Nuclear Data Needs for Generation IV Nuclear Energy Systems

Nuclear Data Needs For Generation IV Nuclear Energy Systems Proceedings of the International Workshop

Editor Peter Rullhusen

Nuclear Data Needs For Generation IV Nuclear Energy Systems Proceedings of the International Workshop

This page is intentionally left blank

Nuclear Data Needs For Generation IV Nuclear Energy Systems Proceedings of the International Workshop

Antwerpen, Belgium

5-7 April 2005

Editor

Peter Rullhusen European Commission, Joint Research Centre, Institute for Reference Materials and Measurements, Belgium

\ p World Scientific NEW JERSEY • LONDON • SINGAPORE • BEIJING • SHANGHAI • HONG KONG • TAIPEI • CHENNAI

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

NUCLEAR DATA NEEDS FOR GENERATION IV NUCLEAR ENERGY SYSTEMS Proceedings of the International Workshop Copyright © 2006 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 981-256-830-1

Printed in Singapore by World Scientific Printers (S) Pte Ltd

INTERNATIONAL ADVISORY COMMITTEE J. Chang M. E. Dunn P. Finck A. Hasegawa R. Jacqmin J. Katakura A. Koning K. Kozier

KAERI ORNL ANL JAERI CEA JAERI NRG AECL

L, Leal A. Nichols D. W. Nigg C. Nordborg P. Oblozinsky E. Pitcher A. Plompen M. Salvatores

LOCAL ORGANISER COMMITTEE

G. Giorginis F.-J. Hambsch G. Lovestam W. Mondelaers S. Oberstedt A. Plompen P. Schillebeeckx P. Siegler

ORNL IAEA-INDC INEEL OECD-NEA BNL LANL IRMM CEA

This page is intentionally left blank

PREFACE

On 6 March 2003 a Technical Exchange and Cooperation Arrangement in the field of Nuclear-Related Technology Research and Development was signed between the Department of Energy of the United States of America and the European Atomic Energy Community. Several joint programs and co-operative research projects have been agreed in the mean time between DOE and Euratom. At the first Steering Committee meeting between DOE and IRC on 23 February 2004 in Brussels, it was decided to organise a workshop on Nuclear Data Needs for Generation IV Nuclear Energy Systems. This workshop was organised jointly by JRC on behalf of Euratom, DOE and AECL on 5-7 April 2005 in Antwerp, Belgium. The workshop was attended by 70 participants from 20 countries. During three consecutive days recent achievements were presented on sensitivity analysis, model calculations, estimates of uncertainties, and the present status of nuclear data bases. Although detailed work on the different Generation IV systems did not yet start at the time of the workshop, it was impressive to see the amount of work and the results already achieved in connection with the development of these new systems. The local organisers are very grateful to all participants contributing with interesting presentations to the workshop and to the session chairmen for monitoring the lively discussions. I would like to thank the international advisory committee and especially Hussein Khalil (DOE) and Robert Speranzini (AECL) for their continuous support.

Peter Rullhusen Conference Chair

vu

This page is intentionally left blank

CONTENTS Organising and Scientific Advisory Committees Preface

v vii

Nuclear Data Needs for Generation IV Systems Future of Nuclear Energy and the Role of Nuclear Data P. Finck

3

Nuclear Data Needs for Generation IV Nuclear Energy SystemsSummary of U.S. Workshop T. A. Taiwo, H. S. Khalil

8

Nuclear Data Needs for the Assessment of Gen. IV Systems G. Rimpault

18

Nuclear Data Needs for Generation IV- Lessons from Benchmarks S. C. van der Marck, A. Hogenbirk, M. C. Duijvestijn

32

Core Design Issues of the Supercritical Water Fast Reactor M. Mori, A. Rineiski, W. Mashek, V. Sinitsa

38

GFR Core Neutronics Studies at CEA J. C. Bosq, V. Brun-Magaud, G. Rimpault, J. Tommasi, A. Conti, J. C. Gamier

46

Comparative Study on Different Phonon Frequency Spectra of Graphite in GCR Young-Sik Cho, Kang-Seog Kim, Do Heon Kim, Young-Ouk Lee, Jonghwa Chang

58

Innovative Fuel Types for Minor Actinides Transmutation D. Haas, A. Fernandez, J. Somers

64

IX

X

The Importance of Nuclear Data in Modeling and Designing Generation IV Fast Reactors K. D. Weaver

70

The GIF and Mexico- "Everything is Possible" C. Arredondo Sanchez

72

Benchmarks, Sensitivity Calculations, Uncertainties Sensitivity of Advanced Reactor and Fuel Cycle Performance Parameters to Nuclear Data Uncertainties G. Aliberti, G. Palmiotti, M. Salvatores, T. K. Kim, T. A. Taiwo, I. Kodeli, E. Sartori, J. C. Bosq, J. Tommasi

81

Sensitivity and Uncertainty Study for Thermal Molten Salt Reactors A. Bidaud, T. Ivanona, V. Mastrangelo, I. Kodeli

101

Integral Reactor Physics Benchmarks- The International Criticality Safety Benchmark Evaluation Project (ICSBEP) and the International Reactor Physics Experiment Evaluation Project (IRPHEP) J. B. Briggs, D. W. Nigg, E. Sartori

113

Computer Model of an Error Propagation Through Micro-Campaign of Fast Neutron Gas Cooled Nuclear Reactor E. lvanov

128

Combining Differential and Integral Experiments on 239Pu for Reducing Uncertainties in Nuclear Data Applications T. Kawano, K. M. Hanson, S. C. Frankle, P. Talou, M. B. Chadwick, R. C. Little

138

Sensitivity of Activation Cross Sections of the Hafnium, Tanatalum and Tungsten Stable Isotopes to Nuclear Reaction Mechanisms V. Avrigeanu, M. Avrigeanu, F. L. Roman, R. A. Forrest, R. Eichin, H. Freiesleben, K Seidel

145

Generating Covariance Data with Nuclear Models A. J. Koning

153

xi

Sensitivity of Candu-SCWR Reactors Physics Calculations to Nuclear Data Files K. S. Kozier, G. R. Dyck

163

The Lead Cooled Fast Reactor Benchmark BREST-300: Analysis with Sensitivity Method V. Smirnov, V. Orlov, A. Mourogov, D. Lecarpentier, T. Ivanovo

173

Sensitivity Analysis of Neutron Cross-Sections Considered for Design and Safety Studies of LFR and SER Generation IV Systems K. Tucek, J. Carlsson, H. Wider

183

Experiments INL Capabilities for Nuclear Data Measurements Using the Argonne Intense Pulsed Neutron Source Facility J. D. Cole, M. W. Drigert, R. Aryaeinejad, D. W. Nigg, R. V. F. Janssens, B. J. Micklich, G. Ter-Akopian

195

Cross-Section Measurements in the Fast Neutron Energy Range A. Plompen

208

Recent Measurements of Neutron Capture Cross Sections for Minor Actinides by a JNC and Kyoto University Group H. Harada, H. Sakane, S. Nakamura, K. Furutaka, J.-I. Hori, T. Jujii, H. Yamana

216

Determination of Minor Actinides Fission Cross Sections by Means of Transfer Reactions M. Aiche, G. Barreau, S. Boyer, S. Czajkowski, D. Dassie, C. Grosjean, A. Guiral, B. Haas, B. Jurado, B. Osmanov, E. Bauge, M. Petit, E. Berthoumieux, F. Gunsing, L. Perrot, C. Theisen, F. Michel-Sendis, A. Billebaud, J. N. Wilson, I. Ahmad, J. P. Greene, R. V. F. Janssens

222

Evaluated Data Libraries Nuclear Data Services from the NEA H. Henriksson, Y. Rugama

235

Nuclear Databases for Energy Applications: An IAEA Perspective R. Capote Noy, A. L. Nichols, A. Trkov

244

xii

Nuclear Data Evaluation for Generation IV G. Noguere, O. Bouland, A. Courcelle, E. Dupont, O. Serot, J. C. Sublet

253

Improved Evaluations of Neutron-Induced Reactions on Americium Isotopes P. Talou, T. Kawano, P. G. Young, M. B. Chadwick, E. J. Pitcher

262

Using Improved ENDF-Based Nuclear Data for Candu Reactor Calculations /. Prodea

270

A Comparative Study on the Graphite-Moderated Reactors Using Different Evaluated Nuclear Data Do Heon Kim, C.-S. Gil, Y.-S. Cho, Y.-O. Lee, J. Chang

278

Author Index

285

NUCLEAR DATA NEEDS FOR GENERATION IV SYSTEMS

This page is intentionally left blank

FUTURE OF NUCLEAR ENERGY AND THE ROLE OF NUCLEAR DATA P. FINCK Argonne National Laboratory

1. Introduction Nuclear energy plays a major role to secure energy supply in future. In fact, nuclear energy provides a unique and cost-effective answer to the problem of C0 2 emission and to the requirement of a sustainable development (see Fig. 1) in particular in view of a potential growth in energy demand. In Fig. 2 the per capita GDP is shown as the per capita energy consumption, with evidence for a potential spectacular growth in energy demand due to population and GDP increase in developing countries (e. g. China, India). At present, there are indications for a nuclear energy revival, which is based on the consolidation and evolution of present reactor concepts, but also on the definition of new requirements for both future reactors and their associated fuel cycles. The key issues for nuclear energy development are: • Economics • Safety • Proliferation resistance • Waste minimization • Supplies of uranium These last two points are crucial to insure sustainability. However, the current once through cycle is not sustainable. Advanced technologies (closed fuel cycles) are required. As one example, in Fig. 3 it is shown the impact on spent fuel stockpiles, of the introduction of advanced technologies, leading to the deployment of closed fuel cycles based on fast reactors (LMR's), according to different hypothesis on energy demand growth. In this respect, the first phase of the Generation IV initiative has produced a wide international consensus around six preferred concepts and has underlined the need for improved solutions for the back-end of the fuel cycle, and in particular the essential benefits of the closed fuel cycle option, associated to the development of fast reactors. 3

• No COj emissions and no contribution to Global Warming

An already competitive energy source Safety improvements in 3rd Gen reactors are already significant

Promising assists to produce Hydrojer

Fig. 1 Benefits from using nuclear energy.

Sweden o

Norway

o

.Malaysia -Turkey Thailand • Trie Philippines

Per-capita energy consumption (tor»s|of oWpmstm) Fig. 2 Energy use vs. GDP

Innovative reactor concept and fuel cycles are a potential source for an increased role of reliable and well validated nuclear data, both at the stage of feasibility studies and for more detailed design assessments. However, no new requirement in the nuclear data field will be credible, without a sound assessment of uncertainties and their impact. This type of

5

assessment, which is at the heart of the present workshop, has already started in some laboratories, and first results will be presented. 120,000 -i

2000

.

2010

•

2020

•

. y

2030

1

2040

2050

Fig. 3 Estimated Stockpile of Spent Fuel (MT/HM)

2. The role of nuclear data in developing future generation nuclear energy Some examples have been studied using a simplified approach. The systems chosen for the analysis are: • Extended burnups for LWR's • Very High Temperature Reactors • Fast Reactors: GFR, SFR, LFR • Accelerator Driven Systems The fuel cycle performances were also analyzed. As far as uncertainties, it has to be stressed that since credible and complete covariance data were not available, estimates of these data have been used, based on physics judgement and on the performances of nuclear data in the analysis of selected, high precision, integral experiments. Uncertainties were propagated through standard static and depletion codes using sensitivities coefficients calculated with Generalized Perturbation Theory (GPT).The results should not be used in an absolute sense, but have a relative value, as a first indication of major trends. A summary of the most significant data are given in the following tables (I through III).

6 Table I. Extended Burnup for LWR's Uncertainties on Kerr at Beginning of Cycle (BOC) Fuel Burnups in LWR's Total uncertainty is -510 pcm have been slowly (U235: 350 pcm; U238: 360 pcm) increasing in order to Uncertainties on Kefr at End of Cycle (EOC) reduce costs. Total uncertainty is -1220 pcm Current plans indicate (Pu239: 620 pcm; Pu241: 320 pcm; that burnup might be Pu 240: 620 pcm; U238: 690 pcm) increased in the [50-100] GWd/ton range. Uncertainty on Burnup Swing -2240 pcm, dominated by Pu240 As burnup increases, the Uncertainties on isotopics neutronic contributions of Largest (-6%) on Np237 Pu238 Pu240 transuranics become Am and Cm isotopes predominant Due to capture of U236, Pu240, Am241, 243 and Cm242, 244 Table n. The Very High Temperature Reactor Graphite moderated Uncertainties on Kefr at BOC U235 enrichment > 10% Total uncertainty is ~580 pcm Very High Burnup (U235: 360 pcm; U238: 430 pcm) Very High Thermal Uncertainties on Ketr at EOC Efficiency Total uncertainty is -1070 pcm High Outlet Temperature (Pu240: 630 pcm; Pu239: 570 pcm; U238: 550 pcm) Uncertainty on Burnup Swing -1749 pcm, dominated by Pu240 Note that other data uncertainties can play a significant role: graphite S(a,P); lower resonances of minor actinides if VHTR's are conceived in a burner mode. Table m. Fast Breeder or Burner modes. Fuels, structures, and reflectors might contain new materials (Zr, Si), as in the case of GFR's

Reactors Uncertainties on Ketr: ~2000pcm U238 (inelastic, capture), Pu239 and Pu241 (fission) still predominate but Si (inelastic) contributes 430 pcm in GFR Uncertainties on void worth: -12-20% He void worth: small absolute value < uncertainties. Sign is unclear Uncertainty on Doppler worth: -5-10% Uncertainty on Burnup Swing: less than 1000 pcm

3. Nuclear Data Needs As compared to current concepts, these examples show a slight increase in overall uncertainties. However, they do not indicate an urgent need for large amounts of new data, only isolated data might be needed urgently. But we must recall that predictive codes for existing reactors achieve very low uncertainty, not because of the quality of the nuclear data, but thanks to a series of "adjustments" to their data libraries. These adjustments were made possible by the accumulation of integral data from measurements in reactors and mock-up facilities. These were lengthy and costly, and might not be available in the future. The need to reduce costs and research and development schedules will

7

lead to use powerful simulation techniques to (partially) replace experiments. The required ingredients are: • Better codes, based as far as possible on first principle description of phenomena • Better data. This in turn means: - To reduce uncertainties on key principal reactions - To provide reliable, complete, and systematically derived covariance files. These data are absolutely needed and should be developed in a consistent way. - The need for a few "clean" integral experiments (simple and well documented). 4. Conclusions It is recognized that nuclear energy has a key role in future energy supply. As far as nuclear data, specific nuclear data needs for existing and new concepts can be justified. Covariance files are needed to move the reactor design community towards a more efficient R, D & D model. A systematic approach for defining needs and for fulfilling the related requirements has to be used. In particular, new experiments should address effectively the problem of reducing uncertainties. In fact, required target accuracies can be sometimes smaller than achievable accuracies in current differential cross section measurements and new experiments as well as new approaches should be envisaged and implemented.

NUCLEAR DATA NEEDS FOR GENERATION IV NUCLEAR ENERGY SYSTEMS - SUMMARY OF U.S. WORKSHOP T. A. TAIWO AND H. S. KHALIL Nuclear Engineering Division, Argonne National 9700 South Cass Avenue, Argonne, IL 60439,

Laboratory U.S.A.

A workshop on the data needs for Generation IV Nuclear Energy Systems was held in the U.S. in April 2003. A summary of this workshop is provided in this paper. Discussions during the workshop evolved along the traditional nuclear data topical areas of data needs, measurements, evaluations, processing and validation. Recommendations were made on how the Generation IV needs could be better defined and on approaches for resolving the needs.

1. Introduction The six advanced nuclear energy systems that have been selected for development under the Generation IV program target significant advances over current generation and evolutionary systems in the areas of sustainability, economics, safety and reliability, and proliferation resistance and physical protection. These nuclear systems are the Very High Temperature Reactor (VHTR), the Gas-Cooled Fast Reactor, the Lead-Cooled Fast Reactor (LFR), the Sodium-Cooled Fast Reactor (SFR), the Supercritical-Water-Cooled Reactor (SCWR), and the Molten-Salt-Cooled Reactor (MSR). As a group, the six systems employ a large spectrum of new fuels (particles, dispersion, solid solution), a variety of core geometries (pins, pebbles, hexagonal blocks, annular pins, plates), and various coolants (supercritical H20, sodium, lead alloys, C02, and molten salts). These systems introduce modeling challenges that differ from those of current generation Light-Water Reactor (LWR) cores. Therefore, while mature tools and data exist for the analysis of LWRs, the ability of these tools to model accurately the advanced systems has to be assessed systematically. Initial U.S. efforts directed to improving capabilities for modeling the physics of the Generation IV systems have focused on evaluating the applicability of existing tools and data, identification of the additional needs, and identification of integral benchmarks that can be used for verification and validation of existing and emerging tools. Several workshops were organized to

8

9

identify analysis needs of Generation IV systems and relevant activities directed to meeting these needs. [1,2,3] Table 1. Topical Areas for the U.S. Workshop* Topical Area Title of Presentation Nuclear Data Needs for Gen IV Program Overview Needs for Gas-Cooled Reactors (VHTR, GFR) Generation IV Nuclear Data Needs for Liquid Metal Reactors (SFR and LFR) Energy Systems Perspective on Nuclear Data Needs for Advanced Nuclear Systems Nuclear Data Processing, Application of Covariance Data to Reactor Design Applications, and ORNL Cross-Section Processing Capabilities and Validation Experience for Advanced Reactor Applications Performance of New Los Alamos Actinide Evaluations Covariances in JENDL-3.3 and covariance evaluation with the KALMAN system Theory and Evaluation Status of Work on the ENDF/B-VII Library EMPIRE: Advanced Tool for Nuclear Reaction Data Evaluation Systematic Re-Evaluation of Neutron Resonance Parameters Development and Validation of Temperature Dependent Neutron Scattering Laws ORNL Tools for Cross Section Measurements and Nuclear Data Evaluation in Support of Gen IV Measurements BNEEL Perspective on Nuclear Data Issues for Gen-IV INEEL/ANL Collaborative Program for Advanced Nuclear Data Measurements at ANL/IPNS LANSCE Facility and Capabilities for Nuclear Data Measurements Differential and Quasi-Integral Measurements at the Gaerttner LINAC

This paper summarizes the results of a workshop on nuclear data needs held under the U.S. Generation IV program in 2003. The workshop, held at the U.S. National Nuclear Data Center (NNDC), Brookhaven National Laboratory, was attended by experts from the U.S. national laboratories and universities. [4] The remainder of this paper is divided into four sections, consistently with the organization of the workshop (see Table 1). These sections provide a summary of the conclusions and recommendations that emerged for the following areas: Nuclear Data Needs for Generation IV Nuclear Energy Systems (Section 2); Nuclear Data Processing, Applications, and Validation (Section 3); A working paper on Nuclear Data for Gen IV was submitted by A. J. Koning et al. of the Nuclear Research and Consultancy Group, Petten, The Netherlands. Additionally, a presentation on "Nuclear Data for AFC in LANL/T-16: Brief Overview for Gen-IV Workshop," was submitted by M. Chadwick.

10

Theory and Evaluation (Section 4); and Nuclear Data Measurements (Section 5). Perspectives on the path forward for the nuclear data activities in support of the Gen IV systems are summarized in Section 6. 2. Nuclear Data Needs 2.1. Workshop Conclusions Predictions of the performance of advanced nuclear systems are limited in their precision by the uncertainties in nuclear data for a variety of important nuclides relevant to reactor design. These uncertainties have bearing on the design margins required to ensure system safety and on the performance and economic competitiveness of the systems. Significant uncertainties exist in the data for minor actinides, as well as for some of the more common fissionable materials (in certain energy ranges) and non-fuel materials such as bismuth and lead in particular. Some of these uncertainties can be addressed with integral benchmarks, but additional direct cross section evaluations (and possibly measurements) are also needed. There is also a more general need for nuclear data covariances and for methods to process and utilize them in sensitivity and uncertainty analyses. Reduction of calculational uncertainty is also of interest in the U.S. Advanced Fuel Cycle (AFC) Program, and efforts in that program are being used to leverage the Generation IV program effort. Existing nuclear data are likely sufficient for preconceptual design of the VHTR because it is a thermal spectrum system and the enriched uranium fuel form is being proposed in current designs. Additional data that may be required (e.g., for high burnup and high enrichment designs) need to be identified through a systematic approach. Such an approach would consist of developing a reference configuration, defining target accuracies of relevant core and fuelcycle parameters, evaluating the impact of nuclear data uncertainties on these parameters, and defining necessary re-evaluations or measurements for data contributing a large fraction of the total uncertainty. Fast spectrum systems (GFR, LFR, SFR), which contain significant fractions of transuranics (particularly minor actinides) in their recycled fuel, require some additional data re-evaluations and possibly differential cross-section or integral measurements. For the GFR, such data evaluations might be necessary for the new fuel matrix and reflector materials that have been proposed. These efforts should be directed to obtaining nuclear data that yield better accuracy in the prediction of core and fuel-cycle related parameters (e.g., transmutation rate, criticality state, power distributions, decay heat, radiation doses, and neutroninduced damage, etc). Co variance data require particular attention since this

11

information is largely unavailable in current evaluated nuclear data files. Additional covariance data are required in ENDF/B evaluations. In particular, only 46 out of 300-plus incident neutron evaluations in ENDF/B-VI (release 7) have uncertainty data. Cross section uncertainty data are available for only a few actinides (Am-241, Pu-240, Pu-241, Pu-242, Th-232, U-235, and U-238). The absence of uncertainty data for hydrogen should be remedied. Additionally, resonance-parameter uncertainty data are limited to the resolved energy region. The covariance data for secondary distributions are not addressed and require additional attention. Even where available, covariance data are not in a form readily applied in routine design calculations. Additional effort is therefore required to make this type of data available to core designers in the form useful for routine use. 2.2. Workshop Recommendations Specific nuclear data needs for the Generation IV systems should be further specified using the following systematic process: Preliminary reactor-core reference configurations should be developed and major fuel-cycle operation parameters should be preliminarily defined for each system. In addition to enabling prediction of typical system performance data, this effort should also produce an inventory of the materials in the entire fuel cycle (including those ultimately stored in a geologic repository), to facilitate the assessment of the nuclear data needs for entire systems. Viability phase target accuracies should be defined for core and fuel cycle design parameters based on key attributes of the reference configurations. More stringent target accuracies might be required in later phases of the system development activities, and can be defined later. A strong emphasis should be placed on sensitivity analysis tools that use covariance data and their application to determine the nuclear data uncertainties having greatest impact on performance, safety, and fuel cycle predictions. Uncertainties resulting from basic data (including those for plutonium, minor actinides and unconventional core materials) should be quantified for the selected reference configurations. Sources of the uncertainties should be identified and activities required to improve nuclear data should be defined. Additional experiments should be defined to address significant deficiencies that may exist in the available experimental database; desired accuracies for system performance parameters should be taken into account and additional effort should be made to specify these target accuracies. Nuclear data needs should be defined in more detail, as information from the concept point-design studies becomes available to identify more

12

systematically the specific nuclides and energy ranges for which new differential measurements and/or independent confirmatory measurements are needed. Attention should be given to the potential need for additional integral experiments. This would be a natural complement to the integral data documentation efforts recommended above. Consideration should be given to creating a nuclear data advisory group to ensure that Generation IV resources are used effectively to address the data needs of the program. The formation of the group could be an outgrowth of future planned workshops. 3. Nuclear Data Processing, Applications, and Validation 3.1. Workshop Conclusions The design differences between the current generation reactors and the advanced nuclear systems dictate the need to quantify uncertainties in basic data, particularly for the minor actinides and unconventional core materials. The worldwide database of integral data represents a valuable resource for improving calculated design parameters, and as such, a high priority should be placed on identifying previous integral experiment measurements of greatest relevance to Generation IV systems and on documenting/preserving their specifications and measured results. Additionally, improved covariance data are needed for use in the methodology of data adjustments to reduce design uncertainties. It might be possible to use covariance data present in other data files when such data is absent in the nuclear data library being used for analysis. For example, covariance data are available in the Japanese nuclear data file, JENDL-3.3, and these data may be applicable with cross sections in other evaluated nuclear data files (e.g. ENDF/B-VI and -VII). The consistency checks that are performed by the tools generating the covariance data should ensure, in principle, the accuracy (pedigree) of the data. New actinide evaluations have been performed at LANL for uranium isotopes 232 to 241 (with the exception of U-236 and U-240), Pu-239, and Np237. The methods employed include new fits to experimental data, new theoretical calculations, and use of integral experiments to guide choices of data used in the evaluations. This activity has resulted in generally improved prediction of the criticality state for both fast and thermal systems. The predicted flux is a little soft for fast U-235 systems, and additional re-evaluation with latest methods might further improve the situation. There are integral measurements that are expected to be relevant to Generation IV systems. For instance, ANL has maintained a wealth of integral

13

data from the ZPR/ZPPR fast reactor critical experiments program. The available data include keff (enrichment), reaction rate ratios (breeding ratio, burnup swing), spatial reaction rates (power distribution), control rod worths and sodium void reactivity (safety), Doppler worths and flux ratios. Also of note are the ANL activities on gas-cooled fast systems. Critical experiments were performed for the gas-cooled fast-spectrum systems at the ANL Zero Power Reactor (ZPR-9) facility in the mid to late 1970. Three gas-cooled fast reactor assemblies (GCFR-I, GCFR-II, and GCFR-III) were employed in these experiments. In addition to the traditional measurements on criticality, reflector worth, reaction rates distributions, material worths, control rod worths, etc., the measurements also addressed safety problems such as steam entry and analytical problems such as the impact of neutron streaming in gas-filled channels on various integral parameters. Pu/U oxide driver fuel and U02 blankets were used in some of the experiments. These experiments can be used for verification and validation of deterministic analysis methods and the Monte Carlo tools. 3.2. Workshop Recommendations Processing: Processing codes (e.g., NJOY) should be extended to provide covariances in multigroup format as needed for direct use in reactor system analysis. Data Validation: A high priority must be placed on identifying previous integral experiment measurements of greatest relevance to Generation IV systems and on documenting/preserving their specifications and measured results in a peer-reviewed format; the Generation IV program should coordinate its efforts within the framework of international activities in these areas coordinated by OECD/NEA. 4. Nuclear Theory and Evaluation 4.1. Workshop Conclusions In the U.S., the effort devoted to the development and maintenance of basic nuclear data library [Evaluated Nuclear Data File (ENDF/B)] has been in place since 1966. The work is organized within the Cross Section Evaluation Working Group (CSEWG). This library includes evaluated nuclear cross section data for all nuclides relevant for applied nuclear technology. The CSEWG activities cover all aspects of nuclear data, including, experimental measurements, data evaluation, storage formats, processing, and validation. These activities are ongoing mainly at the national laboratories (ANL, BNL, INEEL, LANL, LLNL, and ORNL). The CSEWG activities are well coordinated with the international

14

nuclear data community through interactions with the IAEA and OECD/NEA. Work is underway by CSEWG to release version VII of the nuclear data file in 2005. This version would retain essentially the same format as version VI, so that current processing codes can be used in a straightforward manner. The U.S. national laboratories are also participants in international organizations involved in activities advancing nuclear data. For example, the Working Party on International Nuclear Data Evaluation Cooperation (WPEC) of the OECD/NEA promotes the exchange of information on nuclear data evaluations, measurements, nuclear model calculations, validation, and related topics, and provides a framework for cooperative activities between the participating projects. The Working Party also assesses needs for nuclear data improvements and addresses those needs by initiating joint evaluation and/or measurement efforts. The WPEC activities involve the evaluation projects in Japan (JENDL), United States (ENDF), Western Europe (JEF), and non-OECD countries (BROND, CENDL, and FENDL). Nuclear cross sections are typically evaluated using a combination of measured data and nuclear model calculations. There are ongoing activities in nuclear data theory and evaluation. Physics tools in this area are used for theoretical predictions needed to fill the gap that exists in experimental measurements of nuclear data. An example of such capability is the EMPIRE code system, which is an integrated package of nuclear reaction model codes, input parameter libraries, differential experimental data library and utility codes. EMPIRE is an advanced tool for nuclear reaction data evaluation for energies above 1 KeV. Other comparably robust tools are the GNASH/McGNASH code at LANL and the European THALYS capability. Of the three codes, EMPIRE is the only one already released. ORNL is also developing a cross section analysis and evaluation code (SAMMY) for simultaneous differential and integral data analysis and evaluation to reduce biases on the data. This code can produce covariance data for both the resolved resonance region (based on the Reich-Moore formulation) and the unresolved regions (based on single-level Breit-Wigner formulation). The incorporation of the covariance data produced by the SAMMY code in the ENDF data files would require changes to the ENDF format. Activities in data testing, including benchmark calculations are also ongoing at ORNL. 4.2. Workshop Recommendations Effort should be devoted to the development of tools for generating covariance data for the evaluated nuclear data files (e.g., the planned ENDF/B-VII file) and in multigroup format of direct use to reactor system designers. These tools must

15

also provide covariance information for the unresolved resonance range and for secondary distributions. The evaluated data file formats for the covariance data should be modified to accommodate the additional data. Because of similarities of modeling methods and commonality of much of the basic data, some economies in this effort may be realized by combining covariance data from one evaluation with cross sections from another evaluation, with some suitable renormalization. The overall effort must encompass validation of the covariance data. New evaluations should be performed as specified above in the section on Data Needs. These new evaluations should be incorporated into evaluated nuclear data libraries. 5. Nuclear Data Measurements 5.1. Workshop Conclusions Facilities for nuclear data measurements include the ORELA facility at ORNL. The facility uses a powerful electron accelerator-based neutron source to provide high neutron flux across the energy range of interest to nuclear reactor analysis (0.001 eV to 100 MeV), with very high resolution for cross-section measurement. Nuclear data activities, including cross section measurements, are also conducted at the LANL LANSCE facility in collaboration with other U.S. national laboratories (LLNL and INEEL), universities, and international collaborators. For higher energy neutrons (-0.1 < En < ~400 MeV), measurement capabilities include the GEANIE spectrometer (gamma and X-ray yield measurements), FIGARO (inelastic neutron scattering and fission neutron and gamma measurements), and a neutron-induced charged particle production measurement facility. For lower energy neutrons, (thermal < En < -100 keV), the DANCE spectrometer is used for total neutron capture cross section measurements, and a Lead Slowing-Down Spectrometer (LSDS) is under construction. Together, the LANSCE neutron sources cover all neutron energies from ultra-cold to -800 MeV. Sample sizes required for these facilities are quite small. For GEANIE and FIGARO, samples weighing on the order of grams are required, while milligram and nanogram samples are required for DANCE and the LSDS facilities, respectively. Both differential and quasi-integral measurements are performed at the Gaerttner LINAC facility in Rensselaer Polytechnic Institute (RPI). This facility is a major research laboratory used for basic and applied research, including neutron cross-section measurements. Nuclear data activities at the facility

16

include transmission measurements (0.001 eV to several keV), capture and scattering measurements (0.005 eV - 0.50 keV), self-indication measurements (0.005 eV - 0.5 keV), and alpha (0.005 eV - 0.5 keV) and fission (0.1 eV - 100 keV) measurements in the lead slowing down spectrometer. 5.2. Workshop Recommendations A strong emphasis should be placed on maintenance of the required experimental capabilities and on development of an integrated collaborative effort, coordinated with relevant international activities that will provide the necessary measurements. This should be done in a manner that makes best use of what will almost certainly be limited financial resources. The available experimental facilities, equipment, accelerator targets, and personnel capabilities, which are required to support necessary differential nuclear data measurement activities, are believed to be sufficient to address the anticipated need for new data. A mechanism should be established to facilitate the acquisition, maintenance, storage, distribution, and community usage of sample targets especially in the case of purified stable isotopes and actinides of interest. Consideration should be given to the application of accelerator mass spectroscopy to the measurement of integral data for minor actinides. These data would augment those obtained from integral measurements (critical experiments or irradiated fuel characterization). 6. Conclusions and Suggested Path Forward A principal conclusion from workshop is that current nuclear data are generally adequate for the initial phases of the development of Generation IV systems. The evaluated nuclear data files provide a reasonably comprehensive source of basic data for use in Generation IV system analysis. Existing evaluations of nuclear data, such as ENDF/B-VI, are probably adequate for early preconceptual design development of the VHTR. However, high burnup operation of the VHTR would require re-evaluation of some transuranics data (cross sections, decay data, and fission yields) that have not typically been important in thermal reactor design analysis. Some new differential measurements may also ultimately be needed for selected nuclides in the case of the VHTR, depending on the detailed spectral conditions and specific fuel cycle strategy selected for the system. The fast spectrum systems (GFR, LFR, and SFR) to be deployed for actinide management within a closed fuel cycle, are expected to require additional data evaluation for transuranics, particularly minor actinides, as well as integral measurements for validation of the basic data and their processing tools. Non-conventional

17

structural or fuel-matrix materials may also necessitate new evaluations or measurements of basic data. A systematic approach based on sensitivity and uncertainty analysis was recommended for further specifying data needs. Implementation of this approach requires estimates of covariance data that are largely missing from current evaluated nuclear data files (e.g., ENDF/B). This covariance information should be developed expeditiously and made available to analysts for the purpose of estimating key uncertainties. Several organizations are engaged in the efforts that meet nuclear data needs identified during the workshop. These efforts include nuclear data sensitivity and uncertainty studies for the Generation IV systems, development of an application oriented covariance matrix for uncertainty analysis, and collaboration in planned integral experiments (e.g., GFR physics experiments). Acknowledgments The authors are greatly indebted to the experts from U.S. national laboratories and universities who participated in the nuclear data workshop that formed the basis for this document. Special thanks are due to the U.S. National Nuclear Data Center at the Brookhaven National Laboratory for hosting the workshop. The work described in this document was supported by the U.S. Department of Energy. References 1. "Minutes of Workshop on Reactor Physics and Analysis Capabilities for Generation IV Nuclear Energy Systems, Argonne National Laboratory, Argonne, February 18-19, 2003." Issued on March 11, 2003. 2. "Minutes of Workshop on Thermal-Hydraulic and Safety Analysis Tools for Generation IV Nuclear Energy Systems, Idaho National Engineering and Environmental Laboratory, Idaho Falls, March 18-19, 2003." Issued on April 23, 2003. 3. "Minutes of International Workshop on Reactor Physics Advances for Design and Analysis of Generation IV Nuclear Energy Systems, Hyatt Regency Hotel, Chicago, IL, April 30, 2004." Issued on June 17, 2004. (http://www.physor2004.anl.gov/workshops.htm). 4. "Minutes of Workshop on Nuclear Data Needs for Generation IV Nuclear Energy Systems, Brookhaven National Laboratory, Upton, New York, April 24-25, 2003." Issued on May 14, 2003.

NUCLEAR DATA NEEDS FOR THE ASSESSMENT OF GEN IV SYSTEMS GERALD RIMPAULT Centre d'Etudes Nucleaires de Cadarache Commissariat a VEnergie Atomique (CEA) 13108 Saint-Paul-lez-Durance CEDEX, France Four of the six nuclear systems identified by the Gen-IV international forum are relying in fast reactors. The high performance required from these future FR's calls for very innovative core characteristics compared with conventional fast reactor designs, which in turns give rise to new challenges for the available neutronics methods and data. The ERANOS "formulaire" was developed for reliable, precise and efficient calculations of sodium-cooled fast neutron reactor cores. Such a "formulaire" enables the prediction of all the neutronic quantities of interest for reactor design, operation and safety studies, along with their corresponding uncertainties. Given its achievement in terms of accuracy for existing sodium cooled reactors, the way the ERANOS "formulaire" has been designed should be looked as an example for developing similar tools for GENIV fast cores. The methodology for defining nuclear data needs is briefly covered. Target accuracies for GEN-IV neutronic characteristics is an important point to start with. The covariance data is of significant importance for quantifying the needs and evaluators should provide them with their nuclear data. Hence, nuclear data requests should be associated to uncertainty values. Integral experiments have a complementary role to differential measurements for meeting some nuclear data needs (for instance Pu239 fission). High Priority Level Requests should include those nuclides accessible with current differential measurement technology and should include less important nuclides A list of potential requests is existing, quantifying their uncertainties remain a significant effort particularly when looking at fuel cycle quantities.

1. Introduction Four of the six nuclear systems identified by the Gen-IV international forum are relying in fast reactors. The high performance required from these future FR's calls for very innovative core characteristics compared with conventional fast reactor designs, which in turns give rise to new challenges for the available neutronics methods and data. The current paper is aiming at describing the methodology for defining nuclear data needs and the short list of high priority ones. In order to achieve such a demonstration, the following items are successively covered: • Target accuracies for GEN-IV neutronic characteristics • The covariance data

18

19

The integral uncertainties with or without use of integral experiments Convergence between adjusted and evaluated Nuclear Data Uncertainties Requested for Nuclear Data Way to classify High and Low priority level Requests List of potential Nuclear Data Needs Conclusion and Perspectives 2. Target accuracies for GEN-IV neutronic characteristics The design of the cores and fuel cycles of the Gen IV systems relies on some neutronic characteristics. Target accuracies are requested for these neutronic characteristics and their values differ at the different stages of the design studies (1st stage: viability; 2nd stage: performance). Table 1. Target Uncertainties for Generic GFR Neutronic Characteristics Uncertainties at 1 a

System Development Phase

Parameter

Viability

Multiplication factor, keff BOL

<0.7%

<0.3%

Local power density

<5%

<3%

Structure Damage

<15%

<9%

Reactivity Swing (keff EOL)

(<1.0%)

(< 0.5%)

Breeding Gain

<+/-0.06

<+/-0.04

Performance

Void Reactivity Effect on each component (leakage; non-leak.)

<16%

<10%

Doppler Reactivity Effect

<16%

<10%

Delayed Neutron Fraction

<13%

<7%

Control Rod Worth

<16%

<10%

Y heating

<16%

<10%

These values have been achieved for SUPER-PHENIX in the past [1] and are hence not an inaccessible challenge to the developers. However, performance values might differ slightly depending of the design being considered. Performance values are hence provisional and might be revised in the future following detailed studies quantifying the impact of parameter uncertainties on a given behavior of the GENIV system in advanced development. 3. Base data sensitivity and integral uncertainty In order to get uncertainty on integral characteristics of the core sensitivities and variance covariance matrices are needed

20

The sensitivity coefficient Si of the integral parameter I due to base parameter Pi is written:

1 1 Pi These sensitivities are then used to calculate the value of a relative uncertainty AI on an integral parameter with the variance-covariance matrix B of the parameters P i : ATp = S pBS p. The variance-covariance matrix B associated to nuclear data is the first step in the way to get access to design oriented values with an uncertainty control at each stage of the calculation. Matrix B is made of elements by bu = ,J

COV(P,,Py) —.

P xP: 'o

Jo

The variance-covariance matrix B associated to nuclear data should be associated to nuclear data evaluation and this is the first request for evaluators since values at the moment are scarce. However, in order to make progress on the design request lines, simple variance- covariance matrices have been estimated for a few nuclides The uncertainty values on nuclear data, are given in a simplified multigroup energy structure i.e. 15 energy groups selected between 20 MeV and E(thermal) to represent: • thresholds of fertile isotope fission cross-sections, and of many inelastic cross-sections, up to 20 MeV • the continuum down to the upper unresolved resonance energy limit, • the unresolved resonance energy region, • the resolved resonance region, • the thermal range. Uncertainty values within these large energy regions are nearly fully correlated 4. Base data sensitivity and integral uncertainty The ERANOS "formulaire" was developed for reliable, precise and efficient calculations of sodium-cooled fast neutron reactor cores. Such a "formulaire" enables the prediction of all the neutronic quantities of interest for reactor design, operation and safety studies, along with their corresponding

21

uncertainties. Given its achievement in terms of accuracy for existing sodium cooled reactors, the way the ERANOS "formulaire" has been designed should be looked as an example for developing similar tools for GENIV fast cores. In addition to the refined calculation methods available in ERANOS, highest quality nuclear data are also needed to reach the target accuracies required by the design. And nuclear data evaluations as such, bring too large uncertainties. Results of clean core measurements (critical mass Mc, K+ for K*=l experiments, buckling, spectral indices) performed in MASURCA, ZEBRA and SNEAK facilities have been selected to assess the performance of this unadjusted library. The results of the comparisons are presented in the following table 2. Table 2. Discrepancies on Neutronic Characteristics when Using JEF2.2 JEF2.2 ECCO Library critical mass M c buckling Bm2

Average

Standard

(C-E)/E

Deviation

+ 323 pcm

1460 pcm

-210pcm

1200 pcm

- 50 pcm

2200 pcm

K-infmity

K+ f(Pu-239)/f(U-235)

1.1%

2.6%

Spectral

f(U-238) / f(U-235)

-1.0%

3.7%

c(U-238) / f(U-235)

1.4%

2.2%

f(Pu-240) / f(U-235)

- 4.0 %

8.6%

f(Pu-241)/f(U-235)

-1.4%

5.0%

f(Pu-241)/f(U-235)

- 5.2 %

8.0%

-2%

2.3%

Indices

c(B-10)/f(U-235)

The prediction of JEF2.2 for the clean core measurements is quite satisfactory globally. However, the results obtained are associated with large uncertainties and hide the existence of compensating errors. The ERANOS library, based on the JEF2.2 evaluations, is further improved by nuclear data adjustment procedures. The experimental basis for this adjustment was provided by a large number of measured values (more than 450) in clean critical cores. The resulting adjusted library was called ERALIB1 [2]. The adjustment of the JEF2.2 library has been performed by a least square method generalised to both integral parameters and nuclear data. The functional F to minimise is: F = (G-o 0 )' M"1 (G-a°) + (E - C)' r 1 (E - C)

(1)

22

where a 0 is the JEF2.2 multigroup cross section value with associated uncertainty M expressed in a covariance matrix form, E-C is the difference between the calculated value using JEF2.2 and the experimental value with associated covariance matrix I, and a is the unknown true multigroup cross section value. Using direct and generalised perturbation theory, the sensitivities S of the integral parameters to multigroup constants has been calculated:

iz£=s£z*L C

(2)

*0

By substituting equation 2 into equation 1 and by minimising F with respect to variations in a the maximum probable a values can be obtained. The quality of these adjusted a values is measured by the %2 value which, after adjustment, should be equal to N±(2N)1/2, with N being the number of degrees of freedom:

x 2 = N±V2Nor^- = l ± J — . N

(3)

VN

The number of degrees of freedom is the number of input values (microscopic "priors" + integral observable) minus the number of adjusted quantities (adjusted microscopic data). In the present case it is equal to the number N of integral data. The %2 value should lie within some theoretical limits. If not, this indicates either inconsistent integral values or nonlinearities. To detect spurious information from the integral database the computed % distribution is compared to the theoretically expected distribution. The reasons for which some of these results cannot enter into the adjustment could be either an approximation in the calculational method, an error in the modelling, an experimental error or a more fundamental problem associated with the adjustment hypothesis (no change in the resonance parameters, the scattering anisotropy and the neutron secondary distribution). This adjustment method gives satisfactory results, although the process of detecting inconsistent information over the whole set of available integral experimental data has required considerable expertise and is not yet finished. The objective of the ERANOS formulaire is currently fulfilled with an adjusted library (ERALIB1) giving the following results presented in Table 3.

23 Table 3. Discrepancies on Neutronic Characteristics when Using the Adjusted ERALIBl library ERALIBl ECCO Library critical mass

M

c

Average

Standard

(C-E)/E

Deviation

4- 83 pern

100 pent

buckling

Bm

260 pem

150 pem

K-infinity

K+

123 pem

240 pem

f(Pu-239)/f(U-235)

0.3%

0.5%

f(U-238) / f(U-235)

-1.0%

0.8%

c(U-238) / f(U-235)

1.0%

0.5%

f(Pu-240) / f(U-235)

-1.3%

1.5%

f(Pu-241)/f(U-235)

0.5%

1.2%

Spectral Indices

f(Pu-241)/f(U-235)

-1.6%

1.3%

c(B-10)/f(U-235)

- 1.3 %

0.8%

The prediction of the ERALIB1 library for the clean core measurements is very satisfactory and the results obtained are associated with significantly reduced uncertainties. 5. Convergence between adjusted and evaluated Nuclear Data An improved prediction has be obtained from adjusted nuclear data which would also help the transposition of such results to power reactor characteristics, and it is the reason that such approach has been used for the ERANOS method and data system. This approach has highlighted the complementary role of differential and integral experiments. Many consistent results have been obtained from the adjustment technique and from recent differential measurements. Illustrative examples are Sodium 23 or Plutonium 240 cross sections. Recent differential measurements give evidence of the validity of such an approach while others show difficulties in achieving the desired goal Convergence on Nuclear Data between values deduced from the adjustment and recent differential measurement is observed. Two examples are illustrated: • Na cross sections • Pu240 cross sections 5.1. Na cross sections Na cross section measurements performed at Geel demonstrate a significant change on the inelastic cross section. That change is of similar sign and magnitude than the adjustment foreseen in ERALIBl for which a significant

24

amount of sodium void reactivity worth measurements were being used [3].Furthermore, the resonance structure coherent with the total cross section measured at ORNL Na Inelastic Cross-section

Figure 1. Na Inelastic Cross-Section for various data sets (JEF2, ERALIB1, JEFF3)

5.2. Pu240 cross sections Recent differential measurements performed at ORNL, Hanford and Geel on Pu240 capture cross sections (now JEFF3.1) show significant differences compared with JEF2.2 evaluations with a much better detailed description of the complete resonance structure as it can be seen in the following figure 2. cross section (barnsl 1.00E+0:

1 00E+02

4.69E-04

1.08E-03

2.49E-03

Enerav fMsVl

Figure 2. Capture Cross Section for Pu240 in the Resonance Region

25

The CAPRA PuN Core ( U free core) [4,5] has been studied as it is a core rather sensitive to Pu240. The comparison has been done for its critical mass , first by looking between results obtained with ERALIB1 and with JEF2.2 and then between results obtained with the New Pu240 evaluation complemented by JEF2.2 evaluations and with the full JEF2.2 evaluations. Values are presented in the following tables 4 and 5. Table 4. Critical Mass of the CAPRA PuN Core perturbation between JEF2.2 and ERALIB1 (in pern) Group Capture Fission Inelastic Elastic Sum 13 to 26

139.30

-160.64

0.04 3.79

3J9

-29.56

SUM

681.31

175.58

23.10

198.71

369.78

Table 5. Critical Mass of the CAPRA PuN core perturbation between JEF2.2 and JEF2.2 + New Pu240 evaluation (in pcm) Group Capture Fission Inelastic Elastic Sum 13 to 26

1001.91

163.32

0.00

SUM

1001.69

164.78

1.06

5.45 5.53

843.88 841.38

Large differencies can be observed for the critical mass. Similar trends can be seen on both comparisons but they are coming from different energy regions (region from group 13 to 26 being the resonance region). Only part of this problem comes from the resonance region but the new evaluation being used is incomplete. It does not include the consequences of such resonance changes on the unresolved resonance region and also a better analysis reevaluation of the continuum range. All these changes are now included in the JEFF3.1 evaluation. It is clear that ERALIB1 adjusted library pointed out the right trend and was justifying at the time the necessity of completing the Pu240 evaluation on the preliminary basis of the time. However, one should see that the consequences on resonance parameter descriptionis not so large as it can be seen on the Doppler Effect being calculated with the different libraries. Table 6. Doppler Effect for the CAPRA PuN core: from 300°C to 700°C with different nuclear data sets (in pcm) Isotope JEF2.2 JEF2.2 ERALIB1 + New Evaluation Pu239 -2.9 -6.9 -5.4 Pu240 -216.3 -228.6 -207.1 Pu241 +4.7 +4.6 +4.6 Pu242 -29.3 -31.4 -28.3 Fe -81.2 -83.3 -85.9 Total -325.0 -345.6 -322.1

26

A better resonance description (New Pu240 evaluation, now JEFF3.1) only account for a 6.3% difference. Consequently, it can be said that Doppler effect is sensitive to large resonances and not so much to small resonances and the relatively small details. Small resonances are useful for a better knowledge of the cross-section level rather than for their resonance structures themselves. Hence, this demonstrate the complementary role of integral and differential measurements. Some details of an evaluation are however not specifically necessary for themselves but for the consequences they may have on the cross section level. 6. Integral uncertainties with or without use of integral experiments If values such as resonance parameters rely entirely on differential measurements, for some nuclides and reactions, such as the Plutonium 239 fission cross section, the desired accuracy was so high that differential measurements alone cannot meet the target and required the help of integral measurements. In the following table 7, one can see the material balance change for Super Phenix (SFR) due to the adjustment Table 7. Material balance change for Super Phenix (SFR) due to the adjustment Isotope

Reaction

Pu239

NU FISSION INELASTIC ELASTIC CAPTURE NU FISSION CAPTURE FISSION CAPTURE FISSION NU FISSION INELASTIC ELASTIC CAPTURE

Pu240

Pu241 Pu242 U238

Adjustment (in pcm) Isotope 82.41 Fe56 774.73 40.4 -1.12 Ni58 -38.56 2.72 Cr52 -75.87 179.01 -24.2 Na23 5.99 -2 016 -37.22 -63.55 -24.08 -40.67 -145.21 Total

Reaction INELASTIC ELASTIC CAPTURE ELASTIC CAPTURE INELASTIC ELASTIC CAPTURE INELASTIC ELASTIC ELASTIC CAPTURE

Adjustment (in pcm) 60.44 -49.02 47.68 16.91 134.02 -7.13 ^3.96 3.74 -38.42 10.88 -32.55 70.68

809.69

The leading role of the Pu239 change is obvious and given the magnitude of the different effects which do not compensate each other significantly, the fact that this approach had enabled to predict the Super Phenix Start Up experiments has somehow confirmed the adjustment. This overall approach had enabled to

27

predict the Super Phenix Start Up experiments and hence this has somehow confirmed the nuclear data adjustment. The method implemented to develop and validate ERANOS is based on a rigorous approach involving the control of errors at each development step. This method makes it possible to predict fast neutron reactor core characteristics such as critical mass and average power per assembly with respective uncertainties of 158 pern and 2.44 % for the SUPER-PHENIX initial core, and 140 pem and 1.86% for PHENIX. New fast reactor core designs falling outside the code validation domain (for example, because of a fuel other than oxide or a coolant other than sodium) will need specific analyses associated to these changes. The uncertainties for a GFR are calculated [6] for: • the reactivity • the (239p u Fission /235u Fission) reaction rate ratio at the core centre • the (238u Fission /235u Fission) reaction rate ratio at the core centre These ratios are respectively representative of intermediate and fast neutrons. Si evaluated variance-covariance matrices have been added to both variance covariance files (JEF2.2 and ERALIB1). Table 8. Uncertainties on GFR characteristics using different nuclear data sets Reactivity (23!,Pu Fiss /235U Fiss) (238U Fiss / ^ U Fiss) JEF2.2 1390 pem 2.4% 3.3% Uncertainty ERALIB1 312 pem 0.5% 1.2% Uncertainty

Integra] uncertainties are consistent with those of standard sodium-cooled oxide fast reactors ones and rather good for the current state of GFR studies.However, the validity of the experiments used for GFR being questionable, other analyses will make use of relevant past experiments if any and will require new integral experiments such as the ENIGMA physics programme at CEA in support to GFR cores. An improvement of the formulaire and hence of the adjusted libray will make use of these new integral experiments but also of the most recent advances in nuclear data evaluations such as the JEFF3.1 ones. 7. Uncertainties Requested for Nuclear Data The uncertainties on nuclear data as they have been obtained with ERALIB1 fulfil the SFR and GFR design requests for those major nuclides of interest Pu239, Pu240, Pu241, U238, U235 , Fe, Cr, Ni, Na, O. It is therefore worth

28

looking at the standard deviations before and after adjustment to find out where new differential measurements are needed. The approach is approximate since correlations within ERALIB 1 offer a way to significantly reduce the impact of standard deviations. And there are still pending problems associated to some ERALIB 1 nuclides such as the one on structural materials (Fe, Cr, Ni) which need to be considered. In the following figures, one can see the improvement associated to the uncertainties on some nuclides. Pu239 Fission XS Standard Deviation 0.080 0.070 c

BJEF2

0.060

D ERALIB1

.2 0.050 ^

0.040

"D

I 0.030 c CO

55 0.020

:

^™Bfflffll

0.010 0.000

,p ,p s

O0 A^

sr s? £" js' s? sr

o°° # v

t»

T,-


fc-


t\/

o,

N^ &
<s» sv

<& #y


\


tx-



<£>
%•


s? , N-

#

Energy (eV)

Figure 3. Pu239 Fission Cross Section Standard Deviation before and after adjustment U238 Capture XS Standard Deviation 0.035

/ / # / / ' # & f r /' f f /" Energy (eV)

Figure 4. U238 Capture Cross Section Standard Deviation before and after adjustment

29 Ni58 Capture XS Standard Deviation

0.000

%

fc''

\"

V'

\~

<),*

Energy (eV)

Figure 5. Ni58 Capture Cross Section Standard Deviation before and after adjustment

These improvements show a reduction of the standard deviation which is only part of the problem since correlations between groups, between reactions and between isotopes are contributing to a significant reduction of the uncertainties. Hence, one can conclude that: • Nuclear Data Requests are of little use if requested uncertainties are not mentioned. • Nuclear Data Variance-covariance matrices should be provided with the evaluation. • Requested uncertainties associated to nuclear data should be spelt out in terms of energy range and reaction For instance, Pu239 fission is requested with a 2% uncertainty above 4.0 eV and 0.3% below; U238 capture is requested with a a 2% uncertainty above 4.0 eV and 0.5% below and Ni58 capture is requested with a a 10% uncertainty above 67.0 KeV and 15% below. Given this approach, little emphasises should be given to highly sensitive nuclear data where integral measurements can be set up to fill the gap but for which required uncertainties are out of reach of differential measurements. Once a complete analysis leading to nuclear data evaluation needs with their associated requested uncertainty has been performed, another analysis should be made on the possibility of meeting these requests through differential measurements. Hence, those requests which can be met with current differential measurement technology should be put on high priority list. Others should be in

30

the standard priority list. The fact that some requests such as the one on Pu239 fission cannot be easily met should not hide the needs of less important nuclides but more achievable. 8. Conclusion and Perspectives The methodology for defining nuclear data needs has been briefly covered. Target accuracies for GEN-IV neutronic characteristics is an important point to start with. Then, the covariance data associated to evaluated data appears of significant importance for quantifying the needs and evaluators should provide them with their nuclear data. When defining nuclear data requests, one should associate them to uncertainty values. Integral experiments have a complementary role to differential measurements for meeting some nuclear data needs (for instance Pu239 fission) which are out of reach of current technologies. High Priority Level Requests should include those nuclides accessible with current differential measurement technology and should include less important nuclides [7,8]. A list of potential requests is existing, quantifying their uncertainties remain a significant effort particularly when looking at fuel cycle quantities. References 1.

G. Rimpault et al, "The ERANOS Code and Data System for Fast Reactor Neutronic Analyses", PHYSOR'02, Seoul, Corea2. E. Fort, W. Assal, G. Rimpault et al. , « Realisation and Performance of the "Adjusted Nuclear Data Library ERALIBI for calculating fast reactor neutronics", International Conference on the Physics of Reactors, PHYSOR96, September 16-20, 1996, MITO, IBARAKI, JAPAN 3. G. Rimpault, H. Oigawa, P.J. Smith, "Assessment of Latest Developments in Sodium Void Reactivity Worth Calculations", International Conference on the Physics of Reactors, PHYSOR96, September 16-20, 1996, MITO, IBARAKI, JAPAN 4. G. Rimpault, R. Jacqmin, M. Martini, R. Soule, P. Smith, S. Ohki, "The CIRANO Experimental Programme for the Characterisation of Highly Enriched Plutonium Oxide Fuel in Fast Reactors", Proc. of the Int. Conf. on the Physics of Nuclear Science and Technology, PHYSOR98, October 5-8, 1998, Long Island, NY, USA 5. P. Smith, G. Rimpault, O. Bouland, E. Fort, R. Soule, S. Ohki, "Integral and Differential Experiments to Assess Fast Reactor Characteristics when using Degraded Plutonium", Proc. of the Int. Conf. on the Physics of Nuclear Science and Technology, PHYSOR98, October 5-8, 1998, Long Island, NY, USA

31

6.

7. 8.

J.C. Bosq, G. Rimpault, "Methodology for a Large Gas-Cooled Fast Reactor Core Design and Associated Neutronic Uncertainties", Proc. Int. Conf. PHYSOR'2004, Chicago, USA, 04/2004 G. Rimpault, "Nuclear Data Needs for Generation IV Nuclear Systems", HPRL group of the OECD, 27 May 2004, Aix en Provence, France G. Rimpault, " Nuclear Data Needs and Developments" in Minutes of Int. Workshop on Reactor Physics Advances for Design and Analysis of Generation IV Nuclear Energy Systems, Chicago, IL, April 30, 2004

N U C L E A R DATA N E E D S FOR G E N E R A T I O N IV LESSONS FROM B E N C H M A R K S ?

S.C. VAN DER MARCK, A. HOGENBIRK, AND M.C. DUIJVESTIJN Nuclear Research and Consultancy Group NRG P.O. Box 25, 1755 ZG Petten, the Netherlands E-mail: vandermarck@nrg-nl. com One of the topics of the International Workshop on Nuclear Data Needs for Generation IV Nuclear Energy Systems was the assessment of the overall quality and validation status of current nuclear data files and processing capabilities. In this paper we report on such an assessment for the latest preliminary releases of ENDF/B-VII and JEFF-3.1.

1. Introduction One of the topics of the International Workshop on Nuclear Data Needs for Generation IV Nuclear Energy Systems was the assessment of the overall quality and validation status of current nuclear data files and processing capabilities. In this paper we report on such an assessment for the latest preliminary releases of ENDF/B-VII and JEFF-3.1. These preliminary releases, denoted here as ENDF/B-VIIp and JEFF3.1T3, have become available at March 11, 2005. We present the results of criticality benchmark calculations with these two new libraries, to evaluate the quality of these new releases for criticality calculations. Also, we present calculations of /?eff for several systems, as tests of the delayed neutron data of the different libraries. Almost all benchmarks were taken from the International Handbook of Evaluated Criticality Safety Benchmark Experiments from the OECD 2 . These benchmarks are subdivided in main categories according to three criteria. (1) The main fissionable isotope. The systems containing uranium-235 are subdivided according to the enrichment in 235U: there is low enriched uranium (LEU: wt% 235U < 10), intermediate enriched (IEU: 10 < wt% 235U < 60) and high enriched (HEU: wt% x 32

33

235U > 60). There are also plutonium systems (PU), mixed uranium/plutonium systems (MIX), and 233U systems (U233). The physical form of the fissile material: there are metal systems (MET), compound (COM), solution (SOL) and miscellaneous systems. The neutron spectrum: if more than half of the fissions occurs for incoming neutron energy below 0.625 eV the spectrum is thermal (THERM), if more than half occurs between 0.625 eV and 100 keV it is intermediate, and if more than half occurs over 100 keV it is fast (FAST). If none of these applies, the spectrum is classified as mixed. The current-day reactor types, such as PWRs and BWRs, are therefore categorized LEU-COMP-THERM. The reactor types envisaged in the Generation IV program are typically in different categories, for which not as many benchmarks are available, such as MIX-COMP-FAST, or MIX-SOL-THERM. 2. R e s u l t s In Table 1 we listed, per category, the number of benchmark cases we have used in this study. Also here, the largest number of cases is in the LEUCOMP-THERM category. Table 1. The number of criticality safety benchmark cases used in this study.

LEU IEU HEU MIX PU U233 total

COMP therm / inter / fast 257/ / 9/ /

/

/

35/

/

/

/

8/ 309 /

/ /

1

MET therm / inter / fast 1/ / 34 / /

/ / 1

35 /

/ / 10 5/13 / 4

i/ /

r 4

6/38

SOL therm 49 81 3 105 5 243

Total 307 19 133 43 113 17 632

The results of these calculations are shown as calculated over benchmark values (C/E) in Pigs. 1 (for LEU-COMP-THERM benchmarks) and 2 (for fast spectrum benchmarks). For most of the LEU-COMP-THERM benchmark cases, the C/E values lie within the benchmark uncertainty range, which is a significant improvement over the previous releases JEFF-3.0 and ENDF/BVI.8. For the fast spectrum cases, the results are not as good. Some

34

HI

(3

Figure 1.

C / E values for LEU-COMP-THERM benchmarks,

cases show a large deviation from the benchmark value, such as PU-METI N T E R - 0 0 2 (ZPR-6/10). Also, for H E U - M E T - I N T E R - 0 0 6 (Zeus, LANL) the results vary significantly from case 1 to case 4, which is probably due to the variations in spectrum. Presumably the description of either U-235 or Cu (used as reflector here) is not accurate enough in the relevant part of the spectrum to calculate this benchmark with good accuracy. Table 2. Average of JEFF-3.1T3 (in pcm)

LEU IEU HEU MIX PU U233

calculated

COMP therm / inter / fast -96/ / -324 / /

/ 523 /

/ / 303

/

/

-466 /

/

— benchmark

MET therm / inter / fast -139 / / / / -102 - 2 6 7 / 105 / 219 / / -167 / 3993 / 290 / / 258

value

for

SOL therm 116 175 150 553 -227

The results for /3eff, used to test the accuracy of the delayed neutron data, are shown as C/E values in Fig. 3. For most cases the results for

35 Experimental uncertainty JENDL-3.3 i ENDF/B-Vllp JEFF-3.1T3 t-

1.05 -

1.04 1.03

1.02

1.01 •• i * • i

^ . • ! 0.99

T—r- in to

0.98

15

10 Figure 2.

20

25

30

11 E E

35

40

C / E values for fast spectrum benchmarks.

Table 3. Average of calculated ENDF/B-VIIp (in pcm)

LEU IEU HEU MIX PU U233

oo o o

JUI uimm

I

COMP therm / inter / fast -146 / / -342 / /

/

/

459 /

/

/

/

-169 /

/

12

— benchmark

value

for

fast

SOL therm 141

222 657 / / 273 / 4585 / 431 / / -135

216 156 526 455

MET therm / inter / -119 / /

/

/

-273 /

33 /

ENDF/B-Vilp and JEFF-3.1T3 are similar. Only for the plutonium dominated systems there are small differences. On the whole, the results for fast spectrum systems are good, but for thermal spectrum systems the results are 5 to 6% overestimated. 3. Conclusions From the results from all these benchmark calculations, some of which are presented in the previous section, we conclude the following.

36

C/E values for beta-effective

1.15

1.05

0.95

0.85

«HI

0.8

Figure 3. C / E values for (3eg in many systems. For the systems on the left the neutron spectrum was thermal, for the systems on the right it was fast.

The benchmark collenction ICSBEP is very useful for data testing. The IRPhE project will add new types of benchmarks to this sort of testing. The new releases of JEFF and ENDF/B, i.e. JEFF-3.1T3 and ENDF/B-VIIp, show significant improvements in many areas. In particular the fcefT prediction for LEU-COMP-THERM benchmarks, which used to be low for JEFF-3.0 and ENDF/B-VI.8, is now within the benchmark uncertainty range for many benchmarks. Thefceffresults for fast spectrum benchmarks are getting better too, although the deviations from the benchmark values are in general larger than for thermal benchmarks. There is not a very clear trend with the uranium concentration in uranium solution benchmarks. For plutonium solution benchmarks there is a weak trend of lower fceff with increasing plutonium concentration. The delayed neutron data yield good results for fast spectrum systems, but yield a 5 to 6% overprediction of /3eg- for thermal systems. The temperature dependence of thermal scattering in ENDF/BVIIp shows unexpected behavior (as it did in ENDF/B-VII.8).

37

References 1. J.F. Briesmeister, MCNP - A General Monte Carlo N-Particle Transport Code, Version 4C, Technical Report LA-13709-M, Los Alamos National Laboratory, USA (2000) J.S. Hendricks, MCNP4C3, Report X-5:RN(U)-JSH-01-17, Los Alamos National Laboratory, USA (2001) 2. J. Blair Briggs (Ed.), International Handbook of evaluated Criticality Safety Benchmark Experiments NEA/NSC/ DOC(95)03/I, Nuclear Energy Agency, Paris (September 2004 Edition) 3. R. Klein Meulekamp and S.C. van der Marck, Nucl. Sci. Eng., to be published (see also S.C. van der Marck and R. Klein Meulekamp, Proc. PHYSOR-2004 (Chicago, 2004)) 4. S.C. van der Marck, R. Klein Meulekamp, A. Hogenbirk, and A.J. Koning, Proc. Nuclear Data 2004 Conference ND2004 (Santa Fe, 2004)

CORE DESIGN ISSUES OF THE SUPERCRITCAL WATER FAST REACTOR3 MAGNUS MORI, ANDREI RINEISKI, WERNER MASCHEK Institutfiir

Kern- und Energietechnik, Forschungszentrum Hermann-von-Helmholtz-Platz, 1 Eggenstein-Leopoldshafen, 76344, Germany

Karlsruhe,

VALENTIN SINITSA IPPE, Bondarenko Sq., 1 Obninsk, Kaluga 249030, Russia

The Super Critical water Fast Reactor is a Generation IV reactor concept, which presents new and challenging design issues. A correct estimation of the void effect for this watercooled pressurized system is of fundamental importance to assess its theoretical feasibility. Hence, in this work an overview of the void effect analysis is shown together with the resulting core design issues. The effect of the application of different cross section libraries and models on the core design is also treated.

1. Introduction Within the framework of the development of Generation IV reactors, the safety analyst has to face more and more challenging tasks related to the correct modeling of e.g.: heterogeneous structures and new fuels. The introduction of MOX fuels and minor actinides targets in reactor cores requires for instance not only an accurate description of the material properties, but often a new core design to make up for the deteriorated safety coefficients and kinetics parameters [1]. A very good example of this is the Supercritical Water Fast Reactor (SCFR). Supercritical water reactors are being extensively studied for their forecasted economic advantages and in the case of the fast neutron spectrum option for its potential use as a plutonium and minor actinides burner. The SCFR therefore presents very appealing characteristics, and thus, in order to assess the basic feasibility of such a design, its safety parameters were carefully investigated. In particular the void reactivity effect was extensively analyzed, as * This work was supported by JAPCO (contract No. XI-0118). 38

39

its negativity is a fundamental requirement for the inherent safety of watercooled reactors under accident conditions. Because of the fuel and coolant characteristics, the epithermal neutron spectrum and the need to have excess neutrons to enable waste transmutation, a highly heterogeneous structure is adopted for this core in order to prevent the deterioration of the void effect [2]. On the other hand, the studies performed in recent years outlined the difficulties related to a correct modeling of the SCFR core and the uncertainties associated with the void effect estimation [3,4]. The models that were adopted for the estimation of the void effect are: a classical R/Z deterministic, 2D Monte-Carlo, and 3D pin-by-pin Monte-Carlo (where the geometry of the pins was described individually, and then, making use of the universe definition feature of MCNP, reproduced several times, assigning to them not only different positions, but also the corresponding physical properties). Finally a fully coupled fluid-dynamics/Monte-Carlo procedure was implemented and applied to perform a careful core design analysis using different cross section data libraries and models. 2. The SCFR The SCFR is a once through direct cycle reactor, where all feed water flows through the core to the turbine at supercritical pressure. Two very important characteristics of this system are: the high , temperature rise of the -,-*•-"•-< - w»». coolant in the core, which .J ~ ,. r..^.,^r-. - **" ' '"il \ inlet/outlet coolant density \ ,Q - . / f 1 - ' " * ^ 3 - S s ' *-S ratio of about 9; and a very J X, , , U ' ^ v , %L - J-. j '<-, heterogeneous core X T X \l « p * > - - e0 ^A \ geometry (see Figure 1). S £( Ti \ ££ £\ V The core is in fact ;* subdivided in several zones: v., v<.*y seedi /'••- V the seed areas, the radial \^ ^C*.. • ^ A ^ ^ - ~ ^ .J> U.X* N blanket regions, and the ,i | Seed 3 solid moderator pins. The seed areas are made of highly plutonium-enriched uranium

(-23%).

The

Figure 1. Core axial section.

40

blanket areas are made of depleted uranium with an average 235U content of 0.5%, and finally the solid moderator pins, the thin dark elements within the blankets, are made of zirconium hydride (ZrHi 7 ). The use of such a heterogeneous structure is meant to help achieving a negative void effect by smoothening the neutron spectra hardening and improving the leakage component. Unlike Liquid Metal cooled Fast Reactors (LMFR), the SCFR is a pressurized system, therefore Loss of Coolant Accidents are design basis and the void reactivity coefficient becomes the fundamental parameter for the assessment of the safety of this reactor. 3. Void Effect Analyses With the intention of characterizing and understanding better the distinctive SCFR features, the safety analyses focused at first on the study of the reactor neutron spectra. An R/Z model of the core was then developed and using the Sn code TWODANT the spectrum was evaluated (see Figure 2). This figure compares the integral seed and blanket spectra at nominal (wet), void (dry), and after disintegration of zirconium hydride conditions (no H in the core).

0.1

0.01 o ex {/J

10.001

0.0001 0.1

10

/ WET, SEED / WET, BLANKET / DRY, SEED / DRY, BLANKET / NO H, SEED / NO H, BLANKET 1000 100000 Energy (eV)

Figure 2. Integral seed and blanket spectra at nominal (wet), void (dry), and after disintegration of ZrHu (no H in the core) conditions.

It can be noticed that both for wet and dry conditions, the blanket spectra are softer than the seed spectra because of the presence of the zirconium hydride layers. Voiding in fact, changes the seed spectra more appreciably than in the

41

blanket. Moreover, the removal of the solid moderator hardens significantly both the seed and blanket spectra, bringing them closer to those of a conventional LMFR. These preliminary results outline another relevant and specific attribute of this reactor: the epithermal neutron spectrum. The SCFR is in fact neither a thermal, nor a fast reactor; the main implication of this is that the handling of the cross section data libraries needs special care. The reactor in fact operates across the resonance region (~10-^10000eV [5]), and particularly across the socalled unresolved resonance region, which to be accurately described would require the use of either a very fine ad hoc group structure, or of the continuous energy cross section treatment typical of Monte-Carlo codes. Another relevant outcome of the initial investigations is the strong correlation between the void and the Doppler effect. It was in fact observed that, for an average fuel temperature of 1200K, the void effect is about six times smaller in absolute value than at 300K. This important phenomenon, together with the strong coolant density gradients and the resulting neutron spectrum variations, suggested the need of an additional tool to estimate the physical parameters, and led to the development of a coupled Monte-Carlo fluid dynamics methodology [6]. A summary of these results is given in Table 1, where it should be noted that the approximation levels include also the discrepancies among different cross section libraries (ENDF/B-VI.8, JEFF-3.0, and JENDL-3.3), besides the statistical uncertainties and the application of different models e.g. the use of thermal scattering data for water and zirconium hydride. TABLE 1. Estimation of the void effect (comparison of different methods).

Method

Codes

Void

A

Deterministic [7]

CITATION

~ -1 $"

n.a.

Deterministic

TWODANT

~ -1 $

±5$

Monte-Carlo (MC)

MCNP4C

~ -0.5 $

±2.5$

Coupled MC

MXN/MCNP

~0$

± 1$

3.1. Effects of the cross section models As mentioned, the SCFR presents to the analyst new and unique features, which require the use of new or extremely flexible tools of investigation. Hence the " 1 $ corresponds to 377 pcm for the SCFR core at the beginning of cycle.

42

decision to adopt and develop a coupled fluid-dynamics Monte-Carlo system to take into account as accurately as possible the different phenomena, which determine the void reactivity coefficient. In the course of the study the introduction of a more and more detailed geometrical model, together with a better definition of the thermophysical parameters (e.g. fuel and coolant temperature and density distributions) helped reducing the uncertainties related to the method of analysis. The approximation levels indicated in Table 1 however, depend not only on the method, but also on the accuracy of the chosen cross section set and of the chosen models. Consequently, in order to establish a target value for the void reactivity coefficient that would guarantee a negative feedback beyond the indicated approximation levels, calculations comparing ENDF/B-VI.8, JEFF-3.0, and JENDL-3.3 were performed. The so obtained target value might be very conservative, but the lack of a reference cross section library for this design, or of applicable integral experiments, justify this kind of approach for these preliminary investigations. 4. Core Design Development The last column of Table 1 shows that for the most sophisticated method, the coupled Monte-Carlo fluid-dynamics scheme, the estimated error band is in the order of 1$. Moreover, taking into account the approximations of the method itself and of the adopted models, another dollar should be added to the original estimate. Therefore, the target value for the total void reactivity coefficient was set to -2$. Several design modifications were considered with the aim of improving the void effect of the SCFR. Two strategies were followed: the first one concentrated on the optimization of the geometrical aspects of the core, while the second one focused more on the core materials and particularly on the fuel enrichment distribution. The new suggested geometrical solutions improve the core heterogeneity, the overall amount of solid moderator in the core, and neutron streaming in case of voided conditions, while keeping the original number of seed and blanket subassemblies. Instead, the new enrichment distributions aim at: 1. decreasing the temperatures in the lower part of the core, therefore increasing the relative number of neutron captures in the region with the highest neutron importance, and 2. favoring radial neutron leakage by increasing power in the core outer ring.

43

A summary of the results associated to the just mentioned solutions is shown in Figure 3 (see [8] for a detailed description of the considered design variants). o-

"^r

-100-200-. ~

-300- '

ig

-400- '

approximation

boundary: model only

° —,

•g 8

-500-

*

-600-'

J^

1 -700-' ..

approximation

•»

boundary! model + cross

sections

•800-'' -900 J ' -1000-1

-....„., -

•953!

r

./ / s / / ^ ./ Figure 3. Dependence of the void worth on different design configurations.

The void worth reported in Figure 3 indicate that the target value of -2$ can be achieved, but at the cost of the introduction of a new core subassembly arrangement (see Figure 4) together with a significant redistribution of the plutonium enrichment, and a new blanket subassembly design. 5. Conclusions The estimation of the void reactivity effect for the SCFR requires not only the development of new tools, but also a careful treatment of the nuclear data and models. The neither fast, nor thermal neutron spectrum of this supercritical water-cooled reactor, together with the highly plutonium enriched fuel, make this system rather unique. The use of MCNP and of its continuous energy crosssection treatment gets around some of the modeling difficulties related to the very heterogeneous geometry and the epithermal neutron spectrum. However, a few issues remain to be addressed e.g.: the lack of a scattering model for

44

supercritical water, the relevant disagreements in the void worth predicted by the different data libraries, the lack of detailed LOCA transient analyses.

Figure 4. Advanced heterogeneous model: 'flower model' (central clusters).

In order to overcome the uncertainties associated to these issues and to the chosen evaluation method, a conservative target value for the void worth was estimated and, through careful design, achieved. However, the cost and the engineering feasibility of the described solutions have not been assessed yet. Further attention should then be dedicated to the choice and, eventually, evaluation of the right cross section set, which would significantly improve the margins of confidence of the void worth calculations. Acknowledgments We would like to thank Prof. Thomas Schulenberg, Prof. Gunter Lohnert, Mr. Frank Kretzschmar, and Dr. Xue-Nong Chen for their guidance and support.

45

References 1.

2.

3.

4.

5. 6.

7.

8.

W. Maschek, A. Rineiski, K. Morita, M. Flad, "Inherent and Passive Safety Measures in Accelerator Driven Systems: A Safety Strategy for ADS", Global 2001, Paris, France (2001). T. Jevremovic, Y. Oka, S. Koshizuka, "Effect of Zirconium-Hydride Layers on Reducing Coolant Void Reactivity of Steam Cooled Fast Breeder Reactors", Journal of Nuclear Science and Technology, 30[6], pp. 497-504 (1993). Mg. Mori, V. Sinitsa, "Uncertainties in void effect estimations for a super critical water fast reactor with transmutation capabilities", Jahrestagung Kemtechnik 2003, Berlin, Germany (2003). Mg. Mori, , V. Sinitsa, W. Maschek, "Monte-Carlo and Deterministic Models for Void Effect Calculations in the Supercritical Water Fast Reactor", GENES4/ANP 2003, Kyoto, Japan (2003). A. E. Waltar, A. B. Reynolds, Fast Breeder Reactors, Pergamon Press (1980). Mg. Mori, W. Maschek, X. N. Chen, G. Lohnert, "Advanced Methods of Analysis for the SuperCritical Water Fast Reactor", ANS Transactions, 91, p. 241, USA (2004). T. Jevremovic, Y. Oka, S. Koshizuka, "Core Design of a Direct-Cycle Supercritical-Water-Cooled- Fast Breeder Reactor", Nuclear Technology, 108,24,(1994). Mg. Mori, W. Maschek, A. Rineiski, "Heterogeneous Cores for Improved Safety Performance a Case Study: the Supercritical Water Fast Reactor", to be published in the proceedings of the 13th International Conference on Nuclear Engineering, Beijing, China, (2005).

GFR CORE NEUTRONICS STUDIES AT CEA J.C. BOSQ, V. BRUN-MAGAUD, G. RIMPAULT, J. TOMMASI CEA/DEN/DER/SPRC, CEN CADARACHE 13108, Saint Paul lez Durance, France A. CONTI, J.C. GARNIER CEA/DEN/DER/SESI, CEN CADARACHE 13108, Saint Paul lez Durance, France

The Gas cooled Fast Reactor (GFR) is a high priority in the CEA R&D program on Future Nuclear Energy Systems. After preliminary neutronics and thermo-aerolic studies, a first He-cooled 2400MWth core design based on a series of carbide CERCER plates arranged in an hexagonal wrapper were selected. Although GFR subassembly and core design studies are still at an early stage of development, it is nonetheless possible to identify a number of nuclear data needs that could have some impact on the actual design: new materials, decay heat contributors...

1. Introduction The Gas-cooled Fast Reactor (GFR) is one of the six reactor concepts selected in the framework of the Generation IV initiative and has high priority in the CEA R&D program on Future Nuclear Energy Systems. The GFR design goals and criteria are: • Fast neutron spectrum with zero breeding gain (fuel cycle fed with depleted/natural uranium, homogeneous recycling of actinides), • High coolant outlet temperatures (850 °C and higher), • Use of a refractory fuel favouring close retention of fission products, • High burn-up with an intermediate objective of about 5-8 % FIMA, • Power conversion using direct Brayton cycle, also indirect cycle option with He as primary coolant should be assessed. In a first exploratory step, CEA design studies focused on a unit power of 600 MWth using a very challenging CERCER fuel: only 30 %vol. of matrix for 70% of U,Pu ceramics. A first consistent set of results was found [1][2]. However, several limitations were found with this concept, especially in terms of competitiveness. Therefore, in a second step, it was decided to focus on a larger 46

47

unit power, 2400 MWth, for which a less challenging form of CERCER fuel (higher matrix fraction) can be used [3]. This innovative concept implies specificities, compared to sodium-cooled fast reactors, in particular in the nuclear data field. After a brief description of the concept, several particular aspects are described in this paper. 2. Core Design Studies The GFR fuel subassembly currently under study at CEA consists of an hexagonal wrapper containing three diamond-shaped sub-bundles. Each subbundle contains a series of CERCER plates, having the form of a regular rhombus (30 degrees of inclination), arranged in order to form an hexagon. The plates are held in place by internal devices (Figure 1).

Figure 1. GFR Fuel sub-assembly (plate type element).

Previous neutronics impact studies and thermo-aerolic characterizations have allowed us to classify several material candidates; a (U,Pu)C fuel compound in a SiC inert matrix, which acts as the first barrier, was retained as a reference. After preliminary thermo-mechanical studies, a first design of the CERCER plate has been selected. The core volume fractions are 40% He coolant, 20% structures made of SiCf/SiC, 40% CERCER composed of (U,Pu)C + SiC matrix + He gap. After some iterations, it was decided to consider a proportion of 56% (U,Pu)C within the CERCER to ensure a net breeding gain close to zero. The main neutronics and thermo-aerolic characteristics were calculated using the ERANOS and COPERNIC codes [4] [5]. Two fuel situations were studied: • the first cycle, with a (U,Pu)C initial load,

48

•

the equilibrium state with the TRU elements recycled and an external feed of natural uranium, leading to a proportion of minor actinides in the core around 1-2%.

The main neutronics and thermo-aerolic characteristics are summarized in Table 1. Table 1. Main GFR neutronics thermo-aerolic characteristics. First cycle

Average Pvol (MW.m3) Core Diameter, Height (m) Number of fuel S/A Number of plates per fuel S/A TRU fuel fraction (%) Pu+MA inventory (tons/Gwe) Core management EFPD Average discharge burn up (FTMA) Max damages (dpa SiC) Breeding Gain (BOL/EOL) Doppler Constant (10 5 ) (BOIVEOL) He depressurization (10'5) (BOIVEOL) Delayed neutron fraction (10~5) (BOL/EOL) He Pressure Helium core inlet/outlet temp. (°C) Core pressure drop AP (bar) Max. cladding/fuel temp. BOL (°C)

Equilibrium 100 4.44 / 1,55 387 27 15.2 18.5 7.7 10.1 3x831=2493 10.1 10.1 163 152 -0.07 / -0.04 0.05 / -0.05 -1872 / -1175 -1405 / -968 212 / 253 253 / 257 388 / 344 347 / 332 70 bar 480 / 850 0.62 1075 /1210

Although GFR subassembly and core design studies are still at an early stage of development, it is nonetheless possible to identify a number of nuclear data needs that will be important irrespective of the final design choice. These needs arise in connection with: • the use of composite fuels with new materials (significant amount of inert matrix), compared with standard fast reactors with sodium coolant, which requires a good knowledge of the corresponding nuclear data, with the associated uncertainties. The important proportion of neutrons in the resonance energy region, as a result of scattering on the inert matrix, which increases the sensitivity of neutronic parameters to this energy range (large Doppler effect in particular); • the recycled TRU elements and the possible external feed of minor actinides, which condition the breeding gain, a very important design parameter to be determined to within ± 0.05; hence the need for a sufficiently accurate knowledge of these heavy nuclides cross sections in the [1-800] keV range;

49

•

the decay heat removal evaluation in case of a loss of pressure in the primary circuit, which calls for sufficiently accurate fission yields and decay data.

3. The Specificities of GFR Materials The use of composite fuels with a (U,Pu)C actinide compound in a SiC inert matrix implies new materials compared with standard fast reactors with sodium coolant. The volume fractions of structural materials are quite similar (-26% of SiC in the GFR core, compared with -25% of Iron in Phenix, for example), but the neutronics properties of these materials are very different, because of the more important neutron scattering effect in the GFR. The constraint of self-breeder core also calls for a very strict control of the neutron balance in the core; a very efficient Zr3Si2 reflector material was retained as a reference. The characteristics of this material increase the sensitivity of the k-effective to structural materials, compared with standard fast reactors with sodium coolant. Figures 2-3 compare the sensitivity of the k-effective to absorption and scattering for GFR and Phenix reactors. For the absorption term, the Pu239 contribution is more important for Phenix, due to the higher enrichment but the orders of magnitude are quite similar for the two reactors. For the scattering term, the differences are significant and the scattering effects are more important in the GFR core (inelastic scattering for U238 and elastic scattering for structural materials) although the absorption term remains largely dominant. Note that zirconium in the reflector has a positive contribution. It can be seen in Figure 4 that the scattering energy ranges of importance are very different for these two reactors. The scattering effects of Si, C and Zr increase significantly the sensitivity of the k-effective to this reaction. All these considerations create new needs for the GFR systems in terms of nuclear data cross-sections and associated uncertainties but there are also common needs with other fast reactors (heavy nuclei nuclear data in particular).

50

Sensitivity of the k-effective to JEF2.2 Nuclear Data Absorption (pcnV%)

500 400 300

'

200

lGFR I Phenix

100

-IL

0 -100

s

| g

z>

W g

-200

8 S 5 ? » 5 ° =

=

B

=

=

D.

Q.

0.

Q.

Q.

N

E

o j it

8 S O

I

z

^

-300

Figure 2. Sensitivity of the k-effective to JEF2.2 nuclear data absorption. Sensitivity of the k-effectlve to JEF2.2 Nuclear Data Scattering (pcnV%)

a i 3

a ¥ %5 ^ 5 J

3 Q.

3 0.

3 Q-

3 CL

3 Q.

_ ) ' * 4 s $ s 2 £ S z

C ^

QGFR B Phenix

Figure 3. Sensitivity of the k-effective to JEF2.2 nuclear data scattering. Sensitivity of the k-effective to the JEF2.2 elastic scattering (pcm/%) 15 10 5

^ 5 Z ?^

0

-.«-.-- GFR -m— Phenix

-5 -10 -15 -20 1.E •07

1.E-05

1.E-03

1.E-01

1.E+01

1.E+03

Energy (Mev)

Figure 4. Sensitivity of the k-effective to JEF2.2 elastic scattering.

51

A preliminary study was performed in order to find out if these specificities imply new sources of uncertainties which would impact the preliminary conclusions of the GFR studies. Uncertainties due to nuclear data have been quantified for the reactivity and different reaction rate ratios in the core, representative of various energy domains [3]. Two different nuclear data libraries were used: data based on the JEF2.2 [6] evaluation and ERALIB1 [7] adjusted data. The ERALIB1 data set is based on a formal adjustment procedure of JEF2.2 cross-sections which takes into account a very wide range of integral data. In the available data base, different types of integral data are considered, such as critical masses, bucklings, spectral indices, response function data for neutron transmission... In total, 355 integral parameters from 71 different systems (thermal, epithermal, fast) have been used. It is important to note that this library was not specifically produced for the GFR concept studies. So, even if the integral data base was large, it will be necessary to confirm the validity of this library for this type of reactors. This is one of the goal of an experimental program proposed in MASURCA facility, called ENIGMA (Experimental Neutronic Investigation of Gas-Cooled Fast Reactor Configurations in MASURCA). The ERALIB 1 uncertainties calculated here are only trends which will be validated. The JEF2.2 variance-covariance matrix was generated with data based on evaluations (for U238 and Pu239), data extracted from the literature, and expert judgement. The ERALIB 1 variance-covariance matrix was generated by the adjustment process for the 17 main isotopes and especially for U235, U238, Pu239, Pu240, Pu241, Pu242 and C, which are present in the GFR concept studied. Table 2 shows the uncertainties calculated for the reactivity and the following fission rate ratios : Pu239/U235 and U238 /U235. These ratios are calculated at the centre of the core and are respectively representative of intermediate and fast neutrons. They also concern important isotopes for the studied GFR concept. Table 2. Uncertainties due to nuclear data. Reactivity

Pu239 Fiss /U235 Fiss

U238 Fiss /U235 Fiss

JEF2.2 Uncertainty

1390 pcm

2.4%

3.3%

ERALIB 1 Uncertainty

312 pcm

0.5%

1.2%

The important JEF2.2 uncertainty in reactivity is essentially due to the Pu239 fission for which the sensitivity coefficient is very important. The

52

contribution of this reaction represents 72% of the total uncertainty. With the adjusted data, the contribution of this reaction is decreased and the total uncertainty is better distributed over the different elements. It can be noted that the total uncertainty in reactivity with the adjusted data is slightly greater than the typical uncertainty for the reactivity of standard sodium fast reactors, but remains a low value. Even if the validity of ERALIB1 data must be verified for GFR applications, this level of uncertainty is considered sufficient at the current stage of the GFR studies. The important JEF2.2 uncertainty in the Pu239-over-U235 fission ratio is essentially due to the Pu239 fission which represents the direct effect. This reaction represents 92% of the total uncertainty. With the adjusted data, this direct effect remains the main contribution, but the reduction of the variancecovariance data for this isotope implies an important decrease of the total uncertainty. For the U238-over-U235 fission ratio, the direct effect (U238 fission) represents only 42% of the total uncertainty. Other important contributions are U238 inelastic scattering, Si elastic scattering and Si inelastic scattering. With the adjusted data, the U238 contributions are reduced, but the Si contributions remain important because this isotope was not adjusted. Finally, we can say that these uncertainties are consistent with standard sodium-cooled oxide fast reactors ones and not prohibitive for preliminary GFR studies. The ERALIB 1 uncertainties are consistent with the uncertainty levels suitable for viability studies (-800 pcm for reactivity and 5% for reaction rates) and performance studies (-300 pcm for reactivity and 3% for reaction rates). 4. The breeding gain parameter For achieving proliferation resistance and fuel cycle competitiveness, the search for a core achieving a fissile self-sufficiency without blanket materials is performed. This neutronic constraint is fulfilled by keeping the breeding gain (defined as the change of isotopic concentration expressed in terms of equivalent Pu239 concentration) close to zero. Two fuel situations were studied: • the first cycle, with a (U,Pu)C initial load, • the equilibrium state with the TRU elements recycled and an external feed of natural uranium, leading to a mass proportion of minor actinides in the heavy nuclei around 1-2%. The creation of minor actinides during the irradiation of the fuel increases the breeding gain, leading to a more favourable situation.

53

In a first deployment stage, the Generation IV systems should be able to accept the minor actinides produced by the previous Generation III nuclear plants (water-cooled reactors for example). In order to verify this requirement, the impact of the introduction of different amounts of minor actinides was quantified on the reactivity swing and the reactivity coefficients, considering an estimated isotopic vector available in France around 2035 (76% Am, 17% Np, 7% Cm). It was shown that it was possible to introduce ~3 wt.% of minor actinides in the total heavy nuclei of the core without any modifications of the core design and management, or ~5 wt.% with a decrease of the burn-up in order to limit the helium released in the fuel. The latter case with a 5 wt.% amount of minor actinides will have consequences in terms of nuclear data. In Figure 5, we present the decomposition of the breeding gain in the different contributing heavy nuclei for these various strategies. It appears that this parameter is the sum of negative and positive terms and that important discrepancies (or uncertainties) in these terms could have an important impact on this key parameter. We can note also the substantial influence of the americium isotopes on the breeding gain for a minor actinide loaded fuel. Contributions to the breeding gain

I

•or

4L

w-$s<J^ys&t* 'XMAA^ m BG (1st cycle) H BG (equilibrium) o BG (5%MA)

Figure 5. Contribution of the different elements to the breeding gain.

5. The decay heat removal evaluation The reactor decay heat removal (DHR) is an important issue for gas-cooled fast reactors. The use of the gas coolant circulation as the main mechanism to remove the decay heat (by forced or natural convection) has been clearly identified and the DHR systems and strategies have a significant impact on the overall reactor architecture.

54

The design of these systems needs the precise calculation of the heat source term (target accuracy -10%). This source term has been calculated for various fuel loadings (Figure 6) corresponding to two extreme cases: • a (U,15%Pu,0%MA) fuel irradiated with an average discharge burn up (FIMA) of 10%, • a (U,17%Pu,5%MA) fuel irradiated with a lower average discharge burn up of 5% in order to limit the He release in the fuel.

Figure 6. Decay Heat as a function of time.

The 10%FIMA equilibrium fuel cycle with die TRU elements recycled and an external feed of natural uranium, leading to a ~2 wt.% proportion of minor actinides in the core, is an intermediate case for this aspect. Two periods of time are considered: immediately after the irradiation time for the calculations of accidental transients, for example, and after 24 hours of cooling for the subassembly handling considerations, in particular. 5.1. The 0% minor actinide fuel case At the end of the irradiation time, the fission products represent 92% of the decay heat (Figure 7), the complement being the heavy nuclei. The total decay heat is decomposed in the following terms: 1% for the Alpha emission, 52% for the Beta emission and 47% for the Gamma emission. The Cm242 causes 80% of the Alpha emission (10% for Pu238) and there is no predominant isotope for Beta and Gamma emissions. After a period of 24 hours of decay, the fission products represent 69% of the total source (Figure 8). The total decay heat is the sum of the following terms: 7% for the Alpha emission, 44% for the Beta emission and 49% for the Gamma emission. The Cm242 causes 80% of the Alpha emission (10% for

55

Pu238). Np239 causes 32% and 20% of the Beta and Gamma emission respectively. Lal40 and 1132 are the predominant fission products. Contribution of the different elements to the total decay heat (end of irradiation) (normalized to the 0%MA case)

• 0%MA_151%PU CYC1

10%BU

'•!

Figure 7. Contribution of the different elements to the total decay heat (end of irradiation).

Contribution of the different elements to the total decay heat (after 24 hours) (normalized to the 0%MA case)

i

rjO%MA_l5 1%Pu_CYC1. 10%BU_24h

mmfHiUm ! » S S ! f ', 1 " S

ml

0 5%MA_17%Pu_CYC 1 _5%BU_24h

pn

m n, rn

mfll ~

I n

JI

n1

s j ; -' i i ; "• ! • i ! s E s i i J i s ; 5

Figure 8. Contribution of the different elements to the total decay heat (after 24 hours).

The decay heat is very sensitive to the fission products and that the fission yields and decay data are important parameters for the precise calculation of the decay heat. Even if various fission products are more important than others (Nb, Y, La, I, Cs, Tc), many different fission products contribute significantly to the total decay heat and it is not easy to identify priority requirements for nuclear data. Concerning the heavy nuclei, Cm242, Np239 and Pu238 largely dominate the others.

56

5.2. The 5% minor actinidefuel case In this case, at the end of the irradiation time, the fission products represent 87% of the decay heat (Figure 7) and the decomposition of the total decay heat is significantly different: 1% for the Alpha emission, 49% for the Beta emission and 44% for the Gamma emission. The Cm242 causes 89% of the Alpha emission (6% for Cm244, 3% for Pu238) and there is no predominant isotope for Beta and Gamma emissions. After a period of 24 hours of decay, the fission products represent 41% of the total source (Figure 8). The total decay heat is made of the following terms: 44% for the Alpha emission, 26% for the Beta emission and 30% for the Gamma emission. The Cm242 causes 89% of the Alpha emission (6% for Cm244, 3% for Pu238). Np239 causes 30% and 18% of the Beta and Gamma emission respectively. Lal40 and 1132 are the most important fission products. With an increased amount of minor actinides in the loaded fuel, the contribution of the Cm242 (and Cm244) is significantly more important. 6. Conclusion The GFR initial studies underline two different kinds of nuclear data needs: • improvements in neutron cross sections of heavy nuclides. Such needs are common to various reactor systems, • specific needs, in particular those arising from structural materials (Si, C and Zr) which increase the importance of the neutron scattering processes, and the sensitivity of k-effective to these reactions. First preliminary studies have shown that JEF2.2 uncertainties are not sufficient to achieve uncertainty levels suitable for GFR core design (-800 pcm for the reactivity at the viability stage, for example). Nevertheless, the use of the ERALIBl adjusted nuclear data set is consistent with these levels for the reactivity, the Pu239-over-U235 fission rate ratio and the U238-over-U235 fission rate ratio at the centre of the core. The breeding gain, which is a key parameter for the self-breeding constraint, results from the sum of production and disappearance of various heavy nuclei, in particular in the situation of all the isotopes are multi-recycled or when significant amounts of minor actinides are introduced in the core. Large discrepancies (or uncertainties) in the cross sections of these isotopes can lead to significant modifications of this parameter. The decay heat removal is an important issue for gas-cooled fast reactors and the correct prediction of the source term, at different time step is required

57

with a typical target accuracy of 10%. Precise knowledge of fission yields and decay data for numerous fission products and key heavy nuclei (Cm242 in particular), associated with uncertainties, is therefore important. Finally, it is important to note that the GFR studies are still at a preliminary "viability" stage today, in which the impact of nuclear data uncertainties have not been systematically assessed. This will be done progressively as some design choices are firmed up and target performances specified, together with design margins. Future GFR neutronics studies at CEA will focus on two aspects: - detailed calculations of 3D effects (control rod worth, for example); - fine analysis of specific physical effects and parameters (breeding gain...) with sensitivities to nuclear data and requirements in terms of target accuracies associated. References 1.

2.

3.

4. 5. 6. 7.

J.C. Garnier, C. Poette, B. Mathieu, A. Conti and J.P. Gaillard, "Preliminary design of an advanced Gas cooled Fast Reactor core, fuel forms and primary system concept " ICAPP'03, Cordoba, Spain, May 4-7 (2003). G. Rimpault, J.C. Bosq, H. B. Choi and J.C. Garnier, "A Feasibility Study on a 600 MWth Gas-Cooled Fast Reactor Core", GLOBAL'03, New Orleans, U.S.A., November 16-20 (2003). J.C. Bosq, A. Conti, G. Rimpault and J.C. Garnier, "Methodology for a Large Gas-Cooled Fast Reactor Core Design and Associated Neutronic Uncertainties", PHYSOR'04, Chicago, U.S.A., April 25-29 (2004). G. Rimpault, "The ERANOS Code and Data System for Fast Reactor Neutronic Analyses", PHYSOR2002, Seoul, Korea, October 7-10 (2002). J.P. Gaillard, G. Mignot and A. Conti, "Thermal-Hydraulic Design of a Gas-Cooled Fast Reactor", ICAPP'03, Cordoba, Spain, May 4-7 (2003) C. Nordborg, "Distribution of JEF2.2", JEF/DOC-371-OECD/NEA Data Bank, February (1992) E. Fort, W. Assal, G. Rimpault, J. Rowlands, P. Smith and R. Soule, "Realization and Performance of the Adjusted Nuclear Data Library ERALIB1 for Calculating Fast Reactor Neutronics", PHYSOR'96, Mito, Japan, September

COMPARATIVE STUDY ON DIFFERENT PHONON FREQUENCY SPECTRA OF GRAPHITE IN GCR YOUNG-SIK CHO, KANG-SEOG KIM, DO HEON KIM, YOUNG-OUK LEE AND JONGHWA CHANG Korea Atomic Energy Research Institute, P .0. Box 105, Yuseong, Daejeon,

Korea

A GCR (Gas Cooled Reactor) employs graphite as a neutron moderator and a reflector. At thermal energies, the scattering of the neutrons is affected by the binding characteristics of the scattering nucleus in the moderator. Thus, these effects should be carefully described by well defined scattering laws. The calculations for the scattering laws require an exact shape of the phonon frequency distribution of a material as an input parameter, as well as its lattice structure. Currently several variations of the phonon frequency spectra are available. We have generated different sets of temperature dependent scattering laws for graphite with the module LEAPR of the NJOY using the available phonon frequency spectra. The temperature range of the generated data sets was from 300 to 2000 °K. To observe the effect of these different scattering laws on the criticality of a GCR core, MCNP calculations were carried out and their results were compared with each other. As the basis of a comparison, the keff and the temperature coefficients for the moderator and reflector were used.

1.

Introduction

A GCR (Gas Cooled Reactor) employs graphite as a neutron moderator and a reflector. At thermal energies, the scattering of the neutrons is affected by the binding characteristics of the scattering nucleus in the moderator. Thus, these effects should be carefully described by well defined scattering laws. The calculations for the scattering laws require an exact shape of the phonon frequency distribution of a material as an input parameter, as well as its lattice structure. Currently several variations of the phonon frequency spectra are available. We have generated different sets of temperature dependent scattering laws for graphite with the module LEAPR of the NJOY using the available phonon frequency spectra. To establish the effects of these different scattering laws on the criticality of a GCR core, MCNP calculations were carried out and their results were compared with each other. As the basis of a comparison, the k<.ff and the temperature coefficients for the moderator and reflector were used.

58

59

2. Generation of Cross Section Data There are three available variations of the phonon frequency spectra for graphite. The first one was computed by Young and Koppel (YK) using the bond-bending and bond-stretching model (BBS) in 1965 [1]. The scattering law data for graphite of the current ENDF/B evaluations are based on this YK spectrum. In 1972, Nicklow et al. published a new phonon frequency distribution [2]. They used the axially symmetric model (AS) for the calculation of the spectrum. The last one was computed by Hawari et al. based on the central force dynamical theory [3]. Figure 1 shows the comparison of the phonon frequency spectra. ;

: Young & Kopel@GA Nicklow et al.@ORNL Hawari et al. @ NCSU

\

11 I 1

\Jt

0.00

0.05

.• '•-..... / v „ - - k

0.10

1 0.15

1

L

r 0.20

Energy (eV) Figure 1. The comparison of the phonon frequency spectra

300 2000"K 1200'K Young&Koppel@GA Nicklow etal.@ORNL Hawari etal.@NCSU

Energy (eV) Figure 2. Elastic cross sections for graphite.

60

10"

10'*

10"'

Energy (eV)

Figure 3. Inelastic cross section for graphite.

For the calculations of the scattering law data, the module LEAPR of the NJOY [4] was used. It calculates both the elastic and inelastic scattering effects at the thermal energy region. The scattering law data and cross sections for graphite were prepared at temperatures from 300 to 2000 oK at 100 oK intervals. Figures 2 and 3 show the calculated scattering cross sections for the three different phonon spectra compared at three temperatures. The cross sections for the YK spectrum are largest in the case of an elastic scattering, while they are smallest in the case of an inelastic scattering. For the MCNP [5] calculations of keff, the thermal MCNP library was prepared using the module ACER of NJOY99.90. 3. Calculations and Results To investigate the effect of the different phonon frequency spectra, we have calculated keff at temperatures from 300 to 2000 °K at 100 °K intervals and the

Figure 4. MCNPX model of the GCR core

Figure 5. MCNPX model of the fuel assembly of GCR core

61

temperature coefficients for the moderator and reflector of a gas-cooled reactor and compared those results. The MCNPX model of a gas-cooled reactor is displayed in figures 4 and 5. It was modeled based on the conceptual design of VHTR developed by Idaho national laboratory. Table 1. Effect of moderator with different phonon spectra on the calculations of GCR core, ktff. T(°K)

Young & Koppel @GA

Nicklow et al. @ORNL

Hawari et al. @NCSU

300 400 500 600 700 800 900

1.42599(33) 1.42278(33) 1.41856(33) 1.41488(32) 1.41084(33) 1.40652(33) 1.40257(32) 1.40016(34) 1.39573(33) 1.39159(32) 1.38851(33) 1.38598(33) 1.38294(33) 1.37951(33) 1.37690(33) 1.37462(33) 1.37204(32) 1.36989(33)

-67(33) -45(32) +55(33) +45(33) +44(33) +42(32) +73(33) -103(33) +30(34) +82(31) +87(33) +8(33) +9(33) +46(34) +68(33) +35(33) -31(33) -38(33)

-15(32) -50(32) +86(33) +41(33) -59(33) +5(32) +139(33) -71(33) -8(32) +71(32) +66(33) +5(33) +25(34) +76(32) -17(33) -3(33) +33(33) -64(33)

1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

Table 2. Effect of central reflector with different phonon spectra on the calculations of GCR core, keg. T(°K)

300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

Young & Koppel @GA Nicklow et al. @ORNL 1.42579(32) 1.42980(32) 1.43298(33) 1.43575(32) 1.43742(32) 1.43871(33) 1.44063(32) 1.44260(32) 1.44327(32) 1.44391(31) 1.44591(32) 1.44682(33) 1.44751(32) 1.44728(32) 1.44877(33) 1.44990(34) 1.45026(32) 1.45078(32)

-37(33) -30(33) -25(32) -98(32) -23(33) +13(32) -53(32) +23(32) -23(32) +56(32) -97(33) -17(31) -17(32) +23(33) +6(33) -32(32) -97(33) -96(32)

Hawari et al. @NCSU +22(33) -40(33) -168(32) -96(32) -38(33) -17(32) +3(32) -75(33) -14(32) +83(32) -21(33) -20(31) -53(31) +123(32) -54(32) -32(32) -80(31) -65(31)

62 Table 3. Effect of outer reflector with different phonon spectra on the calculations of GCR core, k^. T(°K)

Young & Koppel @GA

Nicklow et al. @ORNL

300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

1.42608(33) 1.42958(33) 1.43190(33) 1.43420(33) 1.43477(32) 1.43605(32) 1.43744(32) 1.43822(33) 1.43915(32) 1.43941(34) 1.44010(33) 1.44061(32) 1.44184(32) 1.44231(33) 1.44290(32) 1.44272(33) 1.44289(32) 1.44363(33)

-23 (32) -52(31) -58 (33) -52 (33) +25 (33) +72 (32) -42 (33) +96 (33) -22 (32) +51 (33) +84 (32) -41 (33) -50 (33) -54 (32) -57 (31) -32 (32) +11 (32) -77 (32)

Hawaii et al. @ -30 (33) -42 (32) -57 (32) -134 (32) +44 (32) +23 (32) +34 (33) +66 (32) -47 (32) +51 (33) +48 (32) +79 (32) -37 (32) -40 (32) -44 (33) +23 (32) -55 (33) -62 (32)

Tables 1, 2 and 3 show the calculated keff when using the different phonon frequency spectra for a moderator, central and outer reflector, respectively. The numbers in the 3rd and 4th columns represent the differences in pcm and errors in parenthesis correspond to last digits of standard deviations. The calculated values agree well with each other to within 3 a except for some values written in bold. Figures 6, 7 and 8 show the calculated temperature coefficients of the GCR for the three scattering law data. The results show the tendency that the moderator temperature coefficient increases and the reflector temperature coefficient decreases, when the temperature increases. 4. Conclusion To investigate the effect of the different phonon frequency spectra, the keff and temperature coefficients for the moderator and reflector of a gas-cooled reactor were calculated and the results were compared with each other. The results show that different phonon spectra do not make any remarkable differences to the calculations of keff of GCR core, even though there were significant differences between the cross sections for the different phonon spectra. The calculated temperature coefficients showed the tendency that the moderator temperature coefficient increases and the reflector temperature coefficient decreases, when the temperature increases.

63 3.0x10 -

2.0x10"°-

••--

Young&Koppel @ GA Nicklow et al. ©ORNL Hawari etal.@NCSU

0.0.

-1.0X10"°-

^-., -2.0x10""-

/y^^^31^^^

-3.0x10"* •

Moderator temperature (K)

Figure 6. Moderator temperature coefficient in the GCR

1000

1500

Central reflector temperature (K)

Figure 7. Central reflector temperature coefficient in the GCR

References 1. Young et al., Photon Spectrum of Graphite, J. Chem. Phys. 42, p. 357, 1965. 2. Nicklow et al., Lattice Dynamics of Pyrolytic Graphite, Phys. Rev. B 5, p. 4951, 1972. 3. Hawari et al., Ab Initio Generation of Thermal Neutron Scattering Cross Sections, PHYSOR 2004, April 25-29, 2004, Chicago, Illinois. 4. R. E. MacFarlane and D. W. Muir, The NJOY Data Processing System Version 91, LA-12740-M, 1994. 5. H. G. Hughes et al., MCNPX for neutron- proton transport, in Int. Conf. on Mathematics and Computation, Reactor Physics and Environmental Analysis in Nuclear Applications, Madrid, Spain, September 1999, American Nuclear Society ,1999.

INNOVATIVE FUEL TYPES FOR MINOR ACTINIDES TRANSMUTATION D.HAAS, A.FERNANDEZ, J.SOMERS European Commission, Joint Research Centre Insitutefor Transuranium Elements P.O.Box2340, D-76l25,Karlsruhe,Germany

Transmutation of long-lived radio-nuclides is an option for reducing the hazards linked to the back-end of the nuclear fuel cycle. Previous studies have demonstrated that the main contributor to spent fuel radio-toxicity is by far Pu, followed by Am and Cm. Prerequisite for any efficient transmutation strategy is therefore Pu multiple recycling, whereas Am and Cm could be treated in different ways, including multiple recycling or once-through burning in dedicated targets. In all cases, however, the transmutation efficiency must be maximised, a condition best achieved if, firstly, uranium-free fuels are considered, and secondly, if multiple reprocessing and recycling is considered. In Europe, and in particular at the Institute for Transuranium Elements (ITU), extensive experimental work is being performed to develop fabrication processes for these innovative compounds, and to characterise their properties under irradiation. This work is mostly done within European collaborations, and is partially funded under the European Framework Programmes.

1. Introduction Within the framework of European national programs and international collaborations, the emphasis on Partitioning & Transmutation is nowadays placed on the use of fast neutron reactors that would lead to the best efficiency for transmutation of long-lived TRU's. Among those, the ADS (Accelerator Driven System) is now intensively studied, mainly within the 5th and 6th FWP of the EC. Fast Reactor systems (Liquid-metal or gas-cooled), including the ability to recycle the TRU's, are also considered within the Generation 4 Forum. The reference fuels for these reactors will depend on their type and on their main design parameters. Therefore, the choice of materials is broad and difficult to fix. In this paper, we will describe the type of fuels that are presently viewed as favorite candidates for the ADS reactor. As was recommended by the Fuel Subgroup of the ADS Technical Working Group these are U-free oxide fuels, with, as possible back-up, the U-free nitrides. Fuels for the Generation 4 fast

64

65

reactor systems will not be considered here. In some of the cases, however, developments made for ADS fuels could also be applied for Gen 4 systems. In terms of fuel fabrication and characterization, the two major laboratories in Europe are ITU (Karlsruhe) and CEA (Marcoule). At ITU, various fuel concepts for the transmutation of transuranic elements under pellet form are being studied. Given the high radiotoxicity of these elements and in particular the minor actinides, much emphasis is given to the development of clean and, possibly, dust-free fabrication methods for fuels containing americium and curium, so that contamination of the surfaces of the sealed containments and the equipment therein is minimized. At present, the main efforts are put on the development of uranium-free oxides. The fuel compositions that are being or will be tested in the near future are: • homogeneous oxide (ITU): a combination of sol gel and infiltration methods is used to produce the free-flowing and dust-free (Zr,Y,An)C>2 powder (An stands for any combination of transuranium elements), which can be pressed directly into pellets. This material is best suited to once-through transmutation strategies, as the spent pellets exhibit excellent resistance to corrosion during extended geological disposal; however, it could be used for multiple reprocessing if pyro-chemistry methods are successfully applied; • the zirconia-based actinide carrier can be mixed with other compounds (ITU), like metals (typically Mo) or ceramics (like MgO), to produce CERMET and CERCER composite materials, in order to improve the thermal properties of the fuel, thus allowing higher actinide loading, or lower operating temperatures. These fuel forms are considered for multiple recycling, in combination with pyro-metallurgical processes. • similarly, CERMET or CERCER can be done by mixing Mo (ITU) or MgO (CEA) respectively, with (An)02, and the final product might be reprocessable by aqueous methods (still to be determined). The latter compounds are now considered the most promising reference fuels for the ADS reactor. The An(02) phase is obtained by various methods: coprecipitation (CEA), infiltration of Pu02 or sol-gel (ITU). 2. Homogeneous fuels and targets One candidate is the so-called inert-matrix mixed oxide, in which the actinides are incorporated in a partially yttria-stabilised zirconia (YSZ) solid solution. At the ITU a hybrid process consisting of a combination of sol gel and porous bead infiltration techniques has been developed for the fabrication of these types of

66

materials on a laboratory scale [1] but with industrialisation being the long term goal. It meets three critical criteria: (a) it is dust free, (b) there are no liquid wastes, and (c) it minimises the steps in which highly radioactive materials are handled. The infiltration process itself is well known [2, 3] but until now has only been used with porous disks or pellets [4, 5] to produce actinide compacts directly. In contrast, the process described here has several additional advantages, the most important being the higher actinide content achievable and the homogeneity of the final products. This new process has been tested and validated with cerium and plutonium at the laboratory scale. The actual fabrication of Am- and eventually Cm - containing materials is being performed in a set of specially constructed shielded cells installed in the ITU Nuclear Fuel Laboratories [6]. An a-autoradiograph and a ceramograph of a (Pu,Y,Zr)02.x pellet produced by this process are shown in Figure 1.

t ^

•1mm Figure 1: a-autoradiograph and ceramograph of an axial section of a (Puo.o4sYo.i63Zro.792)02 pellet.

3. Heterogeneous fuels A second candidate is the composite dispersion-type fuel or target in which the actinide oxides are dispersed in a ceramic (CERCER) or metallic (CERMET) uranium-free matrix. The fabrication of composite pellets is considerably more difficult than for solid solution oxides pellets. This is a result of the specific requirements of particle size and homogeneous distribution of the dispersed actinide phase. The major problem to be solved is the control of the distribution of the actinide ceramic phase in the matrix, and major efforts have been made to develop procedures to fabricate both CERCER and CERMET pellets. Particles containing the actinide phase are prepared by a combination of the sol gel, and infiltration methods. Thereafter they are mixed with the matrix powder by

67

conventional blending methods. Following compaction into pellets by cold pressing, they are sintered at high temperature [7, 8]. CERCER Pufl.2oY0.i4Zr0.6(i02.x -MgO microdispersed 40-60um

macrodispersed _10Q-.125ym.____

CERMET: P u 0.24lYo.l28Zro.63i02. x

-SSL microdispersed 60-80um

Figure 2 : Microstructure of the different cercer and cermet composite

Both CERCER pellets, in which the actinides are in (near) spherical yttriastabilised zirconia (YSZ) particles, dispersed in ceramic-ceramic a MgO matrix, and CERMET pellets, using Stainless Steel as matrices have been successfully fabricated at the ITU. MgO-(Pu,Y,Zr)02.x CERCER pellets with a Pu density of 0.7 g-cm"3 have been fabricated without cracks and with a random distribution of isolated spheres, for the sphere sizes investigated (Figure 2). A stainless steel based CERMET fuel with a Pu density of 0.9 gem"3 has also been fabricated and characterised for an irradiation experiment (SMART) in the HFR-Petten reactor. The final pellets have a density of 87.2 ± 0.2% TD. As the porosity is principally in the ceramic phase, accommodation of fission gases during irradiation is achieved and therefore swelling should be mitigated. The optical micrograph shown in Figure 2 indicates that all particles are uniformly distributed in the matrix. The main advantage of this fabrication method is its high flexibility to select the particle size and volume fraction of the ceramic phase. In the near future, for the transmutation fuels dedicated to multi-recycling in ADS systems, the actinide phase, namely (Pu,Am,Y,Zr)02-x will be replaced by (Pu,Am)02-x which allows incorporation in the fuel of an increased content of transuranic elements, and potentially a simplified reprocessing ability. In this case the actinide compound will be prepared by two methods: either by the

68

infiltration of Am nitrate in Pu0 2 beads, or by direct preparation of by sol-gel. This material will be characterized within the FP5 Project FUTURE. 4. Fuel irradiation experiments The main irradiation programme foreseen at ITU, in close international collaboration, will consist in the performance of two new experiments, called FUTURIX-FTA in PHENIX [9], and HELIOS in HFR [10]. The goals are to fabricate and characterize the properties of homogeneous and composite oxide fuels in representative fast flux conditions for FUTURIX- FTA, and to study specific irradiation properties (in particular, the Helium formation and release mechanisms) in similar compounds, with in-pile instrumentation allowing to measure on-line the fuel center temperature. These results will be compared with those on nitride and metallic fuels investigated by the US national laboratories. To that purpose, a large Integrated Project, called EUROTRANS/AFTRA, has been set up and has been proposed for financing by the European Commission under the 6th Framework Programme. 5. Conclusions The results obtained have demonstrated the feasibility of the fabrication of (Pu,Y,Zr)02.x based magnesia and stainless steel composite pellets, by a combination of GSP, infiltration and conventional blending techniques. The main advantage of this fabrication method is its high flexibility to select the particle size and volume fraction of the ceramic phase. Materials exhibiting a homogeneous dispersion of the actinide ceramic bearing phase have been produced. The actual fabrication of Am containing materials is now being performed in the minor actinide laboratory. References 1. A. Fernandez, D. Haas, R.J.M. Konings, J. Somers. J. Am.Ceram.Soc, 85 [3] (2002)694-96 2. B.R.Marple, DJ. Green, J. Mater. Sci. 28 (1993) 4637. 3. P.H.Colvin, F.F.Lange, J. Am. Ceram. Soc. 79 (1996) 1810. 4. K.Richter, A.Fernandez, J.Somers, J.Nucl.Mater. 249 (1997) 121. 5. Kinoshita, K.Kuramoto, M.Uno, T.Yanagi, S.Yamanaka, H.Mitamura, T. Banba. J. Am. Ceram. Soc. 83 (2000) 391. 6. Institute for Transuranium Elements Annual Report 2000, EUR 19812, 2001, p l l 5 .

69

7. A. Fernandez, D. Haas, R.J.M. Konings, J. Somers. J.Mater Sci. Lett. 22 (2003) 119-121 8. A. Fernandez, R.J.M. Konings, J. Somers. J.Nucl. Mater. 319 (2003) 44-50 9. S. Pillon, N. Schmidt, E. Abonneau, Y. Croixmarie, L. Donnet, S. Hayes, M. Meyer, K. Chidester, B. Margevicius, A. Fernandez, D. Haas. GLOBAL 2003, 16-20 November 2003, New Orleans, USA 10. D. Warin, L. Brunei, G. Vambenepe, P. Raison, D. Haas, R.J.M. Konings, J. Knebel, R.P.C. Schram. GLOBAL 2003, 16-20 November 2003, New Orleans, USA

THE IMPORTANCE OF NUCLEAR DATA IN MODELING AND DESIGNING GENERATION IV FAST REACTORS KEVAN D. WEAVER Idaho National Laboratory, P.O. Box 1625 Idaho Falls, Idaho 83415-3850, USA

Three of the six Generation IV nuclear energy systems are fast spectrum reactors, and all intend to utilize high concentrations of minor actinides in their fuel either in initial core loads, or as part of the subsequent loadings during recycle. The fuel for each system varies, as do the selected materials for in-core use. These advanced fuels and materials will require nuclear data for analyzing the behavior of the reactor during operation, but there are gaps in the data as it pertains to these designs. It appears that new, more complete nuclear data evaluations are needed to capture the details required in analyzing these systems.

1.

Summary

W i t h i n the G e n e r a t i o n I V International F o r u m ( G I F ) , six c o n c e p t s w e r e c h o s e n

for further development, where three of the concepts utilize a fast neutron spectrum. These are: the Gas-Cooled Fast Reactor (GFR), the Lead-Cooled Fast Reactor (LFR), and the Sodium-Cooled Fast Reactor (SFR), where the key attribute for these fast reactors is sustainability with regard to resource utilization and waste minimization. While each of these reactor concepts use different fuel types and forms, e.g., carbide or nitride fuel in an inert ceramic matrix (or ceramic cladding) for the GFR; oxide, nitride, and possibly metal fuel in a metal (or possibly ceramic) clad for the LFR; and oxide, nitride, and metal fuel in metal cladding for the SFR, each will use the full compliment of recycled actinides in a self-sustaining fuel cycle. The use of recycled fuel is not unique to fast reactor development, and is, in fact, an integral part of designing such reactors. Even in the early days of fast reactor design, a top priority was to obtain as high a breeding ratio as possible to extend the fissile resources, and thus necessitated fuel recycle. However, a paradigm shift has occurred: high breeding ratios are no longer a top priority based on current fissile resources, yet spent nuclear fuel has become a priority, where a fast spectrum is ideal for fissioning the minor actinides (Np, Am, Cm) to reduce the waste. This latter point is important due to the ever-increasing need 70

71

for repository space under the current once-through fuel cycle scenario, where the minor actinides play a major role in the long-term radiotoxicity of the spent fuel. The fast spectrum systems can be used to limit (or possibly eliminate) the build-up of these minor actinides, and thus can play an important role in the future sustainability of nuclear energy. However, the addition of large amounts of minor actinides in recycled fuel can greatly affect the physics of a system. For example, the delayed neutron fraction in standard uranium-plutonium fueled fast reactors is already half that found in thermal reactors (-300 pcm versus 650 pcm), which is an important factor in determining the controllability of a reactor. The addition of minor actinides can further reduce the delayed neutron fraction, where current studies are looking at possible limits to the amount of minor actinides that can be used in the fuel. Also affected are Doppler and void coefficients, which are important during accident scenarios, and the reactivity loss (or gain) during burnup that has been seen with high minor actinide content fuel. Nuclear data for the minor actinides is incomplete at the energy ranges of interest, and little nuclear data exists for many of the high temperature ceramics and intermetallics that are being considered for fuel matrices and in-core materials of the GFR and LFR. Recent results are based on codes that use the most current nuclear data, where the data is not necessarily complete, and would not be adequate for detailed conceptual designs of Generation IV fast spectrum systems using the proposed fuels and materials. New, more complete nuclear data evaluations are needed to capture the details required in analyzing Doppler coefficients, void coefficients, delayed neutron fractions, scattering and capture cross sections, etc., that are necessary in designing these Generation IV systems.

THE GIF AND MEXICO "EVERYTHING BEGINS WITH A WISH" CARLOS ARREDONDO-SANCHEZ* Instituto National de Investigaciones Nucleares. Carretera Mexico-Toluca S/N, La Marquesa Ocoyocac, Mexico, 52750 On February 16, 2005 Mexico signed the Kyoto protocol. This official act is very important and hopefully will trigger in the near future the initiation of a nuclear plan for generation of electricity in Mexico. The research related with the fourth-generation of nuclear reactors should be a key step of that plan. Actually the nuclear energy accounts near six percent of the total electricity production with just one plant with two units, Laguna Verde Nuclear Power Plant (LVNPP). At the National Institute of Nuclear Energy (INrN) organization responsible of the research in the nuclear field, there is a group of researches working in projects to assist the operation of the LVNPP and the analysis of new reactors. Therefore the Generation IV is a key point in research and development.

1. Electricity Status in Mexico (2002) 1.1. Introduction Actually Mexico had installed electric power generating capacity of 46.2 GW, on 2004 was generated around 205.39 TWh of electricity, of which thermal accounts for 73.80%, coal 8.52%, hydro 8.44%, nuclear 5.72%, geothermal 3.52% and wind 0.003%. The only nuclear plant, Laguna Verde Nuclear Power Plant (LVNPP) has two units, the first unit went into commercial operation in 1990 the second five year latter, since that time the contribution to the base load electricity generation had been very successful, even that in this moment there are not contemplated a new nuclear power plant for the next future, that must be changed. Participation in GIF could be one way to turn back the actual situation. 1.2. Electricity Generation Capacity At the end of March 2005 the Federal Electricity Commission (CFE) the electricity generation government company has a installed a gross capacity to generate electricity of 46,171.02 Megawatts (MW), from which: 7,750.90 MW

Also National Polytechnic Institute 72

73

are from independent producers (thermal), 10,258.98 MW are hydroelectric, 23,234.59 MW are thermals from CFE burning hydrocarbons; 2,600.00 MW to carbon; 959.50 MW to geothermal; 1,364.88 MW to the nucleoelectric and 2.18 MW to wind. CFE is owner of LVNPP and also has other 171 power plants, divided in 64 hydroelectric, 96 thermoelectric consuming fossil fuels, 7 geothermal, 2 carboelectric and 2 wind-electric. The increase in the electricity generation capacity (MW) and actual generation (TWh) trough the years 1997 to 2005 (31 march) in Megawatts can be seen in Table 1. Table 1. Increase in the electricity installed capacity and generation. 1997 CFE C

fMW1)y

1998

1999

2000

2001

2002

2003

2004

2005*

33,944 34,384 34,839 35,385 36,236 36,855 36,971 38,422 38,420 ffiPs

"

"

484

1>455

3 495

'

6 756

'

7 265

'

7 751

'

Total 33,944 34,384 34,839 35,869 37,691 40,350 43,727 45,687 46,171 CFE Ge

^ZT?a

EPS

159.83 168.98 179.07 190.00 190.88 177.05 169.32 159.53 -

-

-

1.21

4.04

21.83

31.62

38.02

45.86

11.02

Total 159.83 168.98 179.07 191.20 194.92 198.88 200.94 205.39

49.04

(TWn)

* Including 19 Plant from the Independent Energy Producers (IEPs).

1.3. Nuclear Energy Status In Mexico In Mexico there in only one nuclear plant, Laguna Verde Nuclear Power Plant (LVNPP). The nuclear plant is located near the Veracruz city in the Gulf of Mexico and has two BWR reactors of 682 Mw(e) gross each. The first unit went into commercial operation in July 1990 and the second in April 1995. With the two units in operation, nuclear energy accounts between five and six percent of the total electricity production. The principal characteristic of the nuclear reactor are listed in Table 2. It must be noted that since 1996 the nuclear generation had been maintained almost the same percentage, more than 5 %, to the gross electricity generated, that's due to the upgrade in power (105%) and to the upstanding performance of LVNPP.

74

Nuclear reactor Manufacturer Number of units

Table 2. Nuclear Reactor Characteristics of the LVNPP Units 1 and 2 General Electric two

Thermal power by reactor Nuclear reactor tip

2,021 MW BWR-5 H 2 0 in ebullition

Nuclear fuel Initial load of fuel by reactor

UO2 with 3% enrichment 444 assemblies; 92 fuel tons (U02) at 1.87% U235 average 96 assemblies at 2.71 % de U235 87.85 ton. by unit 109 by unit 71.7 Kg/cm2 3,989 tons. / Hr. 99.7 % 2 by unit 9,600 tons / Hr. 20 by unit 682.44 MWe 655.14 MWe 4,782 GWh, with capacity factor 80% 1 million 96 thousand cubic meters (6 millions 895 thousand barrels). Three of 400 KV to Tecali, Puebla and Poza Rica; two of 230 KV to Veracruz city

Annual refuel by reactor Total weight of uranium Control rods Nominal reactor pressure Vapor flux Vapor quality Recirculation pumps Recirculation flux Intern recirculation jet pumps Electric gross power by unit Electric net power by unit Annual energy generated by unit Annual savings in oil by unit Transmission lines

2. Changing Situation 2.1. ININ Participation on the Third Generation Reactors At the National Institute of Nuclear Energy (ININ) organization responsible of the research in the nuclear generation field, there is a group of researches working in projects linked with a reactor of the third generation. Two engineers are taking part since 2004 in the elaboration of the Preliminary Probabilistic Risk Assessment of the IRIS reactor, they expend six months in U.S. working in that, The PRS for API000 has been used as the basis in preparing the analysis for IRIS, the reason is due to the present design stage and the fact that IRIS project shares a number of design solutions and components, especially outside the reactor vessel, with the Westinghouse passive designs AP 600 and AP1000, so many details of IRIS will be obtained from the AP designs. 2.2. Kyoto Protocol and the Future On February 16 of this year the Mexican government signed the Kyoto Protocol, the inventory of GHG in 1998 shows a CO2 emission due to energy activities of 350.4 thousand of tons from which 29% are from electricity generation resulting in 101.3 tons. Even that Mexico is one of the sixteen countries with more

75

emissions of GHG is not an Annex 1 country so it does not have quantitative compromise. Yet it has some compromises, must formulate annual inventory of GHG and implement programs to mitigate de emissions, confront the challenge to diminish the intensity (emissions/consumed energy) of emissions in it's economy, take opportunities in reductions of emissions market. Hopefully these compromises will trigger a plan to generate electricity with nuclear reactors. Demand for electricity in the Mexico is expected to increase sharply in the 21st century. Even accounting for successful implementation of energy efficiency practices and technologies, forecasts show that to meet rising electricity demand and replace older facilities, mostly fossil fuel fired, that are going to be retired Mexico would need hundreds of new power plants of all types by 2025. In a study prospective the electricity sector for 2004-2013, made by the Energy Secretariat (SENER), even that there is not direct recommendation for the expansion of nuclear energy sector in order to satisfy this need, it is mentioned that 6,696 MW are free. It is here where the opportunity exists to introduce new nuclear power plants if their competitiveness is demonstrated. 2.3. Mexican Institutions with the Capacity to Get Involved on GIF At ININ there is a group of researches working in projects to assist the operation of the LVNPP. The fields of these projects are: Fuel Management, Risk Analysis, Safety Analysis, Design Changes, Technical Specifications, Materials, Electronics, Training and others. As an example of the work that is done at ININ in order to improve the performance of some aspects of nuclear fuel management, it had been provided to LVNPP, a nuclear reactor power BWR, specialized assistance for fuel management inside the core including. Static, dynamic and stability studies of BWR reactors including: criticality predictions, control bars patrons optimization, operational cycle following, multi-cycle analysis, reloading economical assessment, transient event analyses, and stability analysis of the core. Also it had been done a design and constitution of nuclear data banks related with the combustible cycle, such as: design of recharge fuel assemblies, design of recharge cores, design of control bars patrons and generation of nuclear data banks. Other example is the performance of the Internal Flood Analysis as part of the Individual Plant Examination for LVNPP and the Adaptation to the Regulatory Guide 1.174 "An approach for using probabilistic risk assessment in risk-informed decision on plant-specific changes to the licensing basis". There are also other institutions that are working in the nuclear field and in each of them some research had been done for CLV, between them it can

76

mentioned the National Autonomous University of Mexico (UNAM), the National Polytechnic Institute (IPN) and the National Institute of Electric Research. 2.4. Options for the Participation on GIF Taking in consideration the "Charter of the Generation IV International Forum" (CGIVIF), signed on July 2001 by Argentina, Brazil, Canada, France, Japan, Korea, United Kingdom, and the United States. First some of the subsections of the Charter are transcribe in order to sustain the proposals. Charter subsection 1.3.1: Subject to the unanimous consent of the Policy Group, membership in the GIF is open to other governmental entities which sign this Charter. Charter subsection 1.3.2: Members are expected to maintain an appropriate level of active participation in collaborative projects, such as the participation in at least one significant collaborative project. Charter subsection 1.3.3: Non-Members may participate in GIF activities only as noted in this Charter. Charter subsection 1.3.4: Technical and other experts from organizations other than those from countries represented by Members may participate in R&D projects conducted under the auspices of the GIF at the invitation of the Secretariat (see 1.6.6, below) with the approval of the Policy Group. Charter subsection 1.6.6: The principal coordinator of the GIF's communications and activities is the GIF Secretariat. The Secretariat (1) organizes the meetings of the GIF and its sub-groups, (2) arranges special activities such as teleconferences, (3) forwards new member petitions to the Policy Group, (4) coordinates communications with regard to GIF activities and their status, (5) acts as a central source of information for GIF, (6) maintains procedures for its key functions that are approved by the Policy Group, and (7) performs such other tasks as the Policy Group directs. In order to be involved on the work that it is done for the developing of the next generation nuclear energy systems (fourth generation reactors). Two preliminary options can be consider: First option: Get the approval of governmental authorities in order that Mexico officially ask for admission as a member in the GIF (see subsection 1.3.1 of the CGIVIF). Of course this will mean a compromise to maintain an appropriate level of active participation in collaborative projects, such as the participation in at least one significant collaborative (see subsection 1.3.2 of the CGIVIF).

77

Second option: Participation as a Non- Member, by participation of some experts in one or some R&D projects conducted under the auspices of the GIF, of course under the invitation of the Secretariat (see subsection 1.6.6 of the CGIVIF) and with the approval of the Policy Group (see subsection 1.3.4 of the CGIVIF). Of course as is stated above these are preliminary options and must be discussed in the Mexican and International institutions in order to be final to pursue all the necessary steps in order to meet all the requirements by the Mexico Government, the GIF and the institutions involved. 3. Conclusions Up to now the expansion program of CFE do not include nuclear power reactors, but the rising worries for the environment are favorable to energy sources that can satisfy the need for electricity on sustainable base with minimal environmental impact. Expectantly the signature of the Kyoto Protocol will activate a plan to generate electricity with nuclear power plants. In the middle future the compromises to reduce GHG should be an open door for new nuclear reactors. Two preliminary options to be involved in GIF are presented, these options must be evaluated and an official definition should be done by the corresponding authorities.

This page is intentionally left blank

BENCHMARKS, SENSITIVITY CALCULATIONS, UNCERTAINTIES

This page is intentionally left blank

SENSITIVITY OF ADVANCED REACTOR AND FUEL CYCLE PERFORMANCE PARAMETERS TO NUCLEAR DATA UNCERTAINTIES G. ALIBERTI, G. PALMIOTTI, M. SALVATORES, T. K. KIM, T. A. TAIWO Nuclear Engineering Division, Argonne National

Laboratory

I.KODELI, E. SARTORI NEA

Databank

J.C. BOSQ, J. TOMMAS1 DER/SPRC,

CEA-Cadarache

As a contribution to the feasibility assessment of Gen IV and AFCI relevant systems, a sensitivity and uncertainty study has been performed to evaluate the impact of neutron cross section uncertainty on the most significant integral parameters related to the core and fuel cycle. Results of an extensive analysis indicate only a limited number of relevant parameters and do not show any potential major problem due to nuclear data in the assessment of the systems considered. However, the results obtained depend on the uncertainty data used, and it is suggested to focus some future evaluation work on the production of consistent, as far as possible complete and user oriented covariance data.

1. Introduction The choice of the preferred systems for the future has been made, under the auspices of the Gen IV initiative, based on a set of high-level requirements: waste minimization, sustainability, safety, economy, and non-proliferation. At the same time, in the framework of the Advanced Fuel Cycle (AFC) program, several systems have been considered as possible transmuters of minor actinides. The physics of these reactors and their associated fuel cycles is rather well understood. However, their optimization, in order to comply more effectively with the requirements, and their timely deployment, requires focusing the research and development in all fields, in particular for innovative fuel development and processing, and also in the reactor physics field. In this last area, the role of nuclear data is quite significant. Most data are by and large available, but their accuracy and validation is still a major concern.

81

82

In order to make a comprehensive assessment, the tools of sensitivity and uncertainty analysis are needed. These tools have been widely developed in the past, in particular for the assessment in the '70s and '80s of the performance of fast reactors. Recently, we have performed a study on the impact of nuclear data uncertainties on the performance parameters (criticality, reactivity coefficients, irradiated fuel isotopic composition, external source effectiveness, etc.) of a generic Accelerator Driven System, dedicated to waste transmutation [1], which allowed to point out nuclear data uncertainties of relevance and to quantify the requirements for their reduction. We have launched a similar study, devoted to future systems, mainly the ones selected by Gen IV or considered in the AFCI program, but also for other systems like extended burnup LWRs. For this type of study, two major difficulties are encountered. First, it is needed to define at an early stage, representative, i.e., general enough, "images" of "future systems". Second, it is necessary to establish a realistic "compilation" of nuclear data uncertainties and their correlation (variance-covariance matrices). Regarding the first point, we have defined what we consider to the best of our knowledge, reference "images" of a GFR (with full recycling of minor actinides, MA); of a SFR (in a TRU burning configuration i.e., with a conversion ratio <1); of a large SFR (referred in the following as EFR) (with full recycling of MA and CR ~1); of a LFR (as defined recently for an IAEA benchmark); and of a VHTR with particle fuel. Finally, an extended burnup (100 GWd/t) PWR has been studied. As far as the second point, we started from an "educated" guess of uncertainties for all the isotopes of interest (actinides, structural and coolant materials), based, as much as possible, on the nuclear data performance in the analysis of selected, clean integral experiments (irradiated fuel and sample analysis, criticality and fission rates in zero-power critical facilities) [1,2]. For the correlations, we have used as a first guess a very crude hypothesis of partial energy correlations "by energy band" (PEC). We will call this set of uncertainties (diagonal values) and correlations "ANL Covariance Matrix" [3]. However, a substantial effort at the NEA Data Bank has been performed to extract relevant covariance data from current evaluations in major data files and to process them in the same multigroup structure used for sensitivity calculations [4]. In this manner we were able to point out the crucial role of well-established covariance data, and the urgent need to develop a consistent, user-oriented set of covariance data to be associated with the evaluated data files.

83

The analysis has been applied to a large number of integral parameters which characterize the reference systems indicated above, and their associated fuel cycle: criticality (Keff), Doppler and coolant void reactivity coefficients, reactivity swing during burnup, isotope concentrations in the spent fuel, power peak in the core, decay heat of the spent fuel in a repository, neutron source associated to the spent fuel e.g., at fuel fabrication, and dose (radiotoxicity) of the spent fuel or of the wastes in a repository, at selected times after storage. A successive study, based both on this uncertainty analysis and on the definition of target accuracies for the various integral parameters, will provide guidelines on priority requirements for data improvements. 2. 2.1.

The approach and theoretical background Uncertainties and target accuracies

The approach used for this work includes: • Sensitivity studies, using GPT (Generalized Perturbation Theory), on the selected integral parameters of representative models of systems. • Uncertainty assessment using covariance data. Once the sensitivity coefficient matrix for each integral parameter R and the covariance matrix D are available, the uncertainty on the integral parameter can be evaluated as follows: AR%=S+RDSR

with

SR=—-^L

dUj

R

A successive step is the assessment of target accuracy requirements. To establish priorities and target accuracies on data uncertainty reduction, a formal approach can be adopted by defining target accuracy on design parameter and finding out required accuracy on data. In fact, the unknown uncertainty data requirements J, can be obtained (e.g. for parameters i not correlated among themselves), by solving the following minimization problem: £ At I dt = min

i = 1.../

i

(I: total number of parameters) with the following constraints:

I

S2nidf<{RTnJ

84

(N: total number of integral design parameters) where Sni are the sensitivity coefficients for the integral parameter Rn and Rn are the target accuracies on the N integral parameters. lt are "cost" parameters related to each <7, and should give a relative figure of merit of the difficulty of improving that parameter (e.g., reducing uncertainties with an appropriate experiment). 2.2. Sensitivity coefficients Sensitivity coefficients are calculated using GPT methods. A few examples will be given here. A reactivity coefficient (like the Doppler effect) can be expressed as a variation of the reactivity of the unperturbed system (characterized by a value K of the multiplication factor, a Boltzmann operator M, a flux O and an adjoint flux <E>*): [

Kp)

{

K)

K

Kp

Where Kp corresponds to a variation of the Boltzmann operator such that:

M^>Mpl=M+SMp)

®^<&p(=® + S®p)

O*-»$*(= $*+Kp(=K + SKp) The sensitivity coefficients (at first order, and for the so-called "indirect effects") for Ap to variations of the Oj are given as:

where

'/=<®*.™>

and

'M*;.^,).

F being the neutron fission production part of the M{=F-A) operator. In the case of nuclide transmutation (i.e. nuclide densities at end of irradiation) the generic nuclide K transmutation during irradiation can be represented as the nuclide density variation between time to and tF. If we denote nf the "final" density for isotope K, the appropriate sensitivity coefficients are given by: K

dnp 0";

daj

nKF

1 'Fr

nf I

85

where the time dependent equations to obtain n* and n are the classical Bateman equation and its adjoint equation, with appropriate boundary conditions. Finally, in the case of the reactivity loss during irradiation, Apcycle, at first order we have: Apcycle

^

=

M

K

pfc

A

n* = «*

-

n

$

and pK is the reactivity per unit mass associated to the isotope K. The related sensitivity coefficients associated to the variation of a orj, are given by: sCycU_

°j Apcycle

tepcyde d a

.

A p

cycle

E

VK

dnK da

j

TH

dpK

da, . J J

or: cycle

*J

=z—^ k ^ K N>

*+U-(*;.^*,)-T-(**.^*)

3. The covariance data For the "home made" covariance Matrix (ANL, Ref. 3), we started by updating the covariance matrix used in the ADS study [1] by taking into account the results of clean integral experiment analysis, in particular irradiated sample/fuel analysis, which gave valuable information on capture and some (n,2n) crosssections, and fission rate measurements in critical assemblies [2]. The uncertainty values, have been given by "energy band", consistent with multigroup energy structures used for deterministic calculations both of thermal and fast reactors. Fifteen energy groups have been selected between 20 MeV and thermal energy. Two extra groups have been added between 150 MeV and 20 MeV for ADS applications. The uncertainty values are given only for neutron cross-section data of actinides and structural materials. Fission products related uncertainties have not been considered in this study. In the compilation of the covariance matrix there has been the attempt to account for the work done, in particular for major actinides, in the frame of the JEF project, in order to produce "adjusted" cross-sections. For minor actinides, we used the analysis and recommendations of a working group of the WPEC updated, once again, for selected isotopes and reactions, on the basis of integral experiment analysis. For most structural materials, we have often relied on the graphical intercomparison of different data files. The covariance matrix diagonal

86

values have been estimated on the performance of the most recent JEF files in the analysis of a large set of integral experiments in different spectra. However, it was observed that the performance obtained using ENDF/B files was not substantially different. Therefore, the covariance matrix can be applied to both files. At first, only "diagonal" values of the full covariance matrices have been used. Their use implies to neglect all type of correlation (in energy, between different isotopes, among reactions, etc.) and, consequently, to underestimate uncertainties. In a second step, we have introduced partial energy correlations. Our first guess has been to use the same correlations for all isotopes and reactions, under the form of full energy correlation in 5 energy bands. The idea was to single out: 1. the region above the threshold of fertile isotope fission cross-sections, and of many inelastic cross-sections, up to 20 MeV; 2. the region of the continuum down to the upper unresolved resonance energy limit; 3. the unresolved resonance energy region; 4. the resolved resonance region; 5. the thermal range. The correlations used are shown in Fig. 1.

1

2

3

4

5

6

7

8

9

10

II

12

n

14

IS

16

17

ENERGY GROUP NUMBER

Figure 1. Energy correlations used in the ANL Correlation Matrix (PEC)

As indicated in the introduction, available covariance data at the NEA DataBank have been used, to provide a comparison with the ANL Covariance Data. The comparison cannot be complete, but limited to selected isotopes and

87

reactions, since some data used in this analysis, are missing in the available evaluated covariance data. The data selected (NEA-K Covariance Matrix) are shown in Table 1. Ideally the covariance data should be taken from a single evaluation, i.e. the one used in the calculations, in order to assure the consistency among the nuclear data used. At present no evaluation is complete enough to cover all the materials needed in this study and the NEA-K covariance matrices are therefore a selection (more or less consistent) of the data available in the ENDF/B-V, -VI, IRDF-2002, JENDL-3.3 and JEFF-3 evaluations. Table 1. NEA-K Covariance Matrix

U235 U238 Np237 Pu239 Pu240 Pu241 Am241 C H O Cr52 Fe56 Na23 Pb206 Pb207 Pb208 Si Zr90 BIO Ni58

(nu, fission, inelastic, elastic, capture, nxn) (nu, inelastic, elastic, nxn) (fission, capture) (fission) (nu, inelastic, elastic, capture, nxn) (fission) (inelastic, elastic, fission, capture, nxn) (nu, inelastic, elastic, fission, capture, nxn) (fission) (elastic, capture) (elastic, capture) (inelastic, elastic) (inelastic, elastic, capture, nxn) (inelastic, elastic, capture, nxn) (inelastic, elastic, capture, nxn) (inelastic, elastic, capture, nxn) (inelastic, elastic, capture, nxn) (inelastic, elastic, capture, nxn) (inelastic, elastic) (inelastic, capture, nxn) (capture) (inelastic, elastic, capture, nxn)

JENDL3.3 JENDL3.3 IRDF-2002 IRDF-2002 JENDL3.3 IRDF-2002 JENDL3.3 JENDL3.3 BRDF-2002 ENDF/B-V ENDF/B-V ENDF/B-V ENDF/B-VI ENDF/B-VI ENDF/B-VI ENDF/B-VI ENDF/B-VI ENDF/B-VI ENDF/B-VI JENDL3.3 IRDF-2002 JEFF3

Some comparisons of the ANL and NEA-K covariance matrices for selected isotopes and reactions are given in Figs. 2 through 5.

MEA-K

1

ANL

1.E+1 1.E+3 Energy[eV]

ANL Correlation Matrix

NEA-K Correlation Matrix

5 o

.1

ffi 40 u •a •

r

o

r

o

eviation (%)

c

Figure 2. Pu-239(n,f)

ANL

30 20 10 0

;:::":::»

'

I

1.E+1 1.E+3 Energy [eV]

NEA-K Correlation Matrix

Figure 3. Pu-240(n,y)

ANL Correlation Matrix

89

^ 20 c 2

-NEA-K -ANL

1.E+1 1.E*3 Energy[eV]

NEA-K Correlation Matrix

ANL Correlation Matrix

l _

i

Ll Figure 4. Pu-241(n,f)

NEA-K ANL

elative

dDevia

c .2

1.E+1 1.&3 Energy [eV]

NEA-K Correlation Matrix

Figure 5. Am-241(n,f)

1.E+5

1.E+7

ANL Correlation Matrix

90

4. The systems used in the analysis As indicated in the Introduction, six systems, related to Gen-IV and AFCI, have been considered. Their main features are: GFR: 2400 MWe - He cooled; SiC - (U-TRU)C fuel; Zr3Si2 reflector; enrichment: 17%, MA:5%; irradiation cycle: 415 d LFR: 900 MW& - Pb cooled; U-TRU-Zr metallic alloy; Pb reflector; enrichment: 21%, MA: 2%; irradiation cycle: 310 d SFR: (Burner: CR=0.25) 840 MW 4 - Na cooled; U-TRU-Zr metallic alloy; SS reflector; enrichment: 56%, MA: 10%; irradiation cycle: 155 d EFR: 3600 MWj, - Na cooled; U-TRU oxide; U blanket; enrichment: 22.7%, MA: 1%; irradiation cycle: 1700 d VHTR:TRISO fuel; enrichment: 14%; burnup: 90 GW d/Kg Extended BU PWR: enrichment: 8.5%; burnup: 100 GW d/Kg The neutron spectra and adjoint fluxes at core center are shown in Figs. 6, 7. The parameters considered are: Criticality (multiplication factor); Doppler Reactivity Coefficient; Coolant Void Reactivity Coefficient; Reactivity Loss during Irradiation; Transmutation Potential (i.e. nuclide concentrations at end of irradiation); Peak Power Value; Decay Heat in a Repository (at 100 years after disposal); Radiation Source at Fuel Discharge; Radiotoxicity at t= 100000 years after disposal. 5. Results A summary of the results obtained with the ANL Covariance Matrix (both without correlations, and with partial energy correlations, PEC) is shown in in the Annex, Tables II through IV. Uncertainties are significant for Keff for all systems and, for the burn-up reactivity swing in thermal systems and, at a lesser extent, for coolant void coefficients in fast systems and neutron source in thermal systems at fuel unloading. For all the other parameters, uncertainties are well within any anticipated target accuracies. The breakdown of the most significant uncertainties on integral parameters by isotope contributions (Tables V to X), shows that the major contributors in the case of Keff, are still U-238 and Pu-239. Despite the presence of significant amounts of MA in the fuel of the fast systems (in particular in the case of SFR), MA data uncertainties do not play a major role with a few exceptions (Am-241 a c for GFR and Am-242m a f for SFR).

t 1 1—

'Z

i i rpn-i r -\-<

1

£

1 - _ -A. - - -

:::jn:f::i^::

•*

fa"l»

f

...

t—

T

„

rX

t

i

H f r ^ - '"L,

i

• i—.

i

'1

L.

""

».;.".„

""

i i i r i i i i i - ~ i - | - T - i - i - - i - t - - r -A- * --t-i- + -i-i—I-I—i j i - . - - 1 - 1 - + -1-1—i - 1 — M l _ . -i - u j — i _ J—i _ i_ - i p i t r , . j _ i _ i _irrii_L jj_ii,

r- -i - r _t_H - r-

J _ L _L _ l _

[»V]

r

- ,..n

'

rot

I — :

i _ _J i

i...

^ ^ - \-

i — '

1--L

1

1

+—1 r

i

i

i

1

1

1

1

1

1

1

1

1

1 1

U

E n i r l y I.V)

Figure 6. Direct flux distributions of fast neutron systems: a) GFR; b) LFR, c) SFR; d) EFR; e) VHTR; f) high-BU PWR i J_

i

i

i

i

—I — 4- —I — J

i 1 -

\j^ - \- -r,p I

r|T{.n~ r T ~ r T ~(±J±_I__L_J_I_-1_I_ i i i i i i i i

—1

i

-

i

i

i

i

i Ji ri

i

i

i

i

i

11

t - i - t- f f t - i - r_Pi " 1- -I "_rf "I " •^-^r^r -i -

r iIt-

h / i

i i i i i i i - -1 - I - 4- - I - 4- - I - 4- -

_i

J

i

i

t--h—1 — 1— -t--|—1 — _ i _ J L_I„I_J l_l_ i 1 1 1 1 1 11

" i r n -Fi l l T T l ~ l ~ - -r -\ - 1 - h{ -r -1 ~ 1- -r- -1 ~ 1 1 1 1 1 1 1 'I" 1 1 1 1 1 1 1 1 11

t ^~-iir r T ~t" - T I r T ~r - r i J i T ~i ~

mi

1 ~\~ I I n _i~ i i

" T 1

-H

~i~

if

I

Figure 7. Adjoint flux distributions of fast neutron systems; a) GFR; b) LFR, c) SFR; d) EFR; e) VHTR; f) high-BU PWR

92

Tables XI and XII give the uncertainties related to the major isotope density variations AN during the cycle, for the GFR and VHTR respectively. Significant uncertainties are shown for the Pu-239, Am-243, Cm-244 and Pu-241 AN values in the GFR case. The capture of U-238, Pu-240, Pu-242 and Am-243, and the fission of Pu-239 and Pu-241, play the most significant role. In the case of the VHTR, lower uncertainties on isotope density variations are observed, the largest values being related to the AN of Np-237 (due to U-236 capture), Pu-238 (also due to the U-236 capture), Pu-240 (due to the Pu-240 capture), Am-241 (due mostly to Am-241 capture) and Am-243 (due to Pu-242 capture and Am-243 capture). As for the energy breakdown, Table XIII is related to the GFR Keff uncertainty for the major isotopes, Table XIV to the SFR and Table XV to the VHTR Keff uncertainty. The most significant data for the different reactions/isotopes are: Pu-239 fission between 1 MeV and 1 keV and below 1 eV; Pu-239 capture below 1 eV; Pu-240 capture at the first resonance; Pu-241 fission between 1 MeV and 1 keV; U-238 capture between 0.2 MeV and 2 keV and between 400 eV and 10 eV; Am-242m fission between 1 MeV and 10 keV. Finally, Tables XVI and XVII give a comparison between the uncertainties, which result from the use of the ANL, and the NEA-K covariance matrices, both for the GFR and the VHTR. Comparable results are observed in the case of Pu-239 for the GFR integral parameters and of the Pu-240 for the VHTR case. In the case of U-238 Pu-241 and Si, the NEA-K covariance matrix data give lower values (factor -2 to 3). This intercomparison has to be taken with precaution, since the NEA-K does not represent a fully consistent and complete set of data. However, the results seem to indicate that the present covariance evaluations would tend to show significantly lower uncertainties on the integral parameters, which in principle indicate lower needs for data improvement. 6. Lessons drawn and perspectives 1. Data uncertainties, whatever the hypothesis, are significant only for a few parameters: • Kgff for all systems (in the case of thermal systems, at EOC due to high burn up); • Burn up (BU) reactivity swing and related isotope density variations AN during BU;

93

• • •

Some void coefficients in fast systems; At a lesser extent, neutron source (thermal systems) at fuel unloading; For other parameters, the uncertainties are within anticipated target accuracies. It has to be kept in mind that only neutron cross section data have been considered in this study and that no fission product data uncertainty has been considered, thus underestimating somewhat the uncertainty on the reactivity swing during BU and the decay heat. 2. Despite a significant MA recycling in fast systems and extended burn-ups in thermal systems, MA data do not play a major role. Some exceptions: Am-243 capture in the "fast" and "thermal" range; Am-242m fission in the "fast" range. 3. As for major actinides, besides U-238, Pu isotope data uncertainties are very significant: Pu-239 fission between 1 MeV and 1 keV and below 1 eV; Pu-240 capture at the first resonance; Pu-241 fission between lMeV and 1 keV; U-238 capture between 0.2MeV and 2keV and between 400eV and lOeV; U-238 inelastic. 4. As for structural/coolant materials, the most significant data are: Fe inelastic; Na elastic; Pb inelastic; Si inelastic. 5. These indications depend of course on the uncertainty data, which have been used. However, consistent results have been obtained, e.g., for Pu-239 fission and Pu-240 capture (in the thermal range), when the NEA-K data (from IRDF and JENDL-3, respectively) have been used. On the contrary, if JENDL-3 data are used for, e.g. Pu-241 fission and for Pu-240 in the "fast" range, or ENDF/BVI data for Si-28, no further requirement for improvement would appear necessary. 6. Better and more complete covariance matrices are needed. However: • The timescale for the production of a full set of covariance matrices should be 2-3 years, to have impact on the feasibility assessment of most concepts and to provide the basis for potential, high priority new cross-section measurements or nuclear data oriented integral experiments.

94

•

The data should be easy to process in a user oriented format (i.e. by energy group). In particular, uncertainties on some major U and Pu isotope resonances should be processed in the given multigroup structure. • To propose a credible program of new cross-section measurements, one should show the impact on relevant parameters of the existing uncertainties taking into account design parameters target accuracies. This assessment can be made only if the cross-section covariance data are available: they do not need to be perfect, but reliable, consistent, and complete enough to make a relevant point clear. All major sources of uncertainties (cross-sections, secondary energy and angular distributions, fission spectra) should be considered. Finally, it has to be kept in mind that integral experiments and statistical data adjustments are still today a powerful tool to overcome most difficulties, since they provide a global validation of data, and allow to point out data types and energy domains which need improved evaluations. Statistical data adjustments taken "per se" can be used with confidence in design calculations, only if a number of conditions are satisfied: The integral experiments used in the adjustment procedure should be "clean", in the sense that the associated experimental uncertainties are small and well understood; The integral experiment calculated values should not be affected by any significant modeling uncertainty (e.g. geometrical description, number of energy group used in the analysis, etc.), in order to avoid the introduction of systematic errors; The covariance data should be reliable, complete, and consistent. In this case, the new covariance matrix obtained as a result of the adjustment, is a very powerful source of information; Data adjustments should as much as possible relate to the physics parameters which describe the cross-sections, to make adjustments independent from the energy collapsing procedures. In summary, better uncertainty data will play an essential role both in the assessment of potential new data needs with reduced uncertainties, and in design oriented statistical data adjustments.

95

References 1. G. Aliberti, G. Palmiotti, M. Salvatores, C. G. Stenberg, "Transmutation Dedicated Systems: An assessment of Nuclear Data Uncertainty Impact", Nucl. Sci. andEng. 146, 13-50, (2004) 2. G.Palmiotti, G.Aliberti, M.Salvatores, J.Tommasi, "Integral Experiment Analysis for Validation and Improvement of Minor Actinide Data for Transmutation Needs", Proc. Int. Conf. ND-2004, Santa Fe, Sept. 2004 3. G. Palmiotti, M. Salvatores, "Proposal for Nuclear Data Covariance Matrix", JEFDOC 1063 Rev.l, January 2005 4. I. Kodeli, E. Sartori, ZZ-COV-15GROUP-2005, NEA-1730 Package, OECD/NEA Data Bank (in preparation).

Annex. Tables II-XVII Table II. Fast Neutron Systems - Total Uncertainties Power Burnup Reactor Doppler Void Keff Ap (pern) Peak ±1.2 ±3.6 ±4.8 ±240 N C (*> ±1.20 GFR 1.90 PEC 1.8 5.5 7.1 384 1.51 5.2 13.0 NC 0.8 177 LFR PEC 2.26 1.0 7.8 20.6 258 1.10 0.4 4.1 17.8 NC 156 SFR 1.66 6.0 23.4 234 PEC 0.5 1.02 3.4 8.4 NC 0.7 652 EFR 1.57 12.1 PEC 1.1 5.1 989 ( ) * NC: No correlation

(%) Decay Heat ±0.3 0.5 0.3 0.6 0.2 0.4 1.6 2.3

Table III. VHTR. Total Uncertainties (%) Peak Peak Doppler Doppler Burnup Keff Keff Power Power [pcm] BOC EOC BOC EOC BOC EOC ±1484 NC(*» ±0.41 ±0.94 ±1.6 ±1.8 ±3.7 ±5.8 2.1 3.4 1576 PEC 0.58 1.07 1.9 5.6 ( * ) N C : No correlation Table TV. Extended BU PWR. Total Uncertainties (%) Doppler Doppler Decay Burnup Keff BOC EOC BOC EOC [pcm] Heat ±0.97 ±2.1 ±2.3 ±4.0 ±1916 NC <*> ±0.33 3.8 PEC 0.52 1.27 3.1 4.6 2206 ( *' NC: No correlation Keff

Dose ±0.4 0.6 0.4 0.5 0.2 0.2 1.1 1.7

Decay Heat ±2.<> 3.1

Neutron Source ±1.2 1.8 0.8 1.2 0.6 0.9 3.9 6.0

Neutron Source ±1.9 2.6

Dose ±2.3 3.1

±12.2 14.3

Neutron Source ±11.2 13.3

96 Table V. GFR. Uncertainties (%) PEC - Breakdown by Isotope (Major Contributions) Burnup Doppler Void Isotope Keff [pcm] ±1.22 ±3.2 U238 ±3.9 ±63 0.22 0.7 85 Pu238 0.6 1.03 2.6 2.6 203 Pu239 0.29 0.7 14 Pu240 0.7 0.57 1.7 189 Pu241 1.5 0.43 1.2 Am241 1.8 73 0.01 Am242m 0.0 0.0 76 Cm242 0.00 0.0 0.0 90 0.13 0.4 Cm244 0.3 35 0.17 0.4 Cm245 0.5 38 0.31 C 1.9 1.7 8 0.42 Si28 1.2 0.7 12 0.12 Zr90 0.5 0.3 9 Table VI. LFR. Uncertainties (%) PEC - Breakdown by Isotope (Major Contributions) Burnup Isotope Keff Doppler Void [pcm] ±0.73 ±2.2 U238 ±3.7 ±13 0.24 Pu238 0.9 0.5 25 1.50 Pu239 3.4 4.0 213 0.41 Pu240 1.1 0.9 18 Pu241 0.32 0.7 1.0 112 Am241 0.10 0.4 0.3 6 Am242m 0.06 0.2 0.1 14 Cm242 0.02 0.0 0.0 11 Cm244 0.13 0.2 12 0.3 0.21 Cm245 0.4 0.7 34 0.24 Fe56 1.6 2.0 5 Pb206 0.88 3.2 13.4 18 0.80 Pb207 3.4 12.2 16 Pb208 0.49 4.0 7.4 8

M

tO - J W >J U

tO

c c

a

CO

Crt

o CD

"__ H \o w b\ io b\ 9 oe

0 0 \ U i U O O \ i - 0

CO B

B

n o

B

CD

Z

•a

£ 3c

[pc

t O ^ W t O N N J t O H -

U t O W | O i - l > i - ' | + ^ C\ W W W Ul M t o •— •&• o -J

CD

opp

o o

CB

O O O O W U ) t O | +

1

o o

b o o o o o

- ] ^

o o

CB

O

o

*

CO

1

W

P

JT

g

B

n

£

5. < cr a

3

O

O

D

g

C 11

T:
utio

O O O O O O t O | +

—

0 0 0 0 0 0 0 1 + OOO^-bvLnLnP

o o o o o o ^ P

O O O O O O O I +

00 U> *13

^

w ^ w «

T

_

UI

L/I

O ^ H - p o p p O ^ - — > OJ O b\Pis)Wtob«^wi^«oo

O

i — ^ - o p o o p o p H - L ^ o i +0 towlototobi-^^to*©^?

0 0 0 0 0 0 0 0 0 0 ^ - 0 1 + ^ i o s j o o o o o t o u ) t o ^ - P

3

^

^, c_ c,. c_ „c >

L f t - ^ t O t O ^

o z a o o c>

3.1

5?

!

O

CB

n ^

2 v

•-

R

tO OO ^

tO

M

W M

O

f

^-' t—> ^

^

U

U

. C\ Wi " - 00 I ^ OOUl ^

If -

t o ^ - *>. y s u i | | *— O Lfl VO LO >-.

^Loobouj^-toc^wj^Lftbor

H

to to * - * - - © — p o ^ ^ - t o ^ - | + j^^tO'-^^bobs^iijuii-'P

^ . t O - J O N - k - I ^ W W H - '

wi3\u)wb^'-'toul/>bowp

t o Is) N> t o N»

• a > B *p c ^ c ^ c ^ E !5

pppppppppppp{+

3

to t o

0\ K> & ft

zjnnn

3.3

«T

^_ rr*

o

s1

3 o

3

on

o ^

•-t

u 2 H £O g; •—;•

utioi

98

Isotope U235 U238 Pu239 Pu240 Pu241 Pu242 Am243 Cm244

U238

Pu240

Pu242

Am243

Cm244 PEC

ill III III

Am241

-

Burnup [pcm] ±228 391 637 1727 359 111 197 78

Neutron Source ±0.02 2.19 2.07 2.45 2.38 3.49 10.01 6.48

Am243

Cm244

10.10 0.13 0.00 0.25 0.00 0.00 14.21 2.90 0.00

0.38 0.00 0.00

III

Pu241

Table XI. GFR. 8(AN)/AN PEC Uncertainties (%) Pu239 Pu241 Am241 ±15.62 capture fission 0.05 N,2N 0.01 5.44 0.14 9.37 0.00 0.06 0.00 capture 7.92 0.09 fission 0.05 0.00 0.00 0.00 N,2N 2.40 0.03 6.90 0.08 0.13 0.00 capture fission N,2N 5.85 1.38 0.00

(%) PEC

III

Pu239

Table X. Extended Burnup PWR. Uncertainties Breakdown by Isotope (Major Contributions) Keff Doppler EOC EOC BOC BOC ±0.16 ±1.0 ±0.3 ±0.35 3.2 0.69 2.8 0.36 0.00 0.62 0.0 1.7 0.00 0.62 0.0 2.4 0.00 0.32 0.0 0.2 0.00 0.07 0.0 0.0 0.14 0.00 0.0 0.1 0.00 0.06 0.0 0.0

19.03

10.78

6.01

17.69

9.10 0.11 0.00 4.16 4.90 0.02 11.15

99 Table XII. VHTR. S(AN)/AN PEC Uncertainties (%) Np237 Pu238 Pu239 Pu240 Pu241 U235 capture ±1.45 ±1.31 ±0.00 ±0.00 U236 capture 6.97 6.31 0.02 0.01 U238 capture 0.00 0.28 2.52 2.55 2.58 Np237 capture 1.40 2.75 0.01 0.01 Pu238 capture 1.26 0.02 0.01 capture 0.23 1.02 1.83 1.91 Pu239 fission 0.07 1.02 0.94 1.07 Pu240 capture 0.26 5.73 0.94 0.04 capture 0.56 Pu241 fission 0.08 1.10 Pu242 capture Am241 capture 0.43 Am243 capture PEC 7.26 7.15 2.92 6.61 3.69

Pu242

Am241

Am243

2.62

2.60

2.63

2.05 0.80 1.76 2.34 0.88 1.02 0.03

1.98 0.86 1.38 0.50 0.98

4.70

6.69

2.13 0.72 2.26 2.39 0.76 4.12 0.06 4.62 7.85

Table XIII. GFR. Keff Uncertainties. Energy breakdown [pcm] Gr. Energy U^cw U238oin Pu239afis! Pu^'a^ Am241aoaw 1 ±0 ±9 ±0 19.6 MeV ±36 ±5 2 524 6.07 MeV 8 90 54 0 14 92 3 2.23 MeV 292 30 1 4 61 236 23 1.35 MeV 151 76 5 61 18 243 96 43 498 KeV 6 289 74 183 KeV 97 21 139 7 67.4 KeV 157 240 79 1 140 8 180 24.8 KeV 0 201 76 134 134 9 255 9.12 KeV 0 184 107 10 2.03 KeV 77 0 106 84 116 11 19 38 454 eV 0 41 33 12 1 22.6 eV 0 0 1 1 13 4.00 eV 0 0 0 0 0 14 0.54 eV 0 0 0 0 0 15 0.10 eV 0 0 0 0 0 Total [pcm] 314 620 625 353 199

Si28aiD ±26 292 93 0 0 0 0 0 0 0 0 0 0 0 0 307

Gr. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Total

Table XIV. SFR Keff Pu^Of Pu^Of Energy ±7 19.6 MeV ±4 6.07 MeV 36 76 2.23 MeV 40 87 1.35 MeV 113 261 498 KeV 94 351 183 KeV 50 293 148 67.4 KeV 90 118 24.8 KeV 80 9.12 KeV 35 43 2.03 KeV 64 44 454 eV 11 13 22.6 eV 0 1 4.00 eV 0 0 0.54 eV 0 0 0.10 eV 0 0 [pcm] 575 217

Uncertainties Energy breakdown [pcm] „ Cm r'm5<w Of „ P ^ c , Pu^Of Am'242m Of ±9 ±6 +8 ±3 59 81 39 75 89 37 75 38 109 138 189 185 180 42 262 33 9 183 258 18 5 111 152 10 4 6 101 70 1 3 43 29 8 65 47 2 0 0 17 11 0 0 3 1 0 0 0 1 0 0 0 0 0 0 0 0 227 334 434 220

5.50

cJ6 a„m re ±30 111 114 242 0 0 0 0 0 0 0 0 0 0 0 291

Na23 oin ±9 51 42 238 1 0 0 0 0 0 0 0 0 0 0 247

100 Table XV. VHTR. EOC Keff Uncertainties [pern] PllOT"n U Oft U Ocapt Gr. Energy itf ±0 ±0 ±0 1 19.6 MeV ±0 0 2 6.07 MeV 1 0 0 1 0 0 3 2.23 MeV 2 2 0 4 1.35 MeV 498 KeV 2 2 0 5 183 KeV 2 2 0 6 67.4 KeV 2 4 0 7 3 7 0 8 24.8 KeV 1 9.12 KeV 8 10 9 1 2.03 KeV 10 24 10 11 454 eV 51 185 11 3 11 12 22.6 eV 34 337 5 2 13 4.00 eV 38 8 5 0.54 eV 346 14 80 9 360 23 0.10 eV 46 5 21 15 Total [pem] 118 385 361 347

U238 Pu239 Pu240 Pu241 Si Zr

U238 Pu239 Pu240 Pu241

V

pu 2 4 u cw

PUM1^:

±0 0 0 0 0 0 0 0 0

±0

1 11 2 620 24 6 621

Table XVI. GFR. PEC/NEA-K Uncertainties (%) Comparison Keff Doppler Void ±1.22 ±3.2 PEC ±3.9 NEA-K 0.58 1.6 1.5 PEC 1.03 2.6 2.6 NEA-K 2.4 3.3 1.20 0.29 0.7 0.7 PEC NEA-K 0.2 0.08 0.1 PEC 0.57 1.5 1.7 NEA-K 0.11 0.1 0.3 0.42 1.2 PEC 0.7 NEA-K 0.09 0.4 0.3 0.12 PEC 0.3 0.5 NEA-K 0.02 0.0 0.1

1 0

2 4 26 56 6 88 23 110

Burnup [pem] ±63 31 203 160 14 5 189 39 12 3 9 1

Table XVII. VHTR. PEC/NEA-K Uncertainties (%) Comparison KeffBOC Keff EOC Burnup [pem] PEC ±0.43 ±0.55 ±150 0.24 NEA-K 0.19 60 PEC 0.00 0.57 624 0.18 NEA-K 0.00 211 PEC 0.00 0.63 1313 0.00 0.51 NEA-K 1017 PEC 0.00 0.17 222 0.00 0.05 NEA-K 73

SENSITIVITY AND UNCERTAINTY STUDY FOR THERMAL MOLTEN SALT REACTORS ADRIEN B I D A U D 1 2 , T A T I A N A I V A N O N A 2 ' ' , VICTOR M A S T R A N G E L O 2 ' \ IVO KODELI 3 'institut de Physique Nucleaire Orsay/CNRS, 15 rue G. Clemenceau 91406 Orsay 2 CNAM-Laboratoire de Physique 292 rue Saint-Martin 75141 Paris Cedex 03 3 OECD-NEA/DB, 12 Bd des lies 92130 Issy-les-Moulineaux

The Thermal Molten Salt Reactor (TMSR) using the thorium cycle can achieve the GEN IV objectives of economy, safety, non-proliferation and durability. Its low production of higher actinides, coupled with its breeding capabilities - even with a thermal spectrum are very valuable characteristics for an innovative reactor. Furthermore, the thorium cycle is more flexible than the uranium cycle since only a small fissile inventory (<2 tons by GWe) is required to start one reactor. The potential of these reactors is currently being extensively studied at the CNRS and EdF /1,2/. A simplified chemical reprocessing is envisaged compared to that used for the former Molten Salt Breeder Reactor (MSBR). The MSBR concept was developed at Oak Ridge National Laboratory (ORNL) in the 1970's based on the Molten Salt Reactor Experiment (MSRE). The main goals of our current studies are to achieve a reactor concept that enables breeding, improved safety and having chemical reprocessing needs reduced and simplified as much as reasonably possible. The neutronic properties of the new TMSR concept are presented in this paper. As the temperature coefficient is close to zero, we will see that the moderation ratio cannot be chosen to simultaneously achieve a high breeding ratio, long graphite lifetime and low uranium inventory. It is clear that any safety margin taken due to uncertainty in the nuclear data will significantly reduce the capability of this concept, thus a sensitivity analysis is vital to propose measurements which would allow to reduce at present high uncertainties in the design parameters of this reactor. Two methodologies, one based on OECD/NEA deterministic codes and one on IPPE (Obninsk) stochastic code, are compared for kdf sensitivity analysis. The uncertainty analysis of kdf using covariance matrices available in evaluated files has been performed. Furthermore, a comparison of temperature coefficient sensitivity profiles is presented for the most important reactions. These results are used to review the nuclear data needs for the TMSR thorium fuelled reactor.

1. Thermal Molten Salt Reactor Molten salts are very good coolants and their thermo-hydraulic characteristics are close to those of liquid metals.They can also be operated with low pressure and have the further advantage of being transparent. However, one disadvantage is that they can be corrosive. Their chemical compatibility with structural

101

102

material has been investigated in the Oak Ridge Molten Salt Reactor Experiment, an 8 MWth reactor operated between 1964 and 1969 with a 85% load factor. Corrosion resistant material Hastelloy-N showed a very promising behavior. In the MSBR concept (Molten Salt Breeder Reactor) the salt chemical reprocessing requirements were known at the time to be beyond industrial capacity. For this reason, the main objective of LPSC and EdF work is to minimize the salt chemical reprocessing while keeping a near zero regeneration gain. The addition of fertile blankets allows an increase in the breeding ratio while maintaining good safety coefficients. The core consists of hexagonal graphite blocks (pitch is 13.5 cm) with cylindrical channels in which the salt (LiF-ThF4) flows. The channels containing the fissile salt (U/Th) are surrounded by a fertile region (Th). This concept can work in a very wide range of conditions of different power density, temperature, actinide loading, and neutron spectrum. This complicates the studies as there are many different characteristics which can be adjusted by a large number of parameters. The densities at equilibrium are a balance between neutronic production and removal reactions and chemical reprocessing. The equilibrium can take quite a long time to reach. Our study is limited to the sensitivity and uncertainty analysis of the equilibrium state. In the framework of the GEN IV, reactors should be compared on their fulfillment of sustainability requirements, i.e. on their nominal behavior when deployed on large scale. Furthermore, as the spectrum is mostly dependent on moderation and thorium and uranium loadings that are roughly constant, it will not change significantly with burn-up. Thus the main contributors to nuclear data uncertainty are expected to be computed correctly. A small salt channel radius corresponds to a reactor containing more graphite and therefore to a more moderated spectrum. For radii smaller than 10cm the neutron spectrum is thermal, whereas for larger radii the spectrum tends to be fast. Power density, temperature, actinide loading and fuel reprocessing are being conserved. It is important to note that all these parameters cannot be optimal in such a wide range of neutron spectrum when comparing these results. For channel radii smaller than 7 cm breeding ratio is bigger than 1.0 and a long graphite lifetime is achievable, but the global coefficient is unfortunately insufficiently negative. However, the uncertainties of nuclear data can affect this coefficient and this is one of the main reasons the impact is being investigated carefully (see below). 8.5cm is the actual 'reference' radius, which enables a negative temperature coefficient and an acceptable breeding ratio. Graphite lifetime is very short and further study should be carried out to determine its eventual extension as well as the cost in terms of waste management.

103

The total temperature coefficient can be separated into 3 components. The Doppler coefficient corresponds to the usual effect of the broadening of the capture resonances with the increase of fuel temperature and is always negative in this concept. The density effect corresponds to the dilatation of the salt with temperature; Part of the salt moves out of the reactor which tends to decrease keff, however, thermalization is more efficient as neutrons "see" more graphite than salt which tends to increase the keff. The last effect is due to graphite heating. The thermal part of the spectrum shifts with temperature (see Figure 3b) in a region where the fission cross section of 23 3U decreases slower than 232Th capture cross section. Hence, fissions are favoured and k<,ff tends to increase with graphite temperature. Certainly, the more graphite that is present in the reactor (i.e. smaller radius of channels with salt) the more important this coefficient. The graphite coefficient will produce an effect on a different time scale than the fuel temperature coefficients because the moderator is not the coolant in this concept, and because graphite has a higher thermal inertia than salt. So even if the global coefficient is close to zero, the reactor behaviour is expected to be stable in most transients. Nevertheless, only a complete study including coupled thermohydraulic and neutronic simulations of transients can confirm the safety of the concept. The total coefficient is slightly positive for thermal systems and strongly negative elsewhere due to strong Doppler coefficient. 2. Sensitivity analysis methodologies For these analyses the discrete ordinates (Sn) transport codes DANTSYS and DOORS and the SUSD3D code /3/ allowing sensitivity and uncertainty analysis using Perturbation Theory were obtained from the OECD/NEA Data Bank. In addition to the above set of programs the cell code APOLL02 /18/ was used to generate cell averaged multi-group cross sections (XS). Data for the TMSR were prepared at the EdF. These data were converted by a specific C++ routine into FIDO format read by the DORT and TWODANT programs. The results obtained with the afore-mentioned 'classical' approach are compared with the IPPE (Obninsk) sensitivity package based on the MMKKENO multigroup stochastic code /4, 5/. The peculiarity of this algorithm is in the macroscopic cross-section derivatives calculation through ratios of summary weights on the consequential generations. An importance function based on this technique is used in perturbation formulae to evaluate sensitivities. The above sensitivity methods produce energy-dependent sensitivity coefficients that give the relative change of keg, to a relative change of the crosssection data by isotope, reaction and energy. The sensitivity coefficients are typically presented as profiles, where the change in A^due to perturbations of

104

the cross sections is given as a function of energy. It shows the importance of this reaction for kgff as a function of the neutron energy. Integrated over the neutron energy, the sensitivities allows a comparison of the importance of each reaction for each nuclide. 3. Nuclear data uncertainties It is possible to propagate the nuclear data uncertainties (given in the form of covariance matrices) to keff value using a sensitivity analysis. Unfortunately, information on nuclear data uncertainties is very limited and can differ significantly between different nuclear data libraries. We use nuclear data uncertainties coming from different libraries even if the library used for the calculation is not the same. This can be justified by the fact that evaluators share the same basic information (experimental data are being gathered in the EXFOR project 191) and very often the same evaluation tools and nuclear models such as SAMMY /10/ for the resonance parameters or EMPIRE /l 1/ or TALYS 1121 for nuclear model at higher energy. Some parts of evaluations are shared by some libraries because the evaluators of one library did not need or were not able to improve the cross sections for particular isotopes. As the effort is being shared by a reducing community, cases like the 238U resolved resonance parameter sets jointly developed by Oak Ridge national Laboratory and CEA Cadarache and included in both ENDF/B-VI and JEFF-3.1 evaluations could become common practice. Producing covariance matrices is almost as difficult as producing crosssection sets for evaluators. Furthermore, little feedback is obtained from the users as only a few programs allow sensitivity and uncertainty calculations and (maybe therefore) a few users are willing to make the effort of performing such studies. For innovative reactors such as the GEN IV reactors, point wise stochastic tools are often preferred. Another issue concerns the resonance formalism. The last decade has seen the development of the Reich-Moore formalism which is considered more advanced R-Matrix formalism than Breit-Wigner approximation. Even if the ENDF format was able to store the uncertainties on the Reich- Moore resonance parameters, ERRORR -NJOY module used for propagation of evaluated uncertainty to group cross-section uncertainties was not able to treat them. This issue has been addressed recently by JENDL evaluators. They developed the ERRORR-J module 111 which calculates the impact of one resonance parameter on the group cross section by changing the parameter by 1% and recalculating the cross section. Using this methodology the code is able to propagate the uncertainty of any resonance parameter to the cross section in the group

105

containing the resonance. Thus the covariance matrices produced give correlation between reactions (as most of the resonance parameters are shared by different channels). Facing the lack of uncertainty information available in evaluated files, ANL /13/ developed their own uncertainty guess based on their experience of calculation and experiment comparisons. This transposes the problem of evaluating uncertainties on nuclear model parameter to evaluating uncertainties on the cross sections as a final product. Due to its completeness and its likelihood, this uncertainty set is valuable even if quite rough. The only covariance matrix available for 232Th can be found in three different libraries (ENDF/B-VI, IRDF90v2 and JEFF3.0) and is in fact coming from 1977 ENDFB evaluation even though the evaluations have been improved up to the 90's. This matrix is rather simple: only 4 diagonal values with no correlation between them. And the uncertainty values are about 12% in the important energy range. The uncertainty proposed in /13/ is 10% in this region which is compatible with the available uncertainty information. New evaluations, based on new measurements done in GEEL, Belgium and in n-TOF at CERN in the framework of IAEA/NDS/14/ are expected to decrease this uncertainty down to 5% which means that 23 Th cross sections accuracy would still be worse than U by a factor of 2 which is satisfactory, keeping in mind that U is a standard studied and used in power reactors since fifty years. For 233U only one set of covariance matrices can be found in JENDL3.3. No energy group correlation can be found outside the thermal range. Furthermore these uncertainties seem low (0.2% standard deviation up to lOOeV) in comparison with expected value from Ref. /13/: few %. The truth being probably somewhere in the middle, our calculation of the contribution of this isotope is likely to be slightly underestimated. 4. Sensitivity analysis Table 1 compares the energy-integrated sensitivity coefficients obtained with deterministic and stochastic methods. It contains the coefficients for the most relevant nuclides and reaction types. Though two different methods, different nuclear data libraries (ERALIB /15/ and ABBN93 /19/), and different approaches for the whole core and adjoint core calculations (2 dimensional models for the deterministic method and detailed 3 dimensional for the stochastic method) are used, the results are in a good agreement. The deterministic method seems to give slightly higher capture reaction sensitivities and lower fission reaction sensitivities.

106 Table 1. Comparison of integrated sensitivities (%/%) Deterministic Isotope Reaction Method(%/%) (n,elas) 0,340 c (n.gamma) -0,075 U233 (n,fission) 0,370 Nu total 0,895 (n.gamma) -0,019 U235 (n,fission) 0,038 0,094 Nu total Th232 (n,gamma) -0,379 U234 (n,gamma) -0,060 Pa233 (n,gamma) -0,014

Stochastic method 0,310 -0,068 0,385 0,883 -0,019 0,043 0,099 -0,371 -0,058 -0,013

Figure 1 shows the calculated sensitivity coefficient versus the neutron energy. The sensitivity results of the 172 self-shielded group cross sections (provided by the deterministic calculation) seem to be significantly higher in some groups than those obtained with the stochastic 30 energy groups, however, it should keep in mind that the 30 energy groups are averaged over a broader energy range. In this context one should outline that the stochastic core calculations are performed with 299 energy groups and only the sensitivity coefficient are condensed to 30 energy groups. Integrated results over the whole energy range shown in Table 1 are consistent.

0,00

""•^

-0,05 -0,10 -0,15

y+UMV*'*'^ •

i

• •

i i

, <

-0,20

i i

-0,25

•"•"•"""•stochastic method

i i i i

-0,30 -0,35

deterministic method

-0,40 1.E-05

1.E-03

1.E-01

1.E+01

1.E+03

1.E+05

1.E+07

Energy

Figure 1. Sensitivity profile of keg to

Th capture cross section.

1.E+09

107

5. Uncertainty analysis Table 3 shows the results obtained with the covariance matrices found in available evaluated files. The global uncertainty calculated here is about 4000 pcm. This uncertainty is dominated by only one reaction: capture on 232Th. It should be noted that in case of no correlation between different contributors the uncertainty is the root of the sum of squares and thus a contributor that is 3 times higher contributes almost 10 times more to the global uncertainty. It has been pointed out that in some cases the first order sensitivity used for the uncertainty propagation does not agree very well 151 with the direct sensitivity. However, the sensitivity (and so the uncertainty) of keff to cross section is lower than what can be obtained with first order perturbation methods. Using the next generation evaluations, more realistic uncertainties (probably 5% instead of 12% in the resonance range that is important for the thorium capture as shown on Figure 6) and corrected sensitivities, the global uncertainty could be reduced to about 1500 pcm. No uncertainty information for 234U capture has been found although the corresponding sensitivity is quite high. We can estimate its contribution by using the recommendation of ref. /13/ (10% in the range from 5eV to 500 eV containing the resonances that are making the largest part of the sensitivity) and the sensitivity of Table 1 confirmed by Table 2 (0.06%/%). Table 2 Uncertainty analysis Isotope

U235

U233

Th232

Total

Uncertainty

Reaction

Sensitivity(%/%)

Uncertainty(pcm)

(n.gamma)

-1.98E-02

42

(n,fission)

3.67E-02

27

Nu total

9.30E-02

30

(n.gamma)

-8.03E-02

27

(n.fission)

3.36E-01

227

Nu total

8.75E-01

186

(n.gamma)

-3J9E-0I

3891

ENDFB6

(n.fission)

2.83E-02

34

IRDF90v2

Nu total

3,20E-02

0

JEFF3.0

-3900

JENDL3.2

JENDL3.3

108

This gives a 600 pcm contribution to the resulting uncertainty, meaning that even if 234U capture is an important contributor to the total uncertainty, its impact is still an order of magnitude lower than the that of the 2Th capture. In a real reactor the uranium density would be adjusted so that the reactivity is 1.0, therefore the real uncertainty of keff is zero. But the uncertainty has not disappeared: it has moved to an uncertainty on uranium density and on uranium inventory. This will also change the breeding factor and other macroscopic observables of the concept. For Molten Salt Reactors, the continuous chemical reprocessing allows some degree of freedom that solid fuel concept does not have. Whatever parameter is eventually tuned to compensate a discrepancy between real behaviour and numerical prediction, any nuclear data uncertainty impact should be established. This is because it will be paid for by an increased margin (of the capacity of the fuel reprocessing unit for liquid fuel reactors) and design complexity in order to allow the concept to accept such a discrepancy. 6. Preliminary sensitivity study of reactivity change with temperature Reactivity changes with temperatures are particularly important for this concept as the effect of a temperature increase in the graphite can make the total coefficient positive, in particular in configurations that are more neutronically efficient. The sensitivity of the reactivity change to cross sections can be estimated using usual perturbation methods I 111: am dp P dam

P

dp

dkl

dp

dk,

dcr

dk„ d<7

dk0

=

'

\k, [a*

3

*' ) k0 {a"

dk

°)

where indexes 0 and 1 correspond to base and perturbed calculation respectively. Figure 2 shows the sensitivity profile of reactivity change with respect to total temperature change 232Th capture cross sections. The shapes of the sensitivities in the thermal range are explained by the graphite temperature coefficient: if the capture of thorium appears to be more important in the upper part of the thermal range, the total temperature coefficient would be bigger (i.e. more negative) because the graphite temperature coefficient would be smaller. The other ranges important for the calculation of reactivity changes are the resolved and unresolved resonance range, the latter being of particular importance for 232Th capture cross sections.

109 3.5E+00

^^ 3.0E+00

Stochastic method Deterministic Method

HI

ffl

* • *

o

2,5E*00 2.0E+00 1.5E+00

0^

>. >

4-i

W c 0) V)

1,0E+00 5.0E-01 0,0E+00 -5.0E-01

1.E-05

1.E-03

1.E-01

1.E+01

1.E+03

1.E+05

1,E«07

1.E+09

Energy(eV) Figure 2. Sensitivity profile of temperature reactivity change to 232Th fission cross section.

The global trends are the most important for the understanding of the physics of the reactivity change with temperature, and are mostly concordant between the two methodologies. Nevertheless, the consistency between them is not excellent and strong discrepancies such as sign incoherence remain. It should be noted that these sensitivities are the difference of two almost equal terms divided by a small difference in k<,ff. Therefore, even small differences in the data, the data processing and the sensitivity calculations may have a large impact for this kind of calculation. 7. Conclusions Thorium thermal molten salt concepts have very promissing neutronic properties such as low fissile inventory, breeding and minimum waste production. Unfortunately, it seems difficult to design a reactor that gathers all the advantages. In particular, good reactivity coefficients need a faster neutron spectrum, while the graphite lifetime is improved in a more thermal neutron spectrum. Due to nuclear data uncertainties, designers would have to use wide safety margins. These high margins will diminish the potential of the concept by reducing its breeding or its competitiveness because of over-determination of the system parameters. The sensitivity and uncertainty analysis of the thorium cycle reactor concept have been performed in order to estimate the contribution of

110

each nuclear data to the uncertainty in k<,ff and to highlight the most needed measurements. Two different methodologies have been compared. The results can be used to update the high priority request list: Sensitivity and uncertainty analysis of keff has shown that 232Th capture cross section in the resolved resonance range was the main contributor to the uncertainty. Other major contributors are 233U fission and number of neutrons by fission as well as 234U capture. To reach the uncertainty on the keff of the equivalent U/Pu cycle (about 1.5%) the 232Th capture cross section should be measured to an accuracy as low as 3%. Preliminary sensitivity studies of reactivity change with temperature confirm the importance of the upper part of the thermal range for 233U fission and 232Th capture cross sections. The resolved resonance range is very important to estimate the Doppler coefficient. The temperature coefficient is also very sensitive to unresolved resonance of 232Th capture cross section. The sensitivities to the thermal treatment of carbon in graphite are not calculated. Since the reactivity change with temperature is very dependant on the thermal range, the impact is expected to be very strong. Uncertainty analyses using perturbation theory are powerful tools to estimate the uncertainties for the most important parameters of nuclear reactors. They permit to determine the accuracy which can be achieved in the simulations of new concepts. At the same time such analysis is useful to highlight the most urgently needed nuclear data. Its application is limited by the quantity and quality of uncertainties in the basic neutron data in the form of covariance matrices. Therefore, producing reliable covariance matrices should be of high priority. Acknowledgments This work was sponsored by GEDEPEON, the French research group working on waste production and waste management using new options. It is the result of a strong collaboration with David Lecarpentier from EdF R&D and Ludovic Matthieu from LPSC Grenoble/CNRS.

Ill

References 1. A. Nuttin et al., "Potential of Thermal Molten Salt Reactors: Detailed Calculations and Concept Evolution, With a View to Large Scale Energy Production. Prog. In Nuc. Ene., Vol. 46, No. 1, pp. 77-99, 2005 2. D. Lecarpentier, J. Vergnes, "The AMSTER (Actinide Molten Salt TransmutER) Concept", Nucl. Eng. And Des. Vol. 216, pp 43-67, 2002 3. I. Kodeli, "Multidimensional Deterministic Nuclear Data Sensitivity and Uncertainty Code System, Method and Application," Nucl. Sci. Eng. 138, pp.45-66 (2001) 4. A. Bliskavka. About estimation of derivation and disturbances of keff using Monte Carlo method. Preprint IPPE - 920, Obninsk, 1979 (in Russian). 5. A. Bliskavka, G. Manturov, M. Nikolaev, A. Tsiboulia. CONSYST/MMKKENO Codes for Nuclear Reactor Calculation Using Monte Carlo Method in Multigroups with P5 approximation. Preprint IPPE2887, Obninsk, 2001(in Russian). 6. M.L. Williams, BL. Broadhead and C. V. Parks, "Eigenvalue Sensitivity Theory for Resonance-Shielded Cross Sections", Nuclear Science and Engineering, Vol. 138, pp. 177-191 (2001) 7. G. Chiba, ERRORJ Manual, JNC TN9520 2003-008, Sept. 2003. 8. D. Lecarpentier, C. Garzenne, D.Heuer, A.Nuttin, "Temperature Feedbacks of a Thermal Molten Salt Reactor : Compromise between Stability and breeding Performances," Proc of ICAPP'03, Cordoba, Spain, May 4-7, 2003. 9. V. McLane. "EXFOR Exchange Formats Manual", IAEA-NDS-207 Rev.2004/08, August 2004 10. N. Larson et al. "Updated User's guide for SAMMY : Multilevel R-Matrix Fits to Neutron Data Using Bayes'Equation,ORNL/TM-9179,Martin Marietta Energy System, Inc., Oak Ridge National Laboratory - 5th revision No. ORNL/TM-9179/R6.", 2003 11. M. Hertman et al. Recent Developments of the Nuclear Reaction Model Code Empire , Proceedings of the international Conference on Nuclear Data for Science and Technology - ND2004, Sept. 26-Oct. 1, 2004, Santa Fe, USA 12. A. Koning et al., "TALYS Comprehensive Nuclear Reaction Modeling ", Proceedings of the international Conference on Nuclear Data for Science and Technology - ND2004, Sept. 26-Oct. 1,2004, Santa Fe, USA 13. Palmiotti et al. "Proposal for Nuclear Data Dispersion Matrix", /JEF/DOC1063_revl, jan 2005 14. IAEA NDS "thorium cycle nuclear data evaluation" 15. C. Andrieux, Notice d'identification des bibliotheques CEA93.V6 a 172 et 99 groupes, CEA report, SERMA/LENR/RT/99-2724/A, 1999

112

16. "SCALE: A Modular Code System for Performing Standardized Computer Analysis for Licensing Evaluation," NUREG/CR-0200, Rev. 6(ORNL/NUREG/CSD-2/R6), Vols. I, II, and III, May 2000 17. M. Salvatores."La theorie des perturbations et les analyses de sensibilite", Bulletin de la direction des etudes et recherches, EdF, Serie A, N°l, 1988. 18. S. Louviere, R. Sanchez, M. Coste, A. Herbert, Z. Stankovski, APOLL02 Twelve Years Later, International Conference on Mathematics and Computation, Madrid, Spain, 1999. 19. Manturov G.N., Nikolaev M.N., Tsiboulia A.M., ABBN-93 group Data Library. Part 1: Nuclear Data for the Calculation of Neutron and Photon Radiation Fields. - INDC(CCP)-409/L, p.65-110, IAEA, Vienna (1997).

INTEGRAL REACTOR PHYSICS BENCHMARKS THE INTERNATIONAL CRITICALITY SAFETY BENCHMARK EVALUATION PROJECT (ICSBEP) AND THE INTERNATIONAL REACTOR PHYSICS EXPERIMENT EVALUATION PROJECT (IRPHEP) J. BLAIR BRIGGS AND DAVID W. NIGG Idaho National Laboratory,

2525 North Fremont, Idaho Falls, ID 83415-3860,

U.S.A.

ENRICO SARTORI OECD Nuclear Energy Agenc, 12 boulevard des lies, F-92130

Issy-les-Moulineaux

Since the beginning of the nuclear industry, thousands of integral experiments related to reactor physics and criticality safety have been performed. Many of these experiments can be used as benchmarks for validation of calculational techniques and improvements to nuclear data. However, many were performed in direct support of operations and thus were not performed with a high degree of quality assurance and were not well documented. For years, common validation practice included the tedious process of researching integral experiment data scattered throughout journals, transactions, reports, and logbooks. Two projects have been established to help streamline the validation process and preserve valuable integral data: the International Criticality Safety Benchmark Evaluation Project (ICSBEP) and the International Reactor Physics Experiment Evaluation Project (IRPhEP). The two projects are closely coordinated to avoid duplication of effort and to leverage limited resources to achieve a common goal. A short history of these two projects and their common purpose are discussed in this paper. Accomplishments of the ICSBEP are highlighted and the future of the two projects outlined.

1.

History of the ICSBEP and the IRPhEP

The Criticality Safety Benchmark Evaluation Project (CSBEP) was initiated in 1992 by the United States Department of Energy. The purpose of the CSBEP was to identify, verify, evaluate, and formally document a comprehensive and internationally peer-reviewed set of criticality safety benchmark data that could be used for the validation of neutronics codes and nuclear cross section data. Early in the project, the importance of identification and estimation of experimental uncertainties became apparent and was included in the project objectives. It was recognized at the beginning that this project would

113

114

significantly reduce the time, money, and resources expended at the numerous non-reactor nuclear facilities; however, the magnitude of the reduction has far exceeded early expectations. The CSBEP became an official activity of the Organization for Economic Cooperation and Development (OECD) - Nuclear Energy Agency (NEA) Nuclear Science Committee (NSC) in 1995 and the name was changed to the International Criticality Safety Benchmark Evaluation Project (ICSBEP). Representatives from the United States, United Kingdom, France, Japan, Russian Federation, Hungary, Republic of Korea, Slovenia, Serbia and Montenegro, Kazakhstan, Spain, Israel, Brazil, Czech Republic, and Poland are now participating. There are four general types of experimental measurements that have relevance to criticality safety: (1) measurement of critical assemblies, (2) measurement of subcritical assemblies, (3) criticality alarm and shielding measurements, and (4) fundamental physics measurements such as integral measurements of neutron leakage, scattering, and absorption (e.g., NIST iron and water sphere or LLNL pulsed sphere measurements). The ICSBEP has focused primarily on critical assemblies of fissile material; however, some effort has been devoted to subcritical measurements. Future focus of the ICSBEP includes the evaluation of all four types of experiments. The data provided by the ICSBEP are intended primarily for criticality safety practitioners to validate their safety analysis tools; however, the data are also of great value for training, range of applicability determinations, experiment design, nuclear data refinement, and validation and verification by analytical methods development groups. The International Reactor Physics Experiments Evaluation Project (IRPhEP) was initiated by the OECD NEA's NSC in June of 2002, after three years of pilot activities. The IRPhEP focus is on the derivation of internationally peer reviewed benchmark models for several types of integral measurements, in addition to the critical configuration. While the benchmarks produced by the IRPhEP are of primary interest to the Reactor Physics Community, many of the benchmarks can be of significant value to the Criticality Safety and Nuclear Data Communities. Benchmarks that support the Generation-IV Very High Temperature Reactor (VHTR), for example, also support fuel manufacture, handling, transportation, and storage activities and could challenge current analytical methods. The IRPhEP is patterned after the International Criticality Safety Benchmark Evaluation Project (ICSBEP) and is closely coordinated with the ICSBEP in order to avoid duplication of effort and to effectively utilize available resources.

115

The first IRPhEP technical review meeting was held in October of 2004. The next IRPhEP Meeting is scheduled for November 2005 with the first publication scheduled for first quarter of 2006. The work of the ICSBEP and the IRPhEP is depicted graphically in Figure 1.

Criticality Safety and Reactor Physics " Measurement Data

• ICSBEP and IRPhEP -

Externally Available ^ H W . Technical Journals & Reports Reactor Design

Internal reports g ^ ^ Letters & Memos ^ H f ~~"s Logbooks L U

Reactor Licensing

If

Drawings ^

Peer Review (National and International Experts)

Experimenter's Annotated __ Copy of Published Reports Experimenters (Retired or Working on Other Projects)

O

i/ * yr

Analytical Methods Development Nuclear Data Refinement

^

Comprehensive Source of Externally Peer Reviewed Criticality Safety and Reactor Physics Data

Criticality Safety Analysis Range of Applicability and Experiment Design

Figure 1. Graphic Representation of the Work of the ICSBEP and the IRPhEP.

2. Common Purpose of ICSBEP and IRPhEP The ICSBEP and the IRPhEP share a common purpose to: • Identify and evaluate a comprehensive set of benchmark data, • Verify the data, to the extent possible, by reviewing original and subsequently revised documentation, and by talking with the experimenters or individuals who are familiar with the experimenters or the experimental facility, • Evaluate the data and quantify overall uncertainties through various types of sensitivity analyses, • Compile the data into a standardized format, • Perform calculations, when appropriate, using standard criticality safety and reactor physics codes and data, and • Formally document the work into a single source of verified, extensively peer reviewed benchmark data. Short and long-term preservation steps are also taken when data are identified as at risk of being lost.

116

3.

Accomplishments of the ICSBEP

The work performed by the ICSBEP is documented in an OECD NEA handbook entitled, "International Handbook of Evaluated Criticality Safety Benchmark Experiments" (Ref. 1). The 2004 Edition of the Handbook (Figure 2) spans over 30,000 pages and contains benchmark specifications for 3331 critical or near critical configurations. Approximately 554 additional experimental configurations are evaluated, but are categorized as unacceptable for use as benchmark experiments.

Figure 2. The 2004 Edition of the ICSBEP Handbook

Over 250 scientists from around the world have combined their efforts to produce the "International Handbook of Evaluated Criticality Safety Benchmark Experiments." As a result of these efforts, a large portion of the tedious and redundant research and processing of critical experiment data has been eliminated. The necessary step of validating computer codes with benchmark critical data is greatly streamlined, and valuable experimental data are preserved. The work of the ICSBEP has highlighted gaps in data, has retrieved lost data, and has helped to identify inadequacies in basic nuclear data and cross section processing codes. The Handbook is currently being used in 59 different countries. 4.

Quality Assurance

Each experiment evaluation included in ICSBEP Handbook undergoes a thorough internal review by the evaluator's organization. Internal reviewers verify:

117

• • • • •

The accuracy of the descriptive information given in the evaluation by comparison with original documentation (published and unpublished), That the benchmark specification can be derived from the descriptive information given in the evaluation, The completeness of the benchmark specification, The results and conclusions, and Adherence to format.

In addition, each experiment undergoes an independent peer review by another working group member at a different facility. Starting with the evaluator's submittal in the appropriate format, independent peer reviewers verify: • That the benchmark specification can be derived from the descriptive information given in the evaluation, • The completeness of the benchmark specification, • The results and conclusions, and • Adherence to format. A third review by the Working Group verifies that the benchmark specifications and the conclusions are adequately supported. ICSBEP Benchmarks are widely recognized by regulators. The same level of quality assurance is planned for the IRPhEP. 5.

Contents of an IRPhEP Evaluation

5.1. Identification and Types of Measurements Each experiment has a unique identifier that consists of two parts. Part 1 consists of the Reactor Name, Reactor Type, Facility Type and a Three Digit Numerical Identifier. Part 2 of the identifier includes the Measurement Type(s). Identifiers take the following form: (Reactor Name)-(Reactor Type)-(Facility Type)-(Three-Digit Numerical ID) (Measurement Type(s). Tables 1 and 2 give the Identifier elements and their meanings. Table 1. Identifiers for facility types FACILITY Type Experimental Facility EXP Power Reactor POWER Research Reactor RESR

118 Table 2. Identifiers for Reactor and Measurement types REACTOR TYPE MEASUREMENT TYPE Pressurized Water Reactor PWR Critical Configuration VVER VVER Reactors Subcritical Configuration Boiling Water Reactor BWR Buckling & Extrapolation Length Liquid Metal Fast Reactor LMFR Spectral Characteristics Gas Cooled (Thermal) GCR Reactivity Effects Reactor Gas Cooled (Fast) Reactor GCFR Reactivity Coefficients Light Water Moderated LWR Kinetics Measurements Reactor Heavy Water Moderated HWR Reaction-Rate Distributions Reactor Molten Salt Reactor MSR Power Distributions RBMK Reactor RBMK Nuclide Composition Fundamental FUND Other Miscellaneous Types of Measurements

CRIT SUB BUCK SPEC REAC COEF KIN RRATE POWDIS ISO MISC

5.2. Format The format for IRPhEP evaluations is patterned after the format used by the ICSBEP. The general format is: (1) describe the experiments, (2) evaluate the experiments, (3) derive benchmark specifications, and (4) provide results from sample calculations. Code and cross section information, including typical input lists, are provided in Appendix A. Additional information may be provided in subsequent appendices. The format is the same for all evaluations. Seldom, if ever, are all types of measurements made in a particular series of experiments. However, sections for all measurement types are retained in the format and it is simply stated, when applicable, that no such measurements were made. A detailed IRPhEP Evaluation Guide (Ref. 2) can be obtained on the following two Internet Sites: http://nuclear.inl.gov/programs.shtml http://www.nea.fr/lists/irphe/ A brief description of the format is provided in this section. The ICSBEP format (Ref. 2) for critical or subcritical measurements is very well known. Except for the expansion to include other types of measurements, there is only one minor difference between the two formats, a specific section for temperature has been added to Section 1. The types of information and format presentation are essentially the same for each measurement type. Therefore, the details of each subsection is only stated once.

119

SECTION 1.0- Detailed Description: A detailed description of the experiments and all relevant data are provided in the appropriate subsections within this section. The detailed description includes the measurement methods used and the results obtained. Enough information should be given in this section so that the derivation of benchmark-model specifications in Section 3.0 is evident. In general, modeling (idealization, simplification) of the experiment is not discussed here. However, if the exact experimental configuration is unknown (was not reported) or was too complicated to describe in detail and an idealization was provided by the experimenters, then the idealized experiment may also be discussed here, as well as in Section 3.1. Any discussion of an idealized experiment includes an explanation of the assumptions used in going from the real experimental configuration to the idealization. Sources of the data should be indicated. Sources of data include published reports, logbooks, photographs, memos or other records provided by experimenters, and discussions with experimenters. Any inconsistencies in the data are mentioned in this section. A justification as to why the data can still be used is provided in the Evaluation of Experimental Data section (Section 2.0). Uncertainties in the measurements that were assigned by the experimenters, either in published or unpublished (e.g., logbooks) sources, should be included. Details of the main features of an experiment given in Section 1.1 for the critical and / or subcritical configurations are often the same for all other types of measurements. It is not necessary to repeat this information in each subsequent section. However, additions and modifications to the geometry and additional materials that are introduced for each particular measurement type must be described in detail in the appropriate subsections. Descriptive information is provided for each measurement type in the appropriate section, Sections 1.1 through 1.10. In general, the descriptive information includes an overview of the experiment (Section l.X.l), description of the experimental configuration (Section 1.X.2), description of material data (Section 1.X.3), temperature information (Section 1.X.4), and additional information that is relevant to the type of measurement. Drawings and sketches should be used liberally. SECTION 1.1- Description of the Critical and / or Subcritical Configuration: This section contains a detailed description of any critical and / or subcritical measurements that were performed. Uncertainties in the measurements assigned by the experimentalists, either in published or unpublished (e.g., logbooks) sources, should be included. Subsections 1.1.1 through 1.1.5 should contain an overview of the measurements, a description of

120

the geometry of the experimental configurations, a description of the material data, temperature data, and additional information relevant to the critical and / or subcritical measurements, respectively. Detailed descriptions of the methods used to obtain the data should be included in the appropriate subsections. SECTION 1.2- Description of Buckling and Extrapolation Length Measurements: This section contains a detailed description of any buckling and/or extrapolation length measurements. SECTION 1.3- Description of Spectral Characteristics Measurements: This section contains a detailed description of any measurements made to determine spectral characteristics such as neutron spectra or 238Uc/235Uf ratios. SECTION 1.4- Description of Reactivity Effects Measurements: This section contains a detailed description of measurements such as control-rod worth, void effects, small sample worth, fuel substitution, and xenon effects. Values of parameters that were actually measured should be given in this section as well as specific data that were used to transform measured values into other parameters, such as group parameters of delayed neutrons. A clear distinction should be made between measured values, calculated values, and data that were used to process measured results. SECTION 1.5- Description of Reactivity Coefficient Measurements: This section contains a detailed description of measurements such as the temperature coefficient of reactivity 3p/3T, the moderator-height coefficient of reactivity 8p/dH. and soluble boron worth 3p/8CB. SECTION 1.6- Description of Kinetics Measurements: This section contains a detailed description of measurements such as decay constants, peff, or prompt neutron lifetime. SECTION 1.7- Description of Reaction-Rate Distribution Measurements: This section contains a detailed description of reaction rate measurements such as flux maps, fission chamber scans, and wire activation fine-structure and macro-structure measurements. SECTION 1.8- Description of Power Distribution Measurements: This section contains a detailed description of power distribution measurements. SECTION 1.9- Description of Isotopic Measurements: This section contains a detailed description of isotopic measurements of discharged fuel.

121

SECTION 1.10- Other Miscellaneous Types of Measurements: This section contains a detailed description of other miscellaneous types of measurements that do not fit directly into one of the other categories such as conversion or breeding ratio measurements. SECTION 2.0- Evaluation Of Experimental Data: Missing data or weaknesses and inconsistencies in published data are discussed for each measurement type in Sections 2.1 through 2.10. The effects of uncertainties in data on the measured parameters are discussed and, if practical, quantified. All codes and data used for calculations of the effects of uncertainties should be specified. Use of data with large uncertainties or data that require assumptions on the part of the evaluator is justified in this section. If all or part of the data is found to be unacceptable for use as benchmark data, this fact is noted, and the reasons summarized. Unacceptable data are not included in Sections 3.0, 4.0, and Appendix A. If data are provided in Section 1 for a certain measurement type, but have not yet been evaluated, it is so stated in the appropriate evaluation subsection. SECTION 3.0- Benchmark Specifications: Benchmark specifications provide the data necessary to construct a calculational model that represents the important aspects of the experiment. Data that are determined to be acceptable as benchmark-model data are provided in Sections 3.1 through 3.10. In general, the benchmark-model specifications include a description of the calculational methodology (Section 3.X.1); dimensions (Section 3.X.2); material data (Section 3.X.3); temperature data (Section 3.X.4); and the experimental value of each parameter and the benchmark-model value of each parameter with the associated uncertainty (Section 3.X.5). Schematics of the benchmark models should always be included. The benchmark specifications should retain as much detail as necessary to model all important aspects of the actual experiment. When it is necessary or desirable to simplify the representation of the experiment for the benchmark specifications, the benchmark specifications must include the transformations from the measured to the benchmark-model values and the uncertainties associated with these transformations. The transformation and associated bias are addressed in Section 3.X.I. SECTION 4.0- Results of Sample Calculations: Calculated results obtained with the benchmark-model specification data given in Section 3.0 are tabulated for each measurement type in the appropriate subsection, Sections 4.1 through 4.10. Details about the calculations, including code versions, cross sections, and

122

typical input listings, are given in Appendix A (A.l through A. 10). Results should be reported both as obtained directly from calculations and in the form 100(C-E)/E, where C is the calculated result and E is the expected result from a calculation with the benchmark model as given in Section 3.X.5. Benchmark uncertainties should be repeated as percentages for comparison purposes. SECTION 5.0- References: All published documents referenced in the evaluation that contain relevant information about the experiments are listed. Internal documents such as logbooks, memos and internal reports should be included in footnotes. Handbooks and computer code documentation should also be included in footnotes. When a primary reference, internal or published, is available in electronic form, it may be included on the CD or DVD with a hyperlink from the point of reference. APPENDICES: Supplemental information that is useful, but not essential, to the derivation of the benchmark specification or the sample calculations is provided in appendices. Appendices are labeled using letters (e.g., Appendix A). Appendix A is reserved for a description of the codes, cross section data, and typical input listings used in the sample calculations whose results are given in Section 4. Other appendices may be added, as needed, after Appendix A. APPENDIX A: Computer Codes, Cross Sections, and typical input listings: Appendix A provides a description of the codes, options, and cross section data used in the calculations of the results given in Section 4. The following information should be included in Appendix A for each measurement type, X: A.X.I Name(s) of code system(s) used. A.X.2 Bibliographic references for the codes used. The format should be followed, but where certain information or data are determined to be "Not applicable", "Not available", or "Not Significant" it should be so stated. A.X.3 Origin of cross-section data. Nuclear data libraries that were used in the evaluation such as JEF-2.2, ENDF/B-VI, JENDL-3.2 should be specified. Deviations from standard libraries, (e.g., mix of different libraries, details) should be described. A.X.4 Spectral calculations and data reduction methods used. Describe calculational scheme, through a figure and explanatory words that provide essential details about assumptions made such as: Resonance shielding: specify method(s), energy range(s), the nuclides affected (actinides, clad, fission products, oxygen), and which unresolved resonance treatment is used; Describe how mutual shielding (overlapping of resonances) is handled, or not; Fission spectra: specify whether only a single spectrum was used or a weighted mix from all fissile nuclides,

123

explaining the procedure for obtaining the weighted mix; Describe how the (n,2n) reaction was treated (Optional); Weighting spectrum for scattering matrices, e.g., corrections of the out-scatter and self-scatter terms considering the differences between the original weighting spectrum and the actual spectrum (Optional). A.X.5 Number of energy groups or if continuous-energy cross sections are used in the different phases of the calculation. A.X.6 Component calculations. The following information should be provided for each component calculation (pin cell, assembly, etc.) as well as full core calculations: Type of cell calculation (pin cell, assembly, etc.); Geometry; Theory used (diffusion, transport); Method used (finite difference, finite element, nodal, Sn (order), collision probability, Monte Carlo, J+/-, etc.); Calculation characteristics (meshes, elements/assembly, meshes/pin, number of histories, multi-group, continuous energy, etc.). A.X.7 Other assumptions and characteristics. A.X.8 Typical Input Listings for each code system type. Typical input listings used to obtain the results reported in Section 4.0 should be provided. Unique and/or important features of the input may also be discussed just prior to the input listings. Listing titles refer to the case number and number of the table in Section 4.0 that gives the calculated result. 6. Technical Review and the Technical Review Group A Technical Review Group was organized during 2004 and the first IRPhEP Technical Review Meeting was held on October 27 and 28, 2004. Representatives from France, Germany, Hungary, Japan, Russian Federation, United Kingdom, United States and the OECD NEA participated. Eleven evaluations were reviewed in detail at the October meeting. None of the eleven were judged to be ready for final approval and publication, but actions were assigned for each that should, if completed, ensure approval at the 2005 Technical Review Meeting. The learning curve for the IRPhEP should be greatly accelerated by drawing on the experience of the ICSBEP. To put things into prospective, the first ICSBEP Meeting was held in November of 1992, the first formal publication was March of 1995. 7.

Coordination between the ICSBEP and the IRPHEP

A protocol for coordination between the ICSBEP and the IRPhEP was established to maximize the benefits from the efforts of both and to avoid duplication of effort. When data for other measurement types are added to an existing ICSBEP evaluation by the IRPhEP, the original ICSBEP evaluation is

124

incorporated into IRPhEP documents by reference only with a hyperlink to the actual evaluation report that will be duplicated on IRPhEP electronic publications. Similarly, data that are reviewed and approved by the IRPhEP that are of common interest to the ICSBEP will be referenced (an ICSBEP identifier will be assigned and a short summary of the data will be provided) and duplicated on ICSBEP electronic publications (a link to the IRPhEP evaluation will be included in the summary). Major errors, omissions, or duplications that are identified by either group will be formally transmitted to the evaluator for consideration. 8.

The First (2006) Edition of the International Handbook of Evaluated Reactor Physics Benchmark Experiments

The first IRPhEP publication is scheduled for the spring of 2006. It will be published on electronic media (CD-ROM or DVD), as an OECD NEA Handbook entitled, "The International Handbook of Evaluated Reactor Physics Benchmark Experiments". A summary of the experiments that are expected to be reviewed and considered for publication in the 2006 Edition of the IRPhEP Handbook are given in Table 3. Several of these evaluations are of interest to Generation-IV, such as the Initial Critical Configuration for the HTR-10 Pebble Bed Reactor (HTR10GCR-RESR-001-CRIT), measurements made on the Block Type High Temperature Engineering Test Reactor (HTTR-GCR-RESR-001-CRIT-REACCOEF-POWDIS) and the Very High Temperature Reactor Critical Assembly (VHTRC-GCR-RESR-001-CRIT-COEF). (Note, VHTRC data may only be made available through the Generation-IV International Forum.) Other benchmarks of interest include a critical assembly of a full scale model of a BN600 reactor hybrid (U02 + MOX) core (BFS2-LMFR-RESR-001-CRIT-SPECREAC-RRATE) and a model of a sodium-cooled fast reactor fueled by uranium (18.5% U-235) metal (BFS1-LMFR-RESR-001-CRIT-SPEC-COEF-RRATE). Several experiments have already been identified for future evaluation. Those experiments are given in Table 4. An evaluation that should be of particular interest to the Generation-IV and nuclear data communities is an extension of the ZPR-9 Uranium/Iron Benchmark for which the critical configurations has already been published by the ICSBEP (HEU-MET-INTER001). Other planned evaluations of interest to Generation IV include pebble bed data from ASTRA, PROTEUS, and AVR.

125 Table 3. IRPhEP IDENTIFIER ZEBRA-LMFR-RESR-001 CRIT-SPEC-REAC-RRATE DMPLE-RESR-EXP-001 CRIT-BUCK-SPEC-RRATE (LEU-COMP-THERM-048) DIMPLE-RESR-START-002 CRIT-BUCK-SPEC-RRATE (LEU-COMP-THERM-055) CROCUS-LWR-RESR-001 CRIT-KIN PFACILITY-VVER-EXP-001 CR1T-RRATE (LEU-COMP-THERM-061)

VENUS-PWR-RESR-001 CRIT-BUCK-RRATE-POWDIS VENUS-PWR-RESR-002 CRIT-BUCK-RRATE-POWDIS VENUS-PWR-RESR-003 CRIT-BUCK-RRATE-POWDIS ZR6-VVER-EXP-001 CRIT-BUCK-SPEC-REAC-COEF-RRATE (LEU-COMP-THERM-015 and 036) VHTRC-GCR-RESR-001 CRIT-COEF KRITZ-RESR-EXP-001 CRIT-BUCK-REAC-RRATE HTR10-GCR-RESR-001 CRIT HTTR-GCR-RESR-001 CRIT-REAC-COEF-POWDIS ZPPR-LMFR-RESR-001 CRIT-REAC-RRATE JOYO-LMFR-RESR-001 CRIT-REAC-COEF DCA-HWR-RESR-001 CR1T-BUCKLING BFS2-LMFR-RESR-001 CRIT-SPEC-REAC-RRATE BFS1-LMFR-RESR-001 CRIT-SPEC-COEF-RRATE

itions in Progress for 2006 SUMMARY Fast Critical Experiments in Plate and Pin Geometry Form. The ZEBRA CADENZA Cores, Assemblies 22, 23, 24 and 25. Light Water Moderated and Reflected Low Enriched Uranium (3 wt.% 235U) Dioxide Rod Lattices DIMPLE SOI Light Water Moderated and Reflected Low Enriched Uranium (3 wt.% 235U) Dioxide Rod Lattices DIMPLE S06 Kinetic Parameters and Reactivity Effect Experiments in CROCUS VVER Physics Experiments: Hexagonal (1.27vm Pitch) Lattices of U(4.4 wt.% 235 U)0 2 Fuel Rods In Light Water, Perturbed by Boron, Hafnium, or Dysprosium Absorber Rods, or by Water Gap With/Without Aluminium Tubes VENUS-2 PWR MOX Core Measurements VENUS-1 PWR U0 2 Core 2-Dimensional Benchmark Experiment VENUS-3 PWR U0 2 Core 3-Dimensional Benchmark Experiment The VVER Experiments: Regular and Perturbed Hexagonal Lattices of Low-Enriched U0 2 Fuel Rods in Light Water Very High Temperature Reactor Critical Assembly (VHTRC) Temperature Coefficient Benchmark KRITZ-2:19 Experiment on Regular H20/Fuel Pin Lattices with Mixed Oxide Fuel at Temperatures up to 245°C Evaluation of the Initial Critical Configuration for the HTR-10 Pebble Bed Reactor Evaluation of the High Temperature Engineering Test Reactor JNC Large fast reactor experiment ZPPR-10A in JUPITER JNC Experimental Fast Reactor JOYO Mk-I core physics tests JNC Heavy water core critical experiment, DCA Critical Assembly of a Full Scale Model of a BN-600 Reactor Hybrid (U02 + MOX) Core (BFS-62-3A) Model of a Sodium-Cooled Fast Reactor Fueled by Uranium (18.5% U-235) Metal

126 Table 4. IRPhEP Evaluations in Progress for 2007 and Beyond Summary Identifier IPEN/MB01-RESR-EXP-001 The Isothermal Experiment of the IPEN/MB-01 COEF Reactor VENUS-PWR-EXP-004 Experimental Study of the VENUS CRIT-BUCK-RRATE-POWDIS Configuration No. 7 VENUS-PWR-EXP-005 Experimental Study of the VENUS Configuration No. 9 CRIT-BUCK-RRATE-POWDIS Experimental Study of the VENUS VENUS-PWR-EXP-006 Configuration No. 17 CRIT-BUCK-RRATE-POWDIS ASTRA Critical Facility Experiments ASTRA-HTGR-EXP-001 CRIT-REAC-RRATE TER-2 in LWR U0 2 with Soluble Poisons TER2-LWR-EXP-001 Reactivity Worth Measurements and Other STEK-LMFR-EXP-001 Experiments in the Critical Facility STEK CRIT-SPEC-REAC-RRATE Evaluation of HTR- PROTEUS PROTEUS-GCR-RESR-001 CRIT-SPEC-REAC-COEF-KIN-RRATE AVR-GCR-RESR-001 Evaluation of the AVR High Temperature CRIT-COEF-KIN Reactor ZPR9-LMFR-RESR-001 Evaluation of the Uranium Iron Benchmark CRIT-SPEC-REAC-KIN-RRATE-MISC (HEU-MET-INTER-001)

9. Archival of Primary Documentation Since the inception of the IRPhEP, the NEA has been collecting primary documentation and has been transforming those documents into electronic form to facilitate data retrieval and dissemination. An archive of those documents has been established at the NEA and contains the following: - IRPHE/B&W-SS-LATTICE, Spectral Shift Reactor Lattice Experiments - IRPHE/ZEBRA, AEEW Fast Reactor Experiments - IRPHE/JOYO MK-II, core management and characteristics database - IRPHE/JAPAN, Reactor Physics Experiments carried out in Japan - IRPhE/HTR-ARCH-01, Archive of HTR Primary Documents - IRPHE-SNEAK, KFK SNEAK Fast Reactor Experiments - IRPhE/STEK, Experiments from Fast-Thermal Coupled Facility - IRPhE-DRAGON-DPR, OECD High Temperature Reactor Dragon Project - IRPhE/RRR-SEG, Experiments from Fast-Thermal Coupled Facility - Experiments in VENUS- Project on the Physics of Plutonium Recycling - IRPHE/AVR, AVR - Experimental High Temperature Reactor

127

10.

Conclusions

The ICSBEP has eliminated a large portion of the tedious and redundant research and processing of experimental data and has greatly streamlined the validation process. The project has also significantly increased the number of benchmarks available for nuclear data testing. Benchmarks produced by the IRPhEP will provide new dimension to validation efforts and will greatly expand the collection of available integral benchmarks for nuclear data testing and uncertainty determination. The first publication of the "International Handbook of Evaluated Reactor Physics Benchmark Experiments" is scheduled for the first quarter of 2006. Acknowledgments The authors would like to acknowledge the efforts of all past and present contributors to the IRPhEP Executive Committee and Technical Review Group. References 1. "International Handbook Of Evaluated Criticality Safety Benchmark Experiments", NEA/NSC/DOC(95')03/I-VIII. OECD-NEA, September, 2004. 2. "International Reactor Physics Experiments Evaluation Project Evaluation Guide", NEA/NSC/DOC(2005)2.

COMPUTER MODEL OF AN ERROR PROPAGATION THROUGH MICRO-CAMPAIGN OF FAST NEUTRON GAS COOLED NUCLEAR REACTOR EVGUENIIVANOV Russian Research Center "Kurchatov Kurchatov sq, 1, 123182, Moscow,

Institute", Russia

Russian conceptual design development of Gas Cooled Reactor with Fast Neutron spectra (FGR) positions it in kind of source of energy which has capabilities to be used in artificial energy carrier (hydrogen) production and in fissile material breeding. Uncertainty analysis for it is a part of the conceptual design development. Uncertainties of nuclear reactor parameters conventionally can be divided onto two parts where one of them is connected with criticality modeling and the next one - with neutron field and nuclides concentration evolution. It is very conventional division but can be useful for comparisons of individual sources of uncertainty. The evolution of the field of nuclide concentrations in materials of the studied system is described by a dynamic model in lumped parameters. The coefficients of the matrix of the nuclide transmutations are determined from calculation of the steady-state neutron and photon fields and are normalized taking into account the complete capacity of the plant.

1. Introduction Overall world tendency for modeling of the evolution of nuclear systems (including common types of reactors) is preceded through increasing of the part of complex models built from the "first principles" with detailed description of partial physical processes in engineering systems and elements. In the development of operating conditions, construction or very concept of power plant it is necessary to solve immediately two forecast problems. This, in the first place, the prediction of the behavior of installation taking into account possible deviations from the basic operating mode, and, in the second place, the determination of the criteria of the transfer of the results of model studies and studies on the prototypes of installation to the characteristics of project. Solution of these problems requires the substantiation of the accuracy of the prediction of the state of nuclear reactor, especially those states, which cannot be observed under the conditions for finalizing installation on the stands and in the laboratories. The campaign of nuclear reactor is usually represented by dynamic model with the sequential calculation of states. In this case each stage of

128

129

simulation is characterized by the small, but final error, the study by which must compose the noticeable part of the calculated works. The determination of the quality of the calculated code by comparison with the experiment is also nontrivial task, since given and the results of experiment must be processed correspondingly in order to satisfy the requirements for benchmark of model accuracy, adequacy, the estimated uncertainty of the conditions for experiment and the controlled error. Thus, the studies of questions of the simulation of the campaign of nuclear reactors generate the following tasks: - the development of methods and program set for the estimations of the uncertainty of experimental data as a result of technological deviations and simplification in the design models; - the development of the criteria of the transference of given, obtained to some systems, to others or the development of the methods of experimental studies on the reduced temporary base; - the evaluation of the influence of errors in nuclear data on the results of simulation and the solution of reverse problem production of the standards of data for the calculations. The reactivity margin is the determining parameter for any nuclear reactor (at least, from the point of view of the substantiation of safety). The error of the critical parameter of nuclear reactor at a certain moment of time is determined by two components: Sktf = Sksl + <5klr, where &kt/ is the complete variation of multiplication factor; dkal is the error of the immediate determination of multiplication factor and Sklr is the error determined by variations of concentrations of basic nuclides. This dependence can be represented as the form of summarizing of contributions from two sources of the error:

Sk "

=

y * . & + y y *..*.&,

h8e

a)

^f^Sp Se

where 8k is the functional derivation in the stationary reactor; 8p is the deviation as a result of the error of the concentrations of nuclides and 5e is the changing of nuclide's concentration via parameters. For the estimation of error the given values are calculated, the procedure of which is represented in the present work. At studies of nuclear reactors are evaluated parameters enumerated below, for which are determined the deviations, caused by small variations in material densities: • the field distribution of power yield distribution; • neutron spectra and the distribution of the various rates of nuclear reactions; • changing of the nuclides composition in the process of burning out; • critical charge and the reactivity margin.

130

The estimations of variations are accomplished with the aid of the sensitivity indices, and the procedure of their calculation is based on the following principles: • the direct simulation of the deviations of basic reactor functionals without the use of adjoint functions; • it is assumed that the deviations are small, i.e., they are relied by those not correlated at the level of nuclear data and influences on the reactivity; • for the analysis of the influence of the deviations pointed out above are used the methods of slight disturbances in the evolutionary task of the simulation of the transformations of nuclides in the course of campaign 151. 2. An algorithm of an error propagation The model of the transformations of nuclides is described by the system of ordinary differential equations (model in lumped parameters) where the solution can be presented in operator form: ftH - p)k< . p)k< ... p)
ft)°(

n)

Since the operators are piecewise constant, we have for the relative variation with respect to the m-th parameter: ••• P

-XP

dsm +

pm

-N

dem dP{k~l) dem

u r x

x p ( i - i ) ...pm

. jyC) +p<*> X / J O - D . . . P 1

Mm de

(4)

Algorithm has been realized in code PATRICK 151 which is intended for solving of various problems associated with variations in nuclide composition of materials in nuclear systems. Deviation of nuclide concentration is determined by condensed reaction rate and neutron flux. But they are dependent from reactor capacity and, therefore, formulae of derivation has been modified by including of derivation of nuclides' vector by total power yield: dEn Ss„

de_

9

?'

<5>

where E n is virtual nuclide with concentration proportional to integral of power in physical zone.

131

3. The way of small deviations evaluation There are different methods of the analysis of small deviations with respect to different functionals of neutron field III; however, in the present work is used the direct calculation of deviations with the only difference that the was conducted the approximation of the obtained results and the derivative of the obtained functional dependence with the tendency of disturbance toward zero was calculated. The use of adjoint functions IM is justified by the fact that from the dependences of the deviations of various parameters, on the basis of the conditions of orthogonality, it is possible to exclude the members of the second order of smallness. However, the combined solution (value) is determined for a certain predetermined functional (actually this projection on the phase space of smaller dimensionality), but for each specific case the combined vector must be recounted (and reformulated equation). It is necessary to ensure the true (algebraic) coupling of the solution (condition of orthogonality for the scalar product). The condition of algebraic coupling makes it possible to obtain the conditions of orthogonality. In particular, for the conditionally- critical task, this condition pours out the dependence of the form: (%,fi.$J)SS,yCl=SIJ-CJ,

(6)

where *P, is the i-th harmonic of value, F is the fission operator, $ y is the j th harmonic of flux; C, is the value of fission neutrons defined for i-th mode. The condition of coupling can be carried out far from always both because of the limitations of the numerical diagram of the solution and because of the very formulation of the problem. In particular with the use of methods for random testing the result of the solution is the collection of the integral characteristics, which regarding cannot have the combined distribution. Obtained data are approximately described by the polynomial of the second order: M / O = «o +«i '(P. -/Vo> + *2 '(P. -A»,o)2 + - .

(7)

where: jt ( p j is the dependence of multiplication factor on concentration of mth nuclide; pm-pma is a change in concentrations if m-th nuclide (supposedly small); {a-} is the collection of dimensionless coefficients is determined numerically. In our case the derivative by the concentration of nuclide is determined as follows:

132

Sensitivity index to a change in the concentration of one or other nuclide or another is determined by simple multiplication by the appropriate concentration: *.=/>-•#-.

(9)

5

Pm

where s is the sensitivity index to the concentration of nuclide; y is the nuclear concentration of nuclide; m is the index of the nuclide (taking into account not only of name, but also zone, where it is placed). The statistical uncertainty of the obtained result also can be evaluated on the basis of the fact that the unknown coefficient is function from the collection of the calculated multiplication factors and collection of the trial concentrations of the nuclide: 8k - — = ax =fun{kQ,kx,k2,pmQ,pml,pm2), tipm

(10)

where k0,kx,k2 is the value of multiplication factor with a variation in the concentrations of nuclides; \p .

D

,

p

A are variations in the

concentrations, in accordance with the formalism of the transfer of error, formula for evaluating the statistical error appears as follows: (8k

da^k^k^k^p

p

pm2)

tipm

**,)

(11)

Thus, with the attraction of the Monte Carlo method, realizes the direct calculation of sensitivity indices and their uncertainty is evaluated. Since calculation was performed directly, then as the parameters, for which were determined the sensitivity indices, they were examined the multiplication factor, capture rate for the separate isotopes on the chosen zones and the field distribution of neutrons. The represented algorithm is realized for changes in the concentrations of nuclides; however, it makes it possible to estimate the influence of small changes in the geometry. For this they are combined a change in the concentrations of nuclides, which are contained in one or other zone or another, and varies the concentration of pseudo- - isotope, which influences blocking resonance absorption. 4. Information flows in PATRICIAN code package For the realization of the transfer of errors is developed code PERT2k2, which consecutively starting of MMKFK-2 varying the concentrations of nuclides,

133

including pseudo-nuclides (Asc and "nuclides", where the section it is substituted by its dispersion). Results it is processed by the code and, thus, PERT2k2+ MMKFK-2 is calculated entire collection of functionals such as the derivative of multiplication factor of small changes in the concentration of each nuclide <&/<£>„ (including sensitivity to the pseudo-nuclides Sk/Sppseudo ), the velocity derivative of processes of small changes in the concentration of each nuclide Sa,l5pm (including sensitivity to the pseudo-nuclides 6alj5ppseuio), a variation in the flux with a change in the concentrations of basic Sq>kjSpm and pseudo-nuclides {SVkjSPpseudo )• (-)n t n e b a s ' s °f t n e obtained coefficients with the aid of code PATRICK is produced the calculation, etc., which makes it possible then to obtain and the deviation of the multiplication factor and distribution pour on emissions SpJ5em, realizing formula Scef = Skst + Sktr • 5. Peculiarities of mathematical model and calculation schemes If the packaged matrix of nuclides transformation in the active core of FGR taken in canonical form Error! Objects cannot be created from editing field codes, describe graphically as on the figure 1 we can see that it is sparse matrix with complex structure of eigen-value spectra.

Figure 1 Structure of operator of nuclide transition through lifetime performance of nuclear fuel in FGRcap

Correspondent to it spectra of eigen-values for two values of the integral flux are presented in the table 1. It is evident that eigen-values will coincide only if will coincide several elements of matrix transition am = yaa •0} + Am+ ratem, what can occur by fitting neutron flux. This case is it. And in real situation complex part of

134

dynamics will be either unpredictable as inevitable. Indeed appearance of complex eigen-values connected with their degeneration due to overlapping of coefficients of nuclide incineration, which in its order is possible by relevant fitting of neutron flux. In this case coincidence of elements in matrix can occurred incidentally and by unexpected way. Thus in entire model of system of nuclear installation behavior we raise oscillations which are lose of any physical content. To avoid such calculation problem we can consider two ways. One of them is analyzing of obtained matrix of dynamical system in each time step and separate treatment of real and imaging parts of matrix exponent. But even with such measures of precaution problems will save in nearest neighborhood of degenerating eigen-values. ible 1 Chan ging of eigen -values of dynamics model via changing of neutron intens: ~1015 ~10 l f Re Im diag diag Re -2.96 10 8 -0.04079 0 -2.96 10 8 -0.00041 u2m3 5 -1227.21 -0.10653 -1227.21 0 -0.00107 u -379.56 -2.96 10"8 0 -379.56 -2.96 10u236 8

u237 U23S

u239

Np237 Np238 Np239 Pu238 Pu239 Pu240 Pu241 Pu242 Pu243 Am241 Am242 Am242m Am243

-37.5177 -119.215 -15515.5 -107.928 -0.20177 -0.17494 -0.0143 -0.03153 -0.07899 -0.07899 -0.06166 -0.10011 -0.10011 -0.04258 -0.04077 -0.10654

-37.5177 -0.0143 -15515.5 -0.08432 -119.215 -107.928 -0.07944 -0.1008 -0.04274 -0.17492 -0.03154 -1227.21 -0.09353 -379.56 -0.20175 -0.06165

0 0 0 0 0 0 0 0 0.00746 -0.00746 0 0.006259 -0.00626 0 0 0

-37.5177 -119.215 -15515.5 -107.821 -0.04939 -0.00861 -0.00653 -0.00252 -0.00014 -0.00032 -0.00087 -0.00071 -0.00052 -0.00102 -0.00107 -0.00041

-37.5177 -0.00014 -15515.5 -0.00084 -119.215 -107.821 -0.00861 -0.00104 -0.00053 -0.04939 -0.00032 -1227.21 -0.00252 -379.56 -0.00653 -0.00071

And there is another way to avoid the canonical representation of matrix exponent by applying of decomposition of branch model on Bayesian chains:

AL(0 = ! > * ( ' > • ic

where:

(12)

135

*4 „ y* exp(-
(13)

+ 9. f l ^

exp(-ory • t)

f[a XX

TtarY[{a,-a,) i*j

J

j=0

is evaluated for each chain. In this case formula (2)-(4) are simply realized by direct derivation of (13) by parameters /PATRICK/. 6. Application to FGR uncertainty analysis View on the role of Fast Neutron Reactors Gas Cooled (BGR - in Russian transcription) in innovative nuclear system - goals of design and requirements to software (including databases and Nuclear Data) dictates the main principles of the design development. And in this reactor from the point of view if entire energy systems' demands it is established the semi-equilibrium fuel cycle with weak grow of capacity [e(t)J X N EQ = l(e{t)i-R-p) -UNAT)

-(e(t)i-R-p)

l

-U'NAT

(14)

where: N - equilibrium nuclide vector [in entrance of the reactor], U - vector of natural uranium (mining), R - operator of nuclide transformation into reactor [internal fuel cycle], P - operator of nuclide vector transformation in storing and reprocessing, I - unit operator. On the figure 2 values of some nuclides in the core of FGR are shown in assumption of semi-equilibrium inventory. It is easy to see that values even in semi-equilibrium case are not constant that is evident because in spite of liner character of the model for main part of dynamic variables (nuclide and neutron fields) relevant them uncertainties behavior can be described only by nonlinear model.

136 Value of nuclide trough campaign

-0.05 Figure 2 Values of HMs (Pu240, Am241 and U238) through lifetime of the fuel loading in FGR

On figure 3 it is shown the hafnium. Hafnium implementation in Fast Neutron Gas Cooled reactor is: (1) the prospective high temperature and radiation resistant cladding material for micro-coated fuel (Zr+Hf)C [unpurified Zr], (2) the "resonance absorber" for mitigation of flooding reactivity. Thus uncertainties raises by hafnium is essentially FGR proper and they should be studied. Relative value of virtual AHf capture trough campaign p/Pst • 0.02 •

0.015 n n0.005 o -0.005 (I -0.0' 0.015 -0.02 -0.025 • -0.03

-V*/dsc(Hf) &/dsc(Hf)Jfy,

Figure 3 Values of hafnium through lifetime of the fuel loading in FGR

7. Conclusions and unresolved problems Burn-up calculation is the result of modeling of non-linear and coupled slow transient processes. Its accuracy is determined by initial [including nuclear] data and by the used numeric procedure (in form of computing model) Therefore the algorithm of sensitivity analysis has been built on the common-type dynamic model:

137

• •

application of virtual nuclides for balances of capacity and masses of nuclides and for simplified representation of uncertainties, accounting of connected with neutron flux changing feedbacks

The algorithm has been realized in codes PATRICK [is used for MonteCarlo burn-up calculations] and PATRICIAN [can be used with some restriction in neutron data requirements analysis]. Co variance matrix has not been used in the PATRICIAN, only diagonal elements (dispersion) had been used and they are divided on rather small number of interval, i.e. on wide energy ranges. The steady-state neutron transport model (Boltzman's equation) is the linear one but reactor behavior with burn-up is described by non-linear models. Where is the border of adequacy of linear response [including GPT] for holistic analysis of uncertainties in the systems? From the one hand, the neglecting of the accounting of feedbacks allow to use correlation matrices of cross section uncertainties, but, form the another one, restrictions in description of uncertainties (only dispersion) allow us to include all interrelations between different processes. What is better? I think that for further development we need in special tools for conformity analysis of integral reactor experiments with burnup and of transferring of operating experience to new designs References 1. A. Frank- Kamenetskiy, M. Yudkevich. Calculation of prompt neutron life time in reactor using Monte Carlo method", Preprint IAE-2155, Moscow, 1971 (in Russian) 2. V.B. Polevoy, et al. "Base Package of Codes MMKFK-2 for Solving Neutron Transport Tasks in Reactor Physic using Monte Carlo Method", OFAP YR #00371, Moscow, 1996 (in Russian). 3. A. Bliskavka, G. Manturov, M. Nikolaev, A. Tsiboulia. CONSYST/MMKKENO Codes for Nuclear Reactor Calculation Using Monte Carlo Method in Multigroups with P 5 approximation. Preprint IPPE2887, Obninsk, 2001(in Russian). 4. Gandini Augusto, Generalized perturbation Theory. Heuristic approach, Advantages in Nuclear Science and Technology, vol. 19, 1987 5. Evgenij Ivanov, Method of estimating the sensitivity of a calculated nuclide vector to deviations of initial data/ Report IAEA(CCP)-418, December

C O M B I N I N G DIFFERENTIAL A N D I N T E G R A L E X P E R I M E N T S ON 2 3 9 P U FOR R E D U C I N G UNCERTAINTIES IN N U C L E A R DATA APPLICATIONS

T. KAWANO, K. M. HANSON, S. C. FRANKLE, P. TALOU, M. B . C H A D W I C K , R. C. L I T T L E Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA E-mail: [email protected]

We present an approach to uncertainty quantification for nuclear applications, which combines the covariance evaluation of differential cross-sections data and the error propagation from matching a criticality experiment using a neutron transport calculation. We have studied the effect on Pu-239 fission cross sections of using a one-dimensional neutron transport calculation with the PARTISN code. The evaluation of Pu-239 differential cross-section data is combined with a criticality measurement (Jezebel) using a Bayesian method. To perform the uncertainty quantification for such calculations, we generate a set of random samples of cross sections, which is representative of the covariance matrix, and estimate the distribution of calculated quantities, such as criticality. We show that inclusion of the Jezebel data reduces uncertainties in estimating neutron multiplicity.

1. Introducton Advanced technologies of nuclear energy applications require uncertainty quantification to assess the margin in a design and to validate model predictions, such as a neutron transport calculation for criticality. When a covariance of nuclear data is provided, the uncertainty in a calculated quantity (e.g., neutron multiplicity, fceff, for example) can be estimated by propagating the error in the nuclear data to the outputs of the application code. In general, calculation of the error propagation is sometimes very difficult to perform, since many application codes, such as a neutron transport Monte Carlo simulation codes, require tremendous computing capabilities. To perform the uncertainty quantification for such calculations within a realistic computational resource (time), a feasible method must be developed. One possible solution is to perform the calculations with several perturbed cross-section sets (ensemble), then look at the distribution of the calculated 138

139

quantities. In this Monte Carlo technique, we must be careful about the fact that the uncertainties in the cross sections are not independent, and this information must be included into the sample set as a constraint. In this study we evaluated the covariance of 2 3 9 Pu fission cross section, because a large experimental database to evaluate the covariance is available, and a straightforward application of 2 3 9 Pu fission cross section such as Jezebel — a critical assembly at LANL 1 ' 2 — makes it possible to validate our Monte Carlo method. The neutron transport calculations are performed with the PARTISN code 3 . Although we limit ourselves to a computationally fast transport calculation to demonstrate our procedure, the Monte Carlo method can be used in any type of application codes.

2. Estimation of Uncertainties in the Cross Sections 2.1. Differential 239

Data

The Pu fission cross sections evaluated at LANL4 are based on measurements of differential data, and the uncertainties (covariance) of the cross sections can be deduced from the experimental uncertainties. In the IAEA standard evaluation 5 , all existing experimental data used in the evaluation were also examined carefully, and covariance for the measurements in each experiment was assigned. We adopt this experimental database, together with the IAEA interim standard cross section, to generate a covariance matrix for 2 3 9 Pu fission cross section. The 30-group structure adopted in the LANL MENDF6 multi-group data library 6 is used. The group covariance is obtained directly from the experimental data by comparing the group-averaged cross sections with the differential measurements. We use a simple Least-Squares fitting procedure with the SOK code 7 . The evaluated uncertainties for the 30-group structure cross sections are shown in Fig. 1 by the solid line. The uncertainties obtained are 1.4% - 2.4% above the unresolved resonance region (the boundary energy is 30 keV), and are reasonable because each data point has typically a 2-6% uncertainty, and this uncertainty is reduced by the group-averaging procedure. The correlation matrix reveals positive correlation coefficients among the group cross sections. In general, the experimental data may have an unknown normalization arising from those systematic errors. This systematic component in the overall uncertainty leads to a positive correlation among the evaluated cross sections.

140 T " I

1 I I I ITT

Differential Data Only Integral Data Included Sensitivity o CD CO

0.75

s I

> 0.5

CD CO

CD

en 0.25 CO t:

o c 3

0 0.001

•

• • • • • « - !

0.01

0.1 1 Neutron Energy [MeV]

10

Figure 1. RMS uncertainties in the 2 3 9 P u fission cross sections. The solid line is the result that includes differential measurements only, the dotted line is the case in which Jezebel integral data are added, and the dot-dashed line shows a relative sensitivity of group-cross sections to feeff value (right axis).

2.2. Integral

Data

It is known that the 2 3 9 Pu fission cross section and average number of prompt neutrons vv in the LANL nuclear data library predict the fceff value from Jezebel very well. To include this information in the cross section covariance, we adopt a Bayesian update technique, which combines information from differential fission data and the integral data. The KALMAN code 8 is used to calculate the covariance matrix. A sensitivity matrix is needed to perform the KALMAN calculation. Frankle 9 obtained the sensitivity matrix of PARTISN calculations for Jezebel fceff by calculating derivatives numerically. The uncertainties calculated with the KALMAN code are shown in Fig. 1 by the dotted line. A relatively large reduction in the uncertainty is seen near 1 MeV. In the other energy regions, however, the reduction is not so large. While the uncertainties in the cross sections in individual groups are only modestly reduced, the biggest benefit of including the integral constraint will be realized for simulations of similar physical setups involving 239 Pu and a hard fast-neutron spectrum. A remarkable change can be seen

141

in the correlation matrix shown in Fig. 2, in which negative correlations appear in the fast energy range. These negative correlations come from the constraint imposed by the integral data.

0.001

0.01

0.1 1 Neutron Energy [MeV]

10

Figure 2. Correlation matrix for the 2 3 9 P u fission cross sections, evaluated based on both differential and integral data. The areas surrounded by solid rectangles represent negative correlation.

3. Uncertainty Quantification of a Simulation Using Monte Carlo The covariance matrices obtained in the preceding section provides a description of the uncertainties in the group cross sections. With these covariances, we can use a Monte-Carlo technique to estimate the uncertainty in simulations of related physical phenomena. The general approach is to draw random samples of cross sections from its probability distribution, and use each sample in the simulation to calculate its associated output. The ensemble of outputs can be used to characterize their uncertainties.

142

The covariance matrix P can be diagonalized as: U- 1 PU = diag(A 1 ,A 2 ) ...,A m ),

(1)

where A's are the eigenvalues of the covariance matrix P, and the matrix U contains the eigenvector (uu,U2i,...,umi,)' for the i-th eigenvalue Aj. In the eigenvector space, there is no off-diagonal element in the covariance matrix, so we sample the eigenvalues & from the Gaussian distributions with the standard deviation of \A7i and the center value of zero. The distribution of & is given by

•<«**-VKS-»(-&)*• with (&) = 0 and (£?) = A*. When the sampled values & are provided, an element of P, ptj cov(<Ti,crj), can be calculated as Pij = ^UikUjkXk

=

k

(2)

=

y^fcMjfc(gfc) k

= (Yl

Uik k

^ Yl, uik^k) •

(3)

\ k k I Here we used (uiki&iUjka&a) = ik1Ujk2(£.k1€k2) = 0 for ki =£ k2. From Eq. (3), X}fcuifc£fc is a deviation from an averaged cross-section when the eigenvalue ^ is sampled : u

o'i = Gi + 5(Ti = Gi + Y^ UikZ,k,

(4)

k

where a\ can be used as the sampled set. We have generated 30 samples to reproduce the uncertainty distribution of fceff, which is shown in Fig. 3. The solid histogram shows the distribution of fceff when only the differential data are included. The dotted histogram represents this distribution when integral data are included. Both distributions can be approximated by a Gaussian of a =0.8% and 0.2%, respectively. The uncertainty in the Jezebel experiment, 0.2%, is well reproduced by this simulation. We also calculated the uncertainties in keff using the standard error propagation technique, and obtained exactly the same answers.

143

30 samples, for fission data in simulation 0.5 0.45

•••••,,••,„. r., |

|

|

r

- | •• -f•••• |

i

i

i — i — | — i — r

i n — | — i

i

i

Using Differential Data Only Differential and Integral Data

0.4 0.35 £

0.3

§ 0.25 o £

0.2 0.15 0.1 0.05

0 0.97

0.98

0.99 1 1.01 1.02 Predicted keff for JEZEBEL

1.03

Figure 3. Distributions of calculated keg values for Jezebel with the PARTISN code. The solid histogram shows the distribution of keg when the prior covariance matrix X is used. The dotted histogram is for the posterior matrix P.

4. Conclusion We presented a Monte Carlo technique to perform uncertainty quantification for nuclear applications, combining the covariance evaluation of differential cross-sections data and the error propagation from matching the Jezebel experiment using the PARTISN neutron transport calculation. The covariance of 2 3 9 Pu fission cross section was estimated from the experimental differential data. The evaluation of 2 3 9 Pu differential cross-section data was then combined with the integral data of the Jezebel critical assembly using the Bayesian method, hence reducing slightly the uncertainties in the fission cross sections. Random samples of the fission cross sections were generated based on the covariance matrices obtained, and distributions of the neutron multiplicity, fceff, were estimated. It was shown that the experimental uncertainty of Jezebel, which is 0.2%, was well reproduced by including the integral data into the covariance of 2 3 9 Pu fission cross section.

144

References 1. S.D. Clement, Nucl. Sci. Eng., 145, 72 (2003). 2. "International Handbook of Evaluated Criticality Safety Benchmark Experiment," NEA/NSC/DOC(95)03, Organisation for Economic Co-operation and Development / Nuclear Energy Agency (2002). 3. "PARTISN," Transport Methods Group, CCS-4, Los Alamos National Laboratory, Los Alamos National Laboratory, LA-UR-02-5633 (2002). 4. P. Talou, M.B. Chadwick, D. Madland, and P.G. Young, "Modification of ENDF/B-VI Release 5 (MOD 5) (pu2391a7d)," Los Alamos National Laboratory (2003) [private communication]. 5. A.D. Carlson, G.M. Hale, and V.G. Pronyaev, "Summary Report of the Second Research Co-ordination Meetings on Improvement of the Standard Cross Sections for Light Elements," Gaithersburg, MD, USA, 13-17 October 2003, INDC(NDS)-453 (2004). 6. R.C. Little, "MENDF6: A 30-Group Neutron Cross-Section Library Based on ENDF/B-VI," Los Alamos National Laboratory, LA-UR-98-545 (1998). 7. T. Kawano, H. Matsunobu, T. Murata, A. Zukeran, Y. Nakajima, M. Kawai, O. Iwamoto, K. Shibata, T. Nakagawa, T. Ohsawa, M. Baba, and T. Yoshida, J. Nucl. Sci. Technol, 37, 327 (2000). 8. T. Kawano and K. Shibata, "Covariance Evaluation System," JAERIData/Code 97-037, (1997) [in Japanese]. 9. S.C. Frankle, "Results of Initial Jezebel Calculations for Testing the Sensitivity of fceff Calculations to Changes in Voj" Los Alamos National Laboratory, LA-UR-05-1457 (2005).

SENSITIVITY OF ACTIVATION CROSS SECTIONS OF THE HAFNIUM, TANTALUM AND TUNGSTEN STABLE ISOTOPES TO NUCLEAR REACTION MECHANISMS V. AVRIGEANU, M. AVRIGEANU, F.L. ROMAN EURATOM-MEC Fusion Association, "Horia Hulubei" National Institute for Physics and Nuclear Engineering (IFIN-HH), P.O.Box MG-6, 76900 Bucharest, Romania R.A. FORREST EURATOM-UKAEA Fusion Association, Culham Science Center, Abingdon OX143DB, United Kingdom R. EICHIN, H. FREIESLEBEN, K. SEIDEL EURATOM-FZK/TUD Fusion Association, Technische Universitdt Dresden, D-01062, Dresden, Germany

1. Introduction A detailed analysis of the fast-neutron activation cross sections for W isotopes, relevant to safety aspects and waste management of power plants, has already been reported [1] with the aim to study the ratios of calculated-to-experimental activity (C7E) for several of the dominantly produced radionuclides. These ratios were found significantly above unity within a benchmark experiment with pure W irradiated by 14 MeV neutrons [2], when calculated with the European Activation System (EASY) [3]. The above-mentioned analysis, concerning all activation data for the stable W isotopes, was carried out using the computer codes EMPIRE-II [4] and TALYS [5] as well as a local parameter set in one updated version of STAPRE-H code [6]. The consistent input-parameter set adopted in the last case made use of recent neutron total cross sections of the m ,i83,i84,i86\V nuclei for analysis of deformed optical potential within the coupled-channels (CC) model, proton reaction cross sections, low-lying level and resonance data used for determination of level density parameters within a realistic approach recently developed [7], and y-ray strength functions /Ei(Ey) based on description of the corresponding capture data. The sensitivity of the calculated cross sections with respect to reaction mechanisms modeled by the

145

146

three codes, also related to the sensitivity to the main pre-equilibrium emission (PE) parameter - the single-particle level (s.p.l.) density g - is outlined in Fig. 1.

'•"""IO

is

E„ (MeV)

20

"

io

is

En (MeV)

Fig. 1. Sensitivity of calculated (n,p) and (n,2n) reaction cross sections to s.p.l. density global value g=A/14 MeV"1 and the values g=A/13 MeV"1 [4] and g=A/l5 MeV"' [5] used for model predictions.

147

2. Calculated cross-section sensitivity to model parameters for

Ta

A similar or even the same consistent parameter set has been involved for calculation of the activation cross sections for the I81Ta nucleus (standing actually for the whole Ta element) and their comparison with all available experimental data, especially for the same residual nuclei which are now in the neutron channel but have been also in the charged-particle channels for the neutron-induced reactions on the target nuclei 180'182'183.184'186\y. For the TALYS calculations we have used the code version TALYS-0.64 [5] which has recently been made available. In order to understand the particular points of various target nuclei and reaction channels we have used the code STAPRE-H and a consistent local parameter set. The neutron transmission coefficients provided by the code EMPIRE-II, already corrected for the direct inelastic scattering, have been taken as input in the TALYS and STAPRE-H calculations. Actually, the neutron optical potentials for both cases of the 181Ta and Hf stable isotopes (the 178Hf nucleus) have been analyzed at the same time in order to find a systematic trend of the OMP parameters. The PE model Geometry-Dependent Hybrid (GDH) has been used in the last case along with the CC method and statistical Hauser-Feshbach models to analyze experimental data of fast-neutron interaction with the 181Ta isotope, fully representative for the Ta element. No free parameter was involved in the GDH calculations while the same common parameters concerning OMP and nuclear level density have been used in the CC, GDH and HF model calculations. Therefore, since the semi-classical calculations can be applied rather easily for all reaction channel excitation functions, including reactions with the same residual nuclei in different channels, a proper description of a large body of data without free parameters may obviously validate both the adopted nuclear model assumptions and parameter set. The particle-hole state density with energy-dependent s.p.l. densities [8] was included within the GDH model as shown elsewhere [9]. The main point concerns the GDH specific account of the nuclear-density distribution by means of average quantities over the densities corresponding to the entrance-channel trajectory. Thus, local-density Fermi energies, which depend by the incident energy and angular momentum, are considered within the GDH formalism. Among other basic model assumptions the calculation of the intra-nuclear transition rates is based on the average imaginary optical-model potential, so that no free parameter comes in but the a-particle state density ga=A/10.36 MeV"1 [10] and the value (p=0.12 of the a-particle pre-formation probability [6]. The

148

brief comparison of the calculated and available measured activation cross 181. sections of the i51Ta nucleus is shown in Fig. 2. ,':-|ii!Ta(n,2n) 180 Ta

181

Ta(n,p)181Hf

*.,,«•

H/2- 42.39d]

•

/

0.015 * v D o a •-•

0.010

0.005

# l H

•

: ?#T A

XiOTgzhong+(i9S2) Begun+{1M9) Rlalerkm+OSSS) niatef*OT+(20O1) Begun-(20O1) EMPIRE-B EMPIRE-H: g=W14 TALV50.64 TALYSO,M:g=A/14 STAPRE-H

0.5t. ' jj ~£.

0.000

ff i3 {3 fl 3 fl S // fl i

•

o m • c o

I

Mogharrabt(1972) \ j \ Vaasar+ (1977) W»k Frehaut+(1980) V.^ Calkal (1982) 1^* TakahasM+(1989) N £ N EMPJRE-f! " § ^ EMPIRE-II, g=A/14 ^ c TALYS-0.64 TALYS-0.64, g=A/14 STAPRE-H i . . , i .—•—• 1 i

h4|; 25

0.003

Ta(n,a) [1+ 28.4

min]

//

Ta(n,2n)1808Ta

Lu/ ;' ;

+ 8.154h]

Begun+(1999) Fi!atenkov+(2001) '. Begun+ (2002) TALYS-0.64 - TALYS-0.64, g=A/14 i

-STAPRB^L——^;

0.001 f-

•

o.oog

~ '

^*'L*-^"^

/ / / / i f / / / / if // jl /

...;-:•_-•!—-:---..

* T v ^ • o • ,

LatehmanaDa&t-(197e) Ryvea* (1980) lkedaf(19fla) Hanl&M-(1989) Kasugak-(1992) Filatenkov+ (1899) Rlalenkov+(2001) TALYS-0.64 TALY&0.64,g=A/14 STAPRE-H

&

0.0020 179-,

^TaMnr^^pi 0.0010 o

Veeser+(1977) EMPIRE-II EMPIRE-II, g=A/14 TALYS-0.64 TALYS-0.64, g=A/14 STAPRE-H

0.0000

En (MeV) Fig. 2. Comparison of measured reaction cross sections for target nucleus 181Ta and calculated values using EMPIRE-II and TALYS codes with default global parameters (except local OMPs and s.p.l. value g=A/14 MeV"1) and STAPRE-H with the local parameter set of this work. A normahzation factor of 1.078 is applied to data of Frehaut etal. [11], according to Vonach etal. [12].

149

The global predictions have been obtained by using the corresponding default s.p.l. values, i.e. A/13 MeV"1 by EMPIRE-II and A/15 MeV"1 by TALYS, respectively, as well as the alternate value A/14 MeV"1 which proved to be most suitable in the case of the W isotopes analysis [1], Actually the present comparison with especially the predictions of TALYS is shown in order to underline the physical reason of some still existing discrepancies. Illustrative in this respect is the only large disagreement between the calculated and experimental data for the reaction 181Tafn,pj181Hf. A similar disagreement is proved by the global predictions of EMPIRE-II and TALYS corresponding to the default s.p.l. densities as well as to the particular value of A/14 MeV"1 which is leading to a much narrow uncertainty area of model-calculated cross sections.

15

En (MeV)

Fig. 3. Comparison of calculated and evaluated (ENDF/B-VI.8, JENDL-3.3) cross sections for 181Ta.

On the other hand the systematic smaller experimental cross sections suggest a common cause of the large C/E ratios, maybe due to the actual knowledge of the 181 Hf decay scheme. Thus, there is a tentatively assigned 9/2+ level at 68 keV within the rather obsolete evaluated level scheme [13], with no observed decay to the 1/2" ground state. This could be an isomeric state whose population would increase with the energy increase, and may explain the overprediction of the g.s. population. However, additional recent results are not confirming this level [14] but short-life isomers at 0.595 MeV (9/2+), 1.040 MeV (17/2+), and 1.738 MeV

150

(25/2") [15]. At the same time the evaluated data libraries ENDF/B-VI.8 and JENDL-3.3 are obtained based only on the measured data, so that no real disagreement exists between data and them (Fig. 3). 3. Calculated cross-section sensitivity to level schemes of Hf isotopes A most interesting case is that of the Hafnium isomers, which could be produced after a few reactions on W and Ta isotopes in the first-wall material of fusion reactors.

r=66±5 meV [RIPL2] r=71 meV 66meV 61 meV 51 meV

,8

%, #

°Hf( %Y) 181 Hf

*

'

Siddappaf (1974) Zu-Mng+ (1984) TrofimovJM (1987) Trofimov#2(19B7) Bckhovkot- (1996) Jlnxiang+-(1997)

^

r r = 5 0 ± 5 meV [RIPL2] $ •'••••• ( r = 5 5 m e V

SOmeV 45meV 35 meV

%™ 3 ^ J ^ r ^ i

0.1

E (MeV)

V

,

181

^^> • 0 * o _•

Sddappa* (1974) Poenitt(1975) Undnert (1976) Kononov* (1977) Yamamurot- (1978)

» v a A •

Allent (1B81) Volgnler-f (1986) Haishant (1936) Haiahan* (1958) Bokhovkft* (1996)

Ta(n,Y) 182 Ta r =57+3meV[RIPL2] - - r=60meV — 57 meV : — 54 meV

En°(MeV)

Fig.4. Comparison of calculated and neutron capture cross sections of Hf and m T a nuclei, with respect to experimental average radiative widths Tyo of the s-wave neutron resonances.

In order to get confidence in the calculated isomer cross sections, the calculated and experimental cross sections of the (n, y) reactions on 176180Hf nuclei were firstly compared (Fig. 4). The RIPL values for F^ lead to /Ei(Ey) strength

151

functions which are sometime too large, so that appropriate values have been established in these cases and used in the following activation calculations. 176

Hf(n,2n) 175 Hf

Qaim (1974) Lakahmana Da&+ (1974)\ Patrick+(1990) Meadow3+(1996) Kong+ (1398) Kiraly+(2001) -CC+GDH+HF

15

179

Hf(n,2n) 178m1 Hf

179

Prasad+(1966) - CC+GDH+HF

180

Patrick+(1990) lkeda+(1991) Webdang+(1992) »angzhong+(199S) Wei-»ang+(1995) lkeda+ - CC+GDH+HF

Hf(n,2n) , 7 9 m , Hf [1/2- 18.7S]

• • • •

Hf(n,2n) 178m2 Hf

,80

Hf(n,2n)' 79nn *Hf

^

[25/2- 25.1D1

Prasad* (1966) Rurarz+(1970) Sothras(1977) CC+GDH+HF

\

• V •

Patrick+(1990) Weiwang+(1992) Konno+(1993) CC+GDH+HF

/ f

: •

E (MeV)

E (MeV)

Fig. 5. Comparison of the calculated and experimental (n,2n) reaction cross sections of the 174 ' 176179180 Hf nuclei.

The agreement of the STAPRE-H results and measured lowest-lying and low spin isomer cross sections (Fig. 5) is much decreased by the poor knowledge of the decay schemes. These schemes play however no role for the high-spin isomers, while their continuum feeding seems to be rather well described. Finally, an overview of the present work shows that the differences between various evaluated data are rather similar to those between model calculations using global as well as local parameter sets. At the same time, the differences within the latter class are larger than changes of the calculated results due to the variation of the model parameters within the limits allowed by the actual knowledge. On the other hand, the correctness of the last ones depends critically

152

by the accuracy of the other independent data used for the consistent determination or validation of the model parameters. Acknowledgements Thanks are addressed to Arjan Koning, Roberto Capote and Andrej Trkov for helpful advice during the work with the codes TALYS-0.64 and EMPIRE-II. References 1. V. Avrigeanu, R. Eichin, R. A. Forrest, H. Freiesleben, M. Herman, A. J. Koning, and K. Seidel, Int. Conf. on Nuclear Data for Science and Technology, Sept. 26 - Oct. 1, 2004, Santa Fe, New Mexico (in press). 2. R. Eichin, R.A. Forrest, H. Freiesleben S.A. Goncharov, V.D. Kovalchuk, D.M. Markovskij, K. Seidel, and S. Unholzer, Int. Workshop on Fast Neutron Physics, Sept. 5-7, 2002, Dresden, Germany. 3. R.A. Forrest et al., Report UKAEA FUS-500, 2003, Culham Science Centre, Abingdon, UK. 4. M. Herman, EMPIRE-H v.2.18, http://www-nds.iaea.org/empire/. 5. A.J. Koning, Report INDC(NDS)-431, IAEA, Vienna, 2002, p. 117; A.J. Koning and M.C. Duijvestijn, Nucl. Phys. A 744, 15 (2004). 6. M. Avrigeanu and V. Avrigeanu, IPNE Report NP-86-1995, Bucharest, 1995, and references therein; News NEA Data Bank 17, 22 (1995). 7. V. Avrigeanu, T. Glodariu, AJ.M. Plompen, and H. Weigmann, J. Nucl. Sci. Tech. S2, 746 (2002). 8. M. Avrigeanu and V. Avrigeanu, Comp. Phys. Comm. 112, 191 (1998); A. Harangozo, I. Stetcu, M. Avrigeanu, and V. Avrigeanu, Phys. Rev. C 58, 295 (1998). 9. M. Avrigeanu and V. Avrigeanu, J. Phys. G: 20, 613 (1994); M. Avrigeanu, V. Avrigeanu, and A.J.M. Plompen, J. Nucl. Sci. Technol. S2, 803 (2002). 10. E. Gadioli and E. Gadioli-Erba, Z. Phys. A 299, 1 (1981). 11. J. Frehaut, A. Bertin, R. Bois, and J. Jary, Report BNL-NCS-51245, BNL, 1980, vol. I, p. 399. 12. H. Vonach, A. Pavlik, and B. Strohmaier, Nucl. Sci. Eng. 106,409 (1990). 13. R.B. Firestone, Nucl. Data Sheets 62, 101 (1991). 14. C. Gunther et al., Phys. Rev. C 65, 047301 (2002); V. Bondarenko et al., Nucl. Phys. A 709, 3 (2002). 15. R.D. 'Alarcao et al., Phys. Rev. C 59, R1227 (1999); I. Shestakova et al., Phys. Rev. C 64, 054307 (2001).

G E N E R A T I N G COVARIANCE DATA W I T H N U C L E A R MODELS

A.J. KONING NRG,

Nuclear

Research and Consultancy Group Westerduinweg 3 P.O. Box 25, NL-1755 ZG Petten The Netherlands E-mail: koning@nrg-nl. com

A reliable assessment of the uncertainties in calculated integral reactor parameters depends directly on the uncertainties of the underlying nuclear data. Unfortunately, covariance nuclear data are scarce, not only because a significant experimental database for the isotope under consideration must be available, but also because the covariance evaluation process can be rather complex and timeconsuming. We attack this problem with a systematical approach and developed, following the initial ideas of D. Smith (ANL), a method to produce a complete covariance matrix for evaluated data files on the basis of uncertainties of nuclear model parameters. This is accomplished by subjecting the nuclear model code TALYS to a Monte Carlo method for perturbing input parameters, an approach that is now possible with the available computer power. After establishing uncertainties for parameters of the optical model, level densities, gamma-ray strength functions, fission barriers etc., we produce random input files for the TALYS code. These deliver, provided enough calculations (samples) are performed, uncertainties + all off-diagonal elements for all open reaction channels. The uncertainties of the nuclear model parameter are tuned such that the calculated cross section uncertainties coincide, to a reasonable extent, with uncertainties obtained from covariance evaluations based on experimental data. If this method proves to be successful, and we will show here that we are not too far off, it will enable mass production of credible covariance data for isotopes for which no covariance data exists and this constitutes a very significant part of the periodic table of elements.

1. Introduction Nuclear data, such as cross sections, resonance parameters, energy spectra and angular distributions are of prime importance to the computational simulation of any nuclear design. To make the most impact, these nuclear data should include uncertainties. Starting from the covariances of the basic data, error propagation in transport, transmutation, etc. codes enables to estimate the uncertainties of calculated design parameters, which may have 153

154

a profound impact on issues of general concern such as safety and economy. Unfortunately, while the good habit of delivering error bars is followed by most experimental nuclear physicists, only a few data evaluators and practically no theoreticians provide uncertainties, let alone a full covariance matrix, with their results. Especially for the last group, the nuclear model community, there seems to be no excuse: as long as (a) the exact form for the strong interaction, and (b) the exact solution for the many-body problem, have not been uncovered we know that model-predicted nuclear data only have one certainty: they are wrong. So why not providing error bars with the results? We can think of a few reasons why the task of assigning uncertainties to nuclear model calculations is difficult: 1) In general, it is not straightforward, or at least ambiguous, to propagate input uncertainties of model parameters through a nuclear model calculation to obtain the final cross section uncertainties. 2) It is very difficult, if not impossible, to disentangle uncertainties of model parameters from the uncertainties of the models themselves. 3) Many nuclear models (level densities, optical model, preequilibrium, fission barriers, etc.) contribute to the description of one reaction channel. This leads unavoidably to parameter correlations, not only within one model (e.g. the well-known W 2 ambiguity of the optical model) but also between parameters of different models (e.g. level density parameters and the pre-equilibrium matrix element) that have the same effect on a calculated quantity. Using uncorrelated parameters may yield unrealistic uncertainties for the calculated cross sections. With the help of a flexible nuclear model code and the current-day computer power, we can now take a pragmatical approach to attack some of these problems, and in this paper we start with issue 1. We note that the basic ideas of this approach have already been outlined by Smith 1 . The present paper shows that the method is actually feasible. The essential idea is to assume that each nuclear model parameter has its own uncertainty, where for the moment the uncertainty distribution is assumed to have a Gaussian shape. Running a nuclear model code many times, whereby each time all elements of the parameter vector are randomly sampled from a Gaussian distribution with a specific width for each parameter, provides all needed statistical information to produce a full covariance matrix. The current results should be regarded as "proof of principle" only: various subtleties on correlation of parameters and models have been dis-

155

regarded at this stage of the development. At the moment, we can only provide an indirect proof of the validity of the method: reasonable uncertainties for the simulated cross sections. 2. The nuclear model code TALYS As workhorse we use the recently released TALYS code 3 . It is a computer code system for the prediction and analysis of nuclear reactions. It simulates reactions that involve neutrons, gamma-rays, protons, deuterons, tritons, helions and alpha-particles, in the 1 keV - 200 MeV energy range and for target nuclides of mass 12 and heavier. This is achieved by implementing a suite of nuclear reaction models into a single code system. It enables to evaluate nuclear reactions from the unresolved resonance region up to intermediate energies. The results described in this paper are based on a theoretical analysis that utilizes the optical model, compound nucleus statistical theory, direct reactions and pre-equilibrium processes, in combination with databases and models for nuclear structure. Fig. 1 summarizes the nuclear models implemented in TALYS. The following output

TALYS

Optical Model: * Phenomenology local / global

i Input:

Direct reaction; Spherical OM DWBA Rotational CC * Vibrational CC Giant resonances Weak-coupling

I* Keywords, eg:

Preequilihriimi: * Exciton model - 2-component * p - h LD phenom. - surface effects * Kalbach systematics - angular distribution - cluster emission * 7 - r a y emission

Output: *File 'output' defined by keywords •Dedicated files with spectra, ...

^projectile n \element

fe

mass 56 ;energy 14.

Optional loops * Incident energies * Natural isotopes

Nucl. S t r u c t u r e : * Abundancies * Discrete levels * Deformations * Masses * Level density par. * Resonance par. * Fission barrier par. * Thermal XS * Microscopic LD * Prescission shapes

Compound: * Width fluctuations - Moldauer - GOE triple integr. -HRTW * Hauser-Feshbach * Fission competition - isotopic yields * y -ray emission * GC+ Ignatyuk

Figure 1.

Multiple emission: * Exciton (any order) * Hauser-Feshbach

ENDF: * transport libs * activation libs

* Fission competition - isotopic yields * y -ray cascade * All flux depleted * Exclusive channels * Recoils

Nuclear models in TALYS

is produced by TALYS (and eventually stored in the nuclear data files): • Total, elastic and non-elastic cross sections

156

• Elastic scattering angular distributions • Inelastic scattering cross sections and angular distributions to discrete states • Exclusive channel cross sections, e.g. (n,7), (n,2n), (n,np),.., energy and double-differential spectra • Gamma-ray production for discrete states and continuum • Fission cross sections • Isomeric and ground state cross sections • Residual production cross sections (for E > 20 MeV) • Total particle cross sections, e.g. (n,xn), (n,xp),.., energy and double-differential spectra (for E > 20 MeV) All data is produced in easily readable files, per reaction channel, enabling quick graphical and numerical comparison with measurements. In Fig. 1, the box labeled "Input" may give the obviously false impression that the maximal predictive quality can be obtained by simply specifying the nuclear reaction under consideration with 4 input lines. In fact, this only delivers a so called "blind" set of results. In practice, usually various nuclear model parameters, fed to the code by means of keywords, need to be adjusted to produce the optimal results. 3. Random covariance model A short, simple mathematical formalism is sufficient to explain the method. Let p be a vector of L nuclear model parameters, i.e. p = {p1,...,ph...,pL}.

(1)

This vector contains all adjustable parameters that are relevant to the calculation under consideration. For example, one may start with the optical model parameters p\ = rv (radius), pi = av (diffuseness), etc., followed by the level density parameters and so on. In practice L may take on values from only a few up to several tens, especially if many residual nuclides are involved in a calculation. Next, let cr be a vector of N calculated cross sections, i.e. c = {o-i,...,<Ti,...,crN}.

(2)

We have used the term "cross sections" here, but it should be understood that the definition of a applies to all quantities of interest (angular distributions, differential energy spectra, etc.). For example, o\ = CTtot(£i),-..,cri = d<7 e l /dn(.Ei,0i),...., where Ei and Gi are the first

157

energy and angle respectively, on the grid. It is obvious that a can contain a very large number of elements. For the present paper, we have limited ourselves to cross sections only, for all open reaction channels. The vector a is a function T of the vector p, ° = T(p),

(3)

where T stands for the TALYS code. For the determination of the covariance matrix, we run TALYS with K different sets of model parameters, i.e. vectors p, which we will therefore provide with a superscript (k) to denote the k-th TALYS run, CTW=T(p(fc)).

(4)

Let us designate the "best" set of input parameters, the central values, as p(°) and the corresponding "best" calculated cross section vector as a^0'. These are for example obtained from an appropriate goodness-offit estimator in an optimization scheme with the experimental data of all reaction channels. It can also simply represent a so-called global nuclear model calculation with all parameters equal to their default values. We assume that each nuclear model parameter has its own uncertainty, pi=p0±Api,

l = l,L,

(5)

where it is understood that the uncertainty distribution is assumed to be a Gaussian. The basis of our method is now formed by running TALYS K times, with e.g. K = 1000, whereby each time all L elements of the p vector are randomly sampled from a Gaussian distribution with a specific width Api for each parameter pi. After performing all K calculations, all statistical information is available. First of all, the average calculated cross sections are

°T' = ^ I > i f c ) '

* = !,*•

(6)

/c=l

The average covariance matrix is given by v

« = Jc I > ' ( f c ) - a™*°?

- ^(0))'

" = x«N-

(7)

fc=i

From this, the average relative covariance matrix can be obtained, Rii = V « / ( ^ 0 M 0 ) ) > i,3 = l,N.

(8)

158

The square root of the diagonal elements of this matrix represents the uncertainty. Hence, the final calculated cross sections together with their uncertainties can be expressed as ^final

=

aav(1

±

Jjj^^

i

=

l>N.

(9)

4. Calculations in practice We think it is illustrative to show in more detail how TALYS works for the purpose of evaluating covariances. Here are two input files for TALYS:

Input file 0 (Central values) parameters)

Input file 1 (first set of random

projectile n element fe mass 56 energy energies

projectile n element fe mass 56 energy e n e r g i e s

#General

#General parameters

parameters

# M 2 c o n s t a n t 1. # # Parameters # a 26 gammald 26 gamgam 26 sgr 26

for

57Fe

M 2 c o n s t a n t 1.19515 # # Parameters for 57Fe n

57 57 57 57

6.77226 0.11927 0.92000 8 3 . 9 7 6 El

a gammald gamgam sgr

26 26 26 26

57 6.29336 57 0.17733 57 0.69751 57 1 0 2 . 6 0 0 El

Clearly, these cases concern neutrons on 56 Fe, with the incident energies specified by the user in the file 'energies'. The input file on the left concerns the 'best' input file, i.e. where the parameter set p ° has been adjusted to optimally describe the measurements. A calculation of this type yields all information that can generally be found in a nuclear data file: total, elastic, non-elastic partial cross sections, angular distributions, double-differential cross sections, and gamma-ray production cross sections. The input file on the right has the same structure, only this time all nuclear model parameters relevant for this reaction have been randomly sampled from a Gaussian distribution. For the results presented in this work, 1000 of such random cases have been calculated with TALYS. The calculation of one input file

159

takes approximately 30 seconds on a 1GHz Pentium PC, provided we take 20 MeV as the maximum energy. Hence, a full covariance calculation for one nuclide takes approximately one night. After this, all 1000 sets of
160

- Best TALYS ) i oJEFF-3.0 - Rnal

- B e s l TALYS S

:5"

2

4

"ww

6

8 10 12 Energy (MeV)

o Frehaut et al. (1980) • Simakovetal.{1993) — TALYS

14

16

18

Tefn.a)

20

10 12 Energy (MsV)

Pb(n,2n)

14

Pb(n,p) Tl

j

o Bass and Wechsung (1S6S) • Welch etal. (1981) • Pbmpen et al. (2002) TALYS

ff/' ft

S" 18 Energy (MeV)

20

0

2

4

6

6

10 12 14 Energy (MeV)

16

18

20

22

Figure 2. Calculated cross section including uncertainties, compared with experimental data or with nuclear data files that contain covariance data: (a) 5 6 Fe(n,7) for 30 random parameter sets, (b) the resulting S 6 Fe(n,7) cross section, (c) 5 6 Fe(n,2n), (d) 5 6 Fe(n,a), (e) 2 0 8 Pb(n,2n), (f) 2 0 8 P b ( n , p ) .

161

indeed normally distributed, whether they should actually be log-normally distributed to avoid the issue of negative, unphysical, values for some sampled parameter values, and how one should deal with discrete model parameters (not yet used for the current work). Apart from the model parameter uncertainties, the nuclear models themselves are also uncertain (that's why they are called models), and this has not yet been disentangled from the parameter uncertainties. It is under debate whether this is a practical problem 5 or not 6 . It should also be possible to invert the route we have taken here, i.e. to start from cross section uncertainties and to derive the parameter covariances from this. This is currently under study at TU Wien 5 . For the present work, a priori uncorrelated parameter uncertainties have been used, and the consequences of this approximation for the calculated covariances are unknown. Finally, it may turn out to be very difficult to exactly mimic the experimental uncertainty. It can be imagined that e.g. various high-quality 14-15 MeV measurements of an (n,p) cross section for a nuclide exist and no measurements outside this energy range. This situation should be represented by an uncertainty band which is relative broad up to, say 12 MeV, then starts to narrow down to the average experimental uncertainty at 14 MeV, and then broadens again above 15 MeV. There are no modeling techniques yet to simulate this. Also, it turns out to be difficult to produce large uncertainties in some cases for which we know the results should be uncertain, e.g. the threshold behaviour mentioned before. Even unrealistic parameter choices do not reproduce the experimental uncertainties. Probably, the introduction of model uncertainties should help out here. 7. Conclusions In this paper we aim to demonstrate that it is possible to generate a complete covariance matrix for all neutron-induced reactions using a Monte Carlo method. The only necessary ingredients are a flexible, validated, and robust nuclear model code and enough computer power. Nuclear model parameters, fed to the TALYS code by means of keywords, are randomly sampled from a Gaussian distribution, and the subsequent running of TALYS for many cases leads to a full covariance matrix. The current method should be regarded as "proof of principle" only, since various deficiencies still need to be overcome: credibility of the parameter uncertainties, reproduction of

162

experimental uncertainties, the issue of parameter correlations and (which would be a major asset) a general uncertainty assessment of the TALYS results. This last objective requires a serious review of the E X F O R database. After t h a t , we could perform a comparison of blindly calculated TALYS results against all experimental d a t a of the periodic table of elements. T h e average deviation from experiment would enable us to assess the most pessimistic uncertainties for unmeasured reaction channels. Such a systematical approach towards nuclear d a t a evaluation is as yet non-existent, while it would contribute to a well-founded justification, or decline, for specific new measurements, and could have a significant impact on nuclear applications t h a t require uncertainties. Finally, although this has not been discussed in detail here, we have the entire calculation route under control:

from

parameter uncertainties to a full ENDF-6 formatted nuclear d a t a file t h a t includes the covariance file M F 3 3 (the angular distribution covariances of MF34 are under construction). T h e next step is t o improve t h e quality of the actual numbers of the covariance matrix. Some steps in this direction have been suggested above.

References 1. D. Smith, "Covariance matrices for nuclear cross sections derived from nuclear model calculations", Argonne report ANL/NDM-159 November 2004. 2. H. Vonach, S. Tagesen, M. Wagner and V. Pronyaev,"Evaluation of the fast neutron cross sections of 5 6 Fe including complete covariance information", Physics Data, Fachinformationszentrum, Karlsruhe report 13-7, 1992. 3. A.J. Koning, S. Hilaire and M.C. Duijvestijn, "TALYS: Comprehensive nuclear reaction modeling", Proceedings of the International Conference on Nuclear Data for Science and Technology - ND2004, Sep. 26 - Oct. 1, 2004, Santa Fe, USA. 4. T. Kawano, T. Sanami, M. Baba, and H. Nakashima, Journ. Nucl. Sci. Techn. 36, 256 (1999). 5. H. Leeb, private communication. 6. D. Smith, to be published in Acta Physica et Chimica Debrecina, in honor of Prof. Csikai's 75th birthday.

SENSITIVITY OF CANDU-SCWR REACTOR PHYSICS CALCULATIONS TO NUCLEAR DATA FILES K.S. KOZIER and G.R. DYCK Atomic Energy of Canada Limited, Chalk River Laboratories, Chalk River, Ontario, KOJ 1J0, Canada

A simplified MCNP model of a CANDU-SCWR lattice was used to test the sensitivity of the calculated reactivity to various nuclear data files involving issues of current interest. These tests were performed for cooled and voided conditions, with and without axial neutron leakage, for a uniform lattice of mid-life fuel and a mixed lattice with highburnup and low-burnup fuel in alternate channels. Results were compared using different room-temperature data files for deuterium, various thermal-scattering-law data files for hydrogen bound in light water and deuterium bound in heavy water, and for pre-ENDF/B-Vn and ENDF/B-VI.8 data for uranium. The reactivity differences observed were small (typically <1 mk) and increased with axial neutron leakage.

1. Introduction 1.1. The CANDU-SCWR Generation-IV concept The long-term vision for CANDU®* (CANada Deuterium Uranium) reactor development foresees a continued focus on flexible, thermal-spectrum reactors featuring simple and inexpensive fuel contained in modular pressure tubes and surrounded by neutron-efficient heavy-water (D 2 0) moderator, with an evolution toward higher operating temperature and improved efficiency. The next step in this evolution is the Generation III+ ACR®" (Advanced CANDU Reactor). It features light water (H 2 0) instead of D 2 0 as coolant, slightly enriched uranium instead of natural uranium fuel and a reduced lattice pitch. The Generation IV successor to the ACR is envisaged as a pressure-tube version of the SCWR (Supercritical Water Reactor), the CANDU-SCWR. The CANDU-SCWR concept involves innovative technology, such as the high efficiency channel (HEC) design shown in Figure 1. The pressure tube in the HEC is in contact with the low-temperature D 2 0 moderator and is separated from the low enriched uranium CANFLEX®a fuel bundle by a perforated liner or * CANDU, CANFLEX and ACR are registered trademarks of AECL

163

164

guide tube and a layer of ceramic thermal insulation. The design intent is to achieve both a high outlet coolant temperature during normal operation and passive heat removal to the moderator in die event of a loss of coolant accident. Additionally, the CANDU-SCWR concept achieves relatively uniform core-wide coolant density and neutronic properties using bi-directional coolant flow and on-power fuelling in adjacent fuel channels.

Figure 1. CANFLEX fuel bundle in a high efficiency channel.

1.2. Nuclear data issues of interest The general-purpose Monte Carlo code MCNP5™f [1] was used in the CANDU-SCWR context to examine some nuclear data topics of generic interest. 1.2.1. Nuclear data files for deuterium MCNP5 is distributed with continuous-energy nuclear data files for deuterium (2H or D) at 294 K denoted by the ZAID (Z=atomic #; A=mass #) nuclide identifier 1002.60c, based on the Evaluated Nuclear Data File (ENDF) library ENDF/B-VI.4, and 1002.66c, corresponding to releases ENDF/B-VI.5 through VI.8. Recently, a new data file for deuterium (denoted 1002.64c) was created [2] at Los Alamos National Laboratory (LANL) based on the ENDF/B-VI.4 specifications.

MCNP is a trademark of the Regents of the University of California, Los Alamos National Laboratory

165

A rather large increase in the calculated multiplication constant, k-effective (keff), of about 9.7 mk (1 mk = 0.1% Ak = 100 pern) is observed when 1002.64c (or 1002.60c) is used instead of 1002.66c in MCNP simulations of high enriched uranium (HEU) D 2 0 solution benchmarks HEU-SOL-THERM-004 and -020 [3]. This reactivity change is attributed [4] to slight revisions with 1002.66c to the elastic scattering angular distributions above 20 MeV and below 3.2 MeV to improve agreement with measurements. In contrast, only small Akeff values (-0.4 to +0.5 mk) are observed [4] when 1002.64c is used instead of 1002.66c in MCNP simulations of critical experiments performed for various arrangements of natural uranium (NU) fuel rods in D 2 0 moderator in AECL's ZED-2 (Zero Energy Deuterium) reactor at the Chalk River Laboratories. 1.2.2. Thermal scattering law data for' H bound in H20 and2H in D20 The importance of treating lattice binding effects correctly in water-cooled reactor calculations has been demonstrated recently by Cullen et al [5]. These effects are described in the ENDF/B data libraries by a thermal scattering law, S(a,f3), where a is the momentum transfer and P the energy transfer. In this study, it is assumed that the available S(a,P) for hydrogen bound in H 2 0 are equally applicable for supercritical conditions. MCNP5 is distributed with a new set of S(cc,P) data, labelled SAB2002 [6], including data for 'H bound in H 2 0 (updated for ENDF/B-VI.3) and 2H bound in D 2 0. Although the S(a,P) for 2H in D 2 0 is based on older data evaluations, SAB2002 provides improved resolution in terms of secondary neutron energies (64) and scattering angles (16). An updated model for thermal neutron scattering in D 2 0 had been developed at AECL by J.V. Donnelly in 2000; however, this work has not yet been published and the model has not been widely tested. More recently, the S(a,p) for hydrogen in H 2 0 and deuterium in D 2 0 have been re-evaluated [7] under the auspices of the International Atomic Energy Agency (IAEA). Notably, the IAEA's S(oc,P) for deuterium in D 2 0 addresses certain intermolecular and intramolecular interaction effects that were not considered in other models. 1.2.3. Pre-ENDF/B-VII data files for uranium A preliminary new version of the ENDF/B data library, labelled preENDF/B-VII, is currently undergoing testing and evaluation. It combines uranium cross sections from the T16_2003 library [8] (294 K data denoted by ZAID filename extension .69c) with revised resonance data for " J U and "°U [9] to produce new data files (ZAID extension .00c). In MCNP simulations of the

166

same ZED-2 experiments noted in Section 1.2.3, the revised pre-ENDF/B-VII data for uranium increase kef? by about 4.4 mk. The T16_2003 library also includes new data files for 233U, 234U, 236U, 237Np, 239Pu, 241Am and 243Am. 2. Method 2.1. WIMS-AECL CANDU-SCWR lattice-cell model Currently, only simplified, preliminary reactor-physics assessments have been performed for the CANDU-SCWR concept, based on two-dimensional WIMS-AECL [10] multigroup, neutron-transport calculations. These calculations simulate an infinite array of identical lattice cells (see Figure 1) and determine the fuel composition as a function of fuel burnup for use in subsequent MCNP calculations. The WIMS-AECL calculations used an 89-group nuclear data library that is consistent with ENDF/B-VI.5. The WIMS-AECL case simulated a lattice pitch of 21 cm with 99.9 wt% pure D 2 0 moderator at 373 K and a 7-mm-thick Zr2.5wt%Nb pressure tube with an outside diameter of 13.4 cm. The 6.75-mm-thick porous insulator region was assumed to contain 30wt% Zr0 2 ceramic and 70wt% supercritical H 2 0 coolant. Similarly, the guide tube region was 1.5 mm thick with 55wt% SX403 alloy and 45wt% H 2 0 coolant. The fuel bundle contained U0 2 enriched to 4.25wt% 235U and clad with 0.3-mm-thick type-304 stainless steel. The supercritical H 2 0 coolant was assumed to correspond to core inlet conditions (about 0.444 g/cm3 and 653 K) for the first half of the irradiation and outlet conditions (about 0.241 g/cm3 and 898 K) for the remainder. 2.2. MCNP CANDU-SCWR model A corresponding MCNP model was constructed using a two-by-two arrangement of square lattice cells with specular-reflection boundary conditions at the outermost surfaces in the transverse direction. The axial core length was varied from infinite (modelled as a 100-cm long core with specular reflection at both ends) to as small as 50 cm. Most of the MCNP calculations were performed using a uniform fuel composition corresponding to a mid-life burnup of 22.3 MWd/kgU. But some calculations were also performed using a "mixed" lattice with nearly fresh fuel (0.2 MWd/kgU) and discharge fuel (44.2 MWd/kgU) arranged in alternate lattice sites in a checkerboard pattern. This arrangement is representative of the ends of the core, with strong transverse neutron currents (i.e., inter-cell leakage) from the fresh fuel toward the irradiated fuel.

167

Both the fully cooled (nominal coolant density 0.444 g/cm3) and fully voided (coolant density reduced to 0.0004 g/cm3) core configurations were simulated. However, for simplicity only uniform coolant density cases were considered. For the voided cases, the coolant density in the porous insulator and guide tube regions was also reduced accordingly. The room-temperature (294 K) nuclear data files corresponding to ENDF/B-VI.8 that are distributed with MCNP5 release 1.20 were used as the reference. However, some minor fission products in the WIMS-AECL output (i.e., 97Mo, 100Mo, 102Ru, 107Pd, 115In, 133Xe, 139La, 141Ce, 142Ce, 144Ce, 143Pr, 144 Nd, 146Nd, 148mPm, 151Pm and 156Eu) were not available from this source. To avoid omitting these nuclides, supplementary data files were added from inhouse MCNP data libraries. Most of the additional files corresponded to 300 K. Each MCNP case executed 1050 cycles of 60 000 neutrons with the first 50 cycles being skipped. The largest statistical uncertainty in keff was <0.09 mk ( l a ) . The net neutron leakage values were obtained from the particle weight loss to escape (per source particle) given in the MCNP problem summary tables. 3. Results 3,1. Sensitivity to nuclear data files for deuterium Table 1 lists the MCNP keff and neutron leakage values obtained using the data files for deuterium (i.e., 1002.66c, .60c and .64c) described in Section 1.2.1. For the uniform lattices, ENDF/B-VI.8 data were used for other nuclides, while for the mixed lattices the latest pre-ENDF/B-VII data files were used (i.e., ZAID extensions .00c for U and U and .69c where available for other nuclides). The Akgjf values with respect to the 1002.66c-based MCNP results are shown as a function of l/(axial core length)2 (which is proportional to the axial geometric buckling, apart from the flux extrapolation distance) in Figure 2. The following observations are noted: Only a small increase in keff occurs using 1002.60c at the low axial leakage values applicable for CANDU-SCWR or ZED-2 (roughly <0.4 m"2 in Figure 2). This increases to 2 to 3 mk for the high-leakage cases involving the escape of about 40% of the neutrons (comparable to the leakage in the HEU-SOLTHERM-004 and -020 benchmarks). The reactivity impact differs for cooled and voided cases, with larger differences occurring for the voided cases at >0.5 m"2 in Figure 2. Equivalent results are obtained using 1002.60c or 1002.64, as expected. Similar results are obtained for the mixed lattice, despite high internal leakage between lattice cells.

168 Table 1. Sensitivity of MCNP5 results to various data files for 2H Cooled Core type

2

H data file

1002.66c

Uniform 1002.60c

1002.64c 1002.66c Mixed 1002.64c

Axial length (cm) Infinite 100 70 50 Infinite 100 70 50 Infinite 50 Infinite 50 Infinite 50

Escape fraction

keff

0 0.12062 0.21739 0.35268 0 0.11961 0.21560 0.35039 0 0.35037 0 0.35217 0 0.34996

1.04587 0.91369 0.80843 0.66237 1.04635 0.91905 0.81040 0.66473 1.04633 0.66483 1.07361 0.68644 1.07408 0.68919

k„ff

0 0.16583 0.29121 0.45635 0 0.16411 0.28852 0.45285 0 0.45295

1.07642 0.88840 0.74792 0.56591 1.07655 0.89014 0.75072 0.56930 1.07658 0.56921

-

-

-i

3.0

.;..• i

2.5

?

J

Jr i^ *s^

£• 2.0 1.5

0.5 [

Escape fraction

I

3.5

1.0

Voided

I

.».—cool»d uniform: 1002.60c • -o> - voidod uniform: 1002.60c O

cooJad uniform: 1002,64c

A

voJtted uniform: 1002.64c

&

cooled mlx*d: 1002.64c

0,0 : 1/(oor* Ungth) 1 (m*3)

Figure 2. MCNP5 Akeff relative to 1002.66c versus l/(axial core length)2.

Apart from the increase of -0.5 mk for the cooled case at zero leakage, most of the reactivity increases in Figure 2 can be directly attributed to reductions in neutron leakage when 1002.60c or 1002.64c are used, as shown in Figure 3.

169

" - *

• "O A

cooled uniform: 1002.60c - voided uniform: 1002.00c cooled mixed: 1002.64c

-3.0

-2.5

-2.0

-1.5

-1.0

Change in particle weight loss to escape x 1000

Figure 3. MCNP5 Akeff relative to 1002.66c versus change in escape fraction x 1000.

3.2. Sensitivity to S(a,fi) data files for 1H bound in H20 and 2H in D20 Table 2 lists the MCNP5 results obtained using the reference SAB 2002 data files, various combinations of the IAEA S(oc,P) for 'H in H 2 0 and 2H in D 2 0 and the AECL S(oc,P) for 2H in D 2 0. At zero leakage, the largest change in keff (+1.1 mk) occurs for the cooled configuration when only the S(oc,P) for ! H in H 2 0 is changed to the IAEA data file. As before, the magnitudes of the reactivity differences tend to increase with the magnitude of the change in axial neutron leakage. However, the changes are of opposite sign for the IAEA and AECL S(a,P) cases for D 2 0, which may reflect the additional interactions considered in the IAEA's model.

170 Table 2. Sensitivity of MCNP5 results to various S(a,P) data files. Cooled Voided

S(a,P) data file SAB2002 IAEA-SAB (H 2 0 & DaO) IAEA-SAB (H 2 0 only) IAEA-SAB (D 2 0 only) AECL-SAB (D 2 0 only)

Axial length (cm) Infinite 50 Infinite 50 Infinite 50 Infinite 50

Escape fraction 0 0.35268 0 0.35202

keff

1.04587 0.66237 1.04664 0.66357 1.04694

Escape fraction

keff

1.07642

0 0.45635

0.56591

0 0.45587

1.07660 0.56650

-

0 0.35245

0.66314

-

0 0.35223

1.04586

0

1.07663

0.66283

0.45595

0.56650

Infinite

0

1.04539

0

1.07599

50

0.35380

0.66079

0.45669

0.56529

3.3. Sensitivity to Pre-ENDF/B- VII data files Table 3 lists the MCNP5 results obtained using the reference ENDF/B-VI.8 (i.e., ZAID extension .66c) data files (except 1002.64c was used for 2H), and the pre-ENDF/B-VII data files. Two sets of cases were considered for the latter: .69c data files for available nuclides, including 235U and 238U, and .69c data files for available nuclides with .00c data files for "\J and ZJ8U. Negligible changes in k ^ occur at zero leakage for both the cooled and voided configurations when the .69c data files are used in place of .66c, although kgff increases proportionately with the change in leakage for the high-leakage cases. When the .00c data files are used for 235U and 238U, kgff increases by -1.5 mk at zero leakage, but the Ak<.ff does not depend strongly on leakage and is similar for both the cooled and voided configurations. Table 3. Sensitivity of MCNP5 results to nuclear data libraries Cooled Voided Nuclear data library ENDF/B-VI.8 (.66c) Pre-ENDF/B-Vn (,69c) Pre-ENDF/B-Vn (,00c)

Axial length (cm) Infinite

Escape fraction

50 Infinite 50

keff

Escape fraction

0

1.04633

0

1.07658

0.35037

0.66483

0.45295

0.56921

0

1.04634

0

1.07663

0.34944

0.66572

0.45209

0.57012

keff

Infinite

0

1.04773

0

1.07810

50

0.34920

0.66680

0.45202

0.57082

171

4. Conclusions The MCNP5 reactivity results for the CANDU-SCWR model studied show relatively little sensitivity to the various room-temperature nuclear data files considered, especially for low net-neutron-leakage conditions. The Akeff values for different 2H data files increased with net leakage and were proportional to the change in net leakage, but did not depend on inter-cell leakage. Small systematic differences were observed between the cooled and voided configurations, due either to the associated change in net leakage or to the IAEA's re-evaluation of the S(a,(3) for ! H bound in H 2 0, which would shift the calculated coolant void reactivity accordingly. The pre-ENDF/B-VII data for 235U and 238U increased k ^ by only -1.5 mk, which is likely due to the reduced importance of 238U captures relative to ZED-2 simulations involving NU fuel. Additional work would be needed to extend this analysis to more realistic conditions, e.g., to incorporate data files that adequately reflect operating temperatures, including the neutronically important temperature difference between coolant and moderator in an operating pressure-tube reactor. Acknowledgments The authors wish to thank R.D. Mosteller, LANL; A. Trkov, IAEA; and J.V. Donnelly, Nuclear Safety Solutions Ltd; for their technical assistance and H. Khartabil and R. Speranzini, AECL for funding support. References 1.

X-5 Monte Carlo Team, MCNP — A General Monte Carlo N-Particle Transport Code, Version 5, Los Alamos National Laboratory report LA-UR03-1987 (2003). 2. R.D. Mosteller, J.M. Campbell and R.C. Little, Reactivity Impact of ENDF/B-VI Cross Sections for Deuterium in Heavy-Water Solution Benchmarks, submitted to 2005 Annual Meeting of the American Nuclear Society, June 5 - 9, San Diego, CA (2005). 3. International Handbook of Evaluated Criticality Safety Benchmark Experiments, NEA/NSC/DOC(95)03, OECD Nuclear Energy Agency (2004). 4. R.D. Mosteller, K.S. Kozier, J.M. Campbell and R.C. Little, Reactivity Impact of Deuterium Cross Sections for Heavy-Water Benchmarks, submitted to Int. Conf. on Mathematics and Computation, Supercomputing, Reactor Physics, and Nucl. and Biological Applications, Avignon, France, September 12-15 (2005).

172

5.

6. 7.

8. 9.

10.

D.E. Cullen, R.N. Blomquist, C.E. Dean, D. Heinrichs, M.A. Kalugin, M. Lee, Yi-Kang Lee, R.E. MacFarlane, Y. Nagaya and A. Trkov, How Accurately can we Calculate Thermal Systems?, INDC(USA)-107/UCRLTR-203892, IAEA, Vienna (2004). R. C. Little and R. E. MacFarlane, SAB2002 - An S(a,3) Library for MCNP, Los Alamos National Laboratory report LA-UR-03-808 (2003). A. Trkov and M. Mattes, On the Thermal Scattering Law Data for Reactor Lattice Calculations, Proc. Int. Conf. Nuclear Energy for New Europe, Portoroz, Slovenia, Sept. 6-9 (2004). R. C. Little, MCNP Neutron Library T16_2003, Los Alamos National Laboratory report LA-UR-04-4520 (2004). L. C. Leal, H. Derrien, and N. M. Larson, Recent Cross Section Evaluations in the Resonance Region at the Oak Ridge National Laboratory, Proc. Int. Conf. Nucl. Data Sci. Tech., Santa Fe, New Mexico, Sep. 26 - Oct. 1, 2004 (2005). J.D. Irish and S.R. Douglas, Validation of WIMS-IST, Proc. 23rd Ann. Conf. of the Canadian Nuclear Society, Toronto, Canada, June 2-5 (2002).

THE LEAD COOLED FAST REACTOR BENCHMARK BREST-300: ANALYSIS WITH SENSITIVITY METHOD VALERY SMIRNOV, VICTOR ORLOV NIKIET, Moscow,

Russia

ALEXANDRE MOUROGOV, DAVID LECARPENTIER EDF-R&D, Departement SINETICS, 1 Avenue du General de Gaulle 92140 CLAMART France TATIANA IVANOVA CNAM - Laboratoire de Physique, 292 rue Saint-Martin, 75141-Paris Cedex 03; Institut de Physique Nucleaire Orsay/CNRS, 15 rue Georges Clemenceau 91406 Orsay Cedex, France

Sustainable development of atomic energy will require development of new types of reactors able to exceed the limits of the existing reactor types in terms of optimum use of natural fuel resources, reduction in the production of long-lived radioactive waste, economic and safety competitiveness. Lead cooled fast neutrons reactor is one of the most interesting candidates with a potential to address these needs. BREST-300 is a 300 MWe lead cooled fast reactor developed by the NIKIET (Russia) with a deterministic safety approach which aims to exclude reactivity margins greater than the delayed neutron fraction. The development of innovative reactors (lead coolant, nitride fuel...) and fuel cycles with new constraints such as cycle closure or actinide burning, requires new technologies and new data from various disciplines: fuel types, fuel designs and fuel reprocessing. In this connection, the tool and neutron data used for the calculational analysis of reactor characteristics requires thorough validation, even if computational codes in Russia and France relies to the calculation of fast reactors' parameters and "fast" experiments. NIKIET developed a reactor benchmark fitting of design type calculational tools (including neutron data). In the frame of technical exchanges between the NIKIET and the EDF (Electricite De France), results of this benchmark calculation concerning the principal parameters of fuel evolution and safety parameters has been intercompared , in order to estimate the uncertainties and validate the codes for calculations of these new kind of reactors. Different codes and cross-sections data have been used, and sensitivity studies have been performed to understand and quantify the uncertainties sources.

173

174

1. Introduction This work initiated by NIKTET, has been performed in the frame of a technical partnership between this institute, EDF R&D, and with strong participation of the CNRS. In regard to EDF R&D, the main goal of this work is to create an independent opinion on potential of lead cooled fast reactors (LFR) as a GEN IV representative in general and on BREST concept [1] in particular. The second goal of this benchmark is to evaluate the suitability of European fast reactor calculation tool ERANOS 2.0 [2] code with nuclear data libraries (JEF2.2 or ERALIB1) for LFR simulation. The benchmark calculations have been performed using three different codes: ERANOS, FACT-BR (NIKIET) [3], TRIGEX (IPPE) [4]. A first phase was devoted to the analysis of the discrepancies between the Russian and French results. Afterwards, an attempt was made to find the origin of these differences using sensitivity analysis. The paper restricts this analysis for BREST-OD-300 keff value at the beginning of life (BOL), in spite of all the reactivity coefficients and evolution have been calculated using mentioned calculational chains and compared. This work is one of the first steps undertaken by EDF R&D in LFR study that allows more clearly understand the BREST concept's neutron features. It provides, also, an opportunity to formulate some needs in nuclear data improvements necessary for further study of this reactor type with ERANOS code. Namely this study will be concentrated on a modeling the fuel burn-up, the core reloading, and then, analysis of BREST-1200 concept as a commercial size reactor. 2. Benchmark Initial Data The BREST-OD-300 reactor model is characterized by a thermal power of 700 MW. The core is loaded with nitride fuel and cooled by lead. The main reactor characteristics are given in Table 1. In the BREST reactor the smoothing of the radial power distribution is realized by using three radial zones containing fuel pins with the same composition but different diameters. In the central zone the pins with the smallest diameter are located while the outer zone holds the fuel pins having the largest diameter. The X-Y layout of the reactor is shown on Figure 1. The core is surrounded by a lead reflector without any radial blanket. In this benchmark it was assumed that the reactor zones have homogeneous compositions and all control rods are withdrawn. The reactor's vertical cut is presented on Figure 2 where the temperature of each medium is also indicated. At the BOL the isotopic vector of the fuel is the same for the three core radial

175

zones (see Table 1), the isotopic composition is close to an equilibrium state and does not hold any fission product. Table 1. Main characteristics of BREST-OD-300 benchmark Parameter Value Reactor thermal power, MW 700 Trans-uranium isotopic composition, 66.08/26.97/4.07/1.69/0.70/0.49 239Pu/240Pu/241Pu/242Pu/241Am/243Am (% at) Uranium isotopic composition, 0.30/99.70 235U/238U (% at) Trans-uranium fraction in the fuel 13.45 (% at) Core height, cm 110 Sub-assemblies pitch, cm 16.9 Pin pitch, cm 1.3 Internal core Number of assemblies 45 Fuel pin outer diameter, mm 9.4 Intermediate core Number of assemblies 64 Fuel pin outer diameter, mm 9.8 External core Number of assemblies 36 Fuel pin outer diameter, mm 9.8 (internal rows), 10.5 (external row)

1

Extern IH-Ott *SH1I elleclo

Figure 1. BREST-OD-300 X-Y layout

176

Reactor vessel

Upper reflector 100 813K 813K Expansion chamber

60

Pellet of — WC

--• -h?r,l--r

25 t 110

•J23K

Radial reflector 693K

Bottom reflector

85

693K 0 230 0 326 0 430 Figure 2. BREST-OD-300 vertical layout with medium temperatures indication. The dimensions are given in cm.

3. Calculational Tools In the undertaken study the following two families of calculational tools were used: 1. The European software ERANOS 2.0 2. Russian packages: FACT-BR (NIKIET, Moscow) and TRIGEX (IPPE, Obninsk). The ERANOS includes the cell code ECCO and the core solver VARIANT. ECCO deals with the self shielding phenomena. In this study cross sections of the twenty two main resonant isotopes* were treated in 1968 groups, while 33 groups were used for the other nuclides. Finally, ECCO provides 33 groups nuclear data set for the core calculation. As initial input nuclear data for ECCO calculations two options were used: JEF2.2 and ERALIBl libraries. The * 2 3 5 U, 2 3 8 U, 239 Pu, 2 4 0 Pu, 241 Pu, 61 Ni, 6 2 Ni, M N i , C, 14 N

242

Pu,

24,

Am,

57

Fe, 5 8 Fe, 5 4 Fe, 5 6 Fe, 5 0 Cr,

52

Cr,

53

Cr,

54

Cr,

58

Ni, ^ N i ,

177

ERALIB1 set is an improved version of JEF 2.2 obtained by applying a statistical adjustment procedure. The experimental basis for the adjustment is provided by a large number of integral values (more than 350) measured on clean critical cores. In the frame of this benchmark the core calculation was performed by resolving the neutron transport equation in diffusion approach using 33 energy group structure by the VARIANT code in Cartesian (X-Y-Z) geometry and by applying the Variational Nodal Method. Russian tools TRIGEX and FACT-BR both operate with ABBN-93 crosssection library [5] based on the newest files of evaluated nuclear data contained in FOND2.2 and adjusted with a big number of integral and macroscopic experiments. The 28-, and 299-energy-group (for main isotopes) ABBN-93 cross sections are prepared for the calculations by the CONSYST [5] code, which determines the resonance self-shielding of the cross sections. Using these cross sections with either the FACT-BR and the TRIGEX 3-dimentional diffusion code, benchmark characteristics were calculated in 26 groups representation. In the frame of this benchmark the results predicted by FACT-BR package were considered as the reference ones for lead cooled systems calculations, because this software was qualified on a large number of integral critical experiments including lead [6]. 4. Results Comparison Calculated results of keff value obtained with the above benchmark-model data are presented in Table 2. Table 2. keff values comparison l

l

!

^

VMCT-BR ~ kUCODE ' • P c m k jS_ jS_

Code

Neutron data library

keff

FACT-BR

ABBN-93

1.00398

TRIGEX

ABBN-93

1.00244

-153

ERANOS 2.0

ERALIB 1

1.00039

-357

ERANOS 2.0

JEF2.2

0.99582

-816

MCNP

JEF2.2

0.9950*

-899

To ensure absence of methodological component of bias that comes from ERANOS calculation, a calculation of the same benchmark-model has been performed with MCNP code and JEF2.2 files. This result was provided by O. Laulan (CNRS/LPSC Grenoble). Statistical uncertainty (standard-deviation) of the calculation is 65pcm

178

The discrepancy on keff value between ERANOS code with ERALIB1 library and the reference (FACT-BR) does not exceed 360 pcm. This difference is of the same order of magnitude than the delayed neutron fraction, which is about of 370 pcm for BREST-OD-300 reactor. On the other hand, the discrepancy is more than twice bigger if JEF 2.2 library is used with ERANOS. So the analysis of the average cross sections values is necessary in order to learn the possible origin of the differences. A comparison of these values calculated using ERANOS code with both the ERALIB-1 and JEF 2.2 library with that corresponding to the reference case is given below. The average one group cross sections are calculated as followed : ITT

NG

jIZT

j

g

l

NG

i

s

in which i- isotope index, j - mesh index, IZT - total number of mesh in the zone of interest (the zones of interest: internal core, intermediate core, external core), Vj-mesh volume, g- group index, x- reaction index (capture of fission), o8- group cross section, cp8- group flux.

Table 3 Core: Comparison of average cross sections from ERANOS 2.0/ERALIB1 and FACT-BR calculations Of, barn

Gc, barn ISOTOPE

ERANOS

FACT-BR

U235 U238 PU239 PU240 PTJ241 PU242 AM241 AM243 PB FENAT CRNAT NINAT MO MN WNAT C N14

0.512 0.260 0.471 0.469 0.428 0.386 1.792 1.540 0.004 0.009 0.010 0.026 0.118 0.052 0.242 0.000 0.021

0.444 0.242 0.410 0.452 0.347 0.375 1.483 0.854 0.004 0.009 0.011 0.028 0.118 0.051 0.232 0.000 0.020

Relative difference,

ERANOS

FACT-BR

Relative difference,

1.799 0.034 1.765 0.364 2.349 0.249 0.253 0.198

1.700 0.036 1.688 0.370 2.258 0.260 0.293 0.230

5.84 -4.61 4.54 -1.54 4.03 -4.36 -13.81 -14.03

% 15.23 7.64 14.89 3.68 23.31 2.91 20.85 80.32 2.06 -0.64 -11.99 -8.83 0.40 2.15 4.10 0.00 3.43

179 Table 4 Core: Comparison of average cross sections from ERANOS 2.0 / JEF2.2 and FACT-BR calculations Of, barn

o"c, barn ISOTOPE

ERANOS

FACT-BR

Relative difference, %

U235 U238 PU239 PU240 PU241 PU242 AM241 AM243 PB FENAT CRNAT NINAT MO MN WNAT C N14

0.487 0.261 0.458 0.521 0.517 0.437 1.796 1.543 0.004 0.009 0.010 0.029 0.119 0.052 0.242 0.000 0.021

0.444 0.242 0.410 0.452 0.347 0.375 1.483 0.854 0.004 0.009 0.011 0.028 0.118 0.051 0.232 0.000 0.020

9.79 7.65 11.59 15.35 49.00 16.48 21.12 80.73 2.08 -2.27 -13.33 2.35 0.67 2.60 4.48 0.0 3.96

ERANOS

FACT-BR

1.810 0.035 1.752 0.377 2.400 0.261 0.251 0.197

1.700 0.036 1.688 0.370 2.258 0.260 0.293 0.23

Relative difference, % 6.5 -2.6 3.8 2.0 6.3 0.5 -14.3 -14.5

The differences between cross sections of the most important nuclides (i.e. presented in significant quantity in the core, as 238U, 239Pu, 240Pu, Pb, 14N) are analyzed below through sensitivity coefficients for BREST-OD-300 benchmark-model. They were calculated by TRIGEX code applying first order perturbation theory as the relative change of keff, due to a relative change of the cross sections by isotope, reaction and energy: k

eff I crfx the meaning of i, x, g being as it was noted above. The values of the integrated over the neutron energy sensitivities are presented on Figure 3. It was made attempt to quantify the contribution of difference in each isotope cross section to the difference in calculated keff value. Tables 5 and 6 present these contributions calculated as follows: _

^ai,x a

i,x

The analysis of the values presented in Tables 5 and 6 allows concluding that the effect is provided by 238U and 239Pu mostly.

180

I

It c

£.

i>

£>

Figure 3. Integrated sensitivity coefficients of k^f to neutron cross section for BREST-OD-300 benchmark

The total effect calculated trough the sensitivity coefficients does not fully agree with difference in keff presented in Table 2. Part of this disagreement comes, of course, from the omission of the energy dependence of cross sections, as the exercise was realized with one group cross sections and sensitivity coefficients values. Table 5 Main nuclides contribution to reactivity difference between ERANOS/ERALIB1 and FACT-BR calculation 238U 239Pu 240Pu 4.54 -1.54 -4.6 8o"f/rjf,% 510 40 SCf, pcm/% 90 Fission contribution -414 2315 -62 7.64 14.9 3.7 5o-c / o c , % SCc, pcm/% -230 -30 -10 Capture contribution -1757 -447 -37 1868 -99 Total effect by nuclide, pcm -2171 Total effect, pcm -489

Pb

14N

2 -10 -20 -20

3.4 -20 -68 -68

Table 6 Main nuclides contribution to reactivity difference between ERANOS/JEF2.2 and FACT-BR calculation 238U 239Pu 240Pu Pb 14N 2 -2.6 3.8 8o-f/rjf,% SCf, pcm/% 90 510 40 Fission contribution -234 1938 80 7.65 11.59 15.35 2.1 3.96 8a c /CTC ,% SCc, pcm/% -230 -30 -10 -10 -20 Capture contribution -154 -1760 -348 -21 -79 Total effect by nuclide, pcm -1994 1590 -74 -21 -79 Total effect, pcm -577

181

The tables above do not allow analyzing of Pb inelastic, elastic scattering cross section and mu-bar value that are responsible for its albedo and, accordingly, neutron leakage from the systems with lead in a reflector and large amount of lead in a core. Moreover, inelastic cross section is responsible for neutrons removal under 238U fission cross section threshold. More thorough analysis with sensitivity method performed by CNRS shows importance of lead cross sections evaluation for accurate prediction of BREST-OD-300 type reactor characteristics. On Figures 4,5 lead group-wise inelastic and elastic cross sections available in the JEF-2.2 and ABBN-93 libraries (supposing that the ERALIB1 and JEF2.2 lead cross sections are the same) are presented. 3 2,5

ABBN

i

JEF2.2

i

2

| ~ l _

]

1.5

1 0,5

;

rfC

0

12

I

11

- - - .ABBN

-

icci n

10 •

- • • • •

1 •Q

6 m m 1

1.E+04

1

1

1

.

1

1

1 — I —

1.E+05

|

1.E+06

1

,

,__1_J

i

i

,

1.E+07

Energy, eV

As it was mentioned above, behavior of the cross section in energy region above 300 keV influences the neutron removal under threshold of 23 8U fission reaction that is responsible for a number of fission in the fast energy region. Maximal contribution of this reaction gives inelastic cross section in energy range 1.4 - 6.5

182

MeV. It is seen on Figure 4 above that Pb inelastic cross section from JEF2.2 is overestimated in this region. 5. Conclusion The necessity to meet the requirements of deterministic safety approach realized in the concept under study demands neutron data that can provide high accuracy of its neutron characteristics prediction. Some features of BREST-OD-300 such as mononitride plutonium fuel, lead coolant and square lattice of fuel elements differ the reactor from sodium-cooled fast reactors. This does not permit to extend the existing calculational experience to high accuracy prediction of new concepts characteristics without proving its applicability. The comparison of results shows that the difference on keff value between ERANOS code with ERALIB 1 library and the reference is of the same order of magnitude than the delayed neutron fraction. On the other hand, the discrepancy is more than twice bigger if JEF 2.2 library is used with ERANOS. Analysis of discrepancies in calculation results reveals that the main effect is provided by the difference of nuclear data, namely 238U, 239Pu fission and capture cross sections and lead inelastic cross sections. This analysis does not pretend to be exhaustive, because to conclude about applicability of calculational tools for lead cooled fast reactor simulation, more precise study of appropriate integral and macroscopic experiments on lead contained systems should be performed. References 1. E.O. Adamov, V.V Orlov et al., Naturally safe lead-cooled fast reactor for large-scale nuclear power, Report, Moscow, 2001 2. G. Rimpault et al. The ERANOS code and data system for fast reactor neutronic analysis, Physor 2002, Seul, Korea, October 7-10, 2002 3. S.Barinov, A. Radkevich. Use of multigroup cross sections preparation system CONSYST/ABBN in program complex FACT-BR for threedimensional neutron-physical calculations of BREST-OD-300 reactor. Proceedings of meeting "Neitronika-99", Obninsk, IPPE, 1999 4. Seryogin A., Kislitsina T., Tsiboulia A. Annotation of TRIGEX.04 code system. Preprint IPPE - 2846, Obninsk, 2000 (in Russian) 5. RSICC DLC-182 "ABBN-90: Multigroup Constant Set for Calculation of Neutron and Photon Radiation Fields and Functionals, Including the CONSYST2 Program" 6. Tsiboulia A., Tocheny L., Ivanova T., contribution to these proceedings.

SENSITIVITY ANALYSIS OF NEUTRON CROSS-SECTIONS CONSIDERED FOR DESIGN AND SAFETY STUDIES OF LFR AND SFR GENERATION IV SYSTEMS KAMIL TUCEK, JOHAN CARLSSON, HARTMUT WIDER Joint Research Centre, Institute for Energy, P.O. Box 2 NL-1755 ZG Petten, The Netherlands

We evaluated the sensitivity of several design and safety parameters with regard to five different nuclear data libraries, JEF2.2, JEFF3.0, ENDF/B-VI.8, JENDL3.2, and JENDL3.3. More specifically, the effective multiplication factor, burn-up reactivity swing and decay heat generation in available LFR and SFR designs were estimated. Monte Carlo codes MCNP and MCB were used in the analyses of the neutronic and burn-up performance of the systems. Thermo-hydraulic safety calculations were performed by the STAR-CD CFD code. For the LFR, ENDF/B-VI.8 and JEF2.2 showed to give a harder neutron spectrum than JEFF3.0, JENDL3.2, and JENDL3.3 data due to the lower inelastic scattering cross-section of lead in these libraries. Hence, the neutron economy of the system becomes more favourable and keff is higher when calculated with ENDF/B-VI.8 and JEF2.2 data. As for actinide cross-section data, the uncertainties in the keff values appeared to be mainly due to 239Pu, 240Pu and 241Am. Differences in the estimated burn-up reactivity swings proved to be significant, for an SFR as large as a factor of three (when comparing ENDF/B-VI.8 results to those of JENDL3.2). Uncertainties in the evaluation of short-term decay heat generation showed to be of the order of several per cent. Significant differences were, understandably, observed between decay heat generation data quoted in literature for LWR-UOX and those calculated for an LFR (U,TRU)02 spent fuel. A corresponding difference in calculated core parameters (outlet coolant temperature) during protected total Loss-of-Power was evaluated.

1. Introduction The correct prediction of design and safety parameters (such as criticalities, reactivity coefficients, power profiles, hot channel factors, burn-up reactivity swings, and decay heat generation) is a necessary requirement for a reliable operation of Generation IV systems. In these systems, there is a strong incentive to allow for low burn-up reactivity swing, ideally below 1$ over an extended time. Then, the reactivity that could be accidentally inserted into the reactor is small and the consequences of reactivity-induced accidents can be limited. Thus, the accurate prediction of Akeff is particularly important. The amount of decay heat that has to be removed from the core is important for correct evaluation of protected Loss-of-Heat-Sink (LOHS), and total Loss-of183

184

Power (TLOP) accidents. Removal of decay heat is of primary concern for a Gas-Cooled Fast Reactor (GCFR), where special measures are considered (second guard reactor vessel, cold fingers etc) in order to protect the core from damage due to a depressurization accident. The dependence of calculated core parameters (criticalities, power distributions) on the choice of a cross-section library was shown to be notable, e.g. in the heavy metal cooled system due to the uncertainties in the inelastic scattering cross-section of lead [1]. In this study, we investigated a sensitivity of several design and safety parameters, effective multiplication factor, burn-up reactivity swing, and decay heat generation on the choice of neutron crosssection libraries, JEF2.2, JEFF3.0, ENDF/B-VI.8, JENDL3.2, and JENDL3.3. Using decay heat generation data, a spread in calculated temperatures for a TLOP accident in LFR was also evaluated. 2. Model An efficient use of fissile fuel resources together with the ability to burn its own high-level waste and those coming from LWRs are primary design goals of new reactor designs developed under the auspices of the Generation IV initiative [2]. Both self-breeder (conversion ratio ~ 1) and TRU "burner" design options are hence considered. The former system is intended to operate in a pure fast reactor scheme while the latter will work in concert with LWRs or, in the double-strata scenario also together with dedicated minor actinide burners (ADS). As the purpose of self-breeder and burner is different, their fuel composition will differ and hence also their neutronic and safety characteristics. In self-breeder reactor designs, the inert matrix is omitted and U/TRU ratio is chosen such that the breeding gain is close to zero. In order to keep the Doppler reactivity coefficient high and coolant temperature reactivity coefficient low, the amount of minor actinides in (equilibrium) fuel should be limited to about 2.5% [3]. Design parameters of the considered LFR and SFR self-breeder cores are listed in Table 1. For the LFR, the pellet and pin designs are similar to those proposed for CAPRA cores [4]. The pitch-to-diameter (P/D) of the core lattice is 1.8, allowing for low-pressure drop across the core, thus increasing safety margins in Loss-of-Flow (LOF) accidents [5]. The maximum lead velocities of 2.3 m/s are well below the design limits of 2.5-3 m/s; the maximum steady state cladding temperature is kept under 753 K, ensuring the long-term stability of the structural material. The SFR design parameters are preliminarily taken as those envisioned for the WAC benchmark reactor [6]. In this case, however, the axial and radial breeding blankets were removed. As sodium allows for higher coolant

185

velocities than lead, tight pin lattice with P/D = 1.2 can be applied. However, this has the drawback that it increases the pressure drop. In both cases, the fuel is of the ceramic type, (U,TRU)02. The TRU vector composition is that of spent LWR UOX fuel after it has undergone 30 years of cooling, see Table 2. Depleted uranium (0.3% 235U) is used in the analyses. Table 1. Design parameters of SFR and LFR self-breeder core concepts considered in this study. Parameter LFR SFR Pellet outer radius (mm)

2.4

3.0

Cladding inner radius (mm)

2.5

3.1

Clad outer radius (mm)

3.0

3.45

Pitch-to-diameter ratio

1.8

1.2

20.10

14.66

Pins per S/A

331

271

Length of upper plenum (cm)

100

100

Length of lower plenum (cm)

10

10

Active pin length (cm)

200

100

Number of S/A Number of channels in the individual enrichment zones Averaged linear power (kW/m)

289

216

4/3/3

5/3

7.5

24.3

Peak linear power (kW/m)

12.7

40.4

S/A outer flat-to-flat (cm)

Table 2. Plutonium and minor actinide vector corresponding to the LWR-UOX spent nuclear fuel with burn-up of 41 GWd/tHM after 30 years of cooling. Isotope Isotope fraction 5

U U 237Np ™Pu 239 Pu 240 Pu 241 Pu 242 Pu 241 Am 243 Am 8

0.003 0.997 1.000 0.023 0.599 0.264 0.040 0.074 0.871 0.129

186

3. Method The Monte Carlo code MCNP4C3 [7] was used for the calculation of the effective multiplication factor and neutron spectra. Burn-up reactivity swing and decay heat were evaluated with MCB1C [8]. Nuclear data libraries were adjusted for the temperature dependence by the NJOY99 code. For individual components of the heterogeneous core model, the averaged steady-state temperatures were assumed as follows: 1500 K for fuel, 900 K for cladding, and 600 K for coolant. The thermal hydraulic calculations were performed by the STAR-CD CFD code [9]. 4. Results 4.1. Neutronics The effective multiplication factor (keff) was first evaluated for our model LFR and SFR configurations using different cross-section libraries. The results of this study are summarised in Table 3. For the LFR, a considerable difference in keff is observed when comparing results obtained with JEFF3.0 and JENDL3.2 data to those of JEF2.2 and ENDF/B-VI.8 libraries. To confirm the hypothesis that substantial difference in k<.ff is due to the lead cross-sections, neutron data for lead in JENDL3.2 library were exchanged for those coming from ENDF/B-VI.8. In this case, the calculated value of k^f is 1.00422, much closer to the figure obtained with ENDF/B-VI.8 data (k^f = 1.01198) than with JENDL3.2 library (keff = 0.98885). Note also that in the sodium-cooled systems, the difference in keff between investigated sets of cross-section data is somewhat less pronounced than for LFR.

System LFR SFR

Table 3. Effective multiplication factors (keff) for different crosssection libraries, keff values are given together with 1-a standard deviation. The ones for the SFR are larger due to less calculation time. JENDL3.2 + ENDF/B-VI.8 JEF 2.2 ENDF/B-VI.8 JEFF 3.0 JENDL3.2 (Pb) 1.00836 1.00422 1.01198 0.98528 0.98885 ± 33 pcm ± 37 pcm ± 26 pcm ± 34 pcm ±31 pcm 1.00065 1.01651 1.00399 1.00964 ± 66 pcm ± 60 pcm ± 63 pcm ± 67 pcm

For LFR, ENDF/B-VI.8 & JEF2.2 libraries yield harder neutron spectrum than JENDL3.2 and JEFF3.0, see Figure 1. The largest difference in neutron spectra is observed in the energy range of 1-4 MeV, where inelastic scattering plays an

187

important role in the neutron transport and slowing-down. In light of these results, we observe indeed large differences in the inelastic cross-section for lead isotopes in this energy range, particularly regarding the inelastic cross-section of 207 Pb, see Figure 2. Energy distribution of flux in LFR [relative units] (a)

0.07 0.05

ENDF/B-VI.8 - - - JEF2.2 --•- JEFF3.0 JENDL3.2 c-=V

0.03

_ ^ \

j

Energy [MeV] 0.5

1

/

\ 5

10

&

J

0.02

\

-?a jf

/*v

0.03

r 0.2

ENDF/B-VI.8 - - - JEF2.2 •••• JEFF3.0 JENDL3.2

0.05

j\y-\

0.02

Energy distribution of flux in SFR [relative units] (b)

0.07

0.1

0.2

\ %L

\

\

Energy [MeV] 0.5

1

j 5

10

Figure 1. Energy distribution of the group flux for LFR (a) and SFR (b).

Inelastic scattering for Pb-207 [barn]

/

'

' JEFF3.0

j

- - ENDF/B-VI.8

J/

^S*^N

1 1

/ / Energy [MeV] 10

Figure 2. Inelastic scattering cross-section of 207Pb taken from JEFF3.0 and ENDF/B-VI.8 libraries.

The neutron economy in the system is influenced by changed neutron spectra, particularly due to the alteration of actinide fission-to-absorption probabilities. In Table 6, we display fission probabilities for two even-neutron number actinides present in the fuel, 238U and 241Am. For both nuclides, ENDF/B-VI.8 and JEF2.2 yield higher fission probabilities than for JEFF3.0 and JENDL3.2. For 238U, a spectral nature of this phenomenon becomes discernible, when

188

comparing oYaa for LFR and SFR. However, in the case of 241Am, the uncertainty in its capture cross-section data is of significance, too, see Figure 3. Table 4. Fission-to-absorption ratio for 238U and 241Am for different crosssection libraries. The relative (1-6) standard deviation is less than 1%. System

Isotope ~ mU

JEF 2.2

ENDF/B-VI.8

JEFF 3.0

JENDL3.2

0.103

0.103

0.091

0.090

0.110

0.129

0.099

0.109

0.150

0.155

0.156

0.150

0.129

0.156

0.134

0.142

LFR 'Am SFR 'Am

3 2.5

Am-241 capture cross-section [barn] (a)

Am-241 fission cross-section [barn] (b)

2 1.5

JEFF3.0 - -

1

ENDF/B-Vl.f

0.5 Energy [MeV] 1 Figure 3. Capture (a) and fission (b) cross-section of Am taken from JEFF3.0 and ENDF/B-VI.8 libraries. Note a significant difference in 241Am capture data.

Differences in the neutron spectra and fission-to-absorption ratios then manifest themselves as differences in the actinide fission probabilities. In Table 5, the effective multiplication factors calculated for LFR systems were compared inbetween individual cross-section libraries. Akeff are broken down according to the contributions from the individual actinide isotopes. When comparing ENDF/B-VI.8 results with JEFF3.0 and JENDL3.2, the uncertainties in k ^ are mostly due to even-neutron number nuclides, 238U, ^ ^ u , 241Am, revealing again a partial contribution of the spectral effect. Note that when the lead crosssections in the JENDL3.2 library were exchanged for ENDF/B-VI.8 data (last column of Table 5) the fission probability of 238U becomes similar to that of

189

ENDF/B-VI.8. In all cases, the contribution due to 239Pu remains also important, especially when comparing effective multiplication coefficient corresponding to ENDF/B-VI.8 and JENDL3.2 data to that of JEFF3.0. Table 5. Breakdown of the differences in the effective multiplication coefficient for LFR core into individual actinide isotopes. Values of Akeg are given in pcm. Akeff

235 238

ENDF/B VI.8 - J E F F 3.0

ENDF/B VI.8 - JENDL 3.2

ENDF/B V1.8 -JEF2.2

JENDL 3.2 - JEFF 3.0

ENDF/B VI.8 -(JENDL 3.2 + ENDF/B VI.8(Pb))

u

-3

10

12

-12

20

U

760

740

65

33

24

237

Np

123

133

83

-5

28

238

Pu

-130

64

39

-165

23

836

238

239pu 240

Pu

241 242

-171

613

775

185

-12

279

-92

-122

54

-23

-32

Pu

70

137

42

-54

16

Am

329

415

42

-64

171

Am

36

38

11

-1

9

2670

2313

362

357

776

241 243

Pu

122

741

Total

4.2. Burn-up reactivity swing In both LFR and SFR, the calculated burn-up reactivity swing is distinctly lower for JEFF3.0 and JENDL3.2 data, see Table 7. Note also that for an SFR the burn-up swing differs by more than a factor three (comparing ENDF/B-VI.8 results to those of JENDL3.2). A decomposition of the burn-up reactivity swing into contributions from individual actinide isotopes is currently underway. Note, e.g. that the lower capture rate (higher 0"f/o"a) of 241Am, observed for ENDF/BVI.8 library data, would mean that the breeding/production of 238Pu diminishes, which in turn, exacerbates the differences in the reactivity swing. Table 6. Bum-up reactivity swing for LFR and SFR self-breeders as a function of neutron cross-section libraries used in the calculations. 1-q standard deviation is provided. System JEF2.2 ENDF/B-VI.8 JEFF 3.0 JENDL 3.2 758 492 577 794 LFR 46 pcm ± 42 pcm ± 40 pcm ± 42 pcm 802 318 253 601 SFR 80 pcm ± 89 pcm ± 79 pcm ± 85 pcm

190

4.3. Decay heat generation Uncertainties in the decay heat generation as estimated in our calculations for different cross-section libraries are due to the difference in the composition of the irradiated fuel. The MCB code is using the MCNP4C subroutines for neutron transport calculations and evaluation of all necessary physical quantities required for burn-up calculations. In this code, incident energy and nuclide dependent fission product yields have been prepared based on the Wahl model. The energy of decay is taken from the ORIGEN library and the decay schemes of all possible nuclides and their isomeric states are formed and analysed on the basis of decay data taken both from Table of Isotopes and ORIGEN. The nuclides are divided into few groups depending on their decay half-life. Note that the nuclides that fall into the fast decaying group (half-life < 15 s) are treated in a simplified manner, i.e. assumed to decay immediately after their appearance. The uncertainties in the short-term decay heat generation appeared to be a general function of decay time and they are of the order of several per cent, see Figure 4. 80 r 70

Decay heat after 1-y irradiation time [MW] (a)

10

Difference in decay heat

7.5

60

— -.-....

1111.

JEF2.2/ENDF/B-VI.8 JENDL3.3/JEF2.2 JENDL3.2/JEF2.2 JEFF3.0/ENDF/B-VI.8

— 1 - JEFF3.0/JEF2.2

I%](b) 2.5 :-r---:-r.-rn 5

50 ENDF/B-VI.8

40

JEF2.2

30

2.5

20

-5

10

7.5

Decay time [s] 1

100

10000

6

10

Decay time [s] 1

100

10000

106

Figure 4. Decay heat generated by LFR spent fuel after a 1-y burn-up (a). The uncertainties in the decay heat calculation are given in (b).

In Figure 6(a), we compare decay heat generation as calculated for (U,TRU)02 LFR spent fuel after a 1-y burn-up and that given in literature for LWR-UOX fuel. Understandably, due to higher TRU content in the spent LFR fuel, the decay heat is larger than for LWR-UOX.

191 0.0/

0.05

1400

Decay heat after 1-y irradiation time [relative to nominal power] (a)

0.06 * \ \

Outlet coolant temperatures atTLOP[K](b) 1200

LWR, Pershagen

LFR, 1 -y burnup

_ ^ j g 0 ^ ^ ^ ^ ^

-i^

0.04

1000

\ \

0.03

\

yf

JMP^

LWR. Pershagen

^ U - F R , 1-yburnup

0.02

800

0.01 Decay time [s] 1

100

^ ^ ^ ^ ^ ^ ^ 10000

10B

Time [s] 10B

50000

100000

150000

Figure 5. Decay heat from LWR-UOX and LFR spent fuel (a). Corresponding differences in outlet coolant temperatures during protected total Loss-of-Power accidents are displayed in (b).

A corresponding difference in the STAR-CD calculated outlet coolant temperatures during protected TLOP accidents is given in Figure 6(b). At 150000 s (~ 42 h), core outlet coolant temperatures differ by about 100 K. In this case, we assume that all off-site electricity becomes unavailable and that on-site diesels cannot be started. The reactor is scrammed and neither the coolant pumps nor the heat exchangers are assumed to work. Decay heat is removed by natural air convection and thermal radiation from the vessel outside, by a so-called Reactor Vessel Auxiliary Cooling System (RVACS). 5. Conclusions In this study, we estimated the uncertainty in predicting the effective multiplication factor, burn-up reactivity swing and decay heat generation in model LFR and SFR systems. A strong dependence of the calculated values on the choice of the cross-section library was confirmed. For the LFR, ENDF/BVI.8 and JEF2.2 gave harder neutron spectra due to their lower inelastic scattering cross-section of lead. As for actinide cross-section data, the uncertainties/differences in the keff values appeared to be mainly due to 239Pu, 240 Pu and 241Am. The uncertainty in the calculated reactivity swing was more pronounced for the SFR, being as large as a factor of three for a 1-y burn-up (comparing ENDF/B-VI.8 results to those of JENDL3.2). This implies a large uncertainty in predicting the excess reactivity that need to be present in the reactor at BOC. In terms of the short-term decay heat generation, the uncertainties in the calculated values are of the order of several per cent. Understandably, significant differences between decay heat generation data quoted in literature for LWR-UOX and those explicitly calculated for GEN-IV

192

systems were observed. Corresponding outlet coolant temperatures during protected total Loss-of-Power accidents were calculated. References 1. "Comparison Calculations for an Accelerator-driven Minor Actinide Burner", Technical report NEA/NSC/DOC(2001)13, OECD/NEA (2001). 2. "A Technology Roadmap for Generation IV Nuclear Energy Systems", U.S. DOE Nuclear Energy Research Advisory Committee and the Generation IV International Forum, Technical report GIF-002-00 (2002). 3. "Accelerator-driven Systems (ADS) and Fast Reactors (FR) in Advanced Nuclear Fuel Cycles", A Comparative Study, OECD/NEA (2002). 4. A. Conti, et al., "CAPRA exploratory studies of U-free fast Pu burner cores", Proceedings of the International Conference on Evaluation of Emerging Nuclear Fuel Cycle Systems, GLOBAL'95, ANS, Versailles, France (1995). 5. K. Tucek, et al., "Comparison of Sodium and Lead-cooled Fast Reactors Regarding Severe Safety and Economical Issues", to be published in ICONE13, Beijing, China (2005). 6. H.U. Wider et al., "Comparative analysis of hypothetical loss-of-flow accident in an irradiated LMFBR core using different computer models for a common benchmark problem", Technical report EUR 11925, ISBN 92-8259471-8(1989). 7. J.F. Briesmeister, editor, "MCNP - A general Monte Carlo N-Particle transport code", version 4C, Technical report LA-13709-M, Los Alamos National Laboratory, USA (2000). 8. J. Cetnar, et al., "MCB - a continuous energy Monte Carlo Burnup code", in Fifth international information exchange meeting, OECD/NEA, Mol, Belgium (1998). 9. Computational Dynamics Ltd., Methodology Volume 3.20 (2004).

EXPERIMENTS

This page is intentionally left blank

INL CAPABILITIES FOR NUCLEAR DATA MEASUREMENTS USING THE ARGONNE INTENSE PULSED NEUTRON SOURCE FACILITY J.D. COLE, M.W. DRIGERT, R. ARYAEINEJAD, D.W. NIGG Idaho National Laboratory, Idaho Falls, ID, USA R.V.F. JANSSENS, B.J. MICKLICH Argonne National Laboratory, Chicago, IL, USA G. TER-AKOPIAN Joint Institute for Nuclear Research, Dubna, Russia

1. Introduction The relevant facts concerning the Argonne National Laboratory - Intense Pulsed Neutron Source (ANL/IPNS) and the Idaho National Laboratory (INL) apparatus for use at the ANL/IPNS facility to measure differential neutron interaction cross sections of interest for advanced reactor physics applications are presented. The INL apparatus, which consists of an array of multiple types of multiple detectors operated in coincidence, signal electronics, and a data acquisition system, is presented as an application of new means and methods to measure the relevant parameters described. The immediate measurement goals involve measurement of neutron induced interaction cross sections for 240Pu and 242Pu with 241Pu, 241 Am, with measurements for other nuclides of interest for advanced reactor physics applications to follow later. Specific uncertainties and error limits are presented and methods for controlling these uncertainties are described. The post experiment analysis using data sorts and data selection from a large, selfconsistent data set to produce spectra that will be analyzed for direct results and used to determine cross sections is also discussed. 2. Neutronic Performance of the IPNS Facility The ANL/IPNS facility is a spallation neutron source with a moderated neutron beam that has a neutron spectrum at 12 and 20 meters given in the Figure 1. The curves shown in Figure 1 are the results of average measured intensities and MCNP calculations performed at IPNS1. For perspective, a direct comparison of flux intensity can be made between IPNS and the Los Alamos Lujan Center at 195

196

lower neutron energies where explicit numbers, in similar units, are available, in particular for epithermal neutrons. A measurement of the neutron spectrum at IPNS was performed and the epithermal neutron flux was determined using activation of a gold foil . The value obtained for the epithermal angular current per unit energy at 1.0-eV, for the "F" moderator is 2.91 x 1010 n/sr-uA-sec-eV. Converting this expression to comparable units results in an IPNS flux at epithermal energies of approximately 0.005 n/sr/proton/eV. Recent papers1'2 have reported neutron spectra at the Lujan Center for several beam lines. The reported neutron intensities at leV range from 0.001 to 0.005 n/sr/proton/eV. For the Lujan flight path FP14, the corresponding number is 0.001 n/sr/proton/eV. Moreover, the proton current for IPNS is 15 uA, while that at Lujan is 100 p.A. Consequently, the two facilities are very nearly the same in the epithermal range. Both LANSCE and IPNS are spallation sources that produce around 20 and 9 neutrons respectively for every proton that strikes a nucleus in the target. The plot shown in Figure 1 also provides the referenced spectral information for Lujan. —1FN3FI-20IT1

IFNSF1-12m

—Q02B6V

2NBV

L#BCEFP14-20m —

10H/W

1.CBC8 1.CB07 -ijr1.G&Q5 1.0&O3 %1.0&O1 S 1.C601 1.0ECS 1.0EO5 1.0EO7 1.0EO1

1.0EC2

1.0&00

1.C&C2

1.0&W

1.0BC6

1.0&CE

n a t a l energy^

Figure 1 Plot of the neutron spectrum on the F3 beam line used for the INEEL apparatus for distances of 12 and 20 meters and the extrapolated values for FP143,4 at LANSCE. Energy markers are shown for the readers benefit.

This flux information is used in our calculation of expected event rates, design evaluation, and other efforts to configure the experimental apparatus. In general the neutrons of energy below 0.001 eV and above 10 MeV are of low enough

197

intensity to be neglected in the experiment and are of little interest in fission reactor applications in any event. The low-energy neutrons can also be filtered with a cadmium absorber to remove the neutrons below 0.4 eV if desired. The integrated intensity as flux (neutrons/cm2 s) is given in Table 1. Table 1 values are the actual flux values at the target location as used in the experiment. Table 1. Integrated neutron flux for the F-moderator beam lines for 12 and 20 meters (neutrons/cm2 s) F2

Fl

F3

12 m

20 m

12 m

20 m

12m

20 m

3.73E+08

1.34E+08

3.99E+08

1.44E+08

3.73E+08

1.34E+08

The uncertainty of the energy of a neutron that induces an interaction of interest in the target is determined largely by the uncertainty on the n-TOF from the time width of the proton pulse. In addition to this important parameter, the pulse rate and the flight path length interact to limit the useable energy range of the neutrons. At IPNS the pulse rate is 30 Hz and the proton pulse full width is 70 ns1. All beam lines at IPNS are heavily shielded and evacuated so that backgrounds are reduced. The low background at IPNS is an important factor for the long runs needed for high statistic experiments. 3. The INL Program The INL apparatus was originally installed at IPNS to perform experiments using induced fission of actinide targets for prompt information concerning fission yields by isotope pairs, nuclear structure information for prompt deexcitation of the fission products, multiplicity of both neutron and gamma rays by isotope pairs, and isotope pair distributions for fission cluster models. These efforts are extensions of spontaneous fission studies on Cf and Pu conducted with arrays of HPGe detectors at INL, Oak Ridge National Laboratory (ORNL), Lawrence Berkeley National Laboratory (LBNL), and finally on GAMMASPHERE at both ANL and LBNL. This work has produced over 100 publications on the nuclear structure of fission products prior to beta decay, fission yields by isotope pairs, and explicit neutron multiplicity as correlated to specific fission pairs. Two years ago efforts began to modify the INL apparatus at IPNS to provide the capability to measure neutron interaction cross sections (fission, capture, and inelastic neutron scattering) as a function of incident neutron energy, branching rations for the production of different isotopes by neutron

198

capture or fission, and cross sections for the production of independent yields from actinide fission. All of the work is for actinides, to be produced and burned in a new generation of reactors being developed by the United States Department of Energy (DOE). Some of these reactors are to have harder or higher energy neutron spectra than the current generation of thermal reactors, with much higher anticipated burnup, making the various transuranics of significantly greater importance than has been the case previously. The goal of the work is to produce results that have statistical uncertainties of 3% -5% or less and maximum neutron TOF uncertainties of 10%. In the resonance energy range of a few eV to -100 keV the TOF uncertainties are much smaller, ranging from 0.2% to - 2 % for both fission and capture cross sections. The INEEL apparatus for measuring event spectra for actinide targets of interest at IPNS is composed of an array of detectors of multiple types and multiple numbers of each type. There is a data acquisition system based on VME architecture with standard NIM and CAMAC electronics to acquire data on an event-by-event basis and store it in list mode for later analysis. A separate computer system is used for a multi-stop Time-to-Digital Converter (TDC) to acquire a neutron TOF spectrum from a fission chamber with high purity 235U foils. The fission chamber consists of six thin U foils arranged in a ring surrounding the volume that the neutron beam passes on its way to the experiment target position. Key features of the apparatus and experimental procedures are described in the following sections. 3.1. Online Neutron Flux Measurement The fission chamber and associated TDC are used to determine the incident neutron spectrum from the IPNS neutron production target in the same IPNS beam line as the primary detector array and as a real-time neutron flux monitor. This is a continuous measurement, as IPNS is an accelerator driven facility and over long run times the intensity of the neutron flux varies. The direct neutron spectrum is determined using the ENDF 235U fission cross section. It is possible to report results based on a ratio of the observed neutron event spectrum from the actinide target in the main array and the neutron event spectrum of the fission chamber. For both the incident neutron flux determination and the neutron event spectrum from the fission chamber, the error from the 235U cross section, which ranges from 3% to 5% itself, must be included in the final uncertainty. This is combined with our experimental error as a standard square root of the sum of the squares of the independent error. The statistical error from the fission chamber is less that 1% over the TOF ranges as this fission chamber has higher neutron

199

event rate due to the higher fission chamber efficiency than the multi-detector array. The timing pulses, both start and stops, are generated from a leading edge discriminator set such that the walk or jitter of the start pulse, the to from the accelerator, and of the stops, the pulse from the fission chamber, have the same level and electronic walk. A calibration pulse from the gamma ray flash produced in the neutron-generating target is feed into the TOF spectrum to provide an absolute calibration. The differential linearity of the TDC is < ±1% full range. The resolution in time of each bin can be set to 1 ns and the total number of bins set a 33,000,000 to cover the total time interval between pulses. The important point is that the 33 millisecond time between pulses, i.e. the entire neutron energy range of interest, can be covered in the experiment in one measurement. It is not required to make separate measurements covering only a small time range or use very coarse resolution to cover the energy range in the time interval and then match the independent measurements together. The primary uncertainty in the TOF is determined by the proton pulse width at energies above a few eV, and since data are collected on an event-by-event basis, the uncertainty due to detectors, electronics or cabling can be corrected in software. This equipment and method allows a model-independent measurement of the neutron interaction cross section to be made over a continuous energy range from a few meV to above 2MeV without breaking the measurement into different energy sections with an time resolution as given in Table 2, and the neutron event spectrum uncertainty determined primarily by statistics. With approximately 4000 hours of beam time available for measurements in one year, low statistical error is easily achieved where as continuous and long runs at other facilities are not as readily possible. The integrated neutron flux from the fission chamber is fed into a VME scalar module that is periodically readout by the main data acquisition system and the results are injected into the recorded event stream. This way the incident neutron flux is monitored throughout the course of the cross section measurement. The overall system dead time is determined by comparing the number of master event triggers generated by the front-end electronics with the number of events processed by the data acquisition system. These numbers are recorded in the data acquisition log as well as a periodically updated process monitor file. This monitor file also contains information of various error conditions seen by the data acquisition software. These include bad event rates for the different detectors that make up the measurement system.

200 Table 2. n-TOF and error for selected neutron energies and two flight path lengths at IPNS F3 neutron energy Flight path length (eV) 12_m 20 m flight time (us) pulse width flight time (us) pulse width error error 0.001 2.74E+04 0.17% 4.57E+04 0.10% 1.45E+04 0.01 8.67E+03 0.17% 0.10% 0.1 2.74E+03 0.17% 4.57E+03 0.10% 1 8.67E+02 0.17% 1.45E+03 0.10% 5 3.88E+02 0.17% 6.47E+02 0.10% 2.74E+02 10 0.17% 4.57E+02 0.10% 50 1.23E+02 0.18% 2.04E+02 0.11% 100 8.67E+01 1.45E+02 0.11% 0.19% 500 3.88E+01 0.25% 6.47E+01 0.15% 1000 2.74E+01 0.30% 4.57E+01 0.18% 5000 1.23E+01 0.59% 2.04E+01 0.36% 10,000 8.67E+00 0.82% 1.45E+01 0.49% 50,000 3.88E+00 1.81% 6.47E+00 1.09% 100,000 2.74E+00 2.56% 4.57E+00 1.53% 500,000 1.23E+00 5.71% 2.04E+00 3.43% 1,000,000 8.67E-01 8.07% 4.84% 1.45E+00 2,000,000 6.13E-01 11.41% 1.02E+00 6.85% 5,000,000 3.88E-01 18.05% 6.47E-01 10.83% 7,000,000 3.28E-01 21.35% 5.46E-01 12.81% 8,000,000 3.07E-01 22.83% 5.11E-01 13.70% 10,000,000 2.74E-01 25.52% 4.57E-01 15.31%

3.2. Neutron Event Spectrum Determination A second multi-stop TDC is used obtain the neutron TOF for the neutrons observed in the main array. The multi-stop TDC produces a TOF for each neutron event, a trigger, during the 33-millisecond interval between proton pulses. One or more of the triggers, discussed below, are used as a stop for this second TDC. The TDC has user selectable time bins with a total number of bins of 2 with the minimum time per bin being 1 nanosecond. A spectrum containing 30,000,000, 1-nanosec bins would allow maximum coverage of the time interval and the minimum time per bin. In actual use a smaller number is used depending on the energy resolution for the spectrum that is desired and what energy range of incident neutrons is of interest. The point is that the time bins can be set such that they have minimum impact on the energy resolution and the proton pulse width is still the dominant source of error. Although the time bins can be summed at a later point, the content of these bins, which represents the number of events observed in that time interval, provide the basis of the statistical accuracy for determining the incident neutron event intensity at each point. Where as this spectrum can have several million

201

channels, the determination of the flux spectrum requires only a few thousand points. The condition to be met in the experiment is that this statistical error for each of the compressed spectrum bins is 3-5%. This means that each compressed bin must have on the close order of 1000 counts or more above background. This gives a total event number of -50,000,000 for a 50,000 channel spectrum. This is not a difficult number to obtain with the INL equipment and our goal for any actinide target is -1,000,000,000 events. 3.3. The Detector Array The INL array is composed of 12 Compton suppressed, high purity germanium (CSHPGe) detectors, eight fast neutron detectors (BC501 liquid scintillator), and a stack of 32 Silicon (Si) detectors interleaved with double-sided foils of actinide targets. The trigger electronics starts the digitization process if: two of the CSHPGe; or two neutron detectors; or a CSHPGe and a neutron detector produce a signal within a set coincidence time window. In this way three separate conditions can be used as independent triggers (Tl, T2, and T3 respectively) for determining a neutron interaction has occurred in a target. The coincidence is overlap timing with a time window of 50 to 100 nanoseconds. The 32 Si detectors are used to directly detect the fission fragments as these fragments recoil directly into the Si detectors. The Si detectors and the target foils are interleaved, each actinide foil has a selected thickness that will allow the low energy, light mass fission fragment to escape the target and enter the Si detector, which is in contact with the target material. A discriminator is used to reject a-particles and their pileup signal and accept only the fission fragment signal, which is a factor of ten greater in amplitude. These fission fragment signals in the Si detectors are used as the fourth trigger (T4) in the system. Si detectors are used instead of a fission chamber primarily because the rise time of the output pulse is faster by roughly a factor of ten. In addition, since the Si detector is in contact with the actinide target, gamma rays observed from the fragment have no Doppler shift or broadening due to emission in flight. There are other advantages with size, less support electronics, better a-particle discrimination, and low mass material. The energy output of each Si detector is also digitized and included in the data packet. Figure 2 presents a drawing of a double detector unit on the left and the right shows an assembly of four double detector units. These Si detectors have been made and tested with 252Cf sources. A paper is in preparation but relevant results are that operation has been tested with a-particle doses of greater than 1012 a/cm2 and fission fragment doses of greater than 1010 fragments/cm2 without deterioration of signal and a rise of leakage current increase of less than a factor

202

of ten. The testing of these detectors continues to determine the failure limits but already these numbers indicate that the detectors will survive beyond our experimental statistical goal. Further tests with 239Pu targets and a Si detector stack are currently underway.

Figure 2 The left side shows a pair of Si detectors mounted onto a frame. 1 is the Si detector, 2 the Al frame, 3 the alignment bushing, 4 the hole in the frame, and 5 the detector contacts. The other numbers are the dimensions in millimeters. The right shows a stack of four detector units (two detectors each) with interleaved target foils assembled on a mounting frame for placement in the beam. The current plan would use 16 frames of 32 detectors and interleaved target foils.

The four triggers Tl, T2, T3, and T4 are "OR'ed" logically to produce an event signal. This event signal will be used to trigger data acquisition of digitized detector signals, time relationships between the detector signals, and as a stop on a multi-stop TDC as described above. In this way a multi-parameter data packet is acquired, for each radiation event detected and stored in list mode format. Since the system is configured to respond to coincidence events and the prompt timing of an event is from 10"22 seconds to 10"12 seconds, a single trigger can result in multiple radiation types being included in the event. The simplest example is that of a fission fragment being detected in a Si detector, a T4 event, and single gamma ray or neutron radiation also observed. Consider the following examples of events and what they mean.

203

3.3.1. Gamma rays alone Two or more gamma rays observed in two or more CSHPGe detectors within the time window will result in a valid Tl trigger. If no other triggers are present, this event would be acquired as a simple gamma-ray coincidence that could arise from beta decay of a fission fragment, a neutron capture event, or an inelastic scatter. 3.3.2. Neutrons alone A coincidence pair in the neutron detectors (T2) can be caused by a gamma ray observed in a neutron detector and a scattered neutron. The neutrons and gamma rays in the liquid scintillator detectors are distinguished by pulse shape discrimination in post analysis. This type of event can be due to inelastic scattering of the neutron or simply two coincident gamma rays being seen in the neutron detectors. For this to be a valid event that can be used in post analysis, two neutrons must be observed with a Si detector fission fragment. In other words a T2 must be present with a T4 to be processed in analysis. A usable elastic scatter event needs to be a T3 with the gamma ray observed in a CSHPGe detector. 3.3.3. Neutrons and gamma rays together This is a T3 event but it can be valid without a T4 Si detector signal. For an inelastic event discussed above, it is only used if the gamma ray is observed in a CSHPGe detector. If multiple neutron detectors and CSHPGe detectors are present as a coincidence event, this is valid only if a Si detector also "sees" a fission fragment. 3.3.4. Si Detector Triggers -T4 With the Si detectors and the actinide target material in contact, fission fragments cannot escape the target foils without being observed in a Si detector. This condition is the critical signature for a fission event. A fission event is a high multiplicity event and can result in one or more neutrons or gamma rays being detected in the appropriate detectors. This high multiplicity event has special significance. 3.3.5. Si detectors with other detector present. This is the special case that shows the particular power of the INL apparatus at IPNS. These types of events allow particular information to be gained that is

204

needed in reactor programs. To measure a simple fission cross section requires only a detector to detect the fission fragment and an MCA to store a neutron event spectrum. A neutron capture cross section can be done with a detector like DANCE1 if no branching ratio is desired or no competing reaction channel is present. But if other parameters are desired, additional experiments are needed or different methods must be used. As an example consider the need to measure neutron and gamma ray multiplicities. The method to use arrays of detectors operating in coincidence has been carefully explored1,2 and the multiplicities of the radiations can be determined. The methods work for mixed types of detectors as well as arrays of single types. With the ability to select the events by post experiment sorting the multiplicities of both neutrons and gamma rays can be determined. The production rates of the selected gamma rays arising from the prompt fission fragments or the excited isotopes produced by neutron capture are used to determine the appropriate reaction cross sections as a function of the incident neutron energy, determined from the event-by-event neutron TOF parameter. In addition to information required to extract absolute reaction cross sections, the acquired data sets also will contain the information needed to extract independent fission fragment yields1 that can be used to validate accepted values in a model independent manner. The most important aspect of this powerful ability to select events by sorting and then determine cross sections is to reduce background and events unassociated with the reaction channel of interest. This reduces the error on the cross section by allowing an event set to be selected that only contains events that are from that particular reaction channel. When various cross sections are measured conventionally, a main problem is to select an experimental configuration that will allow radiations from unassociated channels to be reduced. This is the effort in using C6D6 detectors to measure neutron capture cross sections on actinides yet discriminate against radiations from fission. The selection of particular events from a data set allows particularly low uncertainty results to be obtained. This also allows the measurement of branching ratios of reaction channels to be determined from the larger data set. In addition, it also can allow otherwise impossible cross sections and branching ratios to be determined. Consider the example, neutron capture for M1Am to M2Am and 242m Am. For various systems it is important to know the branching ratio or partial cross section for neutron capture of 241Am to 242Am or 242mAm, thus what is the production ratio of 242Am:242mAm? This is a difficult measurement to perform. For the INEEL apparatus this can be done by selecting the events of capture

205

from fission (no Si detector events), and then sorting for particular gamma rays that cascade from the same level to the isomeric 5" level and to the 1" ground state. In point of fact, it is really necessary to measure multi gamma ray cascades to these two levels to determine the population ratios. This is something that is best done with arrays of CSHPGe detectors (consider Gammasphere or Euroball). The problems of measuring the fission cross section for material with high spontaneous fission rates, or spontaneous fission rates that are statistically significant, is again a problem handled by sorting events based on different conditions. This method also reduces the total error whereas the traditional method of beam-on, beam-off does not remove the spontaneous fission events from the beam on data set. Selecting fission events by requiring a Si detector signal, and then sorting on gamma rays from different fission pairs will provide information on contributions for the two types of fission processes. The distribution of neutrons and thus associated fragments pairs are different for the spontaneous versus induced fission. This is caused by the differences in the excitation energy of the fissioning nucleus. In spontaneous fission the nucleus is in its ground state. For induced fission the nucleus will be at an excited state due to the energy brought in by the incident neutron and by the rearrangement of the population of the nuclear orbitals in the nuclear system after the neutron is captured. In the case of ^''Pu which is very important for the reactor programs, separate spontaneous and induced fission measurements are planned. A 240Pu sample has been obtained which will be used to directly measure the spontaneous fission. The induced fission will be measured using a 239Pu target that will be obtained from Russia. 3.4. Two Unique Capabilities The above discussion provides a description of the unique capabilities of the INL apparatus and how the IPNS facility supports these capabilities. The most important is the ability to take coincidence data associated with a particular nuclear event. An array of detectors can be operated in this manner at other facilities but at IPNS two features are important: 1) an intense flux of neutrons, and 2) the availability of the beam for long experimental measurements. These two facts allow the high statistics data needed to be acquired. To achieve the goal of ~109 events to be stored by the data system, requires over 100 days of beam time. This long experimental time is available at IPNS as the INL apparatus is on one beam line and not affecting experiments at other locations in the facility. These exceptionally high statistics are the key factor in reducing the statistical error to the 3-5% range.

206

The other unique feature is due to the nuclear event based data collection and post analysis by sorting data into subsets based on physics conditions. This very approach cannot be done without the high statistical data set described. By imposing multiple conditions and sorts via software or computer processing in selecting data sets for detailed analysis, the "cross talk" between channels can be minimized in ways that are not possible in the hardware of the electronics. The simplest and easiest to understand is the Si detector trigger T4 to separate fission events and all other events. The non-fission events can be further sorted look at other reaction channels. This capability has produced exceptional results in nuclear structure and spontaneous fission studies and is easily applied to the problems of measuring various neutron cross sections of actinide isotopes. This post analysis capability based on data selection of reaction channels provides results that are self-consistent across the larger experimental data set. This means that the ability to use different approaches in sorting can provide results that provide consistency checks that are otherwise not available. An example of this is the case of the determination of a fission cross section by direct selection of the observation from the Si detectors and the cross section determined from sorting on gamma rays from the highest yield fission fragments. Although the statistics in these two subsets of data will be different, the cross section should have the same result in both cases. This is a powerful tool to check the results and provide a consistency not in previous work. 3.5. Actinide Targets Our targets are fabricated in Russia by our collaborators at the Joint Institute for Nuclear Research (JINR) and they have material in usable quantities up through the light californium isotopes. The targets are metal foils, not oxides, on an appropriate backing and of the thickness needed to allow the light fission fragments to escape the target and enter a silicon detector that is in contact with the target material. The removes the need for large corrections of the incident neutron flux that must be done in the case of oxide targets. This removes an additional source of error. In addition, the vapoe deposition of metal onto metal backing gives excellent stability to the targets and reduces the risk of contamination due to targets coming off the backing. The isotopic purity of the targets is greater than 98% for the principal isotope and a detailed chemical analysis is provide for each target batch and individual target characteristics such as mass per unit area, total mass, and other are provided. The targets are delivered to us in a ready to use form that we request such that no preparation other than mounting them in the Si detector stack is required.

207

4. Conclusions The description of the INL apparatus has focused on operational procedures and how errors are minimized through techniques new to cross section measurements but well established in low energy nuclear physics. These are not simply methods of instrumentation but also of computer processing and data analysis. Neutron TOF errors and uncertainties are dominated by the moderated neutron pulse width and the width of the delivered proton pulse width, but software corrections of measured electronic and detector response times allow their contribution to be less than 10% of the total error. The fact that long run times are possible at IPNS without impacting the operation of the rest of the facility, means that the statistical errors, which dominate the total error, can be minimized to reach our goal of less than 3% to 5% over the neutron event spectrum. The error introduced from using 235U as a standard is what is quoted in the ENDF database and cannot be changed at this time. Some consideration is being given to using 3He tubes instead of a fission chamber, but we are going to continue to use the fission chamber for now. The 3 He tubes would allow the use of a simple 1/v relation for a standard. The goal is absolute cross section values but, as has been done in the past, a relative neutron cross section can be normalized to a single point, i.e., a previous thermal-neutron value. These existing thermal values from past years provide a check on the work as the cross sections being measured include the thermal value. References 1. 2. 3. 4. 5. 6. 7. 8.

Private communication, B.J. Micklich, ANL. J.M. Carpenter, et al., Nucl. Instr. and Meth. A234 (1985) 542-551. Muhrer, G.; et. al., Nucl. Instr. and Meth. A 527 (2004) 531-542. Takashi Ino; et. al., Nucl. Instr. and Meth. A 525 (2004) 496-510. "IPNS Progress Report" 1996-2001. Heil, M.; et. al., Nucl. Instr. and Meth. A 459 (2001) 229-246. Werf, S.Y. van der, Nucl. Instr. Meth. A153 (1978) 221-228. Ockles, W.J., Thesis, University of Groningen, The Netherlands (1978); Z. Phys. A286 (1978) 181-185. 9. Ter-Akopian, G.M.; et. al., Phy. Rev. C 55 (1997) 1146-1161.

CROSS-SECTION MEASUREMENTS IN THE FAST NEUTRON ENERGY RANGE

ARJAN PLOMPEN EC-JRC

Institute

for Reference Materials and Measurements Retieseweg 111, 2440 Geel, Belgium E-mail: [email protected]

Generation IV focuses research for advanced nuclear reactors on six concepts. Three of these concepts, the lead, gas and sodium fast reactors (LFR, GFR and SFR) have fast neutron spectra, whereas a fourth, the super-critical water reactor (SCWR), can be configured to have a fast spectrum. Such fast neutron spectra are essential to meet the sustainability objective of GenlV. Nuclear d a t a requirements for GenlV concepts will therefore emphasize the energy region from about 1 keV to 10 MeV. Here, the potential is illustrated of the GELINA neutron time-of-flight facility and the Van de Graaff laboratory at IRMM to measure the relevant nuclear data in this energy range: the total, capture, fission and inelastic-scattering cross sections. In particular, measurement results will be shown for lead and bismuth inelastic scattering for which the need was recently expressed in a quantitative way by Aliberti et al. for Accelerator Driven Systems. Even without completion of the quantitative assessment of the data needs for GenlV concepts at ANL it is clear that this particular effort is of relevance to LFR system studies.

1. Introduction The Generation IV (GenlV) initiative has selected six advanced reactor concepts for research and development efforts that in the long term (2040) should lead to sustainable and economic reactor options that are safe and secure against diversion of sensitive material. Sustainability entails the minimisation of the use of natural resources and environmental impact by substantially reduced demands on mining of ore and on spent fuel storage capacity. Full recycling of the actinide content of spent fuel is envisioned leaving only partioning losses at the percent level for long term geological disposal. Multiply recycled fuel will contain a relatively large fraction of minor actinides 237 Np, 238,240-242pUj 241,242,243Am a n d 2 4 2 _ 2 4 5 C m t h a t would require a fast spectrum to maintain reactivity. Thus, besides the thermal Very High Temperature Reactor (VHTR), Molten Salt Reactor 208

209

(MSR) and Super Critical Water Reactor, four concepts have been chosen that feature fast neutron spectra; the Lead Fast Reactor, the Gas-cooled Fast Reactor, the Sodium Fast Reactor and a fast version of the SCWR. For most concepts studied an important further aspect associated with both sustainability and economy is the strive towards very high fuel burnup. Nuclear data needs for GenlV systems are currently being evaluated in a systematic way by means of sensitivity studies of the different concepts. Key reactor and fuel cycle parameters are investigated for which target accuracies have been formulated. Preliminary results are shown elsewhere in these proceedings1. Clearly an important focus of data needs will be in the fast energy region from 1 eV to 20 MeV with emphasis on the region from 1 keV to 10 MeV. Furthermore, from a similar sensitivity study for accelerator driven systems by Aliberti et al. 2 , we may amongst others anticipate the relevance of improved inelastic scattering data for the isotopes of lead and for bismuth in the case of the LFR. Here, the potential of the GELINA and Van de Graaff laboratories at IRMM is illustrated for measurements in the fast neutron energy region with a particular emphasis on inelastic scattering and (n,2n) reactions for lead and bismuth radiative capture of 2 0 6 Pb and fission of 234 U and 233 Pa. 2. Inelastic scattering and (n,2n) reactions Gamma-ray production cross sections for samples enriched in 206,207,208p|5 and for 209 Bi are being measured at the 200 m flight path of the GELINA facility. The fast neutron spectrum is used which is obtained from the U ( 7 , O T ) and U(7, F) processes with photons produced by electron bremsstrahlung on the uranium target. HPGe detectors are used with an overall 8-10 ns time-resolution 3 and an energy resolution of 1 keV (35 keV) at 1 MeV (10 MeV). Continuous energy excitation functions are obtained from threshold to 20 MeV for gamma-rays associated with inelastic scattering and, in favorable cases, with the (n,2n) reaction. For inelastic scattering at least one gamma is observed for the first 10-15 excited levels. From these total and level inelastic cross sections are deduced, requiring corrections only for the 0.8 s isomer in 2 0 7 Pb and the 125 fis isomer in 2 0 6 Pb. Fig. 1 demonstrates the energy resolution by comparing the resonance structure observed just above the inelastic threshold of 2 0 7 Pb with that a

full width at half maximum

210

in the total cross section. Fig. 2 shows the deduced total inelastic cross section for 209 Bi, which is strictly a lower limit above the highest excitation energy of ~4 MeV for which decay gammas were measured. Finally, Fig. 3 finally shows the yield for the 803 keV transition obtained from an enriched 207 Pb sample. One clearly observes the sudden rise above the threshold of the 2 0 7 Pb(n, 2n) process. The subthreshold yield results from inelastic scattering of the 2 0 6 Pb impurity in the sample. Further details of the methods employed may be found elsewhere4. 1.25

570

620

670

720

770

820

En(keV) 207

Figure 1. Inelastic cross section of P b just above the threshold. Also shown is the total cross section measured by ... et al. 3 .

3. Capture cross section measurements Capture cross sections at IRMM routinely make use of CeD6 detectors to combine the advantage of the total energy method with the low neutron sensitivity of these detectors. The total energy method can be applied to these detectors by means of the pulse height weighting technique and results in a detection efficiency independent of the details of the decay scheme of the compound nucleus. Thus, the detection efficiency can be

211

This work

Talys Lashuk-1994 Simakov-1992 Prokopets-1980 Joensson -1969

Owens-1968 CO .Q

Rosen-1968 ShiXiaMin-1968 Broder-1967

10000

15000

*

20000

En (keV) Figure 2.

Deduced total inelastic cross section of

209

Bi from threshold to 18 MeV.

made independent of incident neutron (resonance) energy. Recently, the accuracy of this method was greatly enhanced by detailed Monte Carlo modeling of the detector response and the impact of photon transport and neutron scattering in the sample 9 . Fig. 4 shows the capture yield that was measured for 2 0 6 Pb at the GELINA 60 m station with four such detectors at 125°. Signal, background and the correction for the residual neutron sensitivity of the detectors are shown. Also shown is detail of the spectrum in a range where previously no measurements were made. One first of all observes the excellent resolution that allows to extract capture areas up to 600 keV. Furthermore, the comparison with ENDF/B-VI shows the limitations of assumptions about the capture width for individual resonances in absence of experimental data. An assumed constant capture width equal to the average over the low energy region is insufficient to reproduce the measured results. The new data extend experimental information about the capture cross section from 150 keV to 600 keV.

212

120 100

f o o

10000 E n (KeV)

15000

20000

Figure 3. Unnormalised yield of the 803 keV transition of 2 0 6 P b from a sample enriched in 2 0 7 P b in function of the neutron energy. Note that the flux declines exponentially above 2 MeV.

4. Fission cross section measurements Fission cross section measurements take place at GELINA and at the VdG laboratory. Fig. 5 shows some results for the subthreshold fission cross section of 234 U obtained in collaboration with U. Gent at GELINA at a 30 m station 6 . Good energy resolution is obtained which is surpassed only at the 200 m station of the n.TOF facility at CERN. The measurements require a very high purity sample to avoid overwhelming the signal by contributions from 233 U or 235 U. Fig. 6 shows the results obtained for 2 3 3 Pa at the VdG laboratory 7 . The single energy points are compared with the continuous energy curve obtained with surrogate neutrons by the Bordeaux group 8 . The IRMM result concerns a direct measurement on a target that was produced at Studsvik. Despite the half-life of 27 days and the ingrowth of 233 U during the measurement, accurate results were obtained.

213

400000 Neutron Energy / eV

425000 450000 Neutron Energy / eV

Figure 4. Left: Capture yield Yc of 2 0 6 P b compared with background Yc^g and the correction YCiP for the neutron sensitivity of the detector. Right: Narrow energy region showing the new data, a new resonance analysis and the ENDF/B-VI evaluation.

5. Outlook Above some examples were given of the measurement capabilities at IRMM for neutron-induced reactions in the fast energy range. More recent examples including total cross section, activation cross section and charged particle cross section measurements can be found elsewhere ? . As indicated new measurement capabilities were developed for inelastic scattering and substantial improvements were made for capture measurements. Further increase of capabilities will come from a systematic exploitation of the use of fast digitisers, increased efficiency for inelastic scattering and capture measurements and new measurement equipment to study the fission process (neutron spectra, fission fragments). In particular, more efficient detection equipment will allow the use of smaller and smaller masses so that measurements on enriched isotopes and radioactive samples are facilitated. A new GELINA target design is being considered to improve the fast flux and the resolution function for fast neutrons. With these new developments we aim to study the cross sections and reaction parameters required for the development of new concepts in nuclear energy such as those under consideration for GenlV. External collaborations to resolve these new and interesting nuclear data issues are envisioned.

214

8

-i

'

IRMM 2003 ORNL1977

550

560

570

580

590

E [eV] Figure 5. Subthreshold fission cross section of 234 U measured in a part of the resonance region. Measurements were made at GELINA (preliminary results). These can benefit from the N U D A M E transnational access scheme b .

A cknow

ledgements

Many t h a n k s to the coworkers and collaborators for the input without which

this short overview would not be possible. References 1. G. Aliberti, G. Palmiotti, M. Salvatores, T.K. Kim, T.A. Taiwo, I. Kodeli, E. Sartori, J.C. Bosq, J. Tommasi, These proceedings 2. G. Aliberti, G. Palmiotti, M. Salvatores, C.G. Stenberg, Nucl. Sci. Eng. 146, 13 (2004) 3. D.J. Horen, J.A. Harvey, and N.W. Hill, Phys. Rev. C 20, 478 (1979); D.J. Horen, J.A. Harvey, and N.W. Hill, Phys. Rev. C 24 1961 (1981) b

http://www. irmm.jrc.be

215 T

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

r

En (MeV) Figure 6. The fission cross section for

233

Pa measured at the VdG.

4. L.C. Mihailescu, L. Olah, C. Borcea and A.J.M. Plompen, Nucl. Instrum. Meth. A 531, 375 (2004) 5. A. Borella, G, Aerts, F. Gunsing, A. Moens, R. Wynants, and P. Schillebeeckx, Determination of the weighting functions and neutron sensitivity for CeD6 detectors by MCNP, Proc. Int. Conf. on Nuclear Data for Science and Technology, Santa Fe, 26 Sep.-l Oct. 2004, USA, in print, (2005) 6. J. Heyse, C. Wagemans, K.W. Chou, L. De Smet, J. Wagemans, and O. Serot, Proc. of the Seminar on Fission, September 2003, Pont d'Oye, Belgium, World Scientific, (2004) 7. F. Tovesson, E. Birgersson, M. Fleneus, B. Fogelberg, V. Fritsch, C. Gustafsson, F.J. Hambsch, A. Oberstedt, S. Oberstedt, E. Ramstrom, A. Tudora, G. Vladuca, Nucl. Phys. A 733, 3 (2004) and references therein. 8. M. Petit, M. Aiche, G. Barreau, et al., Nucl. Phys. A 735, 345 (2004) 9. A.J.M. Plompen, Neutron data measurements for energy applications and nuclear waste transmutation at JRC-IRMM, Proc. Int. Conf. on Nuclear Data for Science and Technology, Santa Fe, 26 Sep.-l Oct. 2004, USA, in print, (2005)

RECENT MEASUREMENTS OF NEUTRON CAPTURE CROSS SECTIONS FOR MINOR ACTINIDES BY A JNC AND KYOTO UNIVERSITY GROUP* HIDEO HARADA, HIROSHI SAKANE, SHOJINAKAMURA, KAZUYOSHI FURUTAKA Japan Nuclear Cycle Development Institute, Muramatsu, Ibaraki, 319-1194, Japan

Tokai-mura,

Naka-gun,

JUN-ICHI HORI, TOSHIYUKI FUJII, HAJIMU YAMANA Research Reactor Institute, Kyoto University, Kumatori-cho, Osaka,590-0494, Japan

Sennan-gun,

Recent activities on the measurement of neutron capture cross sections by a Japan Nuclear Cycle Development Institute (JNC) and Kyoto university group is reviewed focusing on 7Np and 238Np, Firstly, an experimental issue on the measurement of thermal-neutron capture cross section and resonance integral of 237Np is discussed. In second, the effectiveness of utilizing a double-neutron capture reaction for the measurement of the neutron capture cross section for short-lived nuclei 238Np is discussed. In third, the measurements of the energy dependence of the neutron capture cross section of 7Np are reviewed focusing on the experimental progress, including a total-energy detector and a flash-ADC based data-taking system. The experimental results were compared each other and also with some nuclear data libraries to make the problems clear.

1. Thermal neutron capture cross section and resonance integral of The thermal neutron capture cross section (ao) and the resonance integral (I0) of 237 Np have been measured by an activation method to supply basic data for the study of transmutation of nuclear waste. The neutron irradiation of 237Np samples has been done at Kyoto University Reactor (KUR). Samples including about 1 kBq of 237Np were irradiated between two Cd sheets or without a Cd sheet. The thermal neutron flux and the epithermal component of the neutron field were measured using two kinds of monitors, Au/Al alloy wires and Co/Al This work has been carried out in part under the visiting researcher's program of the research reactor institute, Kyoto university. 216

217

alloy wires. Since 237Np has a strong resonance at the energy of 0.49 eV, the Cd cut-off energy was adjusted as low as 0.36 eV(thickness of the Cd sheets: 0.125 mm). A high purity Ge detector was employed for the activity measurement of the irradiated sample. The reaction rate to produce 238Np from 237Np was analyzed by the Westcott's convention. The details on the experiments and analyses were described in ref. [1] Results obtained were 141.7+5.4 barns foroo and 862+51 barns for I0 above 0.358 eV of 273Np. Table 1 shows the present result and previous results on the a0 for 237Np. Table 1 Comparison of g 0 for 237Np Authors (year) o 0 (b)

Measurement Methods

Brown et al. (1956) Smith etal. (1957) Tattersall et al. (1956) Kobayashietal. (1994) JNC-Kyoto (2003)

a-ray spectroscopic method Total cross section Pile oscillation method y-ray spectroscopic method y-ray spectroscopic method

172 ± 7 170 ±22 169 ± 3 158 ± 3 141.7 ±5.4

Recently, the a0oi 237Np has been independently measured using y-ray spectroscopic methods by two groups. Kobayashi et al used two different neutron fields for their irradiations and took into account the resonance at 0.49 eV in their analysis. A JNC-Kyoto group used thin Cd foils to reduce experimentally the effect of the resonance at 0.49 eV. On the other hand, Brown et al used an a-spectroscopic method and neglected the resonance's effect in their analysis, although cadmium boxes of wall-thickness about 1 mm were used. The measurements of the a0 by y-ray spectroscopic method give smaller values than those measured by other methods. The y-ray spectroscopic method requires absolute y-ray intensity of 238Np. Recent evaluated nuclear structure data file2 gave about 10% smaller value for the absolute y-ray intensity of 238Np; if this value, 25.19+0.21 for 984.5 keV y-ray is used for the reduction of the
218

A sample of 237Np of about 100 Bq was irradiated for 10 hours at KUR. Neptunium-238 and -239 were produced simultaneously as are shown in Fig. 1. 2f£Np + 11 T, /2 = 2.14 X IOC y

.

T 1 / 2 = 2.117 d"

2.3565 d

984.5 keV

'

1

' *f4Pu

= 67.74 y

T.„ = 2.41X10
Fig. 1 Reaction scheme of the double neutron capture reaction

237

Np(n,Y)238Np(n,Y)239Np

The neutron flux at the irradiation position was monitored using Au and Co wires. The epithermal index r^JT/T0 in Westcott's convention was measured as 0.03. A high-purity Ge detector with a BGO (Bi4Ge30i2) anti-coincidence shield was used to measure the weaky-rays of 239Np in strong backgroundy-rays of 238 Np. The a of 238Np was deduced from the ratio of these activities. Details on experiments and analyses were described in Ref. 3. The result obtained is 479± 24 b for the a of 238 Np[3]. Table 2 Comparison oioo for238Np Reference (70 ( b ) ENDL-86 ENDF/B-VI JENDL-3.3 JNC-Kyoto

100 202.8 450

h(b)

a (b)

56.8 100 201

100 202.8 450 479±24

The result is compared in Table 2 with data in the evaluated nuclear data libraries, ENDL-86 [4], ENDF/B-VI [5], and JENDL-3.3 [6]. To explain the present result, the evaluated capture cross sections of 238Np have to be increased about 5 times in the case of ENDL-86 and about 2.4 times in the case of ENDF/B-VI. The value in JENDL-3.3 agrees well with the present result. However, the reduced resonance integrals, that is, the resonance integrals minus 1/v parts, of all libraries are equal to zero. Further fine evaluations are required on the resonance integral of 238Np.

219

3. The measurement of neutron capture cross section of method

Np by a TOF

The neutron capture cross section of 237Np has been measured relative to the B(n,a)7Li* cross section by the time-of-flight method in the energy range from 0.02 eV to 100 eV. The 46 MeV electron linear accelerator at the Research Reactor Institute, Kyoto University, was used as a pulsed neutron source. The 16-section BGO scintillation detector shown in Fig. 3 having a total volume of 8.54 1 and the associated 40 MHz flash-ADC-based data taking system have been developed for the measurement of the capture cross section. The BGO detector can be used as a total energy gamma-ray detector [7]. 10

Fig. 3 The 16-section BGO scintillation detector (back-view in left and side-view in right)

Figure 4 shows typical waveforms for y-rays from a Cs source obtained by the flash-ADC system with 100 ns channel width (10 MHz). The input pulse was smoothed and the derived waveform was calculated; this procedure enables to extract pulse-height information even for overlapping pulses. The capture cross section of 237Np in resonance energy range was measured using the total energy gamma-ray detector for the first time,. The result of the present measurement has been compared with the evaluated capture cross sections of ENDF/B-VI and JENDL-3.3, as well as with the data measured by other authors. Figure 5 shows the ratio of the measured capture cross section and the other measurements or evaluated values. To make the comparison easier, our data in fig. 5 was normalized to 181 b at 0.0253 eV and then the capture cross section was averaged. In the energy range from 1 to 100 eV, the energy dependence of the present cross section agrees well with the experimental data by Weston and Todd [8]. However, there is an obvious discrepancy between the data by Kobayashi et al. [9] and the present data. On the otherhand, there is an agreement in the energy range 0.02 -100 eV on the energy dependence of the 237 Np capture cross section between the present measurements [10] and the

220

ENDF/B-VI. As far as the JENDL-3.3 is concerned, there is an obvious disagreement with the present data below ~ 0.3 eV, while above this energy an agreement exists. These comparisons show the necessity of the measurements and re-evaluations of the neutron capture cross section at especially thermal energy region. Gamma source: Cs : FLASH ADC: 10 MHz (t ch =• 100 ns) '

£ 10000

191 pulses in this waveform average counting rale 29,000 cps

illi!||lli(iii!l!ii! 300 no

20000

40000

50000

60000

300

Signal evel, arb. i nits

•»% o •

* •

agog*

i v

88 -

9 <> 6 *

Input waveform Derived waveform

ATA

o

«

t.

^awawSSb

*

. •

100

\

* *&

• V

-

•>

*•

°

*

* ^>

*• . • V

D

•100

10220 10240 10200 10200 10300 10020 10340 10300 10300 10400

182G0

Fig. 4 Waveforms of the BGO detector signal: (top) full-length input waveforms; (down) stretched parts of the input and derived waveforms

' ™»™»»»*

tz

Present / Kobayashl Present / Weston Present/JENDL-3.3 Present/ENDF/B -VI

.«->.., CO

o O 1-0 p

^jtzL

L£ -4JTL i..-«-i=^M

0.8 ID

3 0.6 ' — 0.01

Neutron Energy (eV) Fig. 5 The ratio of the averaged capture cross section measured by a JNC-Kyoto group and previous measurements or evaluated values

221

References 1.

T. Katoh, S. Nakamura, K. Furutaka, H. Harada, K. Fujiwara, T. Fujii, and H. Yamana, J.Nucl. Sci. Technol, 40,559 (2003). 2. Evaluated Nuclear Structure Data File and NUDAT, National Nuclear Data Center, Brookhaven National Laboratory, USA (2002). 3. H. Harada, S. Nakamura, T. Fujii, and H. Yamana, J.Nucl.Sci.Technol., 41, 1 (2004). 4. Evaluated Nuclear Data Library, ENDL-86, LLNL (1986). 5. Evaluated Nuclear Data File/B-VI, ENDF/B-VI, BNL/NNDC (1990). 6. K. Shibata, T. Kawano, T. Nakagawa, et al., J. Nucl. Sci. Technol. 39, 1125 (2002). 7. O. Shcherbakov, K. Furutaka, S. Nakamura, H. Harada, K. Kobayashi, Nucl. Instru. Method, A517, 269 (2004). 8. L. W. Weston, J. H. Todd, Nucl. Sci. Eng., 79, 184 (1981). 9. K. Kobayashi, S. Lee, S. Yamamoto, H. J. Cho, Y. Fujita, " J. Nucl. Sci. Technol., 39, 111 (2002). 10. O. Shcherbakov, K. Furutaka, S. Nakamura, et al., J. Nucl. Sci. Technol, 42,135 (2005).

DETERMINATION OF MINOR ACTINIDES FISSION CROSS SECTIONS BY MEANS OF TRANSFER REACTIONS M. AICHE, G. BARREAU, S. BOYER, S. CZAJKOWSKI, D. DASSIE, C. GROSJEAN, A. GUIRAL, B. HAAS, B. JURADO, B. OSMANOV CENBG (CNRS/IN2P3-Univ. Bordeaux!), Le Haul Vigneau, 33175 Gradignan, France E. BAUGE M. PETIT,

CEA, SPhN, BP12, 91680 Bruyeres-le-Chdtel,

France

E. BERTHOUMIEUX, F. GUNSING, L. PERROT, C. THEISEN, CENSaclay, DSM/DAPNIA/SPhN, 91191 Gif-sur-Yvette, France F. MICHEL-SENDIS,

IPN, 15 rue G. Clemenceau, 91406 Orsay, France A. BILLEBAUD, J. N. WILSON,

LPSC, 53 Avenue des Martyrs, 38026 Grenoble, France I. AHMAD, J.P. GREENE AND R.V.F. JANSSENS ANL, 9700 S. Cass Avenue, Argonne, IL 60439.

Abstract. An inventive method that allows to determine neutron-induced cross sections of very short-lived minor actinides is presented. . We have successfully applied this method, based on the use of transfer reactions, to 233Pa, a key nucleus in the 232Th-233U fuel cycle. A recent experiment using this technique has also been performed in order to obtain the neutron-induced fission cross sections of 242,243,244 Cm and 241Am which are present in the nuclear waste of the current U-Pu fuel cycle. These cross sections are highly relevant for the design of reactors capable to incinerate minor actinides. Preliminary experimental results will be presented.

222

223

1. Introduction Minor actinides transmutation has been proposed as an alternative way to the high level radioactive waste depositary solution. Indeed several studies promote fast neutron reactors in which minor actinides can be incinerated efficiently by fission. In this respect high quality neutron data such as capture and fission cross sections are crucial to provide reliable core design and fuel assessments. However for some nuclei, like for example Americium and Curium isotopes, the existing evaluated data files show that neutron data of interest for incineration are often of poor quality. This is mainly due to their huge radioactivity (some of them are strong a and neutron emitters), to the difficulty of obtaining pure samples of sufficient quantity which complicates considerably direct neutron induced cross section measurements. We present in this work an alternative experimental method that allows overcoming these difficulties. 2. Presentation of the method: the surrogate technique The measurements of reaction cross section ( capture, fission,..) induced by neutrons on short-lived actinides has been a long standing problem. Many of these nuclei are too-short lived to serve as target in present day set-ups. To overcome these problems, an indirect method has been used in the 70's by J. Cramer and H.C. Britt [1]. It consists in measuring the fission decay probability of an excited nucleus produced via a few nucleon transfer. The transfer reaction (the surrogate reaction) is chosen so as to produce the same mass A and charge Z as those of the compound system formed from direct neutron capture. Then, the corresponding (n,f) cross section is obtained by multiplying the measured fission probability by the compound nuclear cross section formation. The latter is calculated by means of an optical model calculation. A schematic representation of the method is shown in fig. 1. for the transfer reactions 3 He(X,x)A—> fission which have been used in this work. Using the surrogate reactions 3He(X,x=p,d,t,a) on 232Th and 243Am targets, we have been able to measure the fission probability of several short lived nuclei. In some cases, the reliability of the method has been tested with existing neutron induced fission cross sections. The table I gives a list of the studied reactions. The last column indicates the corresponding neutron equivalent targets (A-l) and their half-lives.

224

x(p,d,t,..)

Figure 1. The surrogate reaction 3He+X—>A+x(p,d,t,oc) is used to form the excited (A,Z)* nucleus. The same excited nucleus is produced via capture of a neutron (n) by the target nucleus (A-1,Z) and subsequent decay by fission (ory-rays) of the recoiling nucleus A*. The reaction cross section in the channel (n,f) is inferred from the product of the measured decay probability of the excited nucleus A* with the calculated cross section for the formation of the compound A* in the neutron induced reaction .

Table I Surrogate reactions

Neutron target and Tm

232

Th(3He,a)231Th

230

232

Th(3He,t)232Pa

231

232

Th (7.5 104 y) Pa (3.2 104 y)

Th(3He,d)233Pa

232

Pa (1.3 d)

232

Th(3He,p)234Pa

233

Pa (27 d)

243

Am(3He, a)242Am

241

243

Am(3He,t)243Cm

242

Am (432 y) Cm (162.8 d)

243

Am(3He,d)244Cm

243

243

Am(3He,p)245Cm

244

Cm (28.5 y) Cm(18.1y)

225

3. Experimental details Since the experimental technique and the method of data analysis have been described in detail in a previous publication [2], only the highlights of the experimental set-up will be presented. The EPN Orsay tandem Van de Graaff was used to provide 3He beam at 30 MeV bombarding energy. Targets of Th and Am have been used for two series of measurements; they were approximatively 100p.g/cm2 thick deposited on to 50-75ug/cm carbon foils. The experimental configuration is shown on fig. 2. TnealOOMgfcm''

Ttl 1 1.2 Backing 50 ng/cm~ C F ission detectors

5mm

Figure 1. Experimental set up

The identification of the channel reactions is done with two AE-E telescopes placed respectively at 90° and 130° relative to the 3He beam axis. The fission fragment detector system was designed to achieve a large geometrical efficiency (48%) and good granularity for fission fragment angular distribution. The arrangement consists of 15 photovoltaic cells (20x40 mm2) distributed among 5 units, each unit consisting of 3 cells. The whole assembly is placed normal to the reaction plane defined by the two telescopes. In the reaction plane, the central cells are located at 5 cm from the target. Four units were placed in the forward direction with a covering angle from 14° to 125°. The last unit was placed at 180° relative to the foremost one, it adds one more point to the backward angular distribution measurements.

226

4. Fission probabilities using the surrogate transfer reaction The measured fission probabilities for some of the heavy systems investigated in these measurements are displayed in fig 3.

M

'rPa (0n=6,53 MeV)

s

r % I \

I 0.4

l

0.5

?

"Po (Bn-5.5 MeV) f

5;

•'

£ Q.

|

f

0,3

,VAv- .

ji'/ 5' "'"»

•"Pa(Br.-5.22M«V). """l

"»

«'

10

^''""I2'""li""''l4'"

15

ExcUaiion Energy (MeV)

Figure 3. Measured fission probabilities for 2 3 2 ' 2 3 3 2 3 4 p a a s a function of excitation energy (left panel). Fission probabilities of 242Am(a), 244Cm(b) and 243Cm (right panel). Bn indicates the neutron binding energy of the fissioning system

The error bars in these plots represent statistical errors as well as uncertainties related to background subtraction (transfer reactions from the carbon backing). The comparison of the fission probabilities for three Protactinium isotopes (232'233,234Pa) are shown on the left panel. For all of them, the fission probability has been measured for the first time beyond the second chance fission. A salient feature of these results is the low fission probability of 234Pa as compared to the lighter ones. The right panel shows the fission probabilities of 242Am, 243Cm and 244 Cm. These last results should be considered as very preliminary, they have not yet been corrected for expected minor angular distribution effects. Nevertheless and here again, the measurements extend up to the onset of second chance fission. Contrary to the previous Pa isotopes, these heavy actinides are more fissionable as seen by their large fission probabilities (about twice more). For the two fissile nuclei 233Pa and 244Cm, the fission probability has been measured well below the neutron binding energy (Bn), this feature illustrates the usefulness of the transfer reaction technique for investigating the fission barrier in an energy domain which cannot be reached through direct neutron measurements.

227

5. Determination of the neutron induced fission cross sections As stated before, the neutron induced fission cross sections have been deduced from the measured fission probability of nucleus A at an excitation energy E* multiplied by the formation cross section of the compound nucleus A in the reaction (A-1) + n. At the incident neutron energy En, the (n,f) cross section can be approximated by: af (En) =Pf (E* = Sn +^--En)

cCN(En)

This relation supposes that on the average, the fission probability of the nucleus A does not depend on the reaction mechanism. One may question whether the states populated by neutron absorption and transfer reactions are the same. In fact, model calculations show that, for neutron energies below 2 MeV, the average value of the angular momentum distribution populated via neutron absorption is approximately one or two units lower than the average angular momentum induced in transfer reactions, this difference becoming smaller with increasing neutron energy. This means that, in neutron-induced fission reactions, low-lying states might be populated that are not accessible via transfer reactions and vice-versa. Therefore, one would expect to find some difference between fission induced by these two reactions mechanisms at the fission threshold. However, as shown below, the comparison between experimental neutroninduced fission cross sections and transfer-induced fission cross sections shows a very good agreement at the fission threshold. At higher excitation energies, the fission probability is rather insensitive to angular momentum effects. The latter assumption could be justified with the ratio rn/Tf which is closely related to the fission probability. A large body of experimental values for this ratio has been obtained through (n, f)» (Y>f)> («> xnf) reactions measurements. The systematic studies of Vandenbosch [3] and Gavron [4] show that the fission probabilities do not depend strongly on the reaction mechanism involving neutron, photons or light charged particles. Therefore, the reliability of the calculated compound formation cross sectionCTNCcould be also of potential concern. Moreover, over the last 30 years, many of the uncertainties of the old optical model calculations have been clarified using microscopic or semi microscopic nucleon-nucleus optical potentials. The compound nucleus formation cross sections adopted in this work have been calculated using a semi-microscopic optical model potential [5].

228

6. Validation of the method using existing direct (n,f) measurements In order to validate our approach, we have first considered the reactions 230Th(n, f) and 231Pa(n,f) , the cross sections of which have been already measured by neutron capture. The fissioning systems have been formed through the surrogate reactions Th( He,a) and Th( He,t) respectively. The fission cross sections obtained in the present work are displayed in fig 4. In both cases, the comparison is done with the ENDF-B/VI and JENDL-3 libraries[6], as well as with the more recent neutron measurements performed by Meadows [7]( 230 Th(n,f)) and Plattard [8]( 231Pa(n,f)). In the energy range from 0.5 to 7 MeV our determination of the 230Th (n,f) cross section ( amplitude as well as the onset of second chance fission threshold ) is in good agreement with the direct measurements of Meadows and the ENDF-B/VI evaluation, while the JENDL-3 recommended values are on the average 20% lower. Concerning the 231Pa(n,f) case, a similar but inverse trend is observed while comparing the ENDF-B/VI and JENDL-3 data files. Our determination agrees fairly well with the latter as well as with the measurements of Plattard below 2 MeV. At higher energy up to the second chance fission threshold, our values are however on the average 10% to 15% lower than the ones reported by Plattard.

Figure 4. The 23QTh(n,f) cross section compared to the recent evaluations and to the direct measurements of Meadows[5] (left). The same for 231Pa(n,f) compared to the direct measurements of Plattard [6] (right)

229

7. Determination of the 233Pa(n,f) and 233Pa(n,y) cross sections Before this work, very few measurements have been performed for the shortlived (27 days) Pa which is the precursor of the U fissile nucleus. Owing to a renewal of interest for the 232Th-233U fuel cycle, measurements of the capture Pa(n, y) and fission Pa(n, f) cross sections have been requested since the recommended values extracted from the ENDF-B/VI and JENDL-3 libraries show large discrepancies. The recommended nuclear data files differ by a factor of two for the 233Pa(n, f) cross section values. For the first time, both, the fission and capture cross sections have been determined using the surrogate technique. 1.6

s~^

JEHOL-J M

• 1.2

4

ThUwott

A F.Tovasaofl

r

I

• \ .

r

/ i $t S0.8

\

#0.6

i,

0.4

0.2

0.S

i

•

JENDL-3 £MOF«-VI

xa

k

Paintf

A A

1

f A.

\ \ " • — .

0

This wort Imtuired)

A Thts work (calcuiat«l)

i #

2

'-..

//

Pa(n,f)

I-

\

/

m 1

2*

• — EHSF.B-iy

;••

.

k. • • ~

0

1

: .

2

: „

. •

, •

3

:

4

,

• i

5

,

:

• ; .....

6

:..,

j

7

„

8

9

Neutron Energy (MeV)

Figure 5a. The Pa(n,f) cross section as a function of neutron energy, compared to the latest evaluations and measurements.

10

»

0.2

0.4

0.6

0.8

1

Neutron energy (MeV)

Figure 5b. The 233Pa(n,Y) cross section, compared to the latest evaluations (full circles) and our model calculations (full triangles)

Our determination of the 3Pa(n, f) cross section is shown on the left of fig. 5. It is compared to the recent direct measurements of F. Tovesson [9]. A set of 16 values has been obtained between 1 MeV and 8.5 MeV; below the second chance fission, these data are in agreement with ours except at 2 MeV where the directly measured cross section is 25% lower than our determination. One can notice that the onset of the first chance fission is well reproduced using transfer reaction or neutron induced fission. Above 6 MeV, only four points, with large uncertainties (up to 30%) have been measured directly, moreover, the measured cross sections are lower than ours. In this neutron energy domain, the recommended values extracted from the ENDF-B/VI and JENDL-3 data files are largely at variance with the experiments (surrogate

230

technique and direct neutron measurements). Nevertheless, in the neutron energy domain (2.5- 5 MeV) corresponding to the first chance fission, the recommended values extracted from the JENDL-3 library are fairly well reproduced by our measurements. The y-ray emission probability (Py) has been measured with the transfer reaction 232Th(3He,p)234Pa*. The measurements have been performed below the fission threshold (from 0 to 1 MeV neutron energy) in order to avoid the contamination from fission fragment y-rays. The Pa channel has been identified with 4 AE-E telescopes. The fission fragment detector has been replaced by four CgDg liquid scintillator detectors working in coincidence with the telescopes. A detailed report of the experimental procedures can be found in [10]. A full analysis of the results that are unique to this experiment will appear in a forthcoming publication [11]. The corresponding 233Pa(n, y) cross section has been determined in the same way as the Pa(n,f) one. Our results are shown in the right hand side of fig. 5 along with the recommended values extracted from the ENDF-B/VI and JENDL-3 data files. Below 0.5 MeV, our data are significantly higher than these evaluations which both rely on model calculations without reliable experimental data. Taking advantage of our fission measurements, we have performed recently a complete evaluation of the relevant cross sections for the system n+ 233Pa below 6 MeV neutron energy. A statistical model analysis of the 233Pa (n, f) cross section has been developed. Once the model parameters (level densities and fission barriers characteristics) have been fixed, the calculations have been carried out for the two competing channels Pa(n, y) and Pa(n, n') . As seen on the right of fig.5, our prediction of the 233 Pa(n, y) cross section is in reasonable agreement with the measured one obtained from the surrogate technique. This demonstrates that model calculations coupled to the surrogate technique could be a useful approach for predicting reaction cross section channels which are otherwise difficult to measure. 8. Conclusion Neutron reaction cross sections for minor actinides face large challenges mostly due to their huge alpha activity. A surrogate method has been investigated to overcome these difficulties. The fission cross section has been deduced multiplying the fission probability measured with transfer reactions with model calculations for neutron-compound cross section. Comparisons to cases where (n,f) cross sections have been measured directly demonstrate the feasibility of the method for the determination of (n,f) cross sections of actinides that in many cases cannot be measured directly. The method has been used for the

231

determination of the fission and capture cross sections of the short lived (27 days) nucleus 233Pa. This unique set of measurements points out the need for a new evaluation of the 233Pa data files currently available from the ENDF-B/VT and JENDL-3 libraries. The extension of the method to the determination of the neutron induced-fission cross sections of the minor actinides 242, M3' 244Cm is still in progress. Acknowledgments We thank the Orsay tandem accelerator staff for their great support during the experiment. This work was supported by the GEDEPEON program, the Conseil Regional d'Aquitaine, and by the U.S. Department of Energy under contract W31-109-ENG-38. The authors are also indebted for the use of 243Am to the Office of Basic Energy Sciences, U.S. Department of Energy, through the transplutonium element production facilities at Oak Ridge National Laboratory. References 1. Cramer, J. D. and Britt, H. C , Nucl. Sci. Eng. 41, 177 (1970) 2. Petit, M. et al., Nucl. Phys. A 735, 345 (2004) 3. Vandenbosch, R. et al., Nuclear Fission, Academic Press New York and London (1973) 4. Gavron, A. et al. Phys. Rev. C 13, 2374 (1976) 5. Bauge, E. et al. Phys. Rev. C 61, 034306 (2000) 6. NNDC, Brookhaven National Laboratory. 7. Meadows J.W., ANUNDM-83, 183 8. Plattard S. et al. Phys. Rev. Lett. 46, 633, (1981) 9. Tovesson F. et al. Nucl. Phys. A 733, 3, (2004) 10. Wilson J.N. et al. Nucl. Instr. and Meth. A 511, 388, (2003) 11. Boyer, S., PhD Thesis, Universite Bordeaux (2004) and paper in preparation.

This page is intentionally left blank

EVALUATED DATA LIBRARIES

This page is intentionally left blank

NUCLEAR DATA SERVICES FROM THE NEA HANS HENRIKSSON, YOLANDA RUGAMA OECD/NEA Data Bank, 12 Bid des lies, 92130ISSY-LES-MOULINEAUX, FRANCE

The OECD Nuclear Energy Agency (NEA) Data Bank is part of an international network of data centres in charge of the compilation and dissemination of basic nuclear data. Through its activities in the nuclear data field, the NEA participates in the production of data and their distribution to nuclear data users. The high priority request list is an example of such a project. The NEA thus provides an essential link between producers and users of nuclear data. The NEA Data Bank distributes the main computer codes and nuclear databases with bibliographical information, evaluated libraries, e.g. JEFF, and experimental data in the data base EXFOR comprising published neutron induced as well as charged particle induced nuclear reaction data. The new data library JEFF-3.1 will be presented here, as well as the data display tool JANIS. The NEA is also involved in the work in the Generation IV International Forum (GIF) technical working groups that are developing research programs for advanced reactor concepts.

1. Introduction The Nuclear Energy Agency (NEA) is a specialised agency within the Organisation for Economic Co-operation and Development (OECD), an intergovernmental organisation of industrialised countries, based in Paris, France. NEA Data Bank works internationally in a network of several data centres in charge of the compilation and dissemination of basic nuclear data. Through its activities in the nuclear data field, the NEA participates in the production of data and their distribution to nuclear data users. The NEA thus provides an essential link between producers and users of nuclear data. The NEA web site (www.nea.fr) offers interfaces to the main nuclear databases with bibliographical information, evaluated libraries, e.g. JEFF, ENDF/B and JENDL, and experimental data in the data base EXFOR comprising published neutron induced as well as charged particle induced nuclear reaction data. The NEA serves also as the technical secretariat for the Generation IV International Forum (GIF) technical working groups, which are developing research programmes for advanced reactor concepts.

235

236

The display program JANIS (JAva-based Nuclear Information Software) has been developed at the NEA, and its latest version (JANIS-2.1) was released in July 2004. JANIS is designed to facilitate the visualisation and manipulation of nuclear data, and to allow the user access to numerical and graphical representations without prior knowledge of the storage format. In this paper an overview will be given of EXFOR as well as the JEFF evaluation library project including the contents of the new JEFF-3.1 nuclear data library released in May 2005. The JANIS display program will also be presented with examples. 2. The NEA The mission of the NEA is to assist its member countries in maintaining and further developing the scientific, technological and legal bases for safe, environmentally friendly and economical use of nuclear energy for peaceful purposes. The NEA works as a forum for international co-operation and as a centre of excellence which helps member countries with technical expertise. The NEA's current membership consists of 28 countries, in Europe, North America and the Asia-Pacific region. Together these countries account for approximately 85% of the world's installed nuclear capacity. NEA work areas can be divided into nuclear safety and regulation, nuclear energy development, radioactive waste management, radiation protection and public health, nuclear law and liability, information and communication, nuclear science and the nuclear data bank activities. The nuclear data evaluation co-operation activities involve the evaluation projects in the following regions: Japan (JENDL), United States (ENDF), Western Europe (JEFF), and non-OECD countries (BROND, CENDL, and FENDL). The participation of the evaluation projects in non-OECD Member countries will be channelled through the Nuclear Data Section of the International Atomic Energy Agency (IAEA). 2.1. The NEA involvement in GIF The NEA serves as the technical secretariat for the six projects on advanced nuclear reactor concepts within the Generation IV International Forum (GIF). The NEA will offer long term continuity and neutrality in to this project. Another advantage of the NEA in GIF is the long experience in administrating international working groups in the field of nuclear energy. The involvement in the working groups permits also to find new needs of nuclear data, which is of direct importance for the Generation IV reactor concepts. These needs will be collected toghether with all other relevant data requests in a high priority request list (HPRL) project.

237

2.2. High priority request list of nuclear data needs The Nuclear Energy Agency's HPRL project started in the early 1980s by the NEA committees on Reactor Physics and Nuclear Data. The aim was to provide targets for the improvement of nuclear data, primarily for application in the nuclear industry. The users were asked to define their most important requirements, taking into account the limited resources available for measurement and evaluation of data. The request list is collected to provide targets for the improvement of nuclear data, primarily for application in the nuclear industry through the evaluated data projects, and is a compilation of the highest priority nuclear data requirements. The purpose of the list is to provide a guide for those planning measurement, nuclear theory and evaluation programmes. The HPRL is a place where data users meet data producers. The High Priority Request List (HPRL) is in a stage of renewal. A totally new list is going to be presented in 2005, and each year there will be a review of the requests by external referees coordinated by the subgroup C of the OECD NEA Nuclear Science Committee's Working Parties on International Evaluation Co-operation (WPEC). This group consists of both data users and producers from industry, representing Europe, Japan, Russia and USA. The NEA is at the moment collecting new requests for experimental nuclear data. The requests are divided in high priority ones, where a quantitative justification is needed, and general requests where a more qualitative justification is sufficient. All requests need to be tied to a certain project including a project life span, that is to be stated. The list will be maintained by the NEA Data Bank and is presented on the NEA home page. 2.3. The NEA Data Bank Services The Data Bank primary role is to provide scientists in member countries with reliable nuclear data and computer programs for use in different nuclear applications. The services include also thermochemical data for radioactive waste management applications. The 22 member countries of the NEA Data Bank are: Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Italy, Japan, R.o. Korea, Mexico, Netherlands, Norway, Portugal, Slovak Republic, Spain, Sweden, Switzerland, Turkey, United Kingdom. By arrangement with the IAEA, the Data Bank computer program services cover both Data Bank countries and member states of IAEA, except USA and Canada. A separate agreement covers nuclear data and computer program exchanges with the USA

238

and Canada. Users of the Data Bank services include governmental research institutes, industry and universities. The NEA Data Bank administrates and distributes the main evaluated nuclear data libraries in the world on the NEA web site ("www.nea.fr) as well as on CD/DVD on request. The NEA has been using relational databases since 1993 to provide a centralised repository of data and has used web-based technology to allow interactive retrieval of the data. The NEA home page offers interfaces to the main nuclear databases: EVA for evaluated data, CINDA for bibliographical information and EXFOR for experimental data. This latter also includes on-line plotting capabilities. The nuclear data services are also in charge of the collection and validation as well as the distribution of the Joint Evaluated Fusion and Fission (JEFF) library. The JEFF project has evolved from the two separate EFF (Fusion) and JEF (Fission) projects to a joint collaboration in 1995, with a first library, JEFF-3.0 [1], released in 2002. A preliminary version of JEFF-3.1 was tested and evaluated in November 2004 and a final test version was presented in March 2005. The official release of JEFF-3.1 was in May 2005 as scheduled. A summary report of JEFF-3.1 is planned to be published in the end of 2005. The computer program services (CPS) group provides more than 2000 documented packages and group cross-section data sets related to nuclear energy applications. The CPS group publishes news letters regularly (see www.nea.fr/html/dbprog). describing the acquisition of basic nuclear data, computer codes and experimental system data needed over a wide range of nuclear and radiation applications. Independent verification and validation of these data is offered using quality assurance methods, adding value through international benchmark exercises, workshops and meetings and by issuing relevant reports with conclusions and recommendations. The CPS disseminate the different products to authorised establishments in member countries and integrates user feedback (more than 600 establishments are served in member countries and about 80 from other countries through agreement with the IAEA). The services include collection of programs, compilation and verification in an appropriate computer environment, and that the computer program package is complete and adequately documented. The Thermochemical Database (TDB) project is a project to help meeting the specialised modelling requirements for safety assessments of radioactive waste disposal sites and focuses on actinides and fission and activation products. Chemical thermodynamic data are collected and critically evaluated by review teams of experts. The NEA Data Bank acts as Project Co-ordinator. An "Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium,

239

Americium and Technetium" was published in 2003. Reviews of chemical thermodynamic data for compounds and complexes of Selenium and Nickel have been published early in 2005 whereas the inorganic data of Zirconium and selected organic compounds and complexes of Uranium, Neptunium, Plutonium, Americium, Selenium, Nickel, Technetium and Zirconium are planned for publication later in 2005. After an exploratory stage of the TDB project, comprising the collection and review of bibliographic references for thermodynamic data for inorganic complexes and compounds of Thorium, Iron, Tin and Molybdenum as well as exploratory work for the thermodynamics of solid solutions, three review teams have been established. An expert team will prepare guidelines for evaluating thermodynamic data for solid solutions. The review work is expected to be completed by early 2007. More information can be found on the TDB home page: http://www.nea.fr/html/dbtdb. The NEA Data Bank organises seminars and workshops to present information on computer programs or groups of programs that are considered to be of special interest to users, such as the NJOY workshop in May 2005 at the NEA. Training courses on widely used computer programs are organised a few times a year to ensure a correct and effective use of them. Below follows three examples of services from the nuclear data services of the Data Bank, namely international collaboration, collection and maintenance of experimental data in EXFOR, the data manipulation and display tool JANIS, and finally the new evaluated data library JEFF-3.1. 3. EXFOR A comprehensive set of experimental reaction data are stored in a database called EXFOR [2], that was initiated already 1969. EXFOR has always been coordinated through an international network [3], and the other three main nuclear data centres are, besides the OECD/NEA Data Bank, the National Nuclear Data Center (NNDC) at Brookhaven National Laboratory (USA), the Nuclear Data Services (NDS) at IAEA and the Russian Nuclear Data Center (CJD) at the Institute of Physics and Power Engineering (EPPE) in Russia. In addition to storing the experimental data points and their bibliographic information, experimental information including source of uncertainties is also compiled. EXFOR is complete with respect to neutron reaction data, and is intended to also cover all charged particle data up to 12C with incident energies up to about 1 GeV. Selected heavy-ion-induced and photon-induced reaction data are also included. EXFOR contains at present about 15,000 experiments from 1935, divided in 114,000 different reactions with a total of about 8.4 Million data points.

240

The bibliographic database CINDA (Computer Index of microscopic Neutron Data) is closely linked to EXFOR, and contains a complete bibliography of all neutron data published since 1932, as well as an index to corresponding EXFOR entries and evaluated data. Besides neutron data, CINDA also covers photo-neutron, photo-fission and spontaneous fission data. CINDA is available on web retrieval, through the JANIS program, and as a book that can be requested from the NEA. 4. JANIS JANIS (JAva-based Nuclear Information Software) is a display program designed to facilitate the visualisation and manipulation of nuclear data [4]. Its objective is to allow the user of nuclear data to access numerical and graphical representations without prior knowledge of the storage format. It offers maximum flexibility for the comparison of different nuclear data sets. Features included in the latest release are described such as direct access to centralised databases through JAVA Servlet technology. One of the main missions of the OECD/NEA Data Bank is to provide nuclear data services to research laboratories, universities and industry in member countries. In the last decade, these services have developed along two parallel paths, namely, the use of the Internet and the distribution of dedicated software. Each approach has its advantages. For instance, the Internet option enables the user to access centralised (and thus up-to-date) databases. However, the display of the data is limited by the capabilities of web-pages. Likewise, software running on the user's personal computers can implement advanced, user-friendly interfaces enabling the display of complicated structures. JANIS was implemented as application-like software with direct access to large databases. The software is free of charge and can be downloaded or launched from the JANIS home page: http://www.nea.fr/ianis, where the complete manual can be found as well. JANIS accesses data contained in comprehensive databases. The formats supported are ENDF-6 (along with the linearised pointwise option PENDF and the group-wise option GENDF) and the computational format derived from EXFOR. JANIS comprises a number of functionalities. The main browser window shows the nuclide chart where basic isotope data can be shown, from NUBASE or from evaluated data libraries. This gives overall information of the isotope and of the evaluated data libraries, e.g. JEFF, JENDL, CENDL, BROND and ENDF/B. Searches can also be performed in the EXFOR and CINDA databases.

241

In Fig 1, an example is shown on how JAMS displays data on the total cross section of 99Tc from JEFF-3.1 and where the user has compared the results with JEFF-3.0 as well as with a set of data form EXFOR. y> ^ ^ T T r ^ i ^ ^ ' ^ ^ i ^ ^ y g He

Dabba

CROSS SEC

TrW Imw KFA " " * "^'[*i:i:;^"fctsu''*SE^i!«i,;ioii^Kt't^:i

n.

:r-Am.4,

Ins " i^Pllics

?;:« . «rNpi» .;.;.; i^*ija"s.i.;.:;^i*w.;:i; ;';$ Exfor /Cros* »ecBon* I Tc99 * 10Q47iM8-0(6t*VI971)
<1«4.030-Q(«i6oi) I N ^ L ) « - N & « . , E 1 G / C r o s s s e e * m

»

IIWO.txa-OCC^ISSSJW.tOTJ^JOuiBiiettfaii

s.;i_46-:Z:;;.S

-4.*

9/ < J j , r .

!

Show horizontal b»S

Figure 1. JANIS plot window with the 'Chart of Nuclides' window in the background. The total cross section of 99Tc is plotted from JEFF-3.0 and JEFF3.1 and compared with EXFOR data.

A variety of output formats exist in JANIS. For the graphical display, the PS/EPS and PNG formats are possible, and tabular data can be stored in CSV format (Comma Separated Values) for further use in other software (e.g. MS Excel). Updates of the software can be automatically downloaded through the live-update feature. Version 2.2 will be available soon on the JANIS web page. Feedback is appreciated and can be posted at [email protected]. 5. The Joint Evaluated Fission and Fusion (JEFF) project The Joint Evaluated Fission and Fusion (JEFF) project is a collaboration between the countries participating in the NEA Data Bank. The JEFF library comprises of sets of evaluated nuclear data, mainly for fission and fusion applications; it contains a number of different data types, including neutron and

242

proton interaction data, radioactive decay data, fission yield data, thermal scattering law data and photo-atomic interaction data. The JEFF project is managed by the Scientific Co-ordination Group (SCG) that meets biannually. Technical achievements are reviewed at working group meetings in conjunction with the SCG meetings. The topics covered are benchmarking, testing and evaluations, radioactive decay and fission yield data, experimental data needs, fission product data, and fusion relevant data. The JEFF project also consists of the European Fusion File (EFF) and European Activation File (EAF) projects (funded by the EC Fusion Programme). These projects are directed by the EFF/EAF monitor group. This group hold their meetings in conjunction with the other JEFF working group meetings, and joint sessions are held to discuss matters of common interest for all the JEFF working groups. The first joint library was JEFF-3.0 [1] that was released in 2002. The history of the JEFF project is shown in Fig. 2. History 1983 1383 1984 198E. 1

«;

1888 T989 1990 1991 1992 199J 1994 199S 1998 199? T998 1999 2000 3001 2002 2063 200-1 2005

eftlx

Fission Application

•. .11 I'.i | i \ !

JEFF-3.1 Library

JEF-2.2 t ihrary

JEFF-3/A JEFF-3.0 Fusion Neutronic

:. i • f ' l voi.^I

CVnpittio

EFF Libn

(«...II» ; EFF-2.4 Library \

JEFF Project

Library

JEFF-3.1/RDD JEFF-3.1/SFY JEFF-3.1/NFY

1982 1980 1981 1985 198G 1987 1980 1989 1990 1991 199! 1993 1994 1995 1998 1997 1998 1999 2000 2001 2002 20

Figure 2. JEFF history from the fission and fusion projects JEF and EFF in the 1980s to the release of JEFF-3.1 in May 2005.

The JEFF-3.1 Nuclear Data Library is the latest version of the Joint Evaluated Fission and Fusion Library. The complete suite of data was released in May 2005, and contains general purpose nuclear data evaluations compiled at the OECD Nuclear Energy Agency (NEA) Data Bank in co-operation with several laboratories in NEA Data Bank member countries. Within the framework of the JEFF-3 project, the JEFF Working Group on Radioactive Data and Fission Yields decided to produce improved versions of the decay-data and fission-yield libraries with a release in conjunction with the JEFF library. Activation data has also been included in the latest version.

243

JEFF-3.1 combines the efforts of the JEFF and EFF/EAF Working Groups who have contributed to this combined fission and fusion file. The neutron data library covers 381 isotopes or elements, which is an increase from 340 in JEFF3.0. There are 26 isotopes in the proton data library, and 9 materials are covered in the thermal scattering law file. A great achievement was to include covariance data for many isotopes in the neutron data library. All actinides have now extended information on delayed neutron data in that they all are presented in eight-group formalism. The special purpose library on activation data contains 774 target nuclei with over 12600 neutron induced reactions. Included is also radioactive decay data with about 3852 isotopes and spontaneous and neutron induced fission yield data. Processed data for MC applications will be made available during 2005 as well as full documentation of JEFF-3.1. The data can be downloaded from the NEA web site, www.nea.fr/html/dbdata/JEFF. or CDs can be sent on request. Acknowledgments The authors wish to thank Arjan Plompen and Arjan Koning for valuable contributions. The entire JEFF collaboration is also acknowledged. References 1. 2. 3. 4.

NEA Data Bank, 'The JEFF-3.0 Nuclear Data Library', JEFF Report 19, (2005), ISBN 92-64-01046-7 V. McLane, "EXFOR Basics. A Short Guide to the Nuclear Reaction Data Format", report BNL-NCS-63380 (IAEA-NDS-206) (May 2000) O. Schwerer, V. McLane, H. Henriksson and S. Maev, AIP Conf. Proc. 769, 83 (2005), Int. Conf. on Nuclear Data for Science and Industry 2004. N. Soppera, H. Henriksson, A. Nouri, P. Nagel, E. Dupont, AIP Conf. Proc. 769, 557 (2005), Int. Conf. on Nuclear Data for Science and Industry 2004.

NUCLEAR DATABASES FOR ENERGY APPLICATIONS: AN IAEA PERSPECTIVE ROBERTO CAPOTE NOY, ALAN L. NICHOLS, ANDREJ TRKOV International Atomic Energy Agency, Nuclear Data Section Wagramer Strasse 5, A-1400 Vienna, Austria The development and production of various types of nuclear data represents an important activity of the IAEA Nuclear Data Section, as demanded for a wide range of energybased applications. Recent worldwide initiatives to study innovative designs of nuclear power systems require extensive and accurate nuclear data for materials to which little attention has been devoted in the past. Four highly relevant IAEA projects are described in this paper: two are on going, and two are in the planning stage.

1. Introduction The Nuclear Data Section of the International Atomic Energy Agency (IAEANDS) has been active for many years in the development, assembly and dissemination of atomic and nuclear data in forms designed to assist Member States in their studies of nuclear properties across a wide range of energy and non-energy applications As a consequence, NDS staff are constantly monitoring user needs in order to prepare and maintain the most appropriate databases in a timely manner (see http://www-amdis.iaea.org/and http://www-nds.iaea.org/). Data requirements are identified through regular discussions with users and consultants. Under clear and unambiguous circumstances, IAEA-NDS will encourage and effectively sponsor the establishment of suitable teams of recognised specialists in particular technical areas that merit review and database development. A Co-ordinated Research Project (CRP), Consultants' Meeting (CM), direct staff efforts or a combination of all three can be dedicated to the development and assembly of specific data files. Usually such efforts result in the production of a new (or a significant upgrade of an existing) database. Each CRP involves typically 5-15 scientific groups from different countries working together over a period of 3-4 years, maintaining contact throughout the course of the CRP and meeting at regular intervals to discuss their progress and problems. Some of the most important data requirements involve continuous commitment (neutron cross sections, and nuclear structure and decay data), and these databases are maintained through organised networks

244

245

of expert compilers and evaluators under the sponsorship of the IAEA-NDS. Nuclear data for energy applications are particularly important, and some of the most relevant and on-going activities are discussed in the following sections. 2. Reference Input Parameter Library (RIPL) A broad spectrum of applications have to be regularly addressed, from nuclear power reactor and shielding design through to transmutation of nuclear waste, when considering low-energy nuclear reactions induced by light particles such as neutrons, protons, deuterons and photons. Our theoretical understanding of nuclear phenomena has reached a reasonable degree of reliability, and nuclear modelling programs are widely used to assess and guide nuclear data evaluations (with measurements remaining crucial for data testing and benchmarking). Since such modelling codes require a considerable amount of numerical input, the IAEA has instigated extensive efforts to develop a library of validated nuclear-model input parameters, referred to as the Reference Input Parameter Library (RIPL). A second release of the RIPL library for nuclear reaction calculations is now available (RIPL-2), and a third phase of this project is on-going and designed to extend the applicability of the library to cross sections for reactions on exotic nuclei, incident energies up to 200 MeV, and reactions induced by charged particles (RIPL-3). A starter file in electronic format (known as the Reference Input Parameter Library-1 (RIPL-1)) was developed and made available to users throughout the world in 1997 (IAEA, 1998; and http://www-nds.iaea.org/ripl/). Immediately afterwards, a second CRP was initiated on "Nuclear Model Parameter Testing for Nuclear Data Evaluation (Reference Input Parameter Library: Phase II)", and completed in 2002. This CRP resulted in the revision and extension of the original RIPL-1 starter file to produce a consistent RIPL-2 library of recommended input parameters that is targeted at users of nuclear reaction codes interested in low-energy nuclear applications. Incident and outgoing particles include neutrons, protons, deuterons, tritons, 3He, 4He and y, with energies up to approximately 100 MeV. The numerical data and computer codes included in RIPL-2 are arranged in seven segments (or directories): MASSES: directory of atomic masses that contains basic ground-state properties of nuclei, along with two theoretical predictions of masses and deformations (8979 nuclei ranging from 16 0 to A = 339). Evaluated experimental masses of Wapstra et al. (2003) are also included. LEVELS: 110 files (one for each element) with all known level schemes available from ENSDF in 1998. All missing spins were inferred uniquely for

246

each level from spin distributions, and electromagnetic and gamma-ray decay probabilities were estimated. RESONANCES: average resonance parameters were prepared on the basis of the evaluations performed by the Obninsk group, taking into account the analysis of discrepancies between similar evaluations of other groups. OPTICAL: optical model parameter database is provided in two forms: full library (archival form) and abbreviated library with all single-energy potentials removed (user file). When there are insufficient experimental data to define phenomenological OM parameters (Koning and Delaroche, 2003), the evaluator has to resort to either global parameterizations or new microscopic approaches. Density distributions requested as input for microscopic OMP calculations are stored in the MASSES segment. LEVEL DENSITIES: total level density sub-directory contains a revised version of the Back Shifted Fermi Gas (BSFG), Gilbert-Cameron (GC) and Generalized Super-fluid Model (GSM) parameters prepared by the Obninsk group, which are consistent with both the recommended PJPL-2 neutron resonance parameters and the evaluated parameters of the recommended lowlying levels (LEVELS segment). Microscopic HF-BCS calculations of the nuclear level densities are based on the realistic microscopic single-particle level scheme (Goriely et al., 2001). GAMMA: contains parameters that quantify giant resonances, experimental gamma-strength functions and methods for calculating gamma-emission in statistical model codes. Recent analytical expressions for dipole transition y-ray strength functions provide reliable results over a relatively wide range of y-ray energies from zero up to above the giant dipole resonance (GDR) energy. The experimental GDR parameters were provided by the Chinese group, and are represented by Lorentzian fits to the total photo-neutron cross sections for 102 nuclides ranging from 51V to 239Pu as compiled by Dietrich and Berman (1988). FISSION: includes global prescriptions for barriers and nuclear level densities at saddle points. The RIPL-2 database was released in July 2002, and is available on the Web through: http://www-nds.iaea.org/RIPL-2/. This site provides a means of downloading entire RIPL-2 segments, individual files, and selected data. A CDROM with the complete RIPL-2 library can also be requested cost-free from the IAEA Nuclear Data Section. During the development of RIPL-2, several important issues could not be addressed within the work programme. Therefore, a third phase was initiated in 2003 to extend the applicability of the library to cross sections for reactions on nuclei far from the stability line, incident energies beyond 100 MeV, and

247

reactions induced by charged particles (CRP on "Parameters for calculation of nuclear reactions of relevance to non-energy nuclear applications"). Such studies require cross sections, spectra and angular distributions for reactions that were not the prime objective of RIPL-2. The most noteworthy additions in the formulation of RIPL-3 will be as follows: charged-particle reactions; reactions at high energies for ADS (up to 1.5 GeV), production of medical radioisotopes (up to 50 MeV), and radiotherapy (up to 1 GeV); reactions on nuclei far from the stability line for ADS and astrophysics; specific low-energy charged-particle reactions for materials analysis. These requirements pose new challenges that require consideration of the following issues: use of microscopic models to deduce parameters of nuclei far from the stability line in addition to experimental data available for stable nuclei; uncertainty estimates (and/or range of variation) for library parameters; fission at high energies (serious problem in the design of ADS); improved determination of model parameters for deformed nuclei, especially optical model parameters for coupled-channel calculations and radiative strength functions; treatment of collective enhancements of nuclear level densities at high energies; temperature dependence of the giant dipole resonance width; linkage to intra-nuclear cascade model used for cross-section calculations at high energies; use of forthcoming experimental data from n_TOF facility at CERN and HIND AS collaboration for deducing model parameters and library validation. An update of the RIPL-2 optical model segment was recently carried out, addressing some of the REPL-3 requirements. Analytical integration of the dispersive relations was included in the REPL retrieval code, and the format of the optical model database was extended to allow for coupled-channel dispersive potentials and to include recently published optical model parameters. The majority of new data sets are coupled-channel potentials that address the modelling needs for nuclear data production on thorium and light actinides of direct relevance to studies of the Th-U fuel cycle. 3. Nuclear Data for the Th-U Fuel Cycle Emerging nuclear technologies that utilize advanced fuel cycles are based on design criteria encompassing increased inherent safety, reduction of the risk of fissile material proliferation, reduction of the quantity of highly toxic radioactive waste for long-term disposal, and sufficient abundance of natural resources.

248

Fuel cycle concepts based on thorium have many features that meet the above criteria. However, the quality of nuclear data for the relevant nuclides such as 232

Th,

23U33

Pa

and

232,233,234^

is

infcrior

t0

the

equivalent

data

for

conventional fuel cycles based on uranium. Uncertainties in the nuclear data are a factor of three larger in some cases than the target accuracies set by plant designers, and therefore these inadequacies in the nuclear data are being addressed by means of an on-going CRP entitled "Evaluated nuclear data for the Th-U fuel cycle". A co-ordinated research project was initiated in 2003 to produce new evaluations for the above-mentioned nuclides. The second research coordination meeting was held in December 2004, at which good progress was reported on the 232Th evaluation. At present, the following data and modelling capabilities are available: Average resonance parameters in the unresolved resonance range, complete with uncertainties; their evaluation includes the most recent measurements from Geel and the n_TOF collaboration. Resolved resonance parameters derived from all available experimental data, including the most recent measurements from Geel and preliminary data from the n_TOF collaboration. Optical model parameters (OMP) that are applicable to a broad range of actinides (Soukhovitskii et al, 2004), and dispersive optical model parameters for nucleon-induced reactions on 232Th derived recently by Soukhovitskii et al. (2005); these data have been included in the working version of the RIPL-3 database, as maintained by IAEA-NDS staff. A new version of the EMPIRE-II code has been released (see http://wwwnds.iaea.org/empire/) that includes coding to take advantage of the new dispersive coupled channel OMPs, and advanced fission formalism to describe multiple-hump barriers within the optical model for fission. Integral benchmark experiments identified for data validation - preliminary calculations with existing libraries had been performed to set the reference results. For illustration, Figs. 1 and 2 depict the proton- and neutron-induced reaction and total cross section of 232Th as calculated from different OMP data sets in the updated RIPL-3 database, and compared to measured values (from EXFOR). Tasks that remain to be completed include the following: Application of OMP to calculate the cross sections and outgoing particle distributions with EMPIRE-II, and assembly of a complete file (including the newly evaluated resonance parameters) for incident neutron energies up to 60 MeV or more.

249 Extension of the methodology to produce complete evaluations for other relevant nuclides. Validation of the new evaluated data files through modelling of suitable integral benchmarks. 2500

.a JE c o o

-i—•—r

~i—•—i—•—r

2000

CD CO CO

8 c .o o TO CD

1500

Maslov 2004 (RIPL 4603) Soukhovitskii et al (RIPL 5601) Ignatyuk et al (RIPL 4303) • Soukhovitskii et al (RIPL 4608) NEW

1000-

cc 500-

— i —

T—

20

40

80

60

100 120 140 160 180 200

Energy [MeV] Figure 1: Proton-induced reaction cross section of 232Th calculatedfromdifferent OMPs. I 22.0k -

~

£

s

Total cross

•f %

20.0k -

'"I

' ' """

\ \\ •

18.0k-

\

%

16.0k-

TV \ X.

14.0k-

\

12.0k-

*

^^&

10.0k8.0k6.0k-

1E-3

« Maslov (RIPL 603) ^ t • Soukhovitskiiaa;(RIPL2601) 'kkIgnatyuk el al (RIPL 2303) ^ J Soukhovitskii a al (RIPL 608) NEW 1

1

0.01 0.1 Energy [MeV]

1

Figure 2: Neutron-induced total cross section of

1 Energy [MeV] 232

Th calculated from different OMPs.

250

4. Planned Co-ordinated Research Projects 4.1. Minor actinide neutron reaction data Innovative reactor concepts with high burn-up and continuous recycling of heavy actinides require more accurate data for the minor actinides, for which fewer measurements are available because of their relative unimportance in conventional reactor designs. Assessment of available experimental information in combination with modern statistical data evaluation methods and advanced nuclear model codes would permit the preparation of more accurate databases for these materials. As described in Section 3, an earlier CRP on "Evaluated nuclear data for Th-U fuel cycle" was established to address the lighter minor actinides; the presently proposed CRP on "Minor Actinide Neutron Reaction Data for Closed Fuel Cycle Reactor Concepts" would address the heavier minor actinides. The International Nuclear Data Committee (INDC) provides guidance to the IAEANDS and endorsed the proposed CRP, but also recommended that a subgroup of the Working Party on Evaluation Co-operation (WPEC) of the OECD-NEA should assist in defining the precise requirements for such a project. A primary aim will be to produce evaluated nuclear data files of the heavy minor actinides with particular emphasis on the Am and Cm isotopes. Output of the CRP will provide more accurate data needed in the design studies of innovative reactor concepts, such as those arising from the Generation-IV Project. 4.2. Updated decay data library for actinides Actinide nuclides and their decay chains need to be well characterised and their decay parameters quantified with high confidence, particularly for advanced power-related applications involving fuel manufacture, plant operation, reprocessing and waste storage. A previous CRP from 1978 to 1986 addressed the preparation of such a database directly (IAEA, 1986), and provided the catalyst for a series of new measurements that continued well into the 1990s. All of this new work and earlier data need to be compiled and evaluated to produce an updated set of recommended decay data that would replace the current IAEA database assembled in 1985/86. Specific aims of this CRP will include the following: promotion of decay data research and development; recommended decay data files for the actinides of most relevance to facility design, waste management, safety assessments and safeguards/ proliferation issues, along with non-energy applications.

251

Suggested actinides for decay-data evaluation include: 226Ra and daughters (possibly), 232Th and daughters (possibly), 231Th, 231Pa, 233Pa, 232237 u, 239U, 237 N P , 239Np, 238"242Pu, " V M2mAm, 243Am, 242Cm, 244Cm and 252Cf. As with other NDS-generated databases, the new data will undergo critical assessment, and the recommended sets of data will be published as a technical report and be made available through the World-wide web and NDS home page. The institutes proposed to undertake this work are the recognised specialists in this particular field of decay-data evaluation, and are being selected with care. Their numbers are limited and will require some fortification as the programme develops from 2005 to completion in 2009. 5. Concluding Remarks The nuclear data initiatives discussed above represent a healthy fraction of the on-going and planned IAEA CRPs identified specifically with energy applications. However, this brief review is not comprehensive, and note should also be made of a number of relevant database compiled and evaluated in recent years including photonuclear data (IAEA, 2000), update of FENDL-2.1 for fusion applications (Lopez Aldama and Trkov, 2004), and update of X-ray and y-ray decay data standards (Nichols, 2004). The IAEA Nuclear Data Section has to be sufficiently flexible to respond to the demands of Member States, and formulate technical projects that aid atomic and nuclear data users in a timely manner. However, predicting user requirements in the years ahead remains the most difficult challenge to the credibility of both national and international nuclear data centres in the 21 st century. Users' needs are monitored continuously by NDS staff and reviewed regularly with the assistance of external expertise. The various forms of ongoing IAEA data development programme and their resulting databases constitute the initial and final products of such crystal-ball gazing. Acknowledgement The authors would like to acknowledge the work of past and present colleagues of the IAEA Nuclear Data Section in their formulation and assembly of the wide range of nuclear databases alluded to and discussed in this paper.

252

References 1. Dietrich, S.S. and Berman, B.L. 1988. Atlas of photoneutron cross sections obtained with monoenergetic photons. At. Data Nucl. Data Tables 38, 199338. 2. Goriely, S., Tondeur, F. and Pearson, J.M. 2001. A Hartree-Fock nuclear mass table. At. Data Nucl. Data Tables 77, 311-381. 3. IAEA 1986. Decay data of the transactinium nuclides. Technical Reports Series No. 261, Vienna, Austria. 4. IAEA 1998. Handbook for calculations of nuclear reaction data — Reference Input Parameter Library (RIPL). IAEA-TECDOC-1034, Vienna, Austria. 5. IAEA 2000. Handbook on photonuclear data for applications: cross sections and spectra. IAEA-TECDOC-1178, Vienna, Austria. 6. Koning, A.J. and Delaroche, J.P. 2003. Local and global nucleon optical models from 1 keVto 200 MeV. Nucl. Phys. A713, 231-310. 7. Lopez Aldama, D. and Trkov, A. 2004. FENDL-2.1: update of an evaluated nuclear data library for fusion applications. IAEA report INDC(NDS)-467, Vienna, Austria. 8. Nichols, A.L. 2004. IAEA Co-ordinated Research Project: update of X-ray and y-ray decay data standards for detector calibration and other applications. Appl. Radiat. IsoL 60,247-256. 9. Soukhovitskii, E.Sh., Chiba, S., Lee, J-Y., Iwamoto, O. and Fukahori, T. 2004. Global coupled-channel optical potential for nucleon-actinide interaction from 1 keV to 200 MeV. J. Phys. G: Nucl. Part. Phys. 30, 905920 10. Soukhovitskii, E.Sh., Capote, R., Quesada, J.M. and Chiba, S. 2005. Dispersive coupled channel analysis of nucleon scattering from 232Th up to 200 MeV, submitted to Phys. Rev. C. 11. Wapstra, A.H., Audi, G. and Thibault, C. 2003. The AME2003 atomic mass evaluation, (I) Evaluation of input data, adjustment procedures, (II) Tables, graphs and references. Nucl. Phys. A729, 129-676.

NUCLEAR DATA EVALUATION FOR GENERATION IV G. NOGUERE, O. BOULAND, A. COURCELLE, E. DUPONT, O. SEROT, J.C. SUBLET Centre d'Etudes Nucleaires de Cadarache 13108 Saint Paul les Durance, France Evaluation activities are crucial steps to combine experimental data from microscopic experiments into coherent and comprehensive nuclear data sets for nuclear energy system applications. The nuclear data group of the CEA/DEN is involved in the elaboration of the European library JEFF since the beginning of the project. New evaluated files for fission products and actinides rather than successive revisions of previous evaluations are planned to satisfy GEN-IV issues.

1. Introduction Relevant prediction of core characteristics for innovative nuclear power systems requires improved nuclear data libraries. Neutron transport, kinetic studies, core evolution, heating and shielding calculations require the knowledge of a wide number of nuclear data. To achieve the challenging technology goals set by the GENERATION IV International Forum in the sustainability, safety and reliability fields, the nuclear data group of the CEA/DEN-Cadarache focuses some of its evaluation activities on the neutron properties of actinides (Cm, Am, Np, U, Pu) and fission products (Eu, Sm, Cs, Xe) of interest for conventional nuclear reactors and innovative "fast" concepts (GFR, LFR, SFR, EFR). Parallel long-term R&D activities will be launched to develop a methodology for producing covariance matrices on multi-group cross sections from thermal energies up to the upper energy limit of the Unresolved Resonance Range. 2. Experimental Activities on Actinides Experimental programs on minor and major actinides are supported by the CEA to improve the present data and to provide accurate uncertainty information. Present experimental projects aim (1) to reach a better description of the energy dependent 241Am capture isomeric ratio to 242gAm, (2) to improve the most reactor relevant Fission Product Yields produced by 235 U(n,f), 239Pu(n,f) and 241Pu(n,f) reactions and (3) to get consistent sets of actinides fission cross sections. Experimental campaign on 240_244Cm and 253

254 241,243

Am, conducted by the Nuclear Center of Bordeaux-Gradignan (CENBG), is described elsewhere [I]. 2.1. Capture Isomeric Ratio Among Minor Actinides interesting the safety margins issues, the 241Am isotope plays a significant role in conventional nuclear power plants and in the future GEN-IV systems. The 241Am capture Isomeric Ratio (IR) to 242g Am is a key nuclear data for estimating the production of 242mAm, 244Cm and 242Pu. Recent interpretations of independent Post Irradiated Experiments (PIE) performed with the deterministic APOLL02 and ERANOS2 codes have led to a revision of the 241Am low neutron energy cross sections and of the capture Isomeric Ratio. Figure 1 compares the experimental IR data available in the literature with those obtained at Cadarache and at the Los Alamos National Laboratory (LANL). r

...._.,.;,„,.

PIE data Cadarache i •«

/

i

0.9

*-*._

/

.

.a 0.8

'*

i

' o

LANL chemistry

y

\

i

N

\

"!

\

>

/ TALYS

0.7

calculations i

0.6 _a 10

-»;

10^

10"'

10"

10'

1 0 ' 10 J Energy (eV)

10*

» 10 s

10°

10

Figure 1. Experimental energy dependent Am Isomeric Ratio to 242g Am. Post Irradiated Experiments data (PIE) were obtained at the CEA/DEN Cadarache with the deterministic APOLL02 and ERANOS2 codes [3,4]. The dashed curve is the JEFF-3.1 evaluation [2].

In order to confirm the trend given by the PIE integral data from Cadarache, microscopic measurements are planned in the framework of a European collaboration. Raw material (800 mg of A111O2) is expected to be provided

255

by the ATALANTE laboratory of the CEA-Valrho and the sample preparation is foreseen by the Institute for TransUranium elements (ITU). The thermal capture cross section to 242gAm will be investigated at the Prompt Gamma Activation Analysis (PGAA) facility of the research reactor of the Budapest Neutron Centre (BNC). The thermal and epithermal energy ranges will be studied at the Geel LINear Accelerator (GELINA) of the Institute for Reference Materials and Measurements (IRMM) by means of the Time-Of-Flight technique (TOF). The fast energy range will be measured at the VdG accelerator of the IRMM with the activation technique. After analysis, the experimental data will be compiled in ENDF format for the next release of the JEFF library. 2.2. Fission Yields Data Data related to Fission Yield information (FY) are important for reactor and fuel cycle applications. The European Fission Yield library JEFF-3.1 is based on UKFY3.6, produced by Robert Mills at the British Nuclear Fuels company (BNFL). The library contains FY distributions for 21 isotopes (Th, U, Np, Pu, Am, Cm and Cf). A closer inspection of each mass and charge distributions shows some significant discrepancies with the other libraries [5]. Figure 2 represents the charge yields produced by the 241Pu(nth,f) reaction. The right scale indicates the FY values and the left scale stands for the values of the ratio JEFF31/ENDFB6.8- Ideally, the ratio has to be close to unity. Some significant discrepancies for non-negligible FY values can be observed. Specific work is then needed to improve the experimental data of the most reactor relevant FY produced by the 235U(n,f), 239Pu(n,f) and 241Pu(n,f) reactions. The experimental program supported by the CEA/DEN aims to study thermal neutron induced fission reactions at the Lohengrin mass spectrometer facility of the High-Flux ILL Reactor. The data accuracy will be significantly improved with the development of a new detection system involving an ionisation chamber, a beta detector and two clover gamma detectors each composed of four Ge detectors.

256 0.20

Charge Figure 2. Charge distribution of the fission fragments produced by the 241Pu(nth,f) reaction. The right scale gives the JEFF3.1 Fission Yield values. The dashed curve represents the ratio between the European (JEFF3.1) and American (ENDFB6.8) Fission Yield libraries.

3. Fission Product Evaluations A long term R&D activity on Fission Products has been launched since the beginning of the collaboration between IRMM and CEA. Capture and total cross sections of 99Tc, 127'129I and 103Rh have been measured at the GELINA facility and analysed to produce new Evaluated Nuclear Data Files for JEFF3.1 [6-8]. The analysis of the experimental data have been performed to ensure physics consistency over a broad neutron energy range (0-30 MeV). The RMatrix code SAMMY [9], or alternatively the Resonance Shape Analysis code REFIT [10], were used to describe the Resolved Resonance Range (RRR) in term of Reich-Moore parameters. The statistical modelling of the Unresolved Resonance Range (URR) was performed with the FITACS code [11] incorporated into SAMMY. Above the first inelastic threshold ("continuum" region), consistent calculations of all open channels was done with the TALYS code [12] in association with the coupled-channels optical model code ECIS using a semi-microscopic deformed potential SMOMP [13]. Example of the cross section modelling of the 103Rh(n,y) reaction is shown in Figure 3. The RRR has been revised on the basis of TOF data measured at the GELINA facility. The URR and the high neutron energy

257

range have been analysed with experimental data (cross sections, angular distributions...) available in the EXFOR database.

Neutron energy (eV)

Figure 3. Modeling of the Rh(n,Y) reaction [8]. Individual resonance has been interpreted in term of Reich-Moore parameters, the Unresolved Resonance Range was analysed with the Hauser-Feshbach formalism and all open channels above the first inelastic threshold have been calculated with the comprehensive nuclear reaction program TALYS [12].

The quality of the transition between the RRR and the "continuum" region has been assessed with a specific statistical modeling of the URR. The methodology consisted in deducing 1-wave average parameters from the statistical analysis of the RRR. That prior parameters (in association with the Reference Input Parameter Library RIPL-2) were used to extract posterior Hauser-Feshbach parameters from the URR. In the "continuum" region, a few model parameters (such as radiative width, level density parameter, spin cut-off, preequilibrium constant...) were slightly improved to be consistent with experimental data and with the URR evaluation. The evaluation procedure is complete when a satisfactory agreement is obtained between the Reich-Moore, Hauser-Feshbach and Optical Model average parameters. Table 1 shows the consistent I29I s-wave average parameters successively determined in the RRR, URR and in the "Continuum" region. One can observe a slight increase of the S0 values from 0.54xl0"4 to 0.62xl0"4. The former result remains consistent within the limit of the uncertainty determined in the RRR. For the next release of the European neutron library, we will intend to perform fission product evaluations following a similar methodology as used

258 for the technetium, rhodium and iodine isotopes. For that purpose, new Evaluated Nuclear Data Files for Sm, Eu, Xe and Cs are in preparation. Table 1. 129I s-wave average resonance parameters obtained in the Resolved Resonance Range (RRR), Unresolved Resonance Range (URR) and high energy range (Cont.). So stands for the neutron strength function, D„ is the mean level spacing, represents the average radiation width and R' is the effective potential scattering RRR 104S„ D„ (eV) < r y > (meV) R' (fm)

0.54 ± 0.07 27.3 ± 0.9 106.0 ± 15.2 5.6 ± 1.6

URR 0.60 27.3 108.4

Cont. 0.62 27.7 108.0

5.8

5.7

4. Multi-Group Cross Sections and Covariance Matrices The generation of the Multi-group cross sections together with relevant uncertainties is fundamental to assess the quality of neutronic simulations. The three key information that are needed to propagate the microscopic experimental uncertainties up to macroscopic reactor calculations are (1) the experimental covariance matrices, (2) the correlations between the parameters of the model and (3) the covariance matrices for the multi-group cross sections. 4.1. Experimental Covariance Matrix Assigning uncertainties to microscopic experimental data is a recurrent puzzle for evaluators. It became apparent when uncertainty information (covariance file) had to be added to Evaluated Nuclear Data File in response to requests from users involved in fission and fusion reactor development programs. A particular interest in reliable experimental covariance matrices has arisen since the rapid developments in the implementation of appropriate methodologies (IDC option) in the SAMMY code. The IDC approach makes possible to include in the calculations experimental covariance matrices determined with the AGS system [14]. AGS concerns basic operations like spectrum addition or division, dead time correction, non-linear fitting and other operations, preserving all the steps in a single file with full uncertainty propagation. 4.2. Correlations Coefficients in the Resonance Region Programs, like the SAMMY code, adjust nuclear parameters so that the theoretical curve agrees with the observed data within the limit of the uncertainties. In the Resolved Resonance Range, the model parameters are

259

the energies, the partial widths and the effective radii. The so-called datarelated parameters are usually the sample thickness, the effective temperature and the parameters describing the energy resolution of the experimental set-up. The SAMMY code uses the familiar generalized least squares fitting procedure to get an appropriate set of nuclear parameters and to propagate uncertainties for virtually any parameters. The fitting procedure takes into account the experimental covariance matrix (IDC option), the theoretical covariance matrix (prior uncertainties on resonance parameters) and the uncertainties on fixed data-related parameters (PUP option). The final covariance matrix then properly reflects all uncertainty information and could assess realistic long-range correlation coefficients between the model parameters [15]. 4.3. Multi-group Covariance Matrix The last step before running transport calculations with deterministic codes consists in producing multi-group cross sections and multi-group covariance matrices. The diagonal of the matrix is the variance of the cross section for a given group. The off-diagonal elements are the covariances which provide a measure of how strongly correlated the cross section values are. Group-averaged cross sections and their covariances can be produced with the SAMMY (RRR and URR) and TALYS (high energy region) codes. Example of a multi-group covariance matrix in the case of the 127I(n,y) reaction is shown in Figure 4. The covariances obtained with the two codes have been merged in the same figure. The results are expressed in term of correlation matrix according to an arbitrary group structure (139 groups) covering a wide neutron energy range from 0.01 meV to 20 MeV. The SAMMY calculations have been performed over the full RRR (up to 10 keV). Correlation coefficients above 10 keV have been produced by TALYS1. One can distinguish four main broad group structures. The low energy broad group structure stands for the contribution of the bound levels (negative resonances). The second and the third one represent the correlations respectively between the Reich-Moore and the Hauser-Feshbach parameters. The last region, above 10 MeV, is dominated by the preequilibrium gamma emission. This model is significantly different from the

1

A detailed description of the full capability of TALYS in generating covariance data with nuclear model is given elsewhere [16].

260

lower energy Hauser-Feshbach GDR-based capture model. Hence, no correlations could exist between these two formalisms.

Figure 4. I(n,y) multi-group correlation matrix calculated with the SAMMY [9] and TALYS [12] codes. Group-averaged calculations have been performed over an arbitrary group mesh (139 groups) from 0.01 meV to 20 MeV.

In the present approach, the model parameters used to describe the RRR, URR and optical model region are not correlated. For fast reactor applications, RRR and URR should be treated with a similar nuclear model. A specific research program is required to extend the numerical procedure available in SAMMY above the upper energy limit of the resonance region by sampling, with appropriate Monte-Carlo techniques, resonance ladders and realistic uncertainties from conventional statistical resonance parameter laws (Gilbert and Cameron level density formula, Wigner level spacing distribution, Porter-Thomas reduced neutron width distribution). The continuity with the high-energy region will consist in obtaining a satisfactory agreement with the variances calculated with the TALYS code. We will intend to use this methodology in our future evaluation. 5.

Conclusions

The focus of this paper is to give to the nuclear data community an overview of the evaluation activities planned at the CEA/DEN, Cadarache. New investigation of the Unresolved Resonance Range of the main fission product cross sections is a good opportunity to revise the Evaluated Nuclear Data Files with modern tools using the most recent experimental data available in the EXFOR data base. This approach will ensure physics consistency over a broad energy range (0 - 30 MeV). Parallel long-term

261

experimental campaigns dealing with actinides are supported by the CEA/DEN. Fission and capture cross sections as well as independent and cumulated fission yields will be investigated in order to provide accurate experimental information for future evaluations. These activities are planned in the framework of collaborations with the CEA/DAM (Bruyere le Chatel), the CEA/DSM (Saclay), the Institute for Reference Materials and Measurements (IRMM), the Budapest Neutron Center (BNC), the Nuclear Center of Bordeaux-Gradignan (CENBG), the Oak Ridge National Laboratory (ORNL) and with the scientific staff of the High Flux ILL Reactor. Acknowledgments The authors wish to thank N.M Larson and A. Koning for their continuous support in using the SAMMY and TALYS codes. We also express our gratitude to E. Bauge for his valuable work in the high neutron energy range. References: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

G. Barreau et al., this workshop. O. Bouland and D. Bernard, OECD, NEA JEF/DOC-1086 (2005). D. Bernard et al., OECD, NEA JEF/DOC-1026 (2004). J. Tomasi, Tech. Rep. CEA, NT-DER/SPRC/LEPh-04/217 (2004). O. Serot and R. Mills, OECD, NEA JEF/DOC-1000 (2004). F. Gunsing et al., Phys. Rev. C61 (2000) 054608. G. Noguere et al., Int. Conf. on Nuclear Data for Science and Technology, Santa Fe (2004). E. Dupont et al., Int. Conf. on Nuclear Data for Science and Technology, Santa Fe (2004). N.M. Larson, RISC peripheral shielding routine collection, SAMMY ORNL/TM-9179/R7 (2004). M.C. Moxon Tech. Rep. Harwell laboratory, CBNM/ST/90-131/1 (1990). F.H.Frohner, Nucl. Sci. Eng. 103(1989)119. A.J. Koning et al., Int. Conf. on Nuclear Data for Science and Technology ND2004, Santa Fe, USA (2004). E. Bauge et al., Phys. Rev. C63 (2003) 024607. C. Bastian C , Proc. Int. Conf. on Neutrons in Research and Industry, Crete, Grece (1979). L.C. Leal, this workshop. A.J. Koning, this workshop.

I M P R O V E D EVALUATIONS OF N E U T R O N - I N D U C E D REACTIONS O N A M E R I C I U M ISOTOPES

P. TALOU, T. KAWANO, P.G. YOUNG, M.B. CHADWICK, E.J. PITCHER Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA E-mail: [email protected]

We report on our current efforts to re-evaluate neutron-induced reactions on americium isotopes (241, 242g,m and 243) in the fast energy region up to 20-MeV. Modern theoretical modeling was used extensively in all those evaluations, due to the rather limited number of experimental data available. This paper is mostly concerned with Am fission cross sections evaluations.

1. Introduction New nuclear technology needs (e.g., Accelerator-Driven Systems (ADS) or Generation-IV type nuclear reactors) are pushing for more modern and accurate nuclear data evaluations for some reactions and isotopes that have not been commonly studied in the past. This is the case for all minor actinides (MA) and in particular americium isotopes that are produced during the regular operations of a nuclear power plant. They represent some important part of the nuclear waste that can be partitioned and then transmuted in innovative waste burner devices. Historically, the MA including Americium isotopes have received little attention from the nuclear data and applications communities, therefore explaining the rather poor status of the ENDF/B-VI evaluations. With the new needs just mentioned, it has been shown that the current uncertainties on neutron-induced reaction cross sections on specific americium isotopes are too large for some important target accuracies *. In the present work, we report on our recent efforts to re-evaluate neutron reactions on americium isotopes ( 241 Am, 2 4 2 s Am, 2 4 2 m Am, and 243 Am) in the keV to 20-MeV energy region. In the next section, we will briefly introduce the theoretical modeling and evaluation techniques used. Then, we will present some results for several isotopes of interest, and conclude by mentioning a few still needed improvements. 262

263

2. Theoretical Modeling and Evaluation Methodology The GNASH nuclear reaction code 2 was used extensively in the present work. The backbone of GNASH calculations is the Hauser-Feshbach equations that describe the evaporation process following the formation of a compound nucleus in statistical equilibrium, after the initial interaction between the incident particle and the target nucleus. Non-statistical processes like direct reactions and pre-equilibrium emission of high-energy particles were also included in our calculations. The neutron-induced fission (n, / ) reaction is the main mechanism at play in the transmutation of the americium isotopes. This channel is quite important, and even a small error in the determination of its cross section can lead to significant uncertainties in competing channels. In the cases of 242m - 243 Am, experimental fission cross sections were analyzed with a Los Alamos version of the GLUCS Bayesian code 3 , and the resulting best cross section was used to adjust the fission model parameters in GNASH. These parameters were extrapolated to the case of 242fl Am, for which no experimental data exist. The ECIS96 coupled-channels code4 was used to calculate the total, reaction and shape elastic cross sections for all americium isotopes. An isospin-dependent deformed optical model potential was adopted for all isotopes, based on previous studies by Young5. This potential provides a rather good agreement with the experimental n+ 2 4 1 Am total cross section. The channels transmission coefficients used in the GNASH Hauser-Feshbach decay sequence were also obtained with the ECIS96 code. Nuclear level densities were calculated with the Ignatyuk-GilbertCameron prescription 6 . The capture cross sections were calculated using the generalized-Lorentzian strength function formalism of Kopecky-UhI r at low-energies, and direct-semidirect calculations at higher energies8. 3. Evaluation Results The results obtained are discussed in the following sections for each isotope separately. 3.1.

241

Am

The neutron radiative capture of 241 Am produces both 242 Am ground (Ti/2=16 h) and meta-stable (T!/2=141 yr) states. An ambiguity in the branching ratio significantly influences the prediction for the production of Am and Cm isotopes.

264

The ENDF/B-VI capture cross section was evaluated based on the experimental data by Vanpraet et al.9 However, the measurement of the neutron-capture ratio to 197 Au by Wisshak and Kappeler 10 supports the higher values of Gayther and Thomas 11 . We re-evaluated the capture cross section to reproduce the experimental data by Gayther and Thomas 11 , and Wisshak and Kappeler 10 , simultaneously. The results are shown in Fig. 1. The present result is about 15% larger than the ENDF/B-VI data in the hundreds-keV energy region.

10"3

Figure 1. data.

10"2

Comparison of evaluated

241

10"1 10° Energy [MeV]

101

A m capture cross sections with the experimental

We have obtained new values for the isomeric ratio (IR) for the production of Am-242 in its ground- versus isomeric-state 12 . These new values are higher than the evaluated data in JENDL-3.3 and ENDF/B-VI, as shown in Fig. 2. The experimental data by Dovbenko et al.13 and by the radiochemists at LANL (C-INC) are both integral measurements of the isotopic production ratio in a given neutron field. Note that our new higher value for IR is supported by a recent post-irradiation experiment analysis 14 ' 15 . 3.2.

242

»Am

The ground-state of americium-242 is very short-lived (16 h) and is therefore very challenging to measure directly. In fact, no differential measurement exists for any neutron-induced reaction on this nucleus. However, Fioni et

265 I

-=. 0.9 (0

|—iji inn^—i i nun)—riTiiin—i i inin)—i i MIIH|—i i niiiq—TTTTTWJ—i i IIITTI

I I inm| '

iF

o

t ? 0.8 3 O

o 0.7 •(3

cc | 0.6 o c (0

15

0.5

0.4 10"2

Shinohara(1997) Wisshak(1982) Mughabghab(1984) Dovbenko(1961) LANL Chemistry, C-INC ENDF/B-VI JENDL-3.3 Present

"J

10"1

•*

10°

J

r •••••-!—i i IIIIIJ

i • i i * i i—*—i ' a i o

J

•

-1—

101 10 2 10 3 10 4 105 Incident Neutron Energy [eV]

Figure 2. Branching ratio to produce the ground-state of perimental and evaluated data.

242

J

tJLiJMM

10 6

10 7

A m , compared with ex-

al.16 recently reported an experimental value for the thermal-neutron capture cross section of (330 ± 50) barns. This value is in fair agreement with the existing ENDF/B-VI value of 252 barns, and in strong disagreement with the JEF-2.2 (same as JENDL-3.2) value of 5511 barns. As mentioned by the author, such a high value for cr^1 would strongly inhibit the incineration of americium waste in all planned innovative nuclear reactors and ADS designs. The values found in the newer JEF-3.0 and JENDL-3.3 evaluations were modified to reflect the result of Fioni et al. Since no differential measurement exists for the neutron-induced reactions on 242ffAm, our evaluation is almost entirely based on nuclear model calculations. The fission barrier parameters were taken directly from our analysis of the n+ 2 4 2 m Am evaluation. The calculated neutron-induced fission cross section is shown in Fig. 3, along with the recent JENDL-3.3 evaluated fission cross section and the data by Younes et al. obtained from the surrogate reaction 242 Pu( 3 He,
266 i

_

3

|

— Present - - JENDL-3.3 o Younes, 2004

Neutron Energy (MeV) Figure 3. The calculated neutron-induced fission cross section for 2 4 2 f l Am is compared to the recent JENDL-3.3 evaluation, and to the data points from Younes inferred from the surrogate reaction 2 4 2 Pu( 3 He,d/).

3.3. 242m Am

Americium-242 has a very long-lived (141 yr) isomeric state (5~) at an energy Em = 80 keV. If populated, it can then further capture neutrons to produce amounts of 243 Am. A coupled-channels calculation was performed with the ECIS96 code, considering the target nucleus in its 5 _ isomeric state. The three first members of the rotational band built upon the isomeric state were coupled in the calculation.

Thanks to its relatively long half-life, the neutron-induced fission (n, / ) cross section of 2 4 2 m Am has been measured by several experimental groups over the years, from a few keV to about 20 MeV incident energies. We have performed a Bayesian statistical analysis of this cross section and the result is shown in Fig. 4. Because of the somehow smaller uncertainties attached to the Browne (1984) and Fursov (1994) data, these two sets tend to drive most of the evaluated cross section over the considered energy range. Our result is in fair agreement with the JENDL-3.3 evaluation, though not following exactly the same shape over the entire energy range.

267

|

3

ENDF/B-VI Bowman'68 [12572] Fomushkin'81 [41173] Dabbs'83 [12808] Browne'84 [10805] Gokhberg'91 [41108] Fursov'94 [41303] present analysis JENDL-3.3 Younes'04 (surrogate)

c o

E

<

10

Neutron energy (MeV) Figure 4. Neutron-induced fission cross section of 2 4 2 m A m obtained with the GLUCS code, a Bayesian statistical analysis code. It is compared to available experimental data and the ENDF/B-VI and JENDL-3.3 evaluations. The determination of this cross section by Younes from a surrogate reaction is also shown.

3.4.

243

Am

Americium-243 is the most stable of all americium isotopes (half-life of 7,370 yr). The existing ENDF/B-VI evaluated file is of much better quality than the files for the other americium isotopes, thanks to the more recent work of Young 18 . However, several points of this evaluation were revisited in view of new experimental data, and because of systematics learnt from the study of the other americium isotopes. We will only mention here the question of the (n, / ) cross section. Figure 5 displays all known experimental data sets, along with evaluated files. The most striking feature of this plot is the separation of the fission cross section into two distinct groups in the 1-6 MeV range. The most recent experimental data set by Laptev et al.19 is a ratio measurement to 235 U (n,/) cross section, and extends the energy range studied up to 200 MeV. It lies in the higher cross section group following data by Behrens and Browne 20 , therefore higher than the existing ENDF/B-VI values that follow the data by Knitter and Budtz-j0rgensen 21 . However, the two groups seem to only differ by a normalization factor in this energy range. As noted by the authors of Ref.21, averaged cross sections calculated with the ZEBRA reactor spectrum 22 are in agreement with experimental data if the Knitter

268

Kanda, 1987 Fomushkin, 1984 Fomushkin, 1967 (a' Fomushkin, 1967 (b; Adamov, 1983 Knitter, 1988 Butler, 1961 Fursov, 1985 Seeger, 1970 Ya, 1999 Behrens, 1981 Goverdovsky, 1985 Laptev, 2004 Younes, 2004 (surrogate) ENDF/B-VI JENDL-3.3

1.5

0.5

-0.5

0.01

0.1

1

10

Neutron Energy (MeV)

Figure 5.

Neutron-induced fission cross section of

243

Am.

data are used, and about 12% too high if Behrens and Browne data are used instead. The ZEBRA result is most sensitive to the energy region just above the rapid rise of the cross section around 1 MeV. Although the uncertainties reported for the ZEBRA data are of the order of 20%, it tends to favor the lower-cross section group normalization.

4. Conclusion We have re-evaluated several neutron-induced reactions on americium isotopes that are important for emerging nuclear applications such as the Advanced Fuel Cycle Initiative in the United-States and the development of Generation-IV nuclear reactors worldwide. The quality of current ENDF/B-VI americium evaluations is rather poor due to a historic lack of interest for MA. The present work led to the creation of new ENDFformatted files that will be proposed for inclusion in the US ENDF/B-VII library, due by December 2005. Several further improvements will be pursued in the near future, in particular regarding prompt neutrons multiplicities and spectra that have not been revised in the present work. Also, integral data for radiochemical analyses will be used to constrain several evaluated differential reaction cross sections.

269 Acknowledgments T h e authors would like t o acknowledge t h e support of t h e Advanced Fuel Cycle Initiative (AFCI) program at Los Alamos National Laboratory.

References 1. G. Aliberti, G. Palmiotti, M. Salvatores and C.G. Stenberg, Nucl. Sci. Eng. 146, 13 (2004). 2. P.G. Young, E.D. Arthur and M.B. Chadwick, Proc. IAEA Workshop on Nuclear Reaction Data and Nuclear Reactors: Physics, Design and Safety, Trieste, Italy, April 15-May 17, 1996, Ed. A. Gandini and G. Reffo (1998). 3. D.M. Hetrick and C.Y. Fu, ORNL report TM-7341, ENDF-303 (1980). 4. J. Raynal, CEA report CEA-N-2772, Saclay, France (1994). 5. P.G. Young, LANL reports LA-UR-94-3104 (1994); LA-UR-95-3654 (1995). 6. A.V. Ignatyuk, G.N. Smirenkin, and A.S. Tishin, Sov. J. Nucl. Phys. 2 1 , 255 (1975). 7. J. Kopecky and M. Uhl, Phys. Rev. C41, 1941 (1990). 8. T. Kawano, private communication (2004). 9. G. Vanpraet, E. Cornelis, S. Raman, and G. Rohr, Proc. Int. Conf. Nuclear Data for Basic and Applied Science, 493 (1986). 10. K. Wisshak and F.Kappeler, Nucl. Sci. Eng. 76, 148 (1980). 11. D.B. Gayther and B.W. Thomas, Proc. 4th All Union Conf. on Neutron Physics, 18-22 April 1977, Kiev, USSR, 3, 3 (1977). 12. T. Kawano, P. Talou, M.B. Chadwick, R.E. MacFarlane and P.G. Young, LANL report LA-UR-03-7109 (2003). 13. A.G. Dovbenko, V.I. Ivanov, V.E. Kolesov, and V.A. Tosltikov, INDC(CCP)9, 7 (1970). 14. K. Okumura, private communication (2003). 15. S. Ohki, K. Yokoyama, K. Numata and T. Jin, Proc. 2003 Symposium on Nuclear Data, Nov. 27-28, 2003, JAERI, Tokai, Japan, 40 (2004). 16. G. Fioni et al, Nucl. Phys. A693, 546 (2001). 17. W. Younes, H.C. Britt, and J.A. Becker, LLNL YCRL-TR-201913 (2004). 18. P.G. Young, LANL report (1996). 19. A.B. Laptev et al., Nucl. Phys. A734, E45 (2004). 20. J.W. Behrens and J.C. Browne, Nucl. Sci. Eng. 77, 444 (1981). 21. H.H. Knitter and C. Budtz-j0rgensen, Nucl. Sci. Eng. 99, l (1988). 22. D.W. Sweet, technical report AEEW-R-1090 (1977).

USING IMPROVED ENDF-BASED NUCLEAR DATA FOR CANDU REACTOR CALCULATIONS IOSIFPRODEA Institute for Nuclear Research Pitesti, Campuluil, 115400, Mioveni, Romania

1. Introduction Maintaining a complete and well-documented nuclear data library is still a priority of the R&D activities in the Institute for Nuclear Research (INR) Pitesti. As the ENDF (Evaluated Nuclear Data Files) format is widely used all over the world the last years work was dedicated mainly to the improvement of the existing "WIMS-MASTER" library (in use in INR) by the new ENDF data. We used the ENDF/B-6 files"1, as they seem to be more complete (they include data for higher energies, detailed distributions of emerging particles and also data about charged particles interactions). Computer codes NJOY99 and WTLIT were used for data processing i.e. "translating" them in suitable formats for other codes such as WIMS, MCNP, ANISN etc. WILIT code121 developed by A. Holubar is available in INR since 1993, via NEA Data Bank and NJOY99 code[3] were supplied by DOE-ORNL as part of the packet of computer codes, also including the DRAGON 3.2, DOORS 3.2 and DANTSYS 3.0 codes that succesfully adapted on PC in INR. In order to verify (validate) the nuclear data generated by NJOY99 using the ENDF/B-6 files, cell calculations were performed with the WIMS-D4[4] computer code for a typical CANDU cell. The update procedure was: the interesting data are extracted from ENDF/B6 files and converted into WIMS format. Converted data are entered into WIMS Library via WILIT code, the new isotope geting new identification number chosen by the user. Thus, the folowing table was obtained:

270

271 Table 1. Updated isotopes in WIMS-MASTER Library Identifier in Identifier in WIMS- Temperatures Current no. Name of the isotope E N D p / B . 6 MASTER (K) 1

U-238

9237

8888.4

2

U-235

9228

5555.4

900

3

Mo Natural

4200

4200

300, 600, 900

" " "

900

4

Mo92

4225

4292

5

Mo94

4231

4294

6

Mo95

4234

4295

7

Mo96

4237

4296

••

8

Mo97

4240

4297

9

Mo98

4243

4298

10

MolOO

4249

4210

" " "

11

H-l

2,125

3001

296 350, 400

12

D-2

11, 128

3002

"

2. Uranium isotopes (235 and 238) We represented those parameters that had a "visible" evolution with respect to cell burnup. Figure 1 shows the differences (in mk) in cell reactivity induced by using corresponding Uranium isotopes from ENDFB-VI and WIMS-MASTER library, where: delta (mk) = (l/keff.w -l/keff_E)*1000. (1) The indexes "W" (from WIMS-MASTER) and "E" (from ENDF/B-6) show the isotope source.

""

( * * A

*

A A

A

V

1

A ^

If

f

J A

^ ^

^

-o—dif(mK)U-235E

|

AA

A

-^Booo^

tfMKSSoOC* t^O^O-O—<

dif(mK)U-238E

-V^O-TOTTO

if •k

0

4000

8000

12000

16000

B (MWd/t) Fig. 1. Differences (in mk) between corresponding Uranium isotopes

272

The "circle" line representing the U235 differences (between WIMS-MASTER and ENDF/B-6 data) is closed to zero line, which means that the both sets of Uranium 235 data have the same source. U238 differences ("triangle" line) aren't negligible and they grow with respect to burnup. This behaviour of ENDF/B-6 isotope vs. its correspondent one (from WIMS-MASTER library) seems to be caused by the resonance U238 absorption, i.e. the energy range from 10 keV to 4eV (groups 15-27). The Figure 2.a) and b) confirm these significant differences. Res. Micros. Abs. sections (WIMS calc.) D U238W • U238E

Res. Micros. Abs. Sections • U238W (library values) • U238E 160.0 140.0 120.0 ' 100.0 80.0 60.0 40.0 20.0 0.0

I

aAJj,tl,H.H,ftf|.

.-.-."•nil

15 16 17 18 19 20 21 22 23 24 25 26 27

15 16 17 18 19 20 21 22 23 24 25 26 27

WIMS Group #

WIMS Group #

Fig. 2. a) U238 - Microscopic resonance abs. cross-sections (WIMS calculations)

Fig. 2. b) U238- Microscopic resonance absorption cross-sections (library values)

Figure 3 shows how the microscopic cross-sections behaviour is reproduced at macroscopic level, i. e. the ENDF/B-6 U238 isotope is more absorbent than its correspondent in WIMS-MASTER), but the differences didn't get over 2.2%. We used the formula to calculate relative differences (in %): ENDF-6 / yWIMS-MASTER

dif (%) = (!'

1 \ . 1 (V)

(2)

Fast and thermal cell absorption differences (%)

2.50 2.00

^flft^flrtftftrtft ^fr=

- • - U235-thermal

g1.50

- A - U238-thermal

S

-»- U235-fast

LOO

- o - U238-fast 0.50 t

0.00 5000

10000

15000

B (MWdAU)

Fig. 3. Cell macroscopic absorption differences (%) for U235 and U238

273

For the others mentioned parameters, the differences were also small enough, therefore they weren't analyzed. 3. Hydrogen (H) and Deuterium (D) The next step in our demarche was to obtain nuclear data for Hydrogen and Deuterium, the two of most important isotopes in CANDU coolant and moderator. Regarding Hydrogen we used nuclear data evaluated at LANL151 (Los Alamos National Laboratory). For Deuterium we used nuclear data evaluated at LANL and General Atomic. Isotopes #2 and #11 represent Hydrogen, respectively Deuterium bound in water and their source is the IAEA WIMS Library'61 which is based on ENDF/B-6 data files and it is part of the WLUP (WIMS Library Update Project). Again, WIMS calculations were performed to verify these new H and D nuclear data. Cell k-eff (WIMS-D4.1 calculations)

Cell ^ i ^ l i a b s o r p t i Q n (WIMS-D4.1 calculations) 4.E-03

i«WjMSC4

-»-H2_njoy.d4

V

***, 4000

8000

12000

16000

B (MWd/tU)

Fig. 4. K-eff evolution

- * - H2_iaea.d4

E 4.603

- * - H2_njoy.d5

gj 4E-B3: :,:g'5E-03s

—

H2jaea,d5

-HiitiSA

5 *i-oi

::;s

.

1

riSifrREA"''' ''

4E-03I 00

S000 : )2O00flB00p

B(MW#tU) " : *;j,-' s F;

Fig. 5. Thermal absorption evolution

On Fig. 4 it can be seen k-eff evolution for two sets of Deuterium data using two versions of WIMS (WIMS-D4 indexed by ".d4" and WIMS-D5B indexed by ",d5"). On Fig. 5 we represented cell thermal absorbtion behaviour. "Ref.WIMSD4" represents the reference case, i.e. using existing H-2001 and D-6002 from WIMS-MASTER library. As seen on both upper figures the differences are negligible. Remember that the majority of cell parameters were compared (using an automatique procedure written in Visual Basic), especially: k-infinity, k-effectiv, isotope and material cell reaction rates, power peaking factor, cell average fast, thermal and total neutron flux, isotope concentration, the ratio of fast flux to total flux, isotope fission rates. Because the results of this comparisons are very close howsoever which H&D isotopes were used, we didn't analyze them. At the end of this step, for completeness, we performed some core calculations using the tridimensional diffusion code DIREN_MG[7] developed in our institute, that uses finite differences to solve time-independent diffusion

274

equation on 2 or 7 energy groups. We used the burnup simulation applied to the core model from of Cernavoda NPP, Unit 1 Commissioning181. Cross-sections for fuel and incremental cross-sections for reactivity devices were calculated with WIMS_D4 code (7 energy groups collapsation scheme was used), then were introduced in the DIREN_MG input. We considered the Deuterium isotope to be representative for CANDU reactors because the both moderator and coolant contain high purity heavy water (D 2 0). Also, we considered flux and channel power maps to be relevant and we firstly represented thermal fluxes in the two measurement channel (HFDl-orizontal and VP2-vertical). Consequently, we only used different data sets for Deuterium. The global influence of different set for Deuterium (6002 from WIMS-MASTER and 3002 from ENDF/B-6) is shown on Figures 6 and 7. As it can be seen the flux curves are, practically, identical. Therefore, we only represented the relative differences which are quite small (-0.4, +1.4)%, as seen. Horizontal Flux (DIREN_MG estimations along the HFD1 channel) 3.5E+07

Vertical Flux (DIREN_MG estimations)

T

J0^^JHs^\^f%tk

3.0E+07 ;T2.5E+07

Vi E

-<-D2_WMS-MASTER

'"'1.5E+07 E

1.0E+07

-»-D2_ENDF/B-6

50E+06

- * - dif(%) t

j

(

,

\\

\\ ,. ......

'

7*^... 300

400

500

600

70O

BOO

Y (cm)

Fig. 6. Thermal flux shape in HFDlchannel Channel Power Distribution (KW) - r>3002 {WIMS-MASTER)

Fig.7. Thermal flux shape in VP2 channel Rel. Channel Power Differences (%) D2 (ENDF/B-6) vs. D2 (WIMS-MASTBR)

-7000 -6000 -6000 -4000 -3000 -2000 1000

"""^ffillj V f f IflH^77p "'^7^

R T v

Fig. 8. Channel Power distribution - D3002

Fig. 9. Relative Channel Power Differences

In Figure 8 one of the channel power map for D-3002 (ENDF/B-6 files) transfered via NJOY99 code was represented. The corresponding map for D6002 (WIMS-MASTER) wasn't represented because it is very similar. Nearby

275

the "tridimensional" map of relative diferences (%) can be can be seen, which are also very small (under 1.5%). It is worthy to be mentioned that the differences 3-D shape is similar to that of the channel powers. 4.

Molibdenum

The CANDU reactor adjuster rods control system consists of 21 stainless steel "rod-in-tube" devices and has a total reactivity worth of 15 mK. Some of the tubes contain more than 2% Molybdenum, so we need reliable nuclear data for this element, in order to properly evaluate the device reactivity worth. The WIMS-MASTER library only contains 2 isotopes for Mo (a fission product and also an unknown one), so we used the ENDF/B-6 files and the NJOY99 code to produce a complete set of nuclear data for this element. Therefore, we introduced in WIMS-MASTER library (via NJOY code) all natural isotopes of Molybdenum[9]. To verify the nuclear data for Mo we firstly performed a WIMS reactivity estimation for a B-TYPE ADJUSTER ROD. We used one reference and 3 perturbed cases as below: • Reference case (WIMS cell unperturbed, without ADJ rod), • Mo natural (4200, ENDF/B-6 data), • Compounded Molybdenum (considering abundance ratio), • Mo 1095 (old, existing in WIMS-MASTER). The following tables show the results of comparisons. Table 2. Multiplication factors and cell ADJ rod reactivity rho- inf rho - eff K-inf K-eff Case (mk) (mk) . Reference 1.05279 1.02094 Mo Natural

0.97035

0.94275

-80.704

-81.24

dif w (%)

difeir

-

Mo Comp

0.97037

0.94278

-80.674

-81.21

-0.04

-0.04

Mo_1095

0.97040

0.94280

-80.648

-81.18

-0.07

-0.07

rho (mk) = (l/kref-l/kpert)*1000 (ADJ supercell reactivity) difiafsdifeff (%) = (rho pert / rho NatMo )*100 (rel. diff. vs. Mo Natural) Table 3. Absorption cross sections SABSl Case (cm 1 ) 1.6905E-03 Ref

SABS 2 (cm"') 4.0487E-03

dlfi (%)

dif.

(%) -

Mo Natural

1.7079E-03

4.4909E-03

-

Mo Comp

1.7075E-03

4.4910E-03

-0.023

0.002

Mo 1095

1.7074E-03

4.4908E-03

-0.029

-0.002

(3) (4)

276

dif,, dif2 (%) = (SABS,** /

SABSMO Nat

- 1)*100

(5)

It is not necesary to present all computed parameters (diffusion coefficients, scattering, moderation, yield cross-sections, cell fluxes) because the differences are negligible. So, we can affirm that the WIMS material "1095" represents natural Molybdenum. For completness, we performed 3D core analysis by estimating the reactivity worth of the Adjuster Rod System (ADJ) for the Cernavoda NPP, Unit 1. We used the "recipe" WIMS-PIJXYZ-DIREN_MG. The code PIJXYZ1101, was developed in INR for realistic 3D treatement of reactivity devices in a CANDU supecell and it is based on first collision probability. Now, the references were represented by individual reactivity worths of ADJ rods "measured" at Cernavoda NPP Unit 1 Commissioning. The word "measured" indicates that the reactivity worth was estimated by indirect measurement, i. e. the light water level variation (AVZL%) in the Zone Control Units was measured between rod insertion and rod removal moments. Using a known rho (AVZL) curve the reactivity worth was estimated. The numerical methods used in the 3 upper codes are: WIMS-1D transport (Sn), PIJXYZ-3D transport (1 st collision probabilities), DIREN_MG- 3D diffusion. The differences "calculation vs. measured" are situated in the range (-11.57, +11.50) mk. To appreciate how "important" these differences are we compared them to the corresponding range values as mentioned at other CANDU NPP Commissioning. We found out at Point Lepreau Commissioning'111 the range upper mentioned was (-23.69, -10.72) mk, while at Embalse NPP Commissionigf121 the values were (-17.98, -5.27) mk. According to these facts, we consider that our differences are normal and better centered. 5. Conclusions CI. The compatibility of the format ENDF-6 and WIMS was preliminary verified by NJOY99-WILIT transfer "recipe". CI. The incertainities about existing Molybdenum isotopes in the WIMSMASTER library were eliminated. C2. Using the NJOY99-WILIT "recipe" the INR's WDVIS-MASTER library was updated by newest ENDF/B-6 nuclear data for some of the most important CANDU isotopes: Hydrogen, Deuterium, Uranium and Molybdenum. C3. We estimate that the NJOY99 - WILIT - WIMS code system may be used in order to assess any nuclear data, including those that are to be evaluated for the Fourth Generation Systems. C4. Although Romania has only one NPP in operation and another under construction (both of CANDU-6 type), we hope that the future integration in

277

European Union will bring a larger opening to advanced solutions of nuclear reactors. References 1. H.D. Lemmel, P.K. McLaughlin, V.G. Pronyaev, ENDF/B-VI, Release 7, The U.S. Evaluated Nuclear Data Library for Neutron Reaction Data, IAEA-NDS-100, Rev. 10, Jun. 2000 2. A. Holubar, WILIT - A utility program for WIMS Library, NEA DATA BANK, Rez.1989 (updated 1991), received 15.07.1992 3. "Code System for Producing, Pointwise and Multigroup Neutron and Photon Cross Sections from ENDFB Data", RSICC PeripheraShielding Routine Collection, PSR-480 NJOY99.0 4. WIMS-D4 -Winfrith Improved Multigroup Scheme Code System - RSICC Computer Code Collection, CCC-576,ORNL-DOE, Contributed by: Atomic Energy Establishment, Winfrith, Dorchester through the NEA Data Bank, Gif-sur-Yvette Cedex, France, Dec. 1990, Rev.Oct. 1991 5. ENDF Database http://www.nndc.bnl.gov/endf/index.html 6. Cross Section Evaluation Working Group,"ENDF-201 ENDF/B-VI Summary Documentation, Supplement. 1", BNL-NCS-17541,4th Edition, Suppl. 1, Dec. 1996 7. I. Patrulescu et al., "DIREN code development for multigroup core calculations", RI5120, SCN 1997 8. I. Dumitrache, A. Rizoiu, Benchmark problem for local unperturbed CANDU-6 cell calculations, Unit 1, Cernavoda NPP, RI 5933, INR Pitesti, 2000 9. *** "Isotopes Table", Knolls Atomic Power Laboratory, 1996 10. M. Constantin et al., Performance's Evaluation of 3D code PIJXYZ, Institute for Nuclear Research, RI- 4659,1995 11. O. A. Trojan et al., Phase B Physics tests at Point Lepreau NGS, Annual Meeting of Korean Nucl. Soc, 1982. 12. J. C. Vinez, Physical Measurements: Calibration ofLZC, ADJ, SOR, MCA and Banks, IAEA Training Course, LR-56, 17-28 Oct. 1988

A COMPARATIVE STUDY ON THE GRAPHITEMODERATED REACTORS USING DIFFERENT EVALUATED NUCLEAR DATA DO HEON KIM*, C.-S. GIL, Y.-S. CHO, Y.-O. LEE, J. CHANG Korea Atomic Energy Research Institute, P.O. Box 105, Yuseong Daejeon, 305-600, Korea The benchmark calculations for several graphite-moderated reactors have been performed by the Monte Carlo code MCNP4C using the libraries based on the ENDF/BVI.8, JENDL-3.3 and JEFF-3.0. The calculation results for each library have been compared with the experimental values. The isotopic contributions to the keff calculation have been estimated to clarify the causes of the differences among the three libraries.

1. Introduction R&D projects for a very high temperature reactor (VHTR) are being carried out in connection with a hydrogen production as an alternative energy to fossil fuels. Also the innovative gas-cooled fast reactor (GFR) concepts are being actively considered in the Generation IV program. The VHTR and GFR contain relatively large quantities of graphite and/or silicon when compared to the existing light water reactors (LWRs) and heavy water reactors (HWRs). In this study, several graphite-moderated reactors were selected to perform the MCNP4C criticality benchmark calculations. The isotopic contributions to the keff calculation have been estimated through sensitivity calculations for significant isotopes such as graphite, silicon, copper, zirconium, and uranium. The impact of the different thermal scattering law data of graphite has also been investigated for the criticality calculations. 2. Benchmark Problems Several graphite-moderated reactors were selected for the criticality benchmark calculations. The reactors employed are; core 5 (reference state #3) of the HTRPROTEUS within the framework of an International Atomic Energy Agency (IAEA) co-ordinated research program (CRP) [1] and 8 critical benchmarks

f

E-mail: [email protected]

278

279

taken from the International Criticality Safety Benchmark Evaluation Project (ICSBEP) [2]. The HTR experiments with a low-enriched uranium (LEU) fuel were carried out at the PROTEUS facility of the Paul Scherrer Institute (PSI), Switzerland. The experimental data was obtained from 13 pebble-bed arrangements with different moderation ratios and pebble packing geometries. The core 5 (reference state #3) has a deterministic packing arrangement loaded with LEU pebble-type fuels with 16.7% enriched 235U. The ZEUS experiments are a series of on-going critical assemblies at the Los Alamos Critical Experiments Facility (LACEF) at Los Alamos National Laboratory (LANL). The experiments were designed to test the adequacy of the 235 U cross sections in the intermediate energy range. The plates of highly enriched uranium (HEU) metal were interspersed with graphite plates in a cylindrical stack which was completely surrounded by copper reflectors. The critical configuration 1 was selected for the benchmark calculations. The ZEBRA experiments are a series of k^ experiments in fast/intermediate neutron spectra performed at the ZEBRA zero-power fast reactor at Winfrith, United Kingdom. The 8F/2 assembly comprised of a combination of mixed plutonium and uranium oxide, natural uranium oxide, and graphite plates. The HUG (Homogeneous-Uranium-Graphite) experiment is a series of small-sample k^ experiments in intermediate neutron spectra performed at the HECTOR (Hot Enriched Carbon-moderated Thermal Oscillator Reactor) zeropower reactor at Winfrith, United Kingdom. The experiment known as the HISS (HECTOR Intermediate Spectrum Study) was designed to provide a sensitive test of the 235U and 239Pu cross sections over the energy range of 10 eV to 10 keV. The RBMK reactor is a uranium-graphite reactor from Russia. A wide range of neutron experiments were performed at the Russian Research Center "Kurchatov Institute" during the period of the RBMK design development and improvement. The results of 28 critical experiments were surveyed for the ICSBEP. In this study, three uniform configurations (2, 4 and 6) without water in the fuel channels for each of the three enrichments (1.8%, 2%, and 2.4%) were selected. The ROVER program comprised of critical experiments of hexagonal graphite rods containing HEU moderated and reflected by water which were performed at Oak Ridge National Laboratory (ORNL). The experiments were performed with Westinghouse type NRX-A3 and NRX-A4 fuel elements to provide data for the criticality safety analyses of the ROVER fuel element

280

production, storage, and transportation. Two critical configurations (1 and 11) with square-pitched and triangular-pitched lattices were chosen. [2] The ZEUS, ZEBRA and HUG are the "Intermediate" (0.625 eV < Energy < 100 keV) systems and the others are the "Thermal" (Energy < 0.625 eV) systems, as classified by the energy of the majority of neutrons causing a fission. Figure 1 shows the neutron spectra of the selected benchmark problems. For the RBMK, ROVER and HTR-PROTEUS, the neutron spectra are similar to the typical thermal spectra of a pressurized water reactor (PWR), while a hardening of the spectra is apparent in the epithermal and fast regions. 10° 10"'

iff 2

_ CD

a

iff*

-J

10^

D

10""

u. 10"' 10"*

10-' 10'° 10"

10"*

10'7

10""

Iff"

Iff 4

Iff 3

10*

10''

10°

10'

Neutron Energy (MeV) Figure 1. Neutron spectra of selected benchmark problems.

3. Benchmark Calculations and Results The MCNP4C libraries based on the ENDF/B-VI.8, JENDL-3.3 and JEFF-3.0 have been generated through a NJOY99.90 code [3] processing. The benchmark calculations for the selected benchmark problems have been performed by the MCNP4C code using the libraries. 3.1. Comparisons with Benchmark Experiments The benchmark calculation results have been compared with the experimental values, as shown in Figure 2. The ENDF/B-VI.8 results generally agree well with the JEFF-3.0 results except for the ZEBRA, in which the capture/fission reactions of 238U are predominant. The JENDL-3.3 results tend to decrease for the keff except for the ZEUS and ROVER when compared with the other

281

libraries. A maximum underestimation of 6.7 mk was observed for the JENDL3.3 results of the HTR-PROTEUS. 12 9

ST

6

g

"5 OJ

o c

0

s £ -3

5 -6 -9

5* > m

a.

CO

5

m
X DC Z

$

DC 7

co HI I-

O cc GL

Figure 2. Comparisons of calculated keff with experiments.

3.2. Isotopic Contributions to keff The isotopic contributions to the keff, which could be calculated by substituting the JENDL-3.3 and JEFF-3.0 data for the ENDF/B-VI.8-based benchmark calculations, were estimated to clarify the causes of the differences among the three libraries. For the ZEUS and HUG, the reaction rates of 235U are predominant and larger discrepancies of the reaction rates between ENDF/B-VI.8 and JENDL-3.3 are observed. The 235U data of JENDL-3.3 results in a decrease of the k^f for these problems. However a relatively large increase of keff was observed for the ZEUS due to the large amounts of copper used as a reflector material. As a whole, the ZEUS has a tendency to increase the keff when using the JENDL-3.3. For the ZEBRA, the reaction rates of 238U are predominant and the reaction rates from the ENDF/B-VI.8 show large differences with those from the JENDL3.3 and JEFF-3.0. The 238U data of JENDL-3.3 and JEFF-3.0 result in a decrease of the keff for the benchmark problem. For the RBMK and HTR-PROTEUS, the reaction rates of 235U are predominant, while the capture reaction rates of graphite are relatively large when compared to other benchmark problems. The JENDL-3.3 results have a tendency to decrease the keff, which are caused by the increasing capture reaction rates of the graphite. In addition, the contribution of the graphite data of JENDL-

282

3.3 becomes larger for the HTR-PROTEUS. Figure 3 shows the comparisons of the capture reaction rates of graphite in the moderator pebbles of the HTRPROTEUS core.

10"" 10"'° 10""

10"'

10"'

10 J

10"*

10"'

10"'

10' !

10"'

10"

10'

Energy (MeV)

Figure 3. Comparisons of capture reaction rates of graphite.

Figure 4 shows the impact of the different thermal scattering law data of graphite for the criticality calculations. The calculated keff by the TMCCS library from the MCNP4C code package shows small discrepancies within 2 mk when compared with those by the libraries processed with the thermal evaluations for the ENDF/B-VI.3 and JEFF-3.0, respectively.

Figure 4. Impact of different thermal scattering law data of graphite.

283

4. Summary The benchmark calculations for 9 graphite-moderated reactors have been performed by the MCNP4C code with the libraries based on the ENDF/B-VI.8, JENDL-3.3 and JEFF-3.0. The ENDF/B-VI.8 results show good agreements with JEFF-3.0 results except for the ZEBRA. The JENDL-3.3 results have a tendency to decrease the keff except for the ZEUS and ROVER. For the intermediate systems, • the 235U of JENDL-3.3 results in a decrease of the keff for the ZEUS and HUG, • the copper of JENDL-3.3 results in an increase of the k,.ff for the ZEUS, and • the 238U of JENDL-3.3 and JEFF-3.0 results in a decrease of the keff for the ZEBRA. For the thermal systems such as the RBMK and HTR-PROTEUS, the graphite of JENDL-3.3 results in a decrease of the keffAcknowledgments This project has been carried out under the Nuclear Research and Development program by Korea Ministry of Science and Technology. References 1. T. Williams, "LEU-HTR PROTEUS: Configuration Descriptions and Critical Balances for the Cores of the HTR-PROTEUS Experimental Programme," TM-41-95-18, Paul Scherrer Institut, Switzerland (1996). 2. NEA Nuclear Science Committee, "International Handbook of Evaluated Criticality Safety Benchmark Experiments," NEA/NSC/DOC(95)03, Nuclear Energy Agency (2004). 3. R.E. MacFarlane and D.W. Muir, "The NJOY Nuclear Data Processing System, Version 91," LA-12740-M, Los Alamos National Laboratory (1994).

This page is intentionally left blank

AUTHOR INDEX

A Ahmad, I. Aiche, M. Aliberti, G. Arredondo-Sanchez, C. Aryaeinejad, R. Avrigeanu, M. Avrigeanu, V. B Barreau, G. Bauge, E. Berthoumieux, E. Bidaud, A. Billebaud, A. Bosq, J.C. Bouland, 0. Boyer, S. Briggs, J. B. Brun-Magaud, V. C Capote Noy, R. Carlsson, J. Chadwick, M. B. Chang, J. Cho, Y.-S. Cole, J.D. Conti, A. Courcelle, A. Czajkowski, S. D Dassie, D. Drigert, M.W.

Duijvestijn, M. Dupont, E. Dyck, G.R. E Eichin, R. Fernandez, A. F Finck, P. Forrest, R. A. Frankle, S. C. Freiesleben, H. Fujii, T. Furutaka, K. G Gardnier, J. C. Gil, C.-S. Greene, J. P. Grosjean, C. Guiral, A. Gunsing, F. H Haas, B. Haas, D. Hanson, K. M. Harada, H. Henriksson, H. Hogenbirk, A. Hori, J.-I. I Ivanov, E. Ivanova, T.

222 222 81 72 195 145 145 222 222 222 101 222 46,81 253 222 113 46 244 183 138,262 58, 278 58, 278 195 46 253 222 222 195 285

286

J Janssens, R.V. F. Jujii, T. Jurado, B. K Kawano, T. Khalil, H. S. Kim, D. H. Kim, K.-S. Kim, T. K. Kodeli, I. Koning, A. Kozier, K.S. L Lecarpentier, D. Lee, Y.-O. Little, R. C. M Maschek, W. Mastrangelo, V. Michel-Sendis, F. Micklich, B.J. Mori, M. Mourogov, A. N Nakamura, S. Nichols, A. L. Nigg, D. W. Noguere, G. O Orlov, V. Osmanov, B. P Palmiotti, G. Perrot, L. Petit, M.

195, 222 216 222 138, 262 8 58,278 58 81 81,101 153 163 173 58, 278 138 38 101 222 195 38 173 216 244 113, 195 253 173 222 81 222 222

Pitcher, E.J. Plompen, A. Prodea, I. R Rimpault, G. Rineiski, A. Roman, F. L. Rugama, Y. S Sakane, H. Salvatores, M. Sartori, E. Seidel, K. Serot, O. Sinitsa, V. Smirnov, V. Somers, J. Sublet, J.C. T Taiwo, T. A. Talou, P. Ter-Akopian, G. Theisen, C. Tommasi, J. Trkov, A. Tucek, K. V Van Der Marck, S. C. W Weaver, K. D. Wider, H. Wilson, J. N. Y Yamana, H. Young, P.G.

262 208 270 18,46 38 145 235 216 81 81,113 145 253 38 173 64 253 8,81 138,262 195 222 46,81 244 183 32 70 183 222 216 262

Nuclear Data Needs For Generation IV Nuclear Energy Systems

The Generation IV International Forum (GIF) identified six advanced concepts for sustainable nuclear energy production at competitive prices and with advanced safety, with special attention to nuclear nonproliferation and physical protection issues, minimisation of long-lived radiotoxic waste, and optimum natural resource utilisation. System groups have been established for studying in detail these concepts, and nuclear data are an inherent part of these studies. The workshop was attended by 70 participants from 20 countries. During three consecutive days recent achievements were presents on sensitivity analysis, model calculations, estimates of uncertainties, and the present status of nuclear databases. Special attention was given to the identification of nuclear data needs from sensitivity analysis of benchmark experiments and the treatment of uncertainties. Experimental programmes and recent results of interest for the development of Generation IV systems were presented. Representatives from the major nuclear data centres discussed the actual status of the evaluated data libraries.