Liquid Filtration

LIQUID FILTRATION by Nicholas P. Cheremisinoff, Ph.D.

Environmental Policy and Technology Project United States Agency for International Development

Boston Oxford Johannesburg Melbourne New Delhi Singapore

Copyright © 1998 by Butterworth-Heinemann A member of the Reed Elsevier group All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Recognizing the importance of preserving what has been written, Butterworth-Heinemann prints its books on acid-free paper whenever possible. Butterworth-Heinemann supports the efforts of American Forests and the Global ReLeaf program in its campaign for the betterment of trees, forests, and our environment. ISBN: 0-7506-7047-9 The publisher offers special discounts on bulk orders of this book. For information, please contact: Manager of Special Sales Butterworth-Heinemann 225 Wildwood Avenue Woburn,MA01801-2041 Tel: 781-904-2500 Fax: 781-904-2620 For information on all Butterworth-Heinemann publications available, contact our World Wide Web home page at: http://www.bh.com 10 9 8 7 6 5 4 3 2 1 Printed in the United States of America

DEDICATION This volume is dedicated to the memory of Paul N. Cheremisinoff, M.S., P.E., who authored more than 300 technical books over his career as a chemical engineer and was among the pioneers of pollution control and prevention.

This page intentionally left blank

CONTENTS Preface

ix

About the Author

xi

Chapter 1 An Introduction to Liquid Filtration

1

Introduction The Porous Media The Filter Media Liquid Filtration Classification The Formation of Filter Cake Typical Industrial Filtration Conditions Washing and Dewatering Operations General Considerations for Process Engineers The Objectives of Filtration Preparation Stages for Filtration Equipment Selection Methodology Nomenclature Chapter 2

Filter Media and Use of Filter Aids

Introduction Flexible Filter Media Rigid Filter Media Filter Media Selection Criteria Introduction to the Use of Filter Aids Examples of Filter Aids Filter Aid Selection Suggested Readings Nomenclature Chapter 3

Cake Filtration and Filter Media Filtration

Introduction Dynamics of Cake Filtration Constant-Rate Filtration

1 2 9 10 11 12 12 13 14 15 16 18 19 19 20 34 43 47 50 51 57 58 59 59 60 70

Contents

Variable-Rate and -Pressure Filtration Constant-Pressure and -Rate Filtration Filter-Medium Filtration Formulas Constant-Pressure-Drop Filtration Filtration Mechanisms Constant Rate Filtration Suggested Readings Nomenclature

Chapter 4

Industrial Filtration Equipment

Introduction Rotary Drum Filters Cocurrent Filters Cross Mode Filters Cartridge Filters Diaphragm Filters High Pressure, Thin Cake Filters Thickeners Solids Washing Centrifugal Filtration Screw Presses Ultrafiltration Reverse Osmosis Closure Chapter 5

72 75 75 75 81 83 86 87

88 88 89 91 98 103 110 115 117 120 120 123 124 134 141

Application of Filtration to Wastewater Treatment

142

Introduction Granular Media Filtration Bed Regeneration Flocculation Filtration Slow Sand Filtration Rapid Sand Filtration Chemical Mixing, Flocculation, and Solids Contact Processes Suggested Readings

142 142 148 149 151 153 155 162

Chapter 6

Advanced Membrane Technology for Wastewater Treatment

Introduction Overview of Technology Case Study Case Study Specifics Technology Application Mechanisms of Membrane Separations Treatment of Hazardous Wastes

163

Contents

Features of the Hyperfiltration System Process Economics Detailed Process and Technology Description Summary of Case Study Analytical Results Closure Chapter 7

Sludge Dewatering Operations

Introduction Overview of Dewatering Technologies Use of Drying Beds Use of Vacuum Filtration Use of Pressure Filtration Use of Centrifugation Alternative Mechanical Dewatering Techniques Suggested Readings Chapter 8

Industrial Wastewater Sources

Introduction Paper and Allied Products Industry Wastes Dairy Products Industry Wastes Textile Industry Wastes Pharmaceutical Industry Wastes Leather Tanning and Finishing Industry Wastes Petroleum Refining Industry Wastes Food and Meat Packing Industry Wastes Beverages Industry Wastes Plastics and Synthetic Materials Industry Wastes Blast Furnaces, Steel Works, and Rolling and Finishing Wastes Organic Chemicals Industry Wastes Metal Finishing Industry Wastes Closure Suggested Readings Chapter 9

Filtration Equipment and Process Flow Sheets

Introduction Index to Equipment and Flow Sheet Diagrams Index

vii

173 184 193 202 210 211 211 212 217 219 222 223 226 227 229 229 230 232 237 240 243 246 251 254 258 261 265 268 271 271 272 272 272 316

This page intentionally left blank

PREFACE

This volume has been written as an introductory reference and working guide to the subject of liquid filtration engineering. The book is designed to acquaint the newcomer to industry practices, and general design and operating methodology for filtration processes. Emphasis is given to pollution control applications, however the technologies and equipment described herein are equally applicable to product recovery and product purification applications. The information presented in this volume is based largely on the author's collected notes and lectures over the past 15 years. The volume is not intended for researches or equipment developers, but rather for process engineers, plant engineers, and technicians who require basic knowledge of this important unit operation. Much of the design methodology and working equations presented have been tested on pilot plant studies and applied to commercial and semi-commercial operations with success, however, neither the author nor publisher provide written or implied endorsements that these procedures will work in any or all cases. As with any piece of equipment or process, the designer must consult with specific vendors, suppliers and manufacturers, and further, should field test or at a minimum, conduct pilot tests to ensure performance in the intended application. Filtration equipment, operation conditions, and the use of filtration aids are highly dependent upon the properties of the suspension being filtered. Furthermore, overall process constraints and economics can have major impacts on the selection of equipment, their operating modes and characteristics, and efficiency. The author wishes to extend a heartfelt gratitude to Butterworth — Heineinann for their fine production of this volume, and to members of the United States Environmental Protection Agency for their advise and consultation on some of the materials presented herein. Nicholas P. Cheremisinoff, Ph.D.

This page intentionally left blank

ABOUT THE AUTHOR Nicholas P. Cheremisinoff is Director of the Industrial Waste Management Program in Ukraine, which is supported by the United States Agency for International Development, Washington D.C. He has nearly twenty years of applied research and industry experience in the petrochemicals, oil and gas, rubber, and steel industries, and is considered a leading authority on waste management and process design. Dr. Cheremisinoff provides technical consulting services to both private industry and government agencies and has worked extensively in Republics of the former Soviet Union, South America, Korea, the United States, and Western Europe, He is the author, co-author, or editor of over 100 engineering reference books dealing with waste technologies and process designs, including the multivolume Encyclopedia of Fluid Mechanics by Gulf Publishing Company. Dr. Cheremisinoff received his B.S., M.S. and Ph.D. degrees in chemical engineering from Clarkson College of Technology.

This page intentionally left blank

1 AN INTRODUCTION TO LIQUID FILTRATION introduction In the simplest of terms, filtration is a unit operation that is designed to separate suspended particles from a fluid media by passing the solution through a porous membrane or medium. As the fluid or suspension is forced through the voids or pores of the filter medium, the solid particles are retained on the medium's surface or, in some cases, on the walls of the pores, while the fluid, which is referred to as the filtrate, passes through. The flow of fluids through a porous medium is of interest not only to the unit operation of filtration, but to other processes, such as adsorption, chromatography, operations involving the flow of suspensions through packed columns, ion exchange, and various reactor engineering applications. In petroleum engineering applications, interest lies in the displacement of oil with gas, water and miscible solvents (including solutions of surface-active agents), and in reservoir flow problems. In hydrology, interest is in the movement of trace pollutants in water systems, the recovery of water for drinking and irrigation, and saltwater encroachment into freshwater reservoirs. In soil physics, interest lies in the movement of water, nutrients and pollutants into plants. In biophysics, the subject of flow through porous media touches upon life processes such as the flow of fluids in the lungs and the kidney. The physical parameters that relate the porous material to the hydrodynamics of flow are porosity, permeability, tortuosity and connectivity. This chapter discusses the fundamentals of flow through porous media and relates these principles to the industrial operations of filtration. As indicated in the preface of this volume, the subject of filtration is discussed from a process engineering viewpoint, and in particular from that of the chemical engineer. Filtration has a long history in the chemical engineering field both from the standpoint of the production of high purity products, as well as a technology extensively used in pollution control and prevention.

2

Liquid Filtration

The Porous Media A porous medium may be described as a solid containing many holes and tortuous passages. The number of holes or pores is sufficiently great that a volume average is needed to estimate pertinent properties. Pores that occupy a definite fraction of the bulk volume constitute a complex network of voids. The manner in which holes or pores are embedded, the extent of their interconnection, and their location, size and shape characterize the porous medium. The term porosity refers to the fraction of the medium that contains voids. When a fluid is passed over the medium, the fraction of the medium (i.e., the pores) that contributes to the flow is referred to as the effective porosity. There are many materials that can be classified as porous media, however, not all of them are of interest to the subject of filtration. In general, porous media are classified as either unconsolidated and consolidated and/or as ordered and random. Examples of unconsolidated media are sand, glass beads, catalyst pellets, column packings, soil, gravel and packing such as charcoal. Examples of consolidated media are most of the naturally occurring rocks, such as sandstones and limestones. Materials such as concrete, cement, bricks, paper and cloth are manmade consolidated media. Ordered media are regular packings of various types of materials, such as spheres, column packings and wood. Random media have no particular correlating factor. Porous media can be further categorized in terms of geometrical or structural properties as they relate to the matrix that affects flow and in terms of the flow properties that describe the matrix from the standpoint of the contained fluid. Geometrical or structural properties are best represented by average properties, from which these average structural properties are related to flow properties. A microscopic description characterizes the structure of the pores. The objective of pore-structure analysis is to provide a description that relates to the macroscopic or bulk flow properties. The major bulk properties that need to be correlated with pore description or characterization are porosity, permeability, tortuosity and connectivity. In studying different samples of the same medium, it becomes apparent that the number of pore sizes, shapes, orientations and interconnections are enormous. Due to this complexity, pore-structure description is most often a statistical distribution of apparent pore sizes. This distribution is apparent because to convert measurements to pore sizes one must resort to models that provide average or model pore sizes. A common approach to defining a characteristic pore size distribution is to model the porous medium as a bundle of straight cylindrical capillaries. The diameters of the model capillaries are defined on the basis of a convenient distribution function. Pore structure for unconsolidated media is inferred from a particle size distribution, the geometry of the particles and the packing arrangement of particles. The theory of packing is well established for symmetrical geometries such as spheres. Information on particle size, geometry and the theory of packing allows relationships between pore size distributions and particle size distributions to be established.

An Introduction to Liquid Filtration

3

A macroscopic description is based on average or bulk properties at sizes much larger than a single pore. In characterizing a porous medium macroscopically, one must deal with the scale of description. The scale used depends on the manner and size in which we wish to model the porous medium. A simplified, but sometimes accurate, approach is to assume the medium to be ideal; meaning homogeneous, uniform and isotropic. The term reservoir description is applied to characterizing a homogeneous system as opposed to heterogeneous. A reservoir description defines the reservoir at a level where a property changes sufficiently so that more than a single average must be used to model the flow. In this sense, a reservoir composed of a section of coarse gravel and a section of fine sand, where these two materials are separated and have significantly different permeabilities, is heterogeneous in nature. Defining dimensions, locating areas and establishing average properties of the gravel and sand constitutes a reservoir description, and is a satisfactory approach for reservoir-level type problems. Unfortunately, to study the mechanisms of flow, the effects of nonideal media require more specific definitions. Any discussion of flow through porous media inevitably touches upon Darcy's law which is a relationship between the volumetric flowrate of a fluid flowing linearly through a porous medium and the energy loss of the fluid in motion. Darcy's law is expressed as:

Q =

----

A/z

(1)

where A/z = Az + —— + constant P

The parameter, K, is a proportionality constant that is known as the hydraulic conductivity. The relation is usually considered valid for creeping flow where the Reynolds number, as defined for a porous medium, is less than one. The Reynolds number in open conduit flow is the ratio of inertial to viscous forces and is defined in terms of a characteristic length perpendicular to flow for the system. Using four times the hydraulic radius to replace the length perpendicular to flow and correcting the velocity with porosity yields a Reynolds number in the form: D v p

Re = __JLj£L

Darcy's law is considered valid where Re < 1.

(3)

4

Liquid Filtration

The hydraulic conductivity K depends on the properties of the fluid and on the pore structure of the medium. The hydraulic conductivity is temperature-dependent, since the properties of the fluid (density and viscosity) are temperature-dependent. Hydraulic conductivity can be written more specifically in terms of the intrinsic permeability and the properties of the fluid.

K =

(4)

where k is the intrinsic permeability of the porous medium and is a function only of the pore structure. The intrinsic permeability is not temperature-dependent. In differential form, Darcy's equation is:

Q A

J~ A

= /7 "

=

-

k dp f J j. ax

/C\ P)

The minus sign results from the definition of Ap, which is equal to p2 - p l5 a negative quantity. The term q is the seepage velocity and is equivalent to the velocity of approach vro, which is also used in the definition of the Reynolds number. Permeability is normally determined using linear flow in the incompressible or compressible form, depending on whether a liquid or gas is used as the flowing fluid. The volumetric flowrate Q (or Qm) is determined at several pressure drops. Q (or Qm) is plotted versus the average pressure p m . The slope of this line will yield the fluid conductivity K or, if the fluid density and viscosity are known, it provides the intrinsic permeability k. For gases, the fluid conductivity depends on pressure, so that

K = K\ 1+-

where b depends on the fluid and the porous medium. Under such circumstances a straight line results (as with a liquid), but it does not pass through the origin; instead it has a slope of bK and intercept K. The explanation for this phenomenon is that gases do not always stick to the walls of the porous medium. This slippage shows up as an apparent dependence of the permeability on pressure. Heterogeneity, nonuniformity and anisotropy must be defined in the volume-average sense. These terms may be defined at the level of Darcy's law in terms of permeability. Permeability is more sensitive to conductance, mixing and capillary pressure than to porosity. Heterogeneity, nonuniformity and anisotropy are defined as follows. On a macroscopic basis, they imply averaging over elemental volumes of radius e about a point in the media, where e is sufficiently large that Darcy's law can be applied for appropriate Reynolds numbers. In other words, volumes are large relative to that of

An Introduction to Liquid Filtration

5

a single pore. Further, e is the minimum radius that satisfies such a condition. If e is too large, certain nonidealities may be obscured by burying their effects far within the elemental volume. Heterogeneity, nonuniformity and anisotropy are based on the probability density distribution of permeability of random macroscopic elemental volumes selected from the medium, where the permeability is expressed by the one-dimensional form of Darcy's law. As noted earlier, the principal properties of nonideal porous media that establish the nature of the fluid flow are porosity, permeability, tortuosity and connectivity. In a macroscopic sense, porosity characterizes the effective pore volume of the medium. It is directly related to the size of the pores relative to the matrix. When porosity is substituted, the details of the structure are lost. Permeability is the conductance of the medium and has direct relevance to Darcy's law. Permeability is related to the pore size distribution, since the distribution of the sizes of entrances, exits and lengths of the pore walls constitutes the primary resistance to flow. This parameter reflects the conductance of a given pore structure. Permeability and porosity are related; if the porosity is zero the permeability is zero. Although a correlation between these two parameters may exist, permeability cannot be predicted from porosity alone, since additional parameters that contain more information about the pore structure are needed. These additional parameters are tortuosity and connectivity. Tortuosity is defined as the relative average length of a flow path (i.e., the average length of the flow paths to the length of the medium). It is a macroscopic measure of both the sinuosity of the flow path and the variation in pore size along the flow path. Both porosity and tortuosity correlate with permeability, but neither can be used alone to predict permeability. Connectivity defines the arrangement and number of pore connections. For monosize pores, connectivity is the average number of pores per junction. The term represents a macroscopic measure of the number of pores at a junction. Connectivity correlates with permeability, but cannot be used alone to predict permeability except in certain limiting cases. Difficulties in conceptual simplifications result from replacing the real porous medium with macroscopic parameters that are averages and that relate to some idealized model of the medium. Tortuosity and connectivity are different features of the pore structure and are useful to interpret macroscopic flow properties, such as permeability, capillary pressure and dispersion. Porous media is typically characterized as an ensemble of channels of various cross sections of the same length. The Navier-Stokes equations for all channels passing a cross section normal to the flow can be solved to give:

6

Liquid Filtration

Where parameter c is known as the Kozeny constant, which is essentially a shape factor that is assigned different values depending on the configuration of the capillary (c = 0.5 for a circular capillary). S is the specific surface area of the channels. For other than circular capillaries, a shape factor is included:

~> ck r" - —

(8) l '

The specific surface for cylindrical pores is:

„ _ n2nrL _ 2 ^A J7 ~ nnr L r

and

s^22 >-> A

2(j)

Wk

~

(10)

Replacing 2/81/4 with shape parameter c and SA with a specific surface, the well known Kozeny equation is obtained.

Tortuosity T is basically a correction factor applied to the Kozeny equation to account for the fact that in a real medium the pores are not straight (i.e., the length of the most probable flow path is longer than the overall length of the porous medium):

*2 • To determine the average porosity of a homogeneous but nonuniform medium, the correct mean of the distribution of porosity must be evaluated. The porosities of natural and artificial media usually are normally distributed. The average porosity of a heterogeneous nonuniform medium is the volume-weighted average of the number average: m

E E",

An Introduction to Liquid Filtration

1

The average nonuniform permeability is spatially dependent. For a homogeneous but nonuniform medium, the average permeability is the correct mean (first moment) of the permeability distribution function. Permeability for a nonuniform medium is usually skewed. Most data for nonuniform permeability show permeability to be distributed log-normally. The correct average for a homogeneous, nonuniform permeability, assuming it is distributed log-normally, is the geometric mean, defined as:

<*> -

n*,

(14)

1=1

For flow in heterogeneous media, the average permeability depends on the arrangement and geometry of the nonuniform elements, each of which has a different, average permeability. Figure 1 conceptually illustrates nonuniform elements, where the elements are parallel to the flow.

Figure 1. Flow through parallel nonuniform elements of porous media.

Since flow is through parallel elements of different constant area, Darcy's law for each element, assuming the overall length of each element is equal, is:

(15)

The flowrate through the entire system of elements is Q=Q S +Q 2 +. Combining these expressions we obtain:

(16a)

Liquid Filtration

or

(I6b)

This means that the average permeability for this heterogeneous medium is the area-weighted average of the average permeability of each of the elements. If the permeability of each element is log-normally distributed, these are the geometric means. Reservoirs and soils are usually composed of heterogeneities that are nonuniform layers, so that only the thickness of the layers varies. This means that \{kp)} simplifies to: h.(k.) + /L<JL> + . . .

If all the layers have the same thickness, then h

£*, «*», = *

<18)

where n is the number of layers. Permeability is a volume-averaged property for a finite but small volume of a medium. Anisotropy in natural or manmade packed media may result from particle (or grain) orientation, bedding of different sizes of particles or layering of media of different permeability. A dilemma arises when considering whether to treat a directional effect as anisotropy or as an oriented heterogeneity. In an oriented porous medium, the resistance to flow differs depending on the direction. Thus, if there is a pressure gradient between two points and a particular fluid particle is followed, unless the pressure gradient is parallel to oriented flow paths, the fluid particle will not travel from the original point to the point which one would expect. Instead, the particle will drift. Tortuosity and connectivity are difficult to relate to the nonuniformity and anisotropy of a medium. Attempts to predict permeability from a pore structure model require information on tortuosity and connectivity. From an industrial viewpoint, the objective of the unit operation of filtration is the separation of suspended solid particles from a process fluid stream which is accomplished by passing the suspension through a porous medium that is referred to as a filter medium. In forcing the fluid through the voids of the filter medium, fluid alone flows, but the solid particles are retained on the surface and in the medium's pores. The fluid discharging from the medium is called the filtrate. The operation may

An Introduction to Liquid Filtration

9

be performed with either incompressible fluids (liquids) or slightly to highly compressible fluids (gases). The physical mechanisms controlling filtration, although similar, vary with the degree of fluid compressibility. Although there are marked similarities in the particle capture mechanisms between the two fluid types, design methodologies for filtration equipment vary markedly. This reference volume concentrates only on process liquid handling (i.e., incompressible fluid processing).

The Filter Media The filter medium represents the heart of any filtration device. Ideally, solids are collected on the feed side of the plate while filtrate is forced through the plate and carried away on the leeward side. A filter medium is, by nature, inhomogeneous, with pores nonuniform in size, irregular in geometry and unevenly distributed over the surface. Since flow through the medium takes place through the pores only, the micro-rate of liquid flow may result in large differences over the filter surface. This implies that the top layers of the generated filter cake are inhomogeneous and, furthermore, are established based on the structure and properties of the filter medium. Since the number of pore passages in the cake is large in comparison to the number in the filter medium, the cake's primary structure depends strongly on the structure of the initial layers. This means that the cake and filter medium influence each other. Pores with passages extending all the way through the filter medium are capable of capturing solid particles that are smaller than the narrowest cross section of the passage. This is generally attributed to particle bridging or, in some cases, physical adsorption. Depending on the particular filtration technique and intended application, different filter media are employed. Examples of common media are sand, diatomite, coal, cotton or wool fabrics, metallic wire cloth, porous plates of quartz, chamotte, sintered glass, metal powder, and powdered ebonite. The average pore size and configuration (including tortuosity and connectivity) are established from the size and form of individual elements from which the medium is manufactured. On the average, pore sizes are greater for larger medium elements. In addition, pore configuration tends to be more uniform with more uniform medium elements. The fabrication method of the filter medium also affects average pore size and form. For example, pore characteristics are altered when fibrous media are first pressed. Pore characteristics also depend on the properties of fibers in woven fabrics, as well as on the exact methods of sintering glass and metal powders. Some filter media, such as cloths (especially fibrous layers), undergo considerable compression when subjected to typical pressures employed in industrial filtration operations. Other filter media, such as ceramic, sintered plates of glass and metal powders, are stable under the same operating conditions. In addition, pore characteristics are greatly influenced by the separation process occurring within the pore passages, as this leads to a decrease in effective pore size and consequently an increase in flow resistance. This results from particle penetration into the pores of the filter medium.

10

Liquid Filtration

The separation of solid particles from a liquid via filtration is a complicated process. For practical reasons filter medium openings should be larger than the average size of the particles to be filtered. The filter medium chosen should be capable of retaining solids by adsorption. Furthermore, interparticle cohesive forces should be large enough to induce particle flocculation around the pore openings.

Liquid Filtration Classification There are two major types of filtration: "cake" and "filter-medium" filtration. In the former, solid particulates generate a cake on the surface of the filter medium. In filter-medium filtration (also referred to as clarification), solid particulates become entrapped within the complex pore structure of the filter medium. The filter medium for the latter case consists of cartridges or granular media. Examples of granular materials are sand or anthracite coal. Process engineers who specify filtration equipment for an intended application must first account for the parameters governing the application and then select the filtration equipment best suited for the job. There are two important parameters that must be considered, namely the method to be used for forcing liquid through the medium, and the material that will constitute the filter medium. When the resistance opposing fluid flow is small, gravity force effects fluid transport through a porous filter medium. Such a device is simply called a gravity filter. If gravity is insufficient to instigate flow, the pressure of the atmosphere is allowed to act on one side of the filtering medium, while a negative or suction pressure is applied on the discharge side. This type of filtering device is referred to as a vacuum filter. The application of vacuum filters is typically limited to 15 psi pressure. If greater force is required, a positive pressure in excess of atmospheric can be applied to the suspension by a pump. This motive force may be in the form of compressed air introduced in a montejus, or the suspension may be directly forced through a pump acting against the filter medium (as in the case of a filter press), or centrifugal force may be used to drive the suspension through a filter medium as is done in screen centrifuges. Filtration is a hydrodynamic process hi which the fluid's volumetric rate is directly proportional to the existing pressure gradient across the filter medium, and inversely proportional to the flow resistance imposed by the connectivity, tortuosity and size of the medium's pores, and generated filter cake. The pressure gradient constitutes the driving force responsible for the flow of fluid. Regardless of how the pressure gradient is generated, the driving force increases proportionally. However, in most cases, the rate of filtration increases more slowly than the rate at which the pressure gradient rises. The explanation for this phenomenon is that as the gradient rises, the pores of filter medium and cake are compressed and consequently the resistance to flow increases. For highly compressible cakes, both driving force and resistance increase nearly proportionally and any rise in the pressure drop has a minor effect on the filtration rate.

An Introduction to Liquid Filtration

11

The Formation of Filter Cake Filtration operations are capable of handling suspensions of varying characteristics ranging from granular, incompressible, free-filtering materials to slimes and colloidal suspensions in which the cakes are incompressible. These latter materials tend to contaminate or foul the filter medium. The interaction between the particles in suspension and the filter medium determines to a large extent the specific mechanisms responsible for filtration. In practice cake filtration is used more often than filter-medium filtration. Upon achieving a certain thickness, the cake is removed from the medium by various mechanical devices or by reversing the flow of filtrate. To prevent the formation of muddy filtrate at the beginning of the subsequent filtration cycle, a thin layer of residual particles is sometimes deposited onto the filter medium. For the same reason, the filtration cycle is initiated with a low, but gradually increasing pressure gradient at an approximately constant flowrate. The process is then operated at a constant pressure gradient while experiencing a gradual decrease in process rate. The structure of the cake formed and, consequently, its resistance to liquid flow depends on the properties of the solid particles and the liquid phase suspension, as well as on the conditions of filtration. Cake structure is first established by hydrodynamic factors (cake porosity, mean particle size, size distribution, and particle specific surface area and sphericity). It is also strongly influenced by some factors that can conditionally be denoted as physicochemical. These factors are: 1. the rate of coagulation or peptization of solid particles, 2. the presence of tar and colloidal impurities clogging the pores, 3. the influence of electrokinetic potentials at the interphase in the presence of ions, which decreases the effective pore cross section, and 4. the presence of solvate shells on the solid particles (this action is manifested at particle contact during cake formation). Due to the combining effects of hydrodynamic and physicochemical factors, the study of cake structure and resistance is extremely complex, and any mathematical description based on theoretical considerations is at best only descriptive. The influence of physicochemical factors is closely related to surface phenomena at the solid-liquid boundary. It is especially manifested by the presence of small particles in the suspension. Large particle sizes result in an increase in the relative influence of hydrodynamic factors, while smaller sizes contribute to a more dramatic influence from physicochemical factors. No reliable methods exist to predict when the influence of physicochemical factors may be neglected. However, as a general rule, for rough evaluations their influence may be assumed to be most pronounced in the particle size range of 15-20 /mi.

12

Liquid Filtration

Typical Industrial Filtration Conditions Two significant operating parameters influence the process of filtration: the pressure differential across the filtering plate, and the temperature of the suspension. Most cakes may be considered compressible and, in general, their rate of compressibility increases with decreasing particle size. The temperature of the suspension influences the liquid-phase viscosity, which subsequently affects the ability of the filtrate to flow through the pores of the cake and the filter medium. In addition, the filtration process can be affected by particle inhomogeneity and the ability of the particles to undergo deformation when subjected to pressure and settling characteristics due to the influence of gravity. Particle size inhomogeneity influences the geometry of the cake structure not only at the moment of its formation, but also during the filtration process. During filtration, small particles retained on the outer layers of the cake are often entrained by the liquid flow and transported to layers closer to the filter medium, or even into the pores themselves. This results in an increase in the resistances across the filter medium and the cake that is formed, Particles that undergo deformation when subjected to transient or high pressures are usually responsible for the phenomenon known as pore clogging. Fortunately, what nature has sometimes neglected in the filterability of suspensions, man can correct through the addition of coagulating and peptizing agents. These are additives which can drastically alter the cake properties and, subsequently lower flow resistance and ultimately increase the filtration rate and the efficiency of separation. Filter aids may be used to prevent the penetration of fine particles into the pores of a filter plate when processing low concentration suspensions. Filter aids build up a porous, permeable, rigid lattice structure that retains solid particles on the filter medium surface, while permitting liquid to pass through. They are often employed as precoats with the primary aim of protecting the filter medium. They may also be mixed with a suspension of diatomaceous silica type earth (>90% silica content). Cellulose and asbestos fiber pulps were typically employed for many years as well. The discussions of the basic features of filtration given thus far illustrate that the unit operation involves some rather complicated hydrodynamics that depend strongly on the physical properties of both fluid and particles, as well as interaction with a complex porous medium. The process is essentially influenced by two different groups of factors, which can be broadly lumped into macro- and micro-properties. Macrofactors are related to variables such as the area of a filter medium, pressure differences, cake thickness and the viscosity of the liquid phase. Such parameters are readily measured. Micro-factors include the influences of the size and configuration of pores in the cake and filter medium, the thickness of the electrical double layer on the surface of solid particles, and other properties.

Washing and Dewatering Operations When objectionable (i.e., contaminated or polluted), or valuable suspension liquors are present, it becomes necessary to wash the filter cake to effect clean separation of

An Introduction to Liquid Filtration

13

solids from the mother liquor or to recover the mother liquor from the solids. The operation known as de-watering involves forcing a clean fluid through the cake to recover residual liquid retained in the pores, directly after filtering or washing. If the fluid is gas, then liquid is displaced from the pores. Also, by preheating the gas, the hydrodynarnic process is aided by diffusional drying. Dewatering is a complex process on a microscale, because it involves the hydrodynamics of two-phase flow. Although washing and dewatering are performed on a cake with an initially well defined pore structure, the flows become greatly distorted and complex due to changing cake characteristics. The cake structure undergoes compression and disintegration during both operations, thus resulting in a dramatic alteration of the pore structure.

General Considerations for Process Engineers In specifying and designing filtration equipment, primary attention is given to options that will minimize high cake resistance. This resistance is responsible for losses in filtration capacity. One option for achieving a required filtration capacity is the use of a large number of filter modules. Increasing the physical size of equipment is feasible only within certain limitations as dictated by design considerations, allowable operating conditions, and economic constraints. A more flexible option from an operational viewpoint is the implementation of process-oriented enhancements that intensify particle separation. This can be achieved by two different methods. In the first method, the suspension to be separated is pretreated to obtain a cake with minimal resistance. This involves the addition of filter aids, flocculants or electrolytes to the suspension. In the second method, the period during which suspensions are formed provides the opportunity to alter suspension properties or conditions that are more favorable to low-resistance cakes. For example, employing pure initial substances or performing a prefiltration operation under milder conditions tends to minimize the formation of tar and colloids. Similar results may be achieved through temperature control, by limiting the duration of certain operations immediately before filtering such as crystallization, or by controlling the rates and sequence of adding reagents. Filtration equipment selection is often complex and sometimes confusing because of (1) the tremendous variations in suspension properties; (2) the sensitivities of suspension and cake properties to different process conditions; and (3) the variety of filtering equipment available. Generalities in selection criteria are, therefore, few; however, there are some guidelines applicable to certain classes of filtration applications. One example is the choice of a filter whose flow orientation is in the same direction as gravity when handling polydispersed suspensions. Such an arrangement is more favorable than an upflow design, since larger particles will tend to settle first on the filter medium, thus preventing pores from clogging within the medium structure.

14

Liquid Filtration

A further recommendation, depending on the application, is not to increase the pressure difference for the purpose of increasing the filtration rate. The cake may, for example, be highly compressible; thus, increased pressure would result in significant increases in the specific cake resistance. We may generalize the selection process to the extent of applying three rules to all filtration problems: 1. The objectives of a filtration operation should be defined; 2. Physical and/or chemical pretreatment options should be evaluated for the intended application based on their availability, cost, ease of implementation and ability to provide optimum filterability; and 3. Final filtration equipment selection should be based on the ability to meet all objectives of the application within economic constraints.

The Objectives of Filtration The objectives for performing filtration usually fall into one of the following categories: 1. 2. 3. 4.

clarification for liquor purification, separation for solids recovery, separation for both liquid and solids recovery, and/or separation aimed at facilitating or improving other plant operations.

Clarification involves the removal of relatively small amounts of suspended solids from suspension (typically below 0.15% concentration). A first approach to considering any clarification option is to define the required degree of purification. That is, the maximum allowable percentage of solids in the filtrate must be established. Compared with other filter devices, clarifying filters are of lesser importance to pure chemical process work. They are primarily employed in beverage manufacturing and water polishing operations, pharmaceutical filtration, fuel/ lubricating oil clarification, electroplating solution conditioning, and dry-cleaning solvent recovery. They are also heavily employed in fiber spinning and film extrusion. In filtration for solids recovery, the concentration of solids suspension must be high enough to allow the formation of a sufficiently thick cake for discharge in the form of a solid mass before the rate of flow is materially reduced. However, solids concentration alone is not the only criterion for adequate cake formation. For example, an 0.5% suspension of paper pulp may be readily cake-forming whereas a 10% concentration of certain chemicals may require thickening to produce a dischargeable cake. Filtration for both solids and liquid recovery differs from filtration for solids recovery alone in the cake building, washing and drying stages. If the filtrate is a valuable liquor, maximum washing is necessary to prevent its loss; but if it is valueless, excess wash liquor can be applied without regard to quality.

An Introduction to Liquid Filtration

15

Finally, filtration can be applied to facilitate other plant operations. Like other unit operations, filtration has the most immediate relationship to those operations immediately preceding and following it. Ahead of filtration, the step is often one of preparation. These prefiltration steps could include thickening, coagulating, heating, conditioning, pH adjustment or the handling of an unstable flow that must not be broken by rapid pumping or agitation before filtration. Such preparation stages are used to obtain more filterable material. This allows a continuous operating mode, smaller filter areas or both. Figure 2 schematically summarizes the prefiltration and final processing steps. CHEMICALS

SETTLING TANK

SOURCE

FILTER

TO

—I

PROCESS

'FILTER BACKWASH R E C Y C L E TO P L A N T OR CLARIFIERSUROE TANK

DISCHARGE

DEWATERINQ

SOLIDS

DISPOSAL CHEMICAL RECOVERY

Figure 2. Summary of prefiltration and final processing steps in a filtering operation.

Filtration may also serve as the preparatory step for the operation following it. The latter stages may be dry ing or incineration of solids, concentration or direct use of the filtrate. Filtration equipment must be selected on the basis of their ability to deliver the best feed material to the next step. Dry, thin, porous, flaky cakes are best suited for drying where grinding operations are not employed. In such cases, the cake will not ball up, and quick drying can be achieved. A clear, concentrated filtrate often aids downstream treatment, whereby the filter can be operated to increase the efficiency of the downstream equipment without affecting its own efficiency.

Preparation Stages for Filtration A number of preparation steps alluded to earlier assist in achieving optimum filterability. The major ones are briefly described below. Use of Precoat and Filter Aids Where particles of a colloidal nature are encountered in liquor clarification, a precoat and or filter aid are often required to prevent deposited particles from being carried

16

Liquid Filtration

by strearnflow impact into the pores of the filter medium (or filter cake after formation), thus reducing capacity. A precoat serves only as a protective covering over the filter medium to prevent the particles from reaching the pores, while the filter aid added to the influent assists in particle separation and cake formation. Filter aids serve as obstructions, intervening between the particles to prevent their compacting, and producing, under the pressure velocity impact, a more or less impervious layer on the filter medium, or if a precoat is used, on it. In some instances, precoats are used, not because of danger to filter cloth clogging, but to permit the use of a coarser filter medium such as metallic cloths. This can extend operating life or improve corrosion resistance. Coagulation This is another means of dealing with colloidal or semicolloidal particles. It applies particularly to clarification in water and sewage filtration and in the filtration of very fine solids. While flocculation often can be accomplished by agitation, the use of chemical additives results in alteration of the physical structure of the suspended solids to the extent of losing their colloidal nature and becoming more or less crystalline. This is usually accompanied by agglomeration. Clarification by settling may follow, if the specific gravity of the particles is sufficient to provide reasonably quick supernatant clarity. Direct filtration may be applied if the filter area is not excessive or if complete supernatant clarity is needed. Temperature Control Temperature has a direct impact on viscosity, which in turn affects the flowrate. It is an important factor in filtration, since lower viscosity leads to liquor penetration into smaller voids and in shorter times. Occasionally, temperature plays a role in altering the particle form or composition, and this in turn affects the clarification rate. The Control of pH Proper pH control can result in clarification that might otherwise not be feasible, since an increase in alkalinity or acidity may change soft, slimy solids into firm, free-filtering ones. In some cases precoats are employed, not because of the danger of filter cloth clogging, but to allow the use of a coarser filter medium, such as metallic cloth.

Equipment Selection Methodology Equipment selection is seldom based on rigorous equations or elaborate mathematical models. Where equations are used, they function as a directional guide in evaluating data or process arrangements. Projected results are derived most reliably from actual

An Introduction to Liquid Filtration

11

plant operational data and experience where duplication is desired; from standards set up where there are few variations from plant to plant, so that results can be anticipated with an acceptable degree of confidence (as in municipal water filtration); or from pilot or laboratory tests of the actual material to be handled. Pilot plant runs are typically designed for short durations and to closely duplicate actual operations. Proper selection of equipment may be based on experiments performed in the manufacturer's laboratory, although this is not always feasible. Sometimes the material to be handled cannot readily be shipped; its physical or chemical conditions change during the time lag between shipping and testing, or special conditions must be maintained during filtration that cannot be readily duplicated, such as refrigeration, solvent washing and inert gas use. A filter manufacturer's laboratory has the advantage of having numerous types of filters and apparatus available with experienced filtration engineers to evaluate results during and after test runs. The use of pilot-plant filter assemblies is both common and a classical approach to design methodology development. These combine the filter with pumps, receivers, mixers, etc., in a single compact unit and may be rented at a nominal fee from filter manufacturers, who supply operating instructions and sometimes an operator. Preliminary tests are often run at the filter manufacturer's laboratory. Rough tests indicate what filter type to try in the pilot plant. Comparative calculations of specific capacities of different filters or their specific filter areas should be made as part of the evaluation. Such calculations may be performed on the basis of experimental data obtained without using basic filtration equations. In designing a new filtration unit after equipment selection, calculations should be made to determine the specific capacity or specific filtration area. Basic filtration equations may be used for this purpose, with preliminary experimental constants evaluated. These constants contain information on the specific cake resistance and the resistance of the filter medium. The basic equations of filtration cannot always be used without introducing corresponding corrections. This arises from the fact that these equations describe the filtration process partially for ideal conditions when the influence of distorting factors is eliminated. Among these factors are the instability of the cake resistance during operation and the variable resistance of the filter medium, as well as the settling characteristics of solids. In these relationships, it is necessary to use statistically averaged values of both resistances and to introduce corrections to account for particle settling and other factors. In selecting filtration methods and evaluating constants in the process equations, the principles of similarity modeling are relied on heavily. Within the subject of filtration, a distinction is made between micro- and macromodeling. The first one is related to modeling cake formation. The cake is assumed to have a well defined structure, in which the hydrodynamic and physicochemical processes take place. Macromodeling presents few difficulties, because the models are process-oriented (i.e., they are specific to the particular operation or specific equipment). If distorting side effects are not important, the filtration process may be designed according to existing empirical correlations. In

18

Liquid Filtration

practice, filtration, washing and dewatering often deviate substantially from theory. This occurs because of the distorting influences of filter features and the unaccounted for properties of the suspension and cake. Existing statistical methods permit prediction of macroscopic results of the processes without complete description of the microscopic phenomena. They are helpful in establishing the hydrodynamic relations of liquid flow through porous bodies, die evaluation of filtration quality with pore clogging, description of particle distributions and in obtaining geometrical parameters of random layers of solid particles. Nomenclature A b c D Dp g h k K L n p q Q Qm r Re S V vm x z

= = = = = = = = = = = = = = = = = = = = = =

area ( m ) parameter in slip flow expression for K (sec2-m/kg) shape factor, known as Kozeny constant diameter (m) particle diameter (m) acceleration due to gravity (m/sec ) hydraulic head (m) intrinsic permeability ( k ) hydraulic conductivity (m/sec) characteristic macroscopic length (m) number of pore layers pressure (kg/sec -m) seepage velocity (m/sec) volumetric flowrate (m3/sec) volumetric flowrate at average pressure pm (m3/sec) radius Reynolds number specific surface (m2) volume (m) velocity of approach (m/sec) coordinate (m) coordinate (in direction of gravity) (m)

Greek Symbol

fj. p T 4>

= = =

viscosity (kg/m-sec) density (kg/m3) tortuosity porosity

FILTER MEDIA AND USE OF FILTER AIDS Introduction In conventional filter-medium filtration practices, the filter medium may be described as the workhorse of the process. Proper selection is often the most important consideration for assuring efficient suspension separation. A good filter medium should have the following characteristics: The ability to retain a wide size distribution of solid particles from the suspension, Offer minimum hydraulic resistance to the filtrate flow, Allow easy discharge of cake, High resistance to chemical attack, Resist swelling when in contact with filtrate and washing liquid, Display good heat-resistance within the temperature ranges of filtration, Have sufficient strength to withstand filtering pressure and mechanical wear, Capable of avoiding wedging of particles into its pores.

There are many filter media from which to choose from; however, the optimum type often depends on the properties of the suspension and specific process conditions. Filter media may be classified into several groups, however the two most common classes are the surface-type and depth-media-type. Surface-type filter media are distinguished by the fact that the solid particles of suspension on separation are mostly retained on the medium's surface. That is, particles do not penetrate into the pores. Common examples of this type of media are filter paper, filter cloths, and wire mesh. Depth-type filter media are largely used for liquid clarification. They are characterized by the fact that the solid particles penetrate into the pores where they are retained. The pores of such media are considerably larger than the particles of suspension. The suspension's concentration is generally not high enough to promote particle bridging 19

20

Liquid Filtration

inside the pores. Particles are retained on the walls of the pores by adsorption, settling and sticking. As a rule, depth-type filter media cannot retain all suspended particles, and their retention capacity is typically between 90-99%. Sand and filter aids, for example, fall into this category. Some filter media may act as either surface-type or depth-type, depending on the pore size and suspension properties (e.g., particle size, solids concentration and suspension viscosity). It is also common practice to classify filter media by their materials of construction. Examples are cotton, wool, linen, glass fiber, porous carbon, metals and rayons. Such a classification is convenient for selection purposes, especially when resistance to aggressive suspensions is a consideration. We may also classify media according to structure, with typical classes being rigid, flexible and semi-rigid or combination media. Filtration aids are employed to enhance filtration characteristics, particularly for hardto-filter suspensions. These are normally applied as an admix to the suspensions. The role of the filter aid is to built up a porous, permeable and rigid lattice structure that assists in retaining solid particles while allowing liquid to flow through. This chapter provides a working knowledge of the use and selection of filter aids. Further discussions are given in subsequent chapters.

Flexible Filter Media Flexible nonmetallic materials have been widely used as filter media for many years. They are available in the form of fabrics or as preformed unwoven materials, but also in the form of perforated plates. Fabric filter media are characterized by the characteristics of mesh count, mesh opening, yarn size and the type of weave. The mesh count or thread count of a fabric is the number of threads per inch. Thread counts in both warp and weft directions are the same, and are indicated by a single number. Warp threads run lengthwise in a fabric and are parallel to the selvage edge. Weft or filling threads run across the width of a fabric at right angles to the warp. Figure 1 illustrates the important construction parameters that characterize a fiber-based fabric. Note that the space between threads is the mesh opening. It is measured in units of micrometers or inches. Different yarn sizes are normally specified as a measurement of diameter in micrometers or mils (thousandths of an inch). Yarn sizes in the warp and weft directions are normally the same, and are indicated by a single number. Fabrics are available in differing mesh openings, and varying thread diameters. The thread diameter affects the amount of open area in a particular cloth, which in turn determines the filtration flowrate or throughput.

Filter Media and Use of Filter Aids

21

WEFT MESH COUNT

H THREAD DIAMETER

Figure 1. Construction parameters that determine the characteristics of a fiber-based fabric.

A plain weave is the most basic weave, with a weft thread alternately going over one warp thread and then under one warp thread. A twill weave produces a diagonal or twill line across the fabric face. These diagonals are caused by moving the yarn intersections one weft thread higher on successive warp yams. A twill weave is designated 2/1, 2/2, or 3/1 depending on how many weft threads the warp threads go over and under. A satin weave has a smooth surface caused by carrying the warp (or the weft) on the fabric surface over many weft (or warp) yarns. Intersections between warp and weft are kept to a minimum, just sufficient to hold the fabric firmly together and still provide a smooth fabric surface. The percentage of open area in a textile filter indicates the proportion of total fabric area that is open, and can be determined by the following relationship: % open area

(mesh opening)2 x 100 (mesh opening + thread diameter}'

(1)

The following are some examples of different types of common flexible filter media, Glass Cloths Glass cloths are manufactured from glass yarns. They have high thermal resistance, high corrosion resistance and high tensile strength, and are easily handled; the composition and diameter of the fibers can be altered as desired. The disadvantages

22

Liquid Filtration

of glass cloth are the lack of flexibility of individual fibers, causing splits and fractures, and its low resistance to abrasion. However, backing glass cloth with a lead plate, rubber mats or other rigid materials provides for longevity. Backing with cotton or rubber provides about 50% greater life than in cases where no backing is used,

Cotton Cloths Cotton filter cloths are among the most widely used filter media. They have a limited tendency to swell in liquids and are used for the separation of neutral suspensions at temperatures up to 100°C, as well as suspensions containing acids up to 3% or alkalies with concentrations up to 10% at 15-20°C. Hydrochloric acid at 90-100°C destroys cotton fabric in about 1 hour, even at concentrations as low as 1.5%. Nitric acid has the same effect at concentrations of 2.5%, and sulfuric acid at 5%. Phosphoric acid (70%) destroys the cloth in about six days. Water and water solutions of aluminum sulfate cause cotton fabrics to undergo shrinkage. Woven cotton filter cloths comprise ducks, twills, chain weaves, canton flannel and unbleached muslins. Cotton duck is a fabric weave that is a plain cloth with equalthickness threads and texture in the "over one and under one" of the warp and woof. The twill weave is over two and under two with the next filling splitting the warp strands and giving a diagonal rib at 45° if the number of warp and filling threads are equal. Canton flannel is a twill weave in which one surface has been brushed up to give a nap finish. A muslin cloth is a very thin duck weave, which is unbleached for filtering. In chain weave one filling goes over two warp threads and under two, the next reversing this; the third is a true twill sequence, and the next repeats the cycle. A duck may be preferable to a twill of higher porosity, because the hard surface of the duck permits freer cake discharge. Under high increasing pressure a strong, durable cloth (duck) is required, since the first resistance is small as compared with that during cake building. Certain types of filters, such as drum filters, cannot stand uneven shrinkage and, in some cases, cloths must be preshrunk to ensure fitting during the life of the cloth. Nitro-filter (nitrated cotton cloth) cloths are about the same thickness and texture as ordinary cotton filtration cloths, but are distinguished by a harder surface. It is claimed that the cake is easily detached and that clogging is rare. Their tensile strength is 70-80% of that of the specially manufactured cotton cloths from which they are prepared. They are resistant to the corrosive action of sulfuric, nitric, mixed nitration and hydrochloric acids. They are recommended for filtering sulfuric acid solutions to 40% and at temperatures as high as 90°C, with the advantage of removing finely divided amorphous particles, which would quickly clog most ceramic media. Nitrofilter cloths are composed of cellulose nitrate, which is an ester of cellulose. Any chemical compound that will saponify the ester will destroy the cloth. Caustic soda or potash in strengths of 2% at 70°C or over; alkali sulfides, polysuifides and sulfohydrates; or mixtures of ethyl alcohol and ether, ethyl, amyl and butyl acetates, pyridine, ferrous sulfates, and other reducing agents are detrimental to the cloth.

Filter Media and Use of Filter Aids

23

Cellulose nitrate is inflammable and explosive when dry, but when soaked in water it is considered entirely safe if reasonable care is taken in handling. For this reason it is colored red and packed in special containers. Users are cautioned to keep the cloths wet and to handle them carefully. Wool Cloths Wool cloths can be used to handle acid solutions with concentrations up to 5-6%. Wool cloth has a life comparable to that of cotton in neutral liquors. Wool is woven in the duck-like square cloth weave, or with a raised nap; or it may be formed as a felt. Originally the smooth cloth weave was used for filtering electrolytic slimes and similar slurries. The hairlike fibers, as in cotton cloth, ensure good filtrate clarity. Long-nap wool cloth has found wide application in sewage sludge dewatering and in cases where only ferric chloride is used for conditioning. The wool has a long life and it does not clog easily. Wool cloths are sold by weight, usually ranging 10-22 oz/yd2 with the majority at 12 oz/yd2. The clarity through wool cloths is considerably less than through cotton cloths. Paper Pulp and Fiber Cloths Paper pulp and fiber cloths are excellent materials for precoats and filter aids. Paper pulp gives a high rate of flow, is easily discharged and shows little tendency to clog. Paper pulp's disadvantage lies in its preparation. Soda or sulfate pulp, most commonly used, must be disintegrated and kept in suspension by agitation before precoating. This requires considerable auxiliary equipment. Diatomaceous earths, while they should be kept in suspension, are very easy to handle and do not undergo disintegration. Paper pulp compressed into pads is used in pressure filters for beverage clarification. After becoming dirty, as evidenced by decrease in the rate of flow, the paper may be repulped, water-washed and reformed into pads. Although this involves considerable work, excellent clarity and high flowrates are obtained. The impurities do not form a cake as such, but penetrate into the pad and can only be removed by repulping and washing the pad. Pads of a mixture of paper pulp and asbestos fiber are used hi bacteriological filtrations. In sheet form it is employed in the laboratory for all kinds of filtration. Filter papers are made in many grades of porosity for use in porcelain and glass funnels. Industrially, paper in the form of sheets is used directly or as a precoat in filter presses. Used directly in lubricating clarification in a "blotter press", it acts much the same manner as the paper pads, but is much thinner and is not reused. As a precoat, paper protects the filter medium from slimy fines; it may be peeled off and discarded after clogging, leaving the medium underneath clean.

24

Liquid Filtration

Rubber Media

Rubber media appear as porous, flexible rubber sheets and microporous hard rubber sheets. Commercial rubber media have 1100-6400 holes/in.2 with pore diameters of 0,012-0.004 in. They are manufactured out of soft rubber, hard rubber, flexible hard rubber and soft neoprene. The medium is prepared on a master form, consisting of a heavy fabric belt, surfaced on one side with a layer of rubber filled with small round pits uniformly spaced. These pits are 0.020 in. deep, and the number per unit area and their surface diameter determine the porosity of the sheet. A thin layer of latex is fed to the moving belt by a spreader bar so that the latex completely covers the pits, yet does not run into them. This process traps air in each pit. The application of heat to the under-surface of the blanket by a steam plate causes the air to expand, blowing little bubbles in the film of latex. When the bubbles burst, small holes are left, corresponding to the pits. The blown rubber film, after drying, is cooled and the process repeated until the desired thickness of sheet is obtained. The sheet is then stripped off of the master blanket and vulcanized, Approximately 95% of the pits are reproduced as holes in the rubber sheet. The holes are not exactly cylindrical in shape but are reinforced by slight constrictions which contribute to strength and tear resistance. This type is referred to as "plain," and can be made with fabric backing on one or both sides to control stretching characteristics. If the unvulcanized material is first stretched, and then vulcanized while stretched, it is called "expanded." Resulting holes are oval and have a higher porosity (sometimes up to 30%). Special compounds have been formulated for resistance to specific chemicals under high concentrations at elevated temperatures, such as 25% sulfuric acid at 180° F. The smooth surface allows the removal of thinner cakes than is possible with cotton or wool fabrics. Rubber does not show progressive binding and it can be readily cleaned and used in temperatures up to 180°F. On the other hand, because a clear filtrate is difficult to obtain when filtering finely divided solids, a precoat often becomes necessary. Synthetic Fiber Cloths

Cloths from synthetic fibers are superior to many of the natural cloths thus far considered. They do not swell as do natural fibers, are inert in many acid, alkaline and solvent solutions and are resistant to various fungus and bacterial growths (the degree depending on the particular fiber and use). Several synthetic fibers resist relatively high temperatures, and have a smooth surface for easy cleaning and good solids discharge. Some of the most widely used synthetic filter media are nylon, Saran, Dacron, Dynel, Vinyon, Orion, and Acrilan. Table 1 compares the physical properties of several synthetic fiber filter media.

Filter Media and Use of Filter Aids

25

Tightly woven, monofilament (single-strand) yarns consist of small-diameter filaments. They tend to lose their tensile strength, because their small diameters reduce their permeability; thus multifilament yarns are normally used. Monofilament yarns in loose weaves provide high flowrates, good solids discharge, easy washing and high resistance to blinding, but the turbidity of the filtrate is high and recirculation is usually necessary, initially at least. Table 2 provides additional information on various synthetic filter fabrics. Flexible Metallic Media Flexible metallic media are especially suitable for handling corrosive liquors and for high-temperature filtration. They have good durability and are inert to physical changes. Metallic media are fabricated in the form of screens, wire windings, or woven fabrics of steel, copper, bronze, nickel and different alloys. Perforated sheets and screens are used for coarse separation, as supports for filter cloths or as filter aids. Metallic cloths are characterized by the method of wire weaves as well as by the size and form of holes and by the wire thickness. Metallic cloths may be manufactured with more than 50,000 holes/cm2 and with hole sizes less than 20 /am. Table 1. Properties of woven filter cloth fibers.

Fibers

Acids

Alkalies

Solvents

Acrilan Asbestos Cotton Dacron Dyne! Glass Nyion Orion Saran Teflon Wool

Good PooiPoor Fair Good High FailGood Good High Fair

Good PooiFair Fair Good Fair Good Fair Good High Poor

Good PooiGood FailGood FailGood Good Good High Fail-

Fiber Tensile Strength

Temperature Limit

High

275 750 300 350 200

Low High High FailHigh High High High Fail-

Low

600 300 275 240 180 300

MetaHic/Nonmetallic Cloth Combination metallic and nonmetallic cloths consist of metallic wires and weak cloth or asbestos threads. There are some difficulties in weaving when attempting to maintain uniformity between wires and the cloth, and considerable dissatisfaction has been experienced with such construction. While cotton weaves well with the asbestos, the cotton fibers destroy the fabric's resistance to heat and corrosion. Its use is, therefore, quite limited, despite its resistance to high temperatures, acids and mildew.

26

Liquid Filtration

Cotton cloths are sometimes treated with metallic salts (copper sulfate) to improve their corrosion-resistant qualities. Such cloths are in the usual cotton filter cloth grades, and while they are not equivalent to metallic cloths, the treatment does materially prolong the life of the cotton fiber. Non woven Media Nonwoven media are fabricated in the form of belts or sheets from cotton, wool, synthetic and asbestos fibers or their mixtures, as well as from paper mass. They may be used in filters of different designs, for example, in filter presses, filters with horizontal discs and rotary drum vacuum filters for liquid clarification. Most of these applications handle low suspension concentrations; examples are milk, beverages, lacquers and lubricating oils. Individual fibers in nonwoven media are usually connected among them as a result of mechanical treatment. A less common approach is the addition of binding substances. Sometimes the media are protected from both sides by loosely woven cloth. Nonwoven media of various materials and weights, and in several grades of retentiveness per unit weight can be formed, in either absorbent or nonabsorbent material. These filter media retain less dispersed particles (more than 100 jum) on their surface, or close to it, and more dispersed particles within the depths of the media. Nonwoven filter media are mostly used for filter medium filtration with pore clogging. Because of the relatively low cost of this medium, it is often replaced after pore clogging. In some cases, nonwoven media are used for cake filtration. In this case, cake removal is so difficult that it must be removed altogether from the filter medium. Nonwoven filter media can be prepared so that pore sizes decrease in the direction from the surface of the filter media contacting suspension to the surface contacting the supporting device. This decreases the hydraulic resistance of filtration and provides retention of relatively large particles of suspension over the outer layer of the nonwoven medium. Nonwoven filter media of synthetic, mechanically pressed fibers are manufactured by puncturing the fiber layer with needles (about 160 punctures/cm2), and subsequent high temperature treatment with liquid which causes fiber contraction. Such filter media are distinguished by sufficient mechanical strength and low hydraulic resistance, as well as uniform fiber distribution. Filter media from fibers connected by a blinder are manufactured by pressing at 70N/cm2and 150°C. These media have sufficient mechanical strength, low porosity and are corrosion resistant. Filter media may be manufactured by lining a very thin layer of heat-resistant metal (e.g., nickel 360) over a fiber surface of inorganic or organic material. Such filter media may withstand temperatures of 200°C and higher. Of the flexible filter media described, the synthetic fabrics are perhaps the most widely relied on in industrial applications. Each filtration process must meet certain requirements in relation to flowrate, clarity of filtrate, moisture of filter cake, cake release and nonbinding characteristics. The ability of a filter fabric to help meet these criteria, and to resist chemical and physical attack depend on such characteristics as

Style No. Nylon 6,6.6 Warp and Weft Monofilament 11 1-020 111-110 111-150 111-160 111-170 111-180 111-190 111-206 111-220 111-230 111-056 Warp and Weft Monofilament I053 1093 1103 1123 1153 1193 1203 1212 1283

Weave

Weight

(oz/yd*)

Threadslia., Warp x Weft

Plain Piain Plain Plain Plain Plain Twill Ptain Plain Plain Satin

4.5 4.6 3.2 2.4 4.6 2.3 5.5 5.3 2.9 5.7 7.2

22 x 22 50 X 50 62 X 62 107 X 76 29 X 29 66 X 66 38 X 38 183 X 43 80 X 80 40 X 40 109 x 42

Plain Twill Twill Twill Satin Satin Satin Satin Plain

2.1 3.5 3.7 3.2 3.2 5.3 6.8 6.5 1.8

147 X 97 297 X 122 297 X 135 195 X 140 236 X 99 300 x 99 152 X 76 178 x 97 112 x 97

Thread D i m . , Warp x Weft

Mesh opening

(Pm)

( I 4

(fP/min)

850 X 850 300 x 300 250 X 250 250 X 230 600 X 600 210 x 210 420 X 420

NAb NA NA NA NA NA NA

570 350 270 220 450 270 5 30

NA NA

240 450 450

150-200 75-100 40-70 15-25 15-25 45-70 20-30

160 250 250 190 210 300 390 710 130

305 x 200 x 150 X 100 x 250 X 130 X 250 X 150 X 125 X 250 X 205 X

305 200 150 100 250 130 250 150 125 250 300

175 420

X X

175 420

Air Permeability

170-210

350-400

50-80 170-220

Thickness

bd

310

00 N

Air Permeability

whlin)

Style No. 1338 1353 1363 1393 122-053 122-073

Len0

Plain Plain Twill Twill Oxford Warp Monofilament, Weit Spun 1233 Satin Warp and Weft Monofilament and Metal Spun 9165 Twill Nylon 1 1 Warp and Weft Monofilarnent 1656 Satin 1666 Satin 1686 Satin 111-096 Satin Nomex Warp Multifilament. Weft Spun 1513 Plain Polyester Warp and Weft Moxicifilament 1713 Plain 1716 Plain

NA

11.5 9.7 4.7 15.5

7 x 5 64 X 48 69 x 28 236 x 53 80 X 117 72 x 21

50-100 1-3 5- 10 50-80 0.5-2

630 255 660 560 410 720

5 .O

320 x 71

60- 100

410

4.1

297 x 132

25-40

300

8.2 9.0

x 53 x 53

200-300 125-200 125-200 150-300

520 530

7. I 9.5

99 111 107 99

7.7

107

X

50-80

5 10

4. I 4.1

335 x 84 350 x 79

80- 120

190 1x0

5.6

3.2

x <91 x 53 66

180 180 180 205

x x x x

290 290 180 300

10-20

395

490

3

Table 2 Continued. Style No.

Weave

1733 Plain 9656 B Satin Satin 9813 B Satin 9884 B 31 1-010 HSC Plain 311-020 Plain Warp and Weft Multifilament 1813 Plain 1853 Leno Twill 1893 Twill 18% Satin 1933 Twill 1943 Plain 1953 1956 Plain Plain 1973 Plain 1976 Plain 9881 Twill 322-013 Plain 322-033 Plain 322-036 Broken 322-043 Twill Plain 322-070

„. . . . ,„. , .. ... Weight Threads/in., Warp x / / M\ «r L 2 (oz/yd ) Weft 8.8 8.3 11.8 12.7 6.9 6.9

175 104 112 145 38 38

x 36 X 53 x 53 x 53 X 38 X 38

10.3 10.0 4.1 4.1 3.2 6.5 10.0 10.0 3.5 3.5 2.7 3.6 7.1 7.1 5.3

43 7 295 295 290 257 36 36 94 91 152 219 81 81 216

x X x X X x X X X x X x X X x

11.5

30 5 112 112 99 79 30 30 81 84 81 147 56 56 100

52 x 32

Thread Diam., ' Warp * x Weft

200 200 200 250 250

X 300 x 300 X 250 X 250 X 250

Mesh Opening *~—-&

Air Permeability (tf/nun)

Thickness

420 x 420 420 x 420

300-400 300-400 200-250 80-150 NA NA

380 450 510 440 400 400

1-3 NA 40-75 5-10 35-55 10-20 2-4 1-2 20-40 5-15 40-60 5-10 3-7 1-3 30-50

500 800 250 200 190 330 420 380 195 140 270 180 300 280 250

1-2.5

600

Table 2 Continued. Style No,

Weave

322-073 Plain 322-123 Twill Warp Multifilament, Weft Spun 323-013 Twill Warp and Weft Spun 1823 Twill 1873 Plain 1923 Plain 333-020 Plain Acrylic Warp and Weft Multifilament 2013 Plain 2023 Twill Warp Multifilament, Weft Spun 2033 Twill Warp and Weft Spun 2063 Twill 9856 Twill Poly vi ny Ichloride 2973 Twill Polypropylene Warp und Weft Monofllament 2773 Satin 2776 Satin

„, . , , „,, , ,, ,,, Weight Threads/in., Warp x / ijo\ «7 «* (oz/yd ) Weft

Thread Diain., Warp x Weft , , (taa)

Mesh „ . Opening ; , u^m)

,. „ ,,_ Air Permeability f£,3. . , (ft /mm)

Thickness

(fan)

13.0 5.3

54 X 36 216 X 100

0.5-1.5 40-60

580 270

13.2

76 X 53

4-6

580

10.9 10.6 5.6 9.6

51 53 64 47

30-40 7-15 70-100 20-30

770 610 360 570

x X x X

25 30 53 29

7.1 4.4

81 x 33 102 X 69

3-7 15-20

480 250

6.5

102 x 61

25-40

480

9.4 7.1

69 X 41 51 X 64

10-20 40-60

700 580

6.2

66 X 51

60-90

325

9.7 9.7

86 x 38 86 x 38

70-100 45-90

620 680

300 x 250 320 X 290

Style No. 2793 5 11-013 511-015 511-016 511-050 511-066 511-070 HS 511-086 511-090 511-110 HS 511-126 511-136 511-150 Warp Monofilament, Weft

Weave Twill Satin Satin Satin Plain Basket Twill Plain Plain Twill Twill Satin Plain

Weight Threads/in.,Wxp x Weft (oz/yd2) 10.9 8.0 8.2 8.2 3.5 5.9 7.2 5.3 3.7 5.4

8.8 6.5 3.5

Multifilament

Satin Satin Satin Warp Monofilament, Weft Spun 513-016 Satin Warp and Weft Multifilament 2723 Plain 2726 Plain 2843 Twill 512-013 512-015 512-016

9.1

9.9 9.7 11.4 6.5 6.5 10.0

61 X 25 86x46 89 x 46 89 x 46 50 X 50 110 x 45 28 X 28 108 x 32 30 X 30 38 x 33 76 X 28 110 X 60 20 x 20

86 X 30 86 X 30 86 x 30

Thread Diam., Warp weft (Icm) 300 x 500 250 X 250 250 X 250 250 X 250 200 x 200 200 x 200 375 x 375 200 x 200 250 X 250 305 X 305 305 X 305 200 x 200 305 X 305

Mesh

ojx?nng (run)

300 x 300 530 x 530 600

X

600

950 x 950

Air Permeability

(ftJ/min) 300-400 150-300 40-60 100-150 NA 50-100 NA 100-150 NA NA 100-200 20-300 NA

Thickness (/lm) 700 550 500 580 370 450 650 320 490 600 590 450

510

100-150 10-30 40-80

600

36

5-10

680

33 x 20 33 x 20 241 X 43

5- 10 3-8 10-15

470

93

X

740

5 10

425

640

Table 2 Continued. Style No.

Weave

Twill 2853 2836 Twill 9870 Twill 522-013 Plain Twill 522-053 522-060 Plain 522-063 Plain 522-070 Twill Warp Multifiiament, Weft Spun 2873 Twill 523-040 Twill 523-043 Twill Twill 523-053 Warp and Weft Spun 2783 Twill 2833 Twill 2893 Twill Twill 9811 533-013 Plain Twill 533-033 Twill 533-070 Polyethylene Warp and Weft Monofilament 2473 Satin

„, . . . T,, . ,. „, Weight Threads/in., Warp x t(oz/yd*) ,M\ «? L Weft 16.8 16.8 19.8 10.0 18.8 14.3 15.1

8.3

58 58 69 43 70 76 77 58

X 30 X 30 X 30 x 24 X 30 X 21 x 21 x 39

20.6 10.3 11.0 18.6

175 33 70 71 72

X X 38 x 40 X 30

7.7

53 53 61 53 50 58 48

10.6 13.9 19.2 11.7 15.8 14.0

5.3

x X x X X X X

Thread Diara.. „, „. ' v x Weft Warp . . (Mm)

Mesh _ Opening F^«~6 (fan)

Thickness (jan)

4-8 2-4

1200 1100 1250

25 x 300

690 1160

1-2.5 0.5-2 20-40

960 950 630

5-10

1500

4-7 3-5 2-5

840 800

50-80 8-13 6-8 10-15 1-2 3-6 4-6

150 x 150

j5* CL,

1-2 3-6 4-9

33 30 23 23 23 26 31

155 x 58

Air Permeability (fp/min)

200-300

1200

800 1000 1200 1550

850 1200 1050

425

1 te a B

Table 2 Continued. Style No. 2483 2486 2503 411-010 411-013 411-015 Warp Monofilament, Weft Multifilament 2573

Weave

Weight Threads/in., Warp x (oz/yd2) Weft

Satin Satin Leno Plain Plain Plain

8.3 8.3 5.3 7.2 9.0 10.1

Satin

6.3

Warp and Weft Monofllament 2333 Satin 2336 Satin 2363 Plain Twill 611-010 611-020 Plain 611-040 Plain

11.2 11.2 7.7 6.6 7.5 6.5

89 89 6 105 111 112

X

48

X

48 5 42

X X X X

44 50

Thread Diam., Warp x Weft

(MH») 250 x 250 280 x 500 x 220 X 220 x 220 x

270 500 220 220 220

Mesh Opening (Mm) 40 x 300 15 X 190 4200 X 3000

150 X 36

Air Permeability (ft'/min)

Thickness

250-350 100-150 NA 200-300 100-175 50-80

665 500 1500 580 650 650

130-200

540

600-700 200-300 NA NA NA NA

700 370 900 500 490 550

(MH»)

Saran

Courtesy of Industrial Fabrics Corp., Minneapolis, MN. NA - not available. HS - high shrink.

66 66 15 50 30 20

43 43 15 50 X 30 X 20 X

X X X

250 X 280 X 320 x 200 X 250 x 305 x

250 370 380 200 250 305

65 x 340 25 x 180 1050 x 1050 300 x 300 600 x 600 950 x 950

34

Liquid Filtration

fiber type, yarn size, thread count, type of weave, fabric finish and yarn type (monofilament, multifilament or spun). Monofilament yarns consist of a single, continuous filament with a relatively smooth surface. The different sizes are specified as a measurement of the diameter in mils or in micrometers. Multifilament yarns are made from many fine filaments extruded simultaneously. The different sizes are specified by a measurement of weight known as the denier. These yarns are generally used for filter fabrics which require a smooth surface and relatively tight weave. Spun yarns are made from filaments which are chopped in short lengths and then spun or twisted together. Spun yarns are made into filter fabrics with a hairy, dense surface very suitable for filtration of very fine particles. It is necessary to select the type of fiber that will offer the most resistance to breakdown normally caused by chemical, temperature and mechanical conditions of the filter process. Tables 3 through 5 can serve as rough guides to proper media selection. Table 6 provides linear conversion units between mesh size, inches and micrometers.

Rigid Filter Media Fixed Rigid Media Fixed rigid media are available in the forms of disks, pads and cartridges. They are composed of firm, rigid particles set in permanent contact with one another. The media formed have excellent void uniformity, resistance to wear and ease in handling as piece units. Depending on the particle size forming the filter media, temperature, pressure and time for caking, it is possible to manufacture media with different porosities. The higher the pore uniformity, the more uniform the shape of the particles. These media are distinguished by long life, high corrosion resistance and easy cake removal. However, the particles that penetrate inside the pores are very difficult to extract. Metallic Media Metallic filter media are widely used throughout the chemical and process industries in the form of perforated or slotted plates of steel, bronze or other materials. These designs provide for easy removal of coarse particles and for supporting loose rigid media. Powdered metal is a porous medium. The physical characteristics, chemical composition, structure, porosity, strength, ductility, shape and size can be varied to meet special requirements. The porosity ranges up to 50% void by volume, tensile strength up to 10,000 psi, varying inversely with porosity, and ductility of 3-5 % in tension, and higher in compression.

Table 3, Physical properties and chemical resistances of fibers.

Type of Fiber

Moisture ., .. , Absorption at ^AO*^/JCOOIT\ •4U Moo r;, 65%RH

Specific Gravity

,, . Moisture „ Expansion ,„, l '

Ironing Temperature

Maximum Working Temperature

Cotton Silk

1.5 1.37

7-8.5 9-9.5

45 40

200 140-160

80-100 90

Wool

1.3

13-15

42

140-160

80

Glass

2.54

0

0

300

250-300

Steel Fibers (Brunsmet * ) Polyamides PA 6 (Perlon ® )

7.9

0

0

300

500

1.14

4.5

10-14

120

100

PA 6.6 (Nylon)

1.14

4.5

10-14

150

100

PA 11 (Rislan*)

1.04

1.2

120

100

PA 12 (Vestamid « )

1.02

0.95

PA Nomex ®

1.38

5

230

90 Stab. 120 200-300

Polyester

1.38

0.4

3-5

150

140-160

Polyacrylonitrile

1.15

1-1.5

4.5-6

160

145

Melting Point (°C)b

Resistance c to Light and Weather

D>200 B= 200-400 D=170 B= 200-400 D=130 S=815 M=845-1150 M= 1440- 1455

8 C

8 = 170 M=215 8=235 M = 250-255 8 = 160 M= 183-186 8 = 168 M=175 M=375 D = 371 8=230-249 M=256 8=235-250 D-300

C

C R R

C C C C G R

Table 3 Continued,

1.38

Polyvinylidenechloride (Saran *) Polyolefins Polyethylene High-Pressure

1.7

<0.1

0

80

70

0.92

<0.1

0

70

60

Low-Pressure

0.95

<0.1

0

80

70

0.9

<0.1

0

100

90

2.1

0

0

250

200

Type of Fiber

Polypropylene Polytetrafluoroethylene (Teflon*)

0

60-80

MaHnsujn Working Temperature (°C) 60

Moisture Expansion (%)

Ironing Temperature <°C)

Melting Point (°C)b S=70 D=180-190 8 = 115-138 M= 150- 170

8 = 107-110 M= 110- 120 S = 115 M=124-138 8 = 150-155 M=163-175 8=327 D>275

Resistance c to Light and Weather R G

U U C R

Resistance Against Type of Fiber

Insect Proof

Resistance to Aging

Acid

Alkali

Cotton

Medium

Low

Unstable

Silk Wool Glass Steel Fibers (Brunsmet * )

Medium Bad Good Good

Low Low Good Good

Low Low Low Low

Low resistance, swelling Unstable Unstable Unstable Resistant

resistance resistance resistance resistance

Chlorocarbonic Hydride

Ketone

Phenol

Benzene

Resistant

Resistant

Resistant

Resistant

Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Liquid Filtration

Polyvinylchloride

Moisture Absorption at 200C(68°F), 65%RH 0.1

Specific Gravity

Table 3 Continued. Resistance Against

Insect Proof

Type of Fiber PA 6 (Perlon ) PA 6.6 (Nylon) PA 11 (Rislan ) PA 12(Vestamid") PA Nomex Polyester Polyacrylonitrile Polyvinylchloride Polyvinylidenechloride (Saran? Polyolefns Polyethylene High-pressure Low-Pressure Polypropylene Polytetrafluoroethylene (Teflon 4, @

@

Good Good Good

Good Good Good Good Good Good

Good Good

Good Good

Resistance toAgiog Good Good Good Good Good Good

Good

Good Good

Good

Good Good Good

Acid

Alkali

CMorocarbonic Hydride

Ketone

Phenol

Benzene

Resistant Resistant Resistant Swelling Resistant Resistant Resistant Unstable Resistant

Unstable Unstable Unstable Unstable Unstable Unstable Resistant Unstable Unstable

Resistant Resistant Resistant

Resisutiit Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Resistant Resistant Resistant

Unstable Unstable Low resistance Low resistance Low resistance Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant Resistant Low resistance Low resistance Resistant Resistant except NH,,OH

Resistant Resistant Resistant Swelling Resistant Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Resistant Resistant Resistant Resistant

Swelling Swelling Resistant Resistant

B Data compiled from different industry lirerature source&. b B = burning point; D = disintegration, M = melting point. S = softening point. c R = recommended; G = good; S = satisfactory. C = conditional: U = unsatisfacrory.

Swelling

Resistant Resistant Resistant Unstable Resistant

38

Liquid Filtration

Table 4. Physical properties and chemical resistances of polyester fibers used for belt filters. Specific Gravity Moisture Regain At 65%RH and 68°F(20°C)(%) Water Retention Power (%) Tensile Strength cN/dtex Wet in % of dry Elongation at Break % Wet in % of dry Ultraviolet Light Resistance Resistance to Fungus, Rot, and Mildew Resistance to Dry Heat Continuous °F °C Short-Term Exposure op °C Chemical Resistance to Acids

Acetic Acid Concentration Sulfuric Acid 20% Nitric Acid 10% Hydrochloric Acid 25% Alkalies Saturated Sodium Carbonate Chlorine Bleach Concentration Caustic Soda 25% Ammonia Concentration Potassium Permanganate 50% Formaldehyde Concentration Chlorinated Hydrocarbons Benzene Phenol Ketones, Acetone a b c d

Average properties reported as based on typical industry sources, R = recommended C = conditional U = unsatisfactory

1.38 0.4 3-5 7-9.5 95-100

10-20 100-105 R R

302 150 392 200

Cc R R C C C R R Ud U R R R R C R

Table 5. Standard fibers and micrometer ratuigsfor bag filters. Available Micrometer Ratings Construction Felts Multifilament Meshes Monofilament Meshes

Polyester Polypropylene Polyester Nylon (heavy) Nylon Polypropylene

1

3

5

10

15

25

50

75

100

X X

X X

X X

X X

X

X X

X X

X X

X X X

X

X

X

X

X

X

X

125

150

175

250

300

400

X

X

X

X

X

X

X X

X

200

600

800

X

X

X X

X X X

Compatibility and Temperature Limits for Standard Bag Materials Compatibility with Fiber Polyester Polypropylene Nylon

Organic Solvents Excellent Excellent Excellent

Animal, Vegetable, and Petro Oils Excellent Excellent Excellent

Microorganisms

Alkalies

Excellent Excellent Excellent

Good Excellent Good

Oxidizing Agents Good Good Poor

Organic Acids Good Excellent Fair

Mineral Acids Good Good Poor

Bag Capacities and Dimensions Bag Size No. 1 2 1 (inner) 2 (inner)

Fits Rosedale Model No. 815 830 815 830

Bag Surface Area (ft2) 2.0 4.4 1.6 3.6

Bag Volume (gal) 2.1 4.6 1.7 3.8

Bag Dimensions Diameter (in.) 7 7 5.75 5.75

Length (in.) 16.5 32.0 14.5 30.0

Temperature Limits (°F) 300 225 325

40

Liquid Filtration

Table 6. Comparative particle sizes. U.S. Mesh 3 3»/2

4 5 6 7 8 10 12 14 16 18 20 25 30 35 40 45 50 60 70 80 100 120 140 170 200 230 270 325 400

in. 0.265 0.223 0.187 0.157 0.132 0.111 0.0937 0.0787 0.0661 0.0555 0.0469 0.0394 0.0331 0.0280 0.0232 0.0197 0.0165 0.0138 0.0117 0.0098 0.0083 0.0070 0.0059 0.0049 0.0041 0.0035 0.0029 0.0024 0.0021 0.0017 0.0015

IMK

6730 5660 4760 4000 3360 2830 2380 2000 1680 1410 1190 1000 841 707 595 500 420 354 297 250 210 177 149 125 105 88 74 63 53 44 37

Powdered metal cannot be readily ground or machined. It is made in discs, sheets, cones or special shapes for filtering fuel oil, refrigerants, solvents, etc. The smooth surface associated with a perforated plate permits brush cleaning or scrubbing in addition to the naturally easier discharge from such surfaces. The hard metallic material has a long life, not being subject to abrasion or flexing. The size of particles filtered on such plates must be relatively large. Normally, plates are confined to free-filtering materials where there is little danger of clogging. Metallic filter media may be used either for cake filtration or depth filtration, i.e., pore clogging. Regeneration of media may be achieved by dissolving solid particles inside the pores or by back thrust of filtrate.

Filter Media and Use of Filter Aids

41

Ceramic Media Ceramic filter media are manufactured from crushed and screened quartz or chamotte, which is then thoroughly mixed with a binder (for example, with silicate glass) and sintered. Quartz media are resistant to concentrated mineral acids but not resistant to low-concentration alkalies or neutral water solutions of salts. Chamotte media are resistant to dilute and concentrated mineral acids and water solutions of their salts, but have poor resistance to alkali liquids. The rough surface of ceramic filter media promotes adsorption of particles and bridging. Sintering of chamotte with a binder results in large blocks from which filter media of any desired shape can be obtained. Using synthetic polymers as binders, ceramic filter media that do not contain plugged pores are obtained. Diatomaceous Media Diatomaceous media are available in various shapes. Their relatively uniform particle size establishes high efficiency in retaining solid particles of sizes less than I/mi as well as certain types of bacteria. Media in the form of plates and cartridges are manufactured by sintering a mixture of diatomite with a binder. Coal Media Coal media are manufactured by mixing a fraction of crushed coke with an anthracene fraction of coal tar and subsequent forming under pressure, drying and heating in the presence of a reducing flame. These media of high mechanical strength are good for use in acids and alkalies. Ebonite Media Ebonite media are manufactured from partially vulcanized rubber, which is crushed, pressed and vulcanized. These media are resistant to acids, salt solutions and alkalies. They may be used for filtration at temperatures ranging from -10 to 4-110°C. Foam Plastic Media Foam plastic media are manufactured from poly vinyl chloride, polyurethane, polyethylene, polypropylene and the other polymer materials. The foam plastic media are economical. Loose Rigid Media Filter media may also be composed of particles that are rigid in structure, but are applied in bulk loose form. That is, individual particles merely contact each other. This form has the advantage of being cheap and easy to keep clean by rearrangement of the particles. When the proper size and shape of particles are selected, the section of passage may be regulated over extremely wide limits. The disadvantages of a rigid

42

Liquid Filtration

medium in simple contact are that it can be used conveniently only in a horizontal position and that it does not allow removal of thick deposits or surface cleaning except by backwashing, without disturbing the filter bed.

Coal and Coke Coal (hard) and coke are used in water filtration, primarily for the removal of coarse suspensions, care being taken to prevent them from scouring or washing away, because of their relative lightness and fine division. Coal is principally composed of carbon, and is inert to acids and alkalies. Its irregular shapes are advantageous at times over silica sand. Though inert to acids, sand is affected by alkalies, and its spherical particle shape allows deeper solids penetration and quicker clogging than does coal. With the lighter weight of coal (normally 50 lb/ft3, compared with 100 lb/ft3 for sand), a greater surface area is exposed for solids entrapment.

Charcoal Charcoal, whether animal or vegetable, when used as a filter medium, is required to perform the dual services of decoloring or adsorbing and filtering. The char filters used in the sugar industry are largely decoloring agents and the activated carbons used in water clarification are for deodorizing and removal of taste. There are many types of charcoal in use as filter media, ranging from ordinary wood char to specially prepared carbons.

Diatomaceous Earth Diatomaceous earths may resemble the forms of the charcoals. The earths are primarily filter aids, precoats or adsorbents, the function of the filter medium being secondary. Fuller's earth and clays are used for decoloring applications; diatomaceous earths are used for clarification.

Precipitates and Salts Precipitates or salts are used when corrosive liquor must be filtered, and where there is no available medium of sufficient fineness that is corrosion-resistant and will not contaminate the cake. In these cases, precipitates or salts are used on porous supports. In the filtration of caustic liquors, ordinary salt (sodium chloride) is used as the filter medium in the form of a precoat over metallic cloth. This procedure has the advantage that the salt medium will not be detrimental to either the cake or the filtrate if inadvertently mixed with it.

Sand and Gravel Sand and gravel are the most widely used of the rigid media simple contact. Most of the sand used this way is for the clarification of water for drinking or industrial uses. Washed, screened silica sand is sold in standard grades for this work and is used in depths ranging from a few inches to several feet, depending on the type of filter and

Filter Media and Use of Filter Aids

43

clarification requirements. Heavy, irregular grains, such as magnetite, give high rates of flow and low penetration by the solid particles, and are easily cleaned. They are, however, considerably more expensive than silica sand, so their use is limited. Sand beds are often gravel-supported, but gravel alone is seldom used as the filter medium, Crushed Stone or Brick Crashed stone or brick is used for coarse filtration of particularly corrosive liquors. Their use, however, is extremely limited and they are not considered important filter media.

Filter Media Selection Criteria Due to the wide variety of filter media, filter designs, suspension properties, conditions for separation and cost, selection of the optimum filter medium is complex. Filter media selection should be guided by the following rule: a filter medium must incorporate a maximum size of pores while at the same time providing a sufficiently pure filtrate. Fulfilment of this rule invokes difficulties because the increase or decrease in pore size acts in opposite ways on the filtration rate and solids retention capacity. The difficulty becomes accentuated by several other requirements that cannot be achieved through the selection of a single filter medium. Therefore, selection is often reduced to determining the most reasonable compromise between different, mutually contradictory requirements as applied to the filter medium at a specified set of filtration conditions. Because of this, some problems should be solved before final medium selection. For example, should attempts be made to increase filtration rate or filtrate purity? Is cost or medium life more important? In some cases a relatively more expensive filter medium, such as a synthetic cloth, is only suitable under certain filtration conditions, which practically eliminates any cost consideration in the selection process. Thus, the choice may only be made after consideration of all requirements. It is, however, not practical to analyze and compare each requirement with the hope of logically deducing the best choice. There is, unfortunately, no generalized formula for selection that is independent of the details of the intended application. Each cake requires study of the specific considerations, which are determined by the details or the separation process. One can to outline a general approach for medium selection along with a test sequence applicable to a large group of filter media of the same type. There are three methods of filter media tests: laboratory- or bench-scale pilot-unit, and plant tests. The laboratory-scale test is especially rapid and economical, but the results obtained are often not entirely reliable and should only be considered preliminary. Pilot-unit tests provide results that approach plant data. The most reliable results are often obtained from plant trials.

44

Liquid Filtration

Different filter media, regardless of the specific application, are distinguished by a number of properties. The principal properties of interest are the permeability of the medium relative to a pure liquid, its retention capacity relative to solid particles of known size and the pore size distribution. These properties are examined in a laboratory environment and are critical for comparing different filter media. The permeability relative to a pure liquid, usually water, may be determined with the help of different devices that operate on the principle of measurement of filtrate volume obtained over a definite time interval at known pressure drop and filtration area. The permeability is usually expressed in terms of the hydraulic resistance of the filter medium. This value is found from: ,, V

ApS

= ——£

T

>)

When the cake thickness is 0, we may write the equation as: R/f -

y

ft

(3) W

Note that: Ap=Apt-Apf

(4)

where Ap t = pressure difference accounting for the hydrostatic pressure of a liquid column at its flow through the filter medium, supporting structure and device channels Apf = same pressure drop when the flow of liquid is through the supporting structure and device channels Analytical determination of the hydraulic resistance of the medium is difficult. However, for the simplest filter medium structures, certain empirical relationships are available to estimate hydraulic resistance. The relationship of hydraulic resistance of a cloth of monofilament fiber versus fiber diameter and cloth porosity can be based on a fixed-bed model. In evaluation and selection of a filter medium, one should account for the fact that hydraulic resistance increases gradually with time. In particular, the relationship between cloth resistance and the number of filter cycles is defined by: * = RineKN

(5)

The retentivity relative to solid particles (e.g., spherical particles of polystyrene of definite size) is found from experiments determining the amount of these particles in the suspension to be filtered before and after the filter media. The retentivity K is

Filter Media and Use of Filter Aids

determined as follows: where g', g" =

45

amounts of solid particles in liquid sample before and after the medium, respectively.

The pore sizes distribution, as well as the average pore size, is determined by the "bubble" method. The filter medium to be investigated is located over a supporting device under a liquid surface that completely wets the medium material. Air is introduced to the lower surface of the medium. Its pressure is gradually increased, resulting in the formation of single chains of bubbles. This corresponds to air passages through the largest-diameter pores. As pressure is increased, the number of bubble chains increases due to air passing through the smaller pores. In many cases a critical pressure is achieved where the liquid begins to "boil." This means that the filter medium under investigation is characterized by sufficiently uniform pores. If there is no "boiling," the filter medium has pores of widely different sizes. The pore size through which air passes is calculated from known relations. For those pores whose cross section may be assumed close to a triangle, the determining size should be the diameter of a circle that may be inscribed inside the triangle. For orientation in cloth selection for a given process, the following information is essential: filtration objectives (obtaining cake, filtrate or both), and complete data (if possible) on the properties of solid particles (size, shape and density), liquid (acid, alkali or neutral, temperature, viscosity, and density), suspension (ratio of solids to liquid, particle aggregation and viscosity), and cake (specific resistance, compressibility, crystalline, friable, plastic, sticky or slimy). Also, the required capacity must be known as well as what constitutes the driving force for the process (e.g., gravity force, vacuum or pressure). Based on such information, an appropriate cloth that is resistant to chemical, thermal and mechanical aggression may be selected. In selecting a cloth based on specific mechanical properties, the process driving force and filter type must be accounted for. The filter design may determine one or more of the following characteristics of the filter cloth: tensile strength, stability in bending, stability in abrasion, and/or ability of taking the form of a filter-supporting structure. Tensile strength is important, for example, in belt filters. Bending stability is important in applications of metallic woven cloths or synthetic monofilament cloths. If the cloth is subjected to abrasion, then glass cloth cannot be used even though it has good tensile strength. From the viewpoint of accommodation to the filter-supporting structure, some cloths cannot be used, even though the filtering characteristics are excellent. For rotary drum filters, for example, the cloth is pressed onto the drum by the "caulking" method, which uses cords that pass over the drum. In this case, the closely woven cloths manufactured from monofilament polyethylene or polypropylene fiber are less desirable than more flexible cloths of polyfilament fibers or staple cloths.

46

Liquid Filtration

Depending on the type of filter device, additional requirements may be made of the cloth. For example, in a plate-and-frame press, the sealing properties of cloths are very important. In this case, synthetic cloths are more applicable staple cloths, followed by polyfilament and monofilament cloths. In leaf filters operating under vacuum and pressure, the cloth is pulled up onto rigid frames. Since the size of a cloth changes when in contact with the suspension, it should be pretreated to minimize shrinkage. In selecting cloths made from synthetic materials, one must account for the fact that staple cloths provide a good retentivity of solid particles due to the short hairs on their surface. However, cake removal is often difficult from these cloths - more than from cloths of polyfilament and, especially, monofilament fibers. The type of fiber weave and pore size determine the degree of retentivity and permeability. The objective of the process, and the properties of particles, suspension and cake should be accounted for. The cloth selected in this manner should be confirmed or corrected by laboratory tests. Such tests can be performed on a single filter. These tests, however, provide no information on progressive pore plugging and cloth wear. However, they do provide indications of expected filtrate pureness, capacity and final cake wetness. A single-plate filter consists of a hollow flat plate, one side of which is covered by cloth. The unit is connected to a vacuum source and submerged into the suspension (filtration), then suspended in air to remove filtrate, or irrigated by a dispersed liquid (washing). The filter cloth is directed downward or upward or located vertically, depending on the type of filter that is being modeled in the study. The following is a recommended sequence of tests that can assist in cloth selection for continuous vacuum filters. If the cycle consists of only two operations (filtration and dewatering), tests should be conducted to determine the suspension weight concentration after 60 sec of filtration and 120 sec of dewatering. The cake thickness should be measured and the cake should be removed to determine the weight of wet cake and the amount of liquid in it. The weight of filtrate and its purity are also determined. If the cake is poorly removed by the device, it is advisable to increase the dewatering time, yacuum or both. If the cake is poorly removed after an operating regime change, it should be tested with another cloth. If the cake is removed satisfactorily, filtration time should be decreased under increased or decreased vacuum. Note that compressible cakes sometimes plug pores faster at higher vacuum. After the filtration test for a certain filter cycle (which is based on the type of the filter being modeled), the suspension's properties should be examined. Based on the assumed cycle, a new filtration test should be conducted and the characteristics of the process noted. Capacity (N/m2-hr), filtration rate (nrVm2hr) and cake wetness can then be evaluated. Also, if possible, the air rate and dewatering time should be computed. The results of the first two or three tests should not be taken into consideration because they cannot exactly characterize the properties of the cloth. A minimum of four or five tests is generally needed to achieve reproducible results of the filtration rate and cake wetness to within 3-5 %.

Filter Media and Use of Filter Aids

47

When the cycle consists of filtration, washing and dewatering, the tests are considered principally in the same manner. The economic aspects of cloth selection should be considered after complete determination of cloth characteristics.

Introduction to the Use of Filter Aids Filter aids and/or flocculants are employed to improve the filtration characteristics of hard-to-filter suspensions. A filter aid is a finely divided solid material, consisting of hard, strong particles that are, en masse, incompressible. The most common filter aids applied as an admix to the suspension, are diatomaceous earth, expanded perlite, Solkafloc, fly ash or carbon. Filter aids build up a porous, permeable and rigid lattice structure that retains solid particles and allows the liquid to pass through. These materials are applied in small quantities in clarification or in cases where compressible solids have the potential to foul the filter medium. Filter aids may be applied in one of two ways. The first method involves the use of a precoat filter aid, which can be applied as a thin layer over the filter before the suspension is pumped to the apparatus. A precoat prevents fine suspension particles from becoming so entangled in the filter medium that its resistance becomes excessive. Further, it facilitates the removal of cake at the end of the filtration cycle. The second application method involves incorporation of a certain amount of the material with the suspension before introducing it to the filter. The addition of filter aids increases the porosity of the sludge, decreases its compressibility, and reduces the resistance of the cake. In some cases the filter aid displays an adsorption action, which results in particle separation of sizes down to 0.1 /mi. The adsorption ability of certain filter aids, such as bleached earth and activated charcoals, is manifest by a decoloring of the suspension's liquid phase. This practice is widely used for treating fats and oils. Filter aids are also used in processing sugar, beer, wine, gelatin, antibiotics, glycerol, solvents, synthetic tars, caustic, sulfur and uranium salts, for water cleaning, and in preparing galvanic solutions. The properties of these additives are determined by the characteristics of their individual components. For any filter aid, size distribution and the optimal dosage are of great importance. Too low a dosage results in poor clarity; too great a dosage will result in the formation of very thick cakes. In general, a good filter aid should form a cake having high porosity (typically 0.85-0.9), low surface area and good particle-size distribution. An acceptable filter aid should have a much lower filtration resistance than the material with which it is being mixed. It should reduce the filtration resistance by 67-75 % with the addition of no more than 25 % by weight of filter aid as a fraction of the total solids. The addition of only a small amount of filter aid (e.g., 5% of the sludge solids) can actually cause an increase in the filtration resistance. When the amount of filter aid is

48

Liquid Filtration

so small that the particles do not interact, they form a coherent structure, and resistance may be affected adversely. filter Aid Requirements

Filter aids are assessed in terms of the rate of filtration and clarity of filtrate. Finely dispersed filter aids are capable of producing clear filtrate; however, they contribute significantly to the specific resistance of the medium. As such, their application must be made in small doses. Filter aids comprised of coarse particles contribute considerably less specific resistance; consequently, a high filtration rate can be achieved with their use. Their disadvantage is that a "muddy" filtrate is produced. The optimum filter aid should have maximum pore size and ensure a prespecified filtrate clarity. Desirable properties/characteristics for the optimum filter aid include: •

The additive should provide a thin layer of solids having high porosity (0.85-0.90) over the factor medium's external surface. Suspension particles will ideally form a layered cake over the filter aid cake layer. The high porosity or the filter aid layer will ensure a high filtration rate. Porosity is not determined by pore size alone. High porosity is still possible with small size pores,

•

Filter aids should have low specific surface, since hydraulic resistance results from frictional losses incurred as liquid flows past particle surfaces. Specific surface is inversely proportional to particle size. The rate of particle dispersity and the subsequent difference in specific surface determines die deviations in filter aid quality from one material to another. For example, most of the diatomite species have approximately the same porosity; however, the coarser materials experience a smaller hydraulic resistance and have much less specific surface than the finer particle sizes.

•

Filter aids should have a narrow fractional composition. Fine particles increase the hydraulic resistance of the filter aid, whereas, coarse particles exhibit poor separation. Desired particle size distributions are normally prepared by air classification, in which the finer size fractions are removed.

•

In applications where the filter aid layer is to be formed on open-weave synthetic fabric or wire screens, wider size distributions may have to be prepared during operation. Filter aids should therefore have the flexibility to be doped with amounts of coarser sizes. This provides rapid particle bridging and settling or the filter aid layer. For example, diatomite having an average particle size of 8 /urn may be readily applied to a screen with a mesh size of 175 ^m by simply adding a small quantity of filter aid with sizes that are on the same order, but less in size than the mesh openings. Particle sizes typically around 100 jum will readily form bridges over the screen openings and prevent the loss of filter aid in this example.

•

The filter aid should be chemically inert to the liquid phase of the suspension and not decompose or disintegrate in it.

The ability of an admix to be retained on the filter medium depends on both the suspension's concentration and the filtration rate during this initial precoat stage. The same relationships for porosity and the specific resistance of the cake as functions of suspension concentration and filtration rate presented in Chapter 3 apply equally to filter aid applications. Applications Filter aids are added in amounts needed for a suspension to acquire desirable filtering properties and to prepare a homogeneous suspension before actual filtration. Filter aids increase the concentration of solids in the feed suspension. This promotes particle

Filter Media and Use of Filter Aids

49

bridging and creates a rigid lattice structure for the cake. In addition, they decrease the floe deformation tendency. Irregular or angular particles tend to have better bridging characteristics than spherical particles. As a rule, the weight of aid added to the suspension should equal the particle weight in suspension. Typical filter aid additions are in the range of 0.01-4% by weight of suspension; however, the exact amount can only be determined from experiments. Excess amounts of filter aid will decrease the filter rate. Operations based on the addition of admixes to their suspensions may be described by the general equations of filtration with cake formation. A plot of filtration time versus filtrate volume on rectangular coordinates results in a nearly parabolic curve passing through the origin. The same plot on logarithmic coordinates, assuming that the medium resistance may be neglected, results in a straight line. This convenient linear relationship allows results obtained from short-time filtration tests to be extrapolated to long-term operating performance (i.e., for several hours of operation). This reduces the need to make frequent, lengthy tests and saves time in the filter selection process. In precoating, the prime objective is to prevent the filter medium from fouling. The volume of initial precoat normally applied should be 25-50 times greater than that necessary to fill the filter and connecting lines. This amounts to about 5-10 lb/100 ft2 of filter area, which typically results in a V 16 - to ! / g -in, precoat layer over the outer surface of the filter medium. An exception to this rule is in the precoating of continuous rotary drum filters where a 2- to 4-in. cake is deposited before filtration. The recommended application method is to mix the precoat material with clear liquor (which may consist of a portion of the filtrate). This mixture should be recycled until all the precoat has been deposited onto the filter medium. The unfiltered liquor follows through immediately without draining off excess filter aid liquor. This operation continues until a predetermined head loss develops, when the filter is shut down for cleaning and a new cycle. In precoating, regardless of whether the objective is to prevent filter medium clogging or to hold back fines from passing through the medium to contaminate the filtrate, the mechanical function of the precoat is to behave as the actual filter medium. Since it is composed of incompressible, irregularly shaped particles, a high-porosity layer is formed within itself, unless it is impregnated during operation with foreign compressible materials. Ideally, a uniform layer of precoat should be formed on the surface of the filter medium. However, a nonuniform layer of precoat often occurs due to uneven medium resistance or fluctuations in the feed rate of filter aid suspension. Cracks can form on the precoat layer that will allow suspension particles to penetrate into the medium. To prevent cracking, the filter aid may be applied as a compact layer. On a rotating drum filter, for example, this may be accomplished by applying a low concentration of filter aid (2-4%) at the maximum drum rpm. In other filter systems, maintaining a low pressure difference during the initial stages of precoating and then gradually increasing it with increasing layer thickness until the start of filtration will help to minimize cake cracking. Also, with some filter aids (such as diatomite or perlite), the addition of small amounts of fibrous material will produce a more compact precoat cake. At low suspension concentrations (typically 0.01%), filter aids serve as a medium under conditions of gradual pore blocking. In this case the amount of precoat is 10-25 N/m2 of the medium and its thickness is typically 3-10

50

Liquid Filtration

mm. In such cases, the filter aid chosen should have sufficient pore size to allow suspension particle penetration and retention within the precoat layer.

Examples of Filter Aids The following are examples of some of the most common filter aids. Diatomite The most important filter aids from a volume standpoint are the diatomaceous silica type (90% or better silica). These are manufactured from the siliceous fossil remains of tiny marine plants known as diatoms. Diatomaceous filter aids are available in various grades. This is possible because the natural product can be modified by calcining and processing, and because filter aids in different size ranges and size distributions have different properties. The filter aids may be classified on the basis of cake permeability to water and water flowrate. Finer grades are the slower-filtering products; however, they provide better clarification than do faster-filtering grades. Thus a fast-filtering aid may not provide the required clarification. However, by changing the physical character of the impurities (e.g., by proper coagulation), the same clarity may be obtained by using the fast-filtering grades. Calcinated diatomaceous additives are characterized by their high retention ability with relatively low hydraulic resistance. Calcining dramatically affects the physical and chemical properties of diatomite, making it heat-resistant and practically insoluble in strong acids. Perlite Perlite is glasslike volcanic rock, called "volcanic glass," consisting of small particles with cracks that retain 2-4% water and gas. Natural perlite is transformed to a filter aid by heating it to its melting temperature (about 1000°C), where it acquires plastic properties and expands due to the emission of steam and gas. Under these conditions its volume increases by a factor of 20. Beads of the material containing a large number of cells are formed. The processed material is then crushed and classified to provide different grades. The porosity of perlite is 0.85-0.9 and its volumetric weight is 500-1000 N/m3. Compared to diatomite, perlite has a smaller specific weight and comparable filter applications typically require 30% less additive. Perlite is used for filtering glucose solutions, sugar, pharmaceutical substances, natural oils, petroleum products, industrial waters and beverages. The principal advantage of perlite over diatomite is its relative purity. There is a danger that diatomite may foul filtering liquids with dissolved salts and colloidal clays.

Filter Media and Use of Filter Aids

51

Cellulose Cellulose fiber is applied to cover metallic cloths. The fibers form a highly compressed cake with good permeability for liquids, but a smaller retention ability for solid particles than that of diatomite or perlite. The use of cellulose is recommended only in cakes where its specific properties are required. These properties include a lack of ashes and good resistance to alkalies. The cost of cellulose is higher than those of diatomite and perlite. Sawdust This filter aid may be employed in cases where the suspension particles consist of a valuable product that may be roasted. For example, titanium dioxide can be manufactured by calcining a mixture of sawdust and metal titanium acid. The mixture is obtained as a filter cake after separating the corresponding suspension with a layer of filter aid. Charcoal Charcoal is not only employed in activated form for decoloring and adsorbing dissolved admixtures, but also in its unactivated form as a filter aid. It can be used in suspensions consisting of aggressive liquids (e.g., strong acids and alkalies). As with sawdust, it can be used to separate solids that may be roasted. On combustion, the charcoal leaves a residue of roughly 2% ash. Particles of charcoal are porous and form cakes of high density but that have a lesser retention ability than does diatomite. Fly Ash

This material has a number of industrial filtering applications, but primarily is applied to dewatering sewage sludge. The precoat is built up to 2 in. thick from a 60% solid slurry. On untreated sludges, filtration rates of 25 Ib/ft2-hr or higher are obtainable. This rate can be more than doubled with treated sludges. The sludge is reduced from a liquid to a semidry state. Fly ash may also be used as a precoat in the treatment of papermill sludge.

Filter Aid Selection Filter aid selection should be based on laboratory tests. Guidelines for selection may only be applied in the broadest sense, since there is almost an infinite number of combinations of filter media, filter aids and suspensions that will produce varying degrees of separation. The hydrodynamics of any filtration process are highly complex. Filtration is essentially a multiphase system in which interaction takes place between solids from the suspension, filter aid, filter medium, and a liquid phase. Experiments are needed in most applications not only in proper filter aid selection but in defining the method of their administration. Some general guidelines can be applied to such evaluations:

52

Liquid Filtration

•

The filter aid must have the minimum hydraulic resistance and provide the desired rate of separation,

•

An insufficient amount of filter aid leads to a reduction in filtrate quality. Excess amounts result in losses in filtration rate.

•

It is necessary to account for the method of application and characteristics of filter aids.

For precoat applications to rotary filters, the relative usefulness of various filter aids requires experimental data on the effects on the following variables: 1. type and grade of filter aid used as the precoat, 2. drum speed, 3. knife advance rate, 4. vacuum pressure, 5. filter medium, 6. temperature of feed, 7. filter medium submergence, 8. manner of precoat application, 9. sharpness of knife, and 10. knife angle and bevel.

Tests should be conducted under a constant set of operating conditions such as pressure and temperature. Such tests will reveal that the manner in which filter aid is removed depends on its properties and the drum speed. In some cases the filter aid will be removed as flakes, while other tests will show that the remaining portion of the cake will acquire a rough surface. High drum speed and filter aid dispersity are known to enhance the quality of the precoat layer. For a low rate of knife advance the resistance of the filter aid decreases in spite of the decrease in cake thickness. This is explained by the fact that for a single rotation of the drum, the particles of the suspension penetrate into the pores of filter at depths greater than 1.5 mm, and blockage occurs. Rotary drum filler systems are characterized not only by rate, but by the ratio of filtrate volume to the weight of filter aid cut. Some generalizations for this system are: 1.

The greater the permeability of the filter aid, the greater the amounts of filtrate collected.

2.

A gradual decrease in filtration rate can be observed; however, at some point during the operation a large reduction in rate can also be observed.

3.

In the operating regime characterized by a sharp drop in the filtration rate, the permeability of the filter aid is not zero, but has a definite value.

4.

Two operating periods lead to a reduction in permeability. During the first period, filtration primarily proceeds with pore blocking on the outer surface of the filter aid. During the second period, the filter aid pores undergo blocking within the filter aid cake.

Figure 2 illustrates the operation of a precoat filter with cake in place. Point A represents the entrance of the cake into the unfiltered suspension. Point B represents the point of emergence and C is the point at which the cake, together with its accumulated suspension particles, reaches minimum permeability. Maximum filtration

Filter Media and Use of Filter Aids

Figure 2.

Cross-sectional view of a precoat filter.

economy can be achieved by proper combination of submergence depth and rate of rotation so that points B and C coincide. If minimum permeability is not achieved before point C reaches location B, the drum is rotating too rapidly or the submergence is too low, resulting in a portion of the filter aid capacity being wasted. Conversely, if minimum permeability is reached before C reaches B, either the submergence level could be lowered or the drums rotated at a faster rate. Choice of decreasing the submergence level or increasing the rate of rotation would be dictated by the time requirement for the drainage of accumulated filtrate from the filter cake. In accordance with this thinking, if points B and C coincide with 50% submergence and 10 rpm drum speed, a reduction in drum speed to 0.67 rpm would move point C back to C ls a further reduction in drum speed to 0.5 rpm would locate it at C2. If the blinding medium has zero effective permeability, filtrate collection under these circumstances should have a linear relationship with respect to drum speed. However, if the blinding medium has appreciable permeability, additional filtrate will be collected in proportion to that permeability through the distances BC, and BC2 for 0.67- and 0.5-rpm speeds, respectively. A graphical method is available for determining filter aid efficiency. Figure 3 shows a plot of filtrate rate versus cake thickness. Curve AB corresponds to the filtration of clear liquid through a gradually decreasing layer of filter aid when the resistance of the filter medium is negligible. The curve is based on the basic law of filtration using a special permeability for the filter aid. The line EF corresponds to suspension separation through a gradually decreasing layer of filter aid and filter medium of known resistance. Areas ABCD and EFCD are proportional to the filtrate volumes in an ideal and actual process, respectively. Since the areas under the curves are proportional to total output of the cycle, the observed output (CDEF) is much lower than that theoretically obtainable with clear liquid (ABCD). If we arbitrarily call the filter aid 100% rate-efficient when operating with solids-free liquid under ideal conditions, then the rate efficiency with the slurry involved may be calculated as follows:

54

Liquid Filtration

O 3 O -I IL it, O

in *• < K

Figure 3. Relationship between blinding losses and theoretical efficiency.

CDEF area CDEF xlOO = xlOO ABCD area ABCD

The expression for curve AB in terms of a and b is:

e =o 4

(8)

Hence, area ABCD may be computed by integration of the curve's formula. An alternative approach is to draw the equivalent line AB1, so that the cross-hatched areas are equal, from whence the rate efficiency is:

r, = r

AB'CD

The following approximate procedure may also be used. Note in Figure 3 mat the center of the cake is 1.25 in. Areas AB'CD and CDEF can be computed from the product of the intercept of the diagonal line AB1 or EF and base CD.

Filter Media and Use of Filter Aids

MOxCD

55

MO

The location of line AB can be determined for any filter aid from its permeability. During the period when the knife performs a full cut, the filtration rate is at a maximum. During the inactive periods of the knife, the filtration rate is reduced to a level representative of the permeability of the blinding layer. In Figure 3, this results in a filtration rate curve approximately parallel to the theoretical curve and displaced downward. As shown, these curves are nearly linear. This supports the assumption that the resistance of the filter aid cake does not significantly restrict filtrate flow through the filter medium. Note that the total area ABCD theoretically represents 100% filter aid rate efficiency and the area under observed curve EF is the performance attained. Then, the area above the observed curve denotes the loss of rate efficiency due to blinding. Blinding and penetration are rarely separate effects. Figure 4 illustrates a graphic evaluation of the rate efficiency together with applicable permeability corrections for the blinding medium. The total area under theoretical curve AB1 represents 100% rate efficiency. That area under the actual line EF represents the filtration rate efficiency of the filter aid alone. The middle region denotes the loss of filter aid capacity due to blinding (i.e., pore blocking). The top triangular region denotes the loss of capacity of the filter aid due to penetration of solids into the cake beyond the cutting depth of the knife. The average filtration rate and filtrate volume produced per unit weight of filter aid spent may be determined from Figure 4. The combined influence of both factors on the filtration efficiency is different depending on actual operating conditions. The penetration of fine solids through the filter aid results in larger pores and consequently in smaller specific resistances, which are less applicable than those with pores of smaller size where the fine particles do not penetrate. This can be explained by the fact that the specific resistance of the first filter aid where particles penetrated becomes higher than that of the second filter aid. The efficiency of a rotary drum filter is expressed by the product of the average filtration rate and the ratio of filtrate volume to weight of filter aid spent. Efficiency reaches a maximum at different rates of knife motion and varies according to the properties of the specific filter aid. An increase in the thickness of a cake layer cut results in a ratio of filtrate volume to weight of filter aid spent. One should account for the volume weight of different filter aids since this parameter varies considerably even among different grades of the same material.

56

Liquid Filtration

To maintain an acceptable filtration rate, the following general guidelines apply: 1.

If suspension particles are observed to intensely penetrate the filter aid cake, another grade with better solids retention properties should be selected.

2,

If blinding of the outer surface of the filter aid is observed, then the rate of rotation should be increased.

2.0 1.5 1.0 0.5 Septum CAKE THICKNESS, L Figure 4. Graphical evaluation of filter aid efficiency and blinding and penetration losses,

After preliminary selection of the filter aid type, the optimum amount of addition to a suspension and the preferred rate of filtration must be established. Final selection should be based on comparison of the suspension separation efficiency for different filter aids and filter aid grades. Some general industry guidelines for filter aid precoating as applied to batch filters with a flat filter medium operating under pressure are given below. Examples of specific filter types that fall into this category are filter presses and leaf filters. It is generally recommended that a suspension of filter aid be prepared in a pure liquid by agitation before filtration. Delivery to the filter medium can usually be done with a centrifugal pump. The filtrate can be recycled to the filter aid suspension feed until the finished precoat is formed. Depending on the ratio of particle size of the filter aid to the pore diameters of the medium, as well as the rate of suspension feed and its concentration, two processes arise on the filter. First, the medium retains only some particles and the remainder pass through with filtrate. Gradually, the medium's

Filter Media and Use of Filter Aids

57

retention ability increases and the amount of solid particles in the filtrate decreases. A short time after starting the process, the retention ability of the medium becomes so pronounced (for example, due to bridging) that the pure filtrate passes through, In both cases, the suspension concentration in the filter body decreases gradually. The suspension of filter aid is usually delivered on the filter not filled with liquid; however, in some cases the filter is filled with pure liquid during the precoat stage, in such cases, the filter and suspension entering the unit is diluted. Its concentration first increases, reaching a maximum value, then as the particles settle onto the filter, it decreases. In most cases we may assume that the filtrate volume is equal to the suspension volume (the concentration of filter aid in suspension is typically 0.1-0.5%). The cake may be assumed incompressible. Its specific resistance is independent of concentration. Also, within the filter body, the suspension can be assumed to be perfectly mixed, with filtration at a constant rate. The following expressions describe the case where the suspension of filter aid is delivered to a filter not filled with liquid. For filter aid particles completely retained by the medium: C = C0e-»

For the case where the retention ability of the medium increases gradually: nlk V"

v«"r>

cm A — T

(12)

where C = variable concentration of suspension in the filter body (mVm3) C0 = initial concentration of suspension in the mixer (mVm3) n = V/V S = degree of circulation of suspension in the system V = filtrate volume (m3) Vs = initial volume of suspension in the mixer (m3) K= constant A,= Q/VS = coefficient of dissolution (sec"1) Q = amount of filtered liquid (m3/sec)

Suggested Readings 1. 2.

Masschelein, W.S., Unit Process in Drinking Water Treatment, Marcel Decker, Inc., NJ, 1992 Veshilind, P. A, Treatment and Disposal of Wastewater Sludges, Ann Arbor Science, Ann Arbor, MI, 1975.

58

3. 4.

Liquid Filtration

Noyes, R., Handbook of Pollution Control Processes, Noyes Publications, NJ, 1991 Cheremisinoff, N.P. and D.S. Azbel, Liquid Filtration for Process and Pollution Control, SciTech Publishers, Inc, NJ.1989

Nomenclature

a b c c()

= = = =

g', g" = K, K = L = N = p = Q = r0 = R = V =

addition of filter aid (kg) coefficient concentration of filtration-impeding impurities (kg/m3) initial concentration of suspension in the filter body (kg/m3 or m3/m3) amount of solid particles in liquid before and after the medium, respectively (kg/m ) emperical constants or retentivity cake thickness (in.) degree of circulation of suspension in the system pressure (N/m3) volumetric flowrate (m3/sec) specific cake resistance (m"1) total filtration resistance (m/sec) filtrate volume (m3)

Greek Symbols

rjr 1 jU i

= = = =

filtration rate efficiency (%) coefficient of dissolution (sec"1) viscosity (P) time (sec)

3 CAKE FILTRATION AND FILTER MEDIA FILTRATION Introduction Cake filtration is the most common form of filtering employed by the chemical and process industries. This manifests into handling the permeability of a bed of porous material, the schematic of which is shown in Figure 1, With high-solids-concentration suspensions, even relatively small particles (in comparison to the pore size) will not pass through the medium, but tend to remain on the filter surface, forming "bridges" over individual openings in the filter material. Filtrate flows through the filter medium and cake because of an applied pressure, the magnitude of which is proportional to the filtration resistance. This resistance results from the frictional drag on the liquid as it passes through the filter and cake. Hydrostatic pressure varies from a maximum at the point where liquid enters the cake, to zero where liquid is expelled from the medium; consequently, at any point in the cake the two are complementary. That is, the sum of the hydrostatic and compression pressures on the solids always equals the total hydrostatic pressure at the face of the cake. Thus, the compression pressure acting on the solids varies from zero at the face of the cake to a maximum at the filter medium. When solid particles undergo separation from the mother suspension, they are captured both on the surface of the filter medium and within the inner pore passages. The penetration of solid particles into the filter medium increases the flow resistance until the filtration cycle can no longer continue at economical throughput rates, at which time the medium itself must be replaced. This chapter provides a summary of standard calculation methods for assessing cake formation, behavior, and the overall efficiency of the filter-medium filtration process.

60

Liquid Filtration

Dynamics of Cake Filtration When the space above the suspension is subjected to a source of compressed gas (e.g., air) or the space under the filter plate is connected to a vacuum source, filtration is accomplished under a constant pressure differential (the pressure in the receivers is constant). In this case, the rate of the process decreases due to an increase in the cake thickness and, consequently, flow resistance. A similar filtration process results from a pressure difference due to the hydrostatic pressure of a suspension layer of constant thickness located over the filter medium. If the suspension is fed to the filter with a reciprocating pump at constant capacity, filtration is performed under constant flowrate. In this case, the pressure differential increases due to an increase in the cake resistance. If the suspension is fed by a centrifugal pump, its capacity decreases with an increase in cake resistance, and filtration is performed at variable pressure differentials and flowrates. The most favorable filtration operation with cake formation is a process whereby no clogging of the filter medium occurs. Such a process is observed at sufficiently high concentrations of solid particles in suspension. From a practical standpoint this concentration may conditionally be assumed to be in excess of 1 % by volume.

Figure 1. Operating scheme of a filtration process; 1-filter; 2-filter medium; 3-suspension; 4-filtrate; 5-cake.

To prevent pore clogging in the filter medium when handling relatively low solids concentrations (e.g., 0.1-1% by volume), general practice is to increase the solids concentration in thickeners before the suspension is fed to the filter. Filtration is frequently accompanied by hindered or free gravitational settling of solid particles. The relative directions of action between gravity force and filtrate motion

Cake Filtration and Filter Media Filtration

61

may be concurrent, countercurrent or crosscurrent, depending on the orientation of the filter plate, as well as the sludge location above or below the filter plate. The different orientations of gravity force and filtrate motion with their corresponding distribution of cake, suspension, filtrate and clear liquid are illustrated in Figure 2. Particle sedimentation complicates the filtration process and influences the controlling mechanisms. Furthermore, these influences vary depending on the relative directions of gravity force and filtrate motion. If the suspension is above the filter medium (Figure 2A), particle settling leads to more rapid cake formation with a clear filtrate, which can be evacuated from the filter by decanting. If the suspension is under the filter medium (Figure 2B), particle settling will prevent cake formation, and it is necessary to mix the suspension to maintain homogeneity.

r\ .:*

2 3 (A)

(B)

Figure 2. Direction of gravity force action and filtrate motion in filters: A-cocurrent; Bcountercurrent; C-crosscurrent; solid arrow-direction of gravity force action; dashed arrowdirection of filtrate motion; 1 -filter plate; 2-cake; 3-sludge; 4-filtrate; 5-clear liquid.

When the cake structure is composed of particles that are readily deformed or become rearranged under pressure, the resulting cake is characterized as being compressible. Those that are not readily deformed are referred to as semicompressible, and those that deform only slightly are considered incompressible. Porosity (defined as the ratio of pore volume to the volume of cake) does not decrease with increasing pressure drop. The porosity of a compressible cake decreases under pressure, and its hydraulic resistance to the flow of the liquid phase increases with an increase in the pressure differential across the filter media. Cakes containing particles of inorganic substances with sizes in excess of 100 jum may be considered incompressible, for all practical purposes. Examples of incompressible cake-forming materials are sand and crystals of carbonates of calcium and sodium. The cakes containing particles of metal hydroxides, such as ferric hydroxide, cupric hydroxide, aluminum hydroxide, and sediments consisting of easy deforming aggregates, which are formed from primary fine crystals, are usually compressible.

62

Liquid Filtration

At the completion of cake formation, treatment of the cake depends on the specific filtration objectives. For example, the cake itself may have no value, whereas the filtrate may. Depending on the disposal method and the properties of the particulates, the cake may be discarded in a dry form, or as a slurry. In both cases, the cake is usually subjected to washing, either immediately after its formation, or after a period of drying. In some cases, a second washing is required, followed by a drying period where all possible filtrate must be removed from the cake; or where wet discharge is followed by disposal: or where repulping and a second filtration occurs; or where dry cake disposal is preferable. Similar treatment options are employed in cases where the cake is valuable and all contaminating liquors must be removed, or where both cake and filtrate are valuable. In the latter, cake-forming filtration is employed, without washing, to dewater cakes where a valueless, noncontaminating liquor forms the residual suspension in the cake. To understand the dynamics of the filtration process, a conceptual analysis is applied in two parts. The first half considers the mechanism of flow within the cake, while the second examines the external conditions imposed on the cake and pumping system, which brings the results of the analysis of internal flow in accordance with the externally imposed conditions throughout. The characteristics of the pump relate the applied pressure on the cake to the flowrate at the exit face of the filter medium. The cake resistance determines the pressure drop. During filtration, liquid flows through the porous filter cake in the direction of decreasing hydraulic pressure gradient. The porosity (e) is at a minimum at the point of contact between the cake and filter plate (i.e., where x = 0) and at a maximum at the cake surface (x = L) where sludge enters. A schematic definition of this system is illustrated in Figure 3.

SUBSCRIPT

SLURRY

Figure 3. Important parameters in cake formation.

The drag that is imposed on each particle is transmitted to adjacent particles. Therefore, the net solid compressive pressure increases as the filter plate is approached, resulting in a decrease in porosity. Referring to Figure 4A, it may be

Cake Filtration and Filter Media Filtration

63

assumed that particles are in contact at one point only on their surface, and that liquid completely surrounds each particle. Hence, the liquid pressure acts uniformly in a direction along a plane perpendicular to the direction of flow. As the liquid flows past each particle, the integral of the normal component of force leads to form drag, and the integration of the tangential components results in frictional drag. If the particles are non-spherical, we may still assume single-point contacts between adjacent particles as shown in Figure 4B. Now consider flow through a cake (Figure 4C) with the membrane located at a distance x from the filter plate. Neglecting all forces in the cake other than those created by drag and hydraulic pressure, a force balance from x to L gives: Fs + ApL = Ap

(1)

The applied pressure p is a function of time but not of distance x. Fs is the cumulative drag on the particles, increasing in the direction from x = L to x = 0. Since single point contact is assumed, the hydraulic pressure pL is effectively over the entire cross section (A) of the cake; for example, against the fictitious membrane shown in Figure 4B. Dividing Equation 1 by A and denoting the compressive drag pressure by ps = F/A, we obtain:

MEDIUM

SURFACE SLURRY

(BJ

(C )

Figure 4. Frictional drag on particles in compressible cakes.

The term ps is a fictitious pressure, because the cross-sectional area A is not equal to either the surface area of the particles nor the actual contact areas In actual cakes, there is a small area of contact Ac whereby the pressure exerted on the solids may be defined as FS/AC. Taking differentials with respect to x, in the interior of the cake, we obtain:

dp, + dp, = 0

(3)

64

Liquid Filtration

This expression implies that drag pressure increases and hydraulic pressure decreases as fluid moves from the cake's outer surface toward the filter plate. From Darcy's law, the hydraulic pressure gradient is linear through the cake if the porosity (e) and specific resistance (a) are constant. The cake may then be considered incompressible. This is illustrated by the straight line obtained from a plot of flowrate per unit filter area versus pressure drop shown in Figure 5. The variations in porosity and specific resistance are accompanied by varying degrees of compressibility, also shown in Figure 5. As noted in Chapter 1, filtration is primarily an application of fluid mechanics; that is, filtrate flow is induced through a porous filter cake and filter medium. The rate of the filtration process is directly proportional to the driving force and inversely proportional to the resistance.

INCOMPRESSIBLE

NORMALLY COMPRESSIBLE

w (T

//

^^ Affe ^X^ n ft ,-",..%.

i-n-fl

X LU h

I

'

HIGHLY

COMPRESSI8LE

qct

PRESSURE

DROP,

Apc

Figure 5. Flowrate/area versus pressure drop across the cake.

Because pore sizes in the cake and filter medium are small, and the liquid velocity through the pores is low, the filtrate flow may be considered laminar: hence, Poiseuille's law is applicable. Filtration rate is directly proportional to the difference in pressure and inversely proportional to the fluid viscosity and to the hydraulic resistance of the cake and filter medium. Because the pressure and hydraulic resistances of the cake and filter medium change with time, the variable rate of filtration may be expressed as:

Adi

(4)

Cake Filtration and Filter Media Filtration

where

65

V = volume of filtrate (m3) A = filtration area (m2) T = time of filtration (sec)

Assuming laminar flow through the filter channels, the basic equation of filtration as obtained from a force balance is:

u =

where

1 dV Ap ^ A JT p(Re + Rf)

IJ

'

= pressure difference (N/m2) = viscosity of filtrate (N-sec/m2) = filter cake resistance (m"1) = initial filter resistance (resistance of filter plate and filter channels) (m4) u = filtration rate (m/sec), i.e., filtrate flow through cake and filter plate dV/dt = filtration rate (nvVsec), i.e., filtrate flow rate Ap fji, Rc Rf

Filter cake resistance (Rc) is the resistance to filtrate flow per unit area of filtration. Rc increases with increasing cake thickness during filtration. At any instant, RC depends on the mass of solids deposited on the filter plate as a result of the passage of V (m3) filtrate. Rf may be assumed a constant. To determine the relationship between volume and residence time T, Equation 5 must be integrated, which means that Rc must be expressed in terms of V. We denote the ratio of cake volume to filtrate volume as XQ. Hence, the cake volume is x0V. An alternative expression for the cake volume is h c: A; where hc is the cake height in meters. Consequently: xQV = hcA

(6)

Hence, the thickness of the cake, uniformly distributed over the filter plate, is: fa

c

— y-

V °A

/TV (1)

The filter cake resistance may be expressed as: *c = VoT

where r0= specific volumetric cake resistance (m"2).

(8)

66

Liquid Filtration

As follows from Equation 8, r0 characterizes the resistance to liquid flow by a cake having a thickness of 1 m. Substituting for Rc from Equation 8 into Equation 5, we obtain: l^dV_ = A rfi

=

A/? p[rQxQ(V/A)+Rf]

(9)

Filtrate volume, XQ can be expressed in terms of the ratio of the mass of solid particles settled on the filter plate to the filtrate volume (xw) and instead of r0, a specific mass cake resistance r w is used. That is, r w represents the resistance to flow created by a uniformly distributed cake, in the amount of 1 kg/m2. Replacing units of volume by mass, the term r0 XQ in Equation 9 changes to rwxw. Neglecting filter plate resistance (Rf = 0), and taking into account Equation 7, we obtain from Equation 3 the following expression:

At /u, = l N-sec/m2, hj. = 1 m and u = 1 m/sec, r0 = Ap. Thus, the specific cake resistance equals the pressure difference required by the liquid phase (with a viscosity of 1 N-sec/m2) to be filtered at a linear velocity of 1 m/sec through a cake 1 m thick. This hypothetical pressure difference, however, is beyond a practical range. For highly compressible cakes, r0 can exceed 1012m2. Assuming V = 0 (at the start of filtration) where there is no cake over the filter plate. Equation 9 becomes:

*, ~~ & At ju = 1 N-sec/m2 and u = 1 m/sec, Rf = Ap. This means that the filter plate resistance is equal to the pressure difference necessary for the liquid phase (with viscosity of 1 N-sec/m2) to pass through the filter plate at a rate of 1 m/sec. For many filter plates Rf is typically 10'° m"! . For a constant pressure drop and temperature filtration process all the parameters in Equation 9, except V and T, are constant. Integrating Equation 9 over the limits of 0 to V, from 0 to T, we obtain: v

%

v f ( r Jt —- + R ) dV = I bpAdi 0 0

or

A

Jf

J

(12a)

Cake Filtration and Filter Media Filtration

(12b)

2A

Dividing both sides by jur0Xo/2A gives:

V2 + 2

RfA

V =

(13)

Equation 13 is the relationship between filtration time and filtrate volume. The expression is applicable to either incompressible or compressible cakes, since at constant Ap, r0 and XQ are constant. If we assume a definite filtering apparatus and set up a constant temperature and filtration pressure, then the values of Rf, r0, p. and Ap will be constant. The terms in parentheses in Equation 13 are known as the "filtration constants", and are often lumped together as parameters K and C; where:

RA (15)

Hence, a simplified expression may be written to describe the filtration process as follows:

(16)

Filtration constants K and C can be experimentally determined, from which the volume of filtrate obtained over a specified time interval (for a certain filter, at the same pressure and temperature) can be computed. If process parameters are changed, new constants K and C can be estimated from Equations 14 and 15. Equation 16 may be further simplified by denoting TO as a constant that depends on KandC:

Liquid Filtration

K

(17)

Substituting TO into Equation 16, the equation of filtration under constant pressure conditions is: (V+Cf

=

(18)

Equation 18 defines a parabolic relationship between filtrate volume and time. The expression is valid for any type of cake (i.e., compressible and incompressible). From a plot of V + C versus (T+IO), the filtration process may be represented by a parabola with its apex at the origin as illustrated in Figure 6. Moving the axes to distances C and TO provides the characteristic filtration curve for the system in terms of volume versus time. Because the parabola's apex is not located at the origin of this new system, it is clear why the filtration rate at the beginning of the process will have a finite value, which corresponds to actual practice.

Figure 6. Typical filtration curve.

Constants C and TO in Equation 18 have physical interpretations. They are basically equivalent to a fictitious layer of cake having equal resistance. The formation of this fictitious cake follows the same parabolic relationship, where TO denotes the time required for the formation of this fictitious mass, and C is the volume of filtrate required. Differentiating Equation 16 gives:

Cake Filtration and Filter Media Filtration

dV

K 2(V+C)

(19)

And rearranging in the form of a reciprocal relationship: d^ _ 2V dV ~ K

+ +

2C K~

(20)

This form of the equation provides a linear relation as shown by the plot in Figure 7. The expression is that of a straight line having slope 2/K, with intercept C. The experimental determination of dt/dV is made simple by the functional form of this expression. Filtrate volumes V, and V2 should be measured for time intervals T, and T 2 , Then, according to Equation 16:

2C(V2-V,)

VV2 -VM

V -V V

v

K

K (21) 2C K

In examining the right side of this expression, we note that the quotient is equal to the inverse value of the rate at the moment of obtaining the filtrate volume, which is equal to the mean arithmetic value of volumes V, and V 2 :

V - VM V 2

dV

Filtration constants C and K can be determined on the basis of several measurements of filtrate volumes for different time intervals. As follows from Equations 14 and 15, values of C and K depend on r0 (specific volumetric cake resistance), which in turn depends on the pressure drop across the cake. This Ap, especially during the initial stages of filtration, undergoes changes in the cake. When the cake is very thin, the main portion of the total pressure drop is exerted on the filter medium. As the cake becomes thicker, the pressure drop through the cake increases rapidly but then levels off to a constant value. Isobaric filtration shows insignificant deviation from Equation 16. For approximate calculations, it is

70

Liquid Filtration

Figure 7. Plot of Equation 20.

possible to neglect the resistance of the filter plate, provided the cake is not too thin. Then the filter plate resistance Rf = 0 in Equation 15, C= 0 (Equation 15) and TO = 0 (Equation 17). Therefore, the simplified equation of filtration takes the following form: V2 = K-c

(23)

For thick cakes, Equation 23 gives results close to that of Equation 16. Constant-Rate Filtration

When sludge is fed to a filter by a positive-displacement pump, the rate of filtration is nearly constant (i.e., dV/dt = constant). During constant-rate filtration, the pressure increases with an increase in cake thickness. Therefore, the principal variables are pressure and filtrate volume, or pressure and filtration time. Equation 9 is the principal design relation, which may be integrated for a constant-rate process. The derivative, dV/dT, may be replaced simply by V/T: Ap

pR,

At

(24)

The ratios in parentheses express the constant volume rate per unit filter area. Hence, Equation 24 is the relationship between time T and pressure drop Ap. For incompressible cakes, r0 is constant and independent of pressure. For compressible cakes, the relationship between time and pressure at constant-rate filtration is:

Cake Filtration and Filter Media Filtration

(25)

Filtration experiments are typically conducted in pilot scale equipment and generally tests are conducted either at constant pressure or constant rate to determine axo, as well as s and R f, for a given sludge and filter medium. Such tests provide empirical information that will enable the time required tor the pressure drop to reach the desired level for a specified set of operating conditions to be determined. In the initial stages of filtration, the filter medium has no cake. Furthermore, Ap is not zero, but has a value that is a function of the resistance of the medium for a given flowrate. This initial condition can be stated as:

/ v\

(26)

For an incompressible cake (where s = 0), Equation 25 takes the form:

J?( — f \ At

AT I

As noted earlier, for thick cakes, the resistance of the filter medium may be neglected. Hence, for R t = 0, Equation 25 simplifies to:

* i . = pax \ I ---v Iy T

V

Q

\ AT

(28)

An increase in pressure influences not only coefficient r0, but the cake's porosity as well. Since the cake on the filter plate is compressed, residual liquid is squeezed out. Thus, for constant feed, the flowrate through the medium will not be stable, but will fluctuate with time. The weight of dry solids in a cake is:

W = xQV

(29)

where XQ = weight of solids in the cake per unit filtrate volume. The concentration of solids in the feed sludge is expressed by weight fraction c. It is also possible to evaluate experimentally the weight ratio of wet cake to its dry content m. Hence, a unit weight of sludge contains me of wet cake. We denote y as the

72

Liquid Filtration

specific weight of feed sludge. This quantity contains c amount of solids; hence, the ratio of the mass of solids in the cake to the filtrate volume is:

cv 1 - me

(30)

Thus, from the sludge concentration c and the weight of wet cake per kg of dry cake solids m, XQ can be computed. If the suspension is dilute, then c is small; hence, product me is small. This means that XQ will be approximately equal to c. According to Equations 29 and 30, the weight ratio of wet to dry cake will vary. Equation 30 shows also that because XQ depends on the product me, at relatively moderate suspension concentrations this effect will not be great and can, therefore, be neglected. However, when filtering concentrated sludges the above will play some role; that is, at constant feed, the filtrate changes with time. Variable-Rate and -Pressure Filtration The dynamics of variable-rate and -pressure filtrations can be illustrated by pressure profiles that exist across the filter medium. Figure 8 shows the graphical representation of those profiles. According to this plot, the compressed force in the cake section is: P = Pi-Ps,

where

p, ps

(31)

= pressure exerted on the sludge over the entire cake thickness = static pressure over the same section of cake

p corresponds to the local specific cake resistance (rw)x. At the sludge-cake interface pa = Pi and p = 0; and for the interface between the cake and filter plate ps, = ps( and P = Pi -p'st- p'a corresponds to the resistance of filter plate pf, and is expressed by: Ap = pRfW

(32)

where W = rate of filtration (mVm2-sec). Note that Apf is constant during the operation. Pressure p is also the driving force of the process. Therefore, starting from the governing filtration equations, the general expression for an infinitesimal increment of solid particle weight in a cake of unit of area is xwdq (q = filtrate volume obtained from 1 m2 filtering area, m3/m2). The responding increment dp may be expressed as:

Cake Filtration and Filter Media Filtration

FILTRATE

Figure 8, Distribution of static pressure pst in liquid andp along the cake thickness and filter plate: /, // -boundaries between the cake and sludge at i:" and T'; III, IV-boundaries between cake layers or cake and filter plate at T" and T'; V- boundary line between the cake and filter plate or free surface of filter plate; 1,3-curves Pst=f(hJ andp=f(hoc) at rf; 2, 4 -curves Pst=f(h0() andp=f(hoc) at T".

(33)

xw is not sensitive to changes in p. In practice, an average value for xw can be assumed. Note that W is constant for any cross section of the cake. Hence, Equation 33 may he integrated over the cake thickness between the limits of p = 0 and p=p, -p'sl, from q = 0 to q = q:

dp

(34)

Parameters q and W are variables when filtration conditions change. Coefficient (rw) is a function of pressure:

74

Liquid Filtration

MARIONETTE BOTTLE VENT MECHANICAL LOAD TOP L O A D PISTON FIXED CELL BODY POROUS MEDIUM CAKE POROUS MEDIUM BOTTOM FLOATING PISTON TRANSMITTED LOAD Figure 9. Compression-permeability cell.

The exact relationship can be derived from experiments in a device called a compression-permeability cell which is illustrated in Figure 9. Once this relationship is defined, the integral of the right side of Equation 34 may be evaluated analytically (or if the relationship is in the form of a curve, the evaluation may be made graphically). The interrelation between W and P, is established by the pump characteristics, which define q = f(W) in Equation 34. Filtration time may then be determined from the following definition:

(36)

Hence, dq W

(37)

Cake Filtration and Filter Media Filtration

75

Constant-Pressure and -Rate Filtration This mode of operation is achieved when a pure liquid is filtered through a cake of constant thickness at a constant pressure difference. Cake washing by displacement when the washing liquid is located over the cake may be considered to be filtration of washing liquid through a constant cake thickness at constant pressure and flowrate. The rate of washing is related to the rate of filtration during the last stages and may be expressed by Equation 9, where Ap is the pressure at the final moment and V is the filtrate volume obtained during filtration, regardless of the filtration method used (i.e., constant-pressure or constant-rate operation). In the final stages, filtration usually is performed under constant pressure. Then, the rate of this process may be calculated from Equation 19. From filtration constants C and K, at constant pressure for a given system, the filtration rate for the last period is determined. If the washing liquid passes through the filter in the same pore paths as the sludge and filtrate, then the difference between the washing rate and filtration rate for this last period will be mostly due to a difference in the viscosities of the wash liquor and filtrate. Therefore, Equation 19 is applicable using the viscosity of the washing liquid, p^. Denoting the rate of filtration in the last period as (dV/di), the washing rate is:

dV\

—

( dW] I

u '

d-c )w \ dl} ^ j

,

/TO\

(Jo)

Filter-Medium Filtration Formulas Solid particles undergoing separation from the mother suspension may be captured both on the surface of the filter medium and within the inner pore passages. This phenomenon is typical in the separation of low-concentration suspensions, where the suspension consists of viscous liquids such as sugar liquors, textile solutions or transformer oils, with fine particles dispersed throughout. The penetration of solid particles into the filter medium increases the flow resistance until eventually the filtration cycle can no longer proceed at practical throughputs and the medium must be replaced. In this section standard filtration formulas are provided along with discussions aimed at providing a working knowledge of the filter-medium filtration process.

Constant-Pressure-Drop Filtration Constant-pressure drop filtration can result in saturation or blockage of the filter medium. The network of pores within the filter medium can become blocked because of one or a combination of the following situations: 1. 2. 3.

Pores may become blocked by the lodging of single particles in the pore passage. Gradual blockage can occur due to the accumulation of many particles in pore passages. Blockage may occur during intermediate-type filtration.

76

Liquid Filtration

Proper filter medium selection is based on understanding these mechanisms and analyzing the impact each has on the filtration process. In the case of single-particle blockage, we first consider a i m surface of filter medium containing Np number of pores. The average pore radius and length are r p and S. p, respectively. For laminar flow, the Hagen-Poiseuille equation may be applied to calculate the volume of filtrate V passing through a pore in a unit of time:

V1 -

Consequently, the initial filtration rate per unit area of filtration is: W

= V'N

Consider 1m3 of suspension containing n number of suspended particles. If the suspension concentration is low, we may assume the volume of suspension and filtrate to be the same. Hence, after recovering a volume q of filtrate, the number of blocked pores will be nq, and the number unblocked will be (Np - nq). Then the rate of filtration is: (41)

or

W = W. ~kq III

"

(49)

V"T~/

where V'n

(43)

k' is a constant having units of sec"1. It characterizes the decrease in intensity of the filtration rate as a function of the filtrate volume. For constant V, this decrease depends only on the particle number n per unit volume of suspension. The total resistance R may be characterized by the reciprocal of the filtration rate. Thus, W in Equation 42 may be replaced by 1/R (sec/m). Taking the derivative of the modified version of Equation 42 with respect to q, we obtain:

Cake Filtration and Filter Media Filtration

dR

=

k'

Tq 7F~^7

(44>

Comparison with Equation 42 reveals:
or dR = k'R2 dq

Equation 46 states that when complete pore blockage occurs, the intensity of the increase in the total resistance with increasing filtrate volume is proportional to the square of the flow resistance. In the case of multiparticle blockage, as the suspension flows through the medium, the capillary walls of the pores are gradually covered by a uniform layer of particles. This particle layer continues to build up due to mechanical impaction, particle interception and physical adsorption of particles. As the process continues, the available flow area of the pores decreases. Denoting x0 as the ratio of accumulated cake on the inside pore walls to the volume of filtrate recovered, and applying the Hagen-Poiseuille equation, the rate of filtration (per unit area of filter medium) at the start of the process is:

where

When the average pore radius decreases to r, the rate of filtration becomes: W = BNr*

(49)

For a finite filtrate quantity, dq the amount of cake inside the pores is x0dq, and the cake thickness is dr. That is:

78

Liquid Filtration

xodq = -N2nrtdr

(50)

Note that the negative sign indicates that as q increases, the pore radius r decreases. Integrating this expression over the limits of 0 to q, for rp to r we obtain: P

P

(51)

And from Equations 47 and 49, we may define the pore radii as follows: 1/2

W.

(52)

BN,

P/

or simply:

(53)

Substituting these quantities into Equation 51 and simplifying terms, we obtain: 12

where x

o

B

1/2

(55)

It is convenient to define the following constant:

2C vl/2

(56)

From which Equation 54 may be restated as: W=

W.}(l-l/2Kq)2

(57)

Cake Filtration and Filter Media Filtration

79

Since W= 1/R, we may write: W.m(\ - \l2Kq)2

The derivative of this expression with respect to q is:

dR dq

k

=

W.n(l- \l2Kq)3

On some rearranging of terms, we obtain: (60)

or

where K" - K(Wit)m

(62)

Equation 61 states that the intensity of increase in total resistance with increasing filtrate amount is proportional to resistance to the 3/2 power. In this case, the total resistance increases less sensitively than in the case of total pore blockage. As follows from Equations 56 and 62: K" = 2C

(63)

Substituting Equation 55 for C and using Equation 48 for B, the above expression becomes: K" = 2(W,) 1/2

(64)

Note that for constant Win, parameter K" is proportional to the ratio of the settled volume of cake in the pores to the filtrate volume obtained, and is inversely proportional to total pore volume for a unit area of filter medium.

80

Liquid Filtration

Replacing W by dq/dt in Equation 57, we obtain: W. I

(65)

i - —Kq I dq 2

Integration of this equation over the limits from 0 to T for 0 to q we obtain: 2q ~W.n(2-Kq)

and on simplification:

K

_ T _ 1 T =

(67)

Equation 67 may be used to evaluate constants K (m"1) and Win. Finally, for the case of intermediate filtration, the intensity of increase in total resistance with increasing filtrate volume is less than that occurring in the case of gradual pore blocking, but greater than that occurring with cake filtration. It may be assumed that the intensity of increase in total resistance is directly proportional to this resistance: * = K'"R

dq

Integration of this expression between the limits of 0 to q, from Rf to R gives: R

f

Substituting 1/W for R and l/Win for Rf, the last expression becomes:

Win

~W ~

e

K,,,

or

W = W. e ~K""(I

Cake Filtration and Filter Media Filtration

81

Substituting dq/di for W in Equation 71 and integrating over the limits of 0 to t between 0 and q we obtain:

1 eK'"q - 1 W.in K'" Hence, If

''T

~

__-_.—u7~~

-~

/ --7 o \ W

'

•'

Accounting for Equation 70, the final form of this expression becomes: _L = JL + #'"T

w

w.

Filtration Mechanisms To compare the different mechanisms of filtration, the governing equation of filtration must be rearranged. The starting expression is: dV _ A/? Adi p[rox0(V/A) + Rr]

Replacing V by q, and denoting the actual filtration rate (dq/di) as W, the governing filtration equation may be rewritten for a unit area of filtration as follows: W=

^

At the initial moment when q = 0, the filtration rate is '" ~ ,. v

From Equations 76 and 77 we have:

W

W:

l+K'"Winq

(77)

82

Liquid Filtration

where

The numerator of Equation 79 characterizes the cake resistance. The denominator contains information on the driving force of the operation. Constant K'" (sec/m ) characterizes tile intensity at which the filtration rate decreases as a function of increasing filtrate volume. Substituting 1/R for W in Equation 78 and taking the derivative with respect to q, we obtain: dR

(80)

The expression states that the intensity of increase in total resistance for cake filtration is constant with increasing filtrate volume. Replacing W by dq/di in Equation 78 and integrating over the limits of 0 to q between 0 and T we obtain: K'"

T

1

Note that this expression reduces to Equation 74 on substituting expressions for W in (Equation 77) and K'" (Equation 79). Examination of Equations 46, 61, 68 and 80 reveals that the intensity of increase in total resistance with increasing filtrate volume decreases as the filtration process proceeds from total to gradual pore blocking, to intermediate type filtration and finally to cake filtration. Total resistance consists of a portion contributed by the filter medium plus any additional resistance. The source of the additional resistance is established by the type of filtration. For total pore blockage filtration, it is established by solids plugging the pores; during gradual pore blockage filtration, by solid particles retained in pores; and during cake filtration, by particles retained on the surface of the filter medium. The governing equations (Equations 42, 67, 74 and 81) describing the filtration mechanisms are expressed as linear relationships with parameters conveniently grouped into constants that are functions of the specific operating conditions. The exact form of the linear functional relationships depends on the filtration mechanism. Table 1 lists the coordinate systems that will provide linear plots of filtration data depending on the controlling mechanism. In evaluating the process mechanism (assuming that one dominates) filtration data may be massaged graphically to ascertain the most appropriate linear fit and, hence, the

Cake Filtration and Filter Media Filtration

83

type of filtration mechanism controlling the process, according to Table 1. If, for example, a linear regression of the filtration data shows that q = f(t/q) is the best linear correlation, then cake filtration is the controlling mechanism. The four basic equations are by no means the only relationships that describe the filtration mechanisms. Table 1. Coordinates for representing linear filtration relationships.

Type of Filtration

Equation

Coordinates

With Total Pore Blocking

42

q vs W

With Gradual Pore Blocking

67

T VS T/q

Intermediate

74

T VS ]/W

Cake

81

q vs t/q

All the mechanisms of filtration encountered in practice have the functional form: dR VT3b - - KR

(82)

where b typically varies between 0 and 2. Constant Rate Filtration Filtration with gradual pore blocking is most frequently encountered in industrial practice. This process is typically studied under the operating mode of constant rate. We shall assume a unit area of medium which has Np pores, whose average radius and length are rp and 0p, respectively. The pore walls have a uniform layer of particles that build up with time and decrease the pore passage flow area. Filtration must be performed in this case with an increasing pressure difference to compensate for the rise in flow resistance due to pore blockage. If the pores are blocked by a compressible cake, a gradual decrease in porosity occurs, accompanied by an increase in the specific resistance of the deposited particles and a decrease in the ratio of caketo-filtrate volumes. The influence of particle compressibility on the controlling mechanism may be neglected. The reason for this is that the liquid phase primarily flows through the available flow area in the pores, bypassing deposited solids. Thus, the ratio of cake volume to filtrate volume (x0) is not sensitive to the pressure difference even for highly compressible cakes. From the Hagen-Poiseuille relation (Equation 39) replacing Win in Equation 40 with constant filtration rate W and substituting APin for constant pressure drop AP we obtain: W = B'WinN/p

(83)

84

Liquid Filtration

where

The mass of particles deposited on the pore walls will be x0dq, and the thickness of this particle layer in each pore is dr. Hence xadq = -N 2nr( di

(85)

Integration over the limits of 0, q from rp to r yields

N uH

,

(86)

Radii rp and r are defined by Equations 83 and 85, respectively, from which we obtain the following expressions:

w

W

q =

(87)

B'LpN

or ™p

1/2

1/2

A/?

B'

Since q = WT, Equation 88 may be stated in a reduced form as: CT =

1

1/2

/

, \ 1/2

(89)

where 1/2

JV

A plot of Equation 89 on the coordinate of T vs (l/Apin)'/2 - (l/Apj /2 results in a straight line, passing through the origin, with a slope equal to C. Thus, if experimental data correlate using such coordinates, the process is gradual pore blocking. Note that at T= 0, Ap = Apin, which is in agreement with typical process observations.

Cake Filtration and Filter Media Filtration

85

The filtration time corresponding to total pore blockage, when Ap-+°° may be estimated from: T

™

1

1/2

1

(91)

C

To express the relationship between AP and T more directly. Equation 89 is restated in the form: (A-Cif

where 1/2

It is important to note that pore blocking occurs when suspensions have the following characteristics: 1. relatively small particles; 2. high viscosity; and 3. low solids concentrations.

Both particle size and the liquid viscosity affect the rate of particle settling. The rate of settling due to gravitational force decreases with decreasing particle size and increasing viscosity. The process mechanisms are sensitive to the relative rates of filtration and gravity sedimentation. Examination of the manner in which particles accumulate onto a horizontal filter medium assists in understanding the influences that the particle settling velocity and particle concentration have on the controlling mechanisms. The separation process through a cross section of filter medium is illustrated in Figure 10. "Dead zones" exist on the filter medium surface between adjacent pores. In these zones, particle settling onto the medium surface prevails. After sufficient particle accumulation, solids begin to move under the influence of fluid jets in the direction of pore entrances. This leads to favorable conditions for bridging. The conditions for bridge formation become more favorable as the ratio of particle settling to filtration rate increases. An increase in the suspension's particle concentration also enhances accumulation in "dead zones" with subsequent bridging. Hence, both high particle settling velocity increases and higher solids concentrations create favorable conditions for cake filtration. In contrast, low settling velocity and concentration results in favorable conditions for gradual pore blocking.

86

Liquid Filtration

The transition from pore-blocked filtration to more favorable cake filtration can therefore be achieved with a suspension of low settling particles by initially feeding it to the filter medium at a low rate for a time period sufficient to allow surface accumulation. This is essentially the practice that is performed with filter aids.

Figure 10, Suspension flow downward onto a filter medium. An initial accumulation of solids occurs around the pott entrance followed by particle bridging.

Suggested Readings 1. Cheremisinoff, P.N., Wastewater Treatment Pocket Handbook, Pudvan Publishing, Northbrook, IL, 1987 2. Cheremisinoff, P.N., Pocket Handbook for Solid-Liquid Separations, Gulf Publishing Co., Houston, TX, 1984 3. Noyes, R., Unit Operations in Environmental Engineering, Noyes Publishers, NJ, 1994 4. Kirkpatrick, J. , Mathematics for Water and Wastewater Treatment Plant Operators, Ann Arbor Science Pub., Ann Arbor, MI, 1976

Cake Filtration and Filter Media Filtration

87

5. Environmental Law Institute, Clean Water Deskbook, Environmental Law Reports, Washington, DC, 1988. Nomenclature

A = B, B' = C = c = Fs = hc = K,K",K'" = L = !p = n = Np = p = q = r r0 rp rw R R c ,R t u V W XQ

= = = = = = = = = =

area (m2) empirical parameters filtration parameter concentration (kg/m) force (N) cake height (m) filtration constants cake thickness (m) pore length (m) number of suspended particles number of pores pressure (N/m2) filtrate volumer per unit area of filter (m3/m3) or filtrate volume (m3) specific resistance (m"1) specific volumetric cake resistance (kg/m2) pore radius (m) specific mass cake resistance (kg/m2) resistance (m/sec) cake and filter resistances, respectively (m"1) average velocity (m/sec) filtrate volume (m3) mass of dry solids (kg), or rate of filtration (m3/m2-sec) ratio of cake to filtrate volume.

Greek Symbols

e yt, II T TO

= = = = =

porosity viscosity (P) ratio of filtration rate to gravity setting time (sec) time constant (sec)

4 INDUSTRIAL FILTRATION EQUIPMENT Introduction Filtration equipment is commercially available in a wide range. Proper selection must be based on detailed information of the slurry to be handled, cake properties, anticipated capacities and process operating conditions. One may then select the preferred operational mode (batch, semibatch or continuous), and choose a particular system on the above considerations and economic constraints. Continuous filters are comprised of essentially a large number of elemental surfaces, on which different operations are performed. These operations performed in series are solids separation and cake formation, cake washing, cake dewatering and drying, cake removal, and filter media washing. The specific equipment used can be classified into two groups: (1) stationary components (which are the supporting devices such as the suspension vessel); and (2) scraping mechanisms and movable devices (which can be the filter medium, depending on the design). Either continuous or batch filters can be employed in cake filtration. In filter-medium filtration, however, where particulates are retained within the framework of the filter medium, batch systems are the most common. Batch filters may be operated m any filtration regime, whereas continuous filters are most often operated under constant pressure. In an attempt to organize the almost overwhelming number of different types of filtration equipment, two classification schemes have evolved for continuous operations. The first scheme is based on operating pressure differentials and is provided in Table 1. The second scheme is based on the relative difference between gravity force and filtrate motion. Three orientations are possible: forces acting in

88

Industrial Filtration Equipment

8

opposite direction (countercurrent), forces acting in the same direction (cocurrent), and forces acting normal or perpendicular to each other (cross-mode operation). Table 1. Pressure differential

scheme for filtration operating classification. Source

Pressure Differential (N/mz)

Hydrostatic pressure of the suspension layer to be separated

Usually no more than 5

Action of compressors

5-9

Action of pumps

Up to 50 and higher

Because the influence of gravity is so important to most filtration operations, the second classification basis is used in this chapter. The operating principles and important features of filtration equipment are described in this chapter with the intent of providing the reader a background in the versatility and selection options available.

Rotary Drum Filters Rotary-drum filters fall into the category of the countercurrent mode type operation; and are either vacuum- or pressure-operated. They are most frequently operated as vacuum filters. Although operated under pressure, they are rarely subjected to excessive pumping pressures. The principal advantage of these filters is the continuity of their operation. Total filtration cycles are limited to narrow time intervals. This necessitates maintaining nearly constant slurry properties. Changing slurry properties can lead to wide variations in the required times for completing individual operations of the filtration process. For separating low-concentration, stratified suspensions, rotary drum filters are normally specified for a submergence rate of 50 %. Such slurries require only mild mixing to prevent particle settling. These filters are less useful in handling polydispersions containing particles with wide size ranges. Fouling by small solids is a frequent problem in these latter cases. Drum Vacuum Filters with External Filtering Surfaces These filters are characterized by the rate at which the drum is immersed in the suspension. These are perhaps the most widely employed countercurrent operated filters in industry, an example of which is shown in Figure 1. As shown, the design consists of hollow drum-1, with a slotted face, the outer periphery of which contains shallow tray-shaped compartments-2. The filter cloth is supported by a grid or a heavy screen, which lies over these compartments. The drum rotates on a shaft with one end connected to the drive-3, and die otiier to a hollow trunnion adjoining to an automatic valve. The drum surface is partially immersed in the suspension contained in vesscl-6. The cake that is formed on the outer surface of the drum is removed by scrapcr-7 as the drum rotates.

Figure 1. Rotary-drum vacuum filter with external filtration surface: 1 - hollow drum; 2 filtration compartments; 3 - drive; 4 - hollow trunnion; 5 - automatic valve; 6 - tank far suspension; 7 - knife for cake scraping.

Figure 2 illustrates a longitudinal view of the system. Compartments-2 of the drum-1 are connected through the pipe-3, passing through the hollow trunnion-4 of shaft-5, with the automatic valve-6. A stirring device-7 is mounted under the drum to prevent particle settling. A diagrammatic cross section of the filter is illustrated in Figure 3. As the drum rotates clockwise, each compartment is connected by the pipe-2 with different chambers of immobile parts of the automatic valve-4 and passes in series through the following operating zones: filtration, first dewatering, washing, second dewatering, cake removal and cloth regeneration. In the filtration zone, the compartment contacts the suspension in the tank-11, and is connected to the pipe-10 hooked up to a vacuum source. The filtrate is discharged through the pipe and space in the collector and the cake forms on the compartment's surface. In the first dewatering zone the cake comes in contact with the atmosphere, and the compartment is connected to the space-10. Because of the vacuum, air is drawn through the cake, and for maximum filtrate recovery, the compartment remains connected to a collection port on the automatic valve. In the washing zone the cake is washed by the nozzles (or wash headers)-^. The compartment is connected through the port-6, which is also tied into a vacuum source. The wash liquor is removed in the other collector. In the second dewatering zone, the cake is also in contact with the atmosphere, and the compartment is connected with the port-6. Consequently, the washing liquid is displaced from the cake pores and delivered lo the collector. To avoid cake cracking during washing and dewatering, an endless belt-7 is provided, which moves over a set

Industrial Filtration Equipment

Figure 2. Longitudinal section of a rotary-drum vacuum filter with an external fiitrati surface: 1 - drum; 2 - compartment; 3 - connecting pipe; 4 - hollow trunnion; 5 - shaft; < automatic valve; 7 - stirring device. of guide rollers. In the discharge zone, the compartment is connected with the port-5, which is supplied by a compressed air source. This reversal of pressure or "blow" loosens the cake from the filter medium, whence it is removed by a scraper or doctor blacle-3. In the regeneration zone, compressed air is blown through the cloth; the air enters the compartment through the pipe from the port-13. The automatic valve serves to activate the filtering, washing and cake discharge function of the filter sections. It provides separate outlets for the filtrate and wash liquid, and a connection by which the compressed air blowback can be applied.

Cocurrent Filters Cocurrent or top-feed filters employ flat and cylindrical filtering media. In flat designs, the angle between the directions of gravity force and filtrate motion is 0°, but

n

Liquid Filtration

SECOND DEWAT£fttNG

. FIRST

5

WASH LI QUO

4

CA#£

REMOVAL

FILTRATI FILTRATION

Figure 3, Diagrammatic cross section of rotary-drum vacuum filter: 1 - drum; 2 - connecting pipe; 3 -scraper; 4 -automatic valve; 5, 13 -chambers of automatic valve connected with a source of compressed air; 6, 10 -chambers of automatic valve connected with a source of vacuum; 7 -endless belt; 8 - wash header; 9 - guiding roll; 11 - tank for suspension; 12 stirring device. may vary to larger angles. In this class of filter, the directions of gravity force action and filtrate flow are the same. Filter designs in this class are quite different from die counterflow type. They include sophisticated rotary-drum filters, continuous-belt filters, Nutsch batch filters and filter presses with horizontal chambers. These filters are most often used for separating stratified slurries. Separation by filters of the first subgroup is based on intensive slurry mixing by agitators. It is especially advisable to use these filters for the separation of polydispersed systems. In this case, the cake formed is properly stratified with large particles adjacent to the filter medium. Flat filtering surfaces form a cake of uniform thickness and homogeneous structure at any horizontal plane. This permits highly effective washing. Internal Rotary-Drum Filters An example of an internal rotary-drum filter is illustrated in Figure 4. The filter medium is contained on the inner periphery. This design is ideal for rapid-settling slurries that do not require a high degree of washing. Tankless filters of this design consist of multiple-compartment drum vacuum filters. One end is closed and contains an automatic valve with pipe connections to individual

Industrial Filtration Equipment

93

compartments. The other end is open for the feed entrance. The drum is supported on a tire with rigid rollers to effect cake removal. The drum is driven by a motor and speed-reducer connected to a riding roll shaft. The feed slurry is discharged to the bottom of the inside of the drum from the distributor and is maintained as a pool by a baffle ring located around the open end and the closed portion of the outer end. As the drum revolves, the compartments successively pass through the slurry pool, where a vacuum is applied as each compartment becomes submerged. Slurry discharge is accomplished at the top center where the vacuum is cut off and gravity (usually assisted by blowback) allows the solids to drop off onto a trough. From mere, a screw or belt conveyor removes the solids from the drum. This filter is capable of handling heavy, quick-settling materials. Nutsch Filters Nutsch filters are one design type with a flat filtering plate. This configuration basically consists of a large false-bottomed tank with a loose filter medium. Older designs employ sand or other loose, inert materials as die filtering medium, and are still employed in water clarification operations. In vacuum filtration, these falsebottom tanks are of the same general design as the vessels employed for gravity filtration. They are, however, less widely used, being confined for the most part to rather small units, particularly for acid work. Greater strength and more careful construction are necessary to withstand the higher pressure differentials of vacuum over gravity. This naturally increases construction costs. However, when high filtering capacity or rapid handling is required with the use of vacuum, the advantages may more than offset higher costs. Construction of the vacuum false-bottom tank is relatively simple; a single vessel is divided into two chambers by a perforated section. The upper chamber operates under atmospheric pressure and retains the unfiltered slurry. The perforated false bottom supports the filter medium. The lower chamber is designed for negative pressure, and to hold the filtrate. Nutsch filters are capable of providing frequent and uniform washings. A type of continuous filter that essentially consists of a series of Nutsch filters is the rotatingtray horizontal filter. Horizontal Rotary Filters An example of a horizontal rotary filter is illustrated in Figure 5. These machines are well suited to filtering quick-draining crystalline solids. Due to its horizontal surface, solids are prevented from falling off or from being washed off by the wash water. As such, an unusuaJly heavy layer of solids can be tolerated. The basic design consists of" a circular horizontal table that rotates about a center axis. The table is comprised of a number of hollow pie-shaped segments with perforated or woven metal tops. Each of the sections is covered with a suitable filter medium and is connected to a central valve mechanism that appropriately times the removal of filtrate and wash liquids and the

Figure 4. Sectiua view of an interiar medium rolory-drum v a ~ ~ uf&er. m

96

Liquid Filtration

dewatering of the cake during each revolution. Each segment receives the slurry in succession. Wash liquor is sprayed onto each section in two applications. Then the cake is dewatered by passing dry air through it. Finally, the cake is scooped off of the surface by a discharge scroll. Belt Filters Belt filters consist of a series of Nutsch filters moving along a closed path. Nutsch filters are connected as a long chain so that the longitudinal edge of each unit lias the shape of a baffle plate overlapping the edge of the neighboring unit. Each unit is displaced by driving and tensioning drums. Nutsch filters are equipped with supporting perforated partitions covered with the filtering cloth. The washed cake is removed by turning each unit over. Sometimes a shaker mechanism is included to ensure more complete cake removal. In contrast, a belt filter consists of an endless supporting perforated rubber belt covered with the filtering cloth. The basic design is illustrated in Figure 6. Supporting and filtering partitions-1 are displaced by driving drum-2 and maintained in a stretched condition by tensioning the drum-3, which rotates due to friction against the rubber belt. Belt edges (at the upper part of their path) slide over two parallel horizontal guide planks. The elongated chamber-4 is located between the guide planks. The chamber in the upper part has grids with flanges adjoining the lower surface of the rubber belt. The region under the belt is connected by nozzles-5 to the filtrate collector-6, which is attached to a vacuum source. The chamber and collector are divided into sections from which filtrate and washing liquid may be discharged. The sludge is fed by the trough-7. The cake is removed from the drum-2 by gravity or blowing, or sometimes it is washed off by liquid from the distributor nozzle-8. The washing liquid is supplied from the tank-9, which can move along the filtering partition, it can be washed during the belt's motion along the lower path. The filtering partition, illustrated in Figure 7, consists of riffled rubber belt-1 with slots-2, grooves-3 and filter cloth-4, which is fixed in a set of grooves by cords-5. Slots-2, through which the filtrate passes, are located over the grids of the elongated chamber. The edges of the rubber belt are bent upward by guides forming a gutter on die upper path of the belt. The velocity of the filtering partition depends on die physical properties of die sludge and the filter length. The cake thickness may range from 1 to 25 mm. The advantages of belt filters are tiieir simplicity in design compared to filters with automatic valves, and die abilities to provide countercurrent cake waslling and removal of thin layers of cake. Their disadvantages include large area requirements, inefficient use of die total available filter area, and poor washing at die belt edges.

3

Liquid Filtration

Figure 7. Filtering partition for a belt fitter: 1 - rubber belt; 2 - slots; 3 - grooves; 4 -filtering cloth; 5 - cord; 6 - edges of rubber belt.

Cross Mode Filters

Filters of this group have a vertical flat or cylindrical filtering partition. In this case, filtrate may move inside the channels of the filtering elements along the surface of the filtering partition downward under gravity force action, or rise along this partition upward under the action of a pressure differential. In the separation of heterogeneous suspensions, nonuniform cake formation along the height can occur because larger particles tend to settle out first. This often results in poor cake washing due to different specific resistances over the partition height. The cake may creep down along the partition due to gravity; mis is almost inevitable in the absence of a pressure gradient across the filtering partition. The vertical filtering partition makes these filters especially aseful as thickeners, since it is convenient to remove cake by reverse filtrate flow. Filter Presses

The common filter press is the plate-and-frame design, consisting of a metal frame made up of two end supports rigidly held together by two horizontal steel bars. Varying numbers of flat plates containing cloth filter media are positioned on these bars. The number of plates depends on the desired capacity and cake thickness. The plates are clamped together so that their frames are flush against each other, forming a series of hollow chambers. The faces of the plates are grooved: either pyramided or ribbed. The entire plate is covered with cloth, which forms the filtering surface. The filter cloth has holes that register with the connections on the plates and frames, so that when the press is assembled these openings form a continuous channel over the entire length of the press and register with the corresponding connections on the fixed head. The channel opens only into the interior of the frames and has no openings on the plates. At the bottom of the plates, holes are cored so that they connect the faces of die plates to the outlet cocks. As the filterable slurry is pumped through the feed channel, it first fills all of the frames. As the feed pump continues to supply fluid and buildup pressure, the filtrate passes through the cloth, runs down the face of the plate and passes out through the discharge cock. When the press is full, it is opened and dumped. Cake cannot be washed in these units and is therefore discharged containing

Industrial Filtration Equipment

a certain amount of filtrate with whatever valuable or undesirable material it may contain. Each plate discharges a visible stream of filtrate into the collecting launder. Hence, if any cloth breaks or runs cloudy, that plate can be shut off without spoiling the entire batch. If the solids are to be recovered, the cake is usually washed. In this case, the filter has a separate wash feed line, and the plates consist of washing and nonwashing types arranged alternately, starting with the head plate as the first nonwashing plate. The wash liquor moves down the channels along the side of each washing plate, and moves across the filter cake to the opposite plate and drains toward the outlet. This is illustrated in Figure 8. t—WASHING

CLOTH

PLATE-—*

NON-WASHING PLATE

FRAME

WASH WATER IN

HEAD

(MM

Figure 8, Illustrates wash water outlets on a filter press. To simplify assembly, the nonwashing plates are maked with one button and the washing plates with three buttons. The frames carry two buttons. In open-delivery filters the cocks on the one-button plates remain open and those on the three-button plates are closed. In closed-delivery filters a separate wash outlet conduit is provided. Figure 9 illustrates the basic design of a frame, a nonwashing plate and a washing plate. These plates and frames in closed-delivery filters are shown in Figure 10.

100

Liquid Filtration

In terms of initial investment and floor area requirements, plate-and-frame filters are inexpensive in comparison to other filters. They can be operated at full capacity (all frames in use) or at reduced capacity by blanking off some of the frames by dummy plates. They can deliver reasonably well washed and relatively dry cakes. However, the combination of labor charges for removing the cakes and fixed charges for downtime may constitute a high percentage of the total cost per operating cycle. WASH PORT

FEED PORT

PLATE, NONWASHING

FRAME

PLATE, WASHING

Figure 9. Plates and frame of an open-delivery through-washing filter. FEED

PLATE, NONWASHING Figure 10. Plates and frame of a closed-delivery through-washing filter. Leaf Filters

Leaf filters are similar to plate-and-frame filters in that a cake is deposited on each side of a leaf (refer to Figure 11), and the filtrate flows to the outlet in the channels provided by a coarse drainage screen in the leaf between the cakes. The leaves are immersed in the sludge when filtering, and in the wash liquid when washing. Therefore, the leaf assembly may be enclosed in a shell, as in pressure filtration, or simply immersed in sludge contained in an open tank, as in vacuum filtration. In operating a pressure leaf filter, the sludge is fed under pressure from the bottom and equally distributed. The clear filtrate from each leaf is collected in a common manifold and carried away. In filters with an external filtrate manifold (refer to the sketch in Figure 12), die filtrate from each leaf is visible through a respective sightglass. This is not possible when the leaves are mounted on a hollow shaft that serves as an internal filtrate collecting manifold. The filter cakes are built on each side

Industrial Filtration Equipment

of the leaves and filtration is continued until the required cake thickness is achieved. For washing, the excess sludge is usually drained, simultaneously admitting compressed air (3-5 Ib pressure), which serves mainly to prevent the cake from peeling off the leaves.

Guard

Drainage screen

Filter cloth

Frame

Figure 11. Sectional view of a filter leaf showing construction and approximate location of cake. Disk Filters Disk filters consist of a number of concentric disks counted on a horizontal rotary shaft and operate on the same principle as rotary-drum vacuum filters. The basic design is illustrated in Figure 13. The disks are formed by using V-shaped hollow sectors assembled radially around the shaft. Each sector is covered with filter cloth and has an audet nipple connected to a manifold extending along the length of me shaft and leading to a port on the filter valve. Each row of sectors is connected to a separate manifold. The sludge level in the tank should provide complete submergence to the lowest sector of the disks.

102

Liquid Filtration

FILTRATE OR ASH SIGHT GLASS FEED OR WASH

DISCHARGE

Figure 12, Sweetland pressure filter.

Figure 13. Rotary-disk vacuum filter: 1 - section; 2-filtering disks; 3 - automatic valve; 4 manifold for vacuum and filtrate discharge; 5 - piping for compressed air; 6 - doctor's knives for cake removal.

Industrial Filtration Equipment

103

Compared to drum vacuum filters, the greatest advantage of the disk filter is that, for the same filtering area, it occupies considerably less floor space. However, because of vertical filtering surfaces, cake washing is not as efficient as when a drum filter is used. The disk filter is ideal when the cake is not washed and floor space is at a premium. Cartridge Filters

Cartridge filters are normally operated in the countercurrent mode, but because of their extensive use throughout the chemical and process industries in applications ranging from laboratory-scale to commercial operations for flows extending to an excess of 5000 gal/min, a separate discussion is warranted. Typical applications include: remove undispersed solids; remove precipitated solids; protect catalyst beds; protect instruments; remove DE filter carryover; keep spray nozzles open; filter recirculating water; remove particles from coalings; filter cooling tower water; remove char particles; filter condensate; filter bottle and can wash water; filter poultry and meal wash water;

remove oversize particles from slurries; clean electrolytic solutions; filter waste oil for reuse; remove plastic fines from water; filter scrubber water; filter boiler feed water; filter pump seal water; protect glue applicators; protect reverse osmosis systems; protect chiller and air-conditioners; and remove pulp from juices.

Industrial applications of cartridge filters are as follows. The chemical industry uses cartridge filters to handle: acetic acid, calcium carbonate, brine, ethylene glycol, herbicides, hydrochloric acid,

latexes, resins, polymers, sulfuric acid, cooling tower water, and pelletizer water.

The food industry uses cartridge filters to handle: com syrup, dextrose, lard, jelly, juices, milk sugar, edible oils,

tea liquor, city and well water, extracts, chocolates, soybean concentrate, and peanut butter.

104

Liquid Filtration

The paper industry uses cartridge filters to handle: pigtnented coatings, white water, freshwater, size, starch, TiO, slurry. mill water,

dyes, cooling water, pump seal water, decker shower water, wet end additives, and clay slurry.

The petroleum industry uses cartridge filters to handle: a mine, feedstocks, reduced crudes, naphtha, fuel oil, motor oil,

hydraulic oil, injection fluids, completion fluids, cooling tower water, pump seal water, and synthetic lubricants.

Miscellaneous industrial uses of cartridge filters include: adhesives, resins, solvents, paints, shampoo, dyes,

cooling water, Pharmaceuticals, beverages, toothpaste, liquors, and beer.

Table 2 provides a summary typical filtration ranges encountered throughout industry. Early designs, still widely used, consist of a series of thin metal disks that are 3-10 in. in diameter, set in a vertical stack with very narrow uniform spaces between them. The disks are supported on a vertical hollow shaft, and fit into a closed cylindrical casing. Liquid is fed to die casing under pressure, whence it flows inward between the disks to openings in the central shaft and out through the top of the casing. Solids are captured between the disks and remain in the filter. Since most of the solids are removed at the periphery of the disks, the unit is referred to as an edge filter. The accumulated solids are periodically removed from the cartridge. As with any filter, careful media selection is critical. Media that are too coarse, for example, will not provide the needed protection. However, specifying finer media than necessary can add substantially to both equipment and operating costs. Factors to be considered in media selection include solids content, type of contaminant, particle size and shape, amount of contaminant to be removed, viscosity, corrosiveness, abrasiveness, adhesive qualities, liquid temperature, and required flowrate. Typical filter media are wire mesh (typically 10-700 mesh), fabric (30 mesh - 1 /mi), slotted screens (10 mesh - 25 jttm) and perforated stainless steel screens (1030 mesh). Table 3 provides typical particle retention sizes for different media.

Industrial Filtration Equipment Table 2. Typical filtration ranges encountered in industry applications. Industry and Liquid

Typical Filtration Range

Chemical Industry Alum Brine Ethyl Alcohol Ferric Chloride Herbicides/Pesticides Hydrochloric Acid Mineral Oil Nitric Acid Phosphoric Acid Sodium Hydroxide Sodium Hypochlorite Sodium Sulfate Sulfurie Acid Synthetic Oils

60 mesh-60 pm 100-400 mesh 5- 10pm 30-250 mesh 100-700 mesh 100 mesh to 5-10 pm 400 mesh 40 mesh to 5-10 pm 100 mesh to 5- 10pm 1-3 to 5- 10pm 1-3 to 5- 10pm 5- 10 pin 250 mesh to 1-3 pm 25-30 pm

Drugs and Cosmetics Acetic Acid Aerosol Bath Oil Citric Acid Glycerine Lipstick Shampoo Soap Suntan Lotion Tallow Toothpaste

40- 150 mesh 60-200 mesh 400-700 mesh 60 mesh to 1-3 pm 5- 10pm 60- 150 mesh 100-250 mesh 10-250 mesh 15-20 pm 700 mesh to 25-30 pm 100 mesh

Food and Beverage Apple Juice Beer Brine Chocolate Corn Syrup Fructose Syrup Fruit Juices with Pulp Jelly Lard Lemon Effluent Liquors Vegetable Oil Wash Water

5- 10 pm 250-400 mesh 400 mesh to 15-20 pm 10-400 mesh 80 mesh to 5-10 pm 5- 10 to 25-30 pm 10- 100 mesh 700 mesh 500 mesh to 5-10 pm 60- 150 mesh 700 mesh to 15-20 pm 150 mesh to 5-10 pm 20-250 mesh

Petroleum Industry Atmospheric Reduced Crude Completion Fluids Completion Fluids DBA Distil lated Oil Decant Oil Diesel Fuel Gas Oil Gasoline Hydrocarbon Wax

25-75 pm 200 mesh to 1-3 pm 250 mesh to 5- 10 pm 200 mesh 60 mesh 100 mesh 25-75 pm 1-3 pm 25-30 pm

105

106

Liquid Filtration

Table 2. (Continued) Petroleum industry (continued) Isobutane MEA Naphtha Produced Water for Injection Residual Oil Sea water Steam Injection Vacuum Gas Oil

250 mesh 200 mesh to 5-10 fim 25-30 fim 1-3 to 15-20 /xm 25-50 pm 5- 10 jiin 5-10 /xm 25-75 /zm

Pulp and Paper Calcium Carbonate Clarified White Water Dye Freshwater Groundwood Decker Recycle Hot Melt Adhesives Latex Miliwater Paper Coating River Water Starch Size Titanium Dioxide

30- 100 mesh 30- 100 mesh 60-400 mesh 30-200 mesh 20-60 mesh 40-100 mesh 40-100 mesh 60-100 mesh 30-250 mesh 20-400 mesh 20-100 mesh 100-200 mesh

All Industries Adhesives Boiler-feed Water Caustic Soda Chiller Water City Water Clay Slip (ceramic and china) Coal-Based Synluel Condensate Coolant Water Cooling Tower Water Deionized Water Ethylene Glycol Floor Polish Glycerine Inks Liquid Detergent Machine Oil Pelletizer Water Phenolic Resin Binder Photographic Chemicals Pump Seal Water Quench Water Resins Scrubber Water Wax Weil water

30- 150 mesh 5-10 jum 250 mesh 200 mesh 500 mesh to 1-3 fim 20-700 mesh 60 mesh 200 mesh to 5- 10 /im 500 mesh 150-250 mesh 100-250 mesh 100 mesh to 1-3 /im 250 mesh 5-10/iin 40-150 mesh 40 mesh 150 mesh 250 mesh 60 mesh 25-30 ;xm 200meshto5-10ftm 250 mesh 30-150 mesh 40- 100 mesh 20-200 mesh 60 mesh to i-3 urn

Single filters may be piped directly into systems requiring batch or intermittent service. Using quick-coupling connectors, the media can be removed from the housing, inspected or cleaned. Also, filtering elements are interchangeable. Hence, while one is being cleaned, another can be placed into service.

Industrial Filtration Equipment

107

Multiple filters are also common, consisting of two or more single filter units valved in parallel to common headers. The distinguishing feature of these filters is the ability to sequentially backwash each unit in place while the others remain on stream. Hence, these systems are essentially continuous filters. These units can be fully automated to eliminate manual backwashing. Backwashing can be controlled by changes in differential pressure between the inlet and outlet headers. One possible arrangement consists of a controller and solenoid valves that supply air signals to pneumatic valve actuators on each individual filter unit. As solids collect on the filter elements, flow resistance increases. This increases the pressure differential across the elements and, thus, between the inlet and outlet headers on the system. When the pressure drop reaches a preset level, an adjustable differential pressure switch relays information through a programmer to a set of solenoid valves, which in turn sends an air signal to the pneumatic valve actuator. This rotates the necessary valve(s) to backwash the first filter element. When the first element is cleaned and back on stream, each successive filter element is backwashed in sequence until they are all cleaned. The programmer is then automatically reset until the rising differential pressure again initiates the backwashing cycle. Filter cartridges or tubes are made from a variety of materials. Common designs are natural or synthetic fiber wound over a perforated plastic or metal core. A precision winding pattern covers the entire depth of the filter tube with hundreds of funnelshaped tunnels, which become gradually finer from the outer surface to the center of the tube and trap progressively finer particles as the fluid travels to the center. This provides greater solids retention capacity than is associated with surface filter media of the same dimensions. Typical cartridge materials are cotton, Dynel, polypropylene, acetate, porous stone and porous carbon filter lubes. Supporting perforated cores for cotton, Dynel or polypropylene are stainless steel, polypropylene or steel. Supporting cores for acetate tubes are tin-plated copper with voile liner. Porous stone and porous carbon filter tubes do not require supporting cores. Stainless steel cores are recommended for mildly acid and all alkaline solutions, pH 4-14. Polypropylene cores are used where all metal contact must be eliminated or where stainless steel is attacked, such as high chloride and sulfuric acid solutions. It is recommended for all acid and alkaline solutions, pH 0-14. Two types of polypropylene cores are available: mesh polypropylene and rigid perforated polypropylene. Mesh polypropylene is satisfactory for temperatures below 140°F. The more expensive rigid polypropylene cores are used for temperature applications over 140°F, and for double- and tripletiered filter chambers because their greater strength is needed here. Perforated steel cores are used for dilute alkaline solutions, solvents, lacquers, oils, emulsions, etc. Table 4 can serve as a rough guide to filter cartridge selection. The following general guidelines are useful: •

Cotton filter tubes are recommended for moderately acid and alkaline solutions in the pH range 311.

•

Polypropylene, Dynel and porous carbon filter tubes are recommended for concentrated acid and alkaline solutions and for all fluoborate solutions over the entire pH range (0-14).

108

Liquid Filtration

•

Polypropylene filter tubes are also recommended for electropolishing solutions, as well as certain other highly corrosive solutions.

»

Porous stone Filter tubes are recommended for concentrated acid solutions.

»

Acetate filter tubes are recommended for water,

Table 3. Typical filter retentions based on averages reported by different equipment suppliers.

Mesh Equivalent

Nominal Particle Retention in. /*m

Open Area

Wire Mesh

10 20 30 40 60 80 100 150 200 250 400 700

0.065 0.035 0.023 0.015 0.009 0.007 0.0055 0.0046 0.0033 0.0024 0.0018 0.0012

1650 890 585 380 230 180 140 115 84 60 45 30

56 46 41 36 27 32 30 37 33 36 36 25

Perforated

10 20 30

0.063 0.045 0.024

1575 1125 600

15 18 12

Slotted

10 15 20 30 40 60 80 100 120 150 200 325

0.063 0.045 0.035 0.024 0.015 0.009 0.007 0.006 0.005 0.004 0.003 0.002 0.001

1600 1140 890 610 380 230 180 150 125 100 75 50 25

50 43 36 30 20 18 25 13 11 9 7 5 3

Fabric

60 80 100 150 250 500

0.009 0.007 0.0955 0.0046 0.0024 0.0016 0.0010-0.0012 0.0006-0.0008 0.0002-0.0004 0.00004-0.00012

230 180 140 115 60 40 25-30 15-20 5-10 1-3

NA NA NA NA NA NA NA NA NA NA

NA = percentage of open area not applicable to fabric media.

Industrial Filtration Equipment

109

Strainer and Filter Bag Baskets Strainer filter baskets and filter bag baskets are used as prefiltering devices. This prestraining or prefiltering stops the larger contaminated particles and thus extends the life of the entire system. Single-stage strainers and bag filters differ only in the basket design. Strainer baskets have solid flat bottoms, and baskets for filter bags have perforated bottoms to accept standard size filter bags. Dual-stage straining/filtering action is achieved by insertion of a second, inner basket. It is supported on the top flange of the outer basket. Both baskets can be strainers (with or without wire mesh linings) or both can be baskets for filter bags. They may also be a combination: one a strainer basket, the other a filter bag basket. Dual-stage action increases strainer or filter life and reduces servicing needs. Figure 14 shows details of the basket seal, which prevents unfiltered liquid from bypassing the strainer or filter bag basket. The seal is maintained during operation by a hinged basket bail handle being held down under the closed cover, which holds the basket down against a positive stop in the housing. There are a variety of strainer and filter bag basket arrangements. Figure 15 shows different single- and double-stage basket units. Fabric bag filter baskets are capable of providing removal ratings from 20 mesh to nominal 1 /xm, for both Newtonian and viscous liquids. Wire mesh or fabric baskets can be cleaned and reused in many applications, or are disposable when cleaning is not feasible. Side-entry models feature permanent flanged connections, for line pressures to 150 psi. These filters are fabricated to American Society of Mechanical Engineers (ASME) codes for applications that must comply with piping standards established in many processing plants. Top-entry models feature the inlet connection as an integral part of the lid. The inlet can be equipped with different types of quick disconnects for fast basket removal. Strainers should be selected so that the pressure drop incurred does not exceed a specified limit with a clean strainer basket (typically 2 psi). Pressure drop versus flow capacity curves for basket strainers are given in Figure 16. This plot provides gross pressure drop for different capacities of water flow at suitable strainer pipe sizes. The value obtained must be corrected on the basis of the actual fluid viscosity and strainer opening size to be used. These corrections are given in Table 5 and the procedure is as follows: »

Under the pressure drqj value from the bottom scale, with the specified flowrate. Read up to where its vertical line intersects the diagonal representing a strainer pipe size that gives a reasonable pressure drop, which is found by following the horizontal line to the pressure drop scale at the left.

110

Liquid Filtration

•

To correct this figure to match the actual fluid viscosity (and strainer media choice), use Table 5. Read down the appropriate viscosity column, and across from the appropriate strainer media description, to find the correction factor.

•

Multiply die pressure drop figure found in step 2 by the factor found in step 2 to obtain die adjusted pressure drop.

•

If you are using a mesh-lined (not pleated) strainer basket, the pressure drop can be lowered by using a 3Q-in.-deep basket instead of the 15-in.-deep one: divide the pressure drop figure from step 3 by 1.5.

Diaphragm Filters Diaphragm filters are specially designed filter presses. That have the ability to reduce sludge dewatering costs by a squeezing cycle using a diaphragm. Instead of the conventional plate-and-frame unit in which constant pumping pressure is used to force the filtrate through the cloth, diaphragm filters combine an initial pumping followed by a squeezing cycle that can reduce the process cycle time by as much as 80%. The operating cycles for this design are illustrated in Figure 17. During the filtration cycle, sludge is fed at approximately 100 psi into each chamber through an inlet pipe in the bottom portion of the filter plate. The number of chambers can range from a few dozen to more than 100. The sludge feed pump continues to feed sludge into the chamber until a predetermined filtering time has been achieved. Filtrate passes through die cloth on both sides of the chamber. The filtration cycle is completed independently in each chamber. Short filtration cycles produce cake thicknesses of 0.5-0.75 in, (12.7-19.1mm). COVER GASKET CONTAMINAT60 FLUID

HOUSING

BASKET

I

Figure 14. Details of basket seal.

•^___^_

j,:

MillillillllMMlMBllllilillMKi'

I

.

ouTLrr

Industrial Filtration Equipment

111

Table 4. Guide to filter cartridge selection.

Acid Fluoborates Nonfluoborates

Alkaline Cyanide Pyrophosphate Eletroless

Acids

Alkalies

Misc. Chemicals

Organic Liquids

Petroleum Products a

Electroplating Solutions

Filter Tube (Material/Core)

Cu, Fe, Pb, Sn Cu, Sn, Zn: <6 oz/gal H2SO4 Cu, Sn, Zn: >6 oz/gal H2SO4 Cr Au, In, Rli, Pd FeCl2 (190°F) Ni (Woods) Ni (Watts type & bright) Ni (high-chloride) Ni (sulfamate) Electrotype Cu and Ni Sn (stannate) Brass, Cd, Cu, Zna Au, In, Pt, Ag Cu, Fe, Sn, etc. Ni plating: < 140°F Ni plating: > 140°F Cu: <140°F Cu: > 140°F

Polypropylene (PP) or Dynel/PP PP or cotton/PP PP or Dynel/PP PP or Dynel/PP PP or Dynel/PP PP/rigid PP (RPP) or porous stone PP or Dynel/PP PP or cotton/PP PP or cotton/PP PP or cotton/PP ' PP or cotton/PP Cotton/stainless steel (SS) Cotton/SS, PP or Dynel/PP Cotton/SS, PP/PP Cotton/SS or PP Cotton/SS or PP PP/RPP, cotton/SS PP/PP PP/RPP

Chemicals

Uter Tube (Material/Core)

Acetic: dilute Acetic: concentrated Boric Chromic, hydrochloric, nitric, phosphoric, sulfuric Hydrofluoric, fluoboric NaOH or KOH NH4OH: dilute. NH4OH: concentrated Biological solutions Electropolishing solutions Pharmaceutical solutions Photographic solutions Radioactive solutions Ultrasonic cleaning solutions Nickel acetate (190°F) Food products CC14 Dichloroethylene Hydraulic fluids Lacquers Per- and trichloroethylene Solvents Fuel oil, diesel, kerosene, gasoline, lube oil

Cotton/SS, PP/PP PP or Dynel/PP Cotton/SS, PP/PP PP or Dynel/PP, porous stoneb PP or Dynel/PP PP/PP Cotton/SS, PP/PP PP or Dynel/PP Cotton/SS, PP/PP, porous stone11 Porous stone, PP/PP Cotton/SS, PP/PP, porous stone" Cotton/SS, PP/PP Cotton/SS, porous stoneh Cotton special B compound/SS Cotton/SS Cotton/SS, PP/PP Cotton/steel or SS Cotton/steel or SS Cotton/steel or SS Cotton/steel or SS Cotton/steel or SS Cotton/steel or SS Cotton/steel or SS

When operated as high-speed baths at high temperatures (> 140°F) or with high alkali content, use PP or Dynel/PP. '' Porous stone is recommended for all acids except hydrofluoric and fluoboric.

112

Liquid Filtration

Figure 15. Dgferent one- and two-stage basket configurations.

113

Industrial Filtration Equipment

Strainer pine size: 2"

5

3" 4"

5 8 88 S S 8 3 8 Flow, flpm

Figure 16. Pressure drop vsflow capacity for basket strainers. Table 5, Viscosity correction factors. Viscosity (cP) Air unlined baskets, with or without pleated inserts 40- mesh lined 60- mesh lined 80- mesh lined 100-mesh lined

10

50

100

200

400

600

800

1000

2000

0.65 0,73 0.77 0.93 1.00

0.85 0.95 1.00 1.20 1.30

1.00 1.20 1.30 1.50 1.60

1.10 1.40 1.60 1.90 2.20

1.20 1.50 1.70 2.10 2.40

1.40 1.80 2.10 2.40 2.70

1.50 1.90 2.20 2.60 3.00

1.60 2.00 2.30 2.80 3.30

1.80 2.30 2.80 3.50 4.40

Once the filtration cycle is complete, the sludge pump is stopped and a diaphragm in the chamber is expanded by water pressurized up to 213 psi. This compresses the sludge on both sides of the chamber into a thin, uniform cake with a solids content of more than 35%. The uniform water content of the thin cake (no wet cores) results in easier shredding and conveying and makes it much more adaptable to self-sustained thermal destruction or landfill. Optimum filtering and squeezing time cycles vary, depending on the type of sludge, and can be determined accurately by bench tests.

Liquid Filtration

CAKE DISCHARGE

CLOTH WASHING

Figure 17, Operating cycles of a diaphragm press.

Squeezing water is recycled. A hydraulic ram keeps (he chambers in position during both cycles.

On completion of the filtration and squeeze cycles, the chambers are automatically opened and the cakes are discharged, usually onto a belt conveyor. No precoating is required. Two chambers are normally opened at a time in sequence. This reduces the impact loading on the belt conveyor. Any sludge or filtrate remaming in the feed and filtrate lines is automatically purged by high-pressure (100-psi) air before the next cycle begins. This purging prevents wet sludge from discharging and keeps sludge lines from plugging.

115

Industrial Filtration Equipment

Cake discharge from filter presses is fast. After a number of cycles (depending on the sludge type), the filter cloth will require cleaning. This can be accomplished manually or can be performed automatically at preset frequencies with an automatic cloth washer using a jet of 1000-psi wash water. Where even faster cake discharge is desired or where sludge cakes may tend to be sticky, automatic cloth vibrators can be provided. These units help speed mechanical discharge and help remove cakes where poor sludge conditioning causes excessive sticking. This reduces the need for continuous monitoring by operations personnel. Cloth vibrators also simplify cloth selection, since cloths can be selected to assure clearer filtrate or better filtering qualities rather than sacrificing these advantages for a cloth that allows for filter discharge characteristics. Cake discharge is illustrated in Figure 18. Typical capacities and dimensions for diaphragm presses are given in Table 6.

High Pressure, Thin Cake Filters

Thin-cake staged filters have been used effectively at high flowrates per unit area for many years in both Eastern and Western Europe. Use of ultrathin cakes is a useful technique for increasing flowrates. The basic elements of the filter are shown in Figure 19. Filtration surfaces are recessed plates equipped with rotating turbines that maintain permanent precoat-type thin cakes throughout the filter. Cake thickness is prevented from growing beyond the in situ precoat formed during the first few minutes of the operation by blades on a rotating shaft passing through the axis of the filter. Table 6. Capacities and dimensions for diaphragm presses".

Number of Chambers'1 30 40 52 66 78 92 104 118 130

FilterArea0 2

2

Cake 'Volume at 0.7 Thickness

Length"

Weight"

m

ft

liters

ft

kg

tons

mm

ft-in.

115 154 200 254 300 354 400 454 500

1,237 1,656 2,152 2,732 3,228 3,808 4,304 4,884 5,380

1,002 1,337 1,736 2,211 2,611 3,079 3,479 3,947 4,353

36 47 61 78 92 109 123 139 154

53,200 59,400 67,800 79,600 88,700 99,000 110,400 122,700 130,200

58 65

7,355 7,975 8,720 10,035 10,780 11,650 12,895 13,760 14,505

24-1 26-2 28-7 32-11 35-4 38-2 42-3 45-1 47-7

75 88 98 109 121 135 143

" Based on average values reported by different machine suppliers. h Presses available from 30 to 130 chambers in 2-chamber increments. c Nominal plate size is 1500 X 1500 mm (4 tt-11 in. square). d Weight of the press only (without sludge). "Overall length. All presses have an overall width of 3000 mm and an overall height of 4200 mm (9 ft-10 in. X 13 ft-9in.).

IU

Liquid Filtration

VIBRATION UNIT.

CAKE WITH DtA PHRAGM PLATE

Figure 18. Cake discharge from a diaphragm filter press.

PUD

TYPICAL STAGE DELAYED CAKE FILTER-THICKENER

Figure 19. Principal components of a staged, thin-cake filter thickener.

Industrial Filtration Equipment

11"

Slurry flows into the first stage and then flows around the turbines and through the clearances between the shaft and active filter surfaces. As liquid is removed, die thickening slurry moves from stage to stage. The unit acts as a filter-thickener, producing a continuous extruate that may contain a higher solids content than is normally encountered in conventional filters. The turbine plates sweep close to the filter cloths, leaving a thin, permanent cake on each stationary plate (see Figure 19). Even in the last stage of the filter, where the slurry is highly non-Newtonian, a thin, easily identifiable, hard cake is maintained. At low turbine velocities, the blades serve as scrapers that limit the cake thickness to the'dimensions of the clearance. At higher velocities, the cake thickness is reduced and can be as thin as 1.0 mm with 3-mm clearance. For filters that depend entirely on fluid action, shear forces at the cake surface depend on fluid properties and velocity distributions. The combination of high pressure (300 psi) and thin cakes produces high rates. Washing is accomplished either cocurrently or countercurrently. Separate filters in series can be employed in a manner similar to conventional thickeners. Washing may also be performed within a single unit, whereby, an initial portion of the filter must be used to remove liquid. In this case, the final stages are used for concentration. Clean wash liquid may be injected after the initial filtering at one or several intermediate stations. Injection wash tends to increase the overall filtrate rate but decreases the cake output rate.

Thickeners Many existing filtration applications can be greatly enhanced if their present equipment, such as plate-and-frames and rotary vacuums, is used in conjunction with a thickening operation. Table 7 illustrates this point. Case 1 shows that if a feed slurry of 2% is concentrated in a filter to 50% (by volume), a total of 98% fractional removal of water is needed. If, however, a thickener is employed to concentrate from 2 to 10%, the fractional removal, of water is 82%, thus leaving only 16% of the filter. This means that the present filter could be used about three times more effectively it supplemented with a thickener. In Case 2, a 1% slurry is concentrated to 30% solids. A single filter would require 98% fractional removal of the water. By use of a thickener concentrating first from 1 to 7%, we fractionally remove 87% of the water. This leaves only 11% fractional removal of the present filter to go from 7% to the required 30%. Many filter-thickeners are simple settling tanks or decanters. Thickeners of this type are generally large and bulky and have relatively slow rates. Centrifuges have a greater driving force but, in general, are expensive to operate and can deliver cloudy overflow if fine particles are present.

118

Liquid Filtration

High-velocity cross-flow thickeners are available, but operating experience often shows them to be highly dependent on the rheology of the slurry. Sometimes a slight increase in outlet concentration can result in filter blocking. Table 7. Examples of thickener operation improvements to filtration.

Case

Solids (vol%)

i 2

a 11

Void Ratio, e* 2 10 50 1 7 30

49 9 i 99 13.29 2.33

Fractional Removal of Water (F%)b 0 82 16 0 87 11

e= vol liquid/vol solids. F% =( er^^ted,

Dynamic Thickeners

Dynamic thickeners have become a popular machine option. Special dynamic elements housed inside a thickening chamber keep the slurry continuously moving; hence, a concentrate of a paste-like consistency is possible without the danger of filter blocking. High flowrates per unit area, resulting from very thin cake formation, allow such units to be designed of relatively small size. These systems are designed for little or no cake formation at lower levels of concentration. It can be shown that filtration rates increase with a reduction in cake thickness. However, some materials (especially gels such as aluminum hydroxides) are so compressible that 90% of the available pressure drop is absorbed by a "skin layer" formed on top of the filter media, while the remainder of the cake remains soupy and unconsolidated. Consequently, reducing the cake thickness (in examples such as unwashed "gels") in equipment that uses techniques of "thin cakes" would not result in any significant improvements. It is, therefore, advantageous to minimize the formation of a "skin layer". Dynamic thickeners operate in a recycle mode of operation. The feed enters the thickener and the filtrate leaves the filtering plates while the steady-state-rumiing, concentrated paste comes out of the modulating cake valve and reenters the feed tank. When the feed tank solids reach a predetermined concentration, e, the thickening operation is complete. To make the thickening operation continuous, one would install two feed tanks, so that after the first tank is completed the product can then be fed to continuous filtration equipment for further liquid removal as the second tank of feed solution is processed through the thickener. These two feed tanks can be set up with high- and low-level audible signals and automatic switching three-way plug valves so that continuous operations are possible with any continuous filter. The operating scheme is illustrated in Figure 20.

Industrial Filtration Equipment FEED

WASH WATER (IF REQUIRED)

FEED TANK 1

\ RECYCLE

Figure 20. Typical operating scheme for a dynamic thickener.

FEED TANK 2

119

120

Liquid Filtration

Solids Washing Washing of chemical solids in filtration is employed to enhance the purity of the product. High washing efficiency is the ultimate goal, along with minimum use of energy and wash liquid, clear filtrates, maximum flowrates, and a homogeneous washed product. Washing is usually accomplished in conventional cake-forming systems by forcing wash liquid through existing filter cakes. The initial efficiency is high as the mother liquor is being displaced; however, after breakthrough, the process is controlled by the diffusion rate of the solute, which explains why washing efficiency drops so rapidly with time. Cake sagging in plate-and-frame filters reduces the effectiveness of wash as most of the liquid flows through the area having the least solids buildup of cake. In rotary drum filters, an even more detrimental effect is cake cracking. Massive amounts of wash liquid short-circuit directly through the filter cloth, thus, crippling the entire washing process. In contrast, with dynamic thickeners, the slurry or paste is washed instead of a cake. Solubles in the feed are dispersed into die wash liquid by strong agitation, The time required to reduce the solubles in a slurry to a desired level is a function of the feed solids concentration. The optimum washing concentration can be determined from the slurry's filtration characteristics. For example, as shown in Figure 21, washing should begin at the feed concentration if the data produce a concave curve. On the other hand, a convex curve implies that washing at the higher solids concentration is best. Finally, the washing curve may contain an inflection point—an indication that the slurry should be thickened to a predetermined concentration before washing begins. Note that each curve shows its optimum slope line. As the first stage is run, information is automatically recorded (with a data logger) concerning temperature, torque and filtrate weights. Thus, in one run filtrate rates can be obtained as a function of solids concentration.

Centrifugal Filtration Filtering centrifuges are distinguished from standard centrifugation by a filtering medium incorporated into the design. Slurry is fed to a rotating basket or bowl having a slotted or perforated wall covered with a filtering medium such as canvas or metalreinforced cloth. The angular acceleration produces a pressure that transports the liquor through the filtering medium, leaving the solids deposited on the filter medium surface as a cake. When the feed stream is stopped and the cake spun for a short time, residual liquid retained by the solids drains off. This results in final solids that are considerably drier than those obtained from a filter press or vacuum filter.

Industrial Filtration Equipment

121

»*>

Curve with Inflection Point

Figure 21. Optimum wash curves for a dynamic thickener: Wash at feed concentration (top); wash at thickened concentration (middle); wash at semithickened concentration A (bottom). i/ris a function of rate; yis an inverse function of solids concentration.

122

Liquid Filtration

Principal types of filtering centrifuges are suspended batch machines, automatic shortcycle batch machines and continuous conveyor centrifuges. In suspended centrifuges, the filter medium is usually canvas or a similar fabric, or woven metal cloth. Automatic machines employ fine metal screens. The filter medium in conveyor centrifuges is usually the slotted wall of the bowl itself. Figure 22 shows a widely used design. The system combines the features of a centrifuge and a screen. Feed enters the unit at the top and is immediately brought up to speed and distributed outward to the screen surface by a set of vanes. Water or other liquid is forced by the sudden centrifugal action through the screen openings into an effluent housing. As solids accumulate, they are gently moved down the screen by die slightly faster rotating helix. With the increase in screen diameter, higher centrifugal gravities are encountered and solids are dispersed over a gradually increasing area, thus forming a thin, compact cake from which the remaining liquid is extracted. The relatively dry solids are blown out the bottom of the rotor by a set of vanes into a conical collection hopper. LIQUID FILTRATE MOTOft

HOPPER

SOLIDS DlSCHAftQE

Figure 22. Cutaway view of one type of filter centrifuge.

The theory of constant-pressure filtration may approximately be applied to filtration in a centrifuge. The following are assumed:

Industrial Filtration Equipment 1, 2. 3. 4, 5, 6.

123

Effects of gravity and changes in the liquid kinetic energy arc negligible. The pressure drop developed from centrifagal action is equivalent to the drag of die liquid flowing through the cake. Particle voids in the cake are completely tilled with liquid. The resistance of the filter medium is constant. Liquid now is laminar. The cake is incompressible.

Dewatering is not only an important step in a filtration process — it is also one of me primary operations in processing materials. The necessary first step in the efficient drying or processing of many products is the extraction of excess moisture by screening and pressing. Sludge dewatering can be accomplished in several ways. However, in general, pressing tends to be a more energy-efficient operation than evaporation or other heat transfer methods. Multistage screw presses can be used for dewatering chemical cellulose and for the removal of "black liquor" from kraft pulp, employing a recycling system with liquor flow countercurrent to the flow of stock, thus, producing a much higher percentage of solids in the liquor fed to evaporators. Presses can also be employed in the continuous rendering industry, as well as in reconstitution processes, as, for example, flax shive slurries, where four presses are used in conjunction with four slurry blending tanks, operating as a four-stage countercurrent washing or leaching step for upgrading an otherwise waste material. On certain products, continuous four-stage presses can accomplish multistage counterflow washing in a single unit. Screw presses may be used in the diffusion process for sugarcane, wherein the liquids for the diffusion of sugar solutions and fresh makeup water are extracted from the cut cane chips by the single or multistage unit. Some other products that can be handled by continuous screw presses are: reclaimed and synthetic rubber, wood pulp, waste paper pulp, drugs, miscellaneous chemicals, brewer's spent grains and hops, distiller's spent grain, packing house cracklings, paunch manure, soybean and cereal by-products, beet pulp, tomato pulp,

citrus pulp and peels, sweet and white potato pulp, tobacco slurries, cooked fish and fish cannery offal, copra, peat moss, corn germ, nitrocellulose, castor seed or beans, coffee grounds, alpha-cellulose.

Presses are available with many types of casings, designed to suit the characteristics of the material to be pressed, such as: 1. heavy 1-in.-thick carbon or stainless steel slatted casing;

124

2. 3. 4. 5.

Liquid Filtration

3/16-in,-thick naval alloy brass drilled screens, tapered for self-cleaning; stainless steel perforated screens; stainless steel narrow bar type super-drainage casing; or tine mesh Dutch twilled filler cloth in stainless steel.

The narrow bar type drainage casing has approximately three times the drainage area of the steel slatted casing, and twice the drainage area of the perforated or drilled screen casings. Low oil and moisture contents can be obtained with a continuous press, although output and final moisture content vary with the material being pressed, the speed at which the press is rotated, the uniformity of the feed to the press and the manufacturing process. A variable speed mechanical feeder with screw feed is available for forced feed when gravity feed is inadequate. Presses are of extremely rugged construction. Various parts of the press (the screws, for example) may be chrome-plated, of stainless steel or Monel, or furnished in other materials where corrosion and abrasion are severe. Provisions can be made for steam, water, press liquor or other liquids to be injected for cleaning or improved processing results. Figure 23 illustrates a typical screw press design. The material enters the press through the intake hopper from a surge tank or conveyor and drops on the feed flights (with wide pitch) of the screw. The flights of the screw become progressively closer together and the cones of the various stages increase in diameter as they approach die discharge end. Each successive stage presses the material harder; the high pressure extracts the liquid, which passes through the perforated screens or other types of casings and leaves the press in ail directions around the casing.

Ultrafiltration Three kinds of submicrometer semipenneable membranes can be delineated. The type with the largest pores is used for microfiltration (MF). MF typically lies in the range of 0.02-10 /xm. MF separation generally involves removing particles from fluids based on size; osmotic pressure is negligible. Ultrafiltration (UP) generally involves separation of large molecules from smaller molecules, and overlaps somewhat with die porosity range of membranes used for reverse osmosis (RO). RO usually involves purification or concentration of small molecules or ionic constituents in a solvent. Thus, we have microfilters, ultrafilters and membranes used for RO. The overlap of the definitions for RO and UF membranes arises from the following considerations. The "pores" in the skin of a membrane intended for removal of salt by RO are generally larger (e.g., 10-40 A) than the hydrated ions (e.g., Na+ C l , Ca + , SO4 2 ) they are intended to repulse. However, these pores are filled with water that is strongly influenced by the polymeric walls of the pores. Such water becomes "ordered water", which, because of its ordering, has too low a dielectric constant to

Industrial Filtration Equipment

126

Liquid Filtration

dissolve salt ions, in contrast to the bulk water. Thus, salt rejection in a useful RO membrane (e.g., more than 85% salt rejection) is based, on the lack of solubility of the hydrated ions in the ordered water within the pores, not on the size of the pores, it is not hard to imagine that the same membrane, or (at least) an inferior RO membrane (e.g., 5-20% salt rejection), would pass small molecules and reject larger molecules based primarily on size (UF) rather than on solubility (RO), hence the overlap of RO and UF ranges shown in Figure 24.

Figure 24. Chart showing microporous filtration ranges.

RO, UF and MF membranes are generally a tew mils in thickness; however, the discriminatory layer may be either a tight skin supported by an open substructure (i.e., a very thin effective thickness and, thus, low frictional resistance to flow) or it may be the entire thickness of the membrane or gel involved in the pass/rejection mechanism. In the latter case, the friction factors are much higher, i.e., the entire thickness equals the effective thickness. Osmotic pressure across a semipermeable membrane arises from differences in concentration, which in turn arise from relative ratios of the numbers of impermeable individual ions or molecules on the two sides of the membrane. These osmotic pressures are dominant when salts are to be removed by RO. Osmotic pressures vary from 3.5 psi for good tapwater to 350 psi with average sea water, as the number of ions per unit volume is very high (35,000 ppm). At the other extreme (MF), there are essentially no dissolved species that cannot permeate through the membrane; it follows that the osmotic pressures are minimal. UF membranes lie in between, usually with very few impermeable species of very high molecular weight; and therefore, much lower osmotic pressures exist across the membrane. Exceptions can exist involving UF and may be circumvented, e.g., the pervaporation process.

Industrial Filtration Equipment

127

It follows that operations, such as RO, involving high osmotic pressures require higher pressures (i.e., consume more energy) than do low-osmotic-pressure operations. Therefore, very thin effective thicknesses are desirable for practical industrial or commercial RO installations, to cut down on the frictional resistance to flow clue to effective thickness. Concentrated brine is continually swept out of the RO elements and away from the membrane to avoid plugging and concentration polarization. At the other extreme, MF membranes involved in dead-end flow require low driving pressures; therefore, thicker membranes with higher dirt-holding capacities are generally found most useful. Skinned or pseudoskinned varieties of MF membranes plug rapidly and account for some commercial failures of selected microfilters. UF membranes lie in between RO and MF membranes and are of two kinds; both are useful. Of industrial importance, are the thin-skinned membranes, which allow enhanced fiowrates (low friction factors) at given pressure differentials. Such UF membranes have larger pores in the thin skin than most RO membranes, and molecules of different molecular weights may be separated. Shape, size and molecular weight are important. As the osmotic effect is less important with UF membranes than in the case of RO membranes, lower pressures (generally less than 100 psig) are sufficient to promote permeation, and molecules that differ by a factor of ten in their molecular weights may usually be separated. Fractionation of cheese whey into solutions of protein and lactose is one familiar example. Of medical and biotechnical importance are the thicker homogeneous gel membranes, such as Cuprophane™, which are used in the artificial kidney and/or concentration dialysis. With the Cuprophane membranes, diffusional migration, driven by concentration differences across the membrane, effects the transport of the various species across the membrane and little, if any, pressure differential is applied. In kidney dialysis, toxic "middle molecules" diffuse across the Cuprophane membrane and out of the blood, while the larger desirable species are retained. Almost as much salt diffuses out of the blood as diffuses into the blood from the dialysate during (his procedure. A small pressure is imposed that depletes the patient of a few pounds of accumulated water over a period of hours. Such processes are considered to be primarily concentration-driven. These thick gel membranes are biotechnically very important; however, the pressuredriven thin-skinned UF membranes, while perhaps somewhat less selective, produce product streams so much more rapidly that they are the materials of choice for industrial processes. The above discussions primarily considered the physical parameters of the various membranes and their porous properties. Particularly in the case of UF, serious consideration must be given to the species that penetrate or are rejected by the UF membrane. Figure 24 shows that different sources attempt to relate molecular weight and pore size. Note that 10 A is presumably the cutoff point for either 300- or 500-molecular weight molecules, depending on the reference. Both sources could be correct, and the

128

Liquid Filtration

reasons that such uncertainty exists become even more important when larger molecular weights are involved. As examples only, consider the behavior and properties of proteins that one might wish to separate. Proteins are said to have primary, secondary, tertiary and, sometimes, quaternary structures. An oversimplified description of protein configurations in solution is useful. The various individual amino acids (about equal in number to the letters in the alphabet) may be strung together head-to-tail in an almost infinite number of sequences, just as randomly hitting the keys of a typewriter will give nonsensical words hundreds of letters long. Each of these random chains of amino acids (words) would correspond to the primary structure of a different protein. A particular sequence of amino acids depicted in two dimensions is considered the primary structure of a specific protein. There are, moreover, highly selective sites along these chains that are attracted to other specific sites along these same chains, forming loops held together by hydrogen bonds. A rendition of what sites are connected to what other specific sites and hence whether the resulting protein molecule would be forced to assume either helical or pleated sheet configurations reveals the secondary structure. As a result of these same hydrogen-bond interactions, the helical chains or pleated sheets become twisted, coiled chains, rods or globular shapes. This morphology constitutes the three-dimensional or tertiary structure. On occasion, two to four independent chains (based on primary structure) become intertwined via hydrogen bonds and van der Waals forces and these also assume various three-dimensional morphologies. These multiple-strand agglomerates are said to have quaternary structure. To complicate matters still further, these protein molecules may assume different morphologies in different environments or solutions. Table 8 shows the intrinsic viscosities of various proteins where the intrinsic viscosity is defined as volume per mass of a given protein; it may be seen that the molecular weights of proteins bear little relationship to die intrinsic viscosity. Note that ovalbumin (44,000 molecular weight) is a compact globular particle that occupies 3.7 cnrVg; if the sulfur-sulfur bonds are decoupled it further opens to encumber 54 cmVg of protein. Probable dimensions of variously sized particles are listed in Table 9. Further discussion exceeds the scope of this chapter. These examples illustrate that one should not jump to any filtrative conclusions based on molecular weight. Therefore, while it is safe to say that a given UF membrane could separate the much smaller lactose from the much larger protein in whey, it is dangerous to assume fliat selected proteins could be separated from each other without experimental evidence. All microporous filtration (MF, UF and RO) deals with purification, fractionation, concentration or partition. An example of purification is pressure-driven UF removal of particles and liigh-molecular-weight species from water subsequently to be used in hollow-fiber RO desalination. An example of pressure-driven fractionation is separation of protein and lactose from cheese whey for use as food additives (in the

12?

Industrial Filtration Equipment

case of protein) and subsequent fermentation into alcohol (in the case of the lactose). Were the lactose merely defined as waste and dumped into a sewer, the process would be defined as pressure-driven UF protein concentration. Another concentration-driven UF process is the concentration of protein solutions in die laboratory where an aqueous solution of protein is placed in a dialysis bag or tube and left for a period of hours in concentrated salt solution. Kidney dialysis also exemplifies concentration-driven partition filtration. Table 8. Intrinsic viscosities for macromolecules.

Compact Globular Particles Polystyrene Latex Particles Ribonuclease Lysozyme Myoglobin P-Lactoglobulin Ovalbumin Serum Albumin Hemoglobin Liver Alcohol Dehydrogenase Hemerythrin Aldolase Ribosomes (yeast) Bushy Stunt Virus Randomly Coiled Chains Polystyrene in Toluene Reduced Ribonuclease Oxidized Ribonuclease Oxidized Ribonuclease in Urea Ovalbumin in Urea Serum Albumin in Urea Reduced Serum Albumin in Urea Myosin in Guanidine Hydrochloride RNA Heat- denatured DNA Rodlike Particles Fibrinogen Collagen Myosin DNA TMV

Molecular Weight

Intrinsic Viscosity (cm3/g)

109 13,700 14,400 17,000 35,000 44,000 65,000 67,000 83,000 107,000 142,000 3.5 X 106 8.9 X 106

2.4 3.3 3.0 3.1 3.4 4.0 3.7 3.6 4.0 3.6 3.8 5.0 4.0

45,000 70,000 13,700 14,100 14,100 44,000 66,000 66,000 200,000 1.5 X 10* 5 x 106

28 37 14.4 11.6 13.9 34 22 53 93 100 150

330,000 345,000 620,000 5 X 106 4 X 107

27 1150 230 5000 29

Having distinguished between MF, UF and RO, and identified the two prevalent kinds of UF membranes, we will now discuss modes of operation: cross-flow (also tangeutialtlow and/or split-stream) filtration versus dead-end filtration. The numbers of particles per unit volume generally diminish in (he order: RO (ions) > UF (molecules) > MF (bacteria, etc.); when the retained particles are comparatively small in number, as is usually the case in MF filtration, dead-end filtration is suitable (Table 10). At the other

130

Liquid Filtration

extreme, as is the case in RO, concentration buildup always demands cross-flow, tangential-flow or split-stream treatment. The concentrate is continuously swept away, providing a relatively unchanged surface concentration. Pressure-driven UP also uses split-stream filtration to avoid membrane plugging or concentration polarization (also known as gel polarization). More recently, cross-flow filtration coupled with backwashing has also been implemented in MF filtration when the particulate load is particularly heavy or when long lifetimes of the MF membranes are desired. Table 9. Dimensions of various particles. Particle

Dimensions (/im)

Yeasts, Fungi Bacteria Viruses Proteins (lOMO6 mol wt) Enzymes Antibiotics, Pulypeptides Sugars Water

MO 0.3-10 0.03-0.3 0.002-0.1 0.002-0.005 0.0006-0.0012 0.0008-0.001 0.0002

Table 10. Dead-end versus cross-flow filtration.

Dead-end Crosstlow, Tangential Flow or Split Stream

RO -

UF .

+

+

MF + -f (energizing)

All three kinds of membranes (RO, UF and MF) may be manufactured in either flat sheet, tube or hollow tubular form. Generally, the hollow fiber (RO and smaller) or hollow tubular (UF and larger) configurations are less effective per unit area than are the flat sheet configurations, but this is offset by the greater effective area that can be packed into a volume of hollow tubules or fibers. The flat sheet configurations are usually plate-and-frame, spiral-wound or pleated cartridges. The third configuration, large tube intermediate, possessing the performance characteristics of the flat sheet but lacking the surface-to-volume advantage of hollow fibers. Tube configurations can, however, cope with the most contaminated streams, primarily because they can be cleaned mechanically. At the two extremes, tiny hollow tubules and most flat sheet configurations can be cleaned by reverse flow, but certain clogging contaminants are difficult to remove. The previous discussion brings us to one of the most important features of UF: gel polarization, which is important when the separation of macromolecules is involved in either flat sheet, tubular or hollow-fiber UF membrane configurations. As permeate containing the smaller molecules passes through the membrane, a layer of solution containing the larger rejected molecules accumulates adjacent to the membrane surface

131

Industrial Filtration Equipment

and may reduce the flow by plugging or fouling the membrane and/or forming a gelatinous filtration medium in series with the original membrane, increasing frictional resistance and sometimes reducing its effective pore size and not allowing the passage of smaller molecules that were intended lo pass through the unencumbered membrane. In some cases the problem is so severe that UF is precluded. However, three approaches have been used successfully in restoring the utility of the fouled membranes and/or keeping them from becoming fouled. They are (in order of decreasing difficulty of application) periodic purging with cleaning solutions (e.g., chemicals or enzymes), introduction of turbulence (see below) by one of a number of baffling arrangements and periodic backflushing. Backflushing is most readily applied to hollow tubular devices and is responsible in no small part for their growing acceptance. Turbulence promoters, generally inapplicable in hollow tubule devices, are most commonly employed in flat sheet configurations where, for example, Vexar™, a coarse webbing, is placed next to the membrane surface to induce a sweeping action or eddy currents, which promote rapid mixing of the incipient boundary layer back into the bulk fluid. There is, of course, a maximum concentration of potential gel-forming material that can be tolerated, at which point further UF becomes ineffective. Such induced sweeping is employed in plate-and-frame, spiralwound and pleated-membrane devices. Periodically, the cumulative effects of gel polarization, dirt accumulation of biological growth, render it necessary to renovate or clean the UF assemblies. These cleaning or antifouling techniques are of three kinds: chemical, reverse-flow or mechanical. Combinations of these can be used. All are practicable, depending on element, module or cartridge configuration (refer to Table 11). Chemical cleaning techniques are applicable to all configurations, although care must be taken to make certain that the membrane and other materials of construction are compatible with the chemical agents used. Reversing the flow is usually practicable, but with certain flat-sheet, spiral-wound, fluted and tubular configurations, inadequate membrane support during reverse-flow operation may cause problems. Table 11, Cleaning techniques for UF, Reverse-Flow

Chemical Flat Sheet Spiral Wound Fluted Tubular Hollow Fiber

4+ + +

± ± ± ±

Mechanical

± +

Because plate-and-frame and tubular configurations are used with the most contaminated fluids, mechanical cleaning techniques are used. In the case of plate-andfranie systems, the equipment may be disassembled and scrubbed, while in the tube configurations oversized soft foam plugs are driven through the tubes by pressure.

132

Liquid Filtration

The three methods of cleaning fouled UP membranes have been discussed above and while induced mild turbulence (considered below) may be seen as a preventive measure, all of the procedures result from the necessity to counteract the effects of ge! polarization. At least five additional techniques are being investigated that fall into the preventive category: 1.

2.

3. 4. 5.

The tube pinch effect no doubt takes place during rapid laminar flow in hollow tubules where hydrodynamic forces tend to cause particles to migrate toward the centers of the tubules and, hence, away from the walls. Enzymes, which decompose protein deposits, have been incorporated into the UF membranes, either by postimmobilization or by inclusion during the membrane's manufacture. Such membranes may be considered as self-cleaning to some extent. Immobilized positive or negative charges have been attached to UF membranes. By repelling likecharged species, the tendency to foul is diminished (see section on electrodeposition of paints). Electric fields have been imposed such that potentially fouling macromolecules or particles are attracted away (electrophoretically) from the UF membrane surface. Emulsified surfactants are injected into the feed. The surfactants are selected depending on the specific surfactant's enhanced ability to attract specific foulants to the water-surfactant interface rather than to the membrane-water interface.

The hollow tubule configurations with lumens frequently on the order of 0.5-2,0 mm in diameter present a different set of constraints but also present opportunities. Consider a cartridge (Figure 25) composed of a large number of hollow tubules potted at each end and encased in such a manner that the process stream can enter a plenum (A) at either end of the bundle of hollow tubules, proceed through the length of the tubules, losing fluid through the walls (UF), into the encasement and exit into either, a drain or reticule tank (B). Provision is made for removing the permeate (Q. Figure 26 (left) illustrates a similar situation, where 90% of the material issues as permeate (C). When the permeate flow decreases below a certain point due to fouling, the device may be renovated (Figure 26, right) by closing transiently the permeate valve, reducing the average transmembrane pressure to zero, and concoinitantly increasing the fluid through die tubules fourfold. In this fashion die fast flush may remove the accumulated debris. A close look at fast flushing (Figure 26, right) reveals that backflushing is also taking place. The hollow tubular bundle has a substantial friction factor due to die small diameters; hence, there is a pressure drop between A and B. Assuming for convenience a 20-psi pressure drop down the tubules (from A to B) under fast flow conditions, what would be the pressure in the encasement? Assuming symmetry, the pressure would be around the average at A and B. Thus, at the tubules near A there would be a 10-psi pressure drop between A and the encasement (encouraging ultrafiltration permeation), while the pressure would be reversed at the tubule endings near B, encouraging ultrafiltered fluid to backflush the tubules near B. Reversing the "fast flow" direction through the tubules would backflush in turn each end of the device (refer to Figure 27).

Industrial Filtration Equipment

B

PROCESS FLUID OUT

II PROCESS FLUID IN

Figure 25. Hollow tubule uaraftttration. 10% WASTE OUT

BACK FLUSH ZOfte

SINGLE PASS OPERATION

ULTRA* FILTBATIOII ZONfi

A

100* FEED IM

Figure 26. Hollow tubule utirafiUration.

A

100% FEED IN

133

134

Liquid Filtration

FEED OUT

FEED IN

UlTflAFtLTftATlON UPWARD FLOW FILTRATION BACK FtlHWf

IN

FEED OUT "V

Figure 27. Fast flushing.

As backflushing near the middle would be nil from time to time, ultrafiltered fluid or other cleaning solutions could be injected through D and reclaimed or dumped through A and/or B (refer to Figure 28). Although UF was first thought to be primarily applicable to the treatment of waste waters, such as treated sewage, to remove particulate and macromolecular matter, it is now known to be useful industrially in producing high-grade waters, recycling electrocoat paint particles, separations involving whole and skim milk, vegetable protein isolates (especially soybean), fermentation products, fruit juices, biochemicals such as pyrogens, phages in general, and human chorionic gonadrotropin. Reverse Osmosis

Reverse osmosis (RO) for water and wastewater treatment and for reuse at electricitygenerating power plants is a standard application. Uses of this unit operation include: recirculating condenser water, ash sluice water, boiler blowdown, boiler makeup and wet sulfur dioxide scrubber waste. Use of RO for desalination of seawater for boiler makeup is a typical application. The availability of this system has opened up the use of heretofore unavailable water supplies, and it has been used by the industry as a pretreatment to ion exchange demineralization. RO acts as an economical roughing demineralizer, bringing down the overall cost and improving the life of resins and operation of the ion exchange equipment.

Industrial Filtration Equipment

BACKFLUSH IN

WASTE Figure 28. Back/lushing.

As noted earlier, osmosis is the spontaneous passage of a liquid from a dilute to a more concentrated solution across an ideal semipermeable membrane that allows passage of the solvent (water) but not the dissolved solids (solutes) as shown in Figure 29. If an external force is executed on the more concentrated solution, the equilibrium is disturbed and the flow of solvent is reversed. This phenomenon, RO, is depicted in Figure 30. A basic RO treatment system consists of the components illustrated in Figure 31. Feedwater to the RO system is pumped first through a micrometer filter. This is a replaceable-cartridge element filter. The purpose of this filter is to remove any turbidity and particulate matter from the feed water before it enters the RO system. The filtered raw water then flows to a high-pressure pump, which feeds the raw water at a typical pressure of 400 psi through the RO membrane system. Valves and pressure gauges between the micrometer filter, die high-pressure pump and membrane modules control the flow of water through the system and monitor its operation. The RO system consists of two stages. The raw water is pumped through the first stage, which contains twice the number of membrane modules as the second stage. The first stage purifies 50% of the water fed to the system and rejects the remaining

136

Liquid Filtration

50%, which contains all of the contaminants. This reject water from the first stage K then passed through the second stage, which purifies 50% of the water fed to it and rejects the remaining 50% to waste. This second stage reject now contains all of the contaminants removed by both stages. Thus, the total flow through the system is 75% purified product water and 25% reject water.

I

OSMOTIC PRESSURE

SEMIPERMEABLE MEMBRANE

CONCENTRATED •

SOLUTION

.

DILUTE SOLUTION

Figure 29. Osmosis: normal flow from low to high concentration.

SEMIPERMEABLE MEMBRANE

CONCENTRATED '- SOLUTION

DILUTE SOLUTION

Figure 30. Reverse osmosis: flow reversed by application of pressure to high-concentration solution. The RO system removes 90-95% of the dissolved solids in the raw water, together with suspended matter (including colloidal and organic materials). The exact percent of product purity, product recovery and reject water depends on the amount of dissolved solids in the feedwater and the temperature at which the system operates. RO membrane performance in the utility industry is a function of two major factors: the membrane material and the configuration of the membrane module. Of the four

1ST STAGE 4 RQ Module

p—jj High Pressure Switch -»j—i

Micron Filter

O /"^ /

•— *

— ^; '

-.

FEED ^T

V

Pump

O

Pressure Gauoe

CJ Pressure Control

2

4 Total Product ' Flow Meter X__-^vJ RO Module ^-0~1T rft. Inter Stage /-vfr Shut-Off Valve

? ( p~~£*—^— V~V £ Pump Pres sure Control Va ive

Ur-J J,

P"3 Ntsnua! Valve n C*3 Solenoid Valve

u2R,

O —H h^ PRODUCT

LS=

__j RO Module Hr-i O" 2ND STAGE ¥*

9 1 fl

Legend

9

'€

il

x

£^3

I

1

V

-o- REJEC r

Reject *- Reject Control Valve Flow f i/Ieter

Sampling Port XxZ Flow Meter (Water and/Of Pressare! | j

. Typical reverse osmosis process.

• Second Stage Product Flow Meter

I ft B.

138

Liquid Filtration

RO membrane module types, most utility applications use either spiral-wound or hollow-fiber elements. Hollow-fiber elements are particularly prone to fouling and, once fouled, are hard to clean. Thus, applications that employ these fibers require a great deal of pretreatment to remove all suspended and colloidal material in the feed stream. Spiral-wound modules, due to their relative resistance to fouling, have a broader range of applications. A major advantage of the hollow-fiber modules, however, is the fact that they can pack 5000 ft2 of surface area in a 1 f t 3 volume, while a spiral wound module can only contain 300 ft2/ft3. The hollow fine fiber configuration consists of a bundle of porous hollow fine fibers. These fibers are externally coated with the actual membrane and form the support structure for it. Both ends of each fiber are set in a single epoxy tube sheet, which includes an O-ring seal to match the inside diameter of the pressure vessel. Influent water enters one end of the pressure vessel and is evenly distributed along the length of the vessel by a concentric distributor tube. As the water migrates out radially, some of it permeates the fibers and exits the pressure vessel via the tube sheet on the opposite end. The direction of permeate flow is from outside to inside the fibers. The concentrated solution, or reject, completes its radial flow path and leaves the vessel at the same end at which it entered. Figure 32 is a representation of this configuration. For clarity, the vessel and inlet distributor have been omitted. The actual outside diameters of individual fibers range 3-10 mils, depending on manufacturer. Figure 33 depicts a complete module. The spiral-wound configuration consists of a jelly roll-like arrangement of feed transport material, permeate transport material and membrane material. At the heart of the wall is a perforated permeate collector tube. Several rolls are usually placed end to end in a long pressure vessel. Influent water enters one end of the pressure vessel and travels longitudinally down the length of the vessel in the feed transport layer. Direct entry into the permeate transport layer is precluded by sealing this layer at each end of the roll. As the water travels in a longitudinal direction, some of it passes in radiaily through the membrane into the permeate transport layer. Once in the transport layer, the purified water flows spirally into the center collection tube and exits the vessel at each end. The concentrated feed continues along the feed transport material and exits the vessel on the opposite end from which it entered. A cross section of the spiral configuration is depicted in Figure 34 and a typical module assembly is shown in Figure 35. The two types of membrane materials used are cellulose acetate and aromatic polyamide membranes. Cellulose acetate membrane performance is particularly susceptible to annealing temperature, with lower flux and higher rejection rates at higher temperatures. Such membranes are prone to hydrolysis at extreme pH, are subject to compaction at operating pressures, and are sensitive to free chlorine above 1.0 ppm. These membranes generally have a useful life of 2-3 years. Aromatic polyamide membranes are prone to compaction. These fibers are more resistant to hydrolysis than are cellulose acetate membranes, but they are more sensitive to free chlorine.

139

Industrial Filtration Equipment

Pprous Block End Plato

Rotated

Plf»*

Spacer Fiber Bundle Porous Distributor

End Plato

0' Ring

Figure 32. Hollow-fiber module.

PERMEATE TRANSPORT LAYER PERMEATE TUBE

FEED TRANSPORT SPACER GLUE LINE MEMBRANE

Figure 33. Spiral-wound membrane.

141

Industrial Filtration Equipment

Intereonnector ft O-Rlng* Sealed,End Cap

Purified Water

On

Outlet

Product Tube

Concentre! Outlet Purified Water

Outlet

Concentrate Outlet Figure 35. Spiral-wound module.

Closure

The range of different filtration equipment is broad and it is difficult to generalize selection criteria. Machinery selection depends largely on the application, the properties of the slurry, the degree of separation or intended efficiency, throughput capacities and solids loadings, and the economics of the process. Economic considerations should include the capital investment in the filtration unit and supporting equipment, operating costs (in particular, energy costs), maintenance requirements (including estimated life expectancy of parts and costs for replacement), and labor (operator attention time and cost of training qualified operators). Chapter 6 provides an analysis for cost estimating filtration systems. Although the analysis is presented for a specific filtration technology, the reader can readily generalize the analysis for application to other filtration technologies.

5

APPLICATION OF FILTRATION TO WASTEWATER TREATMENT Introduction

This chapter provides an overview of applications of filtration operations to waste water treatment applications. The most widely used filtration application in wastewater treatment is granular media filtration, although other methods are also used. Filtration may be applied as the primary treatment method, or more commonly as both a pre-treatment step and as a final or finishing stage, depending on the cleanup objectives and criteria. When employed as a finishing operation, the filtration process is referred to as polishing. Other operations that are often used with filtration include carbon adsorption, sedimentation, disinfection, biological methods, and others. The reader should consult the list of references at the end of this chapter for discussions on other unit operations used in wastewater treatment.

Granular Media Filtration

Granular media filtration is most often used for treating aqueous waste streams; the filter media consists of a bed of granular particles (typically sand or sand with anthracite or coal). The bed is contained within a basin and is supported by an underdrain system which allows the filtered liquid to be drawn off while retaining the filter media in place. As water containing suspended solids passes through the bed of filter medium, the particles become trapped on top of, and within, the bed. The filtration rate is reduced at a constant pressure unless an increase in the amount of pressure is applied to force the water through the filter. In order to prevent plugging, the filter is backflushed at high velocity to dislodge the particles. The backwash water contains high concentrations of solids and is sent to further treatment steps. 142

Application of Filtration to Wastewater Treatment

143

Filter application is typically applied to handling streams containing less than 100 200 mg/liter suspended solids, depending on the required effluent level. Increasedsuspended solids loading reduces frequent backwashing. The suspended solids concentration of the filtered liquid depends on particle size distribution, but typically, granular media filters are capable of producing a filtered liquid with a suspended solids concentration as low as 1 - 10 mg/1. Large flow variations will affect the effluent's quality. Granular media filters are usually preceeded by sedimentation in order to reduce the suspended solids load on the filter. Granular media filtration can also be installed ahead of biological or activated carbon treatment units to reduce the suspended solids load and in the case of activated carbon to minimize plugging of the carbon columns. Granular media filtration is only marginally effective in treating colloidal size particles in suspensions. Usually these particles can be made larger by flocculation although this will reduce run lengths. In cases where it is not possible to flocculate such particles (as in the case of many oil/water emulsions), other techniques such as ultrafiltration may be nessesary. Filtration is an effective means of removing low levels of solids from wastes provided the solids content does not vary greatly and the filter is backwashed at appropriate intervals. The operation can be easily integrated with other treatment steps, and further, is well suited to mobile treatment systems as well as on-site or fixed installations, A typical physical/chemical treatment system incorporates three "dual" medial (sand anthracite) filters connected in parallel in its treatment train. The major maintenance consideration with granular medial filtration is the handling'of the backwash. The backwash will generally contain a high concentration of contaminants and require subsequent treatment. In this application, the operations of precipitation and flocculation play important roles. Precipitation is a physiochemical process whereby some, or all, of a substance in solution is transformed into a solid phase. It is based on alteration of the chemical equilibrium relationships affecting the solubility of inorganic species. Removal of metals as hydroxides and sulfides is the most common precipitation application in wastewater treatment. Lime or sodium sulfide is added to the wastewater in a rapid mixing tank along with flocculating agents. The wastewater flows to a flocculation chamber in which adequate mixing and retention time is provided for agglomeration of precipitate particles. Agglomerated particles are then separated from the liquid phase by settling in a sedimentation chamber, and/or by other physical processes such as filtration. Precipitation is often applied to the removal of most metals from wastewater including zinc, cadmium, chromium, copper, fluoride, lead, manganese, and mercury. Also, certain anionic species can be removed by precipitation, such as phosphate, sulfate, and fluoride. Note that in some cases, organic compounds may form organometallic complexes with metals, which could inhibit precipitation. Cyanide and other ions in the wastewater may also complex with metals, making treatment by precipitation less efficient.

144

Liquid Filtration

The process of flocculation is applicable to aqueous waste streams where particles must be agglomerated into larger more settleable particles prior to sedimentation or other types of treatment. Highly viscous waste streams will inhibit the settling of solids. In addition to being used to treat waste streams, precipitation can also be used as an in situ process to treat aqueous wastes in surface impoundments. In an in situ application, lime and flocculants are added directly to the lagoon, and mixing, flocculation, and sedimentation are allowed to occur within the lagoon. Precipitation and flocculation can be integrated into more complex treatment systems. The performance and reliability of these processes depends greatly on the variability of the composition of the waste being treated. Chemical addition must be determined using laboratory tests and must be adjusted with compositional changes of the waste being treated or poor performance will result. Precipitation is nonselective in that compounds other than those targeted may be removed. Both precipitation and flocculation are nondestructive and generate a large volume of sludge which must be disposed of. Coagulation, flocculation, sedimentation, and filtration, are typically followed by chlorination in municipal wastewater treatment processes. Coagulation involves the addition of chemicals to alter the physical state of dissolved and suspended solids. This facilitates their removal by sedimentation and filtration. The most common primary coagulants are alum ferric sulfate and ferric chloride. Additional chemicals that may be added to enhance coagulation include activate silica, a complex silicate made from sodium silicate, and charged organic molecules called polyelectrolytes, which include large-molecular-weight polyacryl-amides, dimethyldiallylammonium chloride, polyamines, and starch. These chemicals ensure the aggregation of the suspended solids during the next treatment step-flocculation. Sometimes polyelectrolytes (usually polyacrylamides) are also added after flocculatioe and sedimentation as an aid to the filtration step. Coagulation may also remove dissolved organic and inorganic compounds. The hydrolyzing metal salts may react with the organic matter to form a precipitate, or they may form aluminum hydroxide or ferric hydroxide floe particles on which the organic molecules adsorb. The organic substances are then removed by sedimentation and filtration, or filtration alone if direct filtration or inline filtration is used. Adsorption and precipitation also removes inorganic substances. Note that flocculation is a purely physical process in which the treated water is gently stirred to increase interparticle collisions and, thus, promote the formation of large particles. After adequate flocculation, most of the aggregates will settle out during the 1 - 2 hours of sedimentation. The process of sedimentation involves the separation from water, by gravitational settling of suspended particles that are heavier than water. The resulting effluent is then subject to rapid filtration to separate out solids that are still suspended in the water. Rapid filters typically consist of 24 - 36 inches of 0.5- to 1-mm-diameter sand and/or anthracite. Particles are removed as water is filtered through the media at rates

Application of Filtration to Wastewater Treatment

145

of 1 - 6 gallons/minute/square foot. Rapid filtration is effective in removing most particles that remain after sedimentation. The substances that are removed by coagulation, sedimentation, and filtration accumulate in sludges which must be properly disposed of. Coagulation, flocculation, sedimentation, and filtration will remove many contaminants. Perhaps most important is the reduction of turbidity. This treatment yields water of good clarity and enhances disinfection efficiency. If particles are not removed, they harbor bacteria and make final disinfection more difficult.

FILTER TANK

GRADED GRAVEL

PERFORATED LATERALS FILTER FLOOR CAST-IRON MANIFOLD

Figure 1. Cutaway view of a rapid sand filter.

The hydraulic performances required of the sand with slow filters are inferior to those for rapid filters. In the case of slow filters, one can use fine sand, since the average filtration velocity that is usually necessary lies in the range 2 - 5 in/day. In slow filtration, much of the effect is obtained by the formation of a filtration layer, including the substances that are extracted from the water. At the early stages of the operation, these substances contain microorganisms able to effect, beyond the filtration, biochemical degradation of the organic matter. This effect also depends on the total surface of the grains forming the filter material. The probability of contact between the undesirable constituents of the water and the surface of the filter medium increases in proportion to the size of the total surface of the grains.

146

Liquid Filtration

The actual diameter of the sands used during slow filtration typically lies between 0.15 and 0.35 mm. It is not necessary to use a ganged sand. The minimum thickness of the layer necessary for slow filtration is 0.3 - 0.4m, and the most efficient filtration thickness typically is at 2 - 3 cm. The actual requirements for the sand in slow filtration are chemical in nature. Purity and the absence of undesirable matters are more important than grain-size distribution in the filtration process. On the other hand, the performance of rapid filters requires sands with quite a higher precise grain size. In the case of rapid filtration, the need for hydraulic performances is greater than in slow nitration. This means that the grain-size distribution of the medium is of prime concern in the latter case. Sand often contains undesirable impurities, and additionally it can have broad particle size distributions. Sand that is used in filtration must be free of clay, dust, and other impurities. The ratio of lime, lime-stone, and magnesium oxide will have to be lower than 5 weight percent. The standard guide value of the quality of fresh sand is to be below 2% soluble matter at 20 °C within 24 hours in hydrochloric acid of a 20 weight percent concentration. In waste water treatment plants, the purity of the sand media used must be examined regularly. In addition, both the head loss of the filter beds and an analysis of the wash water during the operation of washing the filters must be checked regularly. Special attention must also be granted to the formation of agglomerates. The presence of agglomerates is indicative of insufficient washing and the possible formation of undesirable microbiological development zones within the filter bed. The primary mechanisms that control the operation of sand filtration are: Straining Settling Centrifugal action Diffusion Mass attraction, or the effect of van der Waals forces Electrostatic attraction Straining action consists of intercepting particles that are larger than the free interstices left between the filtering sand grains. Assuming spherical grains, an evaluation of the interstitial size is made on the basis of the grains' diameter (specific diameter), taking into account the degree of nonhomogeneity of the grains. Porosity constitutes a important criterion in a description based on straining. Porosity is determined by the formula VL/VC, in which Vc is the total or apparent volume limitated by the filter wall and VL is the free volume between the particles. The porosity of a filter layer changes as a function of the operation time of the filters. The grains become thicker because of the adherence of material removed from the water, whether by straining or by some other fixative mechanism of particles on the filtering sand. Simultaneously the interstices between the grains dimmish in size. This effect assists the filtration process, in particular for slow sand filters, where a deposit is

Application of Filtration to Wastewater Treatment

147

formed as a skin or layer of slime that has settled on the bed making up the active filter. Biochemical transformations occur in this layer as well, which are necessary to make slow filters efficient as filters with biological activity. Filtration occurs correctly only after buildup of the sand mass. This formation includes a "swelling" of the grains and, thus, of the total mass volume, with a corresponding reduction in porosity. The increases and swellings are a result of the formation of deposits clinging to the empty zones between grains. The porosity of a filter mass is an important factor. This property is best defined by experiment. A general rule of thumb is that for masses with the effective size greater than 0.4-0.5 mm and a specific maximum diameter below 1.2 mm the porosity is generally between 40 and 55 % of the total volume of the filter mass. Layers with spherical grains are less porous than those with angular material. The second important mechanism in filtration is that of settling. From Stoke's law of laminar particle settling, the settling velocity of a particle is given by :

18 v p

where : p p+Ap D g v

= = = = =

w

volumetic mass density of the water volumetic mass density of the particles in suspension diameter of the particles 9.81m/s2 kinematic viscosity (e.g., 104 m/s at 20°C)

In sedimentation zones the flow conditions are laminar. A place is available for the settling of sludges contained in the water to be filtered. Although the total inner surface that is available for the formation of deposits in a filter sand bed is important, only a part of this is available in the laminar flow zones that promote the formation of deposits. Usually material with a volumetic mass slightly higher than that of water is eliminated by sedimentation during filtration. Such matter could be, for example, organic granules or particles of low density. In contrast, colloidal material of inorganic origin-sludge or clay, for instance—with a diameter of 1 - 1 0 (Jim is only partially eliminated by this process, in which case the settling velocities in regard to the free surface become insufficient for sedimentation. The trajectory followed by water in a filter mass it is not linear. Water is forced to follow the outlines of the grams that delineate the interstices. These changes in direction are also imposed on particles in suspension being transported by the water. This effect leads to the evacuation of particles in the dead flow zones. Centrifugal action is obtained by inertial force during flow, so the particles with the highest volumetic mass are rejected preferentially.

148

Liquid Filtration

Diffusion filtration is another contributor to the process of sand filtration. Diffusion in this case is that of Brownian motion obtained by thermal agitation forces. This compliments the mechanism in sand filtration. Diffusion increases the contact probability between the particles themselves as well as between the latter and the filter mass. This effect occurs both in water in motion and in stagnant water, and is quite important in the mechanisms of agglomeration of particles (e.g., flocculation). The next mechanism to consider is the mass attraction between particles which is due to van der Waals forces. These are universal forces contributing to the transport and fixation mechanism of matter. The greater the inner surface of the filters, the higher is the probability of attractive action. Van der Waals forces imply short molecular distances, and generally play a minor role in the filtration process. Moreover, they decrease very quickly when the distance between supports and particles increases. Nevertheless, the indirect effects, which are able to provoke an agglomeration of particles and, thus, a kind of flocculation, are not to be neglected and may become predominant in the case of flocculation-filtration, or more generally in the case of filtration by flocculation. Electrostatic and electrocinetic effects are also factors contributing to the filtration process. Filter sand has a negative electrostatic charge. Microsand in suspension presents an electrophoretic mobility. The value of the electrophoretic mobility, or of the corresponding zeta potential, depends on the pH of the surrounding medium. Usually a coagulation aid is used to condition the surface of microsand. In filtration without using coagulant aids, other mechanisms may condition the mass more or less successfully. For instance, the formation of deposits of organic matter can modify the electrical properties of the filtering sand surfaces. These modifications promote the fixation of particles by electrokinetic and electrostatic processes, especially coagulation. Also, the addition of a neutral or indifferent electrolyte tends to reduce the surface potential of the filtering sand by compression of the double electric layer, This is based on the principles of electrostatic coagulation. The sand, as the carrier of a negative charge spread over the surface of the filter according to the model of the double layer, will be able to fix the electropositive particles more exhaustively. This has a favorable effect on the efficiency of filtration of precipitated carbonates or of floes of iron or aluminum hydroxide-oxide. Optimal adherence is obtained at the isoelectric point of the filtrated material. In contrast, organic colloidal particle carriers of a negative charge such as bacteria are repulsed by the electrostatic mechanism in a filter with a fresh filter mass. In this case, the negative charges of the sand itself appear unchanged. With a filter that is conditioned in advance, there are sufficient positively charged sites to make it possible to obtain an electrochemical fixation of the negative colloids.

Bed Regeneration In addition to washing the bed, a degradated mass containing agglomerates or fermentation zones (referred to as mud balls) can be regenerated by specific treatment techniques. Among the regeneration techniques that are usually used are sodium

Application of Filtration to Wastewater Treatment

149

chloride, regeneration through application of chlorine, and treatment with potassium permanganate, hydrogen peroxide, or caustic soda. Cleaning methods based on the use of caustic soda are aimed at eliminating thin clay, hydrocarbons, and gelatinous aggregates that form in filtration basins. After the filter has been carefully washed with air and water or only with water, according to its specific operating scheme, a quantity of caustic soda is spread over a water layer approximately 30 cm thick above the filter bed. The solution is then diffused in the mass by slow infiltration. After about 6 - 1 2 hours, the filter is washed very carefully. Sodium chloride is used specifically for rapid filters. The cleaning solution is spread in solution form in a thin layer of water above the freshly washed sand bed. After 2 or 3 hours of stagnation, slow infiltration in the mass is achieved by opening an outlet valve for the filtered water. The brine is then allowed to work for about a 24 hour period. The filter is placed back into service after a thorough washing. Sodium chloride works on proteinic agglomerates, which are bacterial in origin. The use of potassium permanganate (KMnO4) is applied to filters clogged with algae. A concentrated solution containing potassium permanganate is spread at an effective concentration over the surface of the filters to obtain, a characteristic pink-purple color on the top of the mass and allowed to infiltrate the bed for a 24 hour period. After this operation, the filter is carefully washed once again. Hydrogen peroxide is typically used in the range of 10 - 100 ppm. The cleaning method is similar to that used for permanganate. The addition of phosphates or polyphosphates makes it easier to remove ferruginous deposits. This method can be used in situ for surging the isolation sands of the wells. Adjunction of a reductor as bisulfite can be useful to create anaerobic conditions for the elimination of nematodes and their eggs when a filter has been infected. Hydrochloric acid solution is applied to the recurrent cleaning of rapid filters for sand, iron, and manganese removal. This operation has the advantage of causing the formation of chlorine in situ which acts as a disinfectant. Instantaneous cleaning of a filtering sand bed can be accomplished by the use of chlorine. A water layer is typically used as a dispersion medium. Further infiltration of the solution is obtained by percolation into the bed. The action goes on for several hours, after which the filter is washed. Chlorine is used from concentrated solutions of sodium hypochlorite. An alternative method involves the application of dioxide. This method has the advantage of arresting the formation of agglomerates of biological origin by permanent treatment of the filter wash water with chlorine.

Fiocculation Filtration The sand filtration process is normally comprised of a clarification chain including other unit operations which precede filtration in the treatment sequence and can not

150

Liquid Filtration

be conceived of completely independent of the filtration stage. The conventional treatment scheme consists of coagulation-flocculation-settling followed by filtration. When the preceding process, in this case flocculation and/or settling, becomes insufficient, subsequent rapid filtration can be used to ensure a high quality of the effluent treated. However, this action is achieved at the expense of the evolution of filter head loss. Problems in washing and cleanliness of the mass may arise. Filtration is often viewed as serving as a coagulant flocculator. This is referred to as flocculation-filtration. The presence of thin, highly electronegative colloids (e.g., activated carbons) introduced in the form of powder in the settling phase may be a problem for the quality of the settled effluent. The carbon particles, which are smaller than 50 /^m, penetrate deeply into the sand filter beds. They may rapidly provoke leakage of rapid filters. The same holds for small colloids other than activated carbon. Activated silica, which may have a favorable or an unfavorable effect on filtration, is composed of ionized micella formed by polysilicic acid-sodium polysilicate. This become negatively charged colloidal micella. The behavior of activated silicas depends on the conditions of neutralization and the grade of the silicate used in the preparation of the material. Activated silica is a coagulant aid that contributes to coalescence of the particles. Hence, it brings about an improvement in the quality of settled or filtrated water, depending on the point at which it is introduced. Preconditioning of the sand surface of filters by adding polyelectrolytes is an alternative use of sand filters as a coagulator-flocculator. In the treatment of drinking water the method depends on the limitations of these products in foodstuffs. The addition of polyphosphates to a water being subjected to coagulation usually has a negative effect; specifically the breaking of the agglomeration velocity of the particles during flocculation will occur in sand filtration. The addition of polyphosphates simultaneously with phosphates can be of value in controlling corrosion. This sometimes makes it possible to avoid serious calcium carbonate precipitation at the surface of filter grains when handling alkaline water. The application concerns very rapidly incrasting water while maintaining high hardness in solution. The addition of polyphosphates involves deeper penetration of matter into the filter mass. Hence, the breaking of flocculation obtained by the action of polyphosphates enables the thinner matters to penetrate the filters more deeply. These products favor the "in-depth effects" of the filter beds. Their use necessitates carefully checking that they are harmless from a hygienic point of view. The depth penetration of material in coagulation-filtration is almost opposite to the concept of using the filter as a screen. Precipitation initiated by germs plays a significant role. Empirical relations are normally relied on in the design of filters as a function of the penetration in depth of coagulated material. The concentration of those residual matters in filtered water (Cf) depends on several factors: the linear infiltration rate (v t ), the effective size of the filter medium (ES), the porosity of the filter medium (e), the final loss of head of the filter bed (Ah), the depth of penetration of the coagulated matter (/), the concentration of the particles in suspension in the

Application of Filtration to Wastewater Treatment

15!

water to be filtered (C0), and the water height (H). The following generalized relation is often found among the filtration engineer's notes. = / ( vfx(ES)

xQx//

(2)

It should be noted lhat the total loss of head of a filter bed is in inverse ratio to the depth of penetration of the matter in suspension. In a normal wastewater treatment plant, the water is brought onto a series of rapid sand filters and the impurities are removed by coagulation-flocculation-filtration. Backwashing is typically performed in the counterflow mode, using air and water. One type of common filter is illustrated in Figure 2, consisting of closed horizontal pressurized filters. HITRATIQN

PHASE

PHASE

lir *i!ve

/' \

I2L./ I \ X flitcren

nishhnl

Figure 2. Cross section of a typical filtration unit.

Slow Sand Filtration

Slow sand filtration involves removing material in suspension and/or dissolved in water by percolation at slow speed. In principle, a slow filter comprises a certain volume of area! surface, with or without construction of artificial containment, in which filtration sand is placed at a sufficient depth to allow free flow of water through the bed. When the available head loss reaches a limit of approximately 1 m, the filter must be pulled out of service, drained, and cleaned. The thickness of the usual sand layer is approximately of 1 - 1.50 m, but the formation of biochemically active deposits and clogging of the filter beds takes place in the few topmost centimeters of the bed.

152

Liquid Filtration

The filter mass is pored onto gravels of increasing permeability with each layer having a thickness of approximately 10 - 25 cm. The lower-permeability layer can reach a total thickness of 50 - 60 cm. So-called gravels 18 - 36 cm in size are used and their dimensions are gradually diminished to sizes of 10 - 12 cm or less for the upper support layer. The sand filter must be cleaned by removal of a few centimeters of the clogged layer. This layer is washed in a separate installation. The removal of the sand can be done manually or by mechanical means. The removed sand may not be replaced entirely by fresh sand. Placing preconditioned and washed sand is recommended as this takes into account the biochemical aspects involved in slow filtration. An alternative to manual or mechanical removal involves cleaning using a hydraulic system as illustrated in Figure 3. WATER UNDER PRESSURE

11

-*«-—*>»-—*

ASPIRATION

Figure 3. Hydraulic cleaning device for slow sand filters.

Sometimes slow filtration is used without previous coagulation. This is generally practiced with water that does not contain much suspended matter. If the water is loaded (periodically or permanently) with clay particles in suspension, pretreatment by coagulation-flocculation is necessary. Previous adequate oxidation of the water, in this case preozonization producing biodegradable and metabolizable organic derivatives issuing from dissolved substances, can be favorable because of the biochemical activity in slow filters. There are several disadvantages to the use of slow filters. They may require a significant surface area and volume, and may therefore involve high investment costs.

Application of Filtration to Wastewater Treatment

153

They are also not flexibile — mainly during the winter, when the open surface of the water can freeze. During the summer, if the filters are placed in the open air, algae may develop, leading to rapid clogging during a generally critical period of use. Algae often cause taste and odor problems in the filter effluent. Additional construction costs to cover slow filters are often necessary.

Rapid Sand Filtration Rapid filtration is performed either in open gravitational flow filters or in closed pressure filters. Rapid pressure filters have the advantage of being able to be inserted in the pumping system, thus allowing use of a higher effective loading. Note that pressure filters are not subject to development of negative pressure in a lower layer of the filter. These filters generally support higher speeds, as the available pressure allows a more rapid flow through the porous medium made up by the filter sand. Pressure filtration is generally less efficient than the rapid open type with free-flow filtration. Pressure filters have the following disadvantages. The injection of reagents is complicated, and it is more complicated to check the efficiency of backwashing. Work on the filter mass is difficult considering the assembly and disassembly required. Also, the risk of breakthrough by suction increases. Another disadvantages is that pressure filterts need a longer filtration cycle, due to a high loss of head available to overcome clogging of the filter bed. Another option is to use open filters, which are generally constructed in concrete. They are normally rectangular in configuration. The filter mass is posed on a filter bottom, provided with its own drainage system, including bores that are needed for the flow of filtered water as well as for countercurrent washing with water or air. There are several types of washing bottoms. One type consists of porous plates which directly support the filter sand, generally without a layer of support gravel. Even if the system has the advantage of being of simple construction, it nevertheless suffers from incrustation. This is the case for softened water or water containing manganese. Porous filters bottoms are also subject to errosion or disintegration upon the filtration of aggressive water. The filter bottom is often comprised of pipes provided with perforations that are turned toward the underpart of the filter bottom and embedded in gravel. The lower layers are made up of gravel of approximate diameter 35 - 40 mm, decreasing up to 3 mm. The filter sand layer, located above this gravel layer, serves as a support and equalization zone. Several systems of filter bottoms comprise perforated selfsupporting bottoms or false bottoms laid on a supporting basement layer. The former constitutes a series of glazed tiles, which includes bores above which are a series of gravels in successive layers. All these systems are surpassed to some extent by filter bottoms in concrete provided with strainers. The choice of strainers should in part be based on the dimensions of the slits that make it possible to stop the filter sand, which is selected as a function of the filtration goal. Obstruction or clogging occurs only rarely and strainers are sometimes used.

154

Liquid Filtration

Strainers may be of the type with an end that continues under the filter bottom. These do promote the formation of an air space for backwashing with air. If this air space is not formed, it can be replaced by a system of pipes that provide for an equal distribution of the washing fluids. Pressure filters are worth noting. These are usually set up in the form of steel cylinders positioned vertically. Another variation consists of using horizontal filtration groups. This has the drawback that the surface loading is variable in the different layers of the filter bed; moreover, it increases with greater penetration in the filter bed (the infiltration velocity is lowest at the level of the horizontal diameter of the cylinder). The filter bottom usually consists of a number of screens or mesh sieves that decrease in size from top to bottom or, as an alternative, perforated plates supporting gravel similar to that used in the filter bottoms of an open filter system. Filter mass washing can influence the quality of water being filtered. Changes may be consequent to fermentation, agglomeration, or formation of preferential channels liable to occur if backwashing is inadequate. Backwashing requires locating a source that will supply the necessary flow and pressure of wash water. This water can be provided either by a reservoir at a higher location or by a pumping station that pumps treated water. Sometimes an automated system is employed with washing by priming of a partial siphon pumping out the treated water stored in the filter itself. An example is shown in Figure 4. The wash water must have sufficient pressure to assure the necessary flow. Washing of the filter sands is accomplished followed by washing with water and in most cases including a short intermediate phase of simultaneous washing with air and water. Due to greater homogenization of the filter layer and more efficient washing, the formation of fermentation areas and agglomerates in the filter mass of treatment plants for surface water (mud balls) is diminished. The formation of a superficial crust on the filter sand is avoided by washing with air. After washing with air, water flow is gradually superimposed on the air flow. This operational phase ends at the same time that the wash air is terminated, to avoid the filter mass being blown away. The wash water contains materials that eventually require treatment in a sludge treatment plant. Their concentration varies as a function of the washing cycle. Accounting for the superficial load in filtration, velocity of the wash water, and length of the filtration cycle, it may be assumed that the water used for washing will not attain 5 % of the total production. For new installations the first washing cycles result in the removal of fine sand as well as all the other materials usually undesirable in the filter mass, such as particles of bitumen on the inner surface of the water inlet or other residuals from the crushing or straining devices of the filter media. Consequently, it is normal that at the beginning of operation of a filter sand installation, dark colored deposits appear at the surface of the filter mass. In the long term they have no consequence and disappear after a few

Application of Filtration to Wastewater Treatment

Figure 4.

155

Automatic backwashing filter with a partial siphon system: 1-filtered water

(reserve); 2-partial siphoning; 3-initiation; 4-restitution.

filtration and wash cycles. If, after several weeks of filtration, these phenomena have not disappeared, it will be necessary to examine the filter sand. The elimination of fine sand must stop after 1 or 2 months of activity. If this sand continues to be carried away after the first several dozen washings it is necessary to reexamine the hydraulic criteria of the washing conditions, the granulometry of the filter mass, and the filter's resistance to shear and abrasion.

Chemical Mixing, Flocculation and Solids Contact Processes Chemical mixing and flocculation or solids contact are important mechanical steps in the overall coagulation process. Application of the processes to waste water generally follows standard practices and employs basic equipment. Chemical mixing thoroughly disperses coagulants or their hydrolysis products so the maximum possible portion of influent colloidal and fine supracolloidal solids are absorbed and destabilized. Flocculation or solids contact processes increase the natural rate of contacts between particles. This makes it possible, within reasonable detention periods, for destabilized colloidal and fine supracolloidal solids to aggregate into particles large enough for effective separation by gravity processes or media filtration. These processes depend on fluid shear for coagulant dispersal and for promoting particle contacts. Shear is most commonly introduced by mechanical mixing equipment. In certain solids contact processes shear results from fluid passage upward

156

Liquid Filtration

through a blanket of previously settled particles. Some designs have utilized shear resulting from energy losses in pumps or at ports and baffles. Chemical Mixing Chemical mixing facilities should be designed to provide a thorough and complete dispersal of chemical throughout the wastewater being treated to insure uniform exposure to pollutants which are to be removed. The intensity and duration of mixing of coagulants with wastewater must be controlled to avoid overmixing or undermixing. Overniixing excessively disperses newly-formed floe and may rupture existing wastewater solids. Excessive floe dispersal retards effective flocculation and may significantly increase the flocculation period needed to obtain good settling properties. The rupture of incoming wastewater solids may result in less efficient removals of pollutants associated with those solids. Undermixing inadequately disperses coagulants resulting in uneven dosing. This in turn may reduce the efficiency of solids removal while requiring unnecessarily high coagulant dosages. In water treatment practice several types of chemical mixing units are typically used. These include high-speed mixers, in-line blenders and pumps, and baffled mixing compartments or static in-line mixers (baffled piping sections). An example of a highspeed mixer is shown in Figure 5. Designs usually call for a 10-30 second detention times and approximately 300 fps/ft velocity gradient. Variable-speed mixers are recommended to allow varying requirements for optimum mixing. In mineral addition to biological wastewater treatment systems, coagulants may be added directly to mixed biological reactors such as aeration tanks or rotating biological contactors. Based on typical power inputs per unit tank volume, mechanical and diffused aeration equipment and rotating fixed-film biological contactors produce average shear intensities generally in the range suitable for chemical mixing. Localized maximum shear intensities vary widely depending on the speed of rotating equipment or on bubble size for diffused aeration. Flocculation The proper measure of flocculation effectiveness is the performance of subsequent solids separation units in terms of both effluent quality and operating requirements, such as filter backwash frequency. Effluent quality depends greatly on the reduction of residual primary size particles during flocculation, while operating requirements relate more to the floe volume applied to separation units.

Application of Filtration to Wastewater Treatment

157

DRIVE MECHANISM MOTOR

SUPPORT BEAMS

FEED Figure 5. Example of an impeller mixer.

Flocculation units should have multiple compartments and should be equipped with adjustable speed mechanical stirring devices to permit meeting changed conditions. In spite of simplicity and low maintenance, non-mechanical, baffled basins are undesirable because of inflexibility, high head losses, and large space requirements. Mechanical flocculators may consist of rotary, horizontal-shaft reel units as shown in Figure 6. Rotary vertical shaft turbine units as shown in Figure 7 and other rotary or reciprocating equipment are other examples. Tapered flocculation may be obtained by varying reel or paddle size on horizontal common shaft units or by varying speed on units with separate shafts and drives, In applications other than coagulation with alum or iron salts, flocculation parameters may be quite different. Lime precipitates are granular and benefit little from prolonged flocculation.

Polymers which already have a long chain structure may provide a good floe at low mixing rates. Often the turbulence and detention in the clarifier inlet distribution is adequate.

Liquid Filtration

158

CONTROL VALVE

W.L. , PADDLES

INFLUENT

EFFLUENT

jr

irr^jL

..&.„. „..«,

JJ.

JL.

Figure 6. Mechanical flocculation basin horizontal shaft-reel type. MOTORIZED SPEED REDUCER

WAT Cft

PftESSURE tUBRICATCO

Figure 7. Mechanical flocculator vertical shaft-paddle type.

Application of Filtration to Wastewater Treatment

159

Solids Contacting Solids contact processes combine chemical mixing, flocculation and clarification in a single unit designed so that a large volume of previously formed floe is retained in the system. The floe volume may be as much as 100 times that in a "flow-through" system. This greatly increases the rate of agglomeration from particle contacts and may also speed up chemical destabilization reactions. Solids contact units are of two general types: slurry-recirculation and sludge-blanket. In the former, the high floe volume concentration is maintained by recirculation from the clarification to the flocculation zone, as illustrated in Figure 8. In the latter, the floe solids are maintained in a fluidized blanket through which the wastewater under treatment flows upward after leaving the mechanically stirred-flocculating compartment, as illustrated in Figure 9. Some slurry-recirculation units can also be operated with a sludge blanket. Solids contact units have the following advantages: 1. 2. 3. 4.

Reduced size and lower cost result because flocculation proceeds rapidly at high floe volume concentration. Single-compartment flocculation is practical because high reaction rates and the slurry effects overcome short circuiting. Units are available as compact single packages, eliminating separate units. Even distribution of inlet flow and the vertical flow pattern in the clarifier improve clarifier performances

Equipment typically consists of concentric circular compartments for mixing, flocculation and settling. Velocity gradients in the mixing and flocculation compartments are developed by turbine pumping within the unit and by velocity dissipation at baffles. For ideal flexibility it is desirable to independently control the intensity of mixing and sludge scraper drive speed in the different compartments. Operation of slurry-recirculation solids contact units is typically controlled by maintaining steady levels of solids in the reaction zone. Design features of solids contact clarifiers should include: 1.

2.

3. 4.

5.

Rapid and complete mixing of chemicals, feedwater and slurry solids must be provided. This should be comparable to conventional flash mixing capability and should provide for variable control, usually by adjustment of recirculator speed. Mechanical means for controlled circulation of the solids slurry must be provided with at least a 3:1 range of speeds. The maximum peripheral speed of mixer blades should not exceed 6 ft/sec. Means should be provided for measuring and varying the slurry concentration in the contacting zone up to 50 % by volume. Sludge discharge systems should allow for easy automation and variation of volumes discharged. Mechanical scraper tip speed should be less than 1 fpm with speed variation of 3:1. Sludge-blanket levels must be kept a minimum of 5 feet below the water surface.

RAPID MIXING AND RECIRCUUTtON

SLOW MIXING AND FLOC FORMATION

CHEMICAL INTRODUCTION

r •§'

\

a.

TREATED WATER EFFLUENT CLARIFIED WATER

SEDIMENTATION

SLUDGE RECIRCULATION SLUDGE REMOVAL

Figure 8. Solids contact clarifier without sludge blanket filtration.

Application of Filtration to Wastewater Treatment

161

162

6.

Liquid Filtration

Effluent launders should be spaced so as to minimize the horizontal movement of clarified water.

Further considerations include skimmers and weir overflow rates. Skimmers should be provided on all units since even secondary effluents contain some floatable solids and grease. Overflow rates and sludge scraper design should conform to the requirements of other clarification units. The reader may refer to Chapter 9 for examples of typical flow sheets and auxiliary filtration equipment schematics, including process flow sheets for chemical feeding operations described above.

Suggested Readings 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

Anon., Water Sewage Works, 6, 266 (1968). Maeckelburg, D., G.W.F., 119,23 (1978). O'Mella, Ch. R., and O.K. Crapps, J. AWWA, 56,1326 (1964). Drapeau, A.J., and R.A. Laurence, Eau Quebec, 10, 314 (1977). Burman, N.P., H2O, 11, 348, (1978). Cleasby, J.L., J. Arboleda, D.E. Burns, P.W. Prendiville, and E.S. Savage, J. AWWA, 69,115 (1977). Cheremisinoff, P.N., Pollution Engineering Flow Sheets: Waste water Treatment, Pudvan Publishing Co., Northbrook, IL, 1988. Cheremisinoff, N.P., Biotechnology for Waste and Wastewater Treatment, Noyes Publication, Park Ridge, NJ, 1996. Cheremisinoff, N.P. and P.N. Cheremisinoff, Carbon Adsorption for Pollution Control, Prentice Hall Publishers, Inc., Englewood, NJ ,1993. Cheremisinoff, N.P. and P.N. Cheremisinoff, Liquid Filtration for Process and Pollution Control, SciTech Publishers, Inc., Morganville, NJ, 1981. Cheremisinoff, N.P. and P.N. Cheremisinoff, Chemical and Non-Chemical Disinfection, Ann Arbor Science Publishers, Ann Arbor, MI, 1981. Cheremisinoff, P.N. and R.B. Trattner, Fundamentals of Disinfection for Pollution Control, SciTech Publishers, Inc., Morganville, NJ, 1990.

ADVANCED MEMBRANE TECHNOLOGY FOR WASTEWATER TREATMENT Introduction This chapter discusses a new membrane filtration system technology based on using a formed-in-place hyperfiltration membrane. The technology has been used to treat a creosote and pentachlorophenol (PCP) contaminated groundwater. The membrane technology described can be used as an integral part of a remediation system to significantly reduce the volume and toxicity of contaminated wastewater. The technology is particularly suited for the treatment of contaminated groundwater as part of a pump and treat system. The technology reduces risks to human health and the environment by transferring the contaminants to a smaller volume facilitating destruction or detoxification by other technologies. The technology is particularly applicable to the treatment of dilute waste steams, where the concentration of the contaminants into a reduced volume would result in significant cost savings as well as minimize off-site treatment. The reduced-volume concentrated residual could be further treated on-site, or transported off-site for treatment and disposal. The system is simple to operate, reliable and requires a minimum of operator attention or maintenance once the membrane has been formed. The stability of the system makes it particularly suitable for long-term use as is necessary for extended pump and treat remedial programs. The information provided in this chapter was largely obtained from a reported study by the United States Environmental Protection Agency (USEPA) from their Office of Research and Development in Washington, DC. The reader may contact the Risk Reduction Engineering Laboratory of the Office of Research and Development in Cincinnati, Ohio for detailed information on this filtration technology. A specific reference that the reader can refer for detailed information is EPA/540/AR-92/014 -

163

164

Liquid Filtration

August 1993 - Membrane Treatment of Wood Preserving Site Groundwater by SBP Technologies, Inc.: Application Analysis Report.

Overview of Technology Case Study Based on the results of a demonstration project at the American Creosote Works site in Pensacola, Florida and information concerning other studies provided by the vendor, SBP Technologies, Inc., for different wastes at other sites, several conclusions can be drawn. The conclusions are organized based on the evaluation factors of volume reduction and contaminant reduction. These factors are critical in applying the technology to other sites and wastes. The SBP filtration unit (as configured) effectively removed high molecular weight compounds from the feed stream, but smaller molecular weight compounds were not removed. The technology uses a formed-in-place membrane system which is quite effective (92%) at removing polynuclear aromatic hydrocarbons (PAHs) found in creosote from the feed water and producing a permeate with little of these materials. However, the membrane was found not to be very efficient at removing phenolics. Rejections were in the range of 18% for phenolics. Overall, based on a comparison of total concentrations of a pre-designated list of creosote-derived PAH and phenolic semivolatile contaminants in the permeate versus the feed water, the system did not meet the claimed rejection efficiency of 90%. On the basis of the PAH rejections of over 90%, the permeate would be expected to be acceptable for discharge to POTWs (Publically Owned Treatment Works) with little or no polishing. Other pollutants found in contaminated waters at wood treatment facilities (e.g., polychlorinated dioxins and furans) also are concentrated in the reject stream. Other constituents commonly encountered at such sites including colloidal oils and suspended solids are also extensively removed by the membrane process. Removal efficiencies for oil and grease were 93%. Suspended solids were removed to nondetectable levels. These materials did not appear to have an adverse effect on the filtration process. The system was found to effectively concentrate organic contaminants into a concentrate of much smaller volume. The volume of wood preserving waste contaminated wastewater was reduced by over 80%. This means that only 20% of the volume of the feed water would require further treatment to immobilize or destroy the organic contaminants. The filtration unit operated consistently and reliably over a brief testing period. The unit was easy to operate and maintain. The filtration unit operated in a batch mode for six hours each day, for six days, and processed approximately 1000 gallons of feed per day. Over the six day test period, permeate flux was relatively constant. Based on a total membrane area of 300 ft2 for the system, the permeate flow rate for the four module filtration unit averaged 2.6 gpm. Excessive fouling of the membrane, necessitating frequent cleaning or regeneration, was not encountered. However, the

Advanced Membrane Technology for Wastewater Treatment

165

membrane system did exhibit a gradual and controllable fouling which required periodic cleaning. The operating cost for the membrane process as used at American Creosote Works is in the range of $220 - $1,740/1,000 gallons, depending on system size. Major cost contributors are labor and residuals disposal. Labor costs decrease significantly as the scale of the process increases. Auxiliary equipment that could be needed to support this process is comparable to that which would be needed for other above-ground treatment systems such as oil/water separators and clarifiers for pretreatment, and filters, carbon adsorbers, etc, for effluent polishing as required. With membranes similar to those manufactured for the American Creosote Works site, the system could be well suited for the concentration of polynuclear aromatic hydrocarbons from wastewaters (groundwater, process wastes, lagoon leakage, etc.) found at coke plants, wood preserving sites, and some chemical plants. Based on the expected mechanisms of membrane filtration, the technology also may be useful for wastewaters containing other large molecules such as polychlorinated biphenyls (PCBs) and polychlorinated dioxins and furans, particularly where these are associated with oil or particulate matter. It probably is also highly effective for oils, colloidal solids, and greases. According, to the developer, the formed-in-place membrane can be easily modified to conform to waste characteristics and the degree of contaminant removal desired. Therefore, the membrane can be tailored to the unique characteristics of the waste steam. Extensive data were collected on primary pollutants (phenols, and PAHs) and on secondary pollutants (oil, suspended and dissolved solids, COD, dioxins, and VOC's). The results of this project demonstrated the ability of the formed-in-place membrane, operating in a cross-flow mode, to minimize fouling, and to remove polynuclear aromatic hydrocarbons from the contaminated feed water. As operated, rejection of the PAHs appears to increase with the number of aromatic rings. However, similar correlations appear to exist with molecular weight as well as with the partition coefficient reflecting hydrophobicity. The permeate, accounting for approximately 80% of the feedwater, contained only about 12% of the predominant PAHs, naphthalene and phenanthrene. The removal of phenol and methyl phenols was not comparably high under the conditions of the demonstration, with an average rejection of 18%. The concentrations of phenolics in the permeate could present a regulatory problem in the United States, depending on the concentrations in the feedwater and the final disposition of the permeate. However, the vendor states that different membranes and tube configurations could resolve this. Secondary constituents, such as oil, suspended solids, and dissolved solids, did not appear to interfere with the operation of the process at the concentrations present in the waste water studied during the demonstration. Decreases in chemical oxygen demand (COD), total organic carbon (TOC) and oil and grease (O&G) indicated that

166

Liquid Filtration

the system removes other organic species as well as PAHs, but not necessarily with the same efficiency. The SBP membrane process would be most applicable to wastewaters containing large molecular weight organic compounds (PAHs, dioxins/furans, polychlorinated biphenyls, and certain pesticides/herbicides). The system can remove smaller molecular weight compounds (phenols, benzene, toluene, ethylbenzene, xylenes) if larger molecular weight compounds are not abundantly present. Removal of smaller molecular weight compounds can be accomplished by modifying the structure of the formed-in-place membrane. For these applications, the pores of the membrane are reduced, resulting in higher retentions of smaller components as well as a reduction in the flux (throughput) of the system. To compensate for the reduced flux, either additional membrane modules can be added or more time will be required to accomplish the remediation. In either case, the overall cost may be higher. The system may be most suitable to treating relatively dilute, but toxic, waste streams in which the percent reduction of contaminants will allow discharge of the permeate without further treatment. This feature makes the unit highly suitable for polishing effluents as part of a multi-technology treatment train. In this system, the primary treatment technology can be utilized to remove the bulk of the contamination, with the filtration unit being used as a final polishing step. A major attribute of the system is its ability to minimize fouling. The system effectively controlled excessive fouling, in spite of the problematical nature of the wood preserving waste feed, through a combination of cross-flow operation and membrane cleaning. The membrane cleaning process effectively regenerated the membrane to its original clean permeate flux conditions. This enabled the membrane to be reused, without the necessity to reformulate. The ability to repeatedly regenerate the flux after the cleaning procedure is a good indication that the forrned-in-place membrane is stable and can be used over an extended length of time. In the unlikely event of an irreversible fouling, the membrane can be cost-effectively and easily reformed on-site with a minimum of downtime. The technology uses a proprietary formed-in-place membrane technique. The membrane is formed on porous sintered stainless steel tubes by depositing microscopic layers of inorganic and polymeric chemicals. The properties of the formed-in-place membrane can be varied by controlling the type of membrane chemicals used, their thickness, and the number of layers. This important feature allows for customization of the membrane system to a wide variety of waste characteristics and clean-up criteria. The formed-in-place membrane can be quickly and economically reformulated in the field to accommodate changes in waste characteristics or treatment requirements. The formed-in-place membrane is compatible with a wide variety of contaminants often encountered in hazardous wastewater streams. The formed-in-place membrane is stable under most chemical environments and will not degrade even at high contaminant concentrations.

Advanced Membrane Technology for Wastewater Treatment

167

The extent of contaminant reduction required (overall and for individual pollutants) can also be an important factor in system design and operation. This will impact membrane selection, and operational requirements such as the number of cycles necessary to achieve the targeted volume reduction. Generally, as the desired level of volume-reduction increases, the overall quality of the permeate decreases, so a balance must be maintained between throughput and permeate quality. This will also affect the throughput capability (as permeate) for a particularly sized system. Other factors that could affect the removal of PAHs or other contaminants may include the presence of other organics, oil and grease, suspended solids, and dissolved solids in the feed water. While the levels of such contamination encountered in the demonstration project had no apparent adverse effect, it is unclear how much rejection (of PAHs) was due to molecular size or weight and how much was due to solubility in oil that was rejected and coalesced by the membrane. Additional or alternative mechanisms also may be operative. Case Study Specifics The EPA's Office of Solid Waste and Emergency Response (OSWER) and the Office of Research and Development (ORD) established the Superfund Innovative Technology Evaluation (SITE) Program in 1986 to promote the development and use of innovative technologies to clean up Superfund sites across the country. The SITE Program is helping to provide the treatment technologies necessary to meet new federal and state cleanup standards in the United States that are aimed at permanent, rather than temporary, remedies. The SITE Program is composed of two major elements: the Demonstration Program and an Emerging Technologies Program. In addition, the Program includes research on analytical methods that can expedite cleanups at Superfund sites. The USEPA demonstration programs are designed to provide engineering and cost data on selected technologies. EPA and the developers participating in the program share the cost of demonstrating their innovative systems at chosen sites, usually Superfund sites. Developers are responsible for the operation of their equipment (and related costs). EPA is responsible for sampling, analyzing, and evaluating all test results and comparing these results to claims originally defined by the developer. The result is an assessment of the technology's performance, reliability, and cost. In addition to providing the developer with carefully documented information useful in marketing, the information, in conjunction with other data, also will be used to select the most appropriate technologies for the cleanup of other Superfund sites. Developers of innovative technologies apply to the Demonstration Program by responding to EPA's annual solicitation. To qualify for the program, a new technology must have a pilot or full scale unit and offer some measurable advantage over existing technologies. Mobile technologies are of particular interest to EPA. Once EPA has accepted a proposal, EPA and the developer work with the EPA Regional offices and state agencies to identify a site containing wastes suitable for testing the capabilities of the technology. EPA's contractor designs a detailed sampling

168

Liquid Filtration

and analysis plan that will thoroughly evaluate the technology and ensure that the resulting data are reliable. The duration of a demonstration varies from a few days to several months, depending on the type of process and the quantity of waste needed to assess the technology. While meaningful results can be obtained in a demonstration lasting one week with some technologies, others, may require months. On completion of a demonstration, EPA prepares reports. Ultimately, the Demonstration Program leads to an analysis of the technology's overall applicability to Superfund problems. The second principal element of the SITE Program is the Emerging Technologies Program, which fosters the investigation and development of treatment technologies which are still at the laboratory scale. Successful validation of these technologies could lead to the development of systems ready for field demonstration. A third component of the SITE Program, the Measurement and Monitoring Technologies Program, provides assistance in the development and demonstration of innovative techniques and methods for better characterization of Superfund sites. In this study it was demonstrated that SBP's membrane technology can be used as an integral part of a remediation system to significantly reduce the volume and toxicity of contaminated wastewater. The technology is particularly suited for the treatment of contaminated groundwater as part of a pump and treat system. The technology reduces risks to human health and the environment by transferring the contaminants to a smaller volume facilitating destruction or detoxification by other technologies. The vendor uses a proprietary formed-in-place membrane technology. The membrane is formed on porous sintered stainless steel tubes by depositing microscopic layers of inorganic and polymeric chemicals. The properties of the formed-in-place membrane can be varied by controlling the type of membrane chemicals used, their thickness, and the number of layers. This important feature allows for customization of the membrane system to a wide variety of waste characteristics and clean-up criteria. The formed-in- place membrane can be quickly and economically reformulated in the field to accommodate changes in waste characteristics or treatment requirements. Contaminated feedwater is recirculated through the filtration unit until the desired level of volume reduction is attained. The filtration unit generates two process waste streams. A relatively clean stream, called the "permeate", passes through the membrane while a smaller portion of the feedwater, retaining those species that do not pass through the membrane, is retained in a stream called the "concentrate". The permeate stream should be clean enough for disposal as a non-hazardous waste with little or no additional treatment. The concentrate would require further treatment to immobilize or destroy the contaminants.

Technology Application This technology lends itself as a means of concentrating organic contaminants in aqueous waste streams. The prime benefit of concentrating contaminants is to minimize costly treatment of the entire wastestream. In addition, by concentrating the

Advanced Membrane Technology for Wastewater Treatment

169

organic contaminants into a smaller volume, alternative treatment technologies may be feasible based on technical and/or economic criteria. The ability of the filtration unit to concentrate organic contamination from aqueous waste streams was demonstrated on a groundwater contaminated with wood preserving wastes (phenolics, PAHs, and PCP). The results from the demonstration, in conjunction with information supplied by the vendor, were used to assess the applicability of the technology for a variety of waste types and site conditions. The process uses a formed-in-place hyperfiltration membrane on a stainless steel support to separate and concentrate higher molecular weight contaminants. Contaminated groundwater (feed) is pumped through the modules under pressure. A portion of the feed passes through the formed-in-place membrane forming a permeate. The membrane retains certain contaminants resulting in a permeate that is clean relative to the feed. The bulk of the contamination remains in the "concentrate" fraction. The concentrate is recycled through the unit until the desired concentration or level of volume reduction is attained, or the level of contaminants in the recycling concentrate inhibits the filtration process (fouling). The system relies on cross-flow filtration to minimize fouling of the membrane and, thus, maximize throughput. The properties of the two process streams (permeate and concentrate) are of particular importance since these characteristics define waste disposal options. The permeate stream should exhibit significant reductions in contamination so as to allow economical discharge to local wastewater treatment facilities without extensive pretreatment requirements. The concentrate stream should be volumetrically small, relative to the original feed, in order to minimize the volume of waste requiring further treatment prior to disposal. Furthermore, the filtration process should enable the use of additional disposal options for the concentrate (as compared to the raw feed). The following subsections summarize observations and conclusions drawn from the reported study. Included in the discussion are factors such as the application of membrane processes for wastewater reduction, benefits of the system, other applicable waste waters, site characteristics and constraints, and unique handling requirements.

Mechanisms of Membrane Separations Membranes are semi-permeable barriers that are used to isolate and separate constituents from a fluid stream. The separation process can be accomplished through a number of physical and chemical properties of the membrane as well as the material being separated. Separation can occur through processes such as size, ionic charge, solubility, and combinations of several processes. Membranes can remove materials ranging from large visible particles to molecular and ionic chemical species. Membrane materials are diverse and can consist of synthetic polymers, natural fabrics, porous metals, porous ceramics, or liquids. The surface of the membrane can be chemically or biologically altered to perform separations on specific chemical

170

Liquid Filtration

compounds. The interaction of the components of the fluid stream with the membrane is the mechanism controlling the outcome of the separation process. There are two basic modes of membrane separation. In dead-end filtration specific species are trapped within the matrix of the membrane material. The membrane "filters-out" these species producing a relatively clean effluent. In dead-end filtration the components that are trapped are usually not recovered and remain within the membrane matrix. In addition, the membrane eventually becomes plugged necessitating the replacement of the membrane. Dead-end filtration is principally utilized to purify a fluid in applications where the removed species is relatively dilute. In cross-flow filtration the fluid steam is directed parallel to the surface of the membrane. This action inhibits the accumulation of components within the matrix of the membrane. The cross-flow action of the fluid keeps the surface of the membrane clean allowing for the passage of species smaller than the pores of the membrane. Cross-flow filtration produces two effluent streams. The permeate is the steam that passes through the membrane and is relatively depleted in species larger than the pore size of the membrane. The concentrate is the cross-flow stream that contains the larger species that are unable to pass through the membrane and accumulate. The concentrate can be recycled allowing for progressive concentration of species over time. Due to the ability of the cross-flow system to concentrate components from the feed stream, it is commonly used as a method to separate and recover these components. Furthermore, the cross-flow action minimizes plugging of the membrane (fouling) by constantly sweeping the membrane's surface. This cleaning action extends the life of the membrane and minimizes degradation of flow through the membrane. Membrane systems have many applications for the pretreatment and treatment of hazardous wastes. Membrane separation is a volume reduction technology. This technology can separate and concentrate specific contaminants from a waste stream, resulting in a significant reduction in the volume of waste requiring treatment. The concentrated contaminants can then be destroyed or rendered non-toxic. The utility of a membrane based technology is based on its ability to reduce the volume of waste by removing contaminants from the feed stream and producing an effluent stream that would require little or no further treatment. The greater the volume reduction, the more effective the technology is in reducing ultimate disposal costs. However, there is a balance between the magnitude of the volume reduction, the quality of the effluent stream, and the size and operation of the unit. A higher volume reduction would require additional recycling, reducing the overall flow through the system. In addition, higher levels of contaminant removal will usually result in lower fluxes through the membrane requiring either more membrane area or longer processing time. The balance between throughput and effluent quality is dictated by clean-up standards and treatment costs. This balance will impact such factors as the size and type of the equipment, mode of operation, time required for remediation, treatment requirements for the permeate, and ultimate disposal mechanism for the concentrated contaminants.

171

Membrane processes have many applications in the treatment of contaminated waste streams. The most common applications involve the removal and concentration of organic and inorganic contaminants from liquid waste streams. The waste steams can originate from industrial processes, contaminated groundwater, contaminated surface water bodies, or as by-products of other treatment processes. Membrane and filtration processes have historically been utilized for the treatment and purification of drinking water. For this application, filtration is used to remove a wide variety of constituents, ranging from visible particulates (sand filters, refer to Chapter 5) to ionic species (reverse osmosis, refer to Chapter 4). From these conventional applications, new uses of membrane separations have recently been applied to the treatment of hazardous waste streams. Membranes can be used to separate and concentrate organic contaminants from waste streams. In these applications, the organic contaminants are removed based on their size (molecular weight) or polarity. Size separations rely on membranes with specific pore size distributions. The smaller the pores, the greater will be the removal of small molecular weight compounds. However, as the membrane's pore size decreases, the flux (flow per unit membrane area) also decreases impacting the overall economics and efficiency of the process. The polarity of an organic constituent is a measure of it's ability to ionize in solution. Examples of polar molecules are water, alcohols, and compounds with hydroxyl (e.g. phenols) and carboxyl groups (e.g. organic acids). Aliphatic hydrocarbons and polynuclear aromatic hydrocarbons are examples on nonpolar organic molecules. The chemical characteristics of the membrane can be used to separate non-polar constituents in a waste stream from polar constituents. For example, a membrane whose surface is hydrophilic will allow passage of polar components while retaining the non-polar components. These membranes can be used to separate dissolved and emulsified oils from aqueous waste streams. Inorganic contaminants, such as salts and heavy metals, can be removed and concentrated from waste streams by membrane processes. Suspended inorganics can be easily removed through the use of microfiltration membranes. These membranes have pore sizes ranging from as low as 0.01 up to several microns. Dissolved inorganics can be removed either through the use of hyperfiltration (reverse osmosis) membranes, or by precipitation followed by microfiltration. Conventional reverse osmosis membranes may require extensive prefiltration to avoid fouling, and therefore can only be used on relatively clean feed solutions. Chemical precipitation, followed by microfiltration, allows for the use of microfilters which exhibit higher fluxes and are not as sensitive to fouling. Membrane processes can be helpful in solving many remediation problems at hazardous waste sites. Contaminated Groundwater Containment and/or remediation of contaminated aquifers typically utilizes pump and treat technologies to control contaminant plume migration and ultimately restore the

172

Liquid Filtration

quality of the groundwater. The recovered groundwater usually requires treatment prior to discharge. Treatment alternatives for the recovered groundwater are dependent on the nature and extent of the contamination. Membrane systems can be effectively used to significantly reduce the quantity of groundwater requiring costly treatment. The contaminants of concern can be isolated and concentrated into a reduced volume which can be more easily handled. Another potential benefit of the concentration process is that additional destructive treatment alternatives may become feasible. For example, the concentration of hydrocarbons from a contaminated groundwater can produce a reduced volume waste with a high BTU value allowing for fuel blending as a disposal alternative. This not only reduces the quantity of groundwater that must be treated, but also produces a more easily treatable final waste product. As another example, heavy metals can be concentrated from an aqueous stream by membrane processes and immobilized by solidification/stabilization technologies. Membrane processes can be potentially used to recover organic and inorganic constituents for recycle/reuse. In these applications, the separation scheme must be developed to produce a high quality concentrate. Membrane processes can be applied to the removal of many organic contaminants from waste streams. Organic contaminants that can be removed include petroleum derived hydrocarbons (benzene, toluene, ethylbenzene, xylenes), polynuclear aromatic hydrocarbons, PCBs, dioxins/furans, pesticides, and chlorinated hydrocarbons. Generally, membrane process are more easily applied to removing larger molecular weight, non-polar organic components because larger pored membranes can be utilized and surface chemistry interactions can augment size separations. Removal of hazardous inorganic species from contaminated groundwater requires a detailed knowledge of the water chemistry in order to optimize the separation. In many cases, addition of precipitating chemicals must be added in order to induce particulate formation. Furthermore, groundwater containing high concentrations of innocuous inorganic constituents such as iron and divalent cations (e.g., potassium and calcium) may compete with and interfere with the removal of toxic heavy metals. Conventional reverse osmosis membranes are fragile and must be protected from the corrosive nature of many highly contaminated aquifers.

Integration with Other Technologies Membrane processes are particularly amenable to integration with other remedial technologies enabling applications to additional waste matrices. Ease of integration is facilitated by the modular and scalable properties of membrane systems. These systems can be readily integrated with other remedial process equipment to enhance the effectiveness and economy of these systems. Membrane processes can be used as a final polishing tool for remedial technologies involving discharge of process water. In this capacity, the membrane system is utilized to remove contaminants from a relatively dilute waste stream. The benefit of using this

Advanced Membrane Technology for Wastewater Treatment

173

polishing step is to avoid costly overdesign of the primary remedial technology. For example, a membrane system can be implemented as a final polishing step on a bioreactor. The bioreactor can be designed to cost-effectively treat the bulk of the organic contamination, while the polishing membrane can be designed to treat the aqueous phase prior to discharge. Membrane processes can be used as a pre-treatment step for other remedial technologies. The purpose of the pretreatment would be to concentrate the contaminants to a level that is amenable for specific remedial technologies. For example, organic contaminants in dilute aqueous streams (e.g., groundwater, leachate) can be concentrated to a level that could support an efficient biomass for bioremediation technologies. Membranes can be integrated with remedial technologies as a component in the process. For example, membranes can be used to recycle and recover extraction fluids used to concentrate organic and inorganic contaminants in soil extraction technologies.

Features of the Hyperfiltration System The hyperfiltration system has several unique features which provides advantages over conventional membrane processes in wastewater treatment applications. The technology uses a proprietary formed-in-place membrane technique. The membrane is formed on porous sintered stainless steel tubes by depositing microscopic layers of inorganic and polymeric chemicals. The properties of the formed-in-place membrane can be varied by controlling the type of membrane chemicals used, their thickness, and the number of layers. This important feature allows for customization of the membrane system to a wide variety of waste characteristics and clean-up criteria. The formed-in-place membrane can be quickly and economically reformulated in the field to accommodate changes in waste characteristics or treatment requirements. Conventional membranes rely on rigid polymeric, ceramic, or porous stainless steel membranes. These membranes are available in discrete pore sizes and cannot be customized to the characteristics of the feed. Furthermore, once installed on-site it is difficult and costly to modify their separation properties in response to variable feed characteristics. The formed-in-place membrane is compatible with a wide variety of contaminants often encountered in hazardous wastewater steams. Many conventional reverse osmosis membranes are made from materials such as cellulose acetate and exhibit poor compatibility with reactive substances often encountered in hazardous wastes. These conventional membranes will degrade and become inoperative when challenged with many organic compounds. The compatibility problem becomes more critical as the level of concentration increases. The formed-in-place membrane is stable under most chemical environments and will not degrade even at high contaminant concentrations.

174

Liquid Filtration

A major limitation of many membrane systems is their propensity to irreversibly foul. Fouling is the uncontrolled build up of materials on the surface of the membrane. Fouling leads to a loss of flux and eventually results in cessation of flow. If a membrane fouls, it must be cleaned in order to restore flux. If cleaning is unsuccessful, then the membrane is replaced. The technology discussed in this chapter utilizes a cross-flow filtration mechanism to continuously clean the surface of the membrane, hence minimizing fouling. In this mode, the feed stream is directed parallel to the membrane's surface resulting in a cleaning action which minimizes the buildup of materials on the membrane's surface. Since all membranes eventually foul, a cleaning cycle is necessary to restore flux and operability. Many membrane systems have limited abilities to be regenerated due to restrictions in the choice of cleaning chemicals. The formed-in-place membrane is compatible with a wide range of chemical cleaning methods, enabling in-place regeneration of flux. In situations where the membrane becomes irreversibly fouled, the formed-in-place membrane can be stripped and reformulated on-site. The membrane technology can be used as an integral part of a remediation system to significantly reduce the volume and toxicity of contaminated wastewater. The technology is particularly suited for the treatment of contaminated groundwater as part of a pump and treat system. The technology reduces risks to human health and the environment by transferring the contaminants to a smaller volume facilitating destruction or detoxification by other technologies. The system is simple to operate, reliable and requires a minimum of operator attention or maintenance once the membrane has been formed. The stability of the system makes it particularly suitable for long-term use as is necessary for extended pump and treat remedial programs. The demonstration at the American Creosote Works was designed to evaluate the two most critical process parameters for membrane systems; volume reduction and contaminant reduction. A summary of the demonstration results for these critical processes parameters are presented below. A discussion of the demonstration results and process performance, as they relate to applicability to other wastes and sites also follows. The claim that the system can be operated to recover 80% of the feedwater volume as permeate was achieved in the demonstartion program. Average water recovery (volume reduction) for the first five runs was 83 %. The volume reduction for the extended ran was 96%, and represents the maximum volume reduction capability of the unit for the waste steam tested. The process did not achieve the developer's claim of 90% overall removal of the semivolatiles present in the feedwater (on the average, a 74% reduction was achieved). However, the process does effectively remove polynuclear aromatic hydrocarbons from the feedwater and place them in the concentrate. Overall, removal

Advanced Membrane Technology for Wastewater Treatment

175

of polynuclear aromatic hydrocarbons averaged 92%. Removals of individual PAHs range from 78% to well over 94% for individual two, three, and four ring PAHs, Other high molecular weight pollutants, such as oils and dioxins, are also rejected from the permeate with high efficiency (93% for oils and >99% for dioxins). However, removal of low molecular weight phenols is much less effective, with values between 15 and 21%. Depending on how a system is used, i.e., level of volume reduction and quality of permeate, operating plus capital cost could be as low as $200/1,000 gallons. Capital cost for an averaged size system is approximately $300,000. The demonstration was designed to evaluate the innovative features of process as a volume reduction technology. The demonstration took place at the American Creosote Works in Pensacola, Florida and utilized groundwater contaminated with creosote and pentachlorophenol. Creosote was chosen as a testing material for two reasons. 1.

Creosote is a complex mixture of over 250 individual compounds, dominated by polynuclear aromatic hydrocarbons and phenolics, and exhibits a wide range of chemical and physical properties. The wide molecular weight distribution of the organic contaminants is an excellent challenge material for a membrane process, allowing for analysis of removal efficiencies over a wide range of feed characteristics.

2.

Wood preserving waste contaminated aquifers represent a significant and widespread environmental problem. Results from this demonstration could be directly applicable to other wood preserving waste sites.

A pumping well recovered the creosote and contaminated groundwater from the site. The groundwater, which contained aqueous and dense free product fractions, was allowed to settle and the aqueous phase retained for the study. The aqueous phase was diluted with carbon-treated potable water in order to adjust the concentration of the semivolatiles in the feed to fully test the concentrating capabilities of the filtration unit. The utility of a membrane system is its ability to remove contaminants from a waste water stream and concentrate them into a reduced volume. The contaminant reduction is the percent decrease in specific contaminants from the feed to the permeate (discharge). The higher the percent contaminant reduction, the more effective is the membrane at removing contaminants from the waste steam. It is important to note that the applicability of the technology cannot be made solely on the percent contaminant reduction. Since contamination is reduced as a percentage of the concentration in the feed, the quality of the permeate is dependent on feed concentrations. In order to assess applicability, the predicted quality of the permeate can be estimated by calculating contaminant reductions from the feed. The estimated permeate quality can then be compared to site specific discharge standards.

176

Liquid Filtration

For the demonstration, the total concentrations of semivoiatile contaminants for each run are summarized in Table 1 for the feedwater and permeate. The system was evaluated by comparing the total concentrations of these compounds in the feedwater against the permeate. Over the six day period, an average overall rejection of 74% was achieved. Thus, starting with a feedwater containing on the average 90 mg/L of total designated semivoiatile components, the composited permeate, accounting for 80% of the original feedwater volume, contained on the average 23 mg/L. This did not meet the vendor's claim for 90% removal, largely because of the noted inefficiency with phenolics. This is not totally unexpected since the membrane, as formulated, was not expected to remove species with molecular weights less than 200. Table L Feed and permeate semivolatiles — total concentration and contaminant reduction.

Total Semivoiatile Concentrations (mg/L) Contaminant Reduction (%)

Permeate

Feed Run 1

104

18

83

Run 2

91

24

74

Run 3

92

26

72

Run 4

104

22

79

Run5

85

23

73

Run 6

60

24

60

A summary of the average concentrations for individual semivoiatile compounds in the feed and permeate, along with the associated rejections, for the six day demonstration are presented in Table 2. The results of the demonstration indicated that the pilot unit was capable of removing over 94% of some PAHs but only 15 - 21 % of the phenolics. The permeate generated during the process was discharged directly to the local POTW (publically owned treatment works). These results indicate, as expected, that the membrane is more effective in removing larger molecular weight components (PAHs) than the smaller molecular weight molecules (phenolics). With a complex feed such as creosote, it is difficult to achieve high reductions of all components and at the same time deliver adequate throughput. In this application, the membrane was formulated to maximize reduction of the more toxic polynuclear aromatic hydrocarbons. Passage of the phenolic compounds into the permeate did not pose a significant disposal problem since the local POTW could accept the phenols in their treatment system. At other sites, careful attention should be made to local discharge requirements and available treatment facilities.

Advanced Membrane Technology for Wastewater Treatment

177

Table 2. Individual semivolatile concentration and rejections (average of six daily runs).

Analyte

Feed

Permeate

Rejection

Phenol

4.90

3.88

20.8

2-Methyl phenol

2.31

1.93

16.5

4-Methyl phenol

6.92

5.75

16.9

2,4-Dimethyl phenol

1.82

1.54

15.4

Benzole Acid

(1.42)

2.16

-

Pantachlorophenol

(2.42)

1.88

»

Naphthalene

12.87

2.87

77.7

2-Methyl Naphthalene

4.52

0.46

89.8

Acenaphthylefle

(0.14)

(0.02)

*

Acenaphthene

6.84

0.57

91.7

Dibenzofuran

4.88

0.41

91.6

Fluorene

5.92

0.37

93.8

Phenanthrene

17.08

0.59

96.6

Anthracene

1.98

0.07

96.5

Fluoranlhene

7.01

0.10

98.6

Pyrene

4.70

0.05

98.9

Benzo(a)anthracene

1.24

*0.03

>97.6

Chrysene

1.13

"0.03

>97.4

Benzo(b)fl uoranthene

(0.46)

*0.03

Benzo(k)fluoranthene

(0.43)

*0.03

»

Benzo(a)pyrene

(0.31)

'0.03

»

Values in parentheses represent analytes with estimated values that are above instrument limits but below quantitation limits. Analytes not detected are presented by an *, and the values represent one-half the quantitation limit. ;t Individual rejections not calculated due to estimated values.

This type of membrane process would be most applicable to wastewaters containing large molecular weight organic compounds (PAHs, dioxins/furans, polychlorinated biphenyls, and certain pesticides/herbicides). The system can remove smaller molecular weight compounds (phenols, benzene, toluene, ethylbenzene, xylenes) if larger molecular weight compounds are not abundantly present. Removal of smaller molecular weight compounds can be accomplished by modifying the structure of the

178

Liquid Filtration

formed-in-place membrane. For these applications the pores of the membrane are reduced, resulting in higher retentions of smaller components as well as a reduction in the flux (throughput) of the system. To compensate for the reduced flux, either additional membrane modules can be added or more time will be required to accomplish the remediation. In either case, the overall cost may be higher. The system may be most suitable to treating relatively dilute, but toxic, waste streams in which the percent reduction of contaminants will allow discharge of the permeate without further treatment. This feature makes the unit highly suitable for polishing effluents as part of a multi-technology treatment train. In this system, the primary treatment technology can be utilized to remove the bulk of the contamination, with the filtration unit being used as a final polishing step. If the concentration of contaminants in the permeate does not meet clean-up requirements, then the permeate can be recycled back through the membrane to achieve the targeted effluent quality. Recycling of the permeate has the disadvantage of requiring additional membrane modules, or additional time, both of which increase treatment costs. A number of mechanisms could explain the contaminant reduction results, including rejection by the membrane on the basis of molecular weight or molecular size, rejection and coalescence of dispersed oil in which specific components are soluble, or even rejection simply by adsorption of the PAHs on inert suspended solids. Examination of the results for the conventional parameters tested in the feed and permeate (Table 3) provides some insight into the separation mechanism. High concentrations of oil and grease found in the feedwater suggests that considerable oil remained in a dispersed or colloidal form. This oil would be removed by a membrane with ultrafiltration or hyperfiltration characteristics. Since the PAHs are more soluble in oil than in water, concurrent removal of the PAHs entrained within the oil may have occurred. The phenols with relatively high solubility in water are, also as expected, removed more poorly. This also is reflected in the poor rejections calculated for TOC and COD. Other contaminants, not quantified by the semivolatile analysis, also may contribute to the high TOC and COD in the permeate. Tables 3. Conventional parameters (values are averages of six runs).

Analyte

Feed

Permeate

Rejection %

IDS

237

190

20

TSS

34

<4

>88

OIL/GREASE

191

14

94

TOC

121

92

24

COD

379

35

27

Advanced Membrane Technology for Wastewater Treatment

179

Vo/ume Reduction The utility of a membrane separation system for treating hazardous waste streams is also dependent on the magnitude of volume reduction. The volume reduction is a measure of the percent of the feed water that can be generated as cleaner permeate. The higher the volume reduction, the greater the potential utility of the membrane system. Volume reduction cannot be solely used as an indicator of membrane performance. The quality of the permeate must also be considered when evaluating the applicability of the technology. A high volume reduction with low permeate quality is not acceptable since the permeate will not be dischargeable and will require further treatment. When designing a membrane separation system, volume reduction and permeate quality must be balanced in order to develop a cost-effective treatment meeting site-specific clean-up criteria. For the demonstration at the American Creosote Works an 80% volume reduction was achieved each day. This level of volume reduction was set as a target prior to the demonstration and was easily attained. The level of volume reduction was achieved by continuously recirculating the concentrate through the system. On the last day of operation the process was allowed to run until the unit could no longer function, representing the maximum volume reduction for that feed. The maximum volume reduction was 96%. The relationship between volume reduction and permeate quality is exemplified by results from the demonstration. During the demonstration, were grabbed samples of the permeate steam, were collected at the beginning, middle, and end of each run. The purpose of these samples is to document changes in permeate quality during the course of the batch filtration. The analysis of the data reveals an increase in total semivolatile content of the permeate from the beginning to the end of each run. Six day average permeate concentrations of total semivolatiles were 19.24 mg/L at the beginning of the run, 24.17 rng/L in the middle, and 29.95 mg/L at the end of the run. In addition, on day six, when the unit was allowed to ran to a maximal volume reduction of 96%, the final permeate semivolatile concentration was 47.25 mg/L. These changes in permeate quality during the filtration are due to increasing semivolatile contents of the recirculating concentrate. As the batch filtration proceeds, the surface of the membrane is challenged with progressively higher concentration of contaminants. Since the membrane can only reject a certain proportion of the feed stream, the concentration of contamination in the permeate will increase. When applying a membrane solution to a wastewater problem it is crucial to evaluate the balance between permeate quality and volume reduction. Maximizing volume reduction is important since it impacts economics by minimizing the volume of wastewater requiring treatment. However, the quality of the discharged water is critical and must be maintained during the filtration process. Treatability testing is necessary to determine the optimal balance between permeate quality and volume reduction.

180

Liquid Filtration

Fouling Control Fouling is the loss of flux due to the buildup of components on the surface of the membrane. All membranes exhibit some degree of fouling and eventually require cleaning to restore flux. Many membranes foul readily and are not amenable to cleaning for flux restoration. If flux cannot be restored, then the membrane must be replaced resulting in considerable expense and downtime. A major attribute of the technology is its ability to minimize fouling. The process effectively controlled excessive fouling, in spite of the problematical nature of the wood preserving waste feed, through a combination of cross-flow operation and membrane cleaning. Flux and pressure data collected during the demonstration indicated gradual and slight fouling of the membrane. This slight fouling was reversed after each two-run cycle by a membrane cleaning procedure. Analysis of the washwaters from the cleaning process indicated that approximately 8% of the mass of semivolatiles remained in the system and were removed during the washing process. The membrane cleaning process effectively regenerated the membrane to its original clean permeate flux conditions. This enabled the membrane to be reused, without the necessity to reformulate. The ability to repeatedly regenerate the flux after the cleaning procedure is a good indication that the formed-in-place membrane is stable and can be used over an extended length of time. In the unlikely event of an irreversible fouling, the membrane can be cost-effectively and easily reformed on-site with a minimum of downtime. Operational Reliability and Implementability Operational reliability and implementability are important in deciding the applicability of the technology to other waste streams and sites. The system proved to be quite stable and required a minimum of attention over the demonstration period. System performance was relatively constant during the six day test. With feed concentrations of total semivolatiles ranging from 60.4 - 103.8 mg/L, the percent rejection averaged 74%, with a narrow standard deviation of 7.5. Other than adjustment of the pressure to maintain flux and the cleaning of the unit, which consumed about 2 hours every other day, there was little need for an operator. In a commercial installation some means of on-line monitoring (e.g., changes in pressure, contaminant concentration, etc.) could alert the operator to out-of-specification operation or out-of-compliance permeate. It is estimated that the unit could be run by two operators (health and safety requirements). Additional units could easily be operated by the existing personnel. Other than the cleaning operation every other day, there was no downtime during the demonstration. With the exception of the pump there are no moving parts to break down or require service. The process equipment and supplies for the system are commercially available. This includes the filtration modules, membrane forming chemicals, pumps, tanks, process controls, gauges, and flowmeters.

Advanced Membrane Technology for Wastewater Treatment

181

The membrane formation procedure requires a high level of expertise and may require trial and error methods to achieve the desired separation characteristics. However, these are not obstacles to implementation since the process is inexpensive and rapid. The process is easily scalable and can be modified by adding or deleting modules in response to processing requirements. The addition of modules does not affect the mode of operation, except for additional support equipment (pumps, tanks, plumbing). Based on the observations from the demonstration, it is feasible that the membrane system can be effectively and reliably operated over an extended time period as would be necessary for pump and treat remediations. Applicable Wastes Although the hyperfiltration unit was limited to a single wastewater study for the groundwater available at the American Creosote Works site, the results of the study along with other results provided by the vendor suggest that the technology would have applicability to other contaminated groundwaters and process waters. The developer believes the system can teat wastes with 100 - 500 mg/L of COD where the molecular weight of the contaminants to be concentrated are over about 200. However, the characteristics of the membrane can be modified to treat smaller molecular weight compounds. More dilute feedwater will necessitate additional cycles to achieve the desired concentrations in permeate and concentrate streams. However, the more dilute feedwaters would also allow for higher fluxes. Other than having an impact on cost and throughput, this should not adversely affect operation. Waste streams exceeding the target concentration range (100 - 500 mg/L COD) would require reduced cycling to achieve the required level of concentration. The effect of elevated feedwater concentrations on the rejection of individual components may also need to be determined by laboratory testing. Data from this study indicates a reduction in permeate quality as the concentration of the feed increases. Groundwater rich in PAHs would probably be suitable while feedwater where smaller molecular weight compounds are a major pollutant would probably not be appropriate for this technology. However, membranes could be formulated to separate small molecular weight species (BTEX) such as those found in hydrocarbon contaminated waste waters. Cross-flow filtration using the formed-in-place membrane may also be applicable to other waste streams containing different high molecular weight organic contaminants. This might include polychlorinated biphenyls (PCBs) as might be encountered from a spill from a PCB transformer leak, particularly since the same preferential solubility in oil noted earlier may prevail. On the same basis, the system may be useful for separating other emulsified or dispersed organics which do not lend themselves to simple physical phase separation. The system is also well suited to significantly reduce the concentration of dioxins and furans in wastewater. Reduction of dioxins/furans encountered in this demonstration was greater than 99.9%.

182

Liquid Filtration

The developer believes the membrane can be customized to achieve different rejection characteristics that could be applied to a wide range of contaminants. Site Characteristics of the Pilot Demonstration The pilot-plant unit used in the demonstration program required a level base large enough to accommodate the unit, and storage tanks for the feed, concentrate, and wash water. A covered concrete pad is recommended to protect the equipment from the elements as well as contain the accidental release of contaminated materials. Clean water and power are the only utilities needed. If necessary, the relatively small amount of clean water needed for washing of the membrane can be trucked in and power for the compressor can be provided by an on-site generator. While it was not studied, it may be practical to use permeate for washing. Where the unit is being used to treat groundwater, power also would have to be provided for the well pumps. Acquisition of groundwater for the unit may require the development of an extraction well network, consisting of the appropriate pumps, regulators, and plumbing. Permit requirements and the mode in which the filtration unit is operated may make it necessary to have additional space for storage tanks for equalization of the permeate until analyses can confirm acceptability for the POTW or surface water body discharge. Materials Handling Requirements Materials handling requirements for the unit involve 1) the acquisition of feed material for the unit, 2) pretreatment, and 3) residuals (permeate and concentrate) management. If the filtration unit is part of a system used to treat groundwater, the first need is a well drilling rig to provide the well or wells from which the feedwater is to be obtained. Once the wells are drilled and developed, each must be equipped with a pump to draw up the necessary feed water. Local well drilling requirements would have to be taken into consideration. At some sites pretreatment may be necessary to remove free oil and even suspended solids. Since the developer has indicated that the filtration unit is most effective when operating with a feed water having a COD range of 100 - 500 mg/L and is most effective in rejecting materials with molecular weights greater than about 200, pretesting will be necessary to assure that these requirements are consistently met. If the vendor's system is provided with relatively clean ground or process water, no pretreatment may be necessary. The applicability of this membrane technology at a site is dependent on the quality of the permeate, site-specific discharge criteria, and the availability and accessibility of local public or industrial waste water treatment facilities. It is important to conduct a treatability study to assess the quality of the permeate and to determine options for disposal. If the permeate quality is not amenable for discharge to surface waters or local treatment facilities, then the technology is not applicable to the site.

Advanced Membrane Technology for Wastewater Treatment

183

Prior to the initiation of the demonstration at the site, a limited filterability test was conducted on contaminated groundwater to determine if the permeate would be accepted by the local POTW. The permeate was subjected to biological testing (Ceriodaphnia) and chemical analysis to determine its suitability for discharge. The permeate passed the local POTW's criteria and was directly discharged to a local sewer hook-up. Several additional options are available for permeate disposal and are dependent on waste and site conditions, as well as local discharge regulations and treatment options. The permeate quality may meet local standards for direct discharge to local surface water bodies. This would occur only if the level of contaminants in the permeate was extremely low and meeting the strict requirements for surface discharge. The permeate could be treated on-site with additional treatment equipment to reduce contaminant levels for either surface water body discharge or sewer discharge. Treatment, such as with activated carbon, may be necessary to reduce contamination to acceptable limits. The use of additional treatment equipment will increase remediation costs and may necessitate additional disposal requirements. The permeate may be recycled through the filtration unit, or processed through a smaller unit, to further reduce contaminants for surface water body or sewer discharge. The secondary filtration unit may have different membrane characteristics as the primary unit to remove species that were not retained or require greater reductions. This option would also add to the overall cost of the remediation since additional equipment and time would be required. If it is not feasible to reduce contaminant concentrations to levels adequate for on-site discharge, and if no local sewer hook-up is accessible, then it may be necessary to transport the permeate by tanker truck to an acceptable treatment facility. This option would only be economically feasible if the membrane process drastically reduced the volume of a waste stream that is very costly or difficult to teat (e.g., dioxin contaminated wastewater).

Concentrate Disposal Options The membrane process minimizes the quantity of waste requiring extensive treatment by concentrating the contaminants into a reduced volume while producing a cleaner permeate for discharge. Since the contaminants are not destroyed by the process it is necessary to consider disposal options for the reduced volume concentrate stream. If the treatment options for the concentrate steam do not reduce overall treatment costs or provide a reduction in risk to human health and the environment, then the membrane system is not a feasible remedial technology. Optimally, a disposal option that can permanently destroy or immobilize the contaminants in the concentrate stream on-site is preferable to off-site transportation and disposal. A portion of the concentrate from the demonstration was utilized to develop a bioremediation technology that could be coupled to the filtration unit to produce a treatment system for on-site destruction of a major portion of the waste. The system

184

Liquid Filtration

uses a two-stage bioreactor containing several naturally occurring strains of soil bacteria capable of mediating PAH contamination. The membrane system is used to reduce the quantity of wastewater input into the bioreactors and to optimize contaminant concentrations to support the biomass. The use of the concentrate as a feed to the bioreactors extends the utility of this volume limiting technology by reducing the volume of wastewater that must be processed, therefore reducing equipment costs and site space requirements. The concentration process enhances the calorific value of most organic wastes. This enables the utilization of thermal technologies as a means of destroying the organic contaminants. The feasibility of using and choosing a thermal technology is based on the nature of the organic contaminants. Concentrates from petroleum based contamination could be readily used for fuel blending, while concentrates from other sources (such as wood preserving wastes) would require careful testing to determine selection of the appropriate thermal technology. Thermal destruction could be accomplished on-site (mobile units) or transported off-site. Concentrates containing highly toxic constituents, such as PCBs and dioxins/furans, which are not amenable to biodegradation or thermal treatments, can be chemically neutralized by processes such as dechlorination. The neutralized waste could then be disposed of in a conventional manner.

Process Economics The primary purpose of this economic analysis is to estimate costs (excluding profit) for commercial-scale remediation using the filtration unit. With realistic costs and a knowledge of the basis for their determination, it should be possible to estimate the economics for operating similar-sized systems at other sites utilizing scale-up cost formulas. Among such scale-up cost formulas for chemical process plant equipment is the "six-tenths rule". The six-tenths rule is an exponential method for estimating capital costs from existing equipment costs. If the cost of a piece of equipment of size or capacity q, is Cj, then the cost of a similar piece of equipment of size or capacity q2 can be calculated from:

The value for n in this discussion is taken as 0.6. It is assumed that the performance of commercial-scale equipment will be the same as that demonstrated. Cost figures provided here are "order-of-magnitude" estimates, and are representative of charges typically assessed to the client by the vendor, exclusive of profit. The total annual cost to operate a 12-module filtration unit ranges between $514,180 and $1,209,700, depending on whether effluent treatment and costs are considered, the flow rate through the unit, the cleanup requirements, and the cost of effluent treatment

185

and disposal (if required). Effluent treatment and disposal costs, if considered, could account for up to 60% of the total cost. Labor can account for up to 40% of total annual costs. Processing costs are more dependent on labor costs than equipment costs. The cost per 1,000 gallons can be broken down by flow rate as follows (for with and without effluent treatment and disposal costs): With Effluent Treatment Costs 24gpm $228-522/1, OOOgal

12gpm $456-1,44o/l,000gal

7.2 gpm $760-1,7397 l.OOOgal

Without Effluent Treatment Costs 24gpfn $222/1,000 gal

12gdm $444/1,000 gal

7.2gpm $739/1,000 gal

As expected, the cost category having the largest impact and variability on total cost was effluent treatment and disposal. The demonstration used a four-module filtration unit. For a full-scale remediation, twelve of the same modules instead of four would be used with a portable generator for power, a mix tank, and a single pump and motor. No assumptions as to the site size or volume of waste to be treated were made. It was assumed that the same unit would be operated at different flow rates for a one year period to obtain the desired results. For example, at the maximum assumed flow rate of 24 gpm, 2.6 million gallons of waste would be treated in 230 days of operation. The annual cost was then divided by the volume of waste that would be treated at a particular flow rate to obtain $/1,000 gal. No assumptions regarding percent rejection or outlet contaminant concentrations were made. Based on results from the demonstration, a volume reduction of 80% between waste and concentrate was assumed. Costs per 1000 gal, treated were calculated for 24, 12 and 7.2 gpm flow rates; the last corresponding to what was demonstrated in the pilot program. Flow rates, the amount of recycle, and the initial concentration of contaminants may impact costs significantly. One equipment operator/supervisor and one technician will operate the unit and be onsite eight hours per day, although the system will be operated only seven hours per day, five days per week. The extra hour each day will be used for cleaning and maintaining the unit. A site supervisor will visit the site for approximately two to three days each month for oversight purposes. The two-person crew could operate up to three 12-module systems. If more modules are required, additional manpower would be needed. The filtration unit was assumed to be utilized for 230 days out of a possible 365 days a year. Scheduled maintenance was assumed to be performed during normal operating hours.

186

Liquid Filtration

For the purpose of this analysis, capital equipment costs were amortized over a 7-year period with no salvage value. Interest rates, time-value of money, etc, were not taken into account. The following is a list of additional assumptions used in this study. • « • • •

Access to the site is available. Utilities, such as electricity, water, telephone, is easily accessible. The permeate stream will not require further treatment. A hook-up to the appropriate outlet (sanitary sewer, storm sewer, surface water body) is available on or near the site. There are no waste water pre-treatment requirements.

Basis for Economic Analysis In order to compare the cost-effectiveness of technologies in the pilot demonstration program, costs were broken down into 12 categories shown in Table 4 using the assumptions already described. The assumptions used for each cost factor are described in more detail below. Table 4. Estimated costs for the filtration unit studied.

COST COMPONENT

TOTAL

1. Site Preparation Costs * 2. Permitting & Regulatory Costs * 3. Equipment Costs (amortized over 7 years) 4. Startup * 5. Labor 6. Consumables and Supplies Health & Safety Gear Maintenance Supplies 7. Utilities Telephone Electricity Sewer/Water 8. Effluent Treatment & Disposal (Concentrate) 9. Residuals/Waste Shipping, Handling and Transport Costs 10. Analytical Costs 11. Facility Modification, Repair & Replacement 12. Demobilization Costs * TOTAL (without concentrate Disposal) TOTAL (with concentrate disposal)

$85,000 $15,000 $42,850 $5,000 $199,080

* one-time costs

$3,000 $500 $6,600 $2,000 $2,000 $13,915-$695,520 $46,000 $60,000 $37,150 $10,000 $514,180 $528,09541,209,700

Advanced Membrane Technology for Wastewater Treatment

187

Site Preparation Costs - The amount of preliminary preparation will depend on the site and is assumed to be performed by the responsible party (or site owner). Site preparation responsibilities include site design and layout, surveys and site logistics, legal searches of access rights and roads, preparations for support facilities, decontamination facilities, utility connections, and auxiliary buildings. These preparation activities are assumed to be completed in 500 staff hours. At a labor rate of $50/hr. this would equal $25,000. Other significant costs associated with site preparation include construction of a pad and cover, well drilling as well as buying and installing a groundwater pump, holding tanks, and associated plumbing. The cost to construct a concrete pad and cover to support the unit and protect the unit from the elements is estimated to be $20,000. Based on the demonstration, the cost to drill a well was assumed to be $5,000. To achieve the appropriate maximum groundwater extraction rate of 24 gpm, three recovery wells are required, resulting in a cost of approximately $15,000. A 5200 gallon, holding tank cost $5,000. Using the "six-tenths rule" to scale-up, the cost of a 10,000 gallon tank for a full-scale remediation was assumed to cost $7,400. Three tanks will be required, resulting in a cost of $22,200. A V2 horse-power pump cost $1,035 for the demonstration. A pump for each well would cost a total of $3,105. These additional costs amount to about $40,000. Therefore, the total site preparation costs for a full-scale remediation would be about $85,000 as shown in Table 4. Permitting and Regulatory Costs - Permitting and regulatory costs include actual permit costs, system health/safety monitoring, and analytical protocols. Permitting and regulatory costs can vary greatly because they are very site- and waste-specific. For this cost estimate, permitting and regulatory costs are assumed to be 5% of the equipment costs. This assumption is based on operation at a Superfund site. At RCRA (Resource Conservation and Recovery Act) corrective action sites permitting and regulatory costs may be higher and an additional 5 % of the equipment cost should be added. Equipment Costs - Capital equipment costs are for a twelve-module filtration unit equipped with a portable generator for power, a mix tank, and a single pump and motor all mounted on a trailer with associated instrumentation, alarms and controls. Variation in equipment costs from site-to-site should not be significant. However, based on the cleanup requirements and the material being treated, the flow rate through the system may vary dramatically resulting in a wide range of costs per unit treated. Based on a capital cost estimate of $300,000 for 12 modules, each module would cost $25,000. Equipment costs were amortized over 7 years, with no salvage value at the end of that time period, giving an annual cost of $42,850 as shown in Table 4, without any interest factor.

188

Liquid Filtration

Startup ~ Filtration units are mobile and designed to move from site-to-site. Transportation costs are only charged to the client for one direction of travel and are usually included with mobilization rather than demobilization. Transportation costs are variable and dependent on site location as well as on applicable oversize/overweight load permits, which vary from state-to-state. The total cost will depend on how many and which state lines are crossed. The system is designed to be ready to operate as mounted on the trailer so mobilization costs should be primarily the cost of travel and the time to connect the plumbing and adjust the membranes, if necessary. The startup labor cost is included in the total labor cost component and includes relocation and/or hiring expenses. The cost of health monitoring programs has been broken down into two components — OSHA (Occupational Safety and Health Act) training, estimated at $l,000/person, and medical surveillance, estimated at $500/person for a total cost of $l,500/person. For two people, on-site, this would be $3,000. Depending on the site, however, local authorities may impose specific guidelines for monitoring programs. The stringency and frequency of monitoring required may have significant impact on the project cost, A conservative estimate of $5,000 was assumed as shown in Table 4. Labor - Labor costs may be broken down into two major categories: salaries and living expenses. It is estimated that the equipment will require two on-site personnel for operation and maintenance. Due to the extended time requirements for major groundwater restoration projects, plans to hire local operators or relocate personnel to the site may be necessary. These actions would minimize costs associated with living expenses. A cost of $5,000 is estimated for hiring and/or relocation. Site supervision will require periodic visits from the mam or regional office to oversee the progress of the remediation. Per diem is assumed to be $125 per day per person, but may vary widely by location. This rate is a liberal estimate assuming that cleanups may occur in some of the more expensive areas of the country. Travel to and from the site (periodic supervision) is estimated to be $800/visit. One rental car is assumed to be obtained at a rate of $55/day. Supervisory and administrative staff will consist of an off-site program manager at $75/hour. The filtration system will operate 7 hours per day, 5 days per week. One equipment operator/ supervisor at $50/hr. and one technician at $35/hr. will be on-site 8 hr./day. The labor requirements and rates are detailed in Table 5. Consumables and Supplies - There are two items to consider under this cost category. The first is health and safety gear which include hard hats, safety glasses, respirators and cartridges, protective clothing, gloves, safety boots, and a photoionization detector monitor, all estimated at $l,500/person. For two people this totals $3,000. The second item is maintenance supplies (spare parts, oils, greases and other lubricants, etc.) estimated at 1% of the annual amortized capital costs or

Advanced Membrane Technology for Wastewater Treatment

189

approximately $500, The cost of membrane forming chemicals are inconsequential (less than $200), Utilities - Telephone charges are estimated at $500/month plus an additional 10% for fax service or $550/month. This will total $6,600 annually. Electric usage is estimated to cost about $ 10/day or $2,000 annually. Combined sewer and water usage costs is assumed to be about $0.05/1000 L ($0.20 per 1000 gal). Based on the demonstration results, approximately 150 gallons of water were used to flush a 4-module system. Hence a 12-module system was assumed to use three times as much water or about 500 gallons/day. This would cost about $ 10/day or $2,000 a year as well. This does not consider discharge of permeate, which may incur additional cost. Table 5. Labor requirements and rates to operate the filtration unit, Living and Travelling Expenses: 3 days/month for 12 months: Per Diem $125/day/person x 1 person x 3 days/week x 12 weeks = $4,500 Rental Car $55/day x 7 days/week x 52 weeks = $1,980 Travel $800/trip x 12 months = $9,600 Salaries: Program Manager - $75/hr(*} x 8 hr/day x 36days Operator/Supervisor - SSO/hr^ x 8 hr/day x 230 days Technician - $35/hr(*} x 8 hr/day x 230 days Relocation/Hiring Total Labor

=

$21,600

=

$92,000

= =

$64,000 $5,000

=

$199,080

Includes salary, benefits, and administration/overhead costs but excludes profit.

Effluent Treatment and Disposal - Two process streams are produced by the filtration unit. The permeate is considered to be essentially free of contaminants and is assumed to meet standards appropriate for discharge to a POTW. The concentrate is the reduced-volume portion of the waste stream containing the enriched contaminants. This stream would require further treatment such as biological degradation, incineration, fuel-blending, or some other process appropriate to the type and concentration of contaminants. The filtration system is a volume reduction technology, and as such minimizes the volume of wastewater that would require treatment. The technology was demonstrated as a method to reduce the volume of wood preserving waste contaminated groundwater. Therefore, treatment of the concentrate is not part of the demonstrated

190

Liquid Filtration

technology and it is not necessarily appropriate to consider costing for this parameter. However, the cost for treating these effluents can be a substantial factor in designing a remediation program. Based on these issues, overall costing will be calculated both with and without effluent treatment and disposal costs. Two concentrate disposal options are considered in this exercise. The first is bioremediation which provides on-site destruction of PAH contaminants. A projected cost estimate of 10-40 cents per gallon of groundwater contaminated with 100-2000 ppm of PAHs is appropriate for a full-scale bioremediation system. The second disposal option for the concentrate is more conventional. Based on the characteristics of the concentrate, fuel blending is considered a viable disposal option, resulting in a cost of $1.50/gallon. It is important to note that effluent treatment costs can be very high and are dependent on specific waste and site conditions. Cost estimates for this exercise are based on waste and site characteristics of the demonstration. Based on the demonstration, the concentrate accounts for 20% by volume of the contaminated groundwater influent stream to the filtration unit. The volume of concentrate generated each day and the range of costs for the three different flow rates are shown below for the bioremediation system and conventional disposal: 24gpm

12gprn

7.2gpm

Gallons of Waste Treated/Day

10,080

5,040

3,024

Gallons of Concentrate Generated/Day (assumes 20%)

2,016

1,008

605

Annual Treatment Costs Bioremediation

$46,370$185,470

$23,180$92,736

$13,915$55,660

Annual Treatment Costs Conventional

$695,520

$347,760

$208,725

Effluent treatment and disposal costs can range from $14,000-$700,000 depending on the flow rate through the filtration unit, the mode of treatment, and the cost of treatment in the bioremediation system. Residuals/Waste Shipping, Handling and Transport Costs - Waste disposal costs including storage, transportation and treatment costs are assumed to be the obligation of the responsible party (or site owner). It is assumed that residual or solid wastes generated from this process would consist only of contaminated health and safety gear, used materials, etc. Landfilling is the anticipated disposal method for this material and costs were once again derived from the demonstration test. Approximately four drums of solid waste were generated each day of operation. However, due to intensive sampling activities during the demonstration, excessive solid waste was generated.

Advanced Membrane Technology for Wastewater Treatment

191

Under actual remediation conditions, substantially less waste would be generated. It is estimated that approximately one drum of solid waste would be generated each day of operation. At a disposal cost of $200/drum, the total yearly cost of disposal is estimated to be $46,000. Analytical Costs - Standard operating procedures do not require planned sampling and analytical activities. Periodic spot checks may be executed to verify that equipment is performing properly and that cleanup criteria are being met, but costs incurred from these actions are not assessed to the client. The client may elect, or may be required by local authorities, to initiate a sampling and analytical program at their own expense. For this cost analysis, one sample per day for 100 days at $600/sample was assumed to be required by local authorities for monitoring and permitting purposes. This would total approximately $60,000. Facility Modification: Repair and Replacement Costs - Since site preparation costs were assumed to be borne by the responsible party (or site owner), any modification, repair, or replacement to the site was also assumed to be done by the responsible party (or site owner). The annual cost of repairs and maintenance was estimated to be $37,150. Demobilization Costs - Site demobilization will include shutdown of the operation, final decontamination and removal of equipment, site cleanup and restoration, permanent storage costs, and site security. Site demobilization costs will vary depending on whether the treatment operation occurs at a Superfund site or at a RCRA-corrective action site. Demobilization at the latter type of site will require detailed closure and post-closure plans and permits. Demobilization at a Superfund site does not require as extensive post-closure care; for example, 30-year monitoring is not required. This analysis assumed site demobilization costs are limited to the removal of all equipment and facilities from the site. It is estimated that demobilization would take about two weeks and consist primarily of labor charges. Labor costs include salary and living expenses. Demobilization is estimated to be $10,000. Grading or recompaction requirements of the soil will vary depending on the future use of the site and are assumed to be the obligation of the responsible party (or site owner). Overall Economic Analysis Table 4 shows the total annual cleanup cost to range between $514,180 and $1,209,700. This is based on the assumption that the remediation will take one year. Most applications for this technology will require several years, as in pump-and-treat remedial projects. Since many of the cost factors are one-time, the overall $/gallon cost will go down as the length of the project increases. This is illustrated in the hypothetical site example in the subsequent sub-section. The total cost is also highly dependent on whether concentrate treatment and disposal is considered as part of the

192

Liquid Filtration

nitration's technology and responsibility. Concentrate disposal costs can vary widely, and are dependent on technical and regulatory issues related to the waste characteristics. Therefore, if concentrate disposal costs are considered, this category could account for up to 60% of die total costs. Without concentrate disposal, labor is the dominant cost, accounting for approximately 40% of the cost. Equipment costs represent a relatively minor component. Furthermore, the system can be easily scaledup by adding 12-module units. Up to three 12-module units can be operated without adding additional labor. This would significantly reduce overall treatment costs. The smallest cost categories appear to be those associated with startup, and consumables and supplies. All other cost categories appear to contribute to the total cost about equally (i.e., 5-10%). The costs per 1,000 gal is dependent on the flow rate, the duration of the project, whether concentrate disposal is being considered, and the cost of effluent treatment and disposal. These ranges are shown below and are based on a one year project. Cost per thousand gallons of feed 24gpm

12gpm

7.2gpm

With Concentrate Disposal

$2284522

$456$1,044

$760$1,739

Without Concentrate Disposal

$222

$444

$739

In all of the above analyses, it should be remembered that costs for 10 out of the 12 cost components were considered. One of the cost components not included here was permitting and regulatory expenses. Additional effluent treatment and disposal for the permeate was assumed to be not required. If these factors are taken into account, costs could significantly increase. Remediation of a Hypothetical Site The economic analysis presented in the preceding section is based on costs for a one year remedial project. The dominant application of the membrane system is expected to be for groundwater restoration projects. Since groundwater restoration projects can last for ten to twenty years, a hypothetical economic analysis is presented to illustrate the application of the twelve factors in developing a multi-year project. The hypothetical site contains groundwater contaminated with wood preserving wastes in composition and concentrations similar to the feedwater tested in the demonstration. The remedial plan calls for containment of the groundwater plume, with eventual aquifer restoration. A hypothetical model predicts that approximately two million gallons of groundwater is contaminated, and that twenty million gallons must be treated to restore the aquifer. The groundwater will be extracted from the shallow aquifer (ten to thirty feet below surface) through three wells.

Advanced Membrane Technology for Wastewater Treatment

193

The remedial design will utilize three 12-module filtration units operating at 7.2 gpm/unit for a combined throughput of 21.6 gprn. Treatability testing identified that an 80% volume reduction could be achieved, with the permeate meeting discharge standards to the local POTW. The concentrate from the process will be treated on-site by a bioremediation technology at a cost of 40 cents/gallon. Based on these conditions, and the economic assumptions previously stated, the remedial time-frame will be ten years. Approximately 2,100,000 gallons of groundwater will be treated each year by the filtration unit, and 420,000 gallons of concentrate by the bioremediation system. The total volume of groundwater to be treated for the ten year project is 21 million gallons. Table 6 is a summary of the costs for each of the twelve criteria as they relate to the conditions set forth hi the hypothetical analysis. Based on the requirements of the hypothetical site, the overall treatment costs for the remediation is $300/1000 gallons. It is important to note the overall $/gallon treatment cost is highly dependent on the length of the remediation project. The longer the project, the lower the $/gallon treatment cost. Table 6. Hypothetical site cost analysis for a ten year project.

1. Site Preparation 2. Permitting and Regulatory 3. Equipment 4. Startup 5. Labor 6. Consumables and Supplies 7. Utilities 8. Effluent Treatment and Disposal 9. Residuals 10. Analytical I t . Facility Modification Repair and Replacement 12. Demobilization Total

$85,000 $15,000 $900,000 $5,000 $1,990,800 $35,000 $106,000 $1,669,248 $460,000 $600,000 $371,500 $10.000 $6,247,548

Detailed Process and Technology Description Membranes are being used increasingly for the removal of dissolved and colloidal contaminants in wastewater streams. Reverse osmosis (hyperfiltration) is well known for its ability to concentrate ionic species while ultrafiltration has found broad utility for the removal of dispersed colloidal oil, non-settlable suspended solids, and larger organic chemical molecules. One of the major problems these processes have faced is the fouling or blinding of the membranes after limited use. Various approaches have been developed in an effort to minimize this deterrent. Cross-flow filtration, where the contaminants are constantly flushed or washed from the membrane surface by the feedwater stream, is one of these approaches. The unit goes farther. Rather than a thin

194

Liquid Filtration

polymeric membrane requiring careful handling to avoid perforation, a membrane is created on a stainless steel microfilter support by the introduction of a mixture of carefully selected (and proprietary) chemicals. This approach imparts special properties and allows a degree of customization that may be difficult to achieve with conventional membranes. The resulting "formed-in-place" membrane can be designed to provide properties similar to conventional hyperfiltration or ultrafiltration, as needed for a specific application. The filtration unit consists of porous sintered stainless steel tubes arranged in a modular, shell-and-tube configuration. Multi-layered inorganic and polymeric "formed-in-place" membranes are coated at microscopic thickness on the inside diameter of the stainless steel tubing by the recirculation of an aqueous slurry of membrane formation chemicals. This "formed-in-place" membrane functionally acts as a hyperfilter, rejecting species with molecular weights as low as 200. In addition, surface chemistry interactions between the membrane matrix and the components in the feed play a role in the separation process. A relatively clean stream, called the "permeate", passes through the membrane while a smaller portion of the feedwater, retaining those species that do not pass through the membrane, is retained in a stream called the "concentrate" or "reject". For efficient operation of a membrane filtration system, it is necessary to prevent the buildup of dissolved and particulate species on the surface of the membrane and in the membrane pores. The buildup of contaminants, termed "fouling", can lead to a steady decline in the permeate flux (flow per unit area of membrane surface), eventually causing cessation of flow. To prevent, or retard excessive fouling, the filtration unit is operated in a cross-flow mode (as illustrated Figure 1). In cross-flow mode the feed stream is directed parallel to the surface of the membrane. Material larger than the surface porosity is temporarily retarded on the membrane surface and then swept clean by the cross-flow action — if the fluid velocity is sufficient. Meanwhile, the portion of the stream containing the smaller species passes through the membrane. The goal of cross-flow filtration is not to trap components within the pore structure of the membrane. The test unit used operates with four modules aligned in parallel. The filtration unit is approximately 13 feet long, 5 feet wide, and 7 feet high and contains an estimated total membrane area of up to 300 square feet. Automatic level controls provide for unattended operation with continuous feed to the tank. Concentrate recycle flow also can be controlled automatically. Figure 2 provides a schematic of the filtration unit. At the American Creosote Works site, groundwater was pumped from a well to an above ground storage tank where a quiescent period of several hours allowed oil and suspended solids to coalesce and separate. The feedwater stream to the filtration unit was drawn from the mid-section of the storage tank to minimize introduction of these materials. The pump that drew material from the tank also provided the compression for the system to operate, approximately 750 psig. The permeate leaving the filtration unit was sampled as required and then discharged in accordance with permit requirements. The concentrate was collected in a smaller

Advanced Membrane Technology for Wastewater Treatment

195

tank until the desired volume was accumulated. It is then recycled as feed until the desired final concentration and volume are achieved. This mode of operation was selected for the demonstration in anticipation of a companion study of biodegradation process for the concentrate. Alternate operating modes can be used to achieve other goals, depending on disposal plans and options for the permeate and the concentrate.

Permeate o Concentrate

Influent

• *e?e* <%«» e

Figure 1. Illustrates cross-flow filtration mode of operation.

Groundwater

Concentrate

, Membrane 9-^ Membrane %• Filtration -O Filtration Unit Feed Tank ii Permeate 1I*

Wfc^

Qtsvran*

Tank

Disposal V

Concentrate Recycle

1r

Sampling Points

Figure 2. Schematic of the filtration system.

The hyperfiltration system, which consists of porous stainless steel tubes internally coated with specially formulated chemical membranes, has been demonstrated to successfully treat water contaminated with a number of hazardous or toxic materials. In this system, contaminated ground and surface waters are pumped through the

196

Liquid Filtration

filtration system tubes and contaminants are collected inside the tube membrane while "clean" water permeates the membrane and tubes. The system has been shown to be highly versatile and able to effectively remove a variety of materials from contaminated waste streams, including petroleum hydrocarbons, benzene, toluene, ethylbenzene, xylene, dioxins, chlorinated furans, and heavy metals. This extent of versatility is provided largely by the application of several types of chemical membranes with distinct permeability and ion exchange capabilities. In the field study discussed in this chapter, one type of chemical membrane was used to coat the porous stainless steel tubing, with an aim to provide the optimal separation efficiencies for the higher molecular weight contaminants. These are generally regarded to be most "hazardous" molecules and are generally more resistant to degradation in the environment. The results confirmed that this membrane system was very effective in removing more than 95 % of the high molecular weight polynuclear aromatic hydrocarbon (PAH) contaminants, the most carcinogenic components, during the monitored demonstration run. However, on the average, approximately 20% of the lower molecular weight phenolic contaminants were removed by this membrane type. The vendor feels that this percentage can be improved by adjusting membranes and flows in the field. In the study, the relatively heavy concentrations of contaminants in the feed material for the hyperfiltration unit precluded effective use of other "tighter" types of membranes as the first pass barrier due to the potential for fouling. Optimally, a combination system employing the initial membrane used here to remove high molecular weight contaminants, followed by a "tighter" membrane to remove lower weight phenolics from the permeate of the first membrane would have been more likely to provide the full spectrum of contaminant removal desired.

Petroleum Hydrocarbon Hyperfiltration Other waste streams, highly contaminated with lower molecular weight molecules, can also be hyperfiltered and concentrated by using alternate chemical membrane types. One example is the hyperfiltration of waste water from certain oil refining operations containing significant amounts of benzene, toluene, ethylbenzene, and xylenes (BTEX), as well as other petroleum hydrocarbons. A summary of the hyperfiltration results of this treatment is as follows: Contaminant

Original Feed Sample

Permeate Sample

%Removal

Benzene Toluene Ethylbenzene Xylenes

2200ppb 7640ppb 2590ppb 12200ppb 483ppm

34ppb 82ppb 15ppb lOOppb 3.5ppm

98.5% 98.9% 99.4% 99.3% 99.3%

TPH

Note the TPH stands for total petroleum hydrocarbons.

197

Advanced Membrane Technology for Wastewater Treatment

In a separate study, condensate from a methane-recovery operation at a municipal landfill was treated for the removal and concentration of a large spectrum of contaminants, including naphthalene, heavy metals, and BTEX. The results of this are: Contaminant

Original Feed Sample

Permeate Sample1

%Removal

Toluene Ethylbenzene Xylenes Naphthalene

177ppm 268ppm 561ppm 90.4ppm

<5ppm < 5ppm < lOppm < lOppm

>97,2% >98.1% >98.2% >89.0%

All permeate readings were below the detectable limits for the analysis method used in this study.

These results demonstrate the capacity and versatility of the hyperfiltration system in treating a variety of waste steams and achieving effective volume reduction in removal of contamination from groundwater, municipal landfill leachates, or contaminated petroleum waste streams. The value of this technology can be further enhanced by coupling it with a biodegradation process. This can be achieved by using the hyperfiltration concentrate as a bioreactor feed stream, as well as by using the hyperfiltration system to polish bioreactor effluent to yield two streams: one, a clean stream suitable for discharge; and the other, a polished concentrate to feed to the bioreactor. This creates a closed loop for targeted contaminants, and provides for an efficient continuous flow remediation design. In addition to the application of hyperfiltration technology to the remediation of creosote-contaminated groundwater, effective biodegradation of creosote and pentachlorophenol has also been achieved using specially selected, non-engineered microorganisms in a bioreactor system. The combination of these two systems, hyperfiltration and bioremediation, provides a novel and reliable means to first concentrate the waste feed to the bioreactor to an optimal level for efficient bioremediation activity, as well as to provide for a final polishing step using hyperfiltration of bioreactor effluent. For this study, a bi-phasic bioreactor design was utilized, operating in a semicontinuous flow process, having a hydraulic retention time of four days. Groundwater, with contaminant concentrations as high as 7000 ppm creosote, was treated on-site. This demonstration achieved a removal efficiency of greater than 99% for total polynuclear aromatic hydrocarbons (PAHs). This includes a removal rate of 98% for the most recalcitrant, and most hazardous fraction of the PAHs, and 88% for PCP. The field test proved that biotechnology application for hazardous waste remediation can be effective at an actual waste site. PAHs are a widespread contaminant of soil and groundwater typical of creosote wood treating facilities, manufactured gas plants, refineries and related industries. Bioremediation has been attempted for PAH constituents in several studies and field applications, but until now, biodegradation using indigenous bacterial strains has been able to achieve only 50% - 75% removal of PAHs. The untreated portion is generally comprised of the recalcitrant high molecular weight (HMW) PAHs which are those

198

Liquid Filtration

compounds with 4, 5 or 6 fused rings and are the PAH components which are known or suspected carcinogens (see Table 7). Similarly, PCP, a common wood preservative, has proven difficult to biodegrade under field conditions. Table 7. Classes and characteristics of PAHs.

Low Molecular Weight PAHs — relatively non-hazardous, relatively easy to degrade. naphthalene 2-methylnaphthalene 1-methyinaphthalene biphenyl 2, 6-dimethylnaphthalene 2, 3-dimethylnaphthalene Medium Molecular Weight PAHs — some potentially hazardous to human health, more complex but still biodegradable by many bioremediation systems, acenaphthylene anthracene fluorene phenanthrene 2-methylanthracene anthraquinone High Molecular Weight PAHs — several carcinogens, very slow rates of biodegradation without specialized microbes. fluoranthene pyrene benzo(b)fluorene chrysene benzo(a)pyrene benzo(a)anthracene benzo(b)fluoranthene benzo(k)fluoranthene indeno( 1,2,3-c, d)pyrene USEPA scientists have isolated and identified several strains of naturally occurring soil bacteria capable of mediating PAH and Pep degradation at rates in excess of those achieved by undifferentiated communities of indigenous microbes. The strains are identified as: CRE 1-13: low and medium weight PAH degraders comprised of an assemblage of 13 Pseudomonads.

Advanced Membrane Technology for Wastewater Treatment

15*9

EPA 505: a strain of Pseudomonas palleimobilis capable of high rate degradation of HMW PAH constituents. SR3: a strain of Pseudomonas sp. which degrades PCP. All organisms have been shown to mineralize their target contaminants. Additionally, EPA 505 has been patented for use in the degradation of high molecular weight PAH, The bioremediation process was tested on groundwater at a Superfund site where waste liquids from the manufacturing process were placed in unlined surface impoundments on-site. These impoundments often overflowed into drainage ditches which discharged into local waterways. In addition, wastes have migrated into the shallow aquifer, contaminating both soil and groundwater. The Superfund site has large volumes of shallow groundwater contaminated by creosote and Pep. In order to prove the capability of the organisms to degrade these contaminants under field conditions, a highly contaminated groundwater was chosen as the test matrix. Groundwater was pumped from the aquifer via an existing monitoring well. The extracted groundwater was stored in an equalization tank prior to the test. From this tank, the contaminated feed was pumped to the two-stage bioreactor treatment system (that is illustrated in Figure 3). Each bioreactor had a hydraulic capacity of 200 gallons and was designed to provide mixing and up-lift type aeration. The bioreactors were operated sequentially, i.e., the contaminated water was transferred to Bioreactor 1 (BR1) at a pre-set flow rate of four days. After four days, when BR1 was full, the water was allowed to overflow into Bioreactor 2 (BR2). Laboratory grown concentrates of CRE 1-13 (specialized degraders for the low and medium weight PAHs) were added to BR1, along with nutrients and sparged air, and the tank was mixed. Similarly, BR2 was inoculated with EPA 505 (HMW PAHdegrading strain) and SR3 (Pep degrader) during its eight days of operation. Treated flow from BR2 was held in a tank for testing; after testing, the water was discharged to the city sanitary sewer. During operation of the bioreactors, samples were collected for the analysis of key operating parameters, such as dissolved oxygen, nutrient levels, total organic carbon and suspended solids. Microbial analysis was performed to assess the cell concentration of the specialized bacteria being added. All cultures were prepared in advance and added to the bioreactors. Additional samples were collected to measure the contaminant concentration across the bioreactor, as well as in the various portions of the treatment system, in order to calculate a mass balance.

200

Liquid Filtration

Contaminated Groundwater

ORE 1-13

EPA 505 SR3

Nutrients & Air

Bioreactor No. 1

l

Bioreactor No. 2

Discharge to City Sanitary Sewer

Figure 3. Simplified process flow for SBP Technologies Inc. 's bioreactor system.

The overall PAH and PCP degradation performance of SBP's treatment system is shown in Table 8. Table 8. Summary of bioremediation results of PAN and PCP removal.

Contaminant

Influent (mg/L)

Effluent (mg/L)

%RemovaI

Low Molecular Weight PAHs

31

8.1

<99

Med Molecular Weight PAHs

539

1.6

>98

High Molecular Weight PAHs

368

5,2

98

Pentachlorophenol

256

31

88

These results represent a significant advancement in PAH bioremediation. Not only has the total PAH been reduced to <1% of its original concentrations, the HMW PAH, a class of compounds which are typically not removed in biological remediation systems, have been reduced by 98-plus %. This removal has the additional significant

Advanced Membrane Technology for Wastewater Treatment

201

characteristic of having been mediated by an organism specially and specifically isolated, cultured and introduced into the bioremediation system to remove this class of compounds, and this has been accomplished under typical field remediation conditions. To further validate the treatment technology, the contaminated groundwater was subjected to tests for toxicity and teratogenicity before and after treatment (sampled before carbon treatment). As determined by Microtox™, embryonic Menidia beryllina, Mysydopsis bahia, and Ceriodaphnia dubia bioassays, these indicators of potential threat to human health and the environment were significantly reduced. This data shows that the treatment system was effective in removing the hazardous properties of the waste material, while simultaneously demonstrating that no metabolic byproducts or toxic intermediates were created by the microbial biotransformation of the waste constituents. A mass balance was performed by comparing the total influent loading of PAHs to the residual PAHs left in the treatment system at shutdown. The HMW PAHs were found to be 80% removed (20% remained in the residual). In a scaled-up system with higher volume of wastewater treated, the removal efficiently at steady state would approach the percent removal from influent to effluent (98%). The approach to bioremediation of hazardous waste sites is a "multi-phase" strategy, meaning that the technology is adaptable to the treatment of soil, sludges, leachates, groundwater and surface water. By the multi-phase approach, the waste matrix is modified and managed to support the greatest level of specialized organisms survival and highest biodegradation rates. Examples are given below. Bioremediation is applicable to a wide range of organic contamination. SBP has focused on research, development and field implementation of biotechnology-based approaches to some of the more difficult hazardous waste constituents, in particular wood preserving wastes and solvent contamination. SBP has bioremediation solutions for: PAHs (e.g., creosote, coal tar) Chlorinated Aliphatic (e.g., TCE) Pentachlorophenol Petroleum Hydrocarbons Treatment of contaminated liquids such as water, leachate, filtrate, groundwater, storm water, surface water and industrial process waste water can be accomplished by several of SBP's technologies. Certain waste streams can be concentrated by the hyperfiltration units; the permeate is clean water and the "concentrate" contains the reduced volume of the pollutants. These concentrated contaminants can often be bioremediated, thus minimizing the waste stream. Other liquids may be treated directly, either by biological processes, or in the case of volatiles, air stripping with biotreatment of the pollutants in a gas phase bioreactor, or "biofilter".

202

Liquid Filtration

There are several approaches for the handling of contaminated soil. Solid phase systems are used when biostimulation of indigenous bacteria is appropriate, Slurry phase reactors, either in-vessel or in situ are used for soils or sludges requiring better control of microbial systems. A multi-phase approach combining soil washing and hyperfiltration to reduce volume and bioreactor treatment of the wash water is appropriate for heavily contaminated wastes.

Summary of Case Study Analytical Results The goal of the demonstration project was to study the effectiveness of the formed-inplace hyperfiltration membrane unit, operating in a cross-flow mode, for the concentration of PAHs (and PCP) from contaminated groundwater found at a wood treating site. The original plans called for a companion study of a proprietary biodegradation process for the concentrate; however, a decision was made during this study not to carry out that portion of the study. Based on available information, including a Remedial Investigation / Feasibility Study (RI/FS), the American Creosote Works site in Pensacola, Florida was selected for the investigation. The site was included on the National Priorities List on the basis of the RI/FS study which suggested that the groundwater under the site was heavily contaminated with both PAHs and PCP. The American Creosote Works site had been used for wood treatment from 1902 to 1981. Creosote was the preservative used until the 1970's when a shift was made to pentachlorophenol in a light oil. Over the years, wood impregnation was carried out in open toughs, resulting in spills and drippage which seeped into the ground. A well selected for the study yielded sufficient groundwater volume and flow but the output was contaminated with free product. Field COD tests were used to determine that five-fold dilution of the groundwater with city water was necessary to meet the 100-500 mg/L COD input guideline for feedwater to the pilot filtration unit. Approximately 700 gallons of the groundwater was pumped from the well on alternate days. The groundwater was allowed to settle for several hours, and the aqueous phase pumped to a separate tank. The contaminated aqueous phase was diluted with 2800 gallons of carbon-treated city water to bring the contaminated level of the feed in the desired range of 100-500 mg/L COD. Approximately 1000 gallons of this diluted stream was used as the feedwater to the filtration unit per day. The average characteristics of this feed stream are shown in Table 9, and are based on composite samples taken on each day over the six days of the demonstration. To achieve the desired volumes and concentrations, the feedwater was introduced during the first two hours (approximately) of each day's run. For the remainder of each day's run, the reject (concentrate) was recycled, becoming the feed, while the permeate was continuously removed and discharged after samples were composited for analysis. Each day's run was terminated when the flowmeters indicated that the

Advanced Membrane Technology for Wastewater Treatment

203

desired 800 gallons of permeate had been discharged, leaving approximately 200 gallons of concentrate. Table 9. Average characteristics of feed stream to the filtration unit.

Average TDS(ppm) TSS(ppm) TOC(ppm) COD(ppm) Oil/Grease (ppm) Semivolatiles (ppb) Phenol 2-Methylphenol 4-Methylphenol 2,4-Dimethylphenol Pentachlorophenol Benzole Acid Naphthalene 2-Methylnapthalene Acenaphthylene Acenaphthene Dibenzofuran Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo(a)Anthracene Chrysene Benzo(b)Fluoranthene Benzo(k)Fluoranthene Benzo(a)Pyrene ( ) estimated value

4,900 2,308 6,917 1,817 (2,425) (1,421) 12,867 4,525 (138) 6,842 4,875 5,925 17,083 1,983 7,008 4,700 1,235 1,127 (460) (428) (312)

Samples were analyzed for oil and grease, dissolved and suspended solids, volatile organics, TOG, COD, and dioxins/furans in addition to the list of designated semi volatile organics used in evaluating the developer's claims. As noted earlier, the system was operated by introducing approximately 1000 gallons of the diluted feedwater (about 1 volume of groundwater to 4 volumes of city water) over a two hour period at a rate of about 8 gpm. At that time, recycle of the concentrate was initiated and continued until approximately 800 gallons of permeate was generated and approximately 200 gallons of concentrate was collected.

204

Liquid Filtration

Based on operational parameters (flux, pressure), the developer determined that washing of the unit was necessary every other day. At the end of the second day, the fourth day, and the sixth day approximately 200 gallons of city water, to which cleaning chemicals had been added, was flushed through the system and collected This material was also analyzed. Grab samples of the feed stream were collected every 15 minutes during the initial two-hour single pass phase, while permeate grab samples were collected every 45 minutes over the whole run. Laboratory samples were prepared by compositing the grab samples on a flow proportional basis. After the initial single pass, the concentrate was recycled for approximately four hours in each test and one composite sample was taken at the end of each run. All samples were transferred to bottles, inhibited or preserved as called for by the individual test methods, labeled, sealed, and shipped in ice-filled coolers to off-site laboratories by overnight express. Instantaneous flow and accumulated volumes of feed and permeate were measured automatically using calibrated flowmeters. Volume of the final concentrate was determined by the difference between the feed and permeate volumes. The temperature and pH in the three steams were measured in the field and recorded to assure that there was no gross change in characteristics that could affect filtration effectiveness. At the end of the sixth day of operation, after all the required samples had been taken, the process was allowed to continue for several hours, further concentrating the contaminants. Additional samples of the permeate were taken about every 25 gallons to observe whether there was a fall-off in contaminant rejection rate. The filtration unit operated in a batch mode for six hours each day, for six days, and processed approximately 1000 gallons of feed per day. Over the six day test period, permeate flux was a relatively constant 0.0085 gallons/min/ft2 (coefficient of variation < 10%). Based on a total membrane area of 300 ft2 for the system, the permeate flow rate for the four-module filtration unit averaged 2.6 gpm. The unit was cleaned every two days with approximately 200 gallons of city water and membrane cleaning materials. Over the two day period, the surface of the membrane gradually fouled, requiring an increase in feed pressure in order to maintain a constant permeate flux. The cleaning at the end of each two-day period was sufficient to restore the original flux. Semivolatile Contaminant Analysis Contaminant Reduction A selected list of semivolatile organic compounds was used to assess the effectiveness of the filtration unit. The analytes consisted of all quantifiable semivolatile compounds detected in the feed.

Advanced Membrane Technology for Wastewater Treatment

205

Rejection efficiency was defined as the change in total concentration for these contaminants between the feed and the permeate: % Rejection =

c -c

J-L_,_1£ ———

(2)

xlOO

Cf = total concentration of contaminants in feed C p = total concentration of contaminants in permeate On the basis of composite samples, rejection for the total designated contaminants averaged 74% over the six days, as shown in Table 10. Analysis of the data reveals that the system is very effective for the rejection of polynuclear aromatic hydrocarbons and much less effective for phenols. Separated in this fashion, rejection for the PAHs averaged 92% over the six days while rejection for the phenols averaged 18%. Table 10. Overall semivolatile rejection for the filtration unit.

Day

Feed (ppm)

Permeate (ppm)

Concentrate (ppm)

Rejection %

1 2 3 4 5 6

104 91 92 104 85 60

18 24 26 22 23 24

206 585 248 242 199 538

83 74 72 79 73 60

A more accurate means of quantifying and graphically representing rejection is to calculate the reductions on a logarithmic basis. For example, attempting to graph differences between 90% and 99% rejections does not fully depict the fact that a tenfold reduction has occurred. In order to accurately represent this process, it is necessary to calculate the order of magnitude of rejection (ORD) as the log value of the ratio of the contaminants concentration in the feed to the concentration in the permeate. The equation is written as: ORD = Log (CyCp)

(3)

Therefore, an ORD of 2.0 means that a 100 fold decrease in contamination has occurred in the permeate relative to the feed. An ORD of 1.0 means a 10-fold decrease has occurred.

206

Liquid Filtration

For the semivolan'le constituents, it appears that rejection follows the molecular weight of the chemical. Figure 4 presents the average ORD over the six days versus the molecular weight of the individual compounds.

._ .2 u

3 "O 0)

cc

o>

T3 3 +*

"E

O5 CO

0)

"S o

90

110

130 150 170 Molecular Weight

190 210

Figure 4. Molecular weight versus order of magnitude reduction (ORD).

Mass Balance Combining the masses of pollutants in the permeate and the concentrate on a daily basis did not provide a good material balance, relative to the feed. Only when the material in the washwater was also included was all the material accounted for. Based on the volume of washwater used in each washing operation, approximately 200 gallons, about 8% of the masses of the designated pollutants are retained on the membrane or in the liquid in the system. Table 11 provides a summary of the contribution of the different streams for each two-day period between wash operations. The relative distribution of pollutants in the washwater was very similar to mat in the concentrate, with PAHs far exceeding the phenols. However, the concentration of pollutants in the washwater was similar to the feedwater. Extended Operation

After the system yielded the desired 200 gallons of concentrate on the sixth (last day) of the test and all necessary samples had been taken, operation of the system was continued with further recycle of the concentrate as long as possible to observe the behavior of the unit (i.e., the rate of folding and the quality of the permeate). The extended operation resulted in only thirty seven gallons of concentrate remaining,

207

Advanced Membrane Technology for Wastewater Treatment

representing a 97% volume reduction. However, during this time the rejection efficiency decreased, as expected, as concentration continued (Table 12). Table 11. Mass contribution of total semivolatiles in all streams (2 day totals).

Days

Feed (Influent) gms

Perm gms

Cone gms

Wash gms

Total gms

Recovery %

1&2 3&4 5&6

111 778 581

140 159 151

556 320 527

46 84 32

742 563 710

95 72 122

6 day recovery = 94% Table 12. Behavior on extended filtration.

Additional Gallons of Permeate Produced

Semivolatile Rejection % (calculated from initial feed)

25 50 75 100 150

47.5 48.5 60.0 32.3 21.7

Conventional Parameters In addition to the designated semivolatile pollutants, the removal of several conventional parameters from the feedwater was also evaluated. These included total dissolved solids (TDS), total suspended solids (TSS), chemical oxygen demand (COD), total organic carbon (TOC), and oil and grease (O&G). Comparison of feed and permeate confirm that the membrane is capable of removing larger molecules such as oil and suspended solids, but is not nearly as efficient in the removal of dissolved material (TDS) or otherwise-unidentified organic species (COD) and TOC. Such results, summarized in Table 13, suggest that the membrane is operating predominantly as a hyperfiltration membrane. In addition, considerable concentrations of organics remain in the permeate even though the PAHs are efficiently removed. For example, the ratio of semivolatiles/COD decreases significantly, from about 0.2 in the feedwater to about 0.07 in the permeate, demonstrating that semivolatiles are not the major contributor to the COD, and that significant unidentified, smaller molecular weight compounds can pass through the membrane.

208

Liquid Filtration

Pofychforinated

dibenzo-p-dioxins/dibenzofurans

Polychlorinated dioxins (PCDDs) and furans (PCDFs) are often encountered in the manufacture of pentachlorophenol and can remain with the commercial product. To investigate this matter, selected samples of the three streams taken on the first, second, and sixth day of the demonstration were scanned for the various dioxins and furans using high resolution GC coupled with low resolution MS. A number of the more highly chlorinated dioxin and furan species were found in the feed water and in the reject at significantly higher concentrations, but analyses of the permeate indicated efficient removal of these pollutants. The 2,3,7,8-TCDD isomer was not found above the detection limit in any sample of feed, permeate, or concentrate. The 2,3,7,8TCDF isomer was found in only one sample of feed and one sample of concentrate, in both cases the value was <0.5 ng/L (ppt). The major dioxins and furans were the octa species. These results are consistent with rejection by molecular weight or oil solubility, but not by number of rings. Table 13. Dioxins/furans in process streams.

Day 1

Feed (ng/L)

Perm. (ng/L)

. 2378 TCCD 12378 PeCDD 123478 HxCDD 123678 HxCDD 53.0 123789 HxCDD 74.6 1980 1234678 1.6 38.9 HpCDD 42355 OCDD 14870 2378 TCDF 12378 PeCDF 23478 PeCDF 11.2 3.2 123478 HxCDF 8.5 123678 HxCDF 234678 HxCDF 346 123789 HxCDF 1234678 HpCDF 28.3 1234789 HpCDF 3.3 2595 OCDF 1030 An 1 -' indicates the isomer was absent or below detection limits

Cone. (ng/L)

Removal %

. 277.7 13.3 3835 99.7 .21 52 8.7 30.4 1066 76 99.7 in stream.

99.9

209

Advanced Membrane Technology for Wastewater Treatment Table 13 (Continued),

Day 3 2378 TCDD 12378 HxCDD 123478 HxCDD 123678 HxCDD 123789 HxCDD S 234678 HpCDD OCDD 13250 2378 TCDF 12378 PeCDF 23478 PeCDF 123478 HxCDF 123678 HxCDF 234678 HxCDF 123789 HxCDF 1234678 HpCDF 1234789HpCDF OCDF 22.4 * An'-' indicates the isomer was absent

Day 6

Feed (ng/L)

Perm. (ng/L)

Cone. (ng/L)

. -

_ -

-

2.4 47.7 6.7 1480 7.9 .37 8.1 300

Removal %

116 9.9 3630 99.9 5 . 28.4 4.2 17.0 689 40.1

3413 . -

1223 or below detection limits in stream.

Feed (ng/L)

_ 2378 TCDD 12378 PeCDD 1.9 123478 HxCDD 123678 HxCDD 30.6 4.4 123789 HxCDD 1370 1234678 HpCDD 8.8 OCDD 9200 .39 2378 TCDF 12378 PeCDF 23478 PeCDF 5.1 123478 HxCDF 123678 HxCDF 1.5 5.1 234678 HxCDF 123789 HxCDF 198 1234678 HpCDF 1234789 HpCDF 14.1 OCDF 473 .50 An'-' indicates the isomer was absent or below detection

Perm. (ng/L)

Cone. (ng/L)

_ -

_ .

.36 90400 5235 limits in stream.

8.1 334 27.6 8950 99.90 1.5 1.7 111.7 34.2 42.6 2055 126.2 99.89

Removal %

99.97

210

Liquid Filtration

Volatile Organics Analyses for volatile organics (VOCs) were carried out during the runs on days 1,3, and 6 because the developer was somewhat concerned that the high pressures used in the system might force volatiles through the membrane into the permeate. All VOC concentrations were low in the feedwater; the same species were also found in the permeate and the concentrate but at lower concentrations. The principal VOC was acetone, but the expected BTEX species were also present. Refer to the data in Table 14. Table 14. Votatiles in process streams.

Methylene Chloride Acetone Carbon Disulfide 2-Butanone Benzene Toluene Ethylbenzene Styrene Xylenes

Feed (Mg/U * *

Permeate 0*g/L)

Concentrate 0*g/L)

* *

* *

2 32

<5 31

16 20 12 18 43

7

<5 32 <5 <5 <5 <5 <5

16 (7) 7 26

probable laboratory contamination ) estimated value

Closure The technology reviewed in this chapter represents a relatively new process with the intended application of wastewater treatment. The process is best applied, as demonstrated, with a pump and treat system and can be integrated with other technologies such as bioremediation. The economics of the process are favorable for the intended application, however the reader should apply separate cost studies to each intended application to assess the applicability of this technology. Further information can be gained from the United States Environmental Protection Agency.

7 SLUDGE DEWATERING OPERATIONS Introduction This chapter provides a brief overview of alternative dewatering technologies used in wastewater treatment and sludge concentrating applications. Many of the techniques described can be used in conjuction or integrated with filtration operations. The objective of dewatering is to increase the solid content (decrease the water content) of the paniculate matter separated out from filtration for one or more of the following reasons: 1. 2. 3. 4. 5.

Dewatered solids are more easily handled. Dewatering is normally required prior to incineration to reduce fuel requirements. Dewatering reduces the cost of subsequent treatment processes, particularly thermal processes. Dewatering is required prior to land disposal of solids. Dewatering reduces the costs of transporting wastes recovered from filtration to their ultimate disposal by reducing their volume and weight.

In most cases, the percent solids content of a dewatered solids is set by the requirements for subsequent treatment and disposal. Each treatment technology has an optimum range of percent solid, above or below which the technology will not operate efficiently and economically, if at all. For example, combustion requires that the solid content be greater than 24 %, preferably in the 28-30 % range for more economical operation. Particulates will vary in percent solid depending on the technology and application. Dewatering technologies can be subdivided into two general processes: air drying processes and mechanical processes. This chapter provides an overview of these technologies and concentrates on typical sludge dewatering operations used for wastewater treatment. 211

212

Liquid Filtration

Overview of Dewatering Technologies "Air drying" refers to those dewatering techniques by which the moisture is removed by natural evaporation and gravity or by induced drainage. Air drying is less complex, easier to operate, and requires less operational energy than mechanical dewatering. Air drying also can produce a dryer sludge than mechanical dewatering; up to 40 % solids under normal operation and over 60 % solids with additional drying time or with the use of underdrainage systems. Air drying processes do, however, require a larger land area and are more labor intensive than mechanical processes. Contaminant releases via seepage, drainage, and volatilization during the dewatering process must also be considered. The most widely applicable and economical air drying process available for sludges is an appropriately managed confined disposal facility (CDF). CDFs are engineered structures designed to retain solids during dredging and provide storage time for gravity drainage, consolidation, and evaporation. The rate of dewatering may be accelerated by using underdrains, pumps, or wick drains. Mechanical dewatering involves processes in which water is forced out of the sludge through mechanically induced pressures. Mechanical dewatering processes include the following: filtration, including belt filter presses, chamber filtration, and vacuum rotary filtration; centrifuges, including solid bowl and basket; and gravity thickening. Table 1 provides a summary of the primary dewatering techniques. A brief description of each technique for dewatering applications follows. Belt filter presses accomplish dewatering by carrying the sludge between two tensioned porous belts and squeezing out the water as the sediment/belt "sandwich" passes over and under various size rollers. All belt filter presses incorporate the folio wing features: a polymer conditioning zone, a gravity zone, a low pressure zone, and a high pressure zone. Polymer conditioning produces a superflocculation phenomenon that allows water to drain more efficiently. The gravity drainage zone and the low pressure zone prepare the sediment for the high pressure zone, where most of the water removal actually takes place. Belt filter presses are common dewatering choices in Europe and the United States. They can dewater sediment rapidly and do not take as large an area as the air drying processes. Belt filter presses are, thus, often used in confined locations (such as cities) and where large volumes of sediment must be dewatered. They do not dewater as completely as air drying processes, however, and may be limited by the percent solid demands of the treatment to be used. Furthermore, if the sludge is very gritty, the belts may wear out rapidly.

Table 1. Summary of Dewatering Techniques. Technique Confined Disposal Facility

Applications Dewatering sediment of any grain size to a solids content of up to 60 % up to 99 % solids removal. Generally used for large scale dredging operations where land space is available.

Limitations Requires large land areas. Requires long set-up time. Labor costs associated with removal or dewatering sediments are high. Systems using gravity drainage are prone to clogging.

Secondary Impacts Potential for groundwater contamination.

Relative Cost Low to High

Potential for localized odor and air pollution problems.

System using vacuums require considerable maintenance and supervision. Systems based on electroosmosis are costly. Belt Filter Press

Used to dewater fine grained sediments. Capable of obtaining relatively dry filter cake obtaining up to 45 - 70* % solids; able to achieve solids capture of 85 - 95 % . Generally best suited of filtration methods for mobile treatment systems.

Chamber Filtration

Performance is very sensitive to incoming feed characteristics and chemical conditions. Belts can deteriorate quickly in presence or abrasive material.

Used to dewater fine grained sediments.

Costly and energy intensive.

Capable of obtaining a relatively dry filter cake with a solids content up to 50 - 80*%; able to achieve a high solids capture rate of up to 98%.

Replacement of filter media is time consuming.

Generates a substantial amount of waste water that must be treated.

Medium

Generates a wash water that must subsequently be treated.

High

I

o

•g

Table 1 Continued. Vacuum Rotary Filtration

Used to dewater fine grained sediments capable of obtaining a filter cake of up to 35 - 40% solids and a solids capture rate of 88-95%.

Least effective of the filtration methods of dewatering.

Generates a wash water that must be treated.

High

i CL

Energy intensive. Solid Bowl Centrifuge

Thickening of dewatering sediments; able to obtain a dewatered sludge with 15-35% solid; solids capture typically ranges from 90-98%.

Not as effective in dewatering as filtration or lagoons.

Suitable for areas with space limitations.

Process may result in a build up of fines in effluent from centrifuge.

No significant secondary impacts.

Medium to High

No significant secondary impacts.

Medium to High

Scroll is subject to abrasion. Basket Centrifuge

Gravity Thickening

Thickening of dewatering sediments; able to obtain a dewatered sludge with 10-25% solids. Solid capture ranses from 80 98 %.

Not as effective in dewatering as solid bowl centrifuge, filtration, or dewatering lagoons.

Suitable for areas with space limitations.

Process may result in a build up of fines in effluent from centrifuge.

Good for hard-to-dewater sludges.

Units cannot be operated continuously without complex controls.

Thickening of sediments slurries to produce a concentrate that can then be dewatered using filtration or dewatering lagoons. Able to produce a thickened product with a solids concentration of 15-20%.

Least effective method for dewatering sediment slurries. Requires use of a substantial amount of land.

Potential for localized odor and air pollution problems.

Low to Medium

* Percent solids achievable may represent values for optimal conditions and do not necessarily represent normally expected values. Dredged sediments are often fine-grained and difficult to dewater to the maximum indicated values.

Sludge Dewatering Operations

215

Chamber filter presses also use positive pressures, but apply the pressure to the sediments inside rigid, individual filtration chambers operated in parallel. Vacuum rotary filtration uses negative pressure to pull the water to the interior of a drum while the sludges adhere to the exterior. Centrifuge^ use the centrifugal forces created by a rapidly rotating cylindrical drum or bowl to separate solids and liquids based on variations in density. There are two types of dewatering centrifuges: the solid bowl centrifuge and the basket centrifuge. Gravity thickening is accomplished in a continuous flow tank. Sediments settle to the bottom and are removed by gravity or pumping. Water overflows the tank and leaves through an effluent pump. Gravity thickeners are used primarily in tandem with other pre-treatment technologies: they reduce the hydraulic toad to subsequent pre-treatment options. Particle Classification Particle classification separates the slurry according to grain size or removes oversize material that is incompatible with subsequent processes. Classification by grain size is important in the management of sludges contaminated with toxic materials since the contaminants tend to adsorb primarily onto fine grain clay and organic matter. The small grain solids of a specific size or less can be treated while the relatively noncontaminated, coarser soils and solids can be disposed of with minimal or no additional treatment. Separation technology depends on the following: volume of contaminated sludges; composition of the sludge, such as gradation, percent clays, and percent total solids; characterization of the contaminants; and site location and surroundings, including available land area. Particle classification options include .screening processes that depend on size alone, processes that depend on particle size and density or density alone, and processes that depend on conductive or magnetic properties of the particles. Particle classification technologies include: impoundment basins, hydraulic classifiers, hydrocyclones, grizzlies, and screens. Following is a description of these technologies. Applications, limitations, and relative costs for the first three listed particle classification technologies are identified in Table 2. Impoundment basins allow suspended particles to settle by gravity or sedimentation. A slurry of dredged material is introduced at one end of me basin; settling of solids — depending on the particles' diameters and specific gravities — occurs as the slurry flows slowly across the basin. The flow resulting at the opposite end has a greatly reduced solids content. Multiple impoundment basins in a series can separate sediments across a range of sizes. Hydraulic classifiers are rectangular tanks that function similarly to impoundment basins. They have a series of hoppers along the length of a tank which collects sediments of various sizes. Motor-driven vanes sense the level of solids and activate discharge valves as the solids accumulate in each hopper. Hydraulic classifiers may

216

Liquid Filtration

be used in tandem with spiral classifiers to separate fine grained materials such as clay and silt. Portable systems that incorporate hydraulic and spiral classifiers are available. Table 2. Summary of sediment/water separation techniques. Technique

Impoundment Basin

Applications

Used to remove particles down to a grain size of 20 - 30 microns without flocculants, and down to 10 microns with ilocculants.

Limitations

Requires large land areas.

Secondary Impacts

Relative Cost

Potential for ground water contamination.

High

No significant impacts.

Medium

No significant impacts.

Medium

Requires long setup time.

Provide temporary storage of dredged material. Allow classification of sediments by grain size. Hydraulic Classifier

Used to remove particles from slurries in size range of 74 149 microns (fine sand to coarse sand) .

Hydraulic throughput is limited to about 250 - 300 tph regardless of size. Not capable of producing a sharp size distinction. Requires use of large land area for large scale dredging or where solids concentrations are high.

Hydrocyclones

Used to separate and classify solids in size range of 2000 microns or more down to 10 microns or less.

Not suitable for slurries with solids concentrations greater than 10 20 %.

Hydrocyclones use centrifugal force to separate sediments. A hydrocyclone consists of a cone-shaped vessel into which a slurry is fed tangentially, thereby creating a vortex. Heavier particles settle and exit at the bottom while water and sediments exit through an overflow pipe. Hydrocyclones may be useful where a sharp separation by particle size is needed.

Sludge Dewatering Operations

217

Grizzlies are vibrating units reliable in the removal of oversized material, such as bricks and rocks. Grizzlies are very ragged and are useful in reducing the amount of abrasive material in order to minimize wear on subsequent, more delicate, technologies. Screens may be vibrating or stationary and operate by selectively allowing particles to pass through them. As the slurry passes over the screen, fine-grained particles and water sift through the screen and larger particles slide over the screen. Screens come in a variety of types with a variety of applications to contaminated sediments. Handling and Rehandling

The amount of handling and rehandling required by various pretreatment options will also influence pretreatment decisions. Especially with severely contaminated sludges, all equipment mat comes in contact with the sediments will require subsequent decontamination. For example, air drying heavily contaminated sludges requires that the sediments be put in the drying structure and later rehandled when the dewatered sludges are removed in a more highly concentrated form. Rehandling also mechanically disrupts the sediments and increases die probability of introducing contaminants into the environment. Conversely, a series of pretreatment steps requiring rehandling may be the most efficient way of separating the contaminated sediments and preparing them for treatment. For example, the Dutch are using particle classification (hydrocyclones) and dewatering (belts, filter presses, chamber filter presses, and others) to improve the quality of solids prior to treatment or disposal. Although methods of treatment are not addressed in this chapter, the water generated during dewatering generally contains contaminants and suspended solids, and may require further treatment. Use of Drying Beds Drying beds are among the first method of sludge dewatering. They consist simply of shallow ponds with sand bottoms and tile drains as illustrated in Figure 1. Sludge is pumped to these beds at a depth of 6-12 in. and the time required for the sludge to de water to a liftable consistency ranges from several weeks to several months. In northern climates, sand beds can be covered with a greenhouse roof to facilitate evaporation. The removal of water from sludge in drainage beds is in two stages. Initially, the water is drained through the sludge into the sand and out tile drains. This process might last a few days, until the sand is clogged with the fine particles and/or all the free water has drained away. Further dewatering occurs by evaporation. During the initial drainage a considerable portion of the water is first drained by settling of the solids and compaction. Secondary drainage occurs by the formation of channels that facilitate the movement of water.

218

Liquid Filtration

'i

Tf

;

i1

3

p

.

•9

i

TQ J

|«

la 1i II

11

£

11

c

.5 *

i1

3

ii

.,.„-»,„», ..»......,*i L_..«..... »....»...,••* " r »•\ 6 - in ,' vitrified pipe l*kf -^ i with plcttic joints ! 1 \ (1 |1 \ ***».. *•, J*

c

]

»..**»«...».» ™-:-:^{| i" ! j

11

II II

\

E

:i II

1!

H \

i tif

I:

Spl»»h boxii

V A

11

1 \

J

...

JJ

J.

i

ZFSft&l "*fet

|

j-ictr*ct*

ft ^_jTTTTr B .iA "IT». tmn^ou

*s

shear oaie 1

|i

1 J

t A

15

Is

•_ __

UL :; ...

1

CJj| w.,.^ . J

^""^yti""^^

2-lA, PLAN 6->n. fine sand 3-trt coarse sand 3 3 *n medium grivnl 3 to 6-in COWM 9'§w«l

Pip* column for QlMt-OVW

S-in. with open jam** S€CTION A-A

Figure 1. Typical sand drying bed.

As with other sludge handling methods the design of sludge drying beds is based on experience or, if the sludge is available, by scale-up from laboratory or pilot tests. Most often it is necessary to rely on experience because sludges are not available in sufficient quantities for testing. For open beds rainfall must be taken into account. A rule of thumb is that about 57% of the rain is absorbed by the sludge and must later be evaporated, the remainder

Sludge Dewatering Operations

219

draining away. On the average, therefore, 57% of the rainfall must be added to the amount of water requiring evaporation.

Use of Vacuum Filtration Vacuum filters have been used for decades in wastewater treatment for the dewatering of sludges. It is generally necessary to first digest sludge to increase filterability, although this is not always the case. Digestion does, however, reduce the odor problem, thus making the operation less objectionable than it might be otherwise. The filter operates as illustrated in Figure 2. The chemicals used to condition the sludge and, thus, make it easier to filter are combined with the feed sludge and dumped into a trough, which is situated underneath a large rotating drum. The drum is covered with a permeable fabric or other materials. A vacuum is drawn inside the drum, and water is sucked through the fabric into vacuum lines inside the drum and pumped out as the filtrate. The solids that cannot get through the fabric are caught on the surface of the drum and removed as the filter cake. Belt filters (where the fabric is lifted off the drum during cake discharge, as in Figure 2) tend to require greater chemical doses than the older drum filters, where the sludge is scraped off with a doctor blade. The higher doses are required due to problems with cake release. The objectives of vacuum filtration are to obtain a a. high solids concentration in the filler cake, b. clean filtrate, and an c. acceptable filter yield. The principle operational variable is the sludge to be filtered, since chemical conditioning can be considered a means of changing the sludge. The type of sludge filtered is within the control of the operator, and it is also within his province to dictate what happens to the sludge before it gels to the vacuum filter. The sludge solids concentration has a significant influence on the filter yield which is defined as the pounds of sludge obtained from the vacuum filter per square root of filtered area per hour. An increase in the teed sludge solids concentration usually results in a substantial increase in filter yield. The filter yield in terms of pounds per square foot per hour is often numerically equal to the percent of solids in the feed sludge. There is, a practical limit of about 8-10% because at greater solids concentrations, the sludge becomes difficult to pump. An increase in sludge solids concentration also tends to decrease the amount of chemical needed, hence making it doubly important for the treatment plant operator to try to introduce as concentrated a sludge as possible to the vacuum filter. The amount of time that the sludge spends out of the treatment process decreases the filterabilily of a sludge. For example, if aerobically digested sludge is being filtered, removing the sludge from the aeration tank for any length of time tends to decrease sludge filterability. Similarly, digested sludge tends not to filter as well if the sludge

220

o

Liquid Filtration

=

-5

a

•» «3 »•» «*

5 s

I

1 I

Sludge D ewatering Operations

2 21

is first taken out of the digester and allowed to cool. The action of chemical conditioners also deteriorates with time. Dram submergence can be adjusted or the depth at winch the drum is set inside the trough that contains the sludge can be controlled. If the drum is dropped lower into the sludge, a greater percentage of the cycle time is devoted to the pickup of solids from the trough and, thus, thicker but wetter filter cakes result. Decreasing the drum submergence tends to decrease the amount of time that the wet sludge is in contact with the filter, and, thus, produce thinner but drier cakes. By increasing the speed of the drum, the length of time the sludge is in contact with the filter is decreased as is the time for drainage once die sludge has left the pool, thus obtaining a wetter cake. However, this tends to increase the sludge yield. Correspondingly, a slower drum speed tends to result in a drier cake and a lower filter yield. The sludge in the trough must be agitated to prevent solids from settling and the extent of this agitation can be controlled. If the agitation is too violent, the floes will disintegrate and poor filtration will result. If the agitation is insufficient, settling will occur resulting in operational problems. One of the most important considerations is the selection of the proper filter medium. The available media can be categorized as open or tight. Open media have large pores while tight media have small openings. Although a tight filter will remove a higher percentage of the fines, a medium can be so tight as to make filtration impossible, and thus, blind, or stop up completely, preventing further filtration. The types of media in use were described in an earlier chapter. A method of improving filter efficiency is to use compounds that can be mixed with the sludge to improve its filterability. Filter aids were also described earlier. The disadvantage of filter aids is that the quantity of sludge will increase. The advisability of using filter aids is strongly dependent on the local cost, as well as the increase in filtration efficiency. Precoat filters are widely used in the chemical process industry. In this operation, instead of mixing the filter aid with the sludge, the filter aid is placed onto the filter fabric before the sludge is introduced. This prevents rapid premature clogging of the smaller pores and allows for a greater filter yield. Diatomaceous earth is an effective but expensive filter aid for pre-coat filters. In dewateriug digested sludge, vacuum filters are typically able to form cakes of between 20% - 40%. The filtrate quality can vary anywhere from 100-20,000 mg/1 of solids, corresponding to solids recoveries within ranges between 50 - 99%. The filtrate is almost always returned to the head of the plant for processing because it contains significant BOD (biological oxygen demand) and, thus, cannot be discharged with the plant effluent.

222

Liquid Filtration

Waste activated sludge, unless mixed with primary sludges, is not normally dewatered on vacuum filters. An exception to this is pure oxygen sludges that can be filtered successfully with the use of ferric chloride conditioning in a filter leaf. Vacuum filtration is an old established technology. Although for the most part it is a successful technique, it does have several disadvantages. Some of these are high operating costs, and from the cost of chemical conditioners. Many treatment plants have had problems with blinding to the point where vacuum filtration was no longer possible. Another problem associated with vacuum filtration is that the sludge lias a strong odor. The sludge in vacuum filtration is always open to the atmosphere and often warm which are two conditions producing die most severe odor problems.

Use of Pressure Filtration Pressure filtration differs from vacuum filtration in that the liquid is forced through the filter medium by a positive pressure instead of a vacuum. Among the most widely used in the chemical process industry (and widely used in Europe for wastewater treatment) is the filter press. As shown in Figure 3, the filter press operates by pumping the sludge between plates that are covered with a filter cloth. The liquid seeps through the filter cloth leaving the solids behind between the plates. When the spaces between the plates are filled, the treatment plant operator separates the plates and removes the solids. UOVASLC MUD

CLEAR FILTRATE OUTLET

MATERIAL ENTERS JflP™^ UNOCft PftCSCUIIC lli***—

Figure 3. Schematic of a filter press machine.

Filter pressing is a cyclic operation. Different designs enable automatic cake removal. One method is to blow the cake off by sending compressed air through the filtrate tubes. Another method is to make the filter cloth a moving belt. After each operation, the belt is shifted to allow the cake to drop off on either side.

Sludge Dewateriiig Operations

223

Use of Centrifugation The solid bowl centrifuge (or the decanter) is the type of machine used in wastewater treatment. This machine has the attribute of being able to dewater or at least separate out any solid from any liquid, as long as the solids are heavier. It is possible to use a centrifuge for many purposes in a wastewater treatment plant, for example, thickening activated sludge and when this is not needed, for the dewatering of digested sludge. The conventional solid bowl centrifuge is shown schematically illustrated in Figure 4. It consists of an outer bullet-shaped bowl that rotates at high speed. The sludge is pumped through a central pipe into this rotating bowl and because of centrifugal force, this sludge hugs the inside walls of the bowl. The heavier solids will sink to the bottom (that is, to the inner bowl wall) and the lighter liquid will remain pooled on top. The bowl acts as a highly effective settling tank, and nothing more. It is necessary to remove the sludge to make the operation successful and in the solid bowl centrifuge, (his process is accomplished by the scroll, or screw conveyor, that is placed inside the machine and rotates only slightly slower than the bowl. This screw action tends to convey the solids up onto the inclined beach and out the open end. The centrate, or clear liquid, flows out the holes on the other end of the bowl. The objectives of centrifugation are similar to those of other dewatering devices. It is necessary to obtain a dry cake, a clear centrate, and a reasonable throughput or, in the language of filtration, a centrifuge yield. The variables involved in centrifugation are listed in Table 3. These may be classified as machine variables or operational variables, as before. The bowl diameter is a machine variable controlled by the design engineer when the unit is purchased. Increasing the bowl diameter (and maintaining the same centrifugal force, that is, slowing down the machine) will result in a longer retention time within the machine. This results in a higher solids recovery, much as it would in a settling tank, but this solids recovery is at the expense of cake dryness. As the solids recovery is increased, the smaller particles will also escape as cake, increasing the moisture content of the cake and decreasing the solids concentration. Increasing the bowl length will also increase the residence time which will generally result in a high solids recovery, but the changes in cake dryness are not always predictable. It is also possible under some circumstances, to also increase the cake solids by increasing the bowl length. Bowl speed is one variable that can be designed into the machine simply by selecting the proper gear ratio. Increasing the bowl speed increases the centrifugal force and the solids recovery, with possibly a concurrent increase in the cake solids as well. This will occur only if solids recovery already approximates 100%, that is, when almost aU the solids are driven out the cake end. An increase in centrifugal force at that point will then increase the cake solids concentration, but this increase is often marginal, depending on the sludge.

224

Liquid Filtration

Table 3. Variables affecting centrifuge performance.

Machine Variables Bowl diameter Bowl length Bowl rotational speed Beach angle Beach length Pool depth Scroll rotational speed Scroll pitch Feed point of sludge Feed point of chemicals Condition of scroll blades

Operational Variables Residence time Sludge characteristics (including sludge conditioning)

MAIN SCREW

EFFLUENT CAKE Figure 4. Two solid bowl centrifuge

configurations.

Sludge Dewatering Operations

225

The pool depth is a variable that can be varied while die machine is in operation. Some machines require the removal of plugs at the end of the bowl to increase or decrease the pool depth. An increase in pool depth will result in higher retention time, hence better solids recovery and a wetter cake. Increasing the conveyor speed will force the solids out of the machine more quickly, thus leaving behind some of the wetter solids and increasing the solids concentration in the cake while decreasing solids recovery. Similarly, the conveyor pitch will influence solids recovery and cake dryness. If the scroll pitch is increased, the solids will be moved out faster but only the larger, heavier solids will he pushed out leaving the wetter solids in the centrate. Increasing the number of leads in the conveyor will likewise increase cake dryness at the expense of solids recovery. It is also possible to change the feed point of the sludge within the machine. If the feed point is changed toward the beach, a wetter cake will often result while solids recovery will increase; this is because the sludge will have a longer travel time to die end of the machine where the centrate exists. A drawback of centrifugation is that the conveyor will be subjected to quite severe wear because of its grinding against the sludge, especially if die sludge contains sancl or other gritty materials. A worn conveyor will result in a greater cake dryness since only the heavier particles are moved out, but this is, of course, at the expense of solids recovery, The bowl angle, or the angle at which the conical section comes off the cylinder, also has a strong influence on both cake dryness and solids recovery if the sludge to be centriraged comprises soft, fluffy materials. An increase in the angle will prevent some of the softer, fluffier solids from making their way up the beach, and these solids are hence discharged in the centrate, thus, decreasing solids recovery and increasing the cake solids concentration. The sludge on the inclined beach lias, in addition to the force directed radially outward, a force component called slippage force that pushes die sludge back into the cylindrical section of the bowl. If the solids still can flow, they will move under the conveyor blades, back to the cylinder, and not be expelled as cake. The critical point of solids slip is reached when the solids emerge from the pool. Because of this problem, machines that are designed to handle soft sludges have small beach angles, or the machines operate with the pool level raised above the outlet. This method, called super pool, utilizes hydraulic force to help the solids out. The centrifuge operates because two basic processes occur within its bowl. One is clarification, or the settling of solids from the mother liquid. The second is die successful movement of these solids out of the bowl. If either process is not performed successfully, the centrifuge will not work. It is necessary then to measure in the laboratory test how a sludge will clarify as well as to estimate how well a sludge will be moved out of the bowl.

226

Liquid Filtration

Centrifugation has many advantages over other dewatering methods. In addition to lower capital cost, a reasonable relative operating cost can be achieved, especially if labor costs are taken into account. A centrifuge is an enclosed unit and is, thus, odor free and can be used for dewatering such materials as heat treated sludges, which are obnoxious. Centrifuges can be accommodated in small buildings, or for that matter, outside. Chemicals are not necessary for operation but are generally recommended for continuous good operation and to avoid the buildup of fine solids in a treatment plait. A centrifuge is a flexible unit and is easy to start and clean up by running clean water through the machine. The disadvantages of centrifuges include the following. When chemicals are needed for operation they can be costly. Trash can become embedded in the sludge and clogg the machine very badly, causing many hours of downtime. Maintenance, especially with the wear and tear on the conveyor, can also lead to substantial maintenance costs.

Alternative Mechanical Dewatering Techniques Other centrifugal type devices can be used for sludge diickening and/or dewatering. The cyclone consists of a cone into which the sludge is pumped tangentially as illustrated in Figure 5. The heavy sludge solids tend to move to the inside wall, down, and drop out die bottom while the clean liquid moves up and out the center. Cyclones are quite useful to recover heavy materials and have found a use in degritting of sludge prior to centrifugation. The disc centrifuge machine has also been applied to thicken activated sludge. The disc machine, shown in Figure 6A has the advantage of being compact and efficient in its solids recovery of slow solids feeds due to high hydraulic capacity. The solids are settled outwardly, and the lighter solids separated by settling against the underside of the discs, gaining in density, and also sliding outward. The clarified liquid flows out through the disc stack. The solids are removed through nozzles in the bowl. These nozzles can be so constructed to open intermittently, expel the solids, and close before the liquid can escape, or a portion of die solids can be recycled through the machine, thus, maintaining a high solids inventory. The latter arrangement results in better solids control as well as permitting the use of larger nozzles, thus, reducing the chance of clogging. The disadvantage of the disc machine lies in the clogging problem. Trash in the sludge and unexpected power failures can cause clogged internals, and cleaning the machine is a tedious process. The stack must be disassembled one disc at a time. Another centrifuge of some interest in wastewater treatment is the basket centrifuge machine shown in Figure 6B. The basket is simply a vertical rotating bowl without a conveyor. Periodically, as the solids collect on the sides of the bowl, a scoop comes by and scrapes out die solids as the cake.

Sludge Dewatering Operations

SAND REMOVAL TANK

l

',M

SANDY WATER

Figure 5. Cyclones used for degritting operations.

B

Disk Centrifuge

Basket

Centrifuge

Figure 6. Disc and basket centrifuge basket systems.

Suggested Readings 1.

2.

Francingues, N.R., M.R. Palermo, C.R. Lee, andR.K. Peddicord. Management Strategy for Disposal of Dredged Material: Contaminant Testing and Controls. Miscellaneous Paper D-85-1, USAGE Waterways Experiment Station. Vicksburg, MS, 1985. Averett, D.E., B.D. Perry, and EJ. Torrey. Review of Removal, Containment and Treatment Technologies for Remediation of Contaminated Sediments in the

228

3. 4.

5.

6.

7.

8.

9.

10.

11.

Liquid Filtration

Great Lakes. Miscellaneous Paper EL90-25, U.S. Army Corps of Engineers, Vicksburg, MS, 1990. Fairweatlier, V. The Dredging Dilemma. Civil Engineering, August, pp. 4043, (1990). Palermo, M.R. et al. Evaluation of Dredged Material Disposal Alternatives for U.S. Navy Homeport at Everett, Washington. EL-89-1, U.S. Army Corps of Engineers, Vicksburg, MS, January, 1989. Cullinane, M.J., Jr., D.E. Averett, R.A. Sharer, J.W. Male, C.L. Truitt, and M.R. Bradbury. Alternatives for Control/Treatment of Contaminated Dredged Material. In: Contaminated Marine Sediments — Assessment and Remediation, National Academy Press, Washington, D.C., pp. 221-238, (1989). Truitt, C.L. Engineering Considerations for Capping Subaqueous Dredged Material Deposits - Background and Preliminary Planning: Environmental Effects of Dredging, Technical Notes. EEDP-09-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, 1987. Truitt, C.L. Engineering Considerations for Capping Subaqueous Dredged Material Deposits - Design Concepts and Placement Teclniiques: Environmental Effects of Dredging, Technical Notes. EEDP-09-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, 1987. Palermo, M.R. Design Requirements for Capping, Environmental Effects of Dredging, Technical Notes. In preparation, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, 1991. Culiinane, M.J., D.E. Averett, R.A. Schafer, J.W. Male, C.L. Truitt, and M. R. Bradbury. Guidelines for Selecting Control and Treatment Options for Contaminated Dredged Material Requiring Restrictions. Final Report, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS, September, 1986. Sukol, R.B. and G.D. McNelly. Workshop on Innovative Technologies for Treatment of Contaminated Sediments, June 13-14, 1990. EPA/600/2-900/054, U.S. Environmental Protection Agency, Cincinnati, OH, 1990. Miller, J.A. Confined Disposal Facilities on the Great Lakes. U.S. Army Corps of Engineers, Chicago, IL, February, 1990.

8 INDUSTRIAL WASTEWATER SOURCES Introduction General discussions on the applications of filtration to wastewater flows were given in previous chapters. These discussions have provided the reader with an understanding of the objectives of wastewater treatment and how filtration can be used as one of the unit operations for solids removal. Filtration must, however, be used in conjunction with other treatment technologies or operations, and does not necessarily represent the primary treatment method. It is beyond the scope of this volume to discuss the other techniques used in the treatment of wastewater flows. For general information and background on other unit operations and processes used with filtration in the treatment of waste waters, the reader may refer to the suggested references at the end of this chapter. The decision as to what type of filtration equipment to use, and indeed the decision on whether filtration is even applicable in the treatment of a particular waste stream, depends on the characteristics of the stream and the requirements for treatment which, in fact, are often legal standards and specific to the wastes. This in part requires knowledge of the pollutants generated during a manufacturing operation. Legal standards for wastewater discharges exist in many countries. For the purposes of discussions in this chapter and for the newcomer to this subject, these will be referred to as priority pollutants. They include biochemical oxygen demand (BOD), suspended solids (SS), total dissolved solids (TSD), pH, fecal coliform bacteria, chemical oxygen demand (COD), total organic carbon (TOC), nitrogen and nitrogen compounds. There may also be concern for such pollutants in wastewater streams as fats, oils and greases of animal or vegetable origin because these could interfere with the operations of certain treatment works. This chapter provides an overview of the characteristics of typical waste streams from major industry sectors. A review of the discussions in this chapter will provide the reader with the understanding of the sources of these wastes from each industry sector, 229

230

Liquid Filtration

their general characteristics and types of pollutants, and when filtration is applicable. In most of the cases presented, filtration will play a major role in the pretreatment of waste streams and also, to a lesser degree, as a finishing stage for final solids removal. Industry terminology typically refers to this as polishing. Paper and Allied Products Industry Wastes Industry Description This industry sector includes the manufacture of pulp from wood, rag, and other cellulose fibers and the manufacture of paper, paperboard, and building products. The manufacture of converted paper and paperboard products from purchased paper is not included since it involves a relatively simple process, whose wastewater flows and loadings are not generally significant to the design of treatment systems. Therefore, the plants making converted paper and paperboard products are excluded from discussions. The manufacture of paper and allied products involves the preparation of wood and other raw materials, separation and recovery of cellulose fibers, and blending of the fibers with proper additives to produce "furnish", which is formed into paper. The additives include: sizing materials such as alum and resins, sodium aluminate, and wax emulsions; synthetics, such as acrylics, isocyanates, and fluocarbons; and fillers such as clays, calcium carbonate and sulfate, talc, barium sulfate, aluminum compounds, and titanium oxide. When fillers are used, retention aids (starches or synthetic resins) are added to increase the retention of the filler. The principal operations involved in the manufacture of pulp and paper are: Wood Preparation Pulping (mechanical, chemical, semi-chemical, and deinking) Pulp Washing and Screening Stock Preparation Papermaking The pretreatment sub-groups for this industry areas follows: •

Integrated pulp and paper mills using mechanical pulping processes (bleached and unbleached) Integrated pulp and paper mills using chemical pulping processes (unbleached) Integrated pulp and paper mills using chemical pulping processes (bleached) Integrated pulp and paper mills using deinked pulp Paper and paperboard mills Building products mills

The following industrial practices can significantly influence pretreatment:

Industrial Wastewater Sources

231

Pulping Process — The pulping process determines the pulp yield and quality, and the probable organics loss in the wastewater from a pulp mill. (Mechanical pulping results in minimum dissolution of wood components, while chemical pulping solubilizes the non-cellulose components of wood tannins, lignins, wood sugars, and hemicellulose). Recovery and Reuse of Spent Cooking Liquor — Most chemical pulping processes involve recovery and reuse of spent cooking liquor and, therefore, do not generate significant quantities of wastewater. The dissolved organics present in the liquor usually are burned in the chemical recovery furnace. The only pulping process where the spent cooking liquor is not suitable for recovery is the calcium-base acid sulfite process. The spent cooking liquor from this process, with the dissolved organics, is generally discharged with other process wastes. Bleaching -— When the desired quality of the final product requires bleaching of pulp recovered from wood, it is usually done by the addition of oxidizing chemicals, such as chlorine, chlorine compounds, peroxides, and hydrosulfites. The oxidizing chemicals react with the non-cellulose portion of the pulp, rendering it soluble in water or in alkaline solutions. As a result, the bleaching step adds to the wastewater volume and pollutant loading. Plant Pollution Control Methods — Control of spills and leaks, and recovery and reuse of chemicals constitute the major pollution control practices within the paper and allied products industry. The extent of pretreatment required is largely dependent on the extent and effectiveness of the in-plant control processes adopted,

Wastewater Characteristics The characteristics of the process wastewaters from each pretreatment sub-groups are listed in Table 1. Integrated pulp and paper mills generally operate continuously throughout the year except for the shutdowns for preventive maintenance and equipment repair and replacement. Modem practice is to employ continuous pulping processes, however, some mills around the world are still using batch pulping processes which result in intermittent discharges of wastewater. The overall wastewater characteristics from wood pulping processes may vary seasonally because of the changes in characteristics of wood and variations in the temperature of the water. The volume and characteristics of the process wastewater depend upon the degree of water reuse, chemical recovery systems, and the type and quality of paper involved. The wastewaters generated from the paper and allied products industry contain BOD, COD, suspended solids, dissolved solids, color, acidity or alkalinity, and heat. Chemical pulping processes may produce wastewaters with heavy metals (Cr, Ni, Hg, Pb, Zn). If pulp bleaching is part of the operation, the wastewaters may contain additional heavy metals (Hg) and dissolved solids (chlorides). Mercury may be present in the caustic used in pulping and bleaching operations. Zinc is used in the bleaching of ground wood pulp. Chromium, nickel, and iron may be introduced from the corrosion of process equipment.

232

Liquid Filtration

When spent cooking liquor recovery is not practiced, the wastewaters may be acidic (pH in the range of 2 - 3) and have high concentrations of dissolved organics and inorganics. The solubilized organics and the type of cooking liquor will determine the characteristics of the wastewater. Pretreatment Operations The pretreatment unit operations which may be necessary for various types of joint treatment processes are summarized in Table 2. The neutralization requirements depend primarily on the pulping process and on the related recovery and reuse practices. When spent cooking liquor is discharged to municipal sewers, special care must be taken to insure control of pH, organic shock loads, and color (if present). If bleaching is part of the operation, pH adjustment may be necessary before mixing with other wastes. Spills of spent liquors and pulp washing water may introduce shock loads of pH and organic material if their discharge into the municipal sewer system is not carefully controlled. In the absence of effective in-plant control procedures, adequate equalization and extensive pH and conductivity control may be required to protect the operation of a joint treatment facility such as POTW which handle both municipal and industrial wastes from paper mills. The heavy metal concentrations in the paper and allied products industry wastewaters are very low and generally do not require pretreatment. Nevertheless, their levels should be determined to insure that effluent limitations are not exceeded where heavy metals are a significant factor. Dairy Products Industry Wastes Industry Description This industry sector includes bulk handling, packaging, and processing (pasteurizing, homogenizing, and vitaminizing) of milk, and the manufacture of dairy products including butter, cheese, ice cream, condensed evaporated milk, and dry milk and whey. The manufacture of dairy products involves receiving and storing raw milk, separation of excess cream, pasteurization and homogenization, fluid milk packaging, and making butter, ice cream, and cheese. In the separation step, excess cream may be skimmed off in order to standardize the butter fat content, or the raw milk may be separated by centrifuge into cream and skim milk. Separated cream is then used in butter or ice cream making, while the skim milk may be used in the production of cottage cheese and non-fat dry milk solids. Natural cheese (i.e., not cottage cheese) is made with whole rnilk. Some of these processes generate by-products which may be recovered and utilized in other food manufacturing operations. Buttermilk, skim milk, and whey are produced from butter and cheese making. The regional market potential for these byproducts often determines the amount of in-plant recovery.

Table 1. Wastewater Characteristics Paper and Allied Products. Characteristics

Chemical Pulping (Bleached)

Mechanical Pulping (Bleached & Unbleached) Year-round

Chemical Pulping (Unbleached) Year-round

Year-round

Year-roimd

Year-round

Year-round

BOD TSS IDS COD

Continuous Extremely High High t Average High

Continuous Average-Ext, High Low-High High High

Continuous Average-Ext. High Low-High High High

Continuous High High High High

Continuous Average-High Average-High Low-average High

Continuous Extremely High Extremely High

Grit Cyanide Chlorine Demand pH Color

Present Absent High Neutral

Present Absent High Acidic High

Present Absent High Alkaline

Low

Present Absent High Acid-alkaline High

Present Absent High Neutral

Present Absent High Neutral

Low

Low

Low

Turbidity Explosives Dissolved Gases Detergents Foaming

High Absent Absent Absent Present

High Absent Present Absent Present

High Absent Present Absent Present

High Absent Absent Absent Present

High Absent Absent Absent Present

Very High Absent Absent Absent Present

Heavy Metals Colloidal Solids Volatile Organics Pesticides

Absent Present Absent4 Absent

Present Present Presen^ Absent

Present Pesent Pesent4 Absent

Present Present Absent Absent

Absent Present Absent Absent

Absent Present Absent Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Deficient Deficient High Absent Absent

Deficient Deficient High Absent Absent

Deficient Deficient High Absent Absent

Deficient Deficient High Absent Absent

Deficient Deficient High Absent Absent

Deficient Deficient High Absent Absent

Oil & Grease Coliform (Total)

Present Average

Present Average

Present Average

Present Average

Present Average

Present Average

Industry Operation Flow

High if bleaching of pulp is practiced. ~ Acidic if bleaching of pulp is practiced. Present if bleaching of pulp is practiced. Present only from log washing operations.

Deinking Pulp

Building Products

Paper and Paperboard

Low High

Table 2. Pretreatment Unit Operations for the Paper and AlUed Products Industry. Suspended Biological System Mechanical Pulping (Unbleached)

Coarse Solids Separation + Grit Removal

Mechanical Pulping (Bleached)

Coarse Solids Separation 4- Grit Removal + Neutralization

Chemical Pulping (Unbleached)

Coarse Solids Separation + Grit Removal + Neutralization

Chemical Pulping (Bleached)

Coarse Solids Separation +Grit Removal + Neutralization

Deinking Pulp 1

Coarse Solids Separation + Neutralization

Paper & Paper Board

Coarse Solids Separation

Building Products

Coarse Solids Separation

Fixed Biological System

Independent Physical-Chemical System

Where pulp and paper waste waters constitute more than about 10 % of the total wastewater flow, fixed biological systems or independent physical chemical systems normally are not used. Where the pulp and paper wastewater constitute less than this proportion, the pretreatment requirements will be the same as for suspended biological systems.

Equalization may be required in addition to those shown when batch pulping processes are used.

& SX

Industrial Wastewater Sources

235

The dairy industry is a year-round operation. Raw milk receiving stations handle milk from the local farms for subsequent transfer to tank trucks. These stations are operated on an intermittent daily basis. The rest of the industry will operate either continuously or on an intermittent basis depending on the economics of the individual facility. In general, larger processing plants tend to be integrated plants producing more than one product. The pretreatment sub-groups for this industry are as follows: • •

Cottage and Natural Cheese Products Milk Handling and Products.

Product recovery is the major method for reducing waste water loadings. The dairy industry, however, must maintain sanitary conditions and this tends to limit the amount of wastewater recycle which is practicable. The following industrial practices can have a major impact on the wastewater characteristics: Whey Handling — Whey can be condensed and dried, and used as a food and feed supplement. However, there are inherent difficulties in drying the acid whey derived from cottage cheese because of its lactic acid content. Operating Procedures — Spillage, overflow, and leakage caused by improperly maintained equipment and poor operating procedures can result in major pollutional loads due to the concentrated nature of the dairy products, e.g., whole milk has a BOD over 100,000 mg/L (0.8 pounds of BOD for every gallon of milk lost). Cleaning — High efficiency in sanitizing operations is important to minimize the usage of sanitizers and detergents. In addition, rinses can be collected and used as make-up water for sanitizers. Milk processing lines should be sloped to central collection points so that the milk product may be collected before the lines are sanitized. Wastewater Characteristics The characteristics of the process wastewaters from each pretreatment sub-group are shown in Table 3. Daily wastewater flows are characterized as intermittent because some major unit processes, e.g., cheese and butter making, are batch, and because milk processing equipment must be shut down daily for sanitizing to maintain rigid health standards. Relatively clean water may be a substantial portion of the total wastewater from a dairy plant. These waters are from condensers, refrigeration compressors, milk coolers, and air conditioning systems. The major types of wastewaters from the dairy industry are: 1.

Wash and rinse water from cans, tank trucks, equipment, and floors. In general, the pH of the wastewater will be affected by the cleaning compound

236

Liquid Filtration

(either acid or alkali) and the proportion of the wash water in the total plant waste water flow. By-products (such as buttermilk, skim milk, or whey) are sewered rather than recovered. Buttermilk and skim milk have BOD values as high as 70,000 mg/L, while the BOD of whey is 50,000 mg/L. Entrainment from evaporators and the sewering of spoiled or damaged products.

2.

3.

Table 3. Wastewater Characteristics Dairy Products. Milk Handling Milk Products

Natural and Cottage Cheese Product

Year-round (Batch) Intermittent Average-High Low-Average Average-High

Year-round (Batch) Intermittent Extremely High Average-Ext. High High

pH

Average-High Present Absent High Acid to Alkaline

Extremely High Present Absent High Acid to Alkaline

Color Turbidity Explosives Dissolved Gases Detergents

High High Absent Absent Present

High High Absent Absent Present

Foaming Heavy Metals Colloidal Solids Volatile Organics Pesticides

Present Absent High Absent Absent

Present Absent High Absent Absent

Phosphorus Nitrogen Temperature Phenol Sulftdes

Present Adequate Normal-High Absent Absent

Present Deficient ? Normal-High" Absent Absent

Oil & Grease Coli form

Present Present

Present Present

Characteristics Industrial Operation FLOW

BOD TSS IDS COD Grit Cyanide Chlorine Demand

'

There are possible bio-static effects in the joint treatment plant attributable to large amounts of sanitizers and detergents in the dairy products wastewater. Temperature equal to or higher than domestic wastewater, may affect design but not harmful to joint treatment processes.

The wastewaters generated in the dairy industry can be characterized generally as containing high concentrations of BOD, COD, and TDS. Settleable solids are not an important consideration in most dairy wastewaters, since all the organic material is in a colloidal or dissolved state. However, filtration practices are still important because sand or other gritty material may be present from tank truck washings. Cheese wastewaters, on the other hand, have higher concentrations of settleable solids due to the presence of curd solids.

Industrial Wastewater Sources

237

Design considerations in the joint treatment of dairy and domestic wastewaters are the high chlorine demand and the presence of surface-active agents. In addition, cheese production wastewaters are usually nitrogen deficient. Septicity should be a consideration in the design of any equalization or clarification facilities. Low-rate trickling filters are generally not effective in treating dairy wastes, because these wastes produce large quantities of biological solids which clog the filters. This problem can be overcome by high hydraulic loading and high recirculation rates.

Pretreatment The pretreatment unit operations which may be necessary for various types of treatment facilities are shown in Table 4. In general, dairy wastes are amenable to biological as well as to chemical treatment if equalization and neutralization are provided as pretreatment for the dairy wastes. Where whey is present, it may be necessary to control the rate of discharge and to provide nutrient supplements at the treatment plant. Table 4. Pretreatment Unit Operations for the Dairy Products Industry. Pretreatment S ub-Group

Suspended Biological System

Fixed Biological System

Independent PhysicalChemical System

Milk Handling Milk Products

Equalization + Neutralization

Equalization -f Neutralization

Equalization + Neutralization

Natural and Cottage Cheese Products

Equalization + Neutralization

Equalization + Neutralization

Equalization + Neutralization

Textile Industry Wastes Industry Description The textile industry involves the manufacture of fabrics from wool, cotton, and synthetic fibers; the synthesis or spinning of synthetic fibers is not included in this group, but rather is included under synthetic organic chemicals. Of the three major textiles, wool represents the smallest market and synthetic textiles the largest. The major unit processes of the woolen textile industry include scouring, dyeing, fulling, carbonizing, bleaching, and weaving. Raw wool is scoured to remove grease and dirt. The process employs a detergent and mild alkali at temperatures of 130°F. This operation is responsible for 55-75 % of the total BOD load from wool finishing. The dyeing process uses dyes, dyeing assistants, (e.g. acetic acid, ammonium sulfate), and dye carriers containing heavy metals. The dye carriers will be present only if the wool is being combined with a synthetic fabric, which requires a dye carrier to facilitate dye penetration. Various chemicals (e.g. sulfuric acid, hydrogen peroxide, and olive oil) may be added before and during the fulling operation. These chemicals

238

Liquid Filtration

then enter the wastewater during a subsequent washing step. Carbonizing impregnates the wool with sulfuric acid to remove any traces of vegetable matter. Bleaching may then be accomplished, with either sulfur dioxide, hydrogen peroxide, or optical brighteners. The major unit processes employed in the cotton and synthetic textile industry include sizing, weaving, desizing, scouring, dyeing, and finishing. Chemicals used in the sizing process include starch, polyvinyl alcohol, carboxymethyl cellulose, and polyacrylic acid. After weaving, the fabric is desized using an acid enzyme reaction. Desizing removes the chemicals added during sizing by hydrolyzing them to a soluble form. During scouring cotton wax and other non-cellulosic components of cotton are removed by using hot alkaline solutions. Synthetic materials require only light scouring because of the absence of chemical impurities. Both cotton and synthetic fabrics are treated with special finishes, using formaldehyde and urea, and with fire retardants, such as triaziridyl phosphine oxide. The pretreatment sub-groups for this industry are: • •

Wool Cotton and Synthetic Fabrics

The following industrial practices characteristics:

can significantly affect

the wastewater

Segregation of Waste Streams — The segregation of waste streams permits recovery of heavy metals, caustic recovery and reuse, and control of toxic spills (such as dieldrin used for moth-proofing). Many of the older textile mills have a common collection system with chemical reuse, but the modern mills have a segregated collection system to permit chemical recovery and reuse. Alkaline Wool Scouring — Alkaline wool scouring may be used in place of neutral scouring. In alkaline scouring, soda ash is added to the wash water and subsequently combines with a portion of the wool grease to form natural soap. This procedure reduces the amount of detergent required and reduces the BOD and the concentration of residual surface-active agents. Chemical Sizing — The substitution of polyvinyl alcohol or carboxy methylcellulose for starch in the sizing of cotton reduces the overall COD in the wastewater. Pressure Dyeing Decks — The use of pressure dyeing decks in the place of atmospheric units permits reduction in the amount of dye carriers required, thereby, reducing the BOD and heavy metal concentrat ions.

Industrial Wastewater Sources

239

Wastewater Characteristics The characteristics of the process wastewaters from each pretreatment sub-group are shown in Table 5. The textile industry generally operates continuously throughout the year except for scheduled shutdowns. However, many of the individual unit operations within the industry are batch oriented. Daily waste water flows are continuous with peak flows occurring if certain batch-type operations are used. The textile industry is usually in a state of flux, with the manufacturing process continually being modified to reflect changes in consumer trends. Table 5. Wastewater Characteristics of the Textile Industry. Wool

Characteristics

Cotton and Synthetics

Industrial Operation

Year-round (batch)

Year-round (batch)

Flow

Intermittent -Continuous High High High High

Intermittent -Continuous Average-High Low-Average High Average-High

Color

Present Absent High Basic High

Absent Absent High Basic High

Turbidity Explosives Dissolved Gases Detergents Foaming

High Absent Absent Present Present

High Absent Absent Present Present

Heavy Metals Colloidal Solids Volatile Organics Pesticides Phosphorus

Present Present Absent Absent Present

Present Present Absent Absent Present

Nitrogen Temperature Phenol Sulfides Oil & Grease Coliform (Fecal)

Deficient Normal-High 2 Absent 3 Absent High 5 Present

Deficient Normal-High 2 Absent 3 Absent Absent-Present Absent

BOD TSS IDS COD Grit Cyanide Chlorine Demand

pH

Wastewater flow characterized by an intermittent pattern over the day. Temperature equal to or higher than domestic wastewater. Hay affect design but not harmful to joint treatment processes. Phenol may be present in dye carriers. Oil present in wastewaters from synthetic textiles only. Wool processing wastewaters contain high concentration of animal grease.

Wastewaters generated in the textile industry are high in BOD, COD, TDS, and color. For synthetics, dyeing results in the largest BOD contribution, attributed to the use of dye carriers such as methyl naphthalene, biphenyl, and orthophenyl phenol, all of

240

Liquid Filtration

which have a high BOD. For cotton finishing, desizing contributes about 45 % of the total BOD, scouring about 30 %, and dyeing about 17%. The most significant difference between wool process wastewaters and those from the rest of the industry is the release of wastewaters with high concentrations of suspended solids and grease from the wool-scouring operation. Textile wastewaters can vary from slightly acid to highly alkaline depending on the individual processes carried out within the plant. They generally are alkaline when caustic scouring or mercerizing is involved. Heavy metals such as copper, chromium and zinc result from the use of certain dye carriers in the dyeing operation of synthetic fabrics and of blended fabrics, e.g., cotton and rayon. The pretreatment unit operations which may be necessary for various types of joint treatment facilities are listed in Table 6. Table 6. Pretreatment Unit Operations for the Textile Industry. Suspended Biological System

Fixed Biological System

Independent PhysicalChemical System

Wool

Coarse Solids Separation + Grease Removal + Chemical Precipitation (color, heavy metals) + Equalization + Neutralization

Coarse Solids Separation + Grease Removal 4Chemical Precipitation (color, heavy metals) + Equalization + Neutralization

Coarse Solids Separation + Grease Removal + Chemical Precipitation (color, heavy metals) + Equalization + Neutralization

Cotton & Synthetics

Coarse Solids Separation + Chemical Precipitation (color, heavy metals) + Equalization -tNeutralization

Coarse Solids Separation + Chemical Precipitation (color, heavy metals) 4Equalization + Neutralization

Coarse Solids Separation 4Chemical Precipitation (color, heavy metals) + Equalization + Neutralization

Pretreatment Sub-Group

Pharmaceutical Industry Wastes industry Description This industry produces medicinal chemicals and pharmaceutical products, including some fine chemicals which are marketed outside the pharmaceutical industry as intermediates. In general, the pharmaceutical industry may be divided into two broad production categories: chemical synthesis products and antibiotics (penicillin and steroids are examples). The manufacturing operations for synthesis products may be either dry or wet. Dry production involves dry mixing, tableting or capsuling, and packaging. Process equipment is generally vacuum cleaned to remove dry solids and then washed down. The production of wet synthesis products and antibiotics is very similar to fine chemicals production, and uses the following major unit processes: reaction, extraction and concentration, separation, solvent recovery, and drying. Wet synthesis reactions generally are batch types followed by extraction of the product. Extraction of the pharmaceutical product is often accomplished through

Industrial Wastewater Sources

241

solvents. The product may then be washed, concentrated and filtered to the desired purity, dried, capsulized, and packaged. Some antibiotics are produced in batch fermentation tanks in the presence of a particular fungus or bacterium. The culture frequently is filtered from the medium and marketed in cake or liquid form as an animal feed supplement. The antibiotic is extracted from the culture medium through the use of solvents, activated carbon, etc. The antibiotic is then washed to remove residual impurities, concentrated, filtered, and packaged. The pretreatment sub-groups for this industry are as follows: • •

Synthesis Fermentation

The following industrial practices can significantly influence the wastewater characteristics: 1. Solvent recovery is practiced in both the synthesis and the fermentation products segment of the industry. Certain products may require a highpurity solvent in order to achieve the required extraction efficiency required. This increases the incentive for making the recovery process highly efficient. 2. Some solvent streams which cannot be recovered economically, are incinerated. Incineration is also used to dispose of "still bottoms" from solvent recovery units.

Wastewater Characteristics The characteristics of the process wastewaters from each pretreatment sub-group are shown in Table 7. The pharmaceutical plants operate continuously throughout the year and are characterized by batch and semi-batch operations with significant variations in pollutional characteristics over any typical operating period. The major sources of wastewaters are product washings, concentration and drying procedures, and equipment washdown. Wastewaters generated from the pharmaceutical industry can be characterized as containing high concentrations of BOD, COD, TSS, and volatile organics. Wastewaters from some wet chemical syntheses may contain heavy metals (Fe, Cu, Ni, V, Ag) or cyanide, and generally have anti-bacterial constituents, which may exert a toxic effect on biological waste treatment processes. Considerations significant to the design of treatment works are the highly variable BOD loadings, high chlorine demand, presence of surface-active agents, and the possibility of nutrient deficiency.

242

Liquid Filtration

Pretreatment The pretreatment unit operations which may be necessary for various types of treatment facilities are shown in Table 8. The specific type and degree of pretreatment for heavy metals and cyanide will be governed by the industrial effluent guidelines for the pharmaceutical industry. Cyanide removal or control is especially important. Pharmaceutical industries generate wastewaters on an intermittent basis and equalization may be needed as pretreatment. When solvents are used for extraction, solvent removal can be accomplished by using gravity separation and skimming. Neutralization may be required to neutralize acidic or alkaline wastewaters generated from the production of specific pharmaceutical products. Table 7. Wastewater Characteristics of the Pharmaceutical Industry. Characteristics

Synthesis

Fermentation

Industrial Operation Flow

Year-round (Batch) Intermittent Average-High High Average-High

Year-round (Batch) Intermittent Extremely High Extremely High High

pH

Average-High Absent Present Average-High Acid-Basic

Extremely High Absent Absent High Acid-Basic

Color Turbidity Explosives Dissolved Gases Detergents

Average-High Average Present Absent Present

Average-High High Present Absent Present

Foaming Heavy Metals Colloidal Solids Volatile Organics Pesticides

Present Present High High Absent

Present Absent High High Absent

Phosophorus Nitrogen Temperature Phenol Su If ides

Deficient Deficient Normal-High Absent Absent

Deficient-High Deficient-High Normal-High Absent Absent

Oil & Grease Colifonn (Total)

Absent-Present Absent

Absent-Present Absent-Present

BOD TSS TDS COD Grit Cyanide Chlorine Demand

Temperature equal to or higher than domestic wastewater. May affect design but not harmful to joint treatment processes.

Industrial Wastewater Sources

243

Table 8. Pretreatment Unit Operations for the Pharmaceutical Industry. Suspended Biological System

Fixed Biological System

Independent Physical-Chemical System

Synthesis

Chemical Precipitation (Heavy Metais) + Solvent Separation \- Neutralization + Cyanide Oxidation

Chemical Precipitation (Heavy Metals) + Solvent Separation + Equalization + Neutralization + Cyanide Oxidation

Chemical Precipitation (Heavy Metals) + Solvent Separation + Equalization + Neutralization + Cyanide Oxidation

Fermentation

Solvent Separation + Equalization + Neutralization

Solvent Separation + Equalization + Neutralization

Solvent Separation + Equalization + Neutralization

Pretreatment Sub-Group

Leather Tanning and Finishing Industry Wastes Industry Description This industry sector includes tanning, curing, and finishing hides and skins into leather. A major portion of the output from the tanneries is from cattle-hide processing, with the other sources being pigskin, calfskin, goatskin, and sheepskin processing. The tanning process involves conversion of animal hide and skins into leather. The grain layer and the cerium portion of the skins constitute the leathermaking material and consist mainly of the protein collagen. During the tanning process, the collagen fibers are reacted with tannin, chromium, alum, or other tanning agents to form the leather. Four basic operations are involved in tanneries: 1. Beam House 2. Tan House 3. Retan, color and fat liquor 4. Finishing The beam house operation involves; storage and trimming of hides; washing and soaking to remove dirt, salt, blood, manure and non-fibrous proteins; green fleshing for the removal of adipose fatty tissues and meat; unhairing to remove epidermis and hair; bating to remove non-collagenous proteins; and pickling in some operations to stabilize and preserve the unhaired stock for subsequent operations. The beam house operation is typical of hide and skin processing with cattle-hide processing being the most important in many parts of the world. The tan house operation consists of preparing the stock for tanning. Pickling is done to make the skin acid enough to prevent precipitation of chromium during tanning. Two types of tanning are common; vegetable tanning and chrome tanning. Vegetable tanning is carried out in a solution containing plant extracts (such as vegetable tannin)

244

Liquid Filtration

to produce heavy leathers such as sole leathers and saddle leathers. Light leathers, such as shoe upper leathers, are usually chrome-tanned by immersion in a bath containing proprietary mixtures of basic chromium sulfate. Hides which have not been fully tanned in the chrome-tanning process may be retanned with either chrome, vegetable, or synthetic tanning agents. The fat liquor process involves the addition of many types of oils and greases to the tanned hides to prevent cracking and to make the leather soft, pliable, strong, and resistant to tearing. Coloring or dyeing of tanned leather may be done either before or after fat liquoring and uses either natural or synthetic dyestuffs. Finishing operations such as drying, coating, staking, and plating follow the foregoing wet processes. The pretreatment sub-groups for this industry are chrome tanning and vegetable tanning. In-plant pollution control techniques and chemical recovery practices in tanneries vary depending on the tanning process and the economics of chemical recovery systems. In vegetable tanning, it is common practice to recycle the tanning solution. (In chrome tanning, tanneries usually practice recycling of the tanning solution.) Recovery of grease is normally practiced in pigskin and sheepskin tanneries. The wastewater characteristics from the unhairing process will depend on whether the industry is practicing a "save hair" or "pulp hair" operation. A low amount of sulfide removes the hair with minimal damage, while a high amount pulps and partially dissolves the hair. The "save hair" operation involves mechanical pulling and recovery of hair. Dissolution of hair through chemical reactions is referred to as "pulping" or "burning", Wastewater Characteristics The characteristics of the process wastewaters from each pretreatment sub-group are shown in Table 9. Most sub-processes within the tanneries are batch operated, and, therefore, the wastewater flow and characteristics fluctuate during the industry operation. In addition, weekend shutdowns in some tanneries will result in wastewater flow only during weekdays. The seasonal variations in wastewater flow are limited to the variations in hide characteristics. In general, the waste characteristics and volume vary widely throughout the day and throughout the week in tanneries. The concentrated waste fractions (lime liquors and spent tan solutions) are derived from batchtype processes. These fractions are therefore discharged intermittently. Liquid process wastes are generated in tanneries from soaking and washing, fleshing, unhairing, bating, pickling, tanning, coloring, and fat liquoring. Auxiliary wastewaters from tanneries result primarily from boiler blowdown and from cooling, and represent only a minor fraction of the total waste load from tanneries. Therefore, only process waste streams are considered for establishing wastewater characteristics and recommending pretreatment.

Industrial Wastewater Sources

245

Table 9, Wastewater Characeristics Leather Tanning and Finishing Industry. Chrome Tanning

Characteristics

Veqetable Tanning

Industry Operation Flow BOD TSS IDS

Year-round (Batch) Intermittent Extremely High Extremely High High

Year-round (Batch) Intermittent Extremely High Extremely High High

COD Grit Cyanide Chlorine Demand pH

Extremely High Present Absent High Acid- Alkaline

Extremely High Present Absent High Acid-Alkaline

Color Turbidity Explosives Dissolved Gases Detergents

Present Present Absent Present Present

Present Present Absent Present Present

Foaming Heavy Metals Colloidal Solids Volatile Organics Pesticides

Absent Present Present Present Absent

Absent Absent Present Present Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Deficient Adequate Normal1 Absent Present

Deficient Adequate Normal 1 Absent Present

Oils & Grease Conform (Total)

HighLow

HighLow

'-Temperature equal to domestic wastewater. •;-Oil and grease (animal origin) are significant only in pigskin and sheepskin processing wastewaters,

The characteristics of wastewaters from tanneries vary according to the type of hide processed and the tanning process (vegetable or chrome tanning) used. The tanning process is more of an art than science, and as a result the wastewater characteristics can vary widely for the same type of hide and tanning process. The process wastewaters from tanneries contain as major constituents: BOD, COD, chromium, oil and grease, sulfide, suspended and dissolved solids, alkalinity, hardness, color, and sodium chloride. Significant pollutants that may be present in tannery wastes include: hair, hide, scraps, bits of flesh, blood, manure, dirt, salts, suspended lime, soluble proteins, sulfites, sulfides, amines, chromium salts, vegetable tannin, soda ash, sugar and starches, oils, fats and grease, surface active agents, mineral acids, and dyes which contribute to the BOD and COD. In general, washing, fleshing, and unhairing operations produce more than half, 56 % of the total volume and approximately 70 % of the poUutional load from tanneries. The tanning process generates from 5 - 20 % of wastewater volume and loading. The

246

Liquid Filtration

dry finishing operations produce only minor quantities of wastewater from clean-up operations. The major wastewater sources from tanneries are the beam house and the tan house operations. The beam house wastewater is highly alkaline due to the large quantities of lime used in the process. The wastewater generated from the tan house is generally acidic due to the discharge of spent tanning solution. Normally, these wastewaters are discharged to a common collection sewer before treatment or discharged to a municipal system. The unregulated batch dumping from beam house and tan house result in a total waste stream with varying pH values. Mixing alkaline waste streams (from the beam house) with the acidic chrome tanning wastes would result in partial precipitation of chromium, which can be removed by clarificaftion. There is also the hazard of the evolution of hydrogen sulfide gas. The other major pollutant in tannery waste is the effluent from the lime-sulfide unhairing operation. The concentration of sulfides in tannery wastes may vary between 30 and 100 mg/L in the total effluent. The consequences of release of H2S gas in the sewer lines and the effect of reducing characteristics of the sulfides on biological treatment processes should be taken into consideration in the design of treatment works. Hydrogen sulfide is readily released from solution as a corrosive and extremely toxic gas with an obnoxious odor. By controlling the pH of the solution above 10.0, the H2S release can be minimized. If the pH of the wastewater is expected to be lower, the sulfide concentration should be reduced below 1.0 mg/L to prevent H2S odor problems,

Pretreatment The pretreatment unit operations which may be necessary for various types of treatment processes are shown in Table 10. Screening to remove debris, equalization to provide uniformity of effluent, and neutralization with precautions for possible generation of hydrogen sulfide gas, to prevent excessively high pH values are generally necessary prior to discharge to a municipal collection system. Chemical precipitation may be needed to reduce the amount of chromium in the effluent. The considerations in Table 10 are based on the assumption of fat and grease recovery as a by-product. Where this is not practiced, grease removal facilities may also be needed.

Petroleum Refining Industry Wastes Industry Description This industry is engaged in producing gasoline, kerosine, fuel oils, residual fuel oils, lubricants, and other products through distillation of crude oil, cracking, or other processes. Petroleum refining is a combination of several interdependent processes, many of which are highly complex. There are more than two dozen separate processes

Industrial Wastewater Sources

247

fundamental to the operation of a refinery producing the full spectrum of products from crude oil. The major operations within a refinery include: crude oil storage; desalting; fractionation by pressure, atmospheric and vacuum distillation; thermal and catalytic cracking; reforming; polymerization; and alkylation. Other operations generally involving separation and finishing of the products to specifications, include acid treatment of lubricating oil stocks, sweetening of gasoline, extraction, and stripping. Storage of crude oil to provide adequate working supplies and to equalize process flow involves the separation of water and suspended solids from crude oil. Table 10. Presentment Unit Operations for the Leather Tanning and Finishing Industry. Suspended Biological System

Pretreatment Group Chrome Tanning

Vegetable Tanning

1

Fixed Biological System

Coarse Solids Separation

Coarse Solids Separation

+

+

Gmt Removal + Equalization + Chemical Precipitation (Heavy Metals) + Solids Separation + Neutralization

Grit Removal + Equalization + Chemical Precipitation (Heavy Metals) + Solids Separation 4Neutralization

Coarse Solids Separation

Coarse Solids Separaration

+

+

Grit Removal + Equalization •f Neutralization

Grit Removal + Equalization 4- Neutralization

Independent Physical-Chemical System Coarse Solids Separation + Grit Removal + Equalization

Coarse Solids Separation + Grit Removal + Equalization

Pretreatment requirements assume fat and grease recovery as saleable by-product.

The first major operation in a refinery is the crude oil desalting process for removing inorganic salts and suspended solids from the crude oil prior to fractionation. Water is used in the desalting process as a sequestering agent. The crude oil after desalting is generally passsed through atomspheric and/or vacuum distillation to separate light overhead products, side-stream distillates, and residual crude oil. Steam is used in this process, and the steam condensate from the overhead accumulation is discharged as waste water. The heavy fractions removed during the crude oil fractionation and distillation process can be cracked using either thermal, catalytic, or hydrocracking processes to yield light oil fractions such as domestic heating oil. The low-octane fractions obtained from the foregoing processes can be converted to yield high-octane gasoline blending stock by reforming, polymerization, and/or alkylation processes. The reforming converts naphthas to finished high-octane gasoline. Reforming is a relatively clean process producing a low volume of dilute wastewater. The polymerization process produces waste waters containing sulfide, mercaptans, high pH materials, and nitrogen compounds. Phosphoric acid or sulfuric acid is used in the polymerization process and generates solid wastes. The alkylation process uses a sulfuric acid or hydrofluoric acid catalyst to convert isoparaffins and olefins into high-octane motor fuel. Solvent refining is used in a refinery to extract lubricating oil fractions and aromatics from feedstocks containing various types of hydrocarbons.

248

Liquid Filtration

The petroleum refining industry uses very large quantities of water for process and cooling purposes. Approximately 90 % of the water used in refineries is for cooling purposes. Lesser water uses include: steam generation (boiler-feed.), direct processing, fire protection, and sanitary uses. Steam is used in stripping and distillation processes, where it comes in contact with petroleum products, thereby contributing to the total waste water flow from refineries. Oily process wastes and oil-free wastes are collected separately in some refineries so that the oily wastes can be treated for oil removal before mixing with other waste streams. The spent caustics and spent acids are generally collected and sold or disposed by other means. Few refineries neutralize these wastes for discharge, to the waste water collection system. The sour water (condensates from various fractionation units) containing sulfides and ammonia is generally steam or air-stripped before being discharged to the sewer lines. Depending on the pH of the sour water, the stripping can reduce the sulfide and ammonia concentrations in the final effluent. In general, the in-plant control methods employed by the industry (sour-water stripping, spent-caustic neutralization and oxidation, slop oil recovery, etc.) will determine the final effluent characteristics and the level of pretreatment required for discharge to the municipal collection system. The following specific in-plant practices are frequently employed: 1. Sour-condensate stripping is used to remove sulfides (as hydrogen sulfide, ammonium sulfide, and polysulfides) before the waste water enters the sewer. The sour water is usually treated by stripping with air, stream, or flue gas. Hydrogen sulfide released from the wastewater can be recovered as sulfuric acid or can be burned in a furnace. Hydrogen sulfides at concentrations in the range of 10 to 15 mg/L can cause upsets in biological treatment plants, and removal of sulfides from the sour water by stripping would prevent such upsets. 2. Spent caustic neutralization is applied to both phenolic and sulfidic waste streams, but oxidation of spent caustics is limited to sulfide waste streams, since phenols inhibit the oxidation of sulfides in spent caustics. 3. Spent acids (generally sulfuric) can be recovered for reuse or sold to acid manufacturers, thereby avoiding their discharge to sewer systems. Spent catalysts such as aluminum chloride and phosphoric acid can either be regenerated for reuse or disposed of as landfill.

Wastewater Characteristics The characteristics of the process wastewaters from the industry are shown in Table 11. Petroleum refineries use gravity oil separators to recover free oil from process effluents. The effluents from gravity oil separators are therefore used to define the wastewater characteristics hi the table. The petroleum refineries operate on a continuous basis throughout the year except for separate process shutdowns for

Industrial Wastewater Sources

249

preventive maintenance and equipment repairs. Many refineries have a segregated wastewater collection system, with separate subsystems for clean and polluted waste streams. The 'clean' waters may include pollution-free cooling waters, boiler blowdown, and cooling tower blowdown. Table 11. Wastewater Characteristics of the Petroleum Refining Industry. Year-Round

Characteristics Industry Operation Flow BOO TSS TDS COD

Continuous Average* Low High High

Grit Cyanide Chlorine Demand pH Color

Present Present High Acid-Alkaline Low

Turbidity Explosives Dissolved Gases Detergents Foaming

Low Present Present Low Absent

Heavy Metals Colloidal Solids Volatile Organics Pesticides

Present Low Present Absent

Phosphorus Mitrogen Temperature Phenol Sulfides

Deficient' Adequate High2 High High

Oil & Grease Coliform (Total)

High Low

After gravity oil separation (API Separators). The refinery wastewaters have high COD/SOD ratios indicating the presence of biologically resistant organic chemicals. : If cooling tower blowdown is also discharged with the process wastewater, phosphates may be present depending on water treatment. 1

The characteristics of wastewater drawn from storage tanks will depend on the quality of the crude oil stored and may contain dissolved inorganics, oil, and suspended solids. The steam condensate from the overhead accumulator can be characterized as having oil, sulfides, mercaptans, and phenol. If barometric condensers are used in vacuum distillation, the condenser water will have very stable oil in emulsion. However, if the barometric condensers are replaced by surface condensers, the condenser water will be essentially free of oil.

250

Liquid Filtration

The major wastewater from the alkylation process is the spent caustic from the neutralization of the hydrocarbon stream from the reactor. Even though the spent acids are recovered as salable by-product, leaks and spills of acid catalysts could reach the sewer lines. The major pollutants from the solvent refining processes include solvents such as phenols, glycols, and amines. Other processes for the manufacture of waxes and asphalt, and for finishing and blending of gasoline, produce relatively low volumes of dilute wastewater. In general, the most significant process waste waters from petroleum refining are: crude oil desalting waste, storage tank draw-off, steam condensates, spent caustics, spent acids, product losses, and leaks and spills of solvents used in extraction processes. The process wastewaters, which come in direct contact with petroleum hydrocarbons, contain free and emulsified oil, sulfides, phenols, ammonia, BOD, COD, heavy metals, and alkalinity as major waste constituents. Refinery wastewaters generally are susceptible to conventional biological treatment methods after adequate pretreatment. In addition, phosphorus supplementation may be required in some cases to provide a nutrient-balanced system for biological treatment. This supplemental nutrient requirement will depend on the phosphorus content of the cooling tower blowdown and its inclusion in the process wastewater. Pretreatment The pretreatment unit operations for various types of wastewater treatment facilities are shown in Table 12. These pretreatment operations assume the following industrial practices: 1. 2. 3. 4. 5.

Sour condensate stripping to reduce sulfides and/or ammonia Spent-caustic neutralizations Spent-acid neutralization and recovery or disposal by other means Separate collection and disposal of acid sludges, clay, and spent catalysts Gravity separation of free oil from process effluents

Table 12. Pretreatment Unit Operations for the Petroleum Refining Industry. Suspended Biological System Equalization + Coagulation Solids Separation1 + Neutralization

Fixed Biological System Equalization + Coagulation Solids Separation + Neutralization

Independent Physical-Chemical System Equalization -fCoagulation Solids Separation2 + Neutralization

| j 1

!

' Pretreatment Unit Operations apply to API Separator Effluent. Combined with oil removal to insure oil concentration below 50 mg/1.

2

Depending upon the in-plant control methods used within a refinery it may be necessary to add sulfide removal and neutralization to the pretreatment operations listed in Table 12. The oil concentration in the wastewater should be reduced to 50 mg/L in order to insure trouble-free operation in secondary biological treatment facilities. In addition, the presence of oil in a sewer would constitute a fire and

Industrial Wastewater Sources

251

explosive hazard. For this reason, sewer ordinances generally have prohibited the discharge of refinery wastewaters to municipal facilities. The heavy metals present in refinery effluents (As, Cd, Cr, Co, Cu, Fe, Pb, Ni, and Zn) are generally in such low concentrations that they do not pose serious problems for conventional treatment methods. If heavy metals reduction should be required by effluent guidelines, provisions should be made for their removal before discharging to municipal systems. The biological sludge developed from refinery wastewaters can be thickened and dewatered by conventional methods such as vacuum filters and centrifugation. The pretreatment unit operations which may be necessary for various types of treatment facilities are shown in Table 12. The most prevalent in-plant treatment methods are sour-water stripping, neutralization and oxidation of spent caustics, ballast water treatment, and slop-oil recovery. These measures substantially reduce the waste loadings and to a significant degree are required to protect subsequent treatment. In addition to these in-plant control methods, refineries use gravity oil separators to recover free oil from process effluents. Food and Meat Packing Industry Wastes Industry Description This industry includes slaughterhouses, packinghouses, processing plants (beef, poultry, hog, and sheep), and rendering plants. Live animals are usually held in holding pens for less than one day prior to slaughter. In the killing area, the animals are slaughtered and the carcasses drained of blood. The processes of skinning, defeathering or dehairing, and eviscerating follow the slaughtering of the animals. Depending on the desired product, carcasses may be cut into smaller pieces, e.g., hogs are cut into parts such as hams, sides, loins, and shoulders. These parts may be further processed (e.g., smoked or pickled), or they may be shipped directly to wholesalers without further processing. Because of the similarity of the wastewaters from meat products industry, there are no separate pretreatment subgroups for this industry. The following industrial practices can significantly influence the wastewater characteristics: Sanitation Requirements - Unlike most other industries, the food industry is required to maintain strict sanitary conditions, which limits the amount of process wastewaters that can be recycled. The industry practices in-plant recovery (e.g. blood or grease) to reduce wastewater strength. In most plants, blood is collected and subsequently processed. The recovery of blood represents an in-plant practice which is extremely desirable since whole blood represents a BOD concentration of over 150,000 mg/L. In addition, the dry handling of such wastes as manure and bedding materials from holding pens, paunch manure from eviscerating, and meat cuttings and trimmings will significantly reduce the quantity of waste materials discharged to the sewers.

252

Liquid Filtration

Implementation of these practices will affect the concentration and quantity of waste constituents, but not the quantity of waste water, Rendering — Rendering is a major unit process where in-plant modifications can significantly influence the pretreatment considerations. Wet and dry rendering are two subprocesses presently used within the industry. In the wet process, the meat byproducts in a batch tank are cooked by direct injection of steam. Dry rendering uses only heat, and little wastewater is produced. In wet rendering, the solids in the water phase are screened out, and the remaining tank water maybe evaporated or sewered. The tank water is a major source of organic pollution, when sewered, and has a high BOD value. Evaporation of wet-rendering tank water and the installation of entrainment separators on barometric condensers may reduce the need for pretreatment of wet-rendering process wastewaters. Wastewater Segregation — Wastewater originating within a meat products plant will generally be made up of wastewater from the operations, sanitary wastes, and wastewaters from auxiliary sources (e.g., cooling water from ammonia condensers in the refrigeration systems). Many large plants providing their own complete or partial treatment have found it economical to segregate wastewaters into blood, clean water, manure-free water, and manure waters.

Wastewater Characteristics The characteristics of wastewaters from the meat products industry are shown in Table 13. The meat products industry is a year-round operation with daily operation on an intermittent basis. Plants usually shut down daily for an extensive clean-up period following the processing period. This practice results in the generation of intermittent wastewater flows. Each plant may have a number of operations, depending on the products and degree of processing. However, the wastewaters generated from any meat products plant can be characterized as containing high concentrations of BOD, COD, TSS, TDS, and grease. Washdown of holding pens, mainly to remove manure and urine, will add to the BOD and suspended solids concentrations. The processes of skinning, defeathering or dehairing, and eviscerating are sources of BOD, grease, and suspended solids. The disposal of paunch manure and the washing of carcasses during eviscerating are particular operations which generate substantial pollutants. The processing of hogs and fowl produce a floating solids problem caused by hair and feathers. Dressing and processing operations are minor wastewater sources compared to the slaughterhouse operations. Wastewaters from these operations contain grease and solids originating from equipment washdown and product losses. Druing these operations, various trimmings, fat, and fleshings are produced. Considerations in the treatment of meat product wastewaters with domestic wastes are the high chlorine demand, the presence of surface-active agents, focal conform,

Industrial Wastewater Sources

253

intermittent flows, and the high septicity potential due to the high organic content of meat product wastewaters. In general, meat product waste waters are amenable to either biological or chemical treatment. Table 13. Wastewater Characteristics Meat Products Industry, Characteristics

Meat Products

Industry Operation Flow BOD TSS IDS

Year-round Intermittent1 High-Ext. High High High

COD Grit Cyanide Chlorine Demand pH

High-Ext. High Absent Absent High Neutral

Color Turbidity Explosives Dissolved Gases Detergents Foaming Heavy Metais Colloidal Solids Volatile Organics Pesticides

High High Absent Absent Present Absent Absent High Absent Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Present Present Normal-High2 Present3 Absent

Oil & Grease Coliform (Fecal)

Present Present

1

Wastewater flow is intermittent over the day or week. ' Temperature equal to higher than domestic wastewater; may affect design but not harmful to joint treatment processes, 3 Phenols may be present in sanitizers used for clean-up.

Pretreatment The pretreatment unit operations which may be necessary for various types of treatment processes are shown in Table 14. In addition to screening and free-floating grease removal, various in-plant control practices, such as blood recovery, and separate handling of paunch manure as solid waste would greatly reduce the waste constituents in process wastewaters. When meat product wastewater is combined with domestic wastewater, pretreatment should include equalization to reduce organic and hydraulic fluctuations. The equalization basin should be aerated to prevent septic conditions.

254

Liquid Filtration

Table 14. Pretreatment Unit Operations for the Meat Products Industry. Suspended Bioloqical System Coarse Solids Separation + Grease Removal2

Fixed Biolodical System Coarse Solids Separation + Grease Removal2

Independent PhysicalChemical System Coarse Solids Separation + Grease Removal

1. Assumes in-plant recovery and separate handling of blood, manure, and paunch manure. 2. Only free-floating oil and grease.

Beverages Industry Wastes Industry Description This industry is engaged primarily in the manufacture of malt, malt beverages (ale, beer, and malt liquors), wines (table wine, dessert wine, and brandy), distilled spirits, bottled and canned soft drinks, and flavoring extracts and syrups. These products can be classified under two major groups according to their basic manufacturing processes as: 1. Fermentation Products (beer, wine, distilled spirits, malt) 2. Extraction Products (soft drinks, flavors, and extracts) The fermentation products are made from grains or fruits, while the extraction products are made from flavor substitutes of oils such as cocoa, vanilla, and orange oil. The fermentation products derived from grains are manufactured by cooking the grains, fermenting the cooking liquor with a yeast culture, and separating the fermented alcohol by clarification and filtration. The manufacturing processes used for the production of flavoring extracts and syrups are normally proprietary in nature. The basic processes for the recovery of natural flavoring can be listed as follows: 1. 2. 3. 4.

Steam distillation and petroleum ether extraction (essential oils). Expression (hydraulic pressing) and petroleum ether extraction (fruit syrup). Expression and evaporation (jams). Alcoholic extraction of vanilla and other tissue.

The soft drink baffling and canning plants use flavor extracts and purchased syrups. The bottling and canning process involves bottle washing and sterilizing, mixing of flavor extracts and syrup, carbonation, and filling. The pretreatment sub-groups for this industry are as follows: • • •

Malt, malt beverage, and distilled spirits (except industrial alcohol). Wine and brandy. Bottled and canned soft drinks, and flavors and syrups.

Industrial Wastewater Sources

255

During the brewing and fermentation process, malt and hops are added to convert starch to sugar and to incorporate a bitter taste to the product. Water is used in the process for cooking, cooling, container washing and other miscellaneous uses. Both solid wastes and liquid wastes are generated in the process. Spent grains, excess yeast, and spent hops are the solid wastes, and are generally hauled away or dried for livestock and poultry feed. A variation of this type of disposal is where certain distilleries manufacturing distilled wine or spirits, the stillage is discharged with other liquid wastes. Liquid process wastes result from fermentation, aging, filtration and evaporation, and washing and clean-up operations. Liquid wastes are also discharged from auxiliary operations such as cooling, boiler blowdown, and water softening. The fermentation process results in the generation of "lees", which is a mixture of wine, yeast cells, and other sediment. The lees is considered a liquid or semi-liquid waste, which is either discharged directly to the sewer or recovered in the case of large wineries. In order to improve the quality of the wine, the fermented liquid is often processed by a sequence of racking, filtration, and fining operations. The wastes from racking (clarification) and filtration often produce a sludge containing significant quantities of wine. When these wastes are sewered, they add significantly to the BOD and solids concentration of the wastewater. Brandy is produced by distillation of wine and condensation of the overhead in order to obtain a beverage with high alcohol content. The stillage from such distillation is a significant liquid waste. The manufacture of flavoring extracts and syrups also generates both liquid and solid wastes. Solid wastes are the residues after extraction of flavors and syrups. Wastewaters from normal extraction operations include: Fruit Expression: 1. Water used for washing fruits. 2, Hydraulic press clean-up. Evaporation: 1. Evaporator condensate. 2. Kettle wash water. Steam Distillation: 1. Boiler blowdown. 2. Bottoms from packed column. The major wastewater sources in the bottling industry are the bottle washing and clean-up operations. Auxiliary wastewaters such as cooling, air conditioning, and boiler blowdown are also generated. The predominant method of disposing of liquid wastes from beverage plants is by discharging to municipal sewerage systems. A typical plant collects all its wastewaters in a common sewer and discharges them to municipal sewers.

256

Liquid Filtration

Wastewater Characteristics The characteristics of process wastewaters from each pretreatment sub-group are shown in Table 15. The beverage industries operate throughout the year. However, the volume of waste production and the loading will vary with the season depending upon product demand. Table 15. Wastewater Characteristics of the Beverages Industry. Wine and Brandy

Soft Drinks Bottling1

Year-round Intermittent-Continuous2 High Low - High High

Seasonal Intermittent High-Ext. High Low - Ext. High High

Year-round Intermittent Average - High Low - High Low - High

pH

High Present Absent No Data Acid-Neutral

High-Ext. High Present Absent No Data Acid-Alkaline3

Average - High Present Absent No Data Alkaline3

Color Turbidity Explosives Dissolved Gases Detergents

Present Present Absent Present4 Present

Present Present Absent Absent Present4

Present Present Absent Present Present4

Foaming Heavy Metals Colloidal Solids Volatile Organics Pesticides

Present Absent Present Present Absent

Present Absent Present Present Absent

Present Absent Present Present Absent

Phosphorus Nitrogen Temperature Phenol Sul fides

Deficient Deficient Normal-High5 Absent Absent

Deficient Deficient High Absent Absent

Deficient Deficient Normal Absent Absent

Oil & Grease Coliform (Fecal) Coliform (Total)

Absent Absent Present

Absent Absent Present

Absent Absent Present

Characteristics Industry Operation Flow

BOD TSS IDS COD Grit Cyanide Chlorine Demand

Mait Beverages and Distilled Spirits

Pollutants characteristics represent only Bottling Industry; no data available for flavors and syrups. Malt beverages generate wastes on a continuous basis; distilled spirits waste flow will be cyclic. 3 Alkaline pH due to caustic detergents used for bottle washing. 4 Surface active agents are discharged primarily from bottle washing. 5 Temperature equal to or higher than domestic wastewater, may affect design but not harmful to joint treatment processes. 1

Major considerations in the treatment of beverage wastewaters are die presence of large paniculate matter in suspension and the fluctuations in hydraulic and organic loads. The following is a brief description of the wastewater from each pretreatment group:

Industrial Wastewater Sources

257

Malt Beverages and Distilled Spirits — The wastewaters generated from the malt and malt beverages industries have as major constituents BOD, SS, pH, and temperature. The waste solids from the malt house, and the excess yeast, spent grains, and spent hops from the malt beverages industry are disposed of either by hauling away or by on-site drying to make cattle feeds. If the spent wet grain is dried in the brewery, the spent grain liquid must be disposed of; generally, it is discharged to municipal sewers without pretreatment. The distilleries produce wastewaters from cooking and fermentation of grains, the stillage or slops from distilling operations, and from washing and baffling operations. The stiliage from the distilleries contain yeast, proteins, and vitamins. Depending upon the size of the plant, complete or partial recovery of stillage is practiced. The major constituents in distillery wastewaters include BOD, suspended solids, acidity, and heat. Wine and Brandy — The wine and brandy industries produce wastewaters from crasher-stemmer, pressing, fermentation, clarification and filtration, distillation, and bottling operations. Brandy is manufactured by distillation of wine and, therefore, results in the generation of stillage or "still slop". The stillage is a significant liquid waste in the manufacture of brandy. The wastewaters are high in organics and have the potential of introducing shock loads in treatment works. Soft Drinks, Flavors, and Syrups — The wastewaters generated from the manufacture of flavor extracts and syrups are generally discharged to municipal sewerage systems without treatment. The bottling and canning of soft drinks generate wastewaters primarily from bottle-washing operations. These wastes contain BOD, suspended solids, and alkalinity. Bottling operating practice involves recirculation of final rinse water for pre-rinsing, thereby reducing the volume of waste water discharged to the sewers.

Typical pretreatment operations include screenings, grit removal, and equalization. The pretreatment unit operations are listed in Table 16. The addition of auxiliary wastes (cooling, boiler blowdown, and water softening) will lower the strength of total effluent from the industry. In general, the wastewaters from the beverage industries are amenable to treatment by conventional processes, such as activated sludge and trickling filters. The pretreatment unit operations listed in Table 16 are based on the assumption that the following in-plant pollution control methods are practiced: 1. Hauling or drying of spent grains, hops, and stillage. 2. Separate solids-handling and disposal of crusher'-stemmer and pressing wastes. Spent grains, hops, stillage, crusher-stammer, and pressing wastes can be characterized as solid wastes rather than liquid wastes. Therefore, it is desirable to collect them separately for disposal.

258

Liquid Filtration

Table 16, Pretreatment Unit Operations for the Bevaerages Industry. Pretreatment SubGroup Malt Beverages

Wine and Brandy

Soft Drinks Bottling

Suspended Biological System

Fixed Biological System

Independant PhysicalChemical System

Coarse Solids Separation + Grit Removal + Equalization + Neutralization

Coarse Solids Separation + Grit Removal + Equalization + Neutralization

Coarse Solids Separation + Grit Removal + Equalization *+ Neutralization

Coarse Solids Separation + Grit Removal + Equalization + Neutralization

Coarse Solids Separation + Grit Removal + Equalization + Neutralization

Coarse Solids Separation + Grit Removal + Equalization + Neutralization

Grit Removal + Neutralization

Grit Removal + Neutralization

Grit Removal + Neutralization

Plastic and Synthetic Materials Industry Wastes Industry Description Discussion on this industry sector covers the manufacture of plastic and synthetic materials, but not the manufacture of monomers, formed plastic products (other than fibers), and paint formulations. The manufacture of resins used in paints is also included. Plastics and resins are chain-like structures known chemically as polymers. Polymers are synthesized by one or more of the following processes: bulk, solution, emulsion, and suspension. After polymerization, the products undergo separation, recovery, and finishing before being marketed. There are numerous plastics and synthetics manufactured in this industry and only a few are covered in this discussion. The industry can be divided into the following pretreatment sub-groups: • • • • •

Rayon Fibers Nylon Fibers High- and Low-Density Polyethylene Resins Urethane Resins, Polyolefin Fibers, Polyvinil Acetate Resins, Poly vinyl Alcohol Resins,Polyester Fibers Cellulosic Resins, Cellophane, Polypropilene Resins, Cellulose Acetate Fibers, Polyvinil Chloride Resins, Polystyrene, ABS, SAN Resins, Phenolic Resins, Nylon Resins, Polyacetal Resins, Acrylic Fibers.

The following industrial practices can significantly influence the wastewater characteristics:

Industrial Wastewater Sources

259

Suspension Polymerization — In suspension polymerization, a monomer is dispersed in a suspending medium consisting of a mixture of water and suspending agents such as poly vinyl alcohol, gelatin, etc. The suspension is heated, and polymerization occurs. The polymer is then separated, washed, and dried. The concentrate from the separation process may contain suspending agents, surface-active agents, catalysts, (e.g., benzoyl, lauroyl) and small amounts of unreacted monomers. Emulsion Polymerization — Emulsion polymerization consists of solubilization and dispersion of the monomer in a solvent (e.g., water, cyclohexane, or tetrahydrofuran) with the appropriate emulsifiers (e.g., soaps or surfactants). Before polymerization occurs, initiators such as persulfates, hydrogen peroxides, etc, are added. When polymerization is complete, the product is a milk-like latex of permanently dispersed polymer, from which the polymer particles are recovered, generally by spray drying or by coagulation and centrifugation. Solution Polymerization — Solution-polymerization relies on a solvent to dissolve the monomer, catalyst and co-catalysts reaction, the polymer is precipitated using an antisolvent (e.g., n-hexane, methanol). The polymer is then filtered and dried. Bulk or Mass Polymerization — Bulk or mass polymerization is different from the foregoing processes in that no carrier liquid is used. Therefore, there is generally little or no process waste water associated with this process. Wastewater Characteristics The characteristics of the process wastewaters from the manufacture of plastic and synthetic materials are shown in Table 17. The plastic and synthetic materials industry is typically a continuous year-round operation. Because it is technically and economically advantageous, many firms manufacture several different, but related chemical products at one location. For example, a typical complex makes ethylene, polyethylene, sulfuric acid, ethyl chloride, ammonia, nitric-acid and phosphoric acid. In the first group the wastewater has relatively high BOD, COD, and TSS; heavy metals (Zn, Cu) and synthetic fiber losses. The second group's wastewater has low BOD and COD, may be either acidic or alkaline. The third group is characterized as either having no process water or as having process wastewater containing virtually no pollutants. The fourth group's wastewater has variable BOD and COD, may be either acidic or alkaline, and contains synthetic fibers. Discharge of faulty batches from synthetic fiber plants may introduce shock loads into the wastewater treatment facility. Conditions significant in the design of treatment facilities include high chlorine demand, the presence of surface-active agents, high solids concentrations, and nutrient deficiency. The process diversity and complexity, as well as the proprietary nature of many of the process chemicals, require that the pretreatment be established on a caseby-case basis after thorough investigation.

260

Liquid Filtration

Table 17. Wastewater Characteristics for Plastics and Synthetic Materials Industries, Characteristics

Sub-Group 1

Sub-Group 3

Sub-Group 2

Sub-Group 3

Year-round Cont.-Variable High High High High

Year-round Continous

Year-round Continous

Low Low Low Low

Low Low Low Low

Absent Absent High Acid-Basic Low-Average

Absent Absent

Absent Absent

Low

Low

Acid-Basic Low-Average

Neutral

Turbidity Explosives Dissolved Gases Detergents Foaming

High Absent Absent Absent Absent

Low

Low

Absent Absent Absent Absent

Absent Absent Absent Absent

Heavy Metals Colloidal Solids Volatile Organics Pesticides

Present High Absent Absent

Absent

Absent

Low

Low

Present Absent

Absent Absent

Absent Average Present Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Deficient Deficient Normal-High Absent Absent

Deficient Deficient Normal-High Absent Absent

Deficient Deficient Normal-High Absent Absent

Deficient Deficient Normal-High Present Absent

Oil & Grease Coliform (Fecal)

Absent Absent

Present Absent

Absent Absent

Absent Absent

Industry Operation Flow

BOD TSS IDS COD Grit Cyanide Chlorine Demand

pH Color

Year-round Continuous Average-High Low-High Low-High Average-High Absent Absent Average-High Acid-Basic Low-Average

Low

Low-High Absent Absent Present Absent

Table 18. Pretreatment Unit Operations for the Plastic and Synthetic Materials Industry. Suspended Biological System

Fixed Biological System

1

Coarse Solids Separation + Neutralization -1- Chemical Precipitation (heavy metals)

Coarse Solids Separation + Neutralization + Chemical Precipitation (heavy metals)

Coarse Solids Separation + Neutralization 4- Chemical Precipitation (heavy metals)

2

Oil Separation + Neutralization

Oil Separation + Neutralization

Oil Separation + Neutralization

3

Pretreatment Not Required

Pretreatment Not Required

Pretreatment Not Required

4

Coarse Solids Separation + Neutralization

Coarse Solids Separation + Neutralization

Coarse Solids Separation + Neutralization

Pretreatment Sub-Group

Independent PhysicalChemical System

Oil separation required to reduce mineral oil (petroleum sources) concentration below 50 mg/L.

Industrial Wastewater Sources

261

Blast Furnaces, Steel Works, and Rolling and Finishing Wastes Industry Description This industry sector includes: pig iron manufacture; manufacture of ferro-alloys from iron ore and from iron and steel scrap; converting pig iron, scrap iron, and scrap steel into steel; hot rolling; and cold finishing. Blast furnaces and by-product (or beehive) coke ovens are also included under this category, although these are almost nonexistent in the United States today. The complex and interdependent operations involved in a steel industry around the world can be listed as follows: 1. 2. 3. 4. 5.

Coke Works Iron Works Steel Works Hot Forming Cold Finishing

Significant quantities of water are used, both for processing and for cooling purposes. The steel industry generates enormous volumes of waste water. In older plants around the world, coke is used in large quantities for the production of pig iron. Old style large iron and steel manufacturing operations include the production of coke from coal. There are two methods generally used for the production of coke: 1) the beehive process; and 2) the by-product or chemical recovery process. The beehive process uses air in the coking oven to oxidize the volatile organics released from coal and to recover the heat for further distillation. The by-product or chemical recovery process is operated in the absence of oxygen, and the heat required for distillation is provided from external fuel sources. In the byproduct process, coal is heated in the absence of air to a temperature at which the volatile matter is driven off. At the end of coking cycle, the hot residual coke is conveyed to a quenching station, where it is cooled with a spray of water. The offgases from the coke oven are cooled in a cooling train, where tar and ammonia liquor separate out. The tar contains a large proportion of phenols removed from the furnace. Iron is manufactured from iron ore (iron oxide) in blast furnaces, with carbon monoxide (from coke) as a reducing agent, again in older plant operations as in Russia or the Ukraine. The major impurity (silica) in the iron ore is removed from the blast furnace as molten slag, through the use of limestone. Steel is manufactured from pig iron by adjusting the carbon content of the alloy to approximately 1 %. The three principal steel-making units are the electric arc furnace, the open-hearth furnace, and the basic oxygen furnace. All three methods use the same raw materials and produce similar wastes. Pure oxygen or air is used to refine the hot iron into steel by oxidizing and removing silicon, phosphorus, manganese, and carbon from the iron.

262

Liquid Filtration

The steel ingot obtained from the furnace is reheated to provide uniform temperature for further processing or hot forming. The ingot steel is generally processed in a blooming mill or slab mill to form plates, sheets, strip, skelp, and bars. The cold finishing operations are used for the conversion of hot-rolled products to give desired surface, shape, or mechanical properties. These operations include pickling, cold rolling, tinplating, coating, shaping, and drawing to make various finished products. Integrated iron and steel mills may operate many different subprocesses generating wastewaters of varying characteristics. The only two types of waste stream susceptible to joint treatment are the coke oven wastewaters and the pickling liquor. Most process wastes occur in such large quantities and contain only suspended impurities that it is more logical to treat them on-site and discharge directly to surface waters. Coke oven wastewaters and the pickling liquor are often found sent to municipal treatment works in many plants of the world. The strong acid pickle liquor, containing iron salts of mineral acids, is usually collected separately for other means of disposal or for use in waste treatment plants. The acid salts of iron are useful either as a flocculant aid, as a precipitating agent for phosphorus removal, or as a neutralizing agent in waste treatment plants. Any such use should be investigated before discharging the spent pickle liquor as a waste stream. Recovery of strong acid pickle liquor for reuse is also practiced for hydrochloric acid systems, although this is rare. Wastewater Characteristics The characteristics of die process wastewaters from various operations within the steel industry are shown in Table 19. The wastewaters generated in the steel industry vary widely between operations and they are generally segregated for treatment. Some older steel mills, however, still have common collection systems for discharging the total plant flow. The steel industry operates throughout the year and generates wastewaters over a 24-hour day. The volume and characteristics of waste water are subject to hourly variations from batch dumping of acid baths and still bottoms. The major constituents present in the wastewater are phenol, cyanides, ammonia, oil, suspended solids, heavy metals (Cr, Mi, Zn, Sn), dissolved solids (chlorides, sulfates), acidity, and heat. The process wastewaters are generally treated on-site before disposal. Joint treatment of these wastes with municipal wastes is limited to small installations. A significant portion of the wastewater generated contains suspended solids and dissolved solids that are inorganic in nature. The only two waste streams generally susceptible to joint treatment are the coke oven wastewaters (ammonia liquor, still bottoms, and light oil recovery wastewaters) and the pickling liquor. The other process wastewaters are high in solids (sub-micron iron oxide dust) and heavy metals. Pretreatment to reduce these waste constituents generally results in an effluent which can be discharged directly to surface waters.

Industrial Wastewater Sources

263

The coke oven process wastewaters are amenable to biological treatment when they constitute only a minor fraction (approximately 25 %) of the total waste water flow to the treatment facility. The cyanide concentration (with its resulting aquatic toxicity characteristics) is the primary consideration in the treatment of coke oven process wastes. Table 19, Wastewater Characteristics for Steelmaking Operations. Characteristics

Coke Works

Iron Works Steel Works Hot Forming

Cold Finishing Year-round Intermittent Low-Average Low-High High

Year-round Continuous Low-Aver. Low-High

Year-round Continuous

Year-round Continuous

Low

Low

Aver. -High

Low-High

Low

Low

Low

Low-Aver. Present Present

Low

Low

Absent Absent

Absent Absent

Low

Low

Low

Low

pH

Low-Average Present Present High Neutral

Neutral

Neutral

Neutral

Acidic

Color Turbidity Explosives Dissolved Gases Detergents

Absent Present Absent Present Absent

Absent Present Absent Present Absent

Absent Present Absent Present Absent

Absent Present Absent Absent Absent

Absent Present Absent Absent Present

Foaming Heavy Metals Colloidal Solids Volatile Organics Pesticides

Absent Absent Present Present Absent

Absent Present Absent Absent Absent

Absent Present Absent Absent Absent

Absent Present Absent Present Absent

Absent Present Present Present Absent

Characteristics Phosphorus Nitrogen Temperature Phenol Su! fides

Coke Works Deficient Adequate High Present Present

Iron Works Deficient Adequate High Present Present

Steel Works Deficient Deficient High Absent Absent

Hot Forming Deficient Deficient High Absent Absent

Cold Finishing Deficient Deficient High Present Absent

Oil & Grease Coliform (Total)

Present Absent

Absent Absent

Absent Absent

Present Absent

Present Absent

Industrial Operation Fl ow

BOD TSS TDS COD Grit Cyanide Chlorine Demand

Year-round Intermittent Low-Average

Low Low

Low-Average Absent Present

The following are characteristics of the process wastewaters: The still bottoms, containing phenol, constitute the major wastewater source from the coke oven process. Since the beehive oven process utilizes the heat value in the offgases, only the quench water is. discharged as wastewater. The gases (CO2, CO, N2, and HCN) leaving the furnace are hot and contain dust particles. The gases also contain water vapor and traces of hydrogen sulflde. In order to clean the exit gas from the blast furnace operations, the gas is generally passed through dust collectors, scrubbers, and coolers. The water used in the scrubbers and coolers is the primary wastewater source in iron manufacture. The waste products from this process are slag and the oxides of iron released as submicron dust particles. Precipitators or venturi scrubbers are used to clean the exit gas, and the characteristics of the wastewater

264

Liquid Filtration

discharged from the process will depend primarily on the gas cleaning system. If scrubbers are used, the wastewater generated will be acidic in nature due to the presence of sulfur oxides in the exit gas. Water is used under high pressure to remove scale and for cooling purposes. The primary wastewaters are the scale-bearing waters and cooling waters containing primarily scale and oil. Steel pickling to remove oxides and scales is accomplished through solutions of H2SO4, HC1, or hydrofluoric acid. The pickled steel is then rinsed with water and coated with oil before proceeding to the next step in the process. The cold rolling process involves passing unheated metal through rolls for reducing size or thickness, and improving the surface finish. Plating of steel products is done electrolytically, and is accomplished in either an alkaline or an acid electrolyte solution. If acid electrolyte is used, the process system will consist of alkaline washing, rinsing, pickling, plating, quenching, chemical treating, rinsing, drying, and oiling. The most commonly used metallic coatings are tin, zinc, nickel, chromium, cadmium, copper, aluminum, silver, gold, and lead. Wastewaters generated from cold finishing operations include: rolling solutions, cooling water, plating wastes, pickling rinse waters, and concentrated waste-acid baths. Rolling solutions and cooling water generally contain oil and suspended solids as pollutants. Plating wastes, pickling rinse, and concentrated acid baths may contain various heavy metals (Cr, Cd, Ni, Zn, Sn) as well as cyanides, acids, and alkali. The pretreatment unit operations which are often employed are listed in Table 20. Note that the wastewaters from the coke oven and cold finishing operations contain cyanide and heavy metals and therefore require special consideration in waster water treatment works. Table 20. Pretreatment Unit Operations for Steelmaking Processes. Pretreatment Sub-Group

Suspended Biological System

Fixed Biological System

Independent PhysicalChemical System

Coke Production

Equalization + Solids Separation

Equalization + Solids Separation

Equalization + Solids Separation

Cold Finishing

Equalization + Oil Separation or Skimming + Chemical Precipitation (heavy metals) + Neutralization

Equalization + Oil Separation or Skimming + Chemical Precipitation (heavy metals) -f Neutralization

Equalization -f Oil Separation or Skimming + Neutralization

Industrial Wastewater Sources

265

Organic Chemicals Industry Wastes Industry Description This industry sector includes the manufacture of a wide variety of products ranging from industrial gases and fertilizers to dyes, pigments, and petroleum compounds. The Organic Chemicals industry consists of such a complex combination of processes and products that a "typical or average" plant exists only in a statistical sense. The product mix and output of an industry depends primarily on the total economic activity and the demand for products. The organic chemicals industry is very dynamic in development of new products and processes. Examples of only a few of the high volume products manufactured are: l.Ethylene 2. Benzene 3. Propylene 4. Ethylene Dichloride 5. Toluene 6. Methanol 7. Ethylbenzene 8. Styrene 9. Formaldehyde 10. Vinyl Chloride 11. Ethylene Oxide 12. Xylene (Mixed) 13. Butadiene 14. Ethylene Glycol

15. Ethanol 16. Isopropanol 17. Acetic Acid 18. Cumene 19. Cyclohexane 20. Phenol 21. Acetaldehyde 22. Acetic Anyhdride 23. Terephthalic Acid 24. Dimethyl Terephthalate 25. Acetone 26. Adipic Acid 27. Acrylonitrile

Depending upon the sequence of production from petroleum sources, chemicals are referred to as either feedstocks or intermediate petrochemicals. Of the 27 example chemicals listed there are 22 intermediate chemicals and five feedstocks (i.e., ethylene, propylene, benzene, toluene, andxylene). A review of wastewater characteristics indicates that certain products can be grouped together on the basis of pollutants present in the wastewater. Accordingly, the 27 product chemicals covered under this category are divided into three subgroups as follows: Sub-Group 1 1. Benzene 2. Toluene 3. Xylene 4. Cyclohexane 5. Adipic Acid 6. Ethylbenzene 7. Styrene 8. Phenol

9. Terephthalic Acid (TPA) 10. Dimethyl Terephthalate (DMT) 11. Ethylene 12. Ethylene Dichloride 13. Vinyl Chloride (Monomer) 14. Ethanol 15. Acetaldehyde 16. Acetic Acid

266

Liquid Filtration

17. 18. 19. 20.

Acetic Anyhdride Propylene Isopropanol Acetone

21. Cumene 22. Ethylene Oxide 23. Ethylene dyed

Sub-Group 2 1. Butadiene 2. Methanol 3. Formaldehyde Sub-Group 3 Acrylonitrile The chemical reactions involved in the production of the foregoing chemicals include petroleum reforming, thermal or catalytic cracking, oxidation, alkylation, dehydrogenatton, hydration, and chlorination. Most processes use proprietary catalysts to increase product yield and to reduce severe operating conditions and pollution. Water is used extensively both in the process and for cooling purposes.

Wastewater Characteristics The characteristics of process wastewaters from the manufacture of products under each pretreatment group are shown in Table 21. The characteristics of wastewaters vary from plant to plant, according to the products and processes used. The organic chemicals plants generally operate 24 hours a day throughout the year. Depending upon the product mix and the manufacturing process, hourly variations in wastewater volume and loading may occur as a result of certain batch operations (filter washing, crystallization, solvent extraction, etc.). The wastewater collection systems are generally segregated, to permit separate collection of process wastewaters and relatively clean cooling waters. The process wastewaters are usually discharged to a common sewerage system for treatment and disposal. The process wastewaters from the manufacture of chemicals under subgroup 1 generally contain free or emulsified oil, while under subgroup 2 generally do not contain oil. Acrylonitrile manufacture (subgroup 3) produces a wastewater containing cyanides and substantial quantities of acids. These wastewaters, in general, contain unreacted raw materials and losses in products, by-products, co-products, and any auxiliary chemicals used in the process. Detailed analyses for every specific chemical present in the wastewater is difficult and are not generally used to describe the characteristics of wastes. In general the wastewaters contain: BOD, COD, oil, suspended solids, acidity, alkalinity, heavy metals, and heat. The wastewaters discharged from the manufacture of products under subgroup 1 may contain oil and grease and a series of heavy metals (Fe, Cd, Cu, Co, V, Pd). The types and amounts of heavy metals in the wastewater depend primarily on the manufacturing process and the amount and type of catalysts lost from the process.

Industrial Wastewater Sources

267

Most catalysts are expensive and, therefore, recovered for reuse. Only unrecoverable catalysts (heavy metals), generally in small concentrations, appear in the waste water. The wastewaters generated from the manufacture of products under subgroup 2 contain: BOD, acidity or alkalinity (pH in the range of 4 to 11), and heavy metals (Cr, Cu, Zn, Hg). These wastewaters are amenable to biological treatment after equalization and neutralization. The production of butadiene may produce a waste water containing free or emulsified oil; an oil separation device may be required as pretreatment when the oil content in the wastewater exceeds 50 rng/L. Only unrecoverable heavy metals (catalysts), generally in small concentrations, appear in the waste water. Table 21. Wastewater Characteristics of the Organic Chemicals Industry. Characteristics

Sub-Group 1

Sub-Group 2

Sub-Group 3

Industrial Operation Flow BOD TSS IDS

Year-round Continuous-Varia ble Average-Ext. High Low-High High

Year-round Continuous-Variable Average-High Low Low-High

Year-round Continuous-Variable Low1 High High

COD Grit Cyanide Chlorine Demand PH

Average-Ext. High Absent Absent High Acidic-Alkaline

Average-High Absent Absent High Acidic-Alkaline

High Absent Present High Acidic

Color Turbidity Explosives Dissolved Gases Detergents

Low-Average Low Absent Present Present

Low-Average Low Absent Present Present

Low Low Absent Present Present

Foaming Heavy Metals Colloidal Solids Volatile Organ ics Pesticides

Present Present Absent Present Absent

Present Present Absent Present Absent

Present Present Absent Present Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Deficient Deficient Normal-High3 Low-High Present

Deficient Deficient2 High' Present Present

Deficient Adequate No Data Absent Absent

Oil & Grease Low-High Low-High Absent Coliform (Total) Low Low Low 1 Low BOD is probably due to the toxicity characteristics of this waste. ! Adequate when butadiene is manufactured. ' Temperature equal to or higher than domestic wastewater; may affect design but not harmful to joint treatment.

The manufacture of acrylonitrile produces a highly toxic wastewater which is difficult to treat biologically. The toxicity characteristics are attributed to the presence of hydrogen cyanide in excessive quantities. In addition, the wastewater is generally acidic and contains high concentrations of organic carbon. These wastewaters are

268

Liquid Filtration

generally segregated from other process wastes and disposed of by other means (e.g., incineration). Pretreatment Table 22 shows the pretreatment unit operations which are often used. The heavy metals present in organic chemical wastes are in many cases so low in concentration that heavy metals removal is not required from the standpoint of treatability characteristics. However, the effluent limitations for heavy metals and toxic pollutants may require additional pretreatment (chemical precipitation) for removal of these materials, The pretreatment unit operations generally consist of equalization, neutralization, and oil separation. In addition, phenol recovery (to reduce the phenol concentration) and spill protection for spent acids and spent caustics may be required in some cases. Table 22. Pretreatment Unit Operations for the Organic Chemicals Industry. Pretreatment Sub-Group

Suspended Biological System

Fixed Biological System

Independent PhysicalChemical System

1

Oil Separation + Equalization + Neutralization + Spill Protection 4 Chemical Precipitation'

Oil Separation + Equalization 4 Neutralization 4Spill Protection + Chemical Precipitation 1

Equalization + Neutralization 4 Chemical Precipitation'

2

Oil Separation" 4 Equalization 4 Neutralization

Oil Separation2 4 Equalization 4 Neutralization

Equalization 4 Neutralization

1. Need for chemical precipitation depends on extent of catalyst recovery. 2. Oil separation required for butadiene manufacture only.

Meta! Finishing Industry Wastes industry Description This industry includes various types of plating, anodizing, coloring, forming, and finishing operations. The metal-finishing industry operations are related closely to those of many other industries, including transportation (automobile parts and accessories), electrical, and jewelry. The metal-finishing operation involves cleaning, conversion coating, organic coating, plating, anodizing, coloring, and case hardening. Acid pickling is the most common type of cleaning of metal being prepared for plating. Sulfuric acid is the most commonly used pickling agent, but phosphoric, hydrochloric, hydrofluoric, and other acids are used as well. Alkalies, dichromates, and numerous proprietary compounds are also used in various combinations for descaling, degreasing, stripping,

Industrial Wastewater Sources

269

brightening, or otherwise preparing different metals (zinc, steel, brass, copper, etc.) for plating or anodizing. The plating solutions for nickel, chromium copper, cadmium, zinc, tin, and silver may be basically cyanide, acid, or alkaline. Anodizing is done either in sulfuric acid or in chroniate solutions. Colorizing is accomplished with dyes, nickel acetate, and chromates. Cyanides are used in case hardening. Wastewater Characteristics The characteristics of the process wastewaters from the industry are shown in Table 23. The metal-finishing industry usually generates a continuous stream of rinse waters containing dilute concentrations of heavy metals and cyanide and intermittent batch dumpings of spend acid and cleaning solutions. The nature of metal-finishing operations and the consequent fluctuating (cyclic) characteristics of the wastewater should be taken into consideration in the design of treatment facilities. Table 23. Wastewater Characteristics of the Metal Finishing Industry . Characteristics Industrial Operation

Year-round (BATCH)

Flow

Continuous-Variable

BOD TSS IDS COD

Average-High High

Low Low

Color

Present High High Acidic Present

Turbidity Explosives Dissolved Gases Detergents Foaming

Present Absent Present Present Absent

Heavy Metals Colloidal Solids Volatile Organics Pesticides

High Absent Present Absent

Phosphorus Nitrogen Temperature Phenol Sulfides

Present Present Neutral

Oil & Grease Coliform (Total)

Present Absent

Grit Cyanide Chlorine Demand

PH

Low Absent

270

Liquid Filtration

Water is used extensively in metal-finishing processes to clean, strip, pickle, and rinse the metal products before and after plating operations. The rinse waters constitute the major volume of wastewaters, while spent solutions discharged intermittently add major pollutants to the total effluent. The wastewaters contain, in general, spent acids, alkalis, oil and grease, detergents, cyanides, and various heavy metals (Cr, Ni, Cu, Ag, Fe, Zn, and Sn). The metal-finishing plants differ from one another with respect to their processes, metals, and chemicals, and the characteristics of waste water may vary widely from one to another. However, their wastewaters all contain primarily inorganic pollutants, particularly heavy metals. In addition, the wastewaters frequently are highly toxic due to the presence of cyanides and heavy metals. In general, the types of wastewaters from metal-finishing industries are: 1. Acid wastes 2. Alkaline wastes 3. Heavy metals wastes 4. Cyanide-bearing wastes 5. Miscellaneous wastes (dyes, soluble and floating oils, etc,) Any of these wastewaters may occur as either dilute rinse waters or concentrated baths. Except for the cyanide-bearing wastes, the wastewaters are generally connected to a common sewerage system for treatment and disposal. The cyanide wastes usually are collected in a segregated sewer system in order to prevent the release of toxic hydrogen cyanide gas under acidic conditions. However, the cyanide wastes can be mixed with other waste streams provided that any acid streams are neutralized prior to mixing with the cyanide waste stream. The major constituants in the wastewaters generated from metal-finishing operations are cyanides, metal ions, (Cr 6+ , Ni, Fe, Cu, Ag, and Sn), oil and grease, organic solvents, acids, and alkalis. The wastewaters characteristically are so toxic and corrosive to sewers and equipment that they require pretreatment before discharge to municipal sewers. A wide variety of processes are used in metal finishing operations, resulting in widely varying wastewater characteristics. Typically, these wastewaters have poor treatability characteristics without adequate pretreatment. Pretreatment The pretreatment unit operations for various types of treatment facilities are shown in Table 24, The pretreatment processes generally involve separate treatment of cyanide wastes and other acid wastes containing metal ions. The cyanide wastes can be treated with ferrous sulfate and lime to convert highly toxic cyanides to less toxic cyanates or cyanide complexes, or can be oxidized to CO2 and N2 with chlorine under alkaline conditions. The acid waste streams are treated first to reduce hexavalent chromium to trivalent chromium, using ferrous sulfate, scrap iron, or sulfur dioxide, and then precipitating the metal ions (Cr3+) as metal hydroxides.

Industrial Wastewater Sources

271

Table 24. Pretreatment Unit Operations for the Metal Finishing Industry. Suspended Biological System Equalization + Neutralization + Cyanide Removal + Chromium Reduction + Chemical Precipitation (Heavy Metals) + Solids Separation

Fixed Biological System

Independent PhysicalChemical System

Equalization 4- Neutralization + Cyanide Removal + Chromium Reduction + Chemical Precipitation (Heavy Metals) + Solids Separation

Equalization +Cyanide Removal + Chemical Precipitation + Neutralization

1. Chemical precipitation may not be needed, depending on the processes used in the independent physical chemical joint treatment plant.

In addition to the effluent limitations and the processes shown in Table 24 the degree of reduction in heavy metals waste loadings should consider the sludge handling and disposal methods used for the metal finishing wastewaters. Some processes (e.g., anaerobic digestion) concentrate these metals, and this can lead to process failure unless adequate pretreatment is provided. Dewatering operations discussed in Chapter 7 should be referred to.

Closure Filtration is an important unit operation hi the treatment of many industrial waste streams. In the industry sources described in this chapter, filtration is often relied upon for the removal of coarse and fine particles, and as both a pretreatment and posttreatment step. One must recognize however, that most filtration operations involve physical separation, or in some instances, are combined as a part of biological treatment or with chemical pretreatment methods as described in earlier chapters. Because many of the waste streams described are highly toxic in nature, chemical techniques including extraction, precipitation and others are the primary methods of treatment. The references cited below provide more in depth coverage of this subject along with additional examples on the use of filtration with chemical treatment methods.

Suggested Readings

1. Profile of the Fabricated Metal Products Industry, US EPA Document 310-R-95007, September 1995 2. Profile of the Non-Fuel, Non-Metal Mining Industry, US EPA Document EPA 310-R-95-011, September, 1995. 3. Profile of the Stone, Clay, Glass, and Concrete Industry, US EPA Document 310R-95-017, September 1995. 4. Profile of the Metal Mining Industry, USEPA Document 310-R-95-008, September 1995. 5. Profile of the Iron and Steel Industry, US EPA Document 310-R-95-005, September 1995.

FILTRATION EQUIPMENT AND PROCESS FLOW SHEETS introduction This chapter provides a compendium of filtration machinery and auxiliary equipment, along with typical process flow sheets in diagrammatic form. The principle design features and configurations for different filter machines are summarized along with examples of flocculators, chemical feedstock systems for conditioning operations, and centrifuges. The material presented in this chapter is designed to aquaint the reader with generic design configurations, operating principles, mode or scheme of operation, and typical process flow sheets. This collection of schematics and process flow sheets will assist the newcomer in establishing preliminary design concepts and process flow system layouts. The author has been careful not to discuss specific manufacturer's equipment because it is not the intent of this book to provide specific endorsements or recommendations for suppliers or vendors. Because of the overwhelming numbers, types, and variations of filtration equipment, not all commercially available systems or process schemes are included, however, the reader will find many examples of the most commonly used systems throughout industry. To effectively use this chapter, the reader should first review the list of schematics below, and then turn to the page and figure number on which the drawing appears. Index to Equipment and Flow Sheet Diagrams Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. 272

Design details of a Rapid Sand Filter Details of a typical Vertical Leaf Vacuum Filter Details of a typical Vertical Leaf Pressure Filter with Vertical Tank Orientation Common Filter Operating Configurations Cross section details of an Upflow Filter System Details of a typical Pressure Filter

274 275 276 277 278 279

Filtration Equipment and Process Flow Sheets

Figure 7. Figures. Figure 9. Figure 10. Figure Figure Figure Figure Figure Figure Figure

11. 12. 13. 14. 15. 16. 17.

273

Flow Control Systems : Influent flow splitting and variable declining rate filtration , Design details of Underdraws for Sand Filters Cutaway view of a Rotary Drum Vacuum Filter Illustration of the Cake Processing Phases of a Rotary Vacuum Filter .' Example of a Laboratory-Scale Vacuum Filter Apparatus . . . . Process flow sheet for a typical Vacuum Filter Process flow sheet for a typical Rotary Vacuum Filter System . Illustration of the Operating Zones of a Vacuum Filter Diagram showing the cross-section details of a Coil Filter . . . Diagram illustrating various Classifications of Centrifuges . . . Diagram of a Disc Type Centrifuge

Figure 18.

Illustration of a Continuous Countercurrent Solid Bowl Conveyor and Discharge Centrifuge

Figure 19.

Illustration showing cross-section of a Countercurrent Flow Solid Bowl Centrifuge Schematic diagram of a Basket Centrifuge Typical process flow sheet for a Filter Press System . . . . . . . Side view of a Filter Press Cutaway view of a Filter Press Cross section of a Belt Filter Example of a Moving Screen Concentrator System Example of a typical Microstrainer Unit Schematic of a Moving Bed Filter Illustrates an approach to upgrading a Low-Rate Trickling Filter to a High-Rate Trickling Filter Wastewater Treatment System after upgrading a two-stage trickling filtration system Illustrates upgrading a single-stage trickling filter to a Two-Stage Filtration System Illustrates upgrading a high-rate trickling filter using a super-rate trickling filter as a Roughing Unit Common flow diagrams for single and two-stage High-Rate Trickling Filters Chart listing types of Chemical Feeders Example of a typical Dry Feeder System Example of a typical Lime Feeder System Example of a Flocculant Diagram , Example of a Manual Dry Feeder System Example of an Automatic Dry Polymer Feed System Example of a Caustic Feed System Examples of alternative Liquid Feeder Systems for overhead and ground storage Examples of a Mechanical Flocculation Basin and Flocculator Example of a Rotary Drum Conditioner

Figure 20. Figure 21. Figure 22. Figure 23. Figure 24. Figure 25. Figure 26. Figure 27. Figure 28. Figure 29. Figure 30. Figure 31. Figure 32. Figure 33. Figure 34. Figure 35. Figure 36. Figure 37. Figure 38. Figure 39. Figure 40. Figure 41. Figure 42.

280 281 282 283 284 285 286 287 288 289 290

. 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315

Rate of flow and toss of head

Filter bed washwater troughs

Operating floor Concrete fitter tank

Pipe gallery floor

Pressure lines to hydraulic vafoes from operating tables Influent to filters

Effluent to clear well

Drain

Perfora laterals Cast-iron, manifold

Figure 1. Design details of a rapid sand filter.

276

Liquid Filtration

Figure 3. Details of a typical vertical leaf pressure filter with vertical tank orientation.

Filtration Equipment and Process Flow Sheets

- OVERFLOW TROUGH EFFLUENT

i INFLUENT

30-402UNDEHDRAIN— CHAMBER INFLUENT

EFFLUENT (a), CONVENTIONAL FILTER

0»), UPFLOW FILTER

(c), U-aOW FILTER

SINGLE MEDIA FILTERS

21-48"

30-40"

GAR NET SAND

(d), DUAL MEDIA FILTERS

{*),MlXID-WiD»A FILTERS CTR1M.E MEDIA)

Figure 4. Common filter operating configurations.

277

278

Liquid Filtration

COVER OPTIONAL (FOR CLOSED SYSTEM)

FILTRATE OUTLET "GRID"

DEEP SAND LAYER

SAND "ARCHES

GRAVEL LAYERS

i SPECIAL VENT

INLET

RAW WATER WASH WATER

AIR FOR SANDFLUSH CLEANING

Figure 5. Cross section details of an upflow filter system.

Filtration Equipment and Process Flow Sheets

I COUPLING AIR RELEASE 50PSIS PRESSURE VESSEL-^.

10 FLAN6E INFLUENT BACKWASH WASTE

2"FLANQE SURFACE WASH MEDIA o a a a a a

12 « IS" MANHOLE ON VERTICAL £ OF TANK

10 FLANQf EFFLUENT AND BACKWASH

o o Q a o o o

2 FILTER DRAIN FILTER SUPPORTS AT 1/4 POINTS ELEVATION

B'-0"0,0, DISTRIBUTOR-

12"* l«" MANHOLE SURFACEWASH MIXED MEDIA SUPPORT 0RAVEL CONCRETE LATERALS

S ECTION Figure 6. Details of a typical pressure filter.

279

280

Liquid Filtration

Flow SpliiiiAg T»nk

•I

f»ltr CM

"I

_*.

finer c«« eillutn! irflttw

OB s

Wttti Trough

wunTraygn

o[ o } M«BM*I

1

"•

f i«« Cta No. J

1

r,iw CM

Open

ell

O

V

41"

Claud

noKtina

1

»^p

Claud '

I

Manujl IKilalin^ Vllv*

I I *"

INFLUENT FLOW SPLITTING

COMMON IMn.UINTHE«OEII PIK OH CHAMHCL

oKinee IH.ATC on SHORT VCNTIMI ran NATtWOI CATION ONL*.

VARIABLE DECLINING RATE FILTRATION

Figure 7. Flow control systems : influent flow splitting and variable declining rate filtration.

O

Filtration Equipment and Process Flow Sheets

A. HEADER LATERALS (COUHfTESf OF THE *WWA)

•/•' DM COMTttOL WWICI1

Ox.tFC* M. rr.

B.

LEOPOLD

BLOCK

(Covrttty

r. B- Leopold

SYSTEM Co-)

Figure 8. Design details of underdrains for sand filters.

281

282

Liquid Filtration

CLOTH CAULKING STRIPS DRUM FILTRATE PIPING

CAKE SCRAPER

AIR AND FILTRATE

AIR BLOW-BACK LINE

SLURRY FEED

Figure 9. Cutaway view of a rotary drum vacuum fitter.

VACUUM FilTER MEDIA CHEMICAL CAKE

CONCRETE

COMB

CONVEYOR

BELT S7«

Figure 10. Illustration of the cake processing phases of a rotary vacuum filter.

SLUDGE

284

Liquid Filtration

ALUMINUM TUBE PERFERATED SUPPORT PLATE

GASKET

LEVELING SCREW VACUUM TIGHT VALVE

TO VACUUM

^ r~ 1

L_1

D

u

1

1

•*• i i ***•

s

V

•«•

"»

—

CALIBRATED * PLEXIGLASS TUBE WITH DRAIN

—.

,»-—

s ^. X. \ ^ pHHrt

^^

Figure 11. Example of a laboratory-scale vacuum filter apparatus.

riLTSITE REfURH TO PUN?

RECEIVER COAGULANT

SLUDGE

91 c»

POLYMER

o

V

^

FILTMTE

FLO* CONTROL

9

9,

s-•*

AIR TO ATMOSPHERE

I

J ^

SILENCER WATER TO PLANT

SLUDQE CONDITIONING TANK*

COUVEIfOB

FILTRATE PUMP

Figure 12. Process flow sheet for a typical vacuum filter.

HATER

VACUUM PUMP

A** TO ATMOSTHHf

SlUNCft

\

VACUUM PUMP

SlUOGE

Figure 13. Process flow sheet for a typical rotary vacuum filter system.

Filtration Equipment and Process Flow Sheets

Figure 14. Illustration of the operating zones of a vacuum fitter.

287

COIL SPRING FILTER MEDIA WASH WATER SPRAY PIPING

DRUM

VACUUM AND FILTRATE OUTLETS

CAKE DISCHARGE AGITATOR DRIVE

Figure 15. Diagram showing the cross-section details of a coil filter.

FUtration Equipment and Process Flow Sheets

r

GEAR BOX

D R I V E SKEIV? FEED

SOLID BOWL CENTRIFUGE

FIID CAKE DISCHARGE «-

• C L A f l l F J E D EFFLUEMT

— FEED i

» EFFLUENT DISCHARGE

LUOOI DI9CMAHOI

DISC TYPE CENTRIFUGE Figure 16. Diagram illustrating various classifications of centrifuges.

289

290

Liquid Filtration

Figure 17. Diagram of a disc type centrifuge.

COVER

DIFFERENTIAL SPEED GEAR BOX

MAIN DRIVE SHEAVE

"~7- - FEED PIPES ~^ (SLUDGE AND CHEMICAL)

U

ROTATING CONVEYOR

CiNTRATC DISCHARGE

JS

E.

SLUDGE CAKE DISCHARGE

Figure 18. Illustration of a continuous countercurrent solid bowl conveyor and discharge centrifuge.

292

Liquid Filtration

•

Filtration Equipment and Process Flow Sheets

FEED POLYMER! SKIMMINGS

CAKE

CAKE

Figure 20. Schematic diagram of a basket centrifuge.

293

10

•c CL

11

1 Sludge In 2 Mechanical Screen 3 Sludge Storage Tank

4 Chemical Storage Tank 5 Chemical Measurement and Dilution Tank

6 Chemical Pumps 7 Conditioning Tank 8 Sludge Pumps

Figure 21. Typical process flow sheet for a filter press system.

9 Filter Presses 10 Cakes Out 11 Filtrate Drain

FIXED END

TRAVELLING END

ELECTRIC CLOSING GEAR

OPERATING HANDLE,

i

Figure 22. Side view of a filter press.

296

Liquid Filtration

FILTER ClOTHS FIXED END

SLUDGE IN

FILTRATE DRAIN HOLES

Figure 23. Cutaway view of a filter press.

Filtration Equipment and Process Flow Sheets

.DRUM

DISCHARGE ROIL

DISCHARGE ZONE

WASH ROU WASH TROUGH

Figure 24. Cross section of a belt filter.

297

298 Liquid Filtration

Filtration Equipment and Process Flow Sheets

D R I V E UNIT

WASH WATER IETS EFFLUENT WEI*

EFFLUENT CHAMBER iHFLUEHT CHAMBER

Figure 26. Example of a typical microstrainer unit.

299

300

Liquid Filtration

INFLUENT

I I A J H WH imt

DISCHIRCt

Figure 27. Schematic of a moving bed filter.

FUtration Equipment and Process Flow Sheets

TRICKLING FILTER

SECONDARY CLARIFIES 'FINAL EFFLUENT

PRIMARY EFFLUENT 185,OOO SLUDGE

TREATMENT SYSTEM BEFORE UPGRADING LOW-RATE TRICKLING FILTER

EXISTING TRICKLING FILTER

NEW RECIRCULATION PUMPING STATION EXISTING SECONDARY CLARIFIER •FINAL EFFLUENT

PRIMARY EFFLUENT -*-y 370.010 GPD

ADDITIONAL REQUIRED CAPACITY

^CIRCULATION 185.000 SLUDGE

TREATMENT SYSTEM AFTER UPGRADING Figure 28. Illustrates an approach to upgrading a low-rate trickling filter to a high-rate trickling filter.

301

302

Liquid Filtration

INTERMEDIATE-RATE TRICKLING FILTER

FINAL EFFLUENT

PRIMARY EFFLUENT 6.0 MGD

SECONDARY CLARIFIES

TREATMENT SYSTEM BEFORE UPGRADING SINGLE-STAGE INTERMEDIATE-RATE TRICKLING FILTER

REC1RCBLATION 7.5 HGD HE* RECIRCULATION PUMPING STATION SLUDGE

PRIMARY EFFLUENT 6.0 MSB 1 ST. STAGE-NEW HIGH-RATE FILTER

HE1 INTERMEDIATE CLARIFIES

2ND STAGE-EX 1ST INC INTERMEDIATE-RATE FILTER

EXISTING SECONDARY CLARIFIER

Figure 29. Wastewater treatment system after upgrading a two-stage trickling filtration system.

Filtration Equipment and Process Flow Sheets

RECIRCUUTION 6.0 KGO SLUDGE PR1«ART EFFIUEST 2.0 MCO

FIHJtl EFFLUENT TRICKLING FILTER

SECQHUW CHRIFIER

TREATMENT SYSTEM BEFORE UPGRADING HIGH-RATE TRICKLING FILTER

r-NE» \ COMPLETELY MIXED/ \ UHADOH TANK

/

A

PR mm

•EXISTING SECONDARY

FINAL EFFLUEMT

EFFUHHT—A

2.0 MOD

EXISTING TRICKLIMO FILTER

1 1 u.- _

1 1 t -«.T

HEN 100* SLUDGE REClfCLE FHILITIES

Figure 30. Illustrates upgrading a single-stage trickling filter to a two-stage filtration system.

303

304

Liquid Filtration

HECIRCUUTIflH 6,8 HCD

RECIRCUUT10N PUMPIHS S T A T I O N

PRIKH8Y EFFLUENT, 2.0 MGD

FIHAL EFFLUENT TRICKLING FILTER

SECONDARY CLARIFIES

TREATMENT SYSTEM BEFORE UPGRADING HIGH-RATE TRICKLING FILTER

NEW RE CIRCULATION PUMPING STATION EXISTINC

NEW

RECIRCULAIIOH PUMP I KG STATION

SYNTHETIC MEDIA ROUGHfm FILTER PRIMiY EFFLUENT 2.0 UGD

FINAL EFFLUENT

1.S MGO

EXISTING TRICKLING FILTER

EXISTING S ECO HO ART CLARIFIED

TREATMENT SYSTEM AFTER UPGRADING ROUGHING FILTER PRECEEDING EXISTING HIGH-RATE FILTER

Figure 31. Illustrates upgrading a high-rate trickling filter using a super-rate trickling filter as a roughing unit.

Filtration Equipment and Process Flow Sheets

SINGLE-STAGE R

i s

c —*i

K> Tia-STAGE

LE6EKO $ SLOOCE RETWN

i

KCitetutiB FUJI

"m: -ars

O FHIMWY CUItrtEl O TRICKL1N8 FILTER IKTERIIEOUTE CUKIFIEU FINil CIMIFIE*

Figure 32. Common flow diagrams for single and two-stage high-rate trickling filters.

305

306

Liquid Filtration

TYPES OF CHEMICAL FEEDERS Type of Feeder

Use

General

Limitations Capacity cu ft/hr

Range

0.01 to 35

40 to 1

0.02 to 100

40 to (

0.01 to 1 0

20 to 1

8 to 2 000 or 7.2 to 300 005 to 18

10 to i or 100 to I 20 to 1

Dry feeder: Volumetric Oscillating pbte Oscillating throat (universal) Rotating disc

Any material, granules or powder. Any material, any particle size. Most materials including NaF, granules or powder.

loader for arching.

Rotation cylinder (star) . . . . powder. Screw powder or granular. Dry, free flowing material, powder, granular, or lumps. Dry, free flowing material up to I'/i-inch size, powder or granular.

Ribbon. B«H .

Gravimetric: Continuous-belt and scale

Lou in weigh; Solution feeder: Nonposilivf displacement: Decanter (lowering pipe) - - Orifice Rotameter (calibrated valve)

Lou in weight (tank with control valve). Positive displacement: Rotating dipper Proportioning pump: Diaphragm

10 to i or 100 to 1

0.02 to 2

100 to 1

Most materials, powder, granular or lumps.

0.02 to 80

100 to 1

Most solutions or light slurries

0.01 to 10 0. 1 6 to 5 0005 to 0.16 or 0.01 to 20 0.002 to 0.20

100 to 1 10 to ! 10 to 1

Dry, free (lowing, granular material, or floodable material.

Most solutions or slurries . . . .

Use hopper agitator to maintain constant density.

Chlorine Sulfur dioxide Carbon dioxide

Carbon dioxide

30 to 1

100 to i

0. 1 to 30

0.004 to 0.1 5

Direct feed.

1

SO to i

O.I to 3,000

for i% slurries.' Most solutions, light slurries. .

Piston G« feeders Solution feed

0.002 to 0.16....

.

0.01 to 170

8000 Ib/day max 2000 Ib/day max 7600 Ib/day max 6000 lb/d«y max 300 Ib/day max 1 20 Ib/day max 1 0.000 Ib/day max

Use special heads and valves for (lurries.

Figure 33. Chart listing types of chemical feeders.

100 to i

20 to 1 20 to 1 20 to 1 20 to i 20 to 1 10 to 1 7 to 1 20lo 1

Filtration Equipment and Process Flow Sheets

COLLECTOR PIPE (PNEUMATIC)

DUST COLLECTOR DAY HOPPER FOR DRY CHEMICAL FROM BAGS OR DRUMS

6 FILL SCREEN WITH BREAKER

BIN GATE FLEXIBLE CONNECTION ALTERNATE SUPPLIES DEPENDING ON STORAGE SCALE OR SAMPLE CHUTE FEEDER DRAIN SOLENOID VALVE CONTROL VALVE __ GRAVITY TO APPLICATION

PUMP TO APPLICATION

Figure 34. Example of a typical dry feeder system.

307

308

Liquid Filtration

DUST COLLECTOR FILL PIPE (PNEUMATIC)

NOTE: VAPOR REMOVER HOT SHOWN FOR CLARITY

UN GATE

OR SAMPLE CHUTE

FLEX I8LE CONNECTION FLOW RECORDER WITH PACIW TRANSMITTER,

ROTAMETERS SUKINQ NfcTEK

GRAVITY FEED RECIRCUUTION BACK PRESSURE VALVE

Figure 35. Example of a typical lime feeder system.

TO CENTRIFUGE 1V

•1

-I

7

f

FLOC MIXING EDUCTOR FUNNEL

ROTOMETER

FLOCCULANT FEED PUMP

FtOCCULANT A MIXING TANK

Figure 36. Example of a flocculant diagram.

FRESH WATER OR PLANT EFFLUENT

310

Liquid Filtration

-DRY FEEDER DISPERSER

WATER SUPPLY MIXER

DISSOLVIHQ-AQIHQ TANK

HOLD IMG TANK

SOLUTION FEEDER

POINT OF APPLICATION

Figure 37. Example of a manual dry feeder system.

Filtration Equipment and Process Flow Sheets

3! 1

-I ft,

312

Liquid Filtration

TRUCK FILL LIKE

VENT, OVERFLOW AND DRAIN SODIUM HYDROXIDE STORAGE TANK

DILUTION WATER

POINT OF APPLICATION Figure 39. Example of a caustic feed system.

Filtration Equipment and Process Flow Sheets

313

OVERHEAD STORAGE TANK

CONTROL VALVE

FLOAT

5~-_

f

c

0

'ROTAMETER

"METERI KG PUMP

N

R0TODIP-TYP£ FEEDER

GRAVITY FEED

G R A V I T Y FEED

GRAVITY FEED

PRESSURE FEED , CONTROL V A L V E

,-ROTODIP-TYPE FE EDER

/

X

n

Hf—

!$£?>— r t^gjhy

/i

r

ROTAMETER

i •
3*-

_i

*-

!f

uu

It *•

o

**=

£

QC 0

X

r

5K

ii / RECIRCUL

VITY FEED

mo*

f /

-> TRANSFER PUMP -J ^

(

x 'uj

GRt)UND STORAGE T4 NX

14J

o _^

D

H 1

n

5

i I PRSURE «

°^

T

»— uJ

! 7

UJ

*

t

C3

1

f

FE ED ,

^^

o

-* m

f K, '

Figure 40. Examples of alternative liquid feeder systems for overhead and ground storage.

314

Liquid Filtration

MECHANICAL FLOCCULATION BASIN HORIZONTAL SHAFT-REEL TYPE MOTORIZED SPCtO NKDUCtft

NANOftAIL

auioe WATER PRESSURE LUBRICATED

MECHANICAL FLOCCULATOR VERTICAL SHAFT-PADDLE TYPE

Figure 41, Examples of a mechanical flocculation basin and flocculator.

Filtration Equipment and Process Flow Sheets

CHEMICAL FEED CONNECTION

CONDITIONING TANK FEED CHUTE CONDITIONING TANK SUPPORT FILTER VAT {REAR SIDE)

Figure 42. Example of a rotary drum conditioner.

315

INDEX

coke media, 42 colloidal clays, 50 colloids, 15 compressible cakes, 63 compression-permeability cell, 74 concentrate disposal, 183-184 cocurrent filters, 91 connectivity, 5 constant pressure drop filtration, 7581 constant pressure filtration, 66 constant rate filtration, 57, 70-72, 83-86 contaminated groundwater, 171-172 corrosion resistance, 16 cotton cloth chemical resistance, 2223 cotton cloth filters, 22-23 cross flow filtration, 170, 181, 195 cross mode filters, 98-103 crushed stone media, 43 cyclones, 227

backflushing, 132-134 backwashing, 155 bed regeneration, 148-149 belt filters, 96-98, 297 belt filter presses, 212 beverage industry wastes, 254-258 biological activity, 147

cake filtration, 59-70 cake filtration dynamics, 60-70 cake formation, 11, 17, 62 cake permeability, 50 cake properties, 13, 46 cake resistance, 65 cake specific resistance, 45 cake structure, 61 cake volume, 65 cartridge filters, 103-108 cellulose, 51 centrifugal filtration, 120, 122-123 centrifugation, 223-226 centrifuges, 120, 215, 289-293 ceramic filter media, 41 chamber filter presses, 215 charcoal, 42, 51 chemical feeders, 306-313 chemical mixing, 155, 156, 159 chemical sizing, 238 clarifiers, 160-161 cloth media, 45 coagulation, 16, 144 coagulation filtration, 150 coal media, 41,42 cocurrent filters, 91-98 coil filter, 288

dairy industry wastes, 232, 235-237 Darcy's law, 3-4, 64 depth type filter media, 19 dewatered cakes, 62 dewatered solids, 211 de watering operations, 12, 13 dewatering technologies, 212-217, 226 diaphragm filters, 110-115 diatomaceous earth, 41, 42 diatomite, 50 diffusion, 146 316

Index

disc filters, 101, 102 drainage beds, 217-219 drum vacuum filters, 89-90 drying beds, 217 drying operations, 14 dynamic thickeners, 118-119, 121

ebonite media, 41 electrokinetic forces, 11 electrostatic attraction, 146 equipment selection, 13, 16-18

fabric filter media, 20 fabric surface properties, 21 fiber cloth filter media, 23 filter aid applications, 48-50 filter aid efficiency, 53-55 filter aid precoating, 56 filter aid requirements, 48 filter aid selection, 51-57 filter aids, 15, 19, 20, 47-57 filter cake resistance, 65 filter centrifuge, 122 filter media, 11, 19 filter media filtration, 59 filter media selection criteria, 43-47 filter media washing, 88 filter medium, 88 filter medium selection, 76 filter plate, 66 filter presses, 92, 98-100, 294-296 filtrate, 8 filtrate motion, 88, 91 filtrate quality, 77 filtration classification, 10 filtration conditions, 12 filtration constants, 67 filtration cycles, 113 filtration equipment, 13, 88 filtration formulas, 75 filtration mechanisms, 81-83 filtration rate, 10 filtration tests, 49 filtration time, 74 finishing industry wastes, 243-246, 247

317

fixed rigid media, 34 flexible filter media, 20-34 flexible metallic media cloths, 25 flocculation, 144, 155, 156, 157-158 flocculation filtration, 149 flocculation units, 157 flow control systems, 280 fly ash, 51 foam plastic media, 41 food industry wastes, 251-254 fouling, 166 fouling control, 180

glass cloths, 21-22 granular media filtration, 142-148 gravity forces, 10, 60 gravity thickening, 215 grizzlies, 217 groundwater remediation, 175

heat-resistant filter media, 26 horizontal rotary filters, 93, 95-96 hydraulic classifiers, 215, 216 hydraulic conductivity, 4 hydraulic resistance, 44, 50, 64 hydroclones, 216 hydrogen peroxide, 149 hyperfiltration, 163, 173-190

impeller mixing, 157 impoundment basins, 215, 216 incompressible fluids, 9 incompressible systems, 68 internal rotary drum filters, 92-93

kinematic viscosity, 148 Kozeny constant, 6

leaf filters, 100-101 leather tanning industry wastes, 243246, 247 loose rigid media, 41

318

Liquid Filtration

manmade packed media, 8 materials handling requirements, 182-183 mechanical flocculation basin, 158 membrane filtration, 163 membrane processes, 172 membrane separations, 169-170 mesh size, 34 metal finishing industry wastes, 268271

metallic cloths, 25-26 metallic filter media, 34, 40 methyl phenols, 165 microporous filtration range, 126 microstrainer units, 299 model pore sizes, 2

nitrated cotton cloths, 22 nitrogen compounds, 229 nonmetallic cloths, 25-26 nonwoven filter media, 26, 34 Nutsch filters, 93, 96 nylon cloth, 24

packing arrangement, 2 paper production wastes, 230-232, 233-234 paper pulp, 23 particle bridging, 19-20 particle classification, 215-217 particle settling, 85 particle size distribution, 11 perlite, 50 permeability, 4, 8, 52 petroleum hydrocarbon hyperfiltration, 196-202 petroleum refinery industry wastes, 246-247, 248-251 pH control, 16 pharmaceutical industry wastes, 240243 pilot plant filter assemblies, 17 plastics industry wastes, 258-260 pollution control, 1 pollution prevention, 1 polychlorinated biphenyls, 165

polyvinyl chloride, 41 pore blocked filtration, 86 pore blocking, 84, 85 pore clogging, 12, 60 pore size distribution, 45 pore structure, 2 porosity, 61, 62 porous media, 1, 2, 3-9, 34 porosity of perlite, 50 potassium permanganate, 149 powdered metal, 34 precoat applications, 15-16, 52 precoat filters, 221 precoating, 49 preconditioning, 150 pressure drop across cakes, 64 pressure filter, 279 pressure filtration, 222 process economics, 184-193 pulping process wastes, 231 pumping wells, 175 purification, 14

rapid sand filters, 145, 274 rapid sand filtration, 153-155 reservoir models, 3 retentivity, 44-45 reverse osmosis, 134, 135-141 Reynolds number, 3 rigid filter media, 34, 40-43 rotary disc vacuum filter, 102 rotary drum conditioner, 315 rotary drum efficiency, 55 rotary drum filters, 49, 89-91, 94 rotary drum vacuum filter, 282-283 roughing units, 304 rubber media filters, 24

sand and gravel media, 42 sand drying bed, 218 sand filters, 281 sawdust, 51 screw presses, 123-124, 125 sedimentation, 144-145 semivolatile contaminants, 204-206 slow sand filtration, 151-153

Index

sludge blanket filtration, 161 sludge dewatering operations, 211228 solid bowl centrifuges, 224 solids contact processes, 155, 159 solids recovery, 14 solids washing, 120 specific capacity, 17 specific resistance, 55 specific volume, 65 spiral wound membranes, 139-140 steel industry wastes, 261-264 strainers, 109-110 straining operations, 146 surface-type filter media, 19 suspended solids, 229 suspension properties, 13 Sweetland pressure filter, 102 synthetic fabrics, 238 synthetic fiber cloths, 24-25

temperature control, 16 tensile strength, 45 textile industry wastes, 237-240 theory of pore packing, 2 thickeners, 116-119 thin-cake filters, 115, 116-117 thin-cake thickener, 116 tortuosity, 6 total dissolved solids, 229 total organic carbon, 229 total suspended solids, 229 tubular ultrafiltration, 133

UF membranes, 124, 126 -134 ultrafiltration, 124, 126-134 unit operations, 1 upflow filter system, 278

vacuum filter operation, 220 vacuum filters, 46, 285-287 vacuum filtration, 219 vacuum rotary filtration, 215 variable pressure filtration, 72-74 variable rate filtration, 72-74 vertical leaf vacuum filter, 275 volcanic glass media, 50 volume reduction, 179

washing techniques, 12 waste shipping, 190-191 wastewater sources, 229-271 waste water treatment technology, 142, 163 wool cloth filter media, 23

319

This page intentionally left blank

Other Books from Butterworth-Heinemann Adsorption Calculations and Modelling by Chi Tien 1995 288pp he 0-7506-9121-2 Chemical Process Equipment Design by Stan Walas 1988 755pp he 0-7506-9385-1 Colloid and Surface Engineering Applications in the Process Industries by R.A. Williams 1994 240pp pb 0-7506-1940-6 Engineering Processesfor Bioseparations by Lawrence Weatherby 1994 25Qpp he 0-7506-1936-8 Purification of Laboratory Chemicals, Fourth Edition by W.L.F. Armarego and D.D. Perrin 1997 512pp pb 0-7506-3761-7 Detailed information on these and all other BH Engineering titles may be found in the BH Engineering catalog (Item #725), To request a copy, call 1-800-366-2665. You can also visit our web site at: http://www.bh.com These books are available from all good bookstores or in case of difficulty call: 1-800-366-2665 in the U.S. or +44-1865-310366 in Europe.

E-Mail Mailing List An e-mail mailing list giving information on latest releases, special promotions, offers and other news relating to BH Engineering titles is available. To subscribe, send an e-mail message to [email protected] Include in message body (not in subject line): subscribe bh-engineering