9 1.8 2.6 Note *: information is presented only for those piles where DLT and SLT were performed.
Figure 2. DLT and SLT load-settlement curves. Table 3. Pile characteristics. Pile Cross section Type (mxm) Concrete precast 0.4x0.4
The foundation project consisted in 408 piles placed under caps in groups of one, two or three units each. All the installed piles had to be controlled through the Sonic Integrity Test (SIT) in order to verify that they were fiee of damages and to select which ones were to be subject for DLT and SLT. It was decided, by the Engineer, that the first three piles were going to be tested statically with the
Table 4. Test programme. Test Quantity Service Type Load (MN) PDA 40 DLT 20 1.2 SI.T 1 1.2
720
Length ( 4 19.0
Max. Test Load (MN)
Maximum Settlement (mm)
2.9 2.4
4.8 3.8
It is interesting to discuss what happened with this correlation. Due to the unreal short period of time allowed to finish all the driving works, the Engineer decided to perform the Dynamic Load Test immediately after driving- although GEOTECNICA CIENTEC intended to persuade him not to do so because a less total resistance was going to be obtained due to the influence of pore pressure. The geotechnical information, described on Figure 3, shows the stratigraphic profile throughout SPT and CPT tests. Because discrepancies were detected during the original borings in different areas of the site, Dutch Cone was used. It was also very useful the Pile Driving Prediction (PDP), performed before the first pile was driven in order to select the most suitable diesel hammer. On Figure 4, the results of the load tests performed on pile No 135 are plotted, where the first DLT carried out did not fidfilled the expectationsdue to what was mentioned above. After the results were available, the Engineer decided to perform the SLT two weeks later, having in mind that the test could not be taken to failure because it was a service pile. Five days after finishing the Static test, the piling rig was used to redrive pile No 135 to monitor its Dynamic behaviour. As it can be clearly observed, the differences between the first DLT and the SLT as well as among the Dynamic ones was caused because it was not taken into account that in cohesive saturated soils, the driving effects on remoulding soil layers and the slow dissipating pore pressure cause a reduction of the bearing capacity.
Conclusions: Dynamic Load Tests are affected by local geotechnical characteristics as well as by soil conditions at the moment of carrying out the measurements. This is the reason why the frst DLT did not provide reliable information, because the behaviour of the pile-soil interaction during testing was different than that at the design stage. It is the authors’ opinion that in order to perform a better geotechnical design of piles, more efficient and foundation works less expensive, it must be to carry out CPTs tests during the soil investigation programme. In this way better Signal Matches will be obtained between calculated and measured Upper Travelling Waves, because more exact soil models can be evaluated.
Figure 4. DLT and SLT load-settlement curves.
2.3 Bridge.
In the Province of Santa Fe, Argentina, due to yearly over floods of the Parana River- caused mainly by heavy rains during the rainy season, many bridges in the area of the lowlands had to be rebuilt and some others expanded. In this case, the project consisted in raising the level of the road as well as increasing the length of the bridge fi-om one span 10m long to a deck of two spans 15m each. The geotechnical design for the foundation was based on the soil strata illustrated on Figure 5 , adopting bored piles casted under bentonite mud. The pile details are included on Table 5. The North, South and Intermediate piers had three piles each, which were tested with SIT, aiming to detect defects such us cracks, soil inclusions, contaminated concrete and their real lengths. The testing programme carried out is shown on Table 6. Table 5. Pile characteristics. Pile Bentonite mud Type Bored J Figure 3. Geotechnical profile.
721
Diameter (m) 1.20
Length (m> 21.5
Figure 6. DLT and SLT load-settlement curves.
Figure 5. Geotechnical profile. Table 6. Test programme. Test Quantity Service Max. Test Maximum Type Load Load Settlement (MN) (MN) (mm> DLT 1* 4.0 6.5 7.3 SLT 1 4.0 7.2 8.6 Note *: Originally 1 SLT and 1 DLT were projected, but as one pile was damaged it was agreed to perform a second DLT.
From the interpretation of the Sonic Integrity Tests signals, it was concluded that pile NO4 had an impedance reduction. Afier a detailed analysis using the Signal Matching technique with TNOWAVE, it was calculated that the abnormality consisted in a reduction of cross section from 1.20m of nominal diameter to 1.12m extended over a length of 0.5m and at a depth of 14.6m. With this result, it was decided to go further on in the investigations, in order to determine the real behaviour of the pile-soil interaction and measure if the foundation element was suitable of transferring loads to the surrounding stratigraphy. A Dynamic and a Static Load Tests were going to be conducted on pile N04.
Due to the local geological conditions, mainly sandy soil, a ‘quick’ Static Load Test was going to be performed because the creep effect was not critical for the bridge behaviour. The results are plotted on Figure 6 jointly with its dynamic counterpart. Conclusions: As it is clearly presented, the Dynamic Load Test can have an accurate static correlation, inclusive for high loads- 7MN as in this case, As in the previous cases, of the building and the industrial facility projects, it is an irrevocable condition - in order to obtain a good match or correlation between dynamic and static measurements of a same phenomenon pile-soil interaction - to have an exhaustive geotechnical investigation, knowledge of the parameters of the materials used for the construction on the foundation and the pile shape. Although some skepticism, that still over flies on some Engineer’s minds, Dynamic Load Testing is reliable if the above conditions are fulfilled, inclusive for high loads in cast in-situ piles. 3 GENERAL CONCLUSIONS. In the above case histories, where piles varied in dimensions and construction methods- fi-om driven precast concrete to bored piles under slurry, throughout Franki type; installed in clay, lime and sand affecting the pile-soil interaction in different ways under small to big loads, they all have a commo n den0 minator. That common denominator is, leaving aside the fact that all tests were properly performed, recorded and post-processed, the quality of the basic inforrnation. In other words, the geotechnical information, dimensions and pile shape, mechanical properties of materials used and the installation procedure. For both types of test, Dynamic (DLT) and Static (SLT). that information is vital. Notable contradiction, because it is required for both designing the test and interpretation of results. It is useless to obtain
722
huge amount of data, in white paper in case of a SLT or a computer file for a DLT, in order to be plotted as the ‘Load-Settlement Curve’ and handed in to the Client as a great achievement certifying that ‘the pile will stand the load’ without the geotechnicalstructural diagnosis. In other words, the interpretation! If not, effort, time and money would have been wrongly spent.. .three not rehndable goods in any aspect of life. But making an abstraction and travelling to the Contractors’ Paradise, where no testing is required and QA has not been invented- not even as a concept. Nevertheless, would not be necessary to know the geotechnical investigation, material properties and selected pile in order to construct the project? In almost all the above case histories, a good correlation between Dynamic and Static Load Tests was obtained, satisfying the expectations for homology of both methods. It is a fact the acceptance for DLT fiom the local market and its growing reliability, due that it provides similar results than the Static Test with almost no delays on site and reduced budget. 4 REFERENCES. ASTM, D 1143-81 (Reapproved 1994). “Standard Test Method for Piles Under Static Axial Compressive Load”. ASTM ASTM, D 4945-89. “Standard Test Method for High-Strain Dynamic Testing of Piles”. ASTM C. R. Mullins (1 992). ‘‘Large diameter pile test project”. Proceedings of the Conference on Recent Larfe Scale Fully Instrumented Piles in Clay, London, June 1992 Carol1 L. Crowther (1988). “Load Testing of Deep Foundations”. German Society for Static and Dynamic Pile Tests (1997). “Recommendations for Static and Dynamic Pile Tests”
723
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Results of an international pile dynamic testing prediction event A. Holeyman Universitk Catolique de Louvain, Louvain-la-Neuve, Belgium
J. Maertens Katholieke UniversiteitLeuven, Leuven, Belgium
N.Huybrechts & C. Legrand Belgian Building Research Institute, Brussels, Belgium
ABSTRACT: An international prediction event was carried out within the framework of a 30-pile testing program organized in Belgium. That program called upon several testing methods: static load tests, Statnamic testing, and dynamic testing. This paper provides a summary of received predictions and results obtained from the static pile load tests, which were carried up to failure of the instrumented piles. The comparison between predictions is made using load-settlement curves with reference to the results of the static load tests. A companion paper reports on the project background information that was required to prepare the prediction event, including a description of the pile types and results of an extensive soil investigation program. al, 1999b). Predictions could be established based on the dynamic load tests, the geotechnical investigation, experience, or a combination of the above. The piles types, layout, and the site investigation are described in a companion paper (Holeyman et al, 2000). The present paper focuses on load tests results and their comparison with results of the static load tests.
1 INTRODUCTION 1.1 Program Background
A national research project has been organized by the Belgian Building Research Institute in order to establish the performance of different types of castin-place ground displacement screwed piles. The program included the installation and testing of 30 test piles that allowed the organization of a prediction event. That prediction effort was undertaken with the hope to document the profession's ability to estimate these new piles behavior based on standard investigation means as well on dynamic testing. Six different types of ground displacement piles were installed (five of each) and tested: one prefab and five cast-in-place screwed types: Atlas, De Waal, Fundex, Olivier, and Omega. An extensive soil investigation was performed as part of the research project, including in situ tests (CPT, PMT, SPT, DMT, SASW) and laboratory tests on undisturbed samples. 1.2 Prediction preparation A project synopsis had been prepared to invite interested parties to make those predictions (Holeyman et al, 1999a). It included a description of the pile types, site characterization, the static load test procedure, the dynamic load test procedure and the format of the prediction. Interested predictors were sent the complete information, available as laboratory and in situ investigation and dynamic load test results (Holeyman et
2 DYNAMIC LOAD TESTS 2.1 Procedure for dynamic load tests The loading device used to impact the 12 piles installed for that purpose was a 4 tons drop hammer operated by a crane. A sequence of several blows was applied to each pile. The drop height sequence most often applied was as follows: 0.40m, 0.80ni, 1.2m, 0.8m, and 1.2 m. Dynamic measurements of strain and acceleration were acquired for all 12 piles using a TNO FPDSS system. In addition 6 piles were also monitored using a PDI PDA-PAK system (Prefab, Fundex, and De Waal piles) and one Omega pile was monitored using a PDI PAL system. A 0.4m-diameter head was cast on July 6'' on top of the 10 cast-in-place piles. The transducers were attached generally 0.8 m from the top of the approximately 1.5 m high head. Displacements were acquired using a laser system. 2.2 Distribution of dynamic load test measurements The results of dynamic load test measurements (pile head force, velocity and displacement) were made 725
Figure l a : Long Piles Load Settlement Curves - Prefab, Fundex and De Waal
726
Figure 1b : Long Piles Load Settlement Curves - Olivier, Omega and Atlas
727
available under a digital format to parties that had expressed an interest to make a prediction on that basis. Also characteristics needed to interpret the measurements were provided in part in Table 1 of the companion paper (in particular dimensions and properties of the pile heads extension) and in part in the files containing the measurements. Additional characteristics (wave propagation speed, density, pile impedance, etc.) needed to further analyze the measurements were distributed together with the measurements. Interested parties obtained the digital files of the events by e-mail, which required the structuring of a vast amount of information, totaling more than 8 Mbytes of digital records. For each pile type a directory was established (for example the Prefab directory), containing subdirectories according for each pile of that specific pile type. The ‘pile number’ subdirectory (for example subdirectory pile A7 in the directory Prefab) contained a word file: (e.g. ‘A7info.doc’) and was further subdivided into the following subdirectories: TNO files, ASCII files, PDA files, and Displacement files. The word file ‘A7info.doc’ gave supplementary information about the dynamic load test on the pile A7 (the blow numbers, the drop height, and field notes). It was thus possible for the predictors to reprocess the raw signals using adjusted pile parameters, number of samples, etc. 2.3 Meas urements nomin a I interpretat ion
The choice of the relevant moduli and sections is often considered as part of the predictors’ art and was purposely left open to some degree, as is usually the case for cast-in place piles. It was emphasized that all files had been uniformly acquired using a nominal modulus of approximately 40,000 MPa and a nominal wave speed of 4,000 m/s at the measurement section (i.e. in the concrete of the cast head). The pile heads were cylinders with a diameter of 0.4 m, except for the Prefab pile where the current 0.35x0.35m section prevailed. It was the predictor’s responsibility to assess the measurement section modulus adequate for his prediction. It was also emphasized that the concrete of the tested pile below the added head had a different modulus and a different section. It was also the predictor’s responsibility to assess the appropriate section and modulus for the shaft. Peripheral information allowing the predictor to perform that important assessment included Table 1 of the companion paper, the digital signal themselves (e.g. impedance match or 2L/c check), integrity tests of the piles. Strength and ultrasonic wave speed measurements on concrete samples cast at the time of installation of the piles and concreting of the pile heads were made available finally. Low strain testing had been performed on all piles both the
BBRI and CEBTP and results provided to predictors. 3 RECIEVED PREDICTIONS 3.1 Reporting format The predictions are reported herein under an anonymous format in order not to stigmatize those with less accurate predictions. Each prediction is however labeled with a code corresponding to the prediction type. Each predictor is thus enabled to position his own prediction within the cluster of results and encouraged publishing his prediction procedure, using the present paper as a reference. 3.2 Predictions types Results from 10 predictors had been received on November S’’ 1999, the ultimate submittal date. Contractors had also predicted the ultimate bearing capacity of their own piles (they were not asked to supply the load-settlement curves.) According to the reference data used to cast those predictions, the following labels have been used: - “CPT” for predictors using the CPT results. - “PMT” for predictors using the PMT results. - “LAB” for predictors using the laboratory results. - “DLT” for predictors using the Dynamic Load Test results. - “STN” for predictors using the Statnamic Test results. The CPT predictors used different methods, including ultimate state design as well as load transfer curves. All the contractors’ predictions were made using CPT results and De Beer’s 1974 method. The PMT predictors used the pressiometric approach that provides stress-displacement relationships for the shaft and the base. The LAB predictor used a loadtransfer functions method based on plasticity indices. The DLT predictors’ methods included either CAPWAP or SIMBAT: the soil parameters in a model are adjusted to get the best match between the measured and the predicted signals of a Dynamic Load Test. SIMBAT is an empirical method converting the dynamic reaction to a static reaction. The STN predictor used the Unloading Point Method (UPM) to predict the static load test. It was mentioned by the predictor that, due to strain rate sensitivity of clayey soils, a 30% reduction coeffcient had to be applied on the usual UPM method. A hyperbolic approximation of that reduced function was then calculated. This is the reason why those predictions are labeled as “0.7 STN”. It should be noted that the STN predictor was not provided with the results of the dynamic load tests, and that no TNO-WAVE prediction was submitted.
728
3.3 Predictions classes Predictions had to be made before static pile load tests were performed in order to qualify as Class A type predictions. If predictions were made after the static pile load tests, they qualified as C type predictions, according to accepted definitions of predictions classes (Lambe, 1973). Each prediction for each pile can be classified according to its submittal date relative to the date of static loading. Table 1 shows that most of the predictions are Class C. The only Class A predictors were CPT 1 (except for piles A1 & A4) and Contractors Atlas and Fundex. Other Class A predictions were those of predictor DLT2 for A3 pile and of predictor PMTl for C2 to C4 piles. 4 STATIC LOAD TESTS 4.1 Procedurefor static load tests The static pile load tests were to comply with the following loading guidelines, referring to Q,,,, the maximum anticipated test load, chosen with the hope to cause bearing failure: - A pre-load stage of maximum 5% of Q,,, was applied in order to check the measurement equipment and the centricity of the applied force, - 10 maintained load steps with equal AQ until Q reaches Q,,, - No intermediate unloading cycles Duration of maintained load step of 60 minutes - Load test was performed until a pile head settlement 2 15% Obase was reached - When the pile head settlement has reached a value of 25 mm, subsequent load steps can be applied using a smaller increment (AQ/2), in order to refine the pile load-settlement curve as it approaches failure, - Unloading in 5 steps of 10 min. each, except for final unloading (30 min at least of monitoring). The system used to apply the maintained loads on the piles called upon a sophisticated hydraulic regulation that guaranteed a tolerance of 5 kN. That system had just been developed by the BBRI. The 3 MN reaction was provided by a kentledge consisting of concrete blocks. Besides load and settlement monitoring, extensometers provided longitudinal strains along 5 to 7 shaft segments along the pile length. The results provided by those more detailed measurements are to be reported elsewhere. Such a procedure requires a value for the ultimate capacity R, of each pile. Those capacities were estimated by the BBRI and the national experts using De Beer’s method based on the CPT tests results (De Beer, 1974). The load increments A Q were actually: - R,,/8 for Atlas, Fundex, Prefab and Olivier piles. - RJ0 for De Waal and Omega piles.
The ultimate capacity was considered reached when the pile head settlement was equal to 10% Obase.It should be noted that a maximum Constant Rate of Penetration ( C W ) of 0.6 m d m i n was enforced towards the end of the loading procedure for all piles (except for piles A1 and A4). 4.2 Results Figures 1 and 2 show the various load (Q) - settlement (s) curves for the long and the short piles, respectively. “SLT” refers to the Static Loading Test. The predictors’ curves are also identified using the labels discussed in Section 2.2. “Contractor” refers to the ultimate capacity predicted by the Contractor. This value is drawn for 30mm < s < 50mm with a bold line. “Target SLT” refers to the ultimate capacity estimated by the BBRI. It is a “box” corresponding - 8AQ
729
Figure 2a : Short Piles Load Settlement Curves - Prefab, Fundex and De Waal
730
Figure 2b : Short Piles Load Settlement Curves - Olivier, Omega and Atlas
731
Table 1. Predictions classes Pile Long Long Long Long Long Long slim ’refab Static Test Date 3018 Submittal Date
5th Sept 99 9th Sept 99 10th Sept 99 1st Oct 99 4th Oct 99 27th Oct 99 28th Oct 99 29th Oct 99 29th Oct 99 5th Nov 99 5th Nov 99 ‘Contractor Fundex Franki Socofonda Olivier De Waal
12th Aug 99 26th Aug 99 3rd Nov 99 4th Nov 99 5th Nov 99
‘unde:
1519
--
)e Wai Olivier
1719
2319
___I
_s_
A
A
C C C C C C C
A
C C C
C C
C C
C C
C C
C C
C C
-
3megz
Atlas
5110
12110
Short
slim
Short Short
Olivier
3megi
2119
3019
2l9
719
1519
Atlas 7110
-- -- -- - A
A
A
__I
m
Prefab -unde: )e Wa;
A C C C C C C
A C C C C C C
C C C C C C C
A
C
-- -- -
A
A
C C C
C C
C C
C
C
C
A
A
C C C C C C
A C C C C
C
_ I
A
-
__.__
C
-
A A
A
C
C
C
C
C C --
C
C
-
while TNO provided an FPDS5 system free of charge. Bachir Dali, Paris, using an in-house laser system he developed, acquired displacements on a voluntary basis. The authors would like to thank Mr. J. -M. Couvreur, Research Assistant at UCL, for his management of communication with the predictors and of the information, including preparation of comparative figures. Finally this exercise would not have been possible without countless hours accrued amongst the several prediction team that bravely took up the challenge. Many thanks to those heroes; some of who will hopefully identify themselves through follow-up publications.
those methods encounter more difficulties in predicting pile behavior under large displacements. The predictions (DLT1 to 4 and 0.7 STN) are quite good within the service load range of the loadsettlement curves. DLTl and DLT 3 are very close to the SLT ultimate capacities even though they did not predict the pile behavior for settlements greater than 20 mm. DLT2 (Capwap method), DLT4 (Simbat) and 0.7 STN overestimates of the ultimate capacities would warrant the following approximate reductions: - DLT2by25%, - DLT4by50%. - 0.7 STN by 25 % (which means the 30 % reduction coefficient initially taken by that predictor should have been 50 %). If such reductions were applied to these predictions, they would however not fit as well the initial part of the SLT curves. The reduction of dynamic soil resistance to its static value still needs to be clarifi ed .
7 REFERENCES Lambe T.W. (1973), Predictions in soil engineering, Giotechnique XXIII, n02, p 149-202. Holeyman, A. et al., Design of axially loaded piles - Belgian practice in Design of axially loaded piles - European Practice, edited by De Cock and Legrand, Balkema, Rotterdam, 1997, pp.57-82 Holeyman et a1 (1999a), Prediction Invitation, Belgian Research Project “Ground Displacement Screwed Piles at Sint-Katelijne Waver”, August 20‘” 1999. Holeyman et a1 (1999b), Additional Reference Document for Predictors, Belgian Research Project “Ground Displacement Screwed Piles at Sint-Katelijne Waver”, August 29“’ 1999. Holeyman et a1 (2000), Preparation of an International Pile Dynamic Testing Prediction Event, Proceedings of the VIth SWC, Sao Paulo.
6 ACKNOWLEDGMENTS
The authors would like to acknowledge the Belgian Ministry of Economic Affairs for its financial support of the testing program (Convention CC CIF - 562). The ‘De Nayer’ Institute at SintKatelijne-Waver made the test site available. Contractors installed their own piles free of charge. Profound performed the Statnamic tests and CEBTP the integrity tests also free of charge. Dr. Klingmuller, of GSP, Mannheim loaned a PDA
732
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Preparation of an international pile dynamic testing prediction event A. Holeyman Universiti Catholique de Louvain, Louvain-la-Neuve, Belgium
J. Maertens Kutholieke Unviversiteit Leuven, Leuven, Belgium
N.Huybrechts & C. Legrand Belgian Building Research Institute, Brussels, Belgium
ABSTRACT: An international prediction event could be carried out on the basis of an extensive pile testing program organized in Belgium. Five units of six pile types were installed, allowing various testing methods to be applied: instrumented static load tests carried up to failure, Statnamic testing, and dynamic testing. This paper provides the project background information that was required to prepare the prediction event. The pile types are fully described while the results of the extensive soil investigation program are summarized. A companion paper reports on the received predictions and results obtained from the static pile load tests. 1 INTRODUCTION
2 PILETYPES
A national research project promoted by the Belgian Building Research Institute (BBRI) has been conducted in order to establish the performance of different types of cast-in-place ground displacement screwed piles. A national advisory committee under chairmanship of the first two authors directed the program, which included installation and testing of 30 test piles. The program has also been followed by an international panel of experts selected amongst members of ITC 18 (International Committee on Pile Foundations, of the International Society of Soil Mechanics and Geotechnical Engineering). A total of 30 loading tests could be performed at a site located in Sint-Katelijne-Waver, some 20 km north of Brussels, according to the following schedule: - 6 Statnamic tests and 12 dynamic tests took place within the first and second week of August 1999; - 12 Static pile tests were performed between September Znd and October 12* 1999. That timing allowed the organization of a Class-A type prediction event (Lambe, 1973), with the view to document the profession’s ability to estimate those new piles behavior based on standard investigation means as well on dynamic testing. An extensive soil investigation was performed as part of the research project, including in situ tests (CPT, PMT, SPT, DMT, SASW, ...) and laboratory tests on undisturbed samples. Whereas the present paper provides a description of the pile types and subsurface conditions, its companion (Holeyman et al, 2000) focuses on load tests results and their comparison.
Six different types of ground displacement piles were installed and tested: one prefab and five castin-place screwed types: - Atlas pile, installed by Franki Co. - De Waal pile, installed by De Waal CO - Fundex pile, installed by Fundex Co. - Olivier pile, installed by Olivier Co. - Omega pile, installed by Socofonda Co. Figures 1 through 5 illustrate the installation process of the five ground displacement screwed piles and define the pile toe level relative to the geometry of the augedscrew tip. Five piles of each type have been installed on the test site to accommodate the following conditions for each pile type: - One short pile for static load testing - One long pile for static load testing - One short pile for dynamic load testing - One long pile for dynamic load testing - One long pile for Statnamic testing The short and long piles had an approximate depth of 7.5 m and 1 1.7 m, respectively. The different pile types, their testing destination, their nominal shaft and base diameters for geotechnical bearing capacity calculations, and their measured pile base depths are listed in Table 1. A total number of 30 piles were thus installed according to the pile layout shown on figure 6. The following load tests were to be performed on the following piles referenced according to their grid line locations :
733
-
12 static load tests on piles A l , A2, A3, A4, B1, B2, B3, B4, C1, C2, C3, and C4 12 dynamic load tests on piles A5, A6, A7, A8, B5, B6, B7, B8, C5, C6, C7, and C8; and
734
-
6 Statnamic load tests on piles D1, D2, D3, D4, D5, and D6.
1
2
3
4
tests and laboratory tests, the locations of which are shown on figure 7: In-situ tests - 30 CPT(Cone Penetration Test)-E with electric cone in the axis of each test pile - 27 CPT-M1 with mechanical M1 cone (standard discontinuous penetration) - 3 CPT-Ml with mechanical M1 cone (alternate continuous penetration) - 4 CPT-M4 with mechanical M4 cone - 4 DMT (Dilatometer test) - 2 borings with PMT (Pressuremeter test) tests at 1 m intervals - 2 borings with SPT (Standard Penetration Test) tests at 1.5 m intervals - 1 boring for undisturbed soil samples - SASW (Spectral Analysis of Surface Waves) tests - Seismic Refraction Tests - Seismic Cone tests Laboratorv tests at several depths: - Grain size distribution - Atterberg Limits - CU - Triaxial tests (consolidated, undrained) - UU - Triaxial tests (unconsolidated, undrained) - Triaxial tests with Bender Elements
5
1. Setting up rig.
2. Screwing in displacement auger tip, optimally with vertical pressure. 3. Bringing in reinforcement (eventually after concreting). 4. Injecting concrete and pulling out auger head, while pursuing clockwise rotation. Lost point at pile base. 5 . Finished pile. Toe level determined by the level of the top lost bottom point
Figure 5. - Installation process of Omega Pile
3.2 Subsurface geology and properties
3 SOIL INVESTIGATION
Borings B1, SPTl and SPT2 revealed the following succession of soil layers : 0 - 0.40 m : rubble 0.40 - 0.65 m : Quaternary loamy sand 0.65 - 13.90 m : Tertiary O.C.Boom Clay
3.1 Overview The extensive soil investigation performed as part
of the research project included the following in situ
Figure 6 - Test Piles Layout
735
Table 1 - Features of installed piles
Pile
TEST
Pile type - Nominal dimensions [cm]
Shaft Diameter ( I ) ( 4 0.395 0.380 0.380 0.395 0.380 0.380 0.395 0.395 0.410 0.410 0.510 0.510 0.410 0.4 10 0.510 0.510 0.410 0.410 0.510 0.510 0.410 0.410 0.510 0.510 0.380 0.395 0.410 0.510 0.410 0.510
Prefab. 35x35 Static A1 Fundex 38/45 Static A2 Fundex 38/45 Static A3 Prefab. 35x35 Static A4 Fundex 3 8/45 Dynamic A5 Fundex 3 8/45 Dynamic A6 Prefab. 35x35 Dynamic A7 Prefab. 35x35 Dynamic A8 De Waal4 1/4 1 Static B1 De Waal4 1141 Static B2 Olivier 3615 1 Static B3 Olivier 36/5 1 Static B4 De Waal4 1141 Dynamic B5 De Waal4 1141 Dynamic B6 Olivier 36/5 1 Dynamic B7 Olivier 3615 1 Dynamic B8 Omega 4 1/4 1 Static c1 Omega 4 1141 Static c2 Atlas 36/5 1 Static c3 Atlas 3615 1 Static c4 Omega 4 1/4 1 Dynamic c5 Omega 4 1141 Dynamic C6 Atlas 36/5 1 Dynamic c7 Atlas 3615 1 Dynamic C8 Statnamic Fundex 3 8145 D1 Statnamic Prefab. 35x35 D2 Statnamic De Waal41/41 03 Statnamic Olivier 36/5 1 D4 Statnamic Omega 4 1141 D5 D6 Statnamic Atlas 3615 1 (1) Diameter governing soil failure along the shaft
Base Diameter (m) 0.395 0.450 0.450 0.395 0.450 0.450 0.395 0.395 0.410 0.410 0.510 0.510 0.410 0.410 0.510 0.510 0.410 0.410 0.510 0.510 0.410 0.410 0.510 0.510 0.450 0.395 0.410 0.510 0.410 0.510
Pile base depth ( I ) (4’ -7.39 -7.38 -11.50 -11.58 -7.39 -11.56 -11.63 -7.44 -7.53 -11.73 -11.68 -7.43 -7.48 -11.74 -11.77 -7.90 -7.67 -11.83 -11.76 -7.72 -7.53 -11.78 -11.61 -7.68 -1 1.54 -11.67 -11.58 -11.55 -11.71 -11.68
Excavation level ( 2 ) [ml -1.04 -1.07 -1.02 -0.98 -0.8 1 -0.75 -0.75 -0.71 -1.07 - 1.07 -0.97 -0.90 -0.76 -0.76 -0.79 -0.75 -1.08 -1.02 -0.95 -0.96 -0.84 -0.84 -0.74 -0.82 -0.84 -0.87 -0.75 -0.80 -0.78 -0.76
End level pile head (’) [ml -0.84 -0.87 -0.82 -0.78 -0.60 -0.53 Continuous Continuous -0.87 -0.87 -0.77 -0.70 -0.58 -0.52 -0.46 -0.58 -0.88 -0.82 -0.75 -0.76 -0.57 -0.53 -0.35 -0.48 -0.66 -0.57 -0.49 -0.66 -0.52 -0.48
Top level pile head ( 2 ) [ml +0.22 +0.27 +0.29 +0.36 +0.9 1 +0.97 +1.37 +056 +0.25 +0.28 +0.31 +0.39 +0.92 +0.94 +1.06 +0.94 +0.32 +0.35 +0.37 +0.39 +0.95 +0.96 +1.01 +0.99 +0.30 1-0.38 +0.43 +0.38 +0.48 +0.44
(2) Measured pile base depth relative to original soil surface level and according to the definition of the pile base level (figures 1 to 5 )
Figure 7 - Site Investigation Plan 0 Pile CPT M1-disc. 0 CPT M4-dis. DMT @
*
CPTM1-cont. A SPTeBoring 736
X
++
CPTE SASW PMT 0---o Seis. Refract
Figure 11 - PMT profiles of creep and limit pressures and pressiometric modulus (PMT2 Location)
Figure 9 - Typical CPT-E Log (EB5 Location) The properties of Boom clay, a stiff fissured and stratified clay belonging to the Oligocene, are well documented in the vicinity (De Beer et al, 1977): Natural water content: w = 22 to 3 1%
Liquid limit: Plastic Limit: Clay fraction Permeability 737
wL= 84 % w,=37% 55 % 1O-l0 m / s
UU triaxial tests performed for the research program confirmed the local variability of the properties resulting from the layered and fissured nature of Boom clay: C, varied between 80 kPa at 4.7m depth to approximately 150 kPa in the 8.5 to 13.9 m depth range. Those results confirmed the trend established by De Beer et a1 (1 9977) at the near-by Kontich site: C,[kPa] = 8 4 + 6 . 5 z[m] More discrepancy was found between the two sites regarding effective strength parameters derived from consolidated undrained triaxial tests conducted with pore pressure measurements: 4' = 27" and c'= 30 kPa for this testing program versus $, = 18" and c'= 11 kPa at the Kontich site. 3.3 In sitii testing
The bulk of the investigation effort was directed towards in situ geotechnical testing, which are generally recognized as the "ad hoc" testing for pile design in Belgium (Holeyman et al, 1997). The site was investigated using the several testing tools available to the profession, with a view to accommodate various geotechnical design cultures around the world. The site is herein characterized from the following angles: CPT-M1, as shown in Fig. 8 CPT-E, as shown in Fig. 9 SPT N-values profiles, as shown in Fig. 10 PMT, as shown in Fig. 11 DMT, as shown in Fig. 12 SASW, as shown in Fig. 13 4 PREDICTION ORGANIZATION
4.1 Prediction preparation On August 20th 1999, a reference document was distributed internationally among interested parties, and in particular to members of ITC 18 (International Committee on Pile Foundations of ISSMGE) and APTLY (Association of Pile Testing Laboratory) to enable them to: - predict the load-bearing behavior of the piles based on the results of the dynamic pile load tests, and - predict the static ultimate pile bearing capacity and the load-bearing behavior of the piles by means of the ground investigation results. A project synopsis had been prepared to invite interested parties to make those predictions (Holeyman et al, 1999a). It included a description of the pile types, site characterization, the static load test procedure, the dynamic load test procedure and the format of the prediction. Interested predictors were asked to fill in an invitation document to accept the information release conditions associated with this prediction event. In particular any publication using part of the data herein and public release of any of the research
Figure 13 - Typical SASW profile (SASW D Trace)
According to geologists, Boom clay was covered, prior to the Continental Pleistocene erosion, by a layer of Neogene sand with a thickness of approximately 40 meters. That layer has been completely eroded at Sint-Kathelijne-Waver. 730
measurements warranted the prior permission of the BBRI. Once this was done, they were sent the complete information, available as laboratory and in situ investigation and dynamic load test results (Holeyman et al, 1999b). 4.2 Prediction Format It was requested that the prediction submittal include: - A description of the used model(s), with a list of governing parameters, - The type of soil investigation method on which the calculations were based, - A detailed calculation methodology, with specific references (data provided, standards, publications, ...) and derivation of governing parameters, - A separation between pile base resistance and shaft resistance; - A criterion for the ultimate pile bearing capacity The predicted static load-settlement behavior of the piles was to be summarized into a table providing the loads corresponding to the following settlements: 1, 2, 4,6, 8, 10, 15, 20, 25, 30, 35, 40,45, 50, 60, 70, 80,90,100, 150, and 200 mm. Predictions could be established based on dynamic load tests, detailed geotechnical information (boring+lab tests, CPT, PMT, ...), experience, or a combination of the above.
7 REFERENCES De Beer, E., Lousberg, E., Wallays, M., Carpentier, R., De Jaeger, J., & Paquay, J. 1977. Bearing capacity of displacement piles in stiff fissured clays. I.R.S.I.A. - I.W.O.N.L., Comptes rendus de Recherches - Verslagen over Navorsingen, No 39, Brussels Lambe T.W. ( 1973), Predictions in soil engineering, Geotechnique XXIII, n02, p 149-202. Holeyman et a1 (1997) Belgian practice in Design of axially loaded piles - European Practice, edited by De Cock and Legrand, Balkema, Rotterdam, 1997, pp.57-82 Holeyman et a1 (1999a), Prediction Invitation, Belgian Research Project “Ground Displacement Screwed Piles at SintKatelijne Waver”, August 20‘h 1999. Holeyman et a1 (1 999b), Additional Reference Document for Predictors, Belgian Research Project “Ground Displacement Screwed Piles at Sint-Katelijne Waver”, August 29Ih 1999. Holyeman et a1 (2000), Results of an International Pile Dynamic Testing Prediction Event, Proceedings of the VIth SWC, Sao Paulo
5 CONCLUSION
The soil investigation performed as part of the program was extensive and included enough elements to allow most geotechnical engineering cultures to have a fair chance in the prediction event. Pile types were varied. Although ground displacement screwed piles are not widely known, the program included a prefabricated concrete pile as a more widely known reference pile. Conditions were set to assess the bearing capacity of those pile types and compare them with international predictions. 6 ACKNOWLEDGMENTS
The authors would like to acknowledge the Belgian Ministry of Economic Affairs for its financial support of the testing program (Convention CC CIF 562). The testing site was made available by the ‘De Nayer’ Institute at Sint-Katelijne-Waver. Contractors installation their own piles free of charge, is also very much appreciated.
739
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds) 02000 Balkema, Rotterdam, ISBN 90 5809 1503
Case studies of dynamic load testing in Japan Yasushi W&ya - Kawasaki Steel Corporation, Tokyo,Japan Kenji Nishiumi -Nippon Steel Corporatioiz,Tokyo,Japan Masahiro Hayashi - NKK Corporation, Kawasaki, Japan Atsushi Shibata - Kubota Corporation, Ichikuwa, Japan Shinji Nishimura - Fugro Geoscience Incorporated, Tokyo,Japan Tatsunori Matsumoto - Department of Civil Engineering, Kunazawa Universig, Japan
ABSTRACT: This paper reviews 69 cases of dynamic load testing performed in Japan on a commercial basis since 1983, and points out some subjects, such as 'set-up' phenomena and hammer driving energy, to be considered in the interpretation of the dynamic load test signals, in order to make the dynamic load test a more useful tool in the design, quality assurance and driving control of piles. The Japanese Association for Steel Pile Piles (JASPP) conducted their own dynamic load tests on a total of 13 steel pipe piles at different site conditions, the results of which are compared with the static load test results. This paper also presents the results of these comparative tests, emphasizing that adequate conditions for the dynamic load test are required to derive accurate estimations of the static load-displacement curve, as well as ultimate pile capacity, from the dynamic load test. driving control of piles. This paper also presents the results of comparative dynamic and static load tests on a total of 13 steel pipe piles at different site conditions, emphasizing that adequate conditions for the dynamic load test, such as proper hammer driving energy and a sufficient time interval after the end of initial pile driving, are required to derive an accurate estimation of the static load-displacement curve as well as the ultimate pile capacity from the dynamic load test.
1 INTRODUCTION
The first dynamic pile load testing in the field in Japan goes back to 1983, in which dynamic signals were recorded for the monitoring of stresses in a pile to prevent excessive impact. After that, research was focused on the estimation of the total resistance of a pile during driving from which only static pile capacity was estimated empirically. Next, researchers focused on the estimation of the static loaddisplacement curve for a pile as well as the ultimate pile capacity. The number of the dynamic load tests performed in a year in Japan has been increasing since the early 1990s. Foundation design in Japan is currently changing from the allowable stress design to the framework of the limit states design and/or the performance based design. In the new design methods, confirmation of the performance of each constructed foundation structure is of vital importance. Therefore, engineers have to select an appropriate test method to confirm the performance of the foundation, considering the balance of accuracy and cost. The dynamic pile load test is one of the promising test methods in this respect, because of its quickness and the relatively low cost of testing. This paper reviews 69 cases of the dynamic load test performed in Japan on a commercial basis since 1983, and points out some subjects, such as 'set-up' phenomena and hammer driving energy, to be considered in the interpretation of the dynamic load test signals, in order to make the dynamic load test a more useful tool in design, quality assurance, and
2 CHRONOLOGIY AND STATISTICS OF DYNAMIC LOAD TESTS I N JAPAN
The application of the dynamic load test at actual construction sites on a commercial basis is increasing in Japan. A chronological review and statistics concerning the dynamic load tests are presented, using the data measured by Fugro Geoscience Inc., Japan. Figure 1 shows the chronological variation of the number of test sites and the accumulated number of the test sites in which dynamic load testing was performed. In the first half of the 1980s, dynamic signals were recorded for monitoring of stresses in piles to prevent excessive impact. The use of the dynamic load test rapidly increased after 1990, and the total number of the test sites reached 69 in 1999, with 6 sites per year in average. Figure 2 shows the chronological variation of the average number of test piles per site, indicating a steady increase in the number of test piles per site 741
after 1990 The large number found in 1997 is associated with big projects involving highways and power plants Figure 3 shows a breakdown of the number of piles with respect to type of tested pile, indicating a large share of steel pipe piles, 89% of the total number Figure 4 shows a breakdown of the number of piles with respect to type of hammer used Hydraulic hammers occupy the largest share, 90% of the total pile number Diesel hammers were not used after 1994, because of problems of noise, vibration and pollution A merit of the use of hydraulic hammers is that the driving energy of hydraulic hammers can be controlled
Figure 4 Breakdonn of pile number nith respect to the type of haininer used
Table 1. Type of superstructure. Super structure Nuinber Pier Higliway bridge Quay wall Plant Building Breakwater etc Total
63 19 37 16 12 5 182
Ratio 35% 27% 20% 9?4" 7% 3% 100%
Location of test pile Onland Offshore 3 60 19 30 15 22 16 0 12 0 3 2 68 114 37% 63%
Table 1 lists the number of superstructures with respect to the type of superstructure. The number of each type of superstructure is hrther divided according to the location of the test pile, onland or offshore. The number of offshore locations of the test piles exceeds 60%. The dynamic load tests are mainly conducted on piles supporting port and harbor structures such as piers, revetments, and highway bridges. Figure 5 shows a breakdown of pile number with respect to the diameter of piles. About 80% of the piles have diameters greater than 1000mm, with a share of 53% by piles having diameters from 1200 to 1400mm. Figure 6 shows test (pile) number with respect to pile length. The number of tested piles having lengths from 40 to 50m is the largest. The maximum pile length until now is 91.5m. As mentioned already, most dynamic load tests have been conducted on offshore steel pipe piles having large diameters and large lengths. In such cases, and execution of static load tests is difficult, because it is considerably time and cost consuming. On the contrary, the merits of the dynamic load test become pronounced in such cases, because piles are installed mainly by driving, and the noise problem is not SO Critical in offshore pile driving conditions.
Figure 3. Bre&down of pile number \&h respect to type of pile.
742
Figure 8. Breakdonm of the pile nuniber with respect to the elapsed time at the dynamic load test after the end of initial pile driving
Figure 7 Breakdown of the pile nuniber vith respect to the soil t?pc of the pile toe bearing stratum
Figure 7 presents a breakdown of the pile number with respect to the soil type of the pile toe bearing stratum. Sand and gravel are the major soil types for the bearing stratum. The piles driven in sand stratum make up about half of the total number of piles, and the piles driven in gravel occupy a quarter of the total number of piles. Figure 8 shows a breakdown of the pile number with respect to the elapsed time at the re-driving test after the end of initial pile driving. The piles tested 1 week or more after the end of initial pile driving make up 73% of the total number of piles. Most dynamic load tests (re-driving tests) could be 743
conducted after a sufficient elapsed time, because these dynamic load tests were performed at largescale construction projects, where testing long after the initial pile driving was practical Figure 9 shows the relationship between the driving energy (ENTHRU), EH, transferred to the pile head and the product of the total resistance, RI, of pile during driving and set per blow, S The total resistance, R,, was estimated by the well-known CASE method Theoretically, R , + Sshould be less than EH Most data satisfies this relation, although several data violate this condition due to inaccurate measurements of set per blow, S It may be seen that the ratio, el = (R,.S)/EH, tends to increase, approaching 1, as the set per plow, S, increases For piles having S greater than lOmm, the ratio, er, is near 1 and greater than 0 6 at the least In such cases, it is thought that the soil resistance is fully mobilized by the pile driving On the other hand, the soil resistance does not seem to be hlly mobilized for piles with small S resulting in small values of er Small e,- would mean that a large proportion of the transferred driving energy is transformed to recoverable energy such as elastic deformations of the pile and the surrounding soil and is not used for the permanent penetration of the pile, because the soil resistance is not fully mobilized Figure 10 shows the relationship between the set per blow and the ratio, R,IR,. The static pile capacities, R,:, were estimated through the wave matching analyses of the dynamic load test signals. The RJRt ratio tends to decrease with increasing S until S reaches about 10 mm, and then to level off after that. As mentioned above, R, is not necessarily the maximum total soil resistance against the pile during driving in the cases of small values of S. In such cases, R, may not reach the yield load of the pile during driving. Hence, higher values of RJR, for small values of S mean that the dynamic component of the total soil resistance is relatively small due to the elastic response of the soil,
ratio is defined as the ratio of the static pile capacit), R,, derived from re-driving test to R, derived fr-oni the dynamic load test at the end of initial driving The plots in Figure 11 were selected from the dynamic load tests in which re-driving tests were conducted after sufficient rest periods from the end of initial driving for a full recovery of the pile capacity The set-up ratio tends to increase with increasing the pile length for pile lengths below about 40m, and level off after that length This result would indicate that the set-up ratio for the shaft resistance is larger than for the toe resistance The set-up ratio attains 4 in piles having length less than 6Om, whereas the set-up ratio ranges from I to 2 for piles having length over 6Om In actual piling sites, re-driving tests are usually performed with the same driving hammer as is used in the initial driving I n usual, a minimuni hammer that is capable to penetrate the pile to the design depth is selected Hence, the soil resistance is not fblly mobilized in cases of re-driving tests due to insufficient dribing energy, resulting in small set-up ratios apparently especially for longer piles
Figurc 10 Thc rclationshlp bctvccn tlic scI pcr blon. S, and tlic ratio U , I<,
whereas the dynamic component becomes larger if the pile reaches the yield load during driving, and slippage of the pile shaft occurs, resulting in high values of S greater than about lOmm It may be said that the static component ranges from 40 to 60% of the total soil resistance if the pile reaches the yield load Based on the results of Figures 9 and 10, S greater than lOmm or more is thought to be an indication of the fit11 mobilization of the soil resistance, although the authors are aware that it is only an approximate value, because the data in Figure 9 were obtained from the dynamic load tests on steel pipe piles having various lengths, diameters driven in various bearing strata. If the static load test results to be compared with the dynamic load test results are not available at the site, the relation shown in Figure 10 is useful in practice to judge roughly during driving whether- the pile has the required pile capacity or not Figure 11 shows the relationship between the pile length and the ‘set-up’ ratio. Here, the set-up
Figurc I 1 Tlic rclationsliip bctwccn sct up ratio and pilc Icngt h
3 COMPARATIVE DYNAMIC AND STATIC
LOAD TESTS ON 13 STEEL PIPE PILES The Japanese Association for Steel Pipe Piles (JASPP) conducted their own dynamic load tests at 13 different test sites from 1991 to 1995, while other engineers conducted static load tests Main purpose of these comparative dynamic and static load tests was to examine the applicability of the dynamic load test to the estimation of the load-displacement curve for a steel pipe pile as well as it’s bearing capacity Table 2 lists the test conditions, such as the hammer used, the test pile specifications and time intervals between the dynamic load test and the static load test, and summarizes the main results of the dynamic and static load tests. The ultimate bearing capacity, Q, the toe capacity, Qll, and the shaft capacity, Qr,. 744
Table 2 Main results of comparative tests of dynamc load test and static load tests Pile@ Total Pile capacity Pile capacity Tram Pile Re- Elap Pile capacity at SLT Elap (kN) time at end of drivlng at re-driving energy set bound time pile dia No Haininer (inin> length (kN) (kN) (kNm) (mm> (mm) (In)
Shaft Toe 1 STM500 2 DSLSO 3 DSL15
2000 62 0 13651 5468 800 7 3 0 800 1 1 5
4 DSL4S 5 DSLIS
609 37.5 609 3 9 0
6 DSL45
609 35.0
2724
411
20 12 1 465 186 1067 206
Total
Shaft Toe Total
19120 19424 5978 25402 3165 7213 141
960
196 7409 19 1009
Shaft 321 0 02
250
02
220
683
651 3679 1214 1923 1273 1861 570 2131
1320 3850 754 4604 11071 10672 2274 12916
28.2 121 2
9 HYD100
1600 52 0 11662 1312
15971 14501 3822 18326
295 3
17
92.8 90.6 86.3 60.6
1.2 1.5 1.6 12.0
51.0 32.0 82.0 18.0
STM Steam hammer
1735 2862 1480 323 1823 5958 1098 892
4711
8722 1998 13720 >14d
Ilh 100 32.5 1.6 10.0 115h 15 3 1111 250 120 64.6 2.0 16.0 105h
800 11.0 1000 7 6 0
700 508 800 400
29067 6213 35280 >14d
4292
118 4110 >11d
4059 6217
596 4655 31tl 515 6762 >14d
17
7 DSL10 8 DSL72
NYDl00 DSL45 HYD100 HYD65
43h
426
284 735 1259 61 8118 2656
10 11 12 13
Toe Total
1019 3925
4596 1803 7781 1989
918 1843
7860 960 8820 3410 823 4234 11731 4067 15798 1401 774 2176
23.0 5.0 0 1 220
66h 66h
4240 1245 5485
27tl
4200 158 4658 12279 1215 over 13524 22510 3920 26160
29d 90d
21 0 15mi n 27.0 67211 6105 2127 8232 24.0 138h L] - 3822 15.0 15h 10780 3920 14700 13.0 720h 1421 735 2156
70d l4cl l1d lltl 90d
DSL Diesel hammer HYD Hydraulic hammer
were estimated through the wave matching analysis of the dynamic load test signals In the cases of the static load tests, Q, and Cl, were separately estimated when axial forces of the pile were measured, otherwise 0 alone was measured Note that Q is defined as the load corresponding to a pile toe displacement of 10% of the pile diameter The load-displacement curve estimated from the wave matching analysis of the re-driving test signals is compared with the load-displacement curve obtained from the static load test in Figures 13 to 19 for 7 tests out of the 13 comparative tests For the 7 selected cases, set per blow, S, was greater than 1 2mm In cases 1 and 2, S was almost equal to 0 indicating insufficient driving energy to estimate the ultimate pile capacity In fact, the ultimate capacity derived from the re-driving tests for cases 1 and 2 are notably smaller than the ultimate capacity obtained from the static load tests In cases 3, 5 and 9, the time intervals between the end of initial pile driving and the re-driving tests were short (14 hours in cases 3 and 5 , 15 min in case 9), not allowing the completion of the set-up phenomena of the piles In these cases, the ultimate capacity derived from the re-driving tests is clearly smaller than those obtained from the static load tests It is interesting to note that the same hammer (DSL45) was used in cases 3, 4 and 5 and an ultimate capacity of 4 9MN was derived in case 4 745
Therefore, the small values for the ultimate capacity derived from the re-driving tests for cases 3 and 5 can be attributed to the fact that the set-up phenomena have not been completed at the time instants of the re-driving tests In case 8, the ultimate capacity was not obtained in the static load test due to insufficient load capacity of the loading device Therefore. case 8 was excluded from the comparison of the load-displacement curves obtained from the dynamic and static load tests Figure 12 shows the change with time of the setup ratio of each pile after the end of initial pile driving The values of set-up ratio are in a wide range from 1 1 to 10, although the set-up ratio is lower than 4 if the re-driving test is conducted within 3 or 4 days after the end of initial pile driving It may be very difficult to predict the set-up ratio accurately at the current stage of the pile technology Figures 13 to 19 show the load-displacement curves obtained from the static load test and derived from the re-driving test, together with the profiles of the soil layers and the SPT N-values, for piles 4, 6, 7 and 10 to 13 where re-driving tests were conducted at time instants sufficiently after the end of initial driving to permit a full recovery of the pile capacity Various commercial computer programs such as CAPWAPC and TNOWAVE, and KWAVE (developed by Matsumoto & Takei (1991)) were
Figure 14. Load- displacement cunes of case No.6
used for the wave matching analyses to derive the static load-displacement curves. However, the empirical soil model developed by Smith (1 960) was used in all of the computer programs. Even with such conditions, the derived load-displacement curve is in good agreement with the static load test results for each case. While some discrepancy between the derived ultimate bearing
capacity and the measured bearing capacity is seen in some cases, the initial portion of the derived loaddisplacement curve is fairly coincident with the measured curve, suggesting negligible influence of the computer program and the operator Table 3 shows a comparison of the calculated yield load and the yield load obtained in the static load test of each pile. The yield load is defined as the load corresponding to the first rapid increase in the pile displacement on the load-displacement curve. The 746
Table 3. Yield pile capacity derived from the dynamic and static load tests. Static yield capacit! (kN) Static load Matching Case No. error Test Analwis
4
4635 4998 429 1 5880 3822 1 1760 2940
6 7
10 11 12 13
4924 4843 3150 8397 4367 12576 2653
5.9% -3 2 % -3 4% 30.8Y" 12.5% -10 8% 6.5%
Table 4 Equations to estimate soil paraiiieters for the rational soil models for shaft iiiodel k , = 2 7% l ( d ). c, = G/175 8(; 3 2G for base model kl, = ____ . Ch =?r(l- c,)I -> 7T( 1 - v)d '
Poisson's ratio of soil. p, soil dciisitj, cJ pile diameter Shear modulus of soil. T ', shear u a\ e 1clocit> of soil
\c
(;
=pl :2
Figure 20 Ratioiial soil models (after Raiidolph & Siiiions 1986. Raiidolph & Decks 19'92)
difference between the derived yield load and the measured yield is within t10%, except for pile 10 where KWAVE program was employed for the wave matching analysis An error of I-tlOO/b may be acceptable, if the safety factor of 3 is used as is in Japan Even if the accuracy of the dynamic load test is less than the accuracy of the static load test, the increase in the number of tests performed at a site allows an increase of the design value for the bearing capacity (Hayashi et a1 2000)
Figure 21 W a ~ cmatching results n i t h the use or Smith iiiodel
Pile No 10 was re-analyzed by means of KWAVE with the use of the rational soil models developed by Randolph 8L Simon (1986) and Randolph & Deeks (1992) The rational soil models have been incorporated in the KWAVE program In the KWAVE program, the wave propagation in the soil 747
The soil density, p, was assumed to be 1800 kg/m3 for all the soils. In the wave matching analyses, the distribution of zm, and q b alone were assumed. The final matching results using the Smith model and the rational soil models are shown in Figure 21 and Figure 22, respectively. A better matching was obtained in the wave matching analysis using the rational soil models. The distributions of rmx derived in both the wave matching analyses are shown and compared with the static load test results in Figure 23. The corresponding static load-displacement curves are shown in Figure 24, together with the static load test results. It can been seen fiom these figures that the wave matching analysis using the rational soil models results in better predictions of the static load test results. Figure 22. Wave matching results with the use of the rational soil models.
4 CONCLUSIONS This paper reviewed dynamic load tests performed in Japan, showing their chronology and statistics. It was shown that most dynamic load tests performed in Japan have been conducted on steel pipe piles in offshore conditions for the acceptance of the tested piles. A total of 13 comparative cases of dynamic load testing and static load testing, which were conducted by JASPP, have been reviewed in detail. The following findings were presented: 1. It is difficult to predict the set-up ratio of a pile accurately. Hence, conducting a re-driving test of the pile is required to estimate the bearing capacity as well as the load-displacement curve for the pile. 2. The load-displacement curve derived fi-om the dynamic load test is in fairly good agreement with
Shaft resisitance, z (kN/m2)
-from Static load test _
I
-
Wave matching analysis with rational soil models
- - -Wave matching analysis with the smith soil mode Figure 23. Distributions oft,, derived from wave matching analyses using the S m i t h model and the rational soil models.
inside the open-ended pipe pile and the mobilization of the internal shaft resistance are taken into account, according to the modeling of the internal soil (soil plug) proposed by Randolph (1987). The soil parameters except for the maximum shaft resistance, zma,, and the maximum toe resistance, q b , were estimated from the equations listed in Table 4. The shear wave velocity, 6,of the soil was estimated Figure 24. Static load-displacement curves derived from the using the following empirical equation (Imai 1977); wave matching analyses and obtained from the static load test.
748
that obtained from the static load test until the yield load is reached, if both the tests are conducted at time points sufficiently after the end of initial pile driving.
REFERENCES Hajaslii. M . Matsuiiioto. T & Suzuki. M . 2000 Dliiaiiiic load testing on 102 steel pipe piles for bridge fouiidations on inudstone Proc 5th Int Conf 011 the .4pplicatioli of rhe Stre,r-Iim’e Theor1 to Pilec. Sao Paulo (to be published) Iiiiai. T 1977 P and S n a \ e lelocities of the ground iii Japan. Proc. 9th I C Y W E . T o b o . 1977 Matsuiiioto. T & Takei. M . 1991 Effects of soil plug on beha1 lour of drn en pipe piles. Soils and Foundations. V0l3 1. NO 2 11-31 Raiidolph. M F . 1987 Modeliiig of the soil plug respoiise duriiig pile d m iiig Proc 8th ,C E .-lsrnti Geotechtiicnl C’onf. Bangkok. Vol 2 6 1-6 1 1 Randolph. M F & Siiiioiis. H A . 1986 An iiiiproved soil inodel for one-diiiiensioiial pile drir ing anal! sis Proc 3rd 1111Coiif on \’zii?? .\Jet11 I I I Oj,?ihore Pilitig, Nantes 1-1 7 Raiidolph. M F and Decks. A J , 1992 Djnaiiiic and static soil inodels for alial pile response Proc of 3rd Iiit C ’ m f oii .4pplrcntion of Stress-11nve Tlieort to Piles. Hague 311
749
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)02000 Balkema, Rotterdam, ISBN 90 5809 150 3
Case studies of high capacity CFA pile testing in Australia S. Baycan Pile Test International,Melbourne, Vic.,Australia
ABSTRACT: Continuous Flight Auger (CFA) piling continues to be a cost effective solution employed throughout Australia to solve many foundation engineering problems. This paper aims to describe two examples of dynamic load testing on recent CFA projects, including a comparison with a static load test, A brief description of pile construction, dynamic and static load testing for piles constructed at the new Sydney International Airport and the new Adelaide Convention Centre will be outlined.
2 GEOTECHNICAL PROFILE
1 INTRODUCTION 1.1 Sydney International Airport
2.1 Sydney International Airport
As part of the upgrading of Facilities at the Sydney International Airport in preparation for the 2000 Olympic Games, Vibro-pile Aust Pty Ltd was commissioned by Transfield Constructions to design and construct foundations for the proposed extension works. The project required constructing foundations into widely variable ground conditions overlying a weathered rock profile. Due to limitations on vibration and disturbance to underground services approximately 500 number CFA piles varying from 0.6 to 0.9m diameter were constructed for the project. The test program consisted of a series of low strain integrity, dynamic and static load testing. Geotechnical information, pile construction records and dynamic test results were used to determine a CAPWAPmodel. Model results compared well with the static load test result.
The idealised geotechnical profile within the area of piling works at Sydney International Airport was characterised by fill, overlying variable relative density sand with interbedded layers of soft clay and sands of medium density. Very stiff sandy clay was found to overly the sandstone bedrock. CPT testing was used extensively to investigate the ground conditions and a summary of ground conditions is shown in Table 1. Depth 0-4m 4-5m 5-8m
Geotechnical Profile Med. Dense - Dense SAND Soft CLAY Med. Dense SAND
1.2 Adelaide Convention Centre A combination of ‘V’ (a Vibropile proprietory displacement pile) and CFA piles were adopted by Vibro-pile Aust Pty Ltd for the foundations at the Adelaide Convention Centre project. This project required careful planning due to the structure straddling a roadway and railway lines. A total number of approximately 300 piles were constructed. An example of a dynamic test result on a 500mm diameter CFA pile and the corresponding CAPWAPmodel for such a test is described.
2.2 Adelaide Convention Centre The geotechnical profile at the Adelaide Convention Centre site consisted of a variable layer of fill up to 3m in depth, overlying very stiff silty clays to approximately 14m, overlying very dense sand and gravel. Residual siltstone - hard clays underly the 75 1
sanddgravels below 18m depth. Site investigation consisted of borehole drilling and the use of SPT testing. The shaft of pile T3 was predominantly in very stiff - hard clays with the base at 16.5m in very dense sandy gravel.
Excess 36 t Rig M R Site WHUENIION, BDElilIDE 2182 Diaieter 588 NI Total Vol 4.551 n3 Borelin 16 nin Start 87% Pile P.TEST length 16\58 n Precharge 8.284 w3 Canclin 9 Rin Stop M:56 Ilate 19/82/88
Bore
Penetrate Toque
(dmin) (wdrev) ($1 8 1 2 3 8 288 8 58
Extract Pressure Oversupply Profile (Wnin) (bar) (11 (RR) 8 2 4 6 8 1 2 3 8 48 148 258 R 258
3 PILE DESIGN AND CONSTRUCTION Careful planning and site specific constraints required consideration from the Vibro-pile designers for the Sydney Airport project. These included piling within 50m of Boeing 747 aircraft and minimising affects on underground fuel lines and other services. CPT investigations were used extensively to establish indicative pile design parameters. Individual piles were designed on a combination of shaft and end bearing. Design parameters up to 45% greater than those recommended in the geotechnical report were adopted. Elastic pile design using PIGLET was used to initially design pile groups. Pile testing was planned early on in the project to justifi seemingly ambitious design assumptions. Pile testing was critical as results were used to back-analyse pile global settlement calculations. Pile construction was digitally monitored and relayed to the office - in real time to provide added assurance of pile construction quality. Vibro-pile used a similar on-board computer system PL20 to monitor pile construction for the Adelaide Convention Centre piles. The computer system is able to display construction parameters such as depth, penetration rate, torque, injection pressure and volumetric oversupply. This data is used to obtain a better CAPWAPpile model. An example printout of the PL20 system is shown in Figure 1.
1
Figure 1 Adelaide Convention Centre PL20 pile record T3
value of 150'?Aos* (where S* is the factored design load in ultimate limit state analysis, which is usually 1.35 x working load). The test acceptance criteria consisted of: (a)
4 PILE TESTING PROGRAMME
(b)
4.1 Sydney International Airport
(c)
Every pile constructed for the new Sydney International Airport upgrade was integrity tested using low strain, frequency response and sonic echo methods. Based on integrity test data, all piles were found to be satisfactorily constructed. A total of eight piles were dynamically load tested, however for brevity the results for pile T5 only will be reviewed here. The test system for the static load test comprised of jacking against steel beams with a lOOOtonne capacity hydraulic actuator. Tension piles were used as reaction. Pile movement was monitored by string POtentiometers, with dial gauges and dumpy level readings used as backup to verify data. The loading cycle was applied in accordance with AS2159 . The load-unload and reload sequence comprised of loading in additional increments of 15% up tc 2 752
pile top settlement 4 0 m m at design working load, Net pile settlement <5mm upon removal of test load, Pile top settlement < 25mm at test load = 2xdesign working load.
The specification included an allowance for elastic shortening of the slender test piles, so that the limit of gross settlement could be increased by the amount of calculated elastic pile compression exceeding 1Omm. All piles tested were constructed using concrete having a 28 day characteristic strength f c 2 40MPa. Pile construction, load requirements and test details are listed in Table 2. 4.2 Adelaide Convention Centre As with the Sydney tests, a purpose built 12tonne air hammer with a seating frame was used to apply the test load. Hammer drops were varied between 0.3
and 1.2m. All dynamic testing was conducted using PDA equipment designed and manufactured by PDI -USA.
Adelaide
Pile
Length
Date
500m
16.5111 19/2/00
12/3/00
2300
5 TEST RESULTS AND ANALYSIS 5.1 Sydney international Airport - Static & Dynamic Load Tests The static load test on pile no T5 was carried out before dynamic load testing in order to be able to calibrate results for application to other piles tested on the project. A maximum load of almost 8900kN was recorded at a load of 150% of S*, which is equivalent to twice the working load. Measured gross settlement at this load was 25mm. As described, the acceptance criterion for the static test included an allowance for elastic pile shortening (19mm for 32.9m pile). Consequently, gross settlement could not exceed (19-10+25)= 34mm i.e. almost 5% of pile diameter. PDA test results were signal matched using CAPWAPand results suggested a mobilised pile resistance of 8000kN, with over 60% of the load carried in shaft resistance. A comparison of the static load test and the CAPWAP prediction is shown in Figure 2. Subsequent to the results, designers were able to be confident in their initial assumptions regarding pile parameters and to further refine their pile design. 5.2 Adelaide Convention Centre Dynamic Load Test This pile was selected to reveal some of the parameters which affect the interpretation of PDA data from a cast-insitu CFA pile test. To enable reliable data to be obtained, four sets of strain gauge and accelerometers were attached to the pile for dynamic load testing. The 500mm CFA pile described here was testing with four sets of PDA strain and accelerometers attached to the extended section of the pile. Other dynamic tests on the site were conducted with gauges attached to the actual pile section with 0.5m of ground around the pile being excavated. That approach was taken where the pile extension section was cast at a different time to that of the actual pile. In this instance, the interpretation of the wave speed was made simpler by using casting the top section of pile T3 at the same time as the actual pile. A wave
753
Figure 3 Adelaide Convention Centre PDA test result CFA 500 pile T3
speed of 3950ds was deduced, with a maximuni transferred energy of 85kJ corresponding to a hammer system efficiency of 60%. A typical set of PDA results for the blow described is shown in Figure 3. It is of interest to note the lack of forcehelocity proportionality at peak. This was concluded to be due primarily to the increase of impedance from a 400mm diameter extension above ground to the actual pile 500mm nominal diameter. The maximum compressive stresses was greater than SOMPa, for a transferred energy of 85kJ applied
out on T3 was not able to mobilise the available base resistance.
1
I
2 Current
6 DISCUSSION 2 Modified
z
The Sydney International Airport static load test proved that an effective, early planned pile-testing program was able to achieve efficient outcomes for a CFA piling contract. By using a static load tested to failure, calibration of dynamic load test results was undertaken. This lead to added confidence in the design assumptions. Back analyses of data allowed design assumptions to be verified. Numerous pile testing programs for socketed piles in rock have shown that unit resistances employed can be significantly greater than those normally adopted in design. Published examples include Tchepak (1998), Baycan (1996) and Seidel et a1 (1 998).
Suggest
Figure 4 Adelaide Convention Centre CAPWAP Impedance Profile Pile T3 Load in
KN
000 6000 8000
0.
5. Ru = 6018.6
KN
= 3912.1 = 2i06.5
kN
10.
Rs
15.
Rb Dy = Dmx =
Displacement in KN/m 800
1
15.0 16.0
7 CONCLUSION
KN mm
Dynamic load testing of piles combined with a static test such as that employed on the Sydney International Airport project, can allow a correlation to be used effectively to save time and money on piling contracts. The interpretation of a dynamic load test on a non-uniform CFA needs to take into account the geotechnical profile, concrete properties of pile and extension, pile construction records and any other effects of the test setup. By making use of such data, a reliable static model of the pile is determined.
mm
mm Shaft Resistance Distribution
Pile Forces at Rut 10000 kN
8 ACKNOWLEDGEMENTS
Figure 5 Adelaide Convention Centre - CAPWAP result pile T3
with a 1.2m drop of the 12 tonne hammer. A mobilised resistance of 5900kN using an assumed Case Damping Factor of Jc=0.7 was achieved. The recorded set for the blow was 2.6mm. Due to the added complication of a variation in impedance and pile concrete properties (and therefore wavespeed), for CFA and for all cast in situ piles, a review of all available piling records and geotechnical profile becomes critical. The assessment and building of a CAPWAPmodel for pile T3 required such an assessment. Using the PL20 record as guidance, the nonuniformity of the CFA pile was modeled in CAPWAP and the adopted profile is shown in Figure 4. Using the adopted parameters, a reasonable signal match was achieved in CAPWAP and results are summarised in Figure 5. It is well known that in order to mobilise the full available base resistance of a pile, sufficient movement is required. It is expected that the test carried
The author gratefully acknowledges the support and advice of Dr Julian Seidel. 9 REFERENCES AS2159, 1995. Australian Standard, Piling - Design and Installation Published by Standard Australia, ISBN 0 72629884 0 Baycan, S. 1996. Field Performance of Expansive Anchors and Piles in Rock. PhD Thesis, Department of Civil Engineering, Monash University, Australia. Seidel, J.P. and Haberfield, C.M., and Baycan, S. 1998. Load displacement performance of bored piles in weak rock. Deep foundations on bored and auger pile, Van Impe (ed), Balkema, Rotterdam. Tchepak, S. 1998. The design and performance of bored piles in shales for the Australia Stadium project. Deep foundations on bored and auger piles, Vam Impe (ed), Balkema, Rotterdam.
754
11 Supplement
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama & Beim (eds)0 2000 Balkema, Rotterdam, ISBN 90 5809 150 3
Keynote lecture: trospective of Sonic Integrity Tests - Its application to the control quality on piles Juan J.Goldemberg Geotecnica Cientic, Buenos Aires, Argentina
ABSTRACT: A retrospective of the past twenty years since the First Stress Wave Conference of Stockholm is made, narrating the evolution of Sonic Integrity Testing since its origins until today from the two major equipment manufacturers. As well, three case histories are presented, highlighting the necessity of introducing SIT tests as a regular quality control in all piled foundation sites. 1 INTRODUCTION.
systems, and the market conditions that encouraged the development and current applications of this technology. In addition, some cases histories are described of Deep Foundation Pathologies, detected through the performance of Sonic Integrity Tests and simulation techniques for defect quantification.
The purpose of this paper is to look at the technical data and documentation on Sonic Integrity Tests (SIT) and the major breakthroughs that have taken place over the last twenty years since the First Conference in Stockholm, Sweden, by the initiative of a few pioneers in Stress-Wave Theory and it Applications to Foundations Piles in 1980. Since then, the number of participants as well as the volume and content of papers submitted to each Conference, even from the so-called “peripheral” countries has been on the increase. Undoubtedly, this reflects the marked interest and concern shown by engineers across the world to be able to determine what happens to deep foundations once they are installed- irrespective of the procedure employed. A great number of applications have been derived from the Strcss Wave Theory, among them: Pile Driving Analysis (PDA) Dynamic Load Test (DLT) Rapid Load Test STATNAMIC (RLT-STN) Vibration Monitoring (VM) Monitoring Hammers (MH) Pile Driving Prediction with Impact Hammers (PDP-IH) Pile Driving Prediction with Vibratory Hammers (PDP-VH) Soil-pile Interaction Modelling and Signal Matching Technique (TNOWAVE, CAPWAP, etc.). Sonic Integrity Test (SIT) This paperlodks at the basic documentation on Sonic Integrity Tests underlying a rational interpretation and evaluation, the fundamental contribution of data processing and compilation means and
2 CONFERENCES ON THE STRESS-WAVE THEORY Conferences on Stress-Wave Theory continuc to be held successfully every four years since 1989 as witnessed by the increasing participation of the international engineering community and a larger number of papers being submitted at each Conference [ 11, [2],[3], [4], [ 5 ] ,[6]. Table 1 shows an increasing interest in Quality Control on deep foundations as a result of a greater knowledge of the available methods, equipment and signal evaluation systems. Table 1. SIT Papers presented at Stress Wave Conferences Conference 1980- Stockholm (Sweden) 1984- Stockholm (Sweden) 1988- Ottawa (Canada) 1992- La Naya (Holland) 1996- Orlando (U.S.A.) 2000- San Pablo (Brazil)
Papers on SIT 5 5 13 16 21 14
Total PaDers 24 54 88 107 101 100
As a result, Sonic Integrity Tests- as they are in most inknown today- have been dustrialised countries and are being done in several countries where SIT^ are beginning to be applied. 757
Another result of these International Conferences are the Alliance of Pile Testing Laboratory Engineers (APTLY), an association recently set up during 1995 by users and manufacturers of these technology, as well as the periodic Seminars (among them, Control Quality on Piles: Uses & Abuses) on the occasion of the annual meetings of the Deep Foundation Institute (DFI). That is how nowadays, the Quality Control through Non Destructive Tests (QC-NDT), has begun to be included in the Technical Specifications for the construction of piled foundations in most Civil Engineering works, by private companies or governments from those countries that have adopted this technology. The availability of this technology has resulted in an early detection of defects likely to jeopardize the stability of deep foundations and the superstructure resting on them and reduced the occurrence of accidents likely to lead to the loss of both lives and assets.
Geotechnical Engineer, particularly in South America, during the Preliminary Project, Project Preparation and Construction stages. In most cases, Geotechnical Engineers participated only in the “soil investigation” and failed to be included in the preparation of a Geotechnical Research Plan necessary for proper subsoil identification based on the available data, subsoil characteristics and site specific requirements. Upon completing of the “soil investigation”with the scope and vision of a merely formal requirement in most cases- specialists are kept away from the tasks within their specific field of expertise, such as the analysis of foundation alternatives, determination of load capacity of foundations and prediction of foundation behaviour, geotechnical and structural design (which is beyond a straightforward determination of the “allowable stress” as is mistakenly believed), selection of the construction method (or alternatives) better suited to local conditions, design of containment and anchor structures, excavation procedures, groundwater abatement project, drainage network and relocation of digging material, soil compaction from mechanical, chemical or thermal causes, soil filling project (including sanitary fillings and embankments). Last but not least, the necessary Quality Control on soil movements and/or upgrading, contention and anchor, underground and open pit excavations, and foundations, among other things. Worse still, some of these tasks are sometimes carried out by the Client without the support of properly equipped laboratories or personnel lacking the necessary experience or post-graduate training and, in some cases, personnel alien to the discipline or lacking the necessary skills. It can then be concluded that Foundation and Contention Project Design and Construction and consequent Safety and Economics are, at least, “unstable”. This is unacceptable for any structure designed under the Static Principles. Because the usual design criteria have failed to catch up with the current building practices in developed countries and the quality precepts associated with safety and economics are slow in being adopted, deep foundations and contention structures continue to be, theoretically, over- (or hyper ?) dimensioned with the resulting loss or poor use of, among other things, intellectual resources, materials, time and labour ... in other words, money. The above circumstances and the many cases of structural and foundation pathologies we were involved in our region, paved the way for a study of the way in which these problems were approached and solved in developed countries. Our study allowed us to be in contact with the European and U.S. experience, scope and limitations of NonDestructive Testing technology derived from the
3 QUALITY CONTROL ON PILES , situation of most In the late ’70s and early ’ ~ O S the industrialised countries and some “developing” countries can be summarised as follows:
3.1 Growth of Construction.
This growth was characterised by a strong flow of capital to the construction and/or upgrading of infrastructure projects, such as roads and bridges, thermal, nuclear and hydro power stations and associated transformer stations and transmission lines, industrial and petrochemical plants, airports, port facilities and housing developments. In many cases, these projects required the construction of piled foundations with high design loads. This requirement led to the introduction of new technology for large diameter cast-in place concrete piles under bentonite mud. 3.2 Inadequate and/or insufficient Quality Control
A long-standing general concern in the construction community has been the safety of foundation works, in particular those associated with excavations and piles. This concern arose from the very nature of these structures, the imperfect knowledge of the soil profile, wide range of building methods and customs, general lack of effective quality control methods, lack of tests and measurements on thc behaviour of the soil-structure interaction and a certain blend of pseudo-confidence and fatalistic resignation (or hope ?) on “...pile contractors surely know what they are doing ...” In addition to the above conditions, there has bcen a marked absence of involvement from the 758
Stress Wave Theory and to decide on its introduction to the services provided by our Company. Thus, in 1986, the first tests were performed by us using equipment developed and manufactured by TNO - Building and Construction Research (DeIft, The Netherlands) [7].
struction” [9] organised by the Institution of Civil Engineers- ICE- in London on November 4, 1998) clearly show the incidence of the possibility of cost optimisation consistent with the Project Development Stage and Costs of the sometimes unavoidable and sometimes unforeseeable -Scope Changes depending on the same parameter. In these cases, while the immediate increased costs are paid by the Owner or Concession Operator, the final actual costs are paid by the Community at large.
3.3 Market Demand: Quality Assurance (QA).
The current trend in most industries is to attain the necessary targets to reassure the industry, its customers and the community at large, and that all quality standards or criteria have been met at each stage. This involves systematic compliance with each and every procedure leading to the achievement of planned objectives within each area in the organisation of the purposes of meeting the Quality Policy pursued by the Company, the Owner and/or the Consumer (or the Community itself), within the scope of adcquate conditions in terms of safety and economics. This approach, which is not opposed to the old concept of Quality Control (somehow associated with an “autopsy”, “death certificate” or “postmortem” investigation) is helpful on the basis of a set of tools and modern techniques based, in most cases, on rapid Non-Destructive Tests (NDTs) sufficiently known all over the industrialised world, numeric analysis methods and simulations, and control procedures developed for rapid accurate diagnosis of the detected anomaly. The fitness of these tools for their intended purpose is dependent, prior to the implementation of a successful Quality Control System, on a strict performance of quality standards associated with the scope of the Architectural and Engineering Preliminary Project, study of the available data, project feasibility study, essential construction service availability study, regulatory and environmental provisions, basic geotechnical studies (including feasibility of using local materials for fillings and construction, and final disposal of digging materials and waste generated by the works), Executive Projcct, analysis of alternatives, as well as geotechnical and structural foundation, excavation and contention and anchor design and, if needed, abatement, drainage and/or digging materials relocation. Successful performance of these tasks will ensure an efficient construction in terms of quality, safety and cost-efficiency. This requires specialists in each trade to meet the Quality Assurance (QA) standards for the project. The increasingly accepted principle of having a geotechnical specialist during each stage of project preparation and implcmentation has led the most developed countries to stage Technical Meetings, Seminars, and Congresses to analyse this issue in greater detail [8]. Figure 1 (quoted by R.P.Thompson during the Seminar on “The Value of Geotechnics in Con-
Figure 1. Changcs in “Reduction Cost Potential” and “Change Order Costs” relative to the time schedule.
3.4 Needs of Industrialized Countries
World concern for the energy crisis generated by geo-political instability in the Middle East attributed great priority to the search, exploration and operation of new oil and gas fields, in particular those located in the North Sea and the Mexican Gulf as a result of their geographical proximity to the natural high demand markets. As an immediate result of the above, there was the urgent need to build “off-shore” and “near-shore’’ platforms which, as usual in most cases, are supported by piles. This drove to rapid reliable solutions to predict and check the bearing capacity of piles under extreme performance requirements and environmental conditions calling for the traditional Static Load Tests (SLT) to check the design assumptions adopted. For this reason, a number of International Conferences and Seminars have been held since the ‘60s to present new developments in Dynamic Tests and Numerical Analysis Methods on pile driving and load capacity associated with off-shore structures. When evidence was obtained of the actual suitability of certain tests and analysis of results, Sonic Integrity Tests (SIT) emerged, almost naturally, as the “younger brother” of other prior tests with similar performance assumptions relative to the Stress Wave Theory.
759
3.5 Scientijk Support Necessary for Technology Development Most likely, the first references to the propagation of lengthwise vibrations along elastic bars inspected during our undergraduate training must have included the works by Love, A.E.H. (IO), Timoshenko, S. & Goodier, J.N. (11) and Timoshenko, S. & Young, D.H. (12). If the cross section of the bar remained flat during deformation (Navier-Stokes assumption), all these works proposed a general equation for movement of the following type: a2U/o't2= c2 a2U/aX2 in which c = (E/p)o's where: U: displacement of the section being studied in direc tion x (longitudinal 1y). c: velocity of the stress wave. E: dynamic elastic module. A: cross section of bar. p: density of material. The general solution of the equation (1) leads to: + f,(x-ct) (3 1
U = f(x+ct)
and using the notation proposed by Voitus van Hamme et al. (1 3), we obtain: u = f (x-ct) + f (x+ct) (4)
a) Free end (F = 0, equivalent to a pile toe on soft soil). b) Fixed end (v = 0, equivalent to a pile toe on rock). c) Toe with some end bearing (F = Fp, equivalent to a intermediate condition). d) Change of impedance (Z1 > 2 2 or Z1 < 22, equivalent to an anomaly detected with SIT). e) Impact (equivalent to the driving condition during PDA or redrive during DLT). 4 REVISION OF THE MAIN TECHNOLOGY DEVELOPMENTS He have considered gathering the information delivered by the two main sources that manufacture equipment, including hardware and software: P The European, represented by TNO Building & Construction Research, Delft, (Holland) and, P The American, represented by PDI- Pile Dynamics, Inc. Cleveland, Ohio (USA). 4.1 The main developments of TNO
The equipment produced by this R&D Institute is characterised by the generic denomination of FPDS- Foundation Pile Diagnostic System and it is described as follows (1 4):
P Equipment FPDS-0.
Likewise, the force applied on the section is given by: F = -EA du/dx F = -EA df /d(x-ct) - EA df /d(x+ct) (5) F=F +F (6) In turn, the velocity is expressed by: v = duldt = v = - c df /d(x-ct) + c df /d(c+ct) v = v +VI
(7) (8)
'These two velocities are linked to their respective f ~ r c e by s the following expressions: (9) v =Fi/Z v = - F /Z (10) where Z = EA/c Z : bar impedance. The application of the above equations to the different situations arising daily in the pile industrycoupled to the powerful new "high tech" developments- have paved the way for a rational analysis of the signals obtained from SITS, PDAs and DLT tests, as a result of which it has been able to differentiate the following cases:
Photograph 1. Equipment FPDS-0 system with an oscilloscope and Polaroid camera
P Equipment FPDS-1. In the mid 80s, the FPDS-I was developed, based on a UNIX workstation and a MC68000 (1 6 bit micro-processor chip), with built-in keyboard, disk drive, graphical display, and printedplotter. The signal was digitised (12 bit) at an early stage after retrieving the signal to ensure that all subsequent processing was of the highest possible standard. The system had certain automatic functions, but the 76@
processor) and a subsystem with an analogue-todigital (A/D) converter board and a specially designed board for integrity testing. As with FPDS-2, the subsystem could accommodate different FPDS applications, such as PDA/DLT, the Statnamic load test (STN), vibration measurements (VIBRA), and others. Each option consisted of specially designed boards, sensors, software, and accessories, an innovation begun with the FPDS-2 system. Automatic signal conditioning freed the operator from bridge balancing and scale selection. To reduce weight while conducting multiple SIT tests, the signal conditioner was built into the GRiD computer (called the Consultant System). With FPDS-3, SIT was also extended to include an optional instrumented hammer (an accelerometer fitted in a hammer head) to record the applied impact blow. The design of the hammer head, impact cap, accelerometer sensitivity, and hammer shaft stiffness produces a load pulse signal perfectly representative of the impact amplitude and phase. Since force is proportional to velocity at the pile head, the shape of the force and velocity traces are identical until a reflection is recorded. By comparing the force signal to the velocity signal, defects in the first 2 meters of the pile are revealed. The force is equivalent to the pile head impedance ’ velocity.
skilled operator could also override, magnify, or clarify a result to facilitate interpretation. The FPDS-1 system is still used today by clients for SIT.
Photograph 2. Equipment FPDS-I .
9 Equipment FPDS-2. With FPDS-2, the computer and signal conditioner were separated, a philosophy continued with subsequent systems. The FPDS-2 system, introduced in 1986, was based on a portable IBM-PC/AT compatible computer running under MS-DOS, equipped with a 20 MB hard disk for program and data storage and a floating point co-processor. For each FPDS application, specific sensors, signal conditioning subsystem, and software were available, such as Pile Driving Analysis and Dynamic Load Testing of foundation piles (PDNDLT). The FFDS-2 system did not require knowledge of electronics, requiring only a limited number of components and no oscilloscope or tape recorder. Real time presentation of signals, automatic controls, warnings, and error messages were a big improvement on earlier systems.
Photograph 4. Equipment FPDS-3 Photograph 3. Equipment FPDS-2.
9 Equipment FPDS-3. At the beginning of the 90s, the FPDS-3 system was developed, based on a GRiD computer (80386
P Equipment FPDS-4. In 1993, FPDS-4 was developed as a very robust and economically designed lunch box type system. The FPDS-4 system includes a 80486 processor, 8°-200 MB hard disk, and a built-in graphical 761
screen and key board, both sealed to protect against rain and dust.
P Equipment FPDS-5. Recently, the FPDS-5 system was developed using a notebook computer and separate signal conditioning and A/D card subsystem. Data communication from subsystem to notebook is via a PCMCIA card and cable. The FPDS-5 system is suited for PDNDLT and STN, but can also be used for SIT.
Photograph 7. Equipment FPDS-6, Field computer with SIT PCMCIA card.
P Years 1960/1979. Hardware: Development of analog instrumentation and computation technology. Software: Digital integration in the laboratory. Beta Method published.
Photograph 6. Equipment FPDS-5, Notebook with SIT PCMCIA card.
P Equipment FPDS-6. As well, TNO's latest FPDS system, the FPDS-6, represents a new approach to foundation testing equipment. The FPDS-6 is a hardware/software package for SIT, designed to work with a wide range of IBM compatible PC computers equipped with a PCMCIA card slot. The SIT hardware consists of a hammer, sensor, cable, and PCMCIA card. The PCMCIA card, measuring just 8.5 cm by 5.5 cm by 5 mm, incorporates all data acquisition (300 kHz, 14 bit) and signal conditioning electronics. Besides the enormous space and weight savings, this innovation dramatically increases overall system reliability and service life. The user can use his own notebook computer or a hand-held Hardbody PC. All input is via an easy to use touch screen. Knowledge of electronics is unnecessary-all electronic functions are controlled by the computer. Advice and warnings are generated by the system. Signals and data are stored automatically and can be easily recalled. Automatic reporting capabilities greatly reduce reporting time.
P Years 1983. Hardware: a) Data acquisition: For low strain, highly sensitive accelerometer and power supply; digital data acquisition and integration with Pile Driving Analyzer (PDA); digital display on oscilloscope. b) Kecording: Oscilloscope, photos of oscilloscope traces, analog FM magnetic tape, X-Y plots. Software: Analysis in time domain
k Year 1985. Hardware: IBM PC or compatible 8087 math coprocessor, 5 12 kRam, A/D board (acceleration records integrated and digitised) , signal amplifier. Software: m e g r i t y Program, exponential amplification, averaging of records, smoothing.
P Year 1989. Hardware: PIT test equipment: PC + signal conditioning; PIT processor and plotter or graphics printer in field; permanent storage on disk. Software: Analysis in frequency domain.
4.2 The main developments of PDI
P Year 1990. Hardware: PIT-SC: battery powered, signal conditioning and printer built in a 80286-based PC computer. Instrumented hammer.
The equipment manufactured by this company and its respective innovations are described in the following list (1 5): 762
5.1.2 Problems associated to drived piles
Software: Built in time and frequency domain software. Impedance profile. PITWAP (calculates pile shape given a soil model, based on wave equation).
5.1.3. Problems associated to external causes in the manufacturing process.
>
A detailed description of the problems and the methods available for their detection has been clearly exposed by M.J. Turner (19) and by J.J. Goldemberg & H. Goldemberg (20). As typical examples of the anomalies detected in foundations built with cast-in-place piles the following deserve to be highlighted:
Year 1992. Hardware: PIT Collector, touch screen, 16 bit A/D converter, high and low pass data filtering; signal amplification; interface to plotter/printer; full day battery; storage of 350 piles. Software: Built in analysis software in time domain. Post processing in frequency domain.
>
>
Year 1996. Hardware: PIT Collector memory expanded; improved display capabilities. Software: PITSTOP (frequency domain, 2 velocity, multiple plots, impedance profile).
Case History I.
Corresponds to a sector belonging to a group of Towers of 1 1 stories destined to economic housing, built in the south zone of Buenos Aires (Argentina) and where, due to contractual reasons, two of the mentioned buildings remained for a long time with their foundations (piles, caps and foundation beams) partially built without their superstructure completed. The geotechnical profile of the building site (see Figure 2) shows a first layer of low plasticity clays (CL) up to 4.00/4.50m depth with N value (SPT) between 3 and 16, laying on a non plastic silty strata (ML) with similar respond to the SPT test, up to a depth of 6.00m. From here onwards, the soils prcsented a strong preconsolidated characteristic with N>40, up to well below 10.00m from zero level. At the moment of the investigations, the water table level was 2.00/32.50m below ground level. The foundation system used consisted of displacement piles by driving a recoverable close ended caisson and casting the concrete while the mould was pulled out with the help of a vibratory hammers. Their nominal diameters were: 300/350 and 400mm and length of 4 to 6m. When the continuation of the building was decided, by the Main Building Contractor’s Structural Advisor, the Quality Control upon the total number of the piles
P Year2000. Hardware: Improved memory (1000 piles) and processing speed. Software: PIT-W (Windows) 5 COMMON PATOLOGIES IN DEEP FOUNDATIONS. In order to describe the principle causes of pathology detected in the foundation structures, it is convenient to list them into: 5. I Problems originated in its construction.
5.2 Problems derivedjkom its use during lif. time of the structure.
5.3 Problems due to geotechnical- environmental phenomenon. In this paper a reference to the problems generated during the construction of cast-in-place piles will be made for being the most common in our deep foundation market and the ones which have more possibilities of showing anomalies, according with the authors experience. In general the problems that can turn out of the construction of different piles and deep foundations, had been described in a number of publications: texts, seminars, conferences, etc. (16), (17), (18). In order to identify those causes of damage, problems can be classified in the following way: 5.1. I Problems piles.
associated
to
cast-in-place
Figure 2. Existent geotechnical profile in the building site.
763
of both buildings through the performance of the Sonic Integrity Tests was arranged. In accordance with it, the 108 piles of each building were tested, upon preparing its in order to eliminate all the material “that could be loose”, concrete contaminated by the soils, cracks caused by the trimming work due to use of inadequate techniques, etc. That could originate false signals. Given that some of the piles had the cap already built, they were tested under two different conditions: a) from the upper face of the cap and b) from a concrete brick expressly attached to the upper sector of the shaft - beneath the lower face of the c a p by means of special epoxy resin of quick curing due to the groundwater proximity. Given that, the equipment FPDS-3 was used as well as its corresponding software developed by TNO - Building & Construction Research; the records obtained were adequately “filtered” to take into consideration the presence of the caps, in the indicated cases. The preliminary analysis of the signals revealed a real alarming picture, as they previously showed a number of piles with anomalies which easily exceeded- the usual expected percentages for this type of piles. In the view of the possibility that the diagnosis were influenced by a mistaken interpretation of the geotechnical data provided by the Customer, the performance of new reliable Geotechnical Studies were arranged, as to being able to apply- in a second stage- the Signal Matching Technique (21). Likewise, and giving the worrying situation (because it could lay open to discussion the diagnosis by the use of technology that was rather unknown in our environment at that time), the following measures were adopted: a) Execution of a new Geotechnical Study, as it is described above. b) Take new signals in the total number of piles. c) Apply a Signal-Matching Technique on some of the observed piles. d) Ask for a second independent opinion to TNO- Building & Construction Research, in accordance with the agreement we kept with the mentioned Institute. As a consequence of the previous, we arrived to the following conclusions: i.- The original geotechnical study contained no mistakes that would have led to erroneous conclusions when modelling the soil to apply the TNOWAVE program option Signal Match. ii.- The repetition of the Sonic Integrity Tests showed that the preliminary diagnosis, derived from the first series of signals, kept its force; it is summarised as Table 2 shows. To illustrate the example, some typical signals corresponding to “piles with anomalies” (Figure 3) and “normal pile” (Figure 4) corresponding to Tower A are enclosed
Table 2. Summary of tests on piles. TOWER PILES No Total w/Anomalies A 108+6* 72+6* 26 B 108 Note: * Piles replaced by order of the Main Building Contractor before SIT testing.
5 7m
A
.
f 57rr
0. Icm/
5’6m
0 1 2 3
4
5
Pile 65 29 Jan 94
6
7
8
9
:
v2 a
G
f 5
4000P/s
I
I
sr
e x p 70
Figure 3. Pile #68; anomaly close to the 3m.
4 7m
I a
P i l e 12 26 Jan 94
1
2
3
4
5
6
7
a
9
4000m/s
__
1
v2 0
0 f.
a
sr
exp. 5
Figure 4. Pile #12; without anomalies.
iii.- From the number of Simulation Analysis done, on the behalf of the TNO specialists as well as the author and his colleagues- in which the signal matching techniques were applied through the TNOWAVE program option SIT-SM -,we arrived to the conclusion that most of the “suspicious” piles presented anomalies that were incompatible with their security (stability and life time). The respective graphics are included (Figures 5 and 6). iv.- Based on the conclusions that were obtained, we came to agreement with the main building contractor: to carry out the excavation and the drain of the staircase and elevator sector of Tower A in order to be able to observe the condition of the piles involved. In spite of the difficult work, they were performed efficiently, allowing the personal observation- on the behalf of all the people involved in the construction- of the anomalies which their prediction, based on the Sonic Integrity Tests and the 764
Figure 5 . Signal Match of pile #68 (anomalous).
Figure 6. Signal Match of pile # I 2 (normal).
adequate application of the Signal Matching Techniques, allowed the early detection of serious problems derived from the lack of geometricstructural integrity of a large number of piles in both buildings (see photographs 8 and 9). v.- In order to solve the structural problems, the replacement of the defective piles through the installation of additional ones using the following construction methodology was determined: a) Installation of a temporary open ended caisson by means of a vibratory hammer up to the indicated depth by the complementary studies. b) Inside cleaning. c) Installation of the reinforcements. d) Casting concrete with tremie pipe. e) Retrieving the caisson with vibratory hammer. f ) Control of the substitute piles through the Sonic Integrity Tests (SIT). vi.- The substitute piles were exposed to the Quality Control Tests by means of the indicated technique, and the obtained results are showed in the table:
Photograph 9. Characteristics of the damages.
Table 3. Summary of the additional piles.
TOWER
N" A B
k Case History I1
PILES Total 49 23
w/Anomalies 0 0
Corresponds to a group of three luxurious buildings of thirty stories high built in the north of Buenos 765
Aires city (Argentina); the Architecture Project foresaw the building of a basement that held the totality of the place (a block). Taking into account the bad quality of the upper soil layers (see Figure 7 ) it was decided that the foundation should be with displacement cast-in-situ piles with extended bulb (Franki type). For building reasons, the piles were installed from ground level and casted up to the basement level, which was located beneath the level of general excavation and close to the groundwater level. Because of budget reasons, excavations for the caps and foundation beams construction were done by reduced sectors- without a general groundwater abatement.
4.61~
0.I c d 4.681
0.lcn/s 0
1
2
3
4
5
6
7
8
P i l e 581
4000a/s
21 dug 94
exp: 20
9
1
v2.0
0 f: I
sr
Figure 8. Pile #55:anomaly at 1.5m
for the already mentioned reasons, was taken as model pile (Figure 9).
3’7m
O.lCrn/
0
1
2
3
4
5
6
7
8
P I ] @ 38
4000m/s
17 ADP 94
exa: 2
9
1
I
v2.0
0
I:
sr
Figure 9. Pile #38: without anomalies. Figure 7. Existent Geotechnical Profile in the building site
That brought the impossibility of an appropriate coordination that allowed carrying out and efficient quality control of the totality of the piles because of which, the Engineer and the Construction Company determined that the work was done in a partial and random way as the progress of the work plan of the building will allow, thing that gave place to a statistical pseudo-control. In accordance to the previously mentioned, a small amount of Sonic Integrity Tests were done through the usage of the already mentioned FPDS-3 equipment. Nevertheless the limited quantity of available signals, made the adoption of the statistical standard signal corresponding to the family of records representative of the construction (or the sector) problems, anomalies in several of the registered signals were detected. The main one is showed next (Figure 8). The Simulation program TNOWAVE option SIT-SM (Signal Matching Technique) was applied on that signal and the pile NO38 that is located near,
766
In accordance with the usual confrontation of opinions (the Engineer-Main Building ContractorFoundations Subcontractor-Geotechnical Consultant-Structural Advisor-etc.) when a customer is informed that in his site a pile with anomalies has been detected. The conclusion obtained was similar to the one obtained previously, as Figures 10 and 11 show.
Figure 10. Signal Match of pile #55 (anomalous).
Not-withstanding the previous, and in the view of the disbelief showed by the Engineer, we proposed the excavation of the pile and it could be seen a series of anomalies (deep craks, loss of coating, etc.) which are indicated in Photograph 10.
quality of the soil in the building site (see Figure 12) the planners decided that the foundation should be piled. Franki type piles were adopted.
Figure 1 1. Signal Match of pile #38 (normal).
Figure 12. Geotechnical profile
Figure 13. Pile #1: anomaly at 5.501~1
+--@-I
+ Photograph 10. Pile #55. View of the observed damages.
er
P Case History 111 This construction site, which took place in the north zone of the Gran Buenos Aires (Argentina), constituted by a number of 10 storey towers destined to economic housing. Taking into account the bad
Figure 14.
767
ano'nalyat 8.10m.
-
The total number of the piles of the construction (420) was tested as the Foundation Company was completing the works corresponding to each of the towers. They had previously done the trimming work and cleaning after seven days of casted, in order to obtain an adequate resistance and to be apt to transmit the stress wave (SW) generated by the impact of the hammer. Only two piles, No 1 of the Tower C2 and the No 154 corresponding to Tower A, thrown anomalous signals so a more detailed analysis through the application of the Signal Matching Technique had to be done. Due to the proximity of the observed piles, Pile 44 from the Tower C2 was adopted as "model pile"; the respective signals are showed next: (see Figures 13 to 15). As a consequence of the application of the TNOWAVE program SIT-SM option, the graphics that shows figures 16 to 20 were obtained. The detailed analysis allowed us arriving to the following conclusions: a) Pile 1 (Figs. 13 and 16). This is a complicated signal because the original pile, as a consequence of having problems with its installation, had to be replaced by redriving the caisson in the same place with the concrete of the first one still fresh; as a consequence, the pile shape turned to be highly irregular showing increases and reductions in its diameter. The signal match determined a variable increase in the pile diameter- from 350mm to 450mm in relation to the depths between 2.5 and 5.2mfollowed by an abrupt reduction to the nominal diameter starting at 5.6m. The corresponding reflection of the pile top is clearly marked at an approximately depth of 13.0, while the signal that is detected at a depth of 11.5m corresponds to the repetition of the previous signal of increase and the subsequent section reduction. b) Pile No 154 (Figs. 14 and 17). The SignalMatch (SM) established the presence of a "larger anomaly" to an approximately 8.1m corresponding to a discontinuity, reason why the stress wave that was generated could not continue its way further than that depth, giving place to the "reflection corresponding to the end"; the rest of the diagram corresponds to the first repetition of the signal. From the Sonic point of view and consequently from its capability to transmit shear stresses, is as if the pile were 8.lm long even how, probably, it was driven to its nominal depth, but that discontinuity could have been a crack or an important soil inclusion as it happened in the Case History I. The problem was solved when the Engineer arranged the replacement of the pile by other two of smaller dimensions and the modification of the pile cap and foundation beams.
I
Figure 15. Pile #44: without anomalies.
-eoo1
.
,
0 0
-----
20
CDlLUldlCB
1 0
an0
- - -
,
,
6 0
6 0
I
100
,
,
,
120
260
160
I
I
I80
200
sec I 10-3 .~
m?aI"PE(I "ClOCillCS
Figure 16. Signal Match of pile # I .
-1 2
t
-1
-2.01 00
,
,
20
.___. C*,C"llrfCd
4 0
an0
- .-
,
,
,
69
8 0
100
12.0
,
,
140
160
P P 0 ~ " P C ~"CIOCIIIES
J ZWO
IS0 5rC
I
10-3
Figure 17. Signal Match of pile #154.
-2 0 0 0
2 0
4 0
8.0
0 0
10.0
I20
140
160
100
a 0
Figure 18. Signal Match of pile #44
768
6 CONCLUSIONS.
neering. Design and practice guide”. (Thomas Telford, London, 1996). (9).- Institution of Civil Engineers: “The Value of Geotechnics in Construction”. (Construction Research Communications, London, 1998). (10) Love, A.E.H.,: “A Treatise on the Mathematical Theory of Elasticity”. (4th. Edition, New York, Dover Publications, 1944). (1 1) Timoshenko, S. & Goodier, J.N.: “Theory of Elasticity”. (2nd. Edition, McGraw-Hill Book Company, Inc. 195 1). (12) Timoshenko, S. & Young, D.H.: “Vibration Problems in Engineering”. (3rd. Edition, D. Van Nostrand Conpany, Inc. 1955). (13) Voitus van Hamme, G.E.J.S.L.; Jansz, J.W.; Bomer, H. & Arentsen, D.: “Hydroblock and improved piledriving anilisis”. (De Ingenieur, no 18, ~01.86,1974). 14) Middendorp, P. Private communication, June 2000. (1 5 ) Beim, G. K., Private communication, June 2000. (1 6) Thorburn, S. & Thorburn, J.Q.: “Review of the problems associated with the construction of cast-in-place concrete piles”. (DOE and CIRIA Piling Development Group, Report PC2, London, 1977). ( I 7) Healy, P.R. & Weltman, A.J.: “Survey of problems associated with the installation of displacement piles”. (DOE and CIRIA Piling Development Group, Report PG8, London, 1980). (18) Fleming, W.G.K.; Weltman, A.J.; Randolph, M.F. & EIson, W.K.: “Piling Engineering”. (Surrey University PresdBlackie and Son Ltd., London, 1985). (1 9) Turner, M.J.: “Integrity testing in piling practice” (CIRIA Report 144, London, 1997). (20) Coldcmberg, J.J. & Goldemberg, H.; “Patologia en fundaciones profundas. Origen y clasificacion”. (Memorias del V Congreso Iberoamericano de Patologia de las Constnicciones, VlI Congreso de Control de Calidad CONPAT 99, Montevideo, Uruguay, 1999). 6) (21) Goldemberg, J.J. & Goldemberg, H.; “Confiabilidad de 10s ensayos de integridad sobre pilotes por mCtodo sonico - SIT”. (Memorias del I Congreso Paraguay0 de Ingenieria Geotecnica, IV Jornadas Gcotkcnicas Estnicturales, 1 ra. Reunion de Ingenieria Geotecnica del MERCOSUR - 1er. COPAINGE, Asuncion, Paraguay, 1997).
The results obtained in the Quality Control done since the beginning of the application of this technology, in Argentina as in the rest of the world, allow establishing- without doubt- that the Sonic Integrity Tests are an essential tool to know the condition of a pile, new ones and old ones (case of pre-existent old foundations recycling or in cases of Forensic Engineering) As in every indirect investigation technique, where the diagnosis is obtained through analyzing a measurements with sensitive instruments or image analysis of signals corresponding to physicmechanic phenomenon, the unavoidable requirements are: > The availability of appropriate equipment, with proved and reliable hardware and software. P Skilled operators P Periodic calibration of sensors. P An exhaustive and reliable Soil Investigation. P An adequate control of the execution of the foundations. P Full knowledge of the Foundation Engineering, Geotechnical Engineering and Foundation Dynamics by the Engineer in charge of the Integrity Tests. P Independent criteria and ethical integrity on its behalf.
7 BIBLIOGRAPHIC REFERENCES. ( I ) Bredenberg, H. (Editor); “Application of Strcss-Wave Theory on Piles” (Proceedings of the International Seminar on the Application of Stress-Wave Theory on Piles, Stockholm, Sweden, 1980). (2) Holm, G.; Bredenberg, H. & Gravare, C.J. (Editors); “Application of Stress-Wave Theory on Piles” (Proceedings of the Second International Conference on the Application of Stress-Wave Theory on Piles, Stockholm, Sweden, 1984). (3) Fellenius, B.H. (Editor); “Application of Stress-Wave Theory to Piles” (Proceedings of the Third International Conference on the Application of Stress-Wave Theory to Piles, Ottawa, Canada, 1988). (4) Barends, F.B.J. (Editor); “Application of Stress-Wave Theory to Piles” (Proceedings of the Fourth International Conference on the Application of Stress-Wave Theory to Piles, The Hague, The Netherlands, 1992). 5 ) Townsend, F.C.; Hussein, M.H. & McVay, M.C. (Editors); “Stresswave ‘96” (Proceedings of the Fifth International Conference on the Application of Stress-Wave Theory to Piles, Orlando, Florida, USA, 1996). (6) PEDIR INFORMACION A SUSUMU. Stresswave ‘2000” (Proceedings of the Sixth International Conference on the Application of Stress-Wave Theory to Piles, San Pablo, Brasil, 2000). (7) - van Koten, H. & Middendorp, P.: “Testing of Foundation Piles” (Delft University of Technology, Delft, The Netherlands, HERON, vol. 26, No 4, 198 1). (X).- Institution of Civil Engineers: “Creating value in Engi-
769
This Page Intentionally Left Blank
Application of Stress-Wave Theory to Piles, Niyama 8 Beim (eds)02000 Balkema, Rotterdam, lSBN 90 5809 150 3
Author index
Abreu, L. 201 Amir, J.M. 313 Andreo, C.S. 261,389 Antoniutti Neto, L. 707 Aoki, N.241,375,457,635 Axelsson, G. 665 Ayasrah, I. M. 233 Aye,Z.Z. 163, 171 Balech, J. 241 Barbosa, 6 . E. 429 Barends, E B. J. 541 Baycan, S. 75 1 Beim, J.W. 127 Benamar, A. 1 17,255 Bernardes, G.P. 261,389,707 Bettess, F? 495 Boonyatee, T. 563,569 Bnino, D. 47 Cai, L.B. 11 Camapum de Carvalho, J. 157 Cannon, J.G. 393,399 Chambers, W.G. 407 Chamecki, E! R. 707 Chen, B. 65 1 Chen, R.P. 29,517 Chen, Y. M. 29,5 17 Chernauskas, L.R. 223 Cho, C.W. 41,47 Cintra, J.C.A. 375,457 Costa, C. M.C. 429,441 Courage, W. M.G. 575 Cunha, R. I? 157 Danziger, B. R. 657 de Albuquerque, F? J. R. 677 de Campos, G.C. 429,435 de Carvalho, D. 677 DiMillio, A. E 223
Dyminsky, A.S. 127
Justason, M. D. 609
Ealy, C.D. 223 Esposito, G. 575
Kalinowski, M. 267 Karkee, M. B. 689 Kawabata, N. 335,583 Kikuchi, Y. 599 Kimura, M. 563,569 Kirsch, E 249 Kita, N. 41 1 Klingberg, D.J. 403,407,715 Korkeakoski, F? 415 Kormann, A.C. M. 707 Kusakabe, 0.335,591
Fellenius, B. H. 3 13 Ferreira, J.S. 657 FOB,S.B. 157 Fujita, K.335,451,469 Goble, G.G. 3,305,327 Goldemberg, H. 345,7 19 Goldemberg, J.J. 345,719,757 Gonqalves, C. 261,389 Gutikrrez, A. 201 Hajduk, E.L. 541 Hannen, W.R. 503 Hart, L.J. 223 Hartikainen, J. 415 Hasard, D. 201 Hayashi, M. 583,697,741 Hilmi Acar, M. 51 1 Hintze, S. 665 Hoffmann, Ch. 201 Holeyman, A.E. 479,725,733 Holscher, P. 541 Horiguchi, T. 689 Huch, T. 249 Husein Malkawi, A. I. 233 Hussein, M. 91 Huybrechts, N. 725,733 Imada, K. 179 Iskandarani, W. M. 91 Janes, M.C. 609 Joer, H.A. 47 Jokiniemi, H. 415 Jonker, G. 135 77 1
Liang, R.Y. 121 Lee, J.-S. 421 Lee, M.W. 41 Lee, S.-B. 99 Lee, W.-J. 99 Lee, Y. -N. 42 1 Legrand, C. 725,733 Liang, L. 53 Liang, R.Y. 461 Likins, G.E. 205,211,327 Lima, E M.A. 375 Liu Xi-An 683 Lucieer. W. J. 65 Mackenzie, P. 403,7 15 Maertens, J. 725,733 Matsuda, Y. 187 Matsumoto, T. 179, 187,335,583, 59 1,697,741 Maung, A.W. 163,171 Michi, Y. 187 Middendorp, E! 55 1,609,617,625 Morgano, C. M. 205 Mukaddam, M.A. 9 I Mullins, A.G. 609
Nakata, Y. 179 Navajas, S. 429,435 Navaneethan, T. 171 Nawari, N.O. 121 Nishimura, S. 335,411,591, 599,741 Nishiumi, K. 591,741 Niyama, S. 429,435 Oh, J.-H. 99 Ohno, M. 469 Okahara, M. 335
Rausche, E 53,59,75,205, 21 1,327 Restrepo, C. 219 Robinson, B. 53,59 Rodatz, W. 249 Romanel, C. 127 Romell, J. 145 Roth, B.C. 503 Russo Neto, L. 707
Padcowsky, S.G. 223,28 1,541 Paraiso, S.C. 429,441 Parente-Ribeiro, E. 127 Pereira, J. H. E 157 Pinto Soares, E. 441 Pinto, EL. 275 Piscsalko, G. 205 PlaBmann, B. 249
Salumoto, J. 41 1 Seidel, J.P. 59, 193,267, 319 Selby, A. R. 495 Shibata, A. 583,741 Skov, R. 107 Stenersen, K.L. 281 Stevens, R.E 17,351 Sugimura, Y. 689 Suzuki, M. 697 Svmkin, M.R. 35,107,113, 503,525
Ramshaw, C. L. 495 Randolph, M.E 41,47
Takeda, T. 41 1 Tanseng, I? 163
772
Tatsuta, M. 599 Thasnanipan, N.163, 171 van Foeken, R. J. 135,575,625 van Ginneken, G.J. J. 6 17,625 Viking, K. 533 Wakiya, Y. 583,741 Wu, J. 673 Wu, S. 383 Xi Liang 369 Xiao, L. 383 Xu, D. 383 Yang Wu 153 Yoshizawa, Y. 583,591 Zhang, F. 563 Zhang, Y.-N. 683 Zheng, J. M. 65 1 Zheng, Y. M. 65 1 Zhou, G. 673 Zhu, B. 5 17