mm 
 
LIBRARY OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 AT URBANA-CHAMPAIGN 
 
 £28 
 Li 65 c 
 no. 57-58 
 
 tuGKEwe uwwj 
 
 ENGINEERING 
 
CTClHEERHfc IRRMW 
 
 The person charging this material is re- 
 
 >pon 
 
 >r its return on or before tin- 
 
 Latest Date stamped below. 
 
 Theft, mutilation, and underlining of books 
 are reasons for disciplinary action and may 
 result in dismissal from the University. 
 
 UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN 
 
 L161— O-1096 
 
* s-fi 
 
 
 UILUENG71 2012 
 
 CIVIL ENGINEERING STUDIES 
 
 SANITARY ENGINEERING SERIES NO. 58 
 
 ENGINEERING LIBRARY 
 UNIVEH 
 
 URBANA, ILLINOIS &1U01 
 
 MECHANISMS OF SLUDGE THICKENING 
 
 HflRHHtt 
 
 QQM 
 
 o. 
 
 
 By 
 RICHARD I. DICK 
 
 Supported By 
 FEDERAL WATER QUALITY 
 
 ADMINISTRATION 
 RESEARCH GRANT 17070 DJR 
 
 DEPARTMENT OF CIVIL ENGINEERING 
 
 UNIVERSITY OF ILLINOIS 
 
 URBANA, ILLINOIS 
 
 FEBRUARY, 1971 
 
MECHANISMS OF SLUDGE THICKENING 
 
 by 
 
 Richard I . Dick 
 
 Professor of Civil Engineering 
 
 Univers i ty of 111 inoi s 
 
 FINAL PROJECT REPORT 
 
 A Summary of Research Conducted 
 
 under Research Grant 17070 DJR from 
 
 the Federal Water Quality Administration 
 
 Environmental Protection Agency 
 
 Urbana , 111 inoi s 
 February, 1971 
 
Digitized by the Internet Archive 
 in 2013 
 
 http://archive.org/details/mechanismsofslud58dick 
 
ABSTRACT 
 
 Two areas were emphasized in this research on gravity thickening of 
 sludges. One was investigation of fundamental thickening properties of 
 sludges. The other was consideration of rational criteria for thickener 
 design and operation consistent with the observed fundamental thickening 
 properties. Most results have been presented in detail in the professional 
 literature and they are summarized and interrelated in this report. 
 
 The batch flux curve method of thickener analysis using settling data 
 obtained with alternative sludge depths is considered to be the most reason- 
 able method of thickener analysis available at present. The approach is 
 based on experimentally determined sludge settling properties, and permits 
 convenient evaluation of alternative design or operating conditions. 
 Extreme care is necessary in measuring settling properties for gross anoma- 
 lies in physical behavior can be created by laboratory test conditions. The 
 maximum concentration which a sludge can reach by gravity thickening is a 
 function of its compressive strength. Compressive strength, and hence the 
 difficulty of thickening, increases exponentially with concentration. 
 Permeability of the sludge bed controls both the rate of escape of clarified 
 water and the portion of the effective weight of sludge solids which are 
 effective in compressing underlying layers. 
 
 Also described in the report are results of work concerning the sig- 
 nificance of sludge volume index measurements, the effect of possible methods 
 of altering sludge settleabi 1 i ty , the relationship between thickening and 
 sludge rheology, the influence of biological variables on the rheology of 
 activated sludge, the changes which occur in activated sludge aggregates 
 during thickening, the in situ measurement of suspended solids, the appli- 
 cation of the method of thickener analysis to full scale thickeners, and 
 
 i i 
 
the implications of the work to design of the activated sludge processes. 
 
 I i I 
 
TABLE OF CONTENTS 
 
 Page 
 
 ABSTRACT j j 
 
 TABLE OF CONTENTS j v 
 
 I. INTRODUCTION 1 
 
 Importance of Thickening in Wastewater Management 1 
 
 Purpose of Project 2 
 
 Project Organization 3 
 
 Nature of This Report k 
 
 II. THICKENING PROPERTIES OF SLUDGES 6 
 
 Introduction 6 
 
 Measurement of Settleabi 1 i ty 6 
 
 Enhancement of Sett leabi 1 i ty 8 
 
 Thickening Mechanisms 10 
 
 Sludge Rheology 13 
 
 III. THICKENER DESIGN 16 
 
 Introduction 16 
 
 Analysis of Possible Approaches 16 
 
 Design Technique 18 
 
 Applications and Extensions of the Technique 20 
 
 IV. SUMMARY AND CONCLUSIONS 22 
 
 REFERENCES 26 
 
 Project Reports and Publications 26 
 
 Other References 28 
 
 APPENDICES 31 
 
 Appendix I - The Sludge Volume Index - What Is It? .... 31 
 
Page 
 
 Appendix II - The Effect of Polymer Floccu lat ion on the . . 39 
 Settling Behavior of Activated Sludge 
 
 Appendix III - Thickening Characteristics of Activated . . 52 
 SI udge 
 
 Appendix IV - Aggregate Size Variation during Thickening . 68 
 of Activated Sludge 
 
 Appendix V - Distribution of Compressive Forces in ... . 87 
 Subsiding Sludge Masses 
 
 Appendix VI - Influence of Biological Variables on the . . 93 
 Physical Properties of Activated Sludge 
 
 Appendix VII - Thickening 118 
 
 Appendix VI I I - Role of Activated Sludge Final Settling . . 131 
 Tanks 
 
I. INTRODUCTION 
 Importance of Thickening in Wastewater Managemen t 
 
 In most waste treatment processes pollutants are concentrated by physical, 
 chemical, or biological means into a settleable form for removal from the 
 liquid waste stream. The solids might, for example, be chemical precipitants 
 formed by reaction with ions in the waste stream, carbonaceous material con- 
 centrated in the form of microbial mass (in the case of biological waste 
 treatment processes), or material contained in the suspended form in the raw 
 waste. 
 
 Effective waste management requires effective treatment, handling, and 
 disposal of the suspension of solids. The efficiency of virtually all methods 
 of sludge treatment, handling, and disposal depends on the concentration of 
 solids in the sludge. For example, the cost of transporting sludges depends 
 almost directly on the dilution factor of the solids (AWTR, 1968) as does the 
 required volume of sludge digesters (Shindala et_ a_l_. , 1970). The performance 
 of sludge dewatering processes such as vacuum filtration and centr ifugation 
 also depends upon the degree to which solids can be concentrated in the feed 
 (Sleeth, 1970). Clearly the economy of sludge combustion depends on achieving 
 a high solids concentration so that the process becomes thermally self- 
 sustaining (Hurwitz and Katz, 1959). 
 
 The most economical way of obtaining large sludge volume reductions is 
 by gravity thickening. Depending on the nature of the sludge, one to tenfold 
 volume reduction might be readily achieved by gravity thickening using rea- 
 sonably simple and low-cost equipment. It normally is much more expensive 
 to achieve similar volume reductions by mechanical means. The importance of 
 thickening and waste management may be appreciated by considering that about 
 25 to 50 percent of the total cost of waste management is attributable to 
 
 1 
 
sludge treatment handling and disposal (Levin, I968) . 
 
 In addition to the importance of gravity thickeners in sludge management 
 schemes, thickening principles are also involved in many waste treatment 
 processes. For example, any sedimentation tank is expected to accomplish a 
 degree of thickening, and failure to consider this requirement in designing 
 the facility can lead to unsatisfactory performance. Also, the performance 
 of many related processes depends upon how well thickening is accomplished 
 in sedimentation tanks. To illustrate, the size of the aeration tank in the 
 activated sludge process depends upon how concentrated the microorganisms 
 are in the sludge returned from the final settling tank. Similarly, the 
 performance of flocculators in water and waste treatment processes depends 
 upon the number of particle collisions which can be caused to occur and under 
 given conditions, this is related to the concentration of particles present. 
 The problem and cost of sludge treatment may be expected to become more 
 significant in the future as higher degrees of treatment result in production 
 of large volumes of flocculent suspensions. In summarizing the importance of 
 thickening research, Vinton Bacon (1966) has written: "Sludge concentration 
 is the largest unsolved research and development problem. The savings that 
 could be effected in this area alone would go a long way in improving other 
 treatment processes and the effluent." 
 Purpose of the Project 
 
 In spite of the importance of gravity thickening and the frequency with 
 which the process is used, little is known about how to design a thickener 
 on the basis of a rational interpretation of the thickening phenomena. Most 
 of the research which has been done on the process has been descriptive in 
 i nature and conventional design procedures which have evolved are often based 
 I on consideration of parameters such as surface loading which do not 
 
 2 
 
necessarily have an influence on the performance of a gravity thickener. 
 The purpose of this work was to investigate the basic nature of the 
 thickening process in order to afford an understanding of the performance 
 of the process and to devise techniques for reliable design of thickeners 
 on the basis of the fundamental properties of suspensions. 
 Project Organization 
 
 The work was supported by Federal Water Quality Administration Research 
 Grant 17070 DJR from September 1, 1966 through August 31, 1 9 69 . Research in 
 progress at the end of the grant period continued with support from the 
 University of Illinois. The work was conducted at the Department of Civil 
 Engineering, University of Illinois, Urbana, Illinois, under the direction 
 of Richard I. Dick, Professor of Civil Engineering. 
 
 Much of the research was carried out by candidates for the Master of 
 Science degree in sanitary engineering. The following graduate students 
 were employed as research assistants on the projects: 
 
 T. R. Wall in September 1 966 - February 1967 
 
 A. R. Javaheri September 1966 - August 1969 
 
 S. K Chakrabarti February I967 - May 1968 
 W. R. Gain February 1968 - August 1 969 
 
 Other graduate students conducted research related to the project in 
 partial fulfillment of master's degree requirement but were not employed 
 with project funds. These students, listed below, derived support from the 
 project in terms of supplies and equipment but received their personal 
 financial support from the training grant program of the Federal Water 
 Quality Administration or the U.S. Public Health Service: 
 
 G. A. Farnsworth September 1966 - February 1 967 
 J. R. Quin June - August 1967 
 
 3 
 
J. I. Barkman February I969 - August I969 
 
 Mr. Barkman also received a supply and travel allowance from the Decatur, 
 Illinois Water Company. 
 
 Other special studies for which the project provided supplies but not 
 salaries included graduate projects by M. C. Babb and R. C. T. Wang, and 
 undergraduate projects by J. C. Gratteau, W. R. Gain, and G. L. Caban. 
 Miss Caban's work also received support from the National Science Foundation. 
 
 Others associated with the project included part-time laboratory tech- 
 nicians and Dr. B. S. Narang who was employed as a part-time Research 
 Associate during the summer of 1 968 . Faculty members from various depart- 
 ments at the University of Illinois, especially Dr. B. B. Ewing of the 
 Department of Civil Engineering, served gratuitously as project consultants. 
 Nature of this Report 
 
 Most of the major accomplishments of the project have been reported in 
 the published literature, or were in preliminary manuscript form at the time 
 this project report was prepared. Many of these papers and manuscripts are 
 included in the appendices. The purpose of this report is to interrelate 
 and summarize these individual publications, but not to repeat in detail work 
 which previously has been described. 
 
 To distinguish between project publications and other publications, 
 the reference list at the end of the report is divided into two sections. 
 In the text, numbers preceded by "PR" are used to refer to references from 
 the project publication list, while the author-year system is used to cite 
 nonproject publications. Of the 31 project reports and publications listed, 
 18 have appeared in the published literature, of which five are included 
 here as appendices . 
 
 Preparation of this report was delayed to permit inclusion of results of 
 
 k 
 
work in progress at the end of the grant period. The research continued with 
 support from the University of Illinois Civil Engineering Department and, 
 later, from the Water Resources Center at the University of Illinois. Much 
 of the work has been completed. However, results bred new ideas, and research 
 is still in progress. Some of the continuing work makes use of equipment and 
 concepts generated by the grant. Under these circumstances, judgments as to 
 which accomplishments deserve inclusion in the final project report become 
 somewhat arbitrary. In general, results of work which was underway at the 
 end of the project period concerning excess hydrostatic pressure in subsiding 
 sludge masses have been included here while those related to analysis of the 
 performance of continuous thickeners have been excluded. 
 
 Because the focus of the research was on interpretation of thickening of 
 sludges on the basis of observed basic physical characteristics and the rela- 
 tion of this fundamental behavior to the design of field scale thickeners, 
 most of the work logically fell within one of two general categories. One 
 category involved study of the basic physical properties of sludges which 
 control its thickening behavior and the other category concerned work where 
 the emphasis was on design of thickeners. 
 
 Research concerned principally with study of the physical characteristics 
 of suspensions is described here in Chapter II entitled "Thickening Properties 
 of Sludges." This chapter also includes studies related to the interpretation 
 of results of laboratory batch sedimentation tests and some methodology associ- 
 ated with developments of the research work. Work which was related more 
 directly to the design of full scale thickeners based on known sedimentation 
 characteristics is included in Chapter III entitled "Thickener Design." 
 
II. THICKENING PROPERTIES OF SLUDGES 
 Introduct ion 
 
 In order to consider rational design and operation of thickeners, it is 
 recessary to be able to measure the physical properties of suspensions which 
 influence their thickening behavior and to understand factors which control 
 the physical properties of the sludge. In addition to providing insight into 
 factors controlling the design and operation of gravity thickeners, knowledge 
 of the interrelationship between physical properties and thickening behavior 
 might permit alteration of the basic properties of suspensions by physical, 
 chemical, or biological methods to improve the amenability of the sludge to 
 thickening by gravitational means. Work related to study of the basic nature 
 of the sludges and to experimental procedures required in evaluating sludge 
 properties are reviewed in this chapter. 
 Measurement of Settleab? 1 i ty 
 
 The property of a sludge most closely identified with its behavior in 
 full-scale thickeners is its subsidence velocity under the influence of 
 gravity. The apparent settling velocity of a particular concentration of 
 sludge can be measured readily be observing the rate of subsidence of the 
 liquid-solid interface following uniform dispersion in a transparent labora- 
 tory settling column. The test is deceptively simple, for serious errors 
 can result because of conditions imposed by the laboratory test procedure. 
 
 Studies on the influence of laboratory test conditions on observed 
 settleab i 1 i ty were conducted as a part of this work, and were extended by 
 cooperation with a related research program at the University of North 
 Carolina. Principal factors found to affect results of laboratory sedimen- 
 ' tation tests were the method for initially dispersing sludge solids, column 
 1 diameter, sludge depth, and the presence or absence of slow speed stirring. 
 
 6 
 
The extent to which these factors influence settling rate is highly dependent 
 on the nature and concentration of the particular sludge solids. 
 
 Results of the work have been reported in detail in two publications 
 (PR 10 and PR 29) and have been summarized in two others (PR 11 and PR 17). 
 In general, with activated sludge, column diameter should be as large as 
 possible and preferably not less than about 3 in., sludge depth should be 
 comparable to the effective depth of the full scale facility, and slow 
 stirring (with about 10 in./min tip speed) is essential. The greatest need 
 for additional work in this area is for study of the relationship between 
 the batch settl ing rate in cylinders to the settling rate of the same con- 
 centration of sludge under conditions which exist in a full-scale conti nuous 
 thickener. 
 
 Realization of the great anomolies in settling behavior which could be 
 caused by improper laboratory test conditions led to a critical evaluation 
 of the test most commonly used to express the sett leabi 1 i ty and physical con- 
 dition of sludge - the sludge volume index. These results also have been 
 published (PR 1 8) and are included here as Appendix I. Basically, it was 
 reported that the significance of sludge volume index values is seriously 
 restricted by the nature of laboratory tests. While the test may have value 
 as a plant operational tool, comparison of sludge volume index measurements 
 from various plants is not meaningful. The test was considered to be wholly 
 inadequate for research purposes, and alternative means for measuring the 
 settleabi 1 ity and physical nature of sludge solids were suggested. 
 
 The experimental program involved with evaluating laboratory settling 
 i tests required a large number of suspended solids determinations. In order 
 'to select a convenient, reliable, and economical technique and to permit 
 ^valuation of the number of duplicate samples required to establish a desired 
 
 7 
 
degree of precision, a study of alternative methods of determining the sus- 
 pended solids concentration of activated sludge was conducted. Results, 
 which have been published (PR 22) show, as a function of sample size and 
 concentration, the coefficient of variation for four different methods of 
 determining suspended solids. For the purpose of this study, the glass fiber 
 filter Gooch crucible method with individual desiccators was considered to be 
 superior. 
 Enhancement of Sett leab i 1 i ty 
 
 Studies of the alteration of settling behavior of sludges were under- 
 taken for two reasons. One reason was to consider the practicality of alter- 
 ing the physical properties of sludges in full-scale applications. The other 
 was to study the mechanisms responsible for the improved sedimentation in 
 order to gain insight into the basic factors controlling sludge thickening. 
 
 The most fruitful of the studies of this type was an evaluation of the 
 influence of polyelectrolytes on the settling behavior of activated sludge. 
 The practicality of this technique in full-scale installations under certain 
 circumstances was, of course, known from the published literature (for 
 example, Jordan and Scherer, 1970) and the thrust of the study was to evalu- 
 ate the basic change in the physical properties of activated sludge respon- 
 sible for the improved sett leabi 1 i ty . 
 
 The major portion of the work was the subject of a master's thesis by 
 G. A. Farnsworth (PR 20). Detailed results have been summarized in the form 
 of a draft of a manuscript for publication (PR 21) which is included here as 
 Appendix II. Settling properties were expressed as the ultimate settling 
 ; velocity (the velocity at which the sludge would settle if it behaved as an 
 > ideal suspension (see Dick and Ewing, 1967b) and the retardation factor (a 
 measure of the extent of deviation from an ideal suspension'*. Polymers were 
 
found to influence primarily the ultimate settling velocity and did not 
 appreciably alter the effect of i nterpart i cle contacts (as measured by the 
 retardation factor). Because of this, the apparent effect of polymer addi- 
 tion may be highly dependent on the sludge depth used in laboratory experi- 
 ments . 
 
 A preliminary evaluation of the feasibility of increasing activated 
 sludge settling rates by imposing an electrical gradient was made (PR 23). 
 While some slight effect could be shown the technique was not considered to 
 be promising either for full-scale application or for more basic laboratory 
 experimentation, and the work was terminated. 
 
 More extensive study was made of the effect of ultrasonic vibrations on 
 the settleabi 1 i ty of activated sludge (PR 30). The beneficial effect of 
 ultrasonic vibrations on sludge sett leabi 1 i ty is well known, although the 
 economic feasibility in full-scale applications remains dubious. The inter- 
 est in this work was in learning more about basic settling characteristics 
 of sludges by inquiring as to why exposure to ultrasonic vibrations improved 
 settleab i 1 i ty . 
 
 It was proposed that the basic mechanism explaining the improved thick- 
 ening characteristics was removal of bound water by ultrasonic vibrations 
 and then to measure its settling velocity and bound water content (using a 
 modification of the technique described by Heukelekian and Weisburg, 1956). 
 Significant reductions in bound water content and appreciable improvement in 
 interface subsidence rate could be shown to be effected by exposure to 
 ultrasonic vibrations prior to sedimentation. However, results were less 
 well ordered than desired and quantitative evaluation was difficult. This 
 ': experimental approach is still considered potentially fruitful, but improve- 
 I ments must first be made in the method for determining bound water content. 
 
 9 
 
Thickening Mechanisms 
 
 Ultimately, it is desirable that rational approach to design and 
 operation of gravity thickeners be founded on a basic understanding of the 
 fundamental mechanisms involved in thickening. Previous work showed that 
 the thickening behavior of activated sludge differed appreciably from that 
 of the ideal suspension considered in basic thickening theory. Work de- 
 scribed here was undertaken to explore causes of this difference and to 
 evaluate changes which occur during the course of sludge thickening. 
 
 In work summarized in proceedings of the 4th International Conference 
 on Water Pollution Research (PR ]h reproduced as Appendix III), a conceptual 
 model of activated sludge thickening was developed. The model was based on 
 mathematical analysis of the fluid and i nterparti cle forces acting on a sub- 
 siding mass of sludge. Model predictions deviated from behavior of an ideal 
 suspension, but agreed closely with the observed settling behavior of acti- 
 vated sludge. The relative magnitude of interpart icle forces as computed 
 from laboratory settling data using the model was related to the yield 
 strength of the sludge as experimentally determined with a viscometer. This 
 tended to confirm that the cause of deviation from the description of 
 thickening offered by Kynch (1952) is the existence in activated sludge of 
 interparticle forces. 
 
 Further examination of data previously reported (Dick and Ewing, 1967b) 
 led to approximation of the concentrations at which interparticle forces 
 begin to cause activated sludge sedimentation to differ from that of ideal 
 suspensions (PR 10). As anticipated the concentration at which a continuous 
 j structure occurred was highly dependent on the nature of the sludge, but in 
 ! each of the three plants studied the concentration was exceeded by the mixed 
 liquor suspended solids concentration, and in one case it was as low as 650 mg/ 
 
 10 
 
In considering basic mechanisms of thickening, it is of interest to know 
 whether water eliminated from sludge during thickening originates predominately 
 from within flocculent masses of sludge solids or from the intersticial spaces 
 between sludge aggregates. To answer this and related questions, activated 
 sludge settling data were analyzed by use of two mathematical models (PR 25). 
 One model (the Ri chardson-Zaki equation) related the settling velocity of a 
 mass of particles to the discrete settling velocity of an individual particle 
 comprising the mass, and the second (the Carmen-Kozeny equation) related flow 
 through a porous bed to the physical properties of the bed. Similar conclu- 
 sions were reached with both models. Results of the work were reported at an 
 annual meeting of the Water Pollution Control Federation, and the complete 
 published findings (PR 26) are included as Appendix IV. 
 
 Briefly, it was found that for activated sludges with good settling pro- 
 perties, thickening occurs primarily by elimination of interstitial water. 
 However, with poorly settling sludges, much of the water removal in the course 
 of thickening comes from inside the aggregates. The fraction of clarified 
 liquid which originated from within aggregates increased as thickening took 
 place. During thickening the aggregates which comprise the sludge are 
 "squeezed" to eliminate water and "split" into smaller, more numerous, and 
 more dense particles. To give some idea of the magnitude of values involved, 
 in thickening a good activated sludge from about 0.5 percent solids to about 
 2.0 percent solids, the effective diameter of sludge solids decreased from 
 3 mm to 0.5 mm and the number of particles increased 30-fold. The ratio of 
 the volume of floe particles to the volume of solids in the particles de- 
 ! creased from 57 to 72 and floe density increased from 1.0015 to 1.0032. At 
 : first, less than 10 percent of the water being eliminated from the sludge 
 . came from inside the aggregates, but at the end, more than 30 percent was 
 
 11 
 
from this source. 
 
 A limitation of the work concerning aggregate changes during activated 
 sludge thickening arose from the fact that when performing the calculations, 
 allowance could not be made for the effect of structural support due to 
 interparti cle contracts. Experimentally, the amount of interpart i cle support 
 in a subsiding suspension can be measured by observing the difference between 
 the effective weight of solids above a point and the excess hydrostatic pres- 
 sure at the point. However, this approach could not be used successfully 
 with activated sludge because the extremely light weight of the sludge solids 
 precluded accurate determination of excess hydrostatic pressure profiles. In 
 work with a denser sludge - that from a water softening plant - it was found 
 that allowance for interparticle support could readily be made by measuring 
 concentration and excess hydrostatic pressure profiles. Preliminary work of 
 this type (PR 27) was extended and is to be summarized in PR 28. A discussion 
 of the experimental technique and a presentation of some of the pertinent re- 
 sults of this work is included as Appendix V. 
 
 The experimental approach afforded considerable insight into the basic 
 mechanisms of thickening. The final solids concentration of a flocculent 
 sludge depends on its compressibility and can be increased by increasing the 
 weight of solids per unit area. The compressive strength of sludge was found 
 to vary exponentially with solids concentration. Hence, it becomes exceed- 
 ingly difficult to reach higher solids concentrations by gravity thickening. 
 During the course of thickening the amount of the total weight of sludge 
 jwhich serves to compress underlying layers is a function of the permeability 
 of the sludge. Thus, sludge with low permeability not only retards egress of 
 clarified water, but also reduces the compressive force available to accom- 
 plish sludge consolidation. Stirring on sludge thickening in laboratory 
 
 12 
 
vessels has been shown to increase compressibility and to reduce permeability, 
 
 The work on the basic nature of aggregate particles and their behavior 
 in concentrated suspensions provided the basis for several contributions to 
 the literature concerning the related work of others. In a discussion of 
 Mueller, Voelkel and Boyle's work ( 1 966) on activated sludge floe diameter 
 (PR 5), the possibility for change of the size and water content of floe 
 exposed to a shear field was discussed. Mechanisms of floe breakdown, sedi- 
 mentation in laboratory vessels, and characterization of floe properties 
 were considered in discussion (PR 12) of work by Ham and Christman (1969). 
 Application of the Ri chardson-Zaki equation to activated sludge solids as 
 described by Edeline, Tesarik, and Vostreil (1970) was discussed at the 4th 
 International Conference on Water Pollution Research (PR 14). The effect of 
 temperature on the rate of escape of water from subsiding sludge masses and 
 on compressibility of sludge solids was considered (PR 13) in discussion of 
 work by Reed and Murphy (1969). 
 Sludge Rheology 
 
 Advances in fundamental understanding of sludge thickening have been 
 handicapped by lack of fundamental measures of the physical properties of 
 sludges. As discussed in previous pages, the measure most commonly used, 
 the sludge volume index, suffers from being influenced by many different 
 physical characteristics of sludges, and it also reflects the nature of lab- 
 oratory test conditions. Use was made in these studies of directly observ- 
 able physical properties such as gravimetric concentration, settling veloc- 
 ity, and specific gravity and of calculated or estimated properties such as 
 volumetric concentration, aggregate size, aggregate density, porosity, and 
 permeability. However, none of these parameters gave a measure of the basic 
 .deformation and flow characteristics of sludge. Rheological measurements 
 
 13 
 
were undertaken for this purpose. 
 
 It had previously been proposed (Dick and Ewing, 1967a) that basic 
 Theological measures should prove useful in study of sludge treatment pro- 
 cesses, and a viscometer suitable for measuring the rheological character- 
 istics of activated sludge was described. For purposes of the work de- 
 scribed in this report, the viscometer was improved. The modified version 
 of the instrument, as described by Wang (PR 30 retained the basic features 
 of the former instrument including outer cylinder rotation, roughening of 
 cylinder surfaces, and use of a wide annular space between cylinders. Im- 
 provements incorporated in the modified version included oil damping of 
 inner cylinder oscillation, use of calibrated torsion wires for measuring 
 torque, and improved concentricity of the two cylinders. While the modified 
 viscometer represents a vast improvement over the original version, additional 
 improvements in range and convenience are desirable. 
 
 As discussed on previous pages, it was shown early in the project period 
 (PR ]k reproduced as Appendix III) that reported deviations in the settling 
 behavior of activated sludge from Kynch's theory could be interpreted in 
 
 • terms of the rheological behavior of the sludge. This was done by using a 
 mathematical model of thickening to compute the relative magnitude of inter- 
 
 '. particle forces. This value, deduced by use of observed sedimentation data, 
 was shown to be related to the yield strength of the sludge as measured in a 
 viscometer. 
 
 Later in the project period, development of the procedure for measuring 
 the absolute values of fluid resistance and interpart icle resistance in dense 
 
 J sludges permitted a more direct comparison of settling behavior and rheologi- 
 
 : cal properties (PR 28 - see Appendix V) . The compressive stress at which 
 failure of sludge took place under the confined conditions of laboratory 
 
 14 
 
 ! 
 
batch sedimentation tests was related to the yield strength of the sludge. 
 In related studies of flotation thickening of activated sludge (Wood, 
 1970) the viscometer was used to measure the rheology of sludges being 
 floated. Wood concluded that yield strength and plastic viscosity were the 
 best parameters for characterizing sludges and for predicting flotation 
 behavior. 
 
 Because activated sludge presents severe thickening problems, because 
 its thickening performance had been shown to be related to its rheological 
 characteristics, and because rheological characteristics were known to be 
 influenced by the nature of the waste treatment plant (Dick and Ewing, 1967a), 
 studies were undertaken (PR 2 and PR 3) to investigate how biological vari- 
 ables influenced sludge rheology. Results have been summarized in a pre- 
 publication manuscript (PR 19) which is included as Appendix VI. 
 
 The yield strength of a particular sludge was shown to be related to 
 solids concentration and organic loading, while plastic viscosity was in- 
 fluenced principally by concentration only. Great reduction in yield 
 strength was produced by aerobic digestion, and dramatic increases occurred 
 shortly after feeding. However, with a constant organic loading and sus- 
 pended solids concentration, changes in biological population could cause 
 pronounced changes in both yield strength and plastic viscosity. Such 
 changes were not necessarily accompanied by changes in performance of the 
 , biological phase of the activated sludge process. In the related studies 
 ! by Wood (1970) it was confirmed that changes in sludge rheology under seem- 
 ingly constant biological conditions were attributable to changes in mor- 
 phological characteristics of the organisms making up the sludge. Such 
 'changes, which significantly affect sludge thickening properties are not 
 reflected in a sensitive or definable way by the conventional sludge volume 
 1 index. 
 
 15 
 
III. THICKENER DES I Gfl 
 I ntroduct ion 
 
 An important aim, indeed the ultimate goal, of the research was to 
 develop procedures for design and operation of gravity thickeners on a 
 rational basis founded on knowledge of the fundamental thickening proper- 
 ties of suspensions. The bridge between thickening theory and practice is 
 a long hazardous one resting on unsure abutments. For the body of knowledge 
 on the basic thickening properties of suspensions is not large and most ap- 
 proaches to thickener design have not been rational. 
 
 However, the laboratory studies on basic settling properties afforded 
 a basis for accepting or rejecting possible design approaches on a rational 
 basis and study of data in the literature and analysis of the performance 
 of full-scale thickeners gave additional basis on which to proceed. The 
 work led to a proposed design technique applicable to the flocculent sludges 
 encountered in sanitary engineering practice which not only permits rational 
 approach to design, but also provides a framework for making the judgments 
 . required in thickener operation. Results of most of the work have been pub- 
 i lished (PR 7, PR 8, PR 9, PR 11, and PR 17) and are summarized in this 
 
 chapter. 
 , Analysis of Possible Approaches 
 
 A large number of approaches to thickener design and operation proposed 
 
 by other workers were considered. Some of these could be rejected without 
 
 ! detailed study whereas others were given careful analysis in order to develop 
 
 a suitable framework for a model for thickener designs and operation. 
 
 Among those approaches which could readily be rejected were the common 
 j 
 i techniques based on hydraulic loading and sludge volume index. While hydraulic 
 
 i loading is a common basis for thickener design (Great Lakes - Upper Mississippi 
 
 16 
 
River Board of Sanitary Engineers, I960), it is only related to solids loading 
 and is in itself wholly inadequate. The sludge volume index has often been 
 advocated as a means for predicting thickener performance - particularly for 
 the final settling tank in the activated sludge process (for example, Stewart, 
 196*0. However, the work on the nature of the SVI test (PR 18) showed that 
 the test would be very unreliable for this purpose. 
 
 Another common design approach is based on somewhat arbitrary establish- 
 ment of retention time and solids loading (American Society of Civil Engineers 
 1959)- This approach is based on parameters which are more closely related 
 to thickener performance, but the method does not permit the designer or 
 operator of thickeners to weigh the consequences of alternative decisions. 
 
 Other approaches evaluated included empirical correlations of various 
 
 variables with thickener performance. One such approach, by Pflanz (1970) 
 
 involved the ise of "solids feed" which was defined as surface settling rate 
 
 times the feed solids concentration. In discussion of this work (PR 15) it 
 
 was argued that interpretation of thickener performance by use of basic 
 
 clarification and thickening theory led to clearer interpretation of the 
 
 full-scale data on which the analysis was based. Fischerstrom et_ aj_. (1967) 
 
 suggested that the product of the thirty minute sediment volume and the 
 
 settling velocity be used to evaluate thickener capacity. A discussion of 
 
 possible inadequacies of this approach was published (PR 6). An extensive 
 I 
 Istudy of the relationship between laboratory and full-scale thickening by 
 
 Edde and Eckenfelder (I967) led to development of a design procedure based 
 
 ,on mass loading and two empirical parameters. Study of the method and the 
 
 nature of the empirical parameters indicated that the approach took Into 
 
 jaccount those basic variables which influence thickening behavior, but that 
 
 'the manner in which the factors were considered was indirect. Hence, the 
 
 17 
 
approach was not considered to provide suitable framework for a rational 
 th i ckeni ng model . 
 
 Probably the most important and most frequently cited work related to 
 thickening is that of Kynch (1952). The applicability of this work to waste 
 sludges was evaluated previously (Dick and Ewing, 1967b). It was found that 
 the work provided a valuable model of the thickening characteristics of ideal 
 suspensions but that the behavior of waste sludges deviated from that of the 
 ideal. Hence, the Talmage and Fitch (1955) geometric procedure for thickener 
 design which is advocated in many text books is not directly applicable to 
 waste sludges because it is based directly on Kynch's work. The approach is 
 further limited by the arbitrary procedure commonly used to identify the 
 limiting concentration. 
 
 It was considered that another approach which developed from Kynch's 
 work showed more promise. This was the use of the batch flux plot as advo- 
 cated by Yoshioka et al . (1957) and Shannon et_ aj_. (1963). A limitation of 
 the approach however would be that it does not take into account the effect 
 of depth on settling rates. 
 
 Another traditional approach to thickener design which required evalu- 
 ation was the procedure for determining thickener depth or volume. Conven- 
 tionally, this has been based on an empirical description of sludge consoli- 
 dation originated by Roberts (193*0- This approach was not considered to 
 
 i be well founded in terms of observed thickening behavior because it involved 
 independent determination of thickener area and thickener depth. It seemed 
 
 : imperative that the design approach take into account the interdependence of 
 the two. 
 Design Technique 
 
 Critical analysis of potential design methods led to selection of an 
 
 18 
 
approach felt to have the greatest utility given to the present state of 
 knowledge of basic thickening properties of waste sludges. The approach 
 is based on earlier work of authors such as Shannon and Tory (1966). The 
 basis of the approach is the rational statement that in a steady state in a 
 continuous thickener, the solids flux, G, is 
 
 G = c.v. + c.u (1) 
 
 where c. is the suspended solids concentration of sludge at any point in the 
 thickener, v. is the gravity settling velocity of the sludge at concentration 
 c. , and u is the downward velocity due to sludge removal. The required solids 
 flux through the thickener is determined by the solids loading and the 
 
 thickener area, A, such that 
 
 c Q 
 
 -2^ (2) 
 
 where c and Q are the feed concentration and feed rate. Hence, the basis 
 o o 
 
 for sizing thickeners is to determine the lowest value of G from equation 1 
 for all concentrations of sludge which could occur in the thickener and to 
 ascertain that sufficient area is provided so that the value of G from 
 equation 2 does not exceed this limiting value. 
 
 This method of thickener analysis is particularly valuable because it 
 takes into account both the settling properties of the sludge and the mode 
 of thickener operation. To explain, the first term in equation 1 is depen- 
 dent entirely on sludge properties while the second term is determined by 
 operating conditions. The value of the underflow velocity, u, is determined 
 
 by the rate of sludge removal which depends, in turn, on the desired degree 
 I 
 of sludge concentration. 
 
 The design approach has been described in more detail in PR 17 which is 
 
 19 
 
included here as Appendix VII. In another publication (PR 11), three pos- 
 sible techniques for solving the basic equations were illustrated, and it 
 was concluded that the approach making use of a batch flux curve had the 
 greatest utility. This paper is reproduced as Appendix VIM. An illustra- 
 tive problem using the technique is solved on page 123 of Appendix VII, and 
 an illustration of the way which the method can be used as a guide to thick- 
 ener operation is included on page 12*t. 
 
 The design technique does not require separate determination of thick- 
 ener depth. Rather, it is necessary that the sludge settling data used in 
 design be representative of those to be expected in the full-scale facility. 
 The effect of depth on the performance or required size of a thickener is 
 evaluated by the use of settling data representative of different sludge 
 depths . 
 Applications and Extensions of the Technique 
 
 The basic thickening model described was used by Barkman ( 1 9 69 ) in 
 analysis of the performance of an existing thickener for waste sludge from 
 a water clarification and softening plant. Predictions based on the model 
 were in reasonable agreement with plant operation. The analysis served to 
 reveal that the thickener was being operated at less than its potential 
 capacity. That is, the value of u in equation 1 was needlessly high be- 
 cause sludge was being withdrawn at a faster rate (and lower concentration) 
 1 than necessary. 
 
 The method of analysis for thickener design and operation was applied 
 to the final settling tank of the activated sludge process (PR 11 reproduced 
 as Appendix VIM). It was shown that use of conventional final settling 
 tank design procedures could result in inability to maintain desired sus- 
 pended solids concentrations in the aeration tank. The approach could be 
 
 20 
 
used as a basis for considering the effects of alternative methods for 
 operating existing tanks or as a basis for optimizing design of new final 
 settling tank. 
 
 Results of the laboratory studies of basic sludge thickening mechanisms 
 and the conceptual model of continuous thickener performance afford bases 
 for interpreting results of full-scale thickening studies and for extending 
 and refining the model. Equation 1 permits comparison of velocities in small 
 laboratory batch conditions with those in full scale continuous thickeners, 
 and interpretation of compression in terms of the combined influences of 
 permeability and compressibility (see Chapter II) provides a basis for evalu- 
 ating the influence of sludge depth and for studying the effects of rakes in 
 thickeners. The first stage of this work is being conducted with a laboratory 
 continuous thickener equipped with means for concentration and excess hydro- 
 static pressure measurement. This research was inspired by results of work 
 conducted as part of this project, but is being carried out with support 
 from the University of Illinois Water Resources Center. 
 
 21 
 
IV. SUMMARY AND CONCLUSIONS 
 
 In spite of the widespread use of sludge thickening in waste treatment, 
 and in spite of the potential for reducing treatment costs by improved thick- 
 ening, design and operation of thickeners has not been accomplished on a 
 rational basis. The basic goal of this research was to investigate basic 
 thickening properties of sludges and to develop thickener design and opera- 
 tion techniques consistent with knowledge of these properties. 
 
 Most results of work carried out in the project have been published in 
 the professional literature. This report serves to summarize and interrelate 
 the various individual reports of work but detailed procedures and results 
 are not repeated. Similarly, general conclusions stemming from the work are 
 presented in this section and more complete conclusions are to be found in 
 the project publications (some of which are included as appendices). 
 
 Serious errors can result from the conventional laboratory test for 
 measuring the settling rate of sludges. The influence of slow stirring in 
 laboratory settling tests is not necessarily a reflection of the benefits 
 to be derived by stirring in sludge thickeners, but is caused by the artifi- 
 cial conditions created by the laboratory test. Laboratory tests should be 
 conducted in cylinders as large in diameter as feasible and preferably not 
 less than about 3 in. Sludge depth should be comparable to the effective 
 
 depth in the full scale facility and a slow stirrer with tip speed of about 
 i 
 j 10 in./min should be provided. 
 
 The sludge volume index is an inadequate, indeed, a misleading, indica- 
 tion of settling characteristics. Its normal use should be restricted to 
 monitoring of gross physical properties of sludge for purposes of routine 
 ; plant operation. Comparisons should not be made between SVI values of dif- 
 ferent sludges and more basic and meaningful measures of the physical nature 
 
 22 
 
of sludges should be used in research work. 
 
 Deviations in the thickening characteristics of flocculent sludges like 
 activated sludges from those of the ideal slurry considered in Kynch's theory 
 are caused by i nterpart i cle contacts. The magnitude of the deviation is 
 related to the yield strength of the sludge. 
 
 The final concentration of sludge solids achievable by gravity thicken- 
 ing is controlled by the compressive strength of the sludge. Compressive 
 strength varies exponentially with suspended solids concentration, and hence 
 high concentrations are difficult to attain. 
 
 Efforts to increase the final concentration of thickened sludges should 
 be directed at reducing the compressive strength of sludges or increasing 
 the applied compressive load. Greater compressive loads may be achieved by 
 increasing the weight of solids per unit of thickener area; however, the 
 portion of these solids effective in compressing underlying solids is a 
 function of sludge permeability. In work with activated sludge, polymer 
 flocculation did not significantly change the magnitude of interparti cle 
 contacts, but rather it altered the discrete settling velocity of the indi- 
 vidual floe particles which comprise the sludge. Stirring reduces compres- 
 sive strength in laboratory settling equipment but may not be of the same 
 significance in full-scale tests. 
 
 The rate at which high concentrations are reached in batch sedimentation 
 is a function both of compressibility and permeability. Reduction in sludge 
 permeability yields dual rewards by increasing settling rates and increasing 
 applied compressive loads on underlying sludges. 
 
 During thickening, aggregate particles are squeezed to eliminate water 
 and also broken apart. The result is the formation of sludge containing 
 smaller, more numerous, and more dense particles. With sludge exhibiting 
 
 23 
 
poor thickening characteristics, much of the water removed by thickening 
 originates from inside sludge particles whereas more of the water from 
 sludges which thicken well comes from between particles. In either case, 
 the fraction of supernatant water originating from within particles increases 
 as thickening progresses. 
 
 With activated sludge, the yield strength in shear (and hence the com- 
 pressive strength) is increased with increased organic loading intensity. 
 However, it is also highly dependent on the morphology of the organisms 
 comprising the sludge, and two sludges of equal solids concentration devel- 
 oped on the same waste at the same loading intensity may have quite differ- 
 ent yield values. Aerobic digestion causes appreciable reduction in yield 
 strength, and pronounced changes occurred following feeding. 
 
 Analyses of proposed methods of thickener design indicated that many 
 are inadequate because they fail to take into account those factors which 
 influence thickener performance while others are unsatisfactory because 
 they are based on suspensions with properties different than those encount- 
 ered in waste treatment. Given the present state of understanding of 
 thickening behavior, the most reliable approach to thickener analysis is a 
 simple statement of continuity in a full-scale thickener. Analyses based on 
 this approach take into account both the settling characteristics of the 
 
 sludge and thickener operating practices. The batch flux curve affords a 
 I 
 
 convenient technique for evaluating alternative thickener designs or 
 
 alternative operational modes. 
 
 Application of the design approach to the final settling tank of the 
 
 ! 
 
 activated sludge process indicates that use of conventional settling tank 
 
 design practices can lead to unsatisfactory performance of the entire pro- 
 
 cess. Even when process performance is not affected, failure to consider 
 
 2k 
 
the interrelationships between the biological and physical phases of the 
 process leads to uneconomical design. 
 
 Full-scale thickener operational practices have an important influence 
 on results. Use of inadequate underflow velocities will become apparent 
 because of solids loss in the effluent. This can be corrected by increasing 
 the rate of sludge removal. However, use of too high an underflow rate 
 must be prevented because it results in the use of facilities at less than 
 f ul 1 capaci ty . 
 
 25 
 
REFERENCES" 
 
 Project Reports and Publications 
 
 PR 1 Babb , M. C.,"An Instrument for the Determination of Sludge Density, 1 ' 
 Unpublished Special Project Report, 11 pp. (Jan. 1968). 
 
 PR 2 Caban, G. L., "Changes in Some Physical Properties of Activated 
 
 Sludge under High Biological Loading Conditions," NSF Undergraduate 
 Research Project Report, University of Illinois, 71 PP- (Aug. I969). 
 
 PR 3 Chakrabarti, S. K. , "Changes in Some Physical Properties of Activated 
 Sludge under Different Biological Loading Conditions," Civil 
 Engineering Studies, Sanitary Engineering Series No. k~] , University 
 of Illinois, Urbana, 65 pp. (June 1 968) . 
 
 PR 4 Dick, R. l.,"Discussionofj_n S i tu Measurement of Solids in Final 
 
 Clarifiers by A. E. Albrecht, R. E. Wul lsch ieger , and W. J. Katz," 
 Journal of Sanitary Engi neeri ng Pi vis ion American Society of Civil 
 Engineers , 92, SA5, 117-119 (1966). 
 
 PR 5 Dick, R. I., "Discussion of Nominal Diameter of Floe Related to Oxygen 
 Transfer by J. A. Mueller, K. G. Voelkel, and W. C. Boyle," Journal 
 Sanitary Engineering Division American Society of Civil Engineers , 
 
 92, SA6, 144-146 (1966) . 
 
 PR 6 Dick, R. I., "Discussion of Settling of Activated Sludge in Horizontal 
 Tanks by C. N. H. Fisherstrom, E. Isgard, and I. Larsen," J ournal 
 Sanitary Engineering Division American Society of Civil Engineers , 
 
 93, SA6, 271-273 (1967). 
 
 PR 7 Dick, R. I., "Gravity Thickening of Sludge," Summer Institute in 
 
 Water Pollution Control - Biological Waste Treatment, Manhattan 
 College, Bronx, New York ( 1968) . 
 
 PR 8 Dick, R. I., "Some Fundamental Aspects of Sedimentation - the Clarifi- 
 cation Function," Water and Wastes Engineering , 6_, 2, 47-50 (1969). 
 
 PR 9 Dick, R. I., "Some Fundamental Aspects of Sedimentation - the Thicken- 
 ing Function," Water and Wastes Engineering , 6_, 3, 44-45 (1969). 
 
 PR 10 Dick, R. I., and Ewing, B. B., "Discussion Closure to Evaluation of 
 
 Activated Sludge Thickening Theories," Journal Sanitary Engineering 
 Division American Society of Civil Engineers , 95_, SA2, 333-3^0 (1969) 
 
 PR 11 Dick, R. I., "Role of Activated Sludge Final Settling Tanks," Journal 
 Sanitary Engineering Division American Society of Civi 1 Engineers , 
 96, SA2, 423-436 (I97O) (See Appendix VIM). ~~""~ 
 
 In the text, references to project reports and publications are by 
 number preceded by "PR" whereas other references are cited by use 
 of the name-date system. 
 
 26 
 
PR 12 Dick, R. I., "Discussion of Agglomerate Size Changes in Coagulation by 
 
 R. K. Ham and R. F. Christman," Journal Sani tary Engi neer i ng Division 
 American Society of Civil Engineer? , 96, SA2 , 624-627 (1970)- 
 
 PR 13 Dick, R. I., "Discussion of Low Temperature Activated Sludge Settling 
 
 by S. C. Reed and R. S. Murphy," Journal Sanitary Engineering Division 
 American Society of Civil Engineers , 96, SA2 , 638-641 (1970). 
 
 PR 14 Dick, R. I., "Thickening Characteristics of Activated Sludge," in Advances 
 in Water Pollution Research , Proceedings of Fourth International 
 Conference on Water Pollution Research, Prague, 1969, 625-642 (1970) 
 (See Appendix III). 
 
 PR 15 Dick, R. I., "Formal Discussion of Sedimentation of Activated Sludge 
 
 in Final Settling Tanks by P. Pflanz," in Advances in Water Pollution 
 Research , Proceedings of Fourth International Conference on Water 
 Pollution Research, Prague, 1969, 583~585 (1970). 
 
 PR 16 Dick, R. I., and Javaheri, A. R. , "Discussion of Fluidization of Floes 
 Produced in Chemical or Biological Treatment Plants by F. Edeline, 
 I. Tesarik, and J. Vostrei 1 ," in Advances in Water Pollution Research , 
 Proceedings of Fourth International Conference on Water Pollution 
 Research, Prague, 1969, 538 (1970). 
 
 PR 1 7 Dick, R. I . , "Th ickeni ng," Advances in Water Quality Improvemen t - 
 
 Physical and Chemical Processes , E. F\ Gloyna and W. W. Eckenfelder, 
 Jr., (editors), University of Texas Press, 358-369 (1970) (See 
 Appendix VII). 
 
 PR 1 
 
 Dick, R. I., and Vesilind, P. A., "The Sludge Volume Index - What Is 
 It?" Journal Water Pollution Control Federation , _4j_, 7, 1285-1291 
 (1969) (See Appendix I) . ---— 
 
 PR 19 Dick, R. I., Chakrabarti, S. K. , and McCutcheon, G. L., "Influence of 
 
 Biological Variables on Rheological Properties of Activated Sludge," 
 Prepubl ication Manuscript (1970) (See Appendix VI). 
 
 PR 20 Farnsworth, G. A., "The Effect of Induced Flocculation on the Settling 
 and Thickening Behavior of Activated Sludge," Civil Engineering 
 Studies, Sanitary Engineering Series No. 42, University of Illinois, 
 Urbana, 54 pp. (Aug. 1967). 
 
 PR 21 Farnsworth, G. A., and Dick, R. I., "The Effect of Polymer Flocculation 
 on the Settling Behavior of Activated Sludge," Prepubl ication 
 Manuscript (1970) (See Appendix II). 
 
 PR 22 Gratteau, J. C, and Dick, R. I., "Activated Sludge Suspended Solids 
 Determinations," Water and Sewage Works , 1 15 , TO, 468-472 (1968) . 
 
 PR 23 Gain, W. R. , "Electrically Induced Settling of Activated Sludge," 
 Unpublished Special Project Report, 29 pp. (Feb. 1968) . 
 
 27 
 
PR 2k Gain, W. R. , "In Situ Measurement of Suspended Solids Profiles in 
 
 Sludge Thickeners," M. S. Thesis, University of Illinois, Urbana 
 (currently being prepared). 
 
 PR 25 Javaheri, A. R., "Applicability of Two Mathematical Models to the 
 Batch Settling of Activated Sludge," Civil Engineering Studies, 
 Sanitary Engineering Series No. 51, University of Illinois, Urbana 
 101 pp. (June 1969) . 
 
 PR 26 Javaheri, A. R., and Dick, R. I., "Aggregate Size Variations During 
 Thickening of Activated Sludge," Journal Water Pollution Control 
 Federati on, 4j_, 5, Part 2, R197-R214 (1969) (See Appendix IV). 
 
 PR 27 Quin, J. R., "Role of Structural Support in Sludge Thickening," 
 
 Civil Engineering Studies, Sanitary Engineering Series No. k$ , 
 University of Illinois, Urbana (May 1968). 
 
 PR 28 Shin, B. S., "Distribution of Compressive Forces in Subsiding Sludge 
 Masses," M. S. Thesis, University of Illinois, Urbana (currently 
 being prepared ) (See Appendix V) . 
 
 PR 29 Vesilind, P. A., and Dick, R. I., "Initial Depth as a Variable in 
 Activated Sludge Settling Tests," Effluent and Water Treatment 
 Journal , 9, 5, 263-268 (1969). 
 
 PR 30 Wallin, T. R. , "The Influence of Ultrasonic Vibrations upon the 
 
 Physical Features of Activated Sludge," Civil Engineering Studies, 
 Sanitary Engineering Series No. ^3, University of Illinois, Urbana, 
 101 pp. (Nov. 1967) • 
 
 PR 31 Wang, R. C. T., "A Viscometer for the Study of the Rheology of 
 Activated Sludge," Unpublished Special Project Report, 25 pp. 
 (June 1967). 
 
 Other References 
 
 American Society of Civil Engineers, "Sewage Treatment Plant Design," Manual 
 of Practice No. 36 (1959) . 
 
 AWTR Summary Report, Advanced Waste Treatment Research Program, July 1964- 
 July 1967, Federal Water Pollution Control Administration 
 Publication WP-20-AWTR- 19 (1968). 
 
 Bacon, V. W. , nd Dalton, F. E., "Chicago Metro Sanitary District Makes no 
 Little Plans," Public Works , 97, 11, 66 (1966) . 
 
 Barkman, J. I., "Gravity Thickening and Mechanical Dewatering of Alum-Lime 
 Sludge," Decatur, Illinois, M. S. Special Problem, University of 
 Illinois, Urbana (1969) . 
 
 28 
 
Dick, R. I., and Ewing, B. B., "The Rheology of Activated Sludge," J ournal 
 Water Pollution Control Federation, 39, *f , 5^3-560 (1967aT. 
 
 Dick, R. I., and Ewing, B. B., "Evaluation of Activated Sludge Thickening 
 Theories, 1 ' Journal Sanitary Engineering C 
 of Civi 1 Engineers, 93, SA*» , 9~29 (1967b) 
 
 Theories, 1 ' Journal Sanitary Engineering Divi s ? on American Soc iety 
 
 Edde, H. J., and Eckenfelder, W. W., Jr., "Theoretical Concepts of Gravity 
 
 Sludge Thickening and Methods of Scale up from Laboratory Units to 
 Prototype Design," Center for Research in Water Resources Report 
 No. 15, University of Texas, Austin, ]kk pp. (1967). 
 
 Edeline, F., Tesarik, I., and Vostril, J., "Fluidizat ion of Floes Produced in 
 Chemical or Biological Treatment Plants," in Advances in Water 
 Pol 1 ut ion Rese arch , Proceedings Fourth International Conference, 
 Prague, 1969 ,~ S. H. Jenkins (editor), 523 (1970). 
 
 Fischerstrom, C. N. H., Isgard, E., and Larsen, I., "Settling of Activated 
 
 Sludge in Horizontal Tanks," J ournal Sa nitary Engineering Pi vi s ion 
 American Society of Civil Engineers , 93, SA3, 73-83~TT9"677.~ 
 
 Great Lakes - Upper Mississippi River Board of Sanitary Engineers, "Recom- 
 mended Standards for Sewage Works" (i960). 
 
 Ham, R. K. , and Christman, R. F., "Agglomerate Size Changes in Coagulation," 
 Journal Sanitary Engin eering Division American Society of Civi l 
 E ngineers , 96, SA3 , 48~l-502 (1969). 
 
 Heukelekian, H., and Weisburg, E., "Bound Water and Activated Sludge Bulking, 
 Sewage and Industrial Wastes, 23 , 558-57** (1956). 
 
 Hurwitz, E., and Katz, W. J., "Concentrating Activated Sludge to a Fuel 
 Value of 4000 BTU per Gallon," Wastes Engineering, 30, 730-733 
 (1959). 
 
 Jordan, V. J., and Scherer, C. H., "Gravity Thickening Techniques of a Water 
 Reclamation Plant," Journal Water Pollution Control Federation , 
 kl, 2, 1 80 (1970) . 
 
 Kynch , G. J., "A Theory of Sedimentation," Transactions Faraday Society , 48_, 
 166-176 (1952). 
 
 Levin, P., "Disposal Systems and Characteristics of Solid Wastes Generated 
 at Waste Water Treatment Plants," P roceedings 10th Sanita ry 
 Engineering Conference , University of Illinois Bulletin £5, 115, 
 21 (1968). 
 
 Mueller, J. A., Voelkel, K. G. , and Boyle, W. C, "Nominal Diameter of Floe 
 Related to Oxygen Transfer," Journal Sanitary Engineering Division 
 American Society of Civil Engineers , 92, SA2 , 9"20 (1966). 
 
 29 
 
Reed, C, and Murphy, R. , "Low Temperature Activated Sludge Settling," 
 
 Journal Sanitary Engineer ing Pi vi s ion Amer i can Societ y of C ivi 1_ 
 Engineers , 95, SWTWFWI (1969). 
 
 Roberts, E. J., "Colloidal Chemistry and Pulp Thickening," Transact ions 
 American Institute of M i nin g and Metallurgical Engineer s, J_l_2, 
 178-188 (193*0 • 
 
 Shannon, P. T., Stroupe, E., and Tory, E. M., "Batch and Continuous Thickening,' 
 I ndustrial and Engineering Chemistry Fundamentals , 2, 203-211 ( 1 9 63) - 
 
 Shannon, P. T., and Tory, E. M., "The Analysis of Continuous Thickening," 
 
 Transactions American Institute of Mining Engineers, 235, 375 _ 382 
 
 TisSST 
 
 Shindala, A., Pust, J. V., and Champion, H. L., "Accelerated Pigestion of 
 
 Concentrated Sludge," W ater and Sewage Works , 1 17 , 9, 329~332 (1970). 
 
 Sleeth, R. E., "Further Experience in the Use of Polyelectrolytes for Sludge 
 Conditioning at Worthing," Effluent and Water Treatment Journal , 
 
 j_o, 10, 582-591 (1970) . 
 
 Stewart, M. J., "Activated Sludge Process Variables - the Complete Spectrum," 
 Water and Sewage Works , R260-262 (196*0 • 
 
 Talmage, W. P., and Fitch, E. B., "Oetermining Thickener Unit Areas," 
 Industrial and Engineering Chemistry , kj_, 38-41 (1955). 
 
 Wood, R. F., "The Effect of Sludge Characteristics upon the Flotation of 
 
 Bulked Activated Sludge," Thesis submitted in partial fulfillment 
 of the requirements for the degree of Poctor of Philosophy, 
 University of Illinois, Urbana, H»3 pp. (1970). 
 
 Yoshioka, N., Hotta, Y., Tanaka, S., Nlaito, S., and Tsugami , S., "Continuous 
 
 Thickening of Homogeneous Flocculated Slurries," Chemical Eng ine ering 
 (Tokyo) , 21, 66-7** (1957). 
 
 30 
 
APPENDIX I 
 
 THE SLUDGE VOLUME INDEX - WHAT IS IT? 
 
 by 
 
 Richard I . Dick 
 
 and 
 
 P. Aarne Vesi 1 i nd 
 
 Reproduced from 
 
 Journal of the Water Pollution Control Federation 
 
 Volume 41, No. 7, Pages 1285-1291 
 
 July, 1969 
 
 31 
 
© Copyright as part of the July 19(59, Journal Watch Poi 
 Federation, Washington, D. C, 20016 
 Printed in U. S. A. 
 
 THE SLUDGE VOLUME INDEX— 
 WHAT IS IT? 
 
 Richard I. Dick and P. Aarne Vesilind 
 
 The sludge volume index (SVI), 
 introduced by Mohlman (1) in 1934, 
 has become the standard measure of 
 the physical characteristics of acti- 
 vated sludge solids. It is defined as 
 "the volume in ml occupied by 1 g 
 activated sludge after settling the 
 aerated liquor for 30 min" (2). The 
 sludge density index, introduced by 
 Donaldson (3), is the reciprocal of 
 the SVI multiplied by 100. Both of 
 these indices originally were intended 
 to be rough measures of sludge settle- 
 ability to be used in the everyday 
 operation of waste treatment plants as 
 means for monitoring the physical con- 
 dition of activated sludge. Because 
 of the simplicity of the SVI test, how- 
 ever, it has been applied widely for 
 purposes for which it was not intended 
 originally. 
 
 The general acceptance of this ar- 
 bitrary parameter as a basic measure 
 of the physical properties of activated 
 sludge solids is indicated by its wide- 
 spread use both in the operation of 
 waste treatment facilities and in re- 
 search on waste treatment. For ex- 
 ample, the SVI commonly is used in 
 research applications to evaluate the 
 effect of biological variables or physi- 
 cal or chemical treatment on the prop- 
 erties of sludge. Also, the SVI has 
 been advocated as a means for estab- 
 
 Bichard I. Dick is Associate Professor of 
 Sanitary Engineering, University of Illinois, 
 Urbana, Illinois, and P. Aarne Vesilind is 
 associated with the Norsk Institutt for Vann- 
 forskning, Oslo, Norway. At the time this 
 paper was prepared, Br. Vesilind was Re- 
 search Associate, Department of Environ- 
 mental Science and Engineering, University 
 of North Carolina, Chapel Hill, North Caro- 
 lina. 
 
 lishing the required sludge recircula- 
 tion rate or for calculating the mixed 
 liquor suspended solids concentration 
 which can be maintained in the aera- 
 tion tank. The most common use of 
 the parameter, of course, has been in 
 monitoring waste treatment plant op- 
 eration and in comparing the settling 
 characteristics of various sludges. 
 
 In the standard SVI test, sludge 
 volume is observed after a uniformly 
 mixed sample of sludge has settled 
 quiescently for 30 min in a standard 
 1-1 graduated cylinder. The vol- 
 ume occupied by the sludge after this 
 period of settling depends on both the 
 initial settling rate and the subsidence 
 characteristics at the higher sludge 
 concentrations. Two different acti- 
 vated sludges, both of which have the 
 same initial suspended solids concen- 
 tration and identical 30-min sediment 
 volumes, will have identical SVI val- 
 ues. However, the settling properties 
 of the two sludges may be grossly dif- 
 ferent (Figure 1). Since the SVI de- 
 fines only one point on the settling 
 curve, it is not a precise measure of 
 settling characteristics. 
 
 If the SVI then is not a measure 
 of the sludge settling characteristics, 
 what is it? What properties of acti- 
 vated sludge influence its magnitude? 
 Does it quantitatively describe physi- 
 cal properties which are indicative of 
 the behavior of the full-scale process? 
 Can meaningful comparisons be made 
 between the values of the SVI in 
 various plants? 
 
 The purpose of this paper is to 
 propose answers to the above ques- 
 tions. The limitations of the SVI 
 test are discussed and alternate means 
 
 32 
 
JOURNAL WPCF 
 
 July ]9G9 
 
 of describing the basic physical prop- 
 erties of activated sludge are sug- 
 gested. 
 
 Factors Influencing the Sludge 
 Volume Index 
 
 Suspended Solids Concentration 
 
 If the sludge volume index were 
 some fundamental measure of the sol- 
 ids which comprise activated sludge, 
 then some orderly relationship between 
 the concentration of sludge solids and 
 the sludge volume index would be ex- 
 pected. Sludge volume index deter- 
 minations were conducted using vari- 
 ous concentrations of sludge from sev- 
 eral activated sludge plants. The re- 
 sults (Figure 2) indicate that no con- 
 sistent relationship seems to exist. 
 
 The rapid increase of the SVI with 
 increasing concentrations is because of 
 the failure of the sludge to agglom- 
 erate into a coarse, open lattice to 
 permit settling. The formation of this 
 open lattice structure, frequently re- 
 ferred to as agglomeration, can be de- 
 termined readily from observation. 
 The failure to agglomerate is an arti- 
 fact of cylinder diameter and does not 
 occur necessarily in the full-scale plant 
 (4). For very high suspended solids 
 concentrations [greater than 6,000 
 mg/1 for plants E and A2 (Figure 2)] 
 the sludge still may not agglomerate ; 
 but since the maximum possible SVI 
 
 10 20 
 TIME.min 
 
 FIGURE 1.— These two sludges, with 
 grossly different settling characteristics, 
 have identical SVI values. 
 
 "0 5000 IOPOO 15,000 
 
 SUSPENDED SOUDS CONCENTRATION, mg/jt 
 
 FIGURE 2.— There is no consistent re- 
 lationship between suspended solids con- 
 centration and SVI. 
 
 of a sludge decreases as concentration 
 increases, the greater concentrations 
 tend to decrease the SVI. Maximum 
 possible SVI values are shown in Fig- 
 ure 2 as a function of concentration. 
 To illustrate, a sludge of 10,000 mg/1 
 solids, even if it did not settle at all, 
 still would have a maximum SVI equal 
 to 1,000 ml/lOg (or 100 ml/g) which 
 generally is considered to be a desir- 
 able SVI value. Clearly, therefore, 
 the SVI of a sludge is highly depen- 
 dent on its suspended solids concen- 
 tration. SVI values measured at vari- 
 ous solids concentrations vary widely 
 (Figure 2). Yet SVI values com- 
 monly are compared without regard to 
 concentration. 
 
 Rheological Characteristics 
 
 Rheological characteristics are fun- 
 damental measures of the physical 
 characteristics of a suspension relat- 
 ing to deformation and flow proper- 
 ties. To determine if these properties 
 consistently influence sludge volume 
 index values, the relationships between 
 the SVI and the yield strength and 
 plastic viscosity of various sludges 
 were determined using the viscometer 
 and procedures described by Dick and 
 Ewing (5) (Figures 3 and 4). Sludge 
 yield strength was not related to SVI 
 values in a consistent fashion (Figure 
 
 33 
 
Vol. 41, No. 7 
 
 SLUDGE VOLUME INDEX 
 
 3). Higher SVI values normally were 
 associated with higher plastic viscosi- 
 ties, but the relationship was not the 
 same for all sludges (Figure 4). Thus 
 it was substantiated that SVI is not a 
 reflection of just these basic physical 
 properties. 
 
 Interface Velocity 
 
 The initial sludge interface velocity 
 obtained in batch settling tests is used 
 widely as an indication of sludge set- 
 tling characteristics. The relationship 
 between the sludge volume index of 
 sludge samples and their initial set- 
 tling velocities in one-liter graduated 
 cylinders was explored, and the re- 
 sults (Figure 5) indicate that there 
 is not a consistent, meaningful rela- 
 tionship between the initial settling 
 velocity and the SVI. Attempts also 
 were made to relate the compression 
 rate constant described by Roberts (6) 
 to the SVI. Again, no meaningful cor- 
 relation was obtained. 
 
 Cylinder Diameter 
 
 Because the SVI values have been 
 used to predict possible underflow 
 solids concentrations and thus required 
 recirculation rates in full-scale plants, 
 it is appropriate to explore the rela- 
 tionship between settling behavior in 
 the SVI test and in full-scale facili- 
 
 600 
 
 600 
 500 
 
 ^400- 
 
 E 
 
 _-300r- 
 
 > 
 
 200 
 100- 
 
 
 
 '0 0.1 0.2 0.3 0.4 0.5 0.6 
 
 YIELD STRENGTH, dynes/sq cm 
 
 FIGURE 3.— Sludge yield strength does 
 not influence SVI in a consistent manner. 
 
 002 004 006 008 010 0.12 
 
 PLASTIC VISCOSITY, dyne sec/sq cm 
 
 FIGURE 4.— The SVI is not related to 
 plastic viscosity in a consistent manner. 
 
 ties. The diameter of the standard 
 one-liter cylinder in which SVI mea- 
 surements are made may influence re- 
 sults. Unless the relationship between 
 settling in small cylinders and settling 
 in full-scale plants can be established, 
 use of SVI values may yield mislead- 
 ing predictions of full-scale settling 
 behavior. 
 
 The results of SVI experiments us- 
 ing various sized cylinders (Figure 6) 
 indicate that SVI values can be ob- 
 tained that are appreciably greater or 
 less than the value associated with the 
 standard cylinder, and that the results 
 from a one-liter cylinder may not be 
 at all indicative of the true sludge 
 settling characteristics. In addition, 
 it does not seem likely that a consis- 
 tent relationship between SVI and set- 
 tling in prototype tanks is possible. 
 In this light, it is interesting to note 
 that researchers working with small 
 volumes of sludge often conduct 
 "SVI" measurements in 100-ml grad- 
 uated cylinders. The effect of cylin- 
 der diameter on settling characteristics 
 has been described more thoroughly 
 by Vesilind (7). 
 
 Initial Depth 
 
 The 1-1 graduated cylinders used 
 for the SVI tests are approximately 
 14 in. (35.6) cm tall. Intial depth, 
 
 3^ 
 
JOURNAL Wl'CK 
 
 July 
 
 however, has a considerable influence 
 on the sludge settling rate. For 
 various concentrations of 3 activated 
 sludges, settling velocities in 14-in. 
 (35.6-cm) columns were found to be 
 from 18 to 84 percent of the settling 
 velocities attainable in much taller 
 cylinders (8). The lower settling 
 velocities in short columns are thought 
 to be caused by the increased support 
 provided by underlying solids. Depth 
 affects different sludges differently, in- 
 dicating that if the settling character- 
 istics of several sludges are to be com- 
 pared, the tests should be conducted 
 in relatively tall cylinders. 
 
 Temperature 
 
 The effect of temperature on the 
 settling of sludge was studied by Ru- 
 dolfs and Lacy in 1934 (9) and their 
 data permit calculation of SVI values 
 (Figure 7). Clearly, the SVI is influ- 
 enced considerably by the temperature 
 under which the tests are conducted, as 
 would be expected because of viscosity 
 changes. Hence, temperature changes 
 alone can be expected to cause appre- 
 ciable changes in SVI values. Two 
 sludges with the same SVI value may 
 
 300 
 
 i i r 
 
 PLANT D 
 
 0.2 0.4 0.6 0.8 10 1.2 
 
 INTERFACE VELOCITY, iaAnin 
 
 FIGURE 5. — The initial sludge interface 
 velocity, used widely as an indication of 
 sludge settling properties, is not related 
 consistently to SVI. 
 
 0.5 1.0 1.5 2.0 2.5 3.0 3.5 
 CYLINDER DIAMETER, in. 
 
 FIGURE 6.— The effect of cylinder diam- 
 eter on SVI is shown. 
 
 not be similar if they are taken from 
 two treatment plants with different 
 waste temperatures. Also, seasonal 
 changes in SVI, not related to sludge 
 solids properties, may occur within a 
 given plant. 
 
 Stirring 
 
 The effect of stirring on sludge set- 
 tling is complex. Stirring is thought 
 to (a) aid in the agglomeration of the 
 sludge and (6) destroy the bridging 
 within a sludge bed in small cylinders. 
 Both of the effects of stirring result 
 in better settling (4). 
 
 Because actual settling basins are 
 not quiescent, stirring in test cylinders 
 may tend to yield more realistic re- 
 sults. The results of a series of ex- 
 periments using slow stirring (1 rpm) 
 (Figure 8) indicate that stirring re- 
 duces SVI values significantly. Note 
 that in the stirred tests the SVI still 
 was not independent of the concen- 
 tration of suspended solids. 
 
 The relationship between the quies- 
 cent and stirred SVI values is shown 
 on Figure 9. Here the SVI values 
 are expressed as a percent of quiescent 
 test SVI with a value less than 100 
 indicating a beneficial effect of stir- 
 ring. All of the sludge tested ex- 
 hibited better settling under stirred 
 conditions. However the degree of 
 improvement varied with different 
 sludges, and no meaningful correla- 
 tion existed between the stirred and 
 quiescent tests. 
 
 35 
 
Vol. 41, No. 7 
 
 SUJPOE YOUJUK INDEX 
 
 Discussion 
 
 The data indicate that sludge vol- 
 ume index is a very nonspecific, arbi- 
 trary measure of the physical charac- 
 teristics of activated sludge. The test 
 does not provide an indication of the 
 settling velocity, since only one point 
 on the settling curve is recorded. 
 Neither is the test a measure of the 
 compactability of the sludge since it 
 is conceivable that, at the end of 30 
 min, the interface still would be set- 
 tling at a constant rate. It therefore 
 is difficult to say exactly what the test 
 does measure, and therein lies its major 
 difficulty. It is not related consist- 
 ently to any basic physical property, 
 but rather represents the combined 
 influence of all of the various physical 
 properties of the sludge. Regrettably, 
 the relative influence of particular 
 physical properties on the SVI changes 
 from sludge to sludge and, indeed, be- 
 tween various concentrations of the 
 same sludge. 
 
 In addition, the sludge volume in- 
 dex is not a representative measure 
 of the settling characteristics of sludge 
 in full-scale settling basins because of 
 
 300 
 
 200- 
 
 100 
 
 '0 10 20 30 40 
 TEMPERATURE °C 
 
 50 
 
 300 
 250 
 200 
 150 
 100 
 50 
 0. 
 
 
 PLANT 0-NOT 
 STIRRED 
 
 ] rPLANT D-STIRRED 
 / / •^ rPLANT E-NOT 
 -<*/ / /X STIRRED 
 
 -PLANT E- STIRRED 
 STIRRING RATE = I rpm 
 
 FIGURE 7.— Temperature affects the SVI 
 value of sludge considerably. 
 
 4000 8000 12000 
 
 SUSPENDED SOUDS CONCENTRATOR mg/fl 
 
 FIGURE 8.— Slow stirring may affect the 
 SVI significantly. 
 
 the artificial settling conditions cre- 
 ated in laboratory settling tests. 
 Therefore, it cannot be used quantita- 
 tively to predict performance of set- 
 tling basins. 
 
 The problems which may result from 
 the misuse of the SVI are numerous. 
 For example, errors inevitably would 
 result if the SVI was used to predict 
 required recycle rates. The selection 
 of design parameters for full-scale 
 treatment plants based on pilot-plant 
 SVI values may result in serious over- 
 or under-design. Research results 
 which rely on the SVI as a primary 
 measure of sludge characteristics may 
 be questionable because of the insensi- 
 tivity of the SVI to actual changes in 
 the sludge physical characteristics. 
 
 What, then, is the value of measur- 
 ing SVI, and what beneficial use can 
 be made of SVI values? The sole vir- 
 tues of the determination would seem 
 to be that it can be conducted with 
 relative ease, and through usage, has 
 acquired a "meaning" which permits 
 its use as an operational tool. The 
 determination provides a very con- 
 venient test for monitoring changes in 
 performance of a particular plant. 
 Comparisons of SVI values from vari- 
 ous plants, however, apparently are 
 meaningless because the test probably 
 measures different properties of dif- 
 ferent sludges. 
 
 36 
 
JOURNAL WPCF 
 
 July LOGO 
 
 Use of the SVI test as an operational 
 tool to monitor changes in sludge char- 
 acteristics in a given plant would seem 
 to be the only valid application of the 
 measurement. Sludge volume index 
 values based on pilot-plant data can- 
 not be used to calculate the thickening 
 performance of final settling basins or 
 to determine required recycle rates. 
 It would seem particularly inappropri- 
 ate to use the SVI in research appli- 
 cations. More fundamental measures 
 of physical characteristics are required 
 in these cases. 
 
 It is suggested that alternate, more 
 meaningful measurements of the phys- 
 ical characteristics of activated sludge 
 be used where possible. One basic 
 measure which can be determined with 
 about the same ease as the .SVI is the 
 initial settling velocity associated with 
 various concentrations of activated 
 sludge solids. This is determined by 
 finding the slope of the interface sub- 
 sidence curve of activated sludge solids 
 in a comparatively large stirred set- 
 tling column. It also may be possible 
 to relate the interface velocities to the 
 concentration by an equation such as 
 v =ae bC where v and C are the inter- 
 face velocity and concentration, re- 
 spectively, e is the base of the natural 
 
 90 
 
 y 8 
 
 i 70 
 S 60- 
 
 50- 
 
 40- 
 
 20 
 10- 
 
 PLANT D 
 PLANT C 
 
 STIRRING RATE 
 
 "0 4000 8000 12,000 
 SUSPENDED SOLIDS CONCENTRATION, mg/t 
 
 FIGURE 9.— AU of the sludges tested 
 exhibited lower SVI values under stirred 
 conditions. 
 
 log, and a and /; are constants (7). 
 Use of these constants may afford a 
 method by which the settling charac- 
 teristics of different sludges may be 
 compared. 
 
 Another possibility, suggested by 
 Dick and Ewing (,9), is that ideologi- 
 cal measures, such as yield strength 
 and plastic viscosity, be used to do- 
 scribe fundamental physical properties 
 of activated sludge. 
 
 It is realized that some of these 
 alternate methods of measuring sludge 
 properties are influenced by variables 
 such as temperature and concentration 
 just as the sludge volume index is. 
 However, the influence is direct and 
 predictable, whereas the effect of such 
 variables on the SVI is indirect and 
 subtle. This is because the sludge vol- 
 ume index may be influenced to vary- 
 ing degrees by several physical prop- 
 erties of the sludge, and each of these 
 properties is influenced in a different 
 way by a change in temperature or 
 concentration. A thorough investiga- 
 tion is required, with many different 
 sludges, to develop a test which will be 
 a true measure of the sludge settling 
 characteristics and physical proper- 
 ties. Until better methods are devel- 
 oped, SVI still is a useful test for in- 
 plant control. The value of SVI in 
 research and design applications, how- 
 ever, seems to be limited, and other, 
 more basic measures should be used. 
 
 The disadvantages of the alternate 
 parameters suggested for measuring 
 sludge characteristics is that 35 yr of 
 experience are not available to aid in 
 interpreting the meaning of individual 
 measurements as is the case with the 
 sludge volume index. Ultimately, 
 adoption of measures more meaning- 
 ful than the sludge volume index may 
 be helpful for use in control and im- 
 provement of the activated sludge 
 process. Accumulation of experience 
 with more basic measures of sludge 
 properties, perhaps obtained along 
 with SVI measurements, would seem 
 desirable. 
 
 37 
 
Vol. 41, No. 7 
 
 SLUDGE VOLUME INDEX 
 
 Summary 
 
 The sludge volume index test does 
 not measure basic physical properties 
 of activated sludge. As an operational 
 tool for in-plant control, the test is 
 useful, but comparisons of sludge vol- 
 ume index measurements from vari- 
 ous plants are not meaningful. Re- 
 sults of SVI tests cannot be used with 
 certainty to predict settling behavior 
 in full-scale plants. 
 
 Even for in-plant control, alternate 
 measures of settling characteristics 
 probably would be more useful. For 
 research applications, alternate mea- 
 sures of physical properties of acti- 
 vated sludge such as settling velocity, 
 yield strength, and plastic viscosity 
 should be used. 
 
 Acknowledgments 
 
 This work was supported at the Uni- 
 versity of Illinois by Research Grant 
 WP 01011 from the Federal Water 
 Pollution Control Administration and 
 at the University of North Carolina by 
 Research Grant WP 00569 from the 
 National Institutes of Health. 
 
 References 
 
 1. Mohlman, F. W., "The Sludge Index." 
 
 Sew. Works Jour., 6, 1, 119 (Jan. 
 1934). 
 
 2. "Standard Methods for the Examination 
 
 of Water and Wastewater." 12th 
 Ed., Amer. Pub. Health Assn., New 
 York, N. Y. (1965). 
 
 3. Donaldson, W., "Some Notes on the Op- 
 
 eration of Sewage Treatment Works." 
 Sew. Works Jour., 4, 1, 48 (Jan. 
 1932). 
 
 4. Vesilind, P. A., "The Influence of Stir- 
 
 ring in the Thickening of Biological 
 Sludge." Ph.D. thesis, University of 
 North Carolina, Chapel Hill, N. C. 
 (19G8). 
 
 5. Dick, R. I., and Ewing, B. B., "The Rhe 
 
 ology of Activated Sludge." Tins 
 Journal, 39, 4, 543 (Apr. 1907). 
 C. Roberts, E. J., "Thickening— Art or Sci- 
 ence?" Mining Eng., 1, 61 (1949). 
 
 7. Vesilind, P. A., "Discussion of Evalua- 
 
 tion of Activated Sludge Thickening 
 Theories by R. I. Dick and B. B. 
 Ewing. ' ' Jour. San Eng. Div., Proc. 
 Amer. Sor. Civil Engr., 94, SA1, 185 
 (1968). 
 
 8. Dick, R. I., and Ewing, B. B., "Evalua- 
 
 tion of Activated Sludge Thicken- 
 ing Theories." Jour. San. Eng. Div., 
 Proc. Amer. Soc. Civil Engr., 93, SA4, 
 9 (1967). 
 
 9. Rudolfs, W., and Lacy, I. O., "Settling 
 
 and Compacting of Activated Sludge." 
 Sew. Works Jour., 6, 4, 647 (July 
 1934). 
 
 38 
 
APPENDIX I 
 
 THE EFFECT OF POLYMER FLOCCULATION 
 
 ON THE SETTLING BEHAVIOR 
 
 OF ACTIVATED SLUDGE 
 
 by 
 
 George A. Farnsworth 
 
 and 
 
 Richard I . Dick 
 
 Prepubl icat ion Manuscript 
 
 33 
 
THE EFFECT OF THE DEGREE OF FLOCCULATION 
 ON THE BASIC SETTLING BEHAVIOR OF ACTIVATED SLUDGE 
 
 George A. Fransworth and Richard I. Dick 
 
 The efficiency, mode of operation, and cost of treatment by the 
 activated sludge process depend to a considerable degree on the efficiency 
 of the final settling operation. It follows that if close control and 
 economical design of the process are to be achieved, settling tank design 
 practices should be based on a firm conception of the fundamental factors 
 governing the settling and thickening of activated sludge. The more prac- 
 ticable aspects of final settling tank design and the influence of the 
 design on the overall activated sludge process have been considered else- 
 where (Dick, 1970b). The work reported here (Farnsworth, I967) was part 
 of a more fundamental investigation of the settling properties of acti- 
 vated sludge. 
 
 ACTIVATED SLUDGE SEDIMENTATION 
 
 Sedimentation of a concentrated "ideal" suspension comprised of par- 
 ticles of uniform size and shape has been described by Kynch (1952). His 
 model has been shown to provide an accurate description of the sedimenta- 
 tion of suspensions such as glass beads (Shannon and Tory, 1965) and sand 
 grains (Dick and Ewing, 1967) ' n which fluid drag and gravity are the only 
 forces acting on the particles. However, considerable deviations from 
 Kynch's theory have been found with flocculant suspensions such as clay 
 (Gaudin and Fuerstena, 1962) and activated sludge (Dick and Ewing, 1967) . 
 
 George A. Farnsworth is Public Works Officer, Centerville Beach Naval 
 Facility, Ferndale, California, and Richard I. Dick is Professor of Civil 
 Engineering at the University of Illinois at Urbana. 
 
 *t0 
 
Dick and Ewing (1967) showed that existence of a continuous structure within 
 activated sludge even at normal mixed liquor suspended solids concentrations 
 caused a dependence of settling velocity upon sludge depth. Experimentally, 
 they found the relationship between sludge depth, D, and settling velocity, 
 v, to be 
 
 where R is the intercept on the ordinate axis of a plot of D/v as a function 
 
 of D and S is the slope of the curve. The form of the relationship has been 
 
 confirmed by Vesilind (1968) and others. When R in equation (l) is zero or 
 as D approaches infinity, 
 
 v - v u - I (2) 
 
 where v is the "ultimate settling velocity" or the velocity at which the 
 interface would subside if it received no support from underlying sludge. 
 Hence, the reciprocal of S is a measure of the settling velocity of sludge 
 if it conformed to Kynch's theory, and R, called the retardation factor, 
 is a measure of the deviation of settling characteristics from those of an 
 ideal suspension. The retardation factor has been found (Dick and Ewing, 
 1967) to vary with suspended solids concentration, e, according to the 
 relationsh ip 
 
 R = ge hc (3) 
 
 where g and h are constants for a given sludge. 
 
 OBJECTIVES 
 The objective of the study was to determine the effects of changes in 
 the degree of flocculation of activated sludge on its settling characteris- 
 tics. An extent of flocculation was adjusted by adding a synthetic organic 
 
 h) 
 
polyelectrolyte and the effect on zone settling velocity as a function of 
 initial sludge depth was observed in order to determine the ultimate settling 
 velocity and retardation factor for each condition of f locculation . 
 
 Inasmuch as the value of g in equation (3) must be zero for an ideal 
 suspension, and because the values of g and h were previously found to be 
 related to the biological (Dick and Ewing, 19&7) ar| d rheological (Dick, 1970a) 
 characteristics of the sludge, it was anticipated that the effect of the 
 degree of f locculation on the settling characteristics of activated sludge 
 might be interpreted in terms of the effects on the constants in equation (3) 
 
 The main purpose of the work was to gain insight into the factors which 
 control the basic physical behavior of activated sludge. Some information 
 on the possible benefits of polymer induced sludge flocculation and on the 
 interpretation of sedimentation tests using polymer conditioned sludge were 
 also obtained. 
 
 PROCEDURES 
 
 A major difficulty in the work was to express quantitatively the degree 
 to which sludge solids were agglomerated. Because the degree of flocculation 
 relates to the extent to which the sludge solids agglomerate, and because 
 this in turn influences the clarity of the liquid eliminated, turbidity of 
 the supernatant liquid was used as a measure of the degree of flocculation 
 This parameter has been found by others to be sensitive to changes in floc- 
 culent dose (for example, by Black and Chen, 1965). It must be admitted 
 however that use of the turbidity measurement was primarily an expedient 
 dictated by the lack, at present, of a better measure of the degree of floc- 
 culation. A nephelometer was used for turbidity measurements. 
 
 Model 1800 manufactured by the Hach Chemical Col, Ames, Iowa. 
 
All sludge used in the study was taken from the Urbana-Champaign, 
 Illinois Sanitary District main treatment plant. The effect of varying 
 doses of a cationic polyelectrolyte, purifloc C 3 1 , on flocculation of 
 the activated sludge is illustrated in Figure 1. Data are from tests in 
 a six-place jar test apparatus using 15 min at 30 rpm followed by 30 min 
 of quiescence. For purposes of this study, flocculant doses in the initial 
 steep portion of the curve preceding the dosage for optimum flocculation 
 were used in order to achieve maximum differences in the degree of floccu- 
 lation. As would be expected, pH exerted a significant influence on 
 flocculation. pH was monitored throughout the study to ascertain that all 
 tests were conducted at the same value (7-1); however, pH adjustment was 
 not found to be necessary. 
 
 Sedimentation tests were conducted in 3-5 in. diameter settling columns 
 using the equipment and procedures described by Dick and Ewing (1967). The 
 suspended solids were determined using the Gooch crucible-glass fiber filter 
 method described by Gratteau and Dick (1968). 
 
 To assure that all tests were conducted on a sludge with the same 
 basic physical properties, it was necessary to obtain all sludge used in 
 the tests at one time and to complete the experiments in as short a period 
 as possible. These requirements limited the study in two ways. Firstly, 
 the number of different flocculent conditions had to be limited to three 
 (0.0, 0.97, and 2.43 mg/purifloc C31/g sludge solids). Secondly, it was 
 necessary to reuse sludge dosed with polymer for all settling tests at a 
 given polymer dosage. Polymer was applied to the most concentrated sludge 
 
 A product of Dow Chemical Co., Midland, Michigan. 
 Manufactured by the Phipps and Byrd Co., Richmond, Virginia 
 
 43 
 
sample and lesser concentrations were achieved by adding supernatant liquid 
 from settled undosed sludge and mixed by aeration. It was recognized that 
 reuse of dosed sludge was highly undesirable; however, random duplication of 
 settling tests gave similar results. 
 
 At each of the three flocculent conditions, settling tests were conducted 
 at four different suspended solids concentrations. At each suspended solids 
 concentration, five or six settling tests at various depths between 1 and 
 6 ft were used to define the relationship between settling velocity and 
 depth. Also, two 3-5 ft control depths were obtained at each concentration 
 to monitor for any temporal variation in sludge quality. 
 RESULTS AND DISCUSSION 
 
 At all degrees of flocculation and at all suspended solids concentrations 
 the data conformed well to equation (l). Values of the retardation factor, 
 as given by application of that formula, are shown for each of the three 
 flocculent conditions in Figure 2. The influence of the degree of floccula- 
 tion on the ultimate settling velocity (1/S) is illustrated in Figure 3. It 
 is seen that the effect of flocculation was on the ideal settling velocity 
 of the sludge particles and not on the extent of deviation from ideal con- 
 ditions (as measured by the retardation factor). This suggests that the struc- 
 tural forces between particles are not significantly influenced by polymers. 
 This is consistent with results reported by Dell and Keleghan (1970) who found 
 that neither flocculent type nor dosage influenced the compressibility of 
 clay suspensions. 
 
 The fact that polymers influence ultimate settling velocity but not the 
 retardation factor means that the beneficial effect of polymers will be 
 masked at shallow sludge depths if sludges have an appreciable retardation 
 factor. This is illustrated in Figures k and 5 where the effect of polymer 
 
dosage on settling velocity at 5.0 and 2,5 ft initial depths is illustrated. 
 Clearly, the difference between flocculated and unf locculated samples de- 
 creased as depth diminished. This means that in laboratory evaluation of 
 the feasibility of polymer addition it would be important to use effective 
 depths comparable to those expected in the field. 
 
 It is interesting to note that the influence of polymers on the set- 
 tling characteristics of the activated sludge was highly dependent upon 
 suspended solids concentration (Figures 3, ^ , and 5). With the particular 
 sludge used, polymers improved settling velocity for concentrations less 
 than about 5250 mg/1 but caused reduction in settling velocity at higher 
 concentrations . 
 
 SUMMARY AND CONCLUSIONS 
 
 The influence of the degree of flocculation of activated sludge was 
 investigated to gain insight into the factors which control the fundamental 
 physical properties of the material. A synthetic organic polyelectrolyte 
 was used to vary the degree of flocculation, and settling velocity was the 
 principal physical property observed. 
 
 Flocculation significantly effects sludge interface settling velocity 
 at large sludge depths and has less effect on settling velocity at low 
 depths. This means that the principal effect of the degree of sludge floc- 
 culation is on the ultimate settling velocity of sludge and not on the mag- 
 nitude of interpart icle forces (or the retardation factor). While the purpose 
 of this study was not to investigate the feasibility of using artificial 
 flocculents in the activated sludge process, several cautions can be offered 
 to those who are working in this area. Influence of polymers on settling 
 velocity depends on sludge depth and hence laboratory tests must be conducted 
 at effective depths comparable to those expected in the field. Furthermore, 
 
 h5 
 
the same polymer may improve settling characteristics of sludge at one con- 
 centration but cause slower settling at another concentration. Hence, an 
 evaluation must extend over a wide range of suspended solids concentrations. 
 
 Implications of the work, are that polymer conditioning of activated 
 sludge will not appreciably alter the concentration of sludge which is ulti- 
 mately achievable by gravity thickening. However, the capacity of a thickener 
 may be increased by polymer flocculation because settling velocities may be 
 enhanced, and the concentration of the underflow from an activated sludge 
 final settling tank may be increased if the tank is not operating near the 
 ultimately achievable concentration. 
 
 ACKNOWLEDGEMENTS 
 This study was supported in part by Research Grant 17070 DJR from the 
 Federal Water Quality Administration and by a traineeship from the U. S. 
 Public Health Service. 
 
 REFERENCES 
 
 Dell, C. C. and Keloghan, W. T., "Compress ibi 1 i ty of Flocculated Clay 
 Suspensions," Filtration and Separation , J_, k, 419-^21 (1970). 
 
 Dick, R. I., "Thickening Characteristics of Activated Sludge," in Advances 
 in Water Pollution Re s earch , Proceedings of Fourth International 
 Conference on Water Pollution Research, Prague, 1969, 625-6^2 (1970a). 
 
 Dick, R. I., "Role of Activated Sludge Final Settling Tanks," Journal 
 Sani t ary Engineering Division American Society of Civil Engineers , 
 H7~SA2, Jt23-436 (1970b). 
 
 Dick, R. I., and Ewing, B. B., "Evaluation of Activated Sludge Thickening 
 Theories," Journal Sanitary Engineering Division American Society of 
 Civil Engineers , 93, SAA, 9-29 (1967). 
 
 Farnsworth, G. A., "The Effect of Induced Flocculation on the Settling and 
 Thickening Behavior of Activated Sludge," Civil Engineering Studies, 
 Sanitary Engineering Series No. hi, 5^ pp (1967). 
 
 k6 
 
Gaudin, A. M., and Fuerstena, N. C, "Experimental and Mathematical Model 
 of Thickening," Trans. Society Mining Eng. , 223 , 122-229 (1962). 
 
 Gratteau, J. C, and Dick, R. I., "Activated Sludge Suspended Solids 
 Determinations," Water and Sewage Works , 115, 10, 468-^72 ( 1 968) . 
 
 Kynch , G. J., "A Theory of Sedimentation," Trans. Faraday Society , 48 , 
 166-176 (1952) . 
 
 Shannon, P. T., and Tory, E. M., "Settling of Slurries," Industria l and 
 E ngineering Chemistry , 57 , 18-25 (1965). 
 
 Vesilind, P. A., "The Influence of Stirring in the Thickening of Biological 
 Sludge," Thesis submitted in partial fulfillment of the requirements 
 for the degree of Doctor of Philosophy, University of North Carolina 
 (1968). 
 
 hi 
 
w 
 3 
 
 200 «oo 6.00 aoo iooo 
 
 Purifioc C3I Dose, mg Purifkx/gm Solids 
 
 Fig. ] Effect of Polymer Dose on Degree 
 of F 1 occu 1 a t ion 
 
 1*8 
 
500 
 
 300 
 
 100 
 
 50 
 
 -O- mg/g Punfloc C 31 
 -A- 97 mg/g Punfloc C3I 
 -O— 2 43 mg/g Punfloc C 31 
 
 2000 4000 
 
 Solids Concentration, mg/£ 
 
 6000 
 
 Fig. 2 Relation Between Suspended Solids 
 
 Concentration and Retardation Factor 
 
 49 
 
?0 
 
 65 
 
 D rr>g/g Puntloc C 3l 
 • '• 9' mg/g Punfloc C 3i 
 T c 4 3 mg/g Pun'ioc C 3' 
 
 mg, g 
 
 20OO 40O0 600O 
 
 Solids Concentration, mg/jC 
 
 Fig. 3 Influence of Degree of Flocculation on 
 Ultimate Settling Velocity 
 
 c 
 
 E 05 
 
 c 
 
 nng 
 
 'g p j 
 
 nfloc C 3! 
 
 
 
 - 
 
 ^ 9 7 
 
 mg/g 
 
 Pjnfioc C 3l 
 
 
 
 :: 
 
 2 43 
 
 mg/g 
 
 Pjr,f,oc C 3> 
 
 
 
 - 
 
 
 
 N \ Vs. 
 
 mg/g \j A 
 
 -0 97 mg, g 
 
 
 
 
 
 \S 
 
 
 
 
 
 Z 43 mg/c 
 
 »A 
 
 On 
 \\ 
 
 2000 4000 6000 
 
 Solids Concentration, mq/JL 
 Fig. k Influence of Degree of Flocculation on 
 Settling Velocity at an Initial Sludge 
 Depth of 5.00 ft 
 
 50 
 
in 
 
 
 mg/g Purifloc C 31 
 
 
 
 A 97 mg/g Purifloc C 31 
 
 
 15 
 
 ~n 2 43 mg/g Purifloc C 31 
 
 
 10 
 
 X > .— 97 
 
 \v 
 
 mg^j — . ^v ^ 
 
 mg/g 
 
 
 
 005 
 
 2 43 mg/g 
 
 \ 
 
 
 
 .. I 1 - 
 
 1 
 
 2000 4000 6000 
 
 Solids Concentration , mq/4 
 
 Fig. 5 Influence of Degree of Flocculation on 
 Settling Velocity at an Initial Sludge 
 Depth of 2.50 ft 
 
 51 
 
APPENDIX I I I 
 
 THICKENING CHARACTERISTICS OF ACTIVATED SLUDGE 
 
 by 
 Richard I . Dick 
 
 Reproduced from 
 
 Advances in Water Pollution Research 
 
 Proceedings of kth International Conference 
 
 on Water Pollution Research 
 
 Held in Prague, Czechoslovakia, April 1969, Pages 625-642 
 
 1970 
 
 52 
 
Reprinted from 
 
 ADVANCES IN WATER POLLUTION RESEARCH 
 
 Proceedings of the 4th International Conference held in 
 
 Prague 1969 
 
 Edited by 
 
 S. H. JENKINS 
 
 PERGAMON PRESS - OXFORD AND NEW YORK 1969 
 
 THICKENING CHARACTERISTICS OF 
 ACTIVATED SLUDGE 
 
 Richard I. Dick 
 
 University of Illinois, Urbana, Illinois, U.S.A. 
 
 INTRODUCTION 
 
 Consolidation of sludge to high concentrations in the final settling tank is necessary 
 for economical performance of the activated sludge process. The amount of organic 
 material which can be applied to an activated sludge treatment plant depends on the 
 mixed liquor suspended solids concentration, and this concentration is related to the 
 degree of thickening accomplished in the final settling tank. Furthermore, the cost of 
 disposing of the excess solids synthesized in the process depends on the concentration 
 at which they are wasted from the final settling tank. 
 
 Improvement in the ability of final settling tanks to accomplish their clarification 
 and thickening functions could decrease significantly the cost of new plants and 
 increase the capacity of existing ones. A basic understanding of the settling behavior 
 of activated sludge must be acquired to permit development of rational and workable 
 design criteria to accomplish such improvement. Attention is given in this paper to 
 basic factors which influence the solids handling capacity of activated sludge. 
 
 CONVENTIONAL THEORIES OF THICKENING 
 
 Coe and Clevenger (1916) introduced the concept of solids handling capacity. 
 Each concentration of a suspension was shown to have a certain capacity to discharge 
 its solids. The minimum, or limiting, solids handling capacity of the suspension must 
 be used to establish the area of a settling basin if thickening of underflow to the 
 desired concentration is to be accomplished. If insufficient area is provided, the solids 
 will queue above the location of the rate-limiting concentration of the suspension, 
 and ultimately pass over the effluent weir along with the "clarified" effluent. 
 
 Confirmation and clarification of Coe and Clevenger's concepts were provided by 
 Kynch (1952). He presented a mathematical analysis of thickening based, not on 
 consideration of fundamental factors controlling subsidence of suspensions, but, 
 rather, on the arbitrary assumption that "at any point in a dispersion the velocity of 
 fall of a particle depends only on the local concentration of particles." The analysis 
 serves as the fundamental basis for present practices for determining the required 
 area of thickeners. 
 
 When concentration increases to the extent that particles begin to rest on one another 
 (when the "compression point" is reached) different thickening concepts are presumed 
 to apply (Roberts, 1949). While the solids handling capacity of the "freely" settling 
 suspension is considered to establish the required area of a thickener, consolidation 
 characteristics are considered to dictate the volume required. This paper only deals 
 with matters related to the area requirement. 
 
 53 
 
Richard I. Dick 
 
 More complete reviews of conventional thickening theories have been presented by 
 Behn (1957), Fitch (1962), and Dick and Ewing (1967b). 
 
 APPLICABILITY OF THICKENING THEORIES 
 
 The Kynch analysis forms a valuable framework for understanding the behavior 
 of freely settling suspensions. For suspensions which conform to Kynch's assumptions, 
 the analysis affords a precise description of settling behavior (Shannon et ai, 1964, 
 Dick and Ewing, 1967b). Suspensions with physical characteristics which are less 
 ideal may not conform to settling behavior predicted by the Kynch analysis because 
 they violate assumptions on which the mathematical analysis is based. Still the Kynch 
 analysis serves as a valuable basis for identifying deviant settling behavior and for 
 intercepting causes of such deviations. 
 
 The settling behavior of activated sludge has been found to be inconsistent with the 
 Kynch analysis (Dick and Ewing, 1967b). The initial "free 1 constant-rate settling 
 velocity of comparatively dilute concentrations of activated sludge varies with sludge 
 depth according to the expression 
 
 n 
 
 (1) 
 
 where D is the initial depth of uniformly dispersed sludge, and R and 5 are constants 
 which characterize the settling behavior of a given concentration of a particular 
 activated sludge. The magnitude of R, the retardation factor, is a measure of the 
 extent to which the settling velocity of activated sludge is retarded by causes not 
 considered in the Kynch analysis. The settling velocity that a sludge would have if 
 it settled like the ideal suspension in prevalent theories of thickening is given by 1 /S 
 and is termed the ultimate settling velocity, v u . If settling velocity depended only on 
 concentration in accordance with prevailing concepts, there would be no dependence 
 on depth ; that is R would be zero. 
 
 An appreciation of the significance of this deviant settling behavior of activated 
 sludge may be gained from Fig. 1 where observed settling velocities for several 
 activated sludges are plotted as a percentage of v u . The purpose of the present study 
 was to investigate the cause of this non-deal settling behavior. 
 
 THICKENING MODEL 
 
 To assume, with Kynch, that the settling velocity of a given suspension is a function 
 only of local particle concentration is equivalent to assuming that only fluid forces 
 act on sludge particles. Although Kynch did not explore the nature of these forces, 
 the relationship between particle concentration and settling velocity of suspensions 
 has been developed by others — notably Vand (1948). Inasmuch as the settling charac- 
 teristics of activated sludge do not conform to these concepts, even at suspended solids 
 concentrations found in mixed liquor, it is necessary to introduce additional forces 
 which act on sludge particles to influence their settling behavior. In this conceptual 
 model, the additional forces of significance are considered to be forces transmitted 
 through interparticle contacts. Conventionally such forces are not considered to 
 exist until the compaction point is reached and compression begins. 
 
 5h 
 
Thickening Characteristics of Activated Sludge 
 
 Figure 2 shows a mass of sludge of unit cross-sectional area and depth, D, along 
 with representation of the forces considered to be acting on the sludge. The free body 
 diagram shows the sludge at the onset of sedimentation when the solids are uniformly 
 distributed. The sludge is considered to settle en masse with floe particles maintaining 
 their relative positions one to another. As sedimentation occurs, the flocculent structure 
 
 - / 
 
 — 1 1 1 — T 
 
 / 4^ 
 
 -1 / 
 
 
 
 i i 
 
 0.5 1.0 1.5 2.0 
 
 DEPTH , m 
 
 Fig. 1. Extent of deviation of activated sludge settling behavior from that 
 of an ideal suspension 
 
 is envisioned to collapse against the bottom of the container. Acceleration is considered 
 to have ceased so that the net force on the sludge mass is zero, or 
 
 Fw = F B + F d + Fs 
 
 (2) 
 
 where F w is the downward force due to weight of sludge solids, F B is the upward 
 force due to buoyancy, Fd is the drag force resisting downward movement of sludge, 
 and Fs is the resistance to subsidence due to the compressive strength of the structure 
 of flocculent particles. 
 
 The forces due to the weight and buoyancy of the sludge may be combined in a single 
 force, F E , representing the effective weight of the solids in the liquid. For a particular 
 
 55 
 
Richard 1. Dick 
 
 sludge, the effective weight of solids per unit cross-sectional area will vary directly 
 with depth and concentration, so that 
 
 F E = K\cD 
 
 (3) 
 
 where the constant, Ku includes basic physical properties of the solids and suspending 
 liquid. 
 
 UNIT AREA 
 
 Fig. 2. Definition sketch for thickening model 
 
 The drag force acting on the descending sludge, or, its equivalent, the resistance 
 offered by the sludge particles to the escape of displaced fluid, depends on the nature 
 of the flow regime. It is estimated that the order of magnitude of Reynolds number 
 for flow through activated sludge is near the approximate upper limit for laminar 
 flow through porous media. On this basis, it would be expected that, in the micro- 
 structure of flow through sludge, localized, turbulence occurs to the extent that 
 inertial forces contribute significantly to the total drag on the sludge mass (Schneebeli, 
 1955). In the absence of precise knowledge of the flow regime, versions of the 
 model were developed for both laminar and turbulent flow to describe the two 
 possible extremes in fluid behavior. The subscripts L and T are used to distinguish 
 the two versions. 
 
 56 
 
Thickening Characteristics of Activated Sludge 
 
 The value of Fn with laminar interstitial flow may be expressed as the sum of drag 
 forces on the individual floe particles which comprise the sludge mass. The general 
 expression for the drag resisting movement of a body through a fluid, 
 
 F D = C D A P -?f- (4) 
 
 may be applied. In laminar flow the coefficient of drag, C D , is related inversely to 
 Reynolds number, and hence drag varies with the first power of velocity. Also, 
 for a single concentration of a particular sludge the collective projected area of the 
 flocculated particles in the sludge mass, A, is a direct function of depth. Hence* 
 
 F Dl = K 2 Dv L (5) 
 
 where K 2 reflects physical properties of the fluid and parameters related to the floccu- 
 Ient nature of the sludge and is constant only for a particular sludge at a single con- 
 centration and temperature, and vl is a representative velocity which may be taken 
 as the interface subsidence rate. 
 
 If, as an extreme, displaced liquid is considered to escape by fully-developed 
 turbulent flow, fluid drag becomes independent of Reynolds number and hence is 
 proportional to the square of the subsidence velocity, orf 
 
 F Dt = K 3 Dv% (6) 
 
 The constant reflects physical properties of the suspending medium and physical 
 characteristics of the solids which establish the size of flow channels and is constant 
 only for one concentration of a given sludge. 
 
 The subsidence rate of the sludge mass may be evaluated by combining equations 
 (2), (3) and (5) (or 6). For the laminar flow model, 
 
 K\ Fs ,-,, 
 
 VL = K, C -K^D (7) 
 
 and for the turbulent flow model, 
 
 V \K 3 K s Dj 
 
 CHARACTERISTICS OF THICKENING MODEL 
 
 In either the laminar or turbulent flow model, settling velocity is dependent solely 
 on particle concentration (in accordance with the Kynch analysis) only when the second 
 term on the right side of equations (7) and (8) is insignificant. Settling velocity is also 
 a function of depth whenever the suspension possesses structure which must collapse 
 to permit subsidence. The extent of retardation caused by a given value of Fs is 
 dependent upon the magnitude of K 2 (or K s in the turbulent version). A suspension 
 
 * If it is preferred to visualize laminar flow in subsiding sludge as analogous to flow through porous 
 
 media, equation (5) may be derived from Darcy's law. 
 
 t Equation (6) may also be derived from the Darcy-Weisbach expression for flow through pipes 
 
 57 
 
Richard I. Dick 
 
 which settles slowly in the absence of retardation due to Fs will be relatively less 
 affected by structural retardation than one which settles rapidly when unretarded. 
 
 Figure 3 shows the basic form of either version of the model. The deviation from 
 prevailing thickening theory when Fs has a finite value and the resemblance of the 
 deviant behavior to observed activated sludge settling characteristics is illustrated. 
 
 /—MODEL WITH F s = 
 
 
 
 
 ^f — MODEL WITH 
 
 
 
 
 ^"F s = FINITE VALUE 
 
 
 // 
 
 
 
 / 
 
 
 
 
 /i 
 
 1 i 
 
 
 FORM 
 
 OF EXPERIMENTAL DATA 
 
 
 
 
 DEPTH, D 
 Fig. 3. General form of thickening model 
 
 The model is, however, somewhat defective in that it indicates that, at some small 
 depth, sludge will stand without settling. This occurs when the depth is so small that 
 the effective weight of the sludge equals the structural support force. Then the drag 
 force — and hence the settling velocity — must equal zero. Although experimental 
 determination of settling velocities of shallow depths of activated sludge is difficult, a 
 zone of clarified water develops in time at the top of the shallowest of sludge depths. 
 
 The defect in the model is understandable. Consolidation of the structure within 
 activated sludge is probably a viscoelastic process which continues at low loadings 
 (i.e. shallow depths) by the process of creep or plastic flow. From time to time con- 
 nections between adjacent floe particles may break and reform in an unstressed posi- 
 tion. This results in a redistribution of interparticle stresses and contributes to 
 subsequent failure of other connections. 
 
 The time-dependent failure of the structure of activated sludge is likely related to its 
 lyophilic nature. Particles in activated sludge probably are not in actual contact, 
 but are separated by liquid films. The deformation characteristics of sludge might, 
 then, be related to the nature of the interparticle films rather than, or in addition to, 
 mechanical properties of the particles. Such films are comprised of water which has a 
 high degree of orientation and exists in a quasi-solid state (Frank and Evans, 1945). 
 A matrix of such quasi-solid films could account for the difference between observed 
 settling behavior and the form of the model, for deformation to permit subsidence 
 would occur under all conditions — not just when the equivalent weight of sludge 
 exceeded Fs. 
 
 58 
 
Thickening Characteristics of Activated Sludge 
 COMPARISON WITH EXPERIMENTAL SETTLING DATA 
 
 For convenience, equations (7) and (8) may be written as 
 
 b L 
 v L = a L -- 
 
 and 
 
 b T \ 
 D 
 
 v T 
 
 j{ GT 
 
 (9) 
 
 (10) 
 
 where the values of a and b are apparent from comparison with the previous equations. 
 Direct determination of the constants a and b is not possible, for Fs, Ki, and K-a 
 cannot be evaluated directly. The values of a and b, however, can be computed from 
 the experimentally determinable parameters R and 5 (equation 1). 
 
 MODEL WITH 
 
 MODEL WITH 
 F„=0 
 
 FORM OF EXPERIMENTAL DATA 
 
 DEPTH, D 
 
 Fig. 4. Form of D/v vs. D plots 
 
 From equation (1) it is seen that S is the slope of a plot of D\v as a function of D, 
 and R is the intercept on the D\v axis as illustrated in Fig. 4. In terms of the laminar 
 flow model, then, 
 
 d ( D\ d 
 
 S L 
 
 dD 
 
 \v L ) dD\ b L ) 
 
 (11) 
 
 Differentiation gives 
 
 S L 
 
 a L D* - 2b L D 
 (a L D - b L f 
 
 (12) 
 
 Because of the defect described in the previous section, Sl is not a constant, but is 
 depth dependent (the Djv vs. D plot for the model only asymptotically approaches 
 
 59 
 
Richard I. Dick 
 
 the straight line found experimentally). The retardation factor, R, for the model may 
 be taken as the back extrapolate of the tangent to any point on the curve in Fig. 4. 
 Thus, 
 
 Rl 
 
 DS l 
 
 Substituting equations (9) and (12) and simplifying gives 
 
 Rl = 
 
 b L D* 
 (a L D - b L f 
 
 (13) 
 
 (14) 
 
 The values of S and R for the turbulent model may be derived in a similar fashion. 
 
 Representative batch settling tests at various initial sludge depths were conducted 
 to generate DjV vs. D plots for various concentrations of several activated sludges. 
 Values of R and S were then determined from the data and a and b were evaluated by 
 simultaneous solution of equations (13) and (14) (or their counterparts in the turbu- 
 lent version) for selected values of D. Typical results are shown in Figs. 5 and 6 for 
 two concentrations of activated sludge from a treatment plant designated as A. 
 
 Because the constants in the model were computed from observed data, the figures 
 do not constitute confirmation of the validity of the model. However, they demon- 
 strate the basic agreement of the form of the model with experimental data. At low 
 degrees of retardation (Fig. 5, R = 0.8 min) differences between the experimental 
 and model curves due to the defect in the model were minimal, while differences be- 
 come more pronounced with higher degrees of retardation (Fig. 6, R = 6.3 min). 
 
 RELATIONSHIP BETWEEN SETTLING AND RHEOLOGICAL BEHAVIOR 
 
 To confirm the validity of the thickening model, it is necessary to show that activated 
 sludge does, indeed, possess internal structure at comparatively dilute suspended 
 solids concentrations and that the effect of such structure is quantitatively described 
 by the model. To do this, sedimentation and rheological properties of activated sludge 
 from three municipal activated sludge plants of widely varying characteristics were 
 examined concurrently. The experiments were conducted over the range of suspended 
 solids concentrations which might occur in the final settling tank in the plant from 
 which each sludge was collected. Table 1 shows the conditions which existed in the 
 three waste treatment plants at the time samples were obtained. 
 
 Table 1 . Properties of Activated Sludges Used 
 
 Source of sludge 
 
 Suspended solids concentration Sludge volume index 
 mg/1. (ml/g) 
 
 Mixed liquor Return sludge 
 
 Organic loading 
 g B.O.D./day 
 
 g sludge solids 
 
 Plant A 
 
 2280 
 
 8230 
 
 75 
 
 0.14 
 
 Plant B 
 
 4350 
 
 7420 
 
 55 
 
 0.06 
 
 Plant C 
 
 1225 
 
 2820 
 
 300 
 
 1.35 
 
 60 
 
Thickening Characteristics of Activated Sludge 
 
 i 1 r 
 
 v 8 
 
 Q Q 
 55 
 
 Q cr -J H 
 
 -«e m o o 
 
 O 2 5 
 
 »- 
 Z 
 
 < < 
 
 Q tf) CO 
 
 UJ 
 
 Ul UJ UJ 
 
 to o o 
 
 oq O O 
 
 OSS 
 
 
 oas/ujo «AllD0n3A 9Nnil3S. 
 
 61 
 
Richard I. Dick 
 
 O Q 
 
 < < 
 
 in v> 
 
 Q Q 
 
 o o 
 
 2 5 
 
 9 m 8 
 
 Q O Q 
 
 (/) Q Q Q 
 fflOOO 
 O S 5 2 
 
 I! 
 
 < 
 
 It 
 
 si 
 
 09S/UJ0 *AllOOn3A 9NH113S 
 
 62 
 
Thickening Characteristics of Activated Sludge 
 
 The sedimentation properties of interest were those which measure the deviations 
 from ideal settling behavior — R and 5. The comparative settling characteristics of the 
 three sludges have been described previously (Dick and Ewing, 1967b). 
 
 The rheological property of interest was one which could account for internal 
 rigidity — the yield strength. Measurements were made with a viscometer adapted for 
 use with activated sludge. Results of the rheological studies are described elsewhere 
 (Dick and Ewing, 1967a). 
 
 For all sludges both the yield strength and retardation factor varied exponentially 
 with suspended solids concentration. It follows that a relationship of the form 
 
 R = mr n y (15) 
 
 relates R and the yield strength, r y , of activated sludges where m and n are constants. 
 Figure 7 illustrates the relationship. The equation does not, however, necessarily 
 show a cause-and-effect relationship but could merely reflect a mutual dependence on 
 suspended solids concentration. 
 
 A more meaningful indication of the relationship between observed settling be- 
 havior and the rheological properties of sludge can be obtained by modifying the 
 mathematical expression of the thickening model to give a relationship between yield 
 strength and thickening properties. In the case of the laminar version of the model, 
 the equality of the denominators in equations (12) and (14) results in 
 
 a L D 2 - 2b L D = b L D 2 
 Sl Rl 
 
 Simplifying and substituting the expressions for at and bh gives 
 
 cDRi 
 
 (16) 
 
 *s L 
 
 Msifk) 
 
 All values on the right side of the equation can be determined experimentally. 
 
 It remains to express the structural support force in terms of yield strength. Although 
 a quantitative relationship between the yield strength of a substance in shear and its 
 confined compressive strength cannot be rigorously formulated, it is reasonable to 
 expect that the same properties which determined the yield strength of a suspension 
 are related to the structural support force, Fs. If the relationship is presumed to be 
 linear 
 
 ^-kA^-^^-A (18) 
 
 / cDR L \ 
 \ S L D + 2R L ) 
 
 in which K4 is K\ modified to account for the relationship between Fs and r y . Similar 
 derivation of the relationship between the settling and rheological parameters in the 
 turbulent version of the model gives 
 
 K 4 
 
 (_2cDR T \ (19) 
 
 63 
 
Richard I. Dick 
 
 The agreement between the rheological properties of activated sludge as measured 
 in the viscometer and as computed from observed settling data based on the thickening 
 model is illustrated in Fig. 8 and 9. The absolute value of the quantity plotted on the 
 abscissa is irrelevant since the value of the constant, Ka, which relates it to r y cannot 
 be measured. It is apparent, however, from the linear nature of the curves that the 
 
 I I I I I I I I I I 
 
 1 1 — I I I I I I I 
 
 0.03 0.05 0.07 0.1 0.3 
 
 YIELD STRENGTH, dynes/sq cm 
 Fig. 7. Relationship between retardation factor and yield strength 
 
 Gh 
 
Thickening Characteristics of Activated Sludge 
 
 0.2 3 0.4 3 
 
 cDRASD + 2R) 
 
 Fig. 8. Relationship between observed yield strength and relative magnitude 
 of structural support — laminar model 
 
 UJ 0.2- 
 
 o.s o.s 
 
 2cOR/(SD + 3R) 
 
 Fig. 9. Relationship between observed yield strength and relative magnitude 
 of structural support — turbulent model 
 
 0.7 08 
 
 1 1 r — i 
 
 — r 1 ■ i 
 
 Plant A^. r 
 
 V A 
 
 \— Plant B 
 
 
 - 
 
 _X v - Plant C 
 
 ...J . 1 . , i , ,. 1 
 
 ill. 
 
 65 
 
Richard I. Dick 
 
 relative magnitude of the structural support force computed from observed settling 
 data is in basic agreement with the measured yield strength. 
 
 The upper portion of the curves in Figs. 8 and 9 (corresponding to higher suspended 
 solids concentrations) is not straight. In computing the relative magnitude of t,, 
 from settling data using equations (18) and (19), the value of D was arbitrarily taken 
 as 1 .22 m (4 ft) — the approximate mean of the settling depths observed experimentally. 
 Because the defect in the model manifested itself at high degrees of retardation (Fig. 6), 
 the effect of this assumption was to reduce the computed structural support force at 
 high suspended solids concentrations but to leave it relatively unaffected at low con- 
 centrations. Thus the nonlinear portion of the curves in Figs. 8 and 9 was expected. 
 
 It will be recalled that the postulated model of sedimentation was something 
 intermediate between the two versions of the mathematical model. The basic correla- 
 tion between yield strength and observed settling characteristics exists for both the 
 laminar (Fig. 8) and turbulent (Fig. 9) versions of the model. This is not to imply 
 that both versions of the model are correct, but that either is plausible. 
 
 SUMMARY AND CONCLUSIONS 
 
 The settling characteristics of activated sludge differ from those of the ideal sus- 
 pension on which gravity thickening theories are based. In this study, the "non- 
 ideal" behavior of activated sludge was interpreted in terms of the rheological proper- 
 ties of the suspension. 
 
 A conceptual model of the hypothesized mechanism of subsidence of activated 
 sludge was developed based on an analysis of forces acting on a mass of sludge. The 
 resulting mathematical expression was inconsistent with prevailing theories of thicken- 
 ing, but was in close agreement with the observed settling behavior of activated sludge. 
 
 The relative magnitude of the interparticle force, as computed from laboratory 
 settling data by use of the mathematical models was shown to be related to the yield 
 strength of the same sludge as measured with a viscometer. It is concluded that the 
 reason activated sludge fails to conform to prevailing theories of thickening is that 
 it has a yield strength at ordinary mixed liquor suspended solids concentrations. This 
 existence of interparticle forces causes a reduction in settling velocity which is not 
 considered in present theories relating to the area requirement for thickeners. 
 
 Contrary to prevailing thickening theories, the area of the thickening portion of a 
 settling basin need not be considered to be inalterably established by the settling 
 velocity of the rate-limiting concentration of activated sludge. The area can be re- 
 duced, or the capacity of an existing basin can be increased, by minimizing the re- 
 duction in settling velocity due to interparticle forces within the sludge. This can be 
 done physically by controlling the depth and mixing conditions in the settling basin, 
 or it can be done biologically by altering conditions in the activated sludge process to 
 minimize the yield strength of the sludge. 
 
 REFERENCES 
 
 Behn, V. C. (1957) Settling behavior of waste suspensions, J. Sanit. Engng. Div. Am. Soc. Civ. Engrs, 
 
 83, SA5, 1-20. 
 Coe, H. S. and Clevenger, G. H. (1916) Methods for determining the capacities of slime settling 
 
 tanks, Trans. Am. Inst. Min. Engrs, 55, 356-84. 
 
 66 
 
Thickening Characteristics of Activated Sludge 
 
 Dick, R. I. and Ewing, B. B. (1967a) The rheology of activated sludge, J. Wat. Pollut. Control Fed 
 
 39, 543-60. 
 Dick, R. I. and Ewing, B. B. (1967b) Evaluation of activated sludge thickening theories, J. Sanit. 
 
 Engng. Div. Am. Soc. civ. Engrs, 93, SA4, 9-29. 
 Fitch, B. (1962) Sedimentation process fundamentals, Trans. Am. hist. Min. Engrs, 223, 129-37. 
 Frank, H. S. and Evans, M. W. (1945) Free volume and entropy in condensed systems, J. Chem. 
 
 Phys. 13, 507-32. 
 Kynch, G. J. (1952) A theory of sedimentation, Trans. Faraday Soc. 48, 166-76. 
 Roberts, E. J. (1949) Thickening — art of science? Min. Engng, 1, 61-64. 
 Schneebeli, G. (1955) Experiences sur la limite de validite de la loi de Darcy et 1'apparition de la 
 
 turbulence dans en ecoulement de filtration, Houille blanche, 10, 141-9. 
 Shannon, P. T., Dehaas, R. D., Stroupe, E. P. and Tory, E. M. (1964) Batch and continuous 
 
 thickening, Ind. Engng Chem. Fundamentals, 3, 250-60. 
 Vand, V. (1948) Viscosity of solutions and suspensions, J. Phys. Colloid Chem. 52, 277-321. 
 
 67 
 
APPENDIX IV 
 
 AGGREGATE SIZE VARIATIONS DURING 
 THICKENING OF ACTIVATED SLUDGE 
 
 by 
 
 Ali R. Javaher i 
 
 and 
 Richard I . Dick 
 
 Reproduced from 
 
 Journal of the Water Pollution Control Federation 
 
 Volume k\ , No. 5, Part 2, Pages R197-R214 
 
 May, 1969 
 
 68 
 
© Copyright as pnrt of the May 1909 Part 2, Joi'Rnal Water Pollution Control 
 
 Federation, Washington, D. C. 20016 
 
 Printed in U. S. A. 
 
 AGGREGATE SIZE VARIATIONS DURING 
 THICKENING OF ACTIVATED SLUDGE 
 
 Ali R. Javaheri and Richard I. Dick 
 
 A significant improvement in the per- 
 formance of the activated sludge process 
 could be realized by an increase in the 
 effectiveness of the solids separation 
 and concentration phases of the proc- 
 ess. The concentration of mixed liquor 
 suspended solids which can be main- 
 tained in aeration tanks and the re- 
 quired size of aeration tanks are de- 
 pendent directly on the concentration 
 at which sludge solids are returned from 
 the final settling tank. Also, the cost 
 of disposing of waste solids produced 
 in the activated sludge process may be 
 minimized if the difficulty of concen- 
 trating the solids into a small volume 
 can be eliminated. While gravity 
 thickening in final tanks or separate 
 thickeners is the most economical way 
 of effecting major reductions in sludge 
 volumes, the effectiveness of thickening 
 has been limited because of the re- 
 sistance of activated sludge to consoli- 
 date to high concentrations. Knowl- 
 edge of the variation of some basic 
 physical parameters which take place 
 during subsidence of activated sludge 
 would be useful in developing methods 
 for improving thickening. 
 
 The purpose of this study was to in- 
 vestigate the fundamental physical 
 behavior of activated sludge during 
 thickening. Specifically, the nature of 
 aggregate size variation and liquid dis- 
 placement from within the aggregates 
 
 Ali R. Javaheri and Richard I. Dick are, 
 respectively, Research Assistant and Associate 
 Professor of Sanitary Engineering, University 
 of Illinois, Urbana, III. 
 
 The paper was presented at the 41st Annual 
 Conference of the Water Pollution Control Fed- 
 eration, Chicago, III, Sept. 22-27, 1968. 
 
 during thickening has been evaluated. 
 In this study, activated sludge solids 
 are envisioned to be organized physi- 
 cally at three different levels — primary 
 particles, floes, and aggregates. The 
 assumption is in accordance with 
 Wold's (1) theoretical analysis of the 
 flocculation and with Michaels' and 
 Bolger's (2) experimental findings. The 
 primary particles in sludge such as 
 microorganisms are considered to ag- 
 glomerate into "floe" particles which 
 serve as the basic building units for 
 further growth of sludge solids. As 
 floes cluster together, they form "ag- 
 gregates." The term "floe" includes 
 the solids and liquid within floe parti- 
 cles and the term "aggregate" includes 
 the constituent floe particles as well as 
 the liquid between them. The "po- 
 rosity" of the suspension is contributed 
 by the interstices between the aggre- 
 
 In this investigation the water con- 
 tent and size of activated sludge aggre- 
 gates have been analyzed by use of the 
 equation for interface subsidence veloc- 
 ity of concentrated suspensions de- 
 veloped by Richardson and Zaki (3). 
 The water content of aggregates also 
 has been analyzed according to the 
 equation for flow through porous media 
 developed by Carmen and Kozeny (4). 
 An appreciable decrease in aggregate 
 size along with a significant decrease in 
 the water content has been found to ac- 
 company thickening. These changes 
 are interpreted to be brought about by 
 two mechanisms, "squeeze" and "split." 
 "Squeeze" refers to the process of 
 squeezing an aggregate to reduce its 
 water content (but not its solids con- 
 
 69 
 
JOURNAL WPCF 
 
 May 
 
 tent). "Split" refers to the process of 
 splitting an aggregate to smaller sized 
 aggregates, each of which contains only 
 a portion of the water and solids from 
 the original aggregate. When both 
 mechanisms exist, the aggregates be- 
 come more numerous, smaller, and 
 denser, and they contain less solids than 
 the original aggregates. The prevalence 
 of the squeeze mechanism has been ana- 
 lyzed for activated sludges with widely 
 differing settling characteristics from 
 three different wastewater treatment 
 plants. 
 
 To explain aggregate size variation 
 in a different manner, two flow patterns 
 have been defined. One type is inter- 
 aggregate flow (flow between the aggre- 
 gates) and the other is intraaggregate 
 flow (flow through the aggregates and 
 between floes in the aggregate). 
 
 Theory and Methods 
 
 Richardson — Zaki Equation 
 
 Stokes (5) found the terminal veloc- 
 ity, V , of a single spherical particle in 
 an infinite fluid under laminar flow 
 conditions to be : 
 
 7 fl 
 
 18m» 
 
 where, 
 
 d = diameter of the spherical particle, 
 p m = density of the particle, 
 p„ = density of the liquid, 
 H w = absolute viscosity of the liquid, 
 and 
 
 g = gravitational constant. 
 
 Stokes' equation is not an adequate 
 description of the settling velocity of 
 particles in concentrated suspensions 
 where particle subsidence is hindered 
 by the presence of neighboring particles. 
 
 Richardson and Zaki (3), who studied 
 the subsidence of the interface of a sus- 
 pension comprised of uniformly sized 
 particles by dimensional analysis, 
 showed that : 
 
 f [Vod Pm d \ 
 
 where, 
 
 V c = suspension interface subsi- 
 dence velocity, 
 V d Pm 
 = Reynolds number, 
 
 D = diameter of settling column, 
 
 and 
 e = suspension porosity. 
 
 Equation 2 is a general expression which 
 is applicable regardless of the nature of 
 flow about particles. They showed that 
 for wholly viscous flow and for fully 
 
 V c 
 developed turbulent conditions, tt- was 
 
 V o 
 independent of Reynolds number, or: 
 
 Vc J d \ 
 
 y.'W 3 
 
 Experimentally, Richardson and Zaki 
 found that subsidence of suspensions of 
 uniformly sized particles could be 
 described by : 
 
 V c 
 
 r." 4 
 
 Note that the suspension interface 
 velocity at a porosity, e, of 1.0 equals 
 the individual particle settling velocity 
 in an infinite fluid. Comparison of 
 Equations 3 and 4 shows that n is a 
 
 function only of jr. They used experi- 
 
 d 
 mental data to relate n and jz. Where 
 
 Reynolds number was less than 0.2, the 
 relationship was : 
 
 n = 4.65 + 19.5 ^ 5 
 
 The value of n decreased to a constant 
 value of 2.39 at Reynolds numbers 
 greater than 500. The variation of n 
 with intermediate Reynolds numbers is 
 illustrated by Richardson's and Zaki's 
 experimental data (Figure 1). 
 
 Richardson's and Zaki's work can be 
 applied to the sedimentation of a sus- 
 pension of activated sludge aggregates 
 of effective diameter, d, in the following 
 way. Assuming laminar flow and 
 
 70 
 
Vol. 41, No. 5, Part 
 
 ACTIVATED SLUDGE 
 
 FIGURE 1.— The variation of n (- -» J with 
 Richardson's and Zaki ! 
 
 intermediate Reynolds numbers is illustrated by 
 s (3) experimental data. 
 
 ignoring the wall effect, Equation 4 
 can be written as : 
 
 = (1 - $ a ) 4 - 65 . 
 
 where $0 is the volumetric aggregate 
 concentration. The settling velocity of 
 an aggregate, V , can be expressed by- 
 Stokes' Law as : 
 
 ;(p» - P»)d> 
 
 V = 
 
 18m. 
 
 where p a is the mass density of aggre- 
 gates. The value of p a is a quantity 
 which cannot be measured readily ; but 
 by means of material balance, its value 
 can be computed. Assuming a unit 
 volume of suspension : 
 
 Suspension wt. = wt. dry solids 
 
 + wt. liquid 
 Suspension wt. = aggregate wt. 
 
 + remaining liquid wt. 
 or; 
 
 (p.)(D =$M + (1 -**)(p») 
 
 (p.)(l) =$a(pa)+ (i -*.)(*.)•• 8 
 
 where, 
 
 p 8 = density of the suspension, 
 $* = volumetric solids concentration, 
 
 and 
 Pk = density of the solids. 
 
 Equation 8 can be simplified to 
 
 (p. - Pw) = $a(pa - P») 
 
 **(p* - p»).-9 
 
 (p a - p.) 
 
 (p* - p») 
 
 (p* - p.) . . 10 
 
 where 4 is the ratio of the volume of 
 aggregates to the volume of dry solids, 
 
 — . Hence, A is a measure of the 
 
 $k 
 
 amount of water associated with solids 
 
 in aggregates, and it will be referred to 
 
 as the "Aggregate Volume Index," 
 
 AVI. Thus 
 
 $a 
 
 A$ k 
 
 11 
 
 Substitution of Equations 7, 10, and 11 
 into Equation 6 gives : 
 
 g (p*- 
 
 *\ 
 
 (i-A$ k y-™...i2 
 
 18 y. w A 
 which can be written as : 
 
 13 
 
 or; 
 
 7,1/4.65 = 7 l/4.65 _ (V l/4.664)$ ti .14 
 
 71 
 
JOURNAL WPCF 
 
 M»y 1969 
 
 The value of the interface subsidence 
 velocity at any specified concentration, 
 V e , can be obtained from laboratory 
 batch settling tests. The value of $*, 
 the volumetric solids concentration, can 
 be determined by use of pycnometers. 
 A plot of Vc 11 * - 65 vs. <J>*: then can be 
 made. If the plot is a straight line, the 
 slope of the line will be [K C 1/466 A] and 
 the ordinate intercept (at $* = 0) will 
 be F 1/4 - 66 . If the aggregate settling 
 velocity and aggregate volume index 
 change with concentration, the plot 
 will be curved. In this case, values of 
 Vo and A at any concentration can be 
 determined in a similar fashion by 
 use of the tangent to the curve at that 
 concentration. In either event, the 
 equivalent aggregate diameter, d, can 
 be found from the calculated value of 
 Vo by use of Equation 12. 
 
 Hence it is seen that Richardson's 
 and Zaki's work provides a convenient 
 basis for investigating the fundamental 
 changes which occur during thickening 
 of activated sludge. Application of 
 Richardson's and Zaki's equation, how- 
 ever, is based on a number of assump- 
 tions, and it is well to review them and 
 their significance. In taking the value 
 of n in Equation 4 to be 4.65, the 
 Reynolds number was assumed to be 
 less than 0.2, and Richardson's and 
 Zaki's correction for the influence of the 
 wall of the settling column was ignored. 
 Based on analysis of the experimentally 
 determined values of the size and set- 
 tling velocity of individual aggregates, 
 the maximum Reynolds number in the 
 present work was approximately 4.0. 
 From Figure 1, the corresponding value 
 of n also would be about 4.0. An error 
 of opposite sign was introduced by not 
 considering the 19.5 d/D term in Equa- 
 tion 5. Based on the column size used 
 and the largest aggregate diameter en- 
 countered, the maximum error in n 
 resulting from this simplification was 
 approximately 10 percent. The ac- 
 curacy of the selected value of n also 
 was influenced by the assumption that 
 aggregates were spherical in shape. 
 Maude and Whitmore (6) found the 
 
 value of n to be 5.85 for cubes and 4.14 
 for discs. In the absence of knowledge 
 of aggregate shape, Krone (7) assumed 
 the value of n for activated sludge to be 
 5. Hence, it is seen that selection of the 
 value of n as 4.65 was somewhat arbi- 
 trary. However, the range of expected 
 variation of n about the value of 4.65 
 was not considered to be sufficient to 
 influence seriously quantitative results. 
 The general conclusions of the study, 
 it is felt, are not limited by the assump- 
 tions relating to selection of the value 
 of n. 
 
 In applying Stokes' law to the sedi- 
 mentation of individual aggregates, the 
 effect of aggregate shape was ignored. 
 Consequently, the computed aggregate 
 diameter, d, was actually the diameter 
 of a hydraulically equivalent sphere. 
 The error in applying Stokes' law in 
 situations where Reynolds number was 
 as high as 4.0 was not significant. 
 
 In using Richardson's and Zaki's 
 equation, it is tacitly assumed that only 
 fluid forces control the subsidence of the 
 suspension. That is, activated sludge 
 is assumed to be an ideal suspension 
 for which interface subsidence velocity 
 is only a function of solids concentra- 
 tion (8). Activated sludge is not an 
 ideal suspension because interparticle 
 forces as well as fluid forces resist sub- 
 sidence (9). The extent to which inter- 
 particle forces influence settling velocity 
 can be minimized by use of deep initial 
 settling depth. Dick and Ewing (9) 
 found that the interface subsidence 
 velocity for initial sludge depths of 3.5 
 ft (1.07 m), such as used in this study, 
 could vary from 40 to 95 percent of the 
 ideal settling velocity, depending on the 
 nature and concentration of the acti- 
 vated sludge. Hence, quantitative 
 determinations based oh application 
 of the Richardson-Zaki equation to 
 observed activated sludge settling be- 
 havior are subject to error because of 
 the non-ideal settling characteristics 
 of the sludge. However, again, the 
 error is not considered sufficient to 
 invalidate the basic conclusions drawn 
 
 72 
 
Vol. 41, No. 5, Part 2 
 
 ACTIVATED SLUDGE 
 
 from application of Richardson's and 
 Zaki's work. 
 
 A final assumption is that the zone 
 settling velocities observed in batch 
 settling tests conducted at various 
 initial solids concentrations are the 
 same as the settling velocities of the 
 suspension at corresponding concen- 
 trations produced during the course of 
 thickening. Talmage and Fitch (10) 
 showed that, for ideal suspensions, this 
 assumption is valid ; and Shannon et al. 
 (11) experimentally confirmed the 
 validity of the assumption by use of a 
 suspension of glass beads. However, 
 Shannon and Torey (12), Hassett (13), 
 and Fitch (14) have reported that the 
 settling velocity of fiocculent suspen- 
 sions depends on the conditions under 
 which the aggregates were formed. 
 That is, aggregates formed by ag- 
 glomeration of a suspension at a specific 
 initial concentration, d, may be differ- 
 ent than the aggregates formed when a 
 dilute suspension with an initial con- 
 centration less than d thickens to the 
 concentration d. Again, this assump- 
 tion is felt to limit the exactness of the 
 quantitative data but not the conclu- 
 sions reached by application of the 
 Richardson and Zaki analysis. 
 
 Carmen — Kozeny Equation 
 
 Carmen (4) demonstrated that, for a 
 porous bed composed of spherical uni- 
 form particles of diameter d, the 
 superficial velocity, u, of liquid through 
 the bed can be expressed as : 
 
 Ap 
 
 15 
 
 _ d* £ 
 
 U ~ K36m„(1 - e) 2 L 
 where, 
 
 K = constant, 
 Ap = pressure difference across length 
 
 of bed, and 
 L = length of porous bed. 
 
 Carmen felt that K could be taken as a 
 constant with an average value of 5.0. 
 Coulson (15), however, noted that K 
 varied between 3.2 and 5 depending on 
 factors such as particle shape. Analy- 
 sis of the data used by Scott (16) indi- 
 
 cates an even wider variation in the 
 value of K. Scott applied the Carmen- 
 Kozeny equation (Equation 15) to 
 sedimentation data for kaolinite clay 
 and assumed K to be a constant of 
 unknown value. The same assumption 
 was made in this study. 
 
 In applying Equation 15 to batch 
 sedimentation of activated sludge, ag- 
 gregates replace the spherical solid 
 particles, and the interstices between 
 the aggregates contribute the porosity 
 of the suspension. The superficial 
 upward velocity of liquid, u, is equal to 
 the interface subsidence velocity, V c . 
 The pressure drop per unit length is 
 also equal to the net weight of solids 
 supported by the unit height of the 
 suspension, or: 
 
 f= ( . 
 
 *)(Pa — Pm)g. 
 
 16 
 
 Substitution of Equation 16 into Equa- 
 tion 15, and considering u = V c : 
 
 and; 
 
 V c 
 
 (Pa 
 
 g(Pa 
 
 36K/x u 
 
 1 - e 
 
 A$ k . 
 
 17 
 
 18 
 
 Pv>) = J (Pk — Pw) 19 
 
 .20 
 
 Equation 17 can be written as : 
 
 v i- g( Pt -p.)tin(i-A$ t )» 
 
 c [ 36K M „A X J $* 
 
 Substituting Ki for the bracketed term 
 in Equation 20 : 
 
 (F c * fc )» = KJQ. - A* fc ). ...21 
 
 or; 
 
 (V c $ k )* = KJ - (KM)$k...22 
 
 As explained previously, the values of 
 V c and $k were obtained experimentally 
 for the suspensions of activated sludge. 
 To apply the Carmen-Kozeny equation, 
 then, (F c 3> fc )* was plotted as a function 
 of $*. From the slope of the plot ob- 
 tained at any specified concentration, 
 
 73 
 
JOURNAL WPCF 
 
 May 1969 
 
 o 
 
 00— L — CX — VCXJ 
 
 Ch»c1l Vol»t-J LThrol'ling Vol»e 
 
 ih volvtd Connection! lo Columni 
 
 FIGURE 2. — The settling tests were conducted in 4 3.5-in. (8.9-cm) diam columns. 
 (Note : Ft X 0.3 = m.) 
 
 the value of aggregate volume index 
 A and K\. could be calculated. Since 
 the value of K (Equation 20) was un- 
 known, the equivalent aggregate diam- 
 eter, d, could not be determined. 
 
 FIGURE 3. — Subsidence of the sludge inter- 
 face is observed on this typical batch ■ettliag 
 curve. (Note : Ft X 0.3 = m.) 
 
 Laboratory Procedure 
 
 Batch settling tests to determine the 
 sludge interface subsidence velocities 
 were made with activated sludge sam- 
 ples taken from three different acti- 
 vated sludge wastewater treatment 
 plants. Plant I was a municipal treat- 
 ment plant using the conventional 
 activated sludge process. The normal 
 sludge settling behavior at Plant I was 
 very satisfactory. Plant II was a munic- 
 ipal wastewater treatment plant using 
 the Kraue and contact stabilization 
 modifications of the activated sludge 
 process. The samples were collected 
 from the contact tanks. Settling be- 
 havior at Plant II also was satisfactory. 
 Plant III was a small municipal contact 
 stabilization plant treating an appre- 
 ciable amount of carbohydrate waste 
 discharged by industry. The settling 
 behavior of this sludge was very poor. 
 
 The settling tests were conducted at 
 a temperature of 20° =h 2°C using the 
 apparatus shown in Figure 2. It con- 
 sisted of 4 3.5-in. (8.9-cm) diam 
 columns, reservoirs, pumps, and a pip- 
 ing system. The sludge was pumped 
 
 7k 
 
Vol. 41, No. 5, Part 
 
 ACTIVATED SLUDGE 
 
 FIGURE 4.— The three activated sludge 
 plants had different interface subsidence 
 velocities. (Note : Ft X 0.3 = m.) 
 
 to a height of 3.5 ft (1.07 m) in the 
 columns at a predetermined rate to in- 
 sure homogeneous distribution of 
 solids. Subsidence of the sludge inter- 
 face then was observed (Figure 3). It 
 is characterized by an initial period 
 with little subsidence followed by a 
 period of settling at a constant rate. 
 The slope of the linear part of the 
 settling curve was the interface settling 
 velocity, V c , for the sludge at its initial 
 concentration. 
 
 Slow stirring was provided to 
 promote agglomeration and to minimize 
 the bridging which occurs in laboratory 
 settling columns when high concentra- 
 tions of activated sludge are used. A 
 comparison of results obtained by 
 analyzing settling data with and with- 
 out stirring by use of the Richardson- 
 Zaki equation afforded a measure of the 
 ■effect of stirring. It should be noted, 
 however, that Vesilind (17) has found 
 that the beneficial effect of stirring is 
 to minimize the artificial effects created 
 in laboratory settling columns. Hence, 
 data from the stirred columns are con- 
 sidered to be more indicative of thick- 
 
 FIGURE 5.— Application of Richardson's 
 and Zaki's equation to sedimentation of acti- 
 vated sludge from Plant I. (Note : Ft X 0.3 
 -m.) 
 
 FIGURE 6.— Th« aggregate volume index, 
 A, was obtained by drawing tangents to the 
 curve (Figure 5) at selected concentrations 
 and solving for the equations of the tangent 
 lines. (Plant I shown.) 
 
 75 
 
JOURNAL Wl'CF 
 
 May 1969 
 
 ening as it actually occurs in full scale 
 settling basins. Small stirrers rotated 
 at 1 rpm were constructed from 
 0.125-in. (3.18-mm) diam aluminum 
 bars (Figure 2). Thirteen angles, 3 in. 
 
 Concentration C, mq/l 
 
 FIGURE 7.— The equivalent aggregate diam- 
 eter, d, was calculated by use of Equation 12. 
 (Plant I shown.) 
 
 X 3 in. (7.62 cm X 7.02 cm) in size, here 
 were welded on 3-in. (7.62-cm) centers 
 at the midpoint of one leg to the main 
 vertical bar. 
 
 The volumetric concentration of 
 solids, $*, and solids density, p k , were 
 determined at 20°C by use of 50-ml 
 pyenometcrs. The volumetric and 
 gravimetric concentrations of a specific 
 sludge were directly proportional be- 
 cause the density of the solids remained 
 constant. 
 
 The relationship between suspended 
 solids concentration and interface set- 
 tling velocity for activated sludge from 
 Plants I, II, and III is shown in Figure 
 4. Significant differences between the 
 settling characteristics of the three 
 activated sludges may be noted. The 
 concentration ranges for the three 
 sludges were appreciably different be- 
 cause of variations in sludge character- 
 istics caused by the particular operating 
 conditions "in the three plants. The 
 concentration range used for each 
 
 FIGURE 8. — As the concentration increased, 
 the individual aggregate volume decreased. 
 (Plant I shown.) 
 
 Concentration C , mg/t 
 
 FIGURE 9. — The volume of solids in each 
 aggregate decreased as the concentration in- 
 creased. (Plant I shown.) 
 
 76 
 
Vol. 41, No. 5, Part 2 
 
 ACTIVATED SLUDGE 
 
 O 5000 10.000 
 
 Concentration C , mg/i 
 
 FIGURE 10.— The number of aggregates 
 increased rapidly as the concentration in- 
 creased. (Plant I shown.) 
 
 sludge was considered to represent the 
 thickening conditions which could exist 
 in the respective plants. 
 
 Application of Richardson's and 
 Zaki's Equation 
 
 The settling data shown in Figure 4 
 were analyzed by use of the Richardson- 
 Zaki equation. Results for- the stirred 
 samples of sludge from Plant I are pre- 
 sented here in detail. Results for Plant 
 II are summarized briefly in the text. 
 Results for both the stirred and un- 
 stirred samples from Plant III are pre 
 sented to illustrate the typical effect of 
 mixing. Figure 5, a plot of V,. 1 ' 4 - 66 vs. 
 the volumetric solids concentration, 
 $k, for sludge from Plant I, establishes 
 the relationship between $ k and the 
 gravimetric concentration, C, as de- 
 termined by use of pycnometers. The 
 fact that the curve in Figure 5 is not 
 linear indicates (Equation 14) that the 
 settling velocity of an individual ag- 
 
 gregate, V , and the aggregate volume 
 index, A, changed with concentration. 
 Values of A and V were obtained by 
 drawing tangents to the curves at se- 
 lected concentrations and solving for 
 the equations of the tangent lines ac- 
 cording to Equation 14. From the 
 
 Concentration C , mg/2 
 
 FIGURE 11. — The aggregate density was cal- 
 culated from Equation 10. (Plant I shown.) 
 
 Concentration C , mg/? 
 
 FIGURE 12.— The amount of water between 
 aggregates can be computed by comparing 
 suspension and aggregate volumes. (Plant I 
 shown.) 
 
 77 
 
JOURNAL WPCF 
 
 May 1969 
 
 values of T r obtained, the equivalent 
 aggregate diameter, d, was calculated at 
 various concentrations by use of Equa- 
 tion 12 (Figures 6 and 7). The figures 
 show that in concentrating from 5,000 
 mg/1 to 19,000 mg/1, the diameter of 
 aggregates was reduced from approxi- 
 
 mately 1.9 mm to 0.5 mm, and the ratio 
 of the volume of water associated with 
 each volume of solids was reduced from 
 56 to 28. 
 
 From these data it was possible to 
 compute the number of aggregates 
 which existed at any concentration, as 
 
 Concentration C , mg/i 
 
 FIGURE 13.— Porosity decreased 28 per- 
 cent as the sludge from Plant I thickened from 
 5,000 mg/1 to 19,000 mg/1. 
 
 Concentration C, mg/Jt 
 
 FIGURE 15.— The aggregate squeeze index 
 indicates the fraction of clarified liquid which 
 originated from within aggregates as thickening 
 took place. (Plant I shown.) 
 
 FIGURE 14. — Fluid displacement rates varied 
 with concentration. (Plant I shown.) 
 
 FIGURE 16.— Richardson's and Zaki's equa- 
 tion was applied to the sedimentation of acti- 
 vated sludge from Plant III. 
 
 78 
 
Vol. 41, No. 5, Part 2 
 
 ACTIVATED SLUDGE 
 
 well as their volume and density. Also, 
 the volume and weight of actual solids 
 in each aggregate could be calculated 
 as could the porosity of the suspension 
 at any concentration. Figures S and 9 
 indicate the decrease in individual ag- 
 gregate volume, V , and the volume of 
 actual solids in each aggregate, V t , 
 which occurred as concentration in- 
 creased. As the concentration of the 
 suspension increased, the number of 
 aggregates, N, increased rapidly (Figure 
 10). Figure 11 shows the increase in 
 the aggregate density, p„, as calculated 
 from Equation 10, which accompanied 
 an increase in concentration. 
 
 The difference between the curves for 
 the total volume of suspension, V», 
 and the total volume of aggregates per 
 gram of suspension, 2 V Q , represents the 
 amount of water between aggregates 
 (Figure 12). The porosity, e, at any 
 
 . . . , V,-2V a 
 concentration is given by — . 
 
 V, 
 
 The relationship between porosity and 
 concentration (Figure 13) indicates that 
 porosity decreased from 70 percent to 
 50 percent as the sludge from Plant I 
 thickened from 5,000 mg/1 to 19,000 
 mg/1. 
 
 It is of interest to know whether the 
 volume reduction which accompanies 
 thickening at any particular concentra- 
 tion is caused by elimination of water 
 
 FIGURE 17. — Aggregate volume index varied 
 with solids concentration. (Plant III shown.) 
 
 FIGURE 18. — As concentration increased, 
 the aggregate diameter decreased. (Plant III 
 shown.) 
 
 from between aggregates or by squeez- 
 ing of water from within aggregates. 
 Figure 14 shows the rate of change of 
 suspension volume, V„, and total ag- 
 gregate volume, 2V«, with respect to 
 concentration as given by the slopes of 
 the curves in Figure 12. To give a mea- 
 sure of the significance of particle 
 squeeze during thickening, the quantity 
 
 A(2V a ) 
 
 — — — is defined as the "Aggregate 
 
 At, 
 
 Squeeze Index," ASI. The ASI indi- 
 cates the fraction of clarified liquid 
 which originated from within aggre- 
 gates as thickening took place. ASI 
 would be for incompressible aggregates 
 and would achieve a maximum value 
 of 1.0 when all displaced fluid originated 
 from within aggregates. For the sludge 
 from Plant I, 15 percent of the total 
 water being displaced from the con- 
 solidating sludge mass was coming from 
 inside aggregates when the suspended 
 solids concentration was 7,000 mg/1 
 (Figure 15). This fraction doubled by 
 the time the sludge concentration 
 reached 15,000 mg/1. 
 
 Figures 8, 9, 10, and 11 show that as 
 thickening took place, i the activated 
 sludge aggregates became smaller, con- 
 tained less solids than the initial aggre- 
 gates, increased in number, and experi- 
 enced an increase in density. It is 
 suggested that these phenomena were 
 
 79 
 
JOURNAL WPCF 
 
 May 19C9 
 
 brought about by the two physical 
 mechanisms, "squeeze" and "split," as 
 defined previously. By the process of 
 
 FIGURE 19. — At a given concentration, the 
 aggregates of a stirred sample were smaller 
 than those of the unstirred sample. (Plant 
 III shown.) 
 
 I 
 
 : 1 
 
 —i 1— iii. 
 
 >' 
 
 
 V-iX^__^ 
 
 I 
 
 s 
 
 ■ 
 
 SltntJ \ 
 
 o 
 
 ■6 
 
 V 
 
 : 
 
 1 1 1 1 1 
 
 squeezing, the water content of acti- 
 vated sludge aggregates is reduced. 
 By the process of splitting, an aggregate 
 is divided into smaller aggregates, but 
 no water is removed. Both the water 
 content and solids content of an aggre- 
 gate become divided by "split." The 
 experimental results show that both 
 mechanisms prevailed during thicken- 
 ing of the activated sludge from Plant 
 I because the aggregates became 
 smaller, denser, and they increased 
 rapidly in number. 
 
 To explain the changes which take 
 place during thickening in a different 
 manner, two different flow patterns 
 may be considered (16). Some of the 
 liquid originating from the body of ag- 
 gregates and from the interstices be- 
 tween aggregates is displaced from the 
 subsiding sludge mass by "interaggre- 
 gate flow" or flow between aggregates. 
 Some of the liquid, however, is dis- 
 placed by flowing through aggregates, 
 and this can be termed "intraaggregate 
 flow." In applying Richardson's and 
 Zaki's equation to settling data, the 
 manifestation of intraaggregate flow is 
 a reduction of the apparent size of the 
 basic settling units or aggregates. 
 
 FIGURE 20. — The volume of solids in an ag- 
 gregate decreased with an increase in solids 
 concentration. (Plant III shown.) 
 
 FIGURE 21.— The number of aggregates in- 
 creased as the solids concentration increased. 
 (Plant III shown.) 
 
 80 
 
Vol. 41, No. 5, Part 
 
 ACTIVATED SLUDGE 
 
 Figure 16 shows the application of 
 Richardson's and Zaki's equation to the 
 sedimentation of stirred and unstirred 
 samples of the poorly settling sludge 
 from Plant III. The curvature of 
 Figure 16 again indicates that the ag- 
 gregate size varied during the thicken- 
 ing process. The decrease in aggregate 
 volume index, A, and aggregate size, d, 
 with increasing concentration is indi- 
 cated in Figures 17 and IS. Similarly, 
 Figures 19 through 26 are plotted for 
 the sample from Plant III by the 
 methods described previously. Note 
 that considerable differences existed be- 
 tween aggregate characteristics in the 
 stirred and unstirred samples. At a 
 given concentration, the aggregates of a 
 stirred sample were smaller, denser, 
 and larger in number than those of the 
 unstirred sample (Figures 19, 21, and 
 22). The suspension porosity and the 
 aggregate squeeze index of the stirred 
 sample increased with concentration; 
 while these variables decreased with 
 concentration for unstirred samples 
 (Figures 24 and 26) . While it may seem 
 unreasonable that porosity should be 
 greater in slow-settling, highly concen- 
 trated suspensions than in more dilute 
 suspensions, it should be noted that 
 
 FIGURE 23. — Volume of suspension and ag- 
 gregate volume for Plant III varied with solids 
 concentration as shown. 
 
 FIGURE 22.— Aggregate density of the stirred 
 samples increased more rapidly than for the 
 unstirred sample with an increase in solids 
 concentration. (Plant III shown.) 
 
 FIGURE 24. — The suspension porosity in- 
 creased for the stirred samples and decreased 
 for the unstirred samples with an increase in 
 concentration. (Plant III shown.) 
 
 FIGURE 25.— Variation of fluid displacement 
 rates with concentration is shown for Plant III. 
 
 81 
 
JOUENAL WPCF 
 
 May 19C9 
 
 the individual aggregate settling veloc- 
 ity, V , decreased at higher concentra- 
 tions and that this change could offset 
 an increase in porosity (Equation 4). 
 Presumably, stirring destroyed bridge 
 networks in the suspension, enhanced 
 squeezing of the aggregates, and pro- 
 duced a greater aggregate squeeze index 
 in the concentrating suspension of 
 
 FIGURE 26.— The poor settling sludge from 
 Plant III had a much larger aggregate squeeze 
 index than the better settling sludge from 
 Plant I (Figure 15). 
 
 FIGURE 27.— Carmen's and Kozeny's equa- 
 tion can be applied to the sedimentation of 
 activated sludge to determine the aggregate 
 volume index. (Plant I shown.) 
 
 BOOp 10.000 
 
 Concentration C , mg/S. 
 
 FIGURE 28. — Carmen's and Kozeny's equa- 
 tion for the aggregate volume index yielded 
 results which corresponded closely with those 
 from the Richardson and Zaki equation. 
 (Plant I shown.) 
 
 activated sludge. As before, both the 
 squeeze and the split mechanisms were 
 prevalent during the thickening of the 
 stirred and the unstirred samples from 
 Plant III. 
 
 Comparison of results for the stirred 
 samples from Plants I and III indicates 
 several significant points. The poor 
 settling sludge from Plant III had a 
 much larger aggregate squeeze index 
 (Figure 26) than the better settling 
 sludge from Plant I (Figure 15). This 
 means that a large fraction of the fluid 
 clarified from the poor settling sludge 
 originated from the bodies of the aggre- 
 gates. The poor settling sludge also 
 had a smaller porosity (Figure 24) than 
 the better settling sludge (Figure 13) in 
 the respective concentration ranges of 
 the two sludges. It seems that a good 
 settling sludge is characterized by its 
 high porosity and low aggregate squeeze 
 index (ASI). When sludge character- 
 
 82 
 
Vol. 41, No. 5, Part 2 
 
 ACTIVATED SLUDGE 
 
 FIGURE 29.— The solids flux curve is used in 
 the analysis of thickening. (Plant I shown.) 
 
 istics are such that squeeze of aggre- 
 gates must be relied on to achieve 
 higher concentrations, thickening does 
 not take place readily. 
 
 A similar analysis was carried out for 
 samples from Plant II. For stirred 
 samples from Plant II the nature of 
 variations of AVI and aggregate diam- 
 eter were comparable to those described 
 for sludge from Plant I. With unstirred 
 samples, a plot of the Richardson-Zaki 
 equation in the form of Equation 14 
 gave a straight line over the range of 
 concentration from 1,970 to 7,190 mg/1, 
 indicating that the aggregate volume 
 index, A, and aggregate diameter, d, 
 were constant for this sludge. 
 
 Application of Carmen's and 
 Kozeny's Equation 
 
 As described previously the values of 
 A (aggregate volume index) and Ki 
 could be determined by plotting (Vc**)* 
 vs. <!>* (Equation 22). Figure 27 shows 
 the plot for the activated sludge sample 
 from Plant I. The decrease in the ag- 
 gregate volume index, A, with concen- 
 tration is indicated in Figure 28. The 
 
 values agreed closely with those ob- 
 tained by use of Richardson's and Zaki's 
 equation (Figure G). Since the value of 
 K (Equation 20) was not known, the 
 equivalent aggregate diameter, d, could 
 not be calculated. This is one of the 
 disadvantages of using the Carmen- 
 Kozeny equation for analysis of settling 
 data. 
 
 FIGURE 30. — Application of the Carmen- 
 Kozeny equation to the settling data from Plant 
 III is shown. 
 
 FIGURE 31. — Variation of aggregate volume 
 index with concentration is shown, based on 
 the Carmen-Kozeny equation. (Plant III 
 shown.) 
 
 83 
 
JOURNAL WPCF 
 
 May 1969 
 
 Another disadvantage of the equation 
 is that as $* approaches zero, (V c $ k )l 
 also approaches zero. Note that V c $ k 
 is proportional directly to solids flux 
 (G = CV C ) which is used commonly in 
 analysis of thickening (IS) (19). The 
 solids flux curve, G vs. C, has the char- 
 acteristic shape shown in Figure 29. 
 Now, (Ve**) 1 is proportional directly 
 to the cube root of the ordinate values 
 in solids flux curve. If a smooth curve 
 has to be drawn in Figure 27 to apply 
 the Carmen-Kozeny equation, the data 
 should represent the values of G follow- 
 ing G m&x . The Carmen-Kozeny equa- 
 tion cannot be applicable, then, to 
 settling data at concentrations corre- 
 sponding to solids^flux values at the 
 left side of G max . The values omitted 
 from the curve in Figure 27 indicate 
 this point. Application of the Carmen- 
 Kozeny equation to the values only 
 on the right side of G max (Figure 27) can 
 provide an estimate of the aggregate 
 volume index as shown later. The dis- 
 advantages mentioned limited the util- 
 ity of the Carmen-Kozeny equation as a 
 means of aggregate size analysis of 
 activated sludge. 
 
 Application of the Carmen-Kozeny 
 equation to the settling data from Plant 
 III is illustrated in Figure 30. The 
 values of the aggregate volume index, 
 A or AVI (Figure 31) compared closely 
 to those obtained by using the Richard- 
 son-Zaki equation (Figure 17). Appli- 
 cation of the equation to data for the 
 unstirred sample from Plant II again 
 indicated a constant value for the ag- 
 gregate volume index, A, in the concen- 
 tration range of 1,970 to 7,190 mg/1; 
 and analysis of the Plant II stirred data 
 indicated that the aggregate volume 
 index decreased as concentration 
 increased. 
 
 The decrease in the aggregate volume 
 index with increasing concentration 
 found by application of the Carmen- 
 Kozeny equation indicated squeezing 
 of the aggregates. Splitting of the ag- 
 gregates could not be shown by the 
 application of the Carmen-Kozeny 
 
 equation, since aggregate diameters 
 could not be calculated. 
 
 Conclusions 
 
 To obtain information regarding the 
 fundamental behavior of activated 
 sludge during thickening, this study was 
 carried out to determine the nature of 
 liquid displacement from aggregate 
 particles and to determine the manner 
 in which aggregate particle size varies 
 as consolidation takes place. Analysis 
 of the experimental data was based pri- 
 marily on an equation for the reduced 
 settling velocity of a suspension. 
 
 The following conclusions may be 
 drawn : 
 
 1. As the concentration of activated 
 sludge increases, the aggregates which 
 comprise the sludge are squeezed to 
 eliminate water and are split into 
 smaller aggregates. The combined 
 effect of "squeeze" and "split" is that 
 aggregates become smaller, more nu- 
 merous, and more dense as thickening 
 takes place. 
 
 2. The fluid eliminated from subsid- 
 ing sludge masses originates from within 
 aggregate particles and from the inter- 
 stices between aggregates. The "ag- 
 gregate squeeze index," ASI, has been 
 introduced to describe the role which 
 aggregate "squeeze" plays in thicken- 
 ing. It is the fraction of the total water 
 being eliminated at any particular con- 
 centration which is originating from 
 within aggregate particles. Based on 
 the settling characteristics of sludges 
 from the plants studied here, activated 
 sludges with good settling characteris- 
 tics have low ASI and high porosity 
 values, and thickening to high concen- 
 trations occurs primarily by elimination 
 of intersticial water. In contrast, 
 sludges with poor settling properties 
 have high ASI values and low porosi- 
 ties. That is, much of the water re- 
 moved in the course of thickening of 
 poor settling sludges comes from inside 
 the aggregates. 
 
 3. Slow stirring during laboratory 
 settling tests appreciably alters the 
 
 84 
 
Vol. 41, No. 5, Part 2 
 
 ACTIVATED SLUDGE 
 
 settling velocity of high concentrations 
 of activated sludge. The faster set- 
 tling velocities commonly observed in 
 stirred columns may be caused by the 
 fact that stirring increases the aggre- 
 gate squeeze index. However, the low 
 ASI values associated with unstirred 
 columns may reflect the bridging in 
 laboratory settling columns and may 
 not be a factor in prototype settling 
 basins. 
 
 4. The thickening of activated 
 sludges brings about two flow patterns 
 within the mass of aggregates. Some 
 of the total displaced fluid which origi- 
 nates from the interstices of the suspen- 
 sion and bodies of the aggregates 
 travels between the subsiding aggre- 
 gates. This is referred to as interag- 
 gregate flow. At the same time, some 
 of the fluid travels through the subsid- 
 ing aggregates. This is referred to as 
 intraaggregate flow. Intraaggregate 
 flow accompanies the split of aggregates 
 to smaller ones. 
 
 Acknowledgment 
 
 This work was supported by Re- 
 search Grant WP 01011 from the 
 Federal Water Pollution Control 
 Administration. 
 
 APPENDIX 
 
 Notation 
 
 The following symbols are used in 
 this paper: 
 
 A = ratio of the aggregate volume to 
 the volume of solids in the 
 aggregate, 
 
 A2V„ 
 aggregate squeeze index 
 
 ASI 
 AVI 
 
 AV.' 
 aggregate volume index, same 
 as A, 
 
 C = gravimetric solids concentra- 
 tion, (F/L 3 ), 
 d = initial gravimetric solids con- 
 centration, (F/L 3 ), 
 
 d = diameter of the spherical par- 
 ticle, (L), 
 
 d = equivalent aggregate diam- 
 eter, (L), 
 
 D = diameter of settling column, 
 
 oo. 
 
 g = gravitational constant, (L/T 2 ), 
 G = solids flux, (F/L 2 T), 
 Gmax = maximum solids flux, (F/L 2 T), 
 K = constant of Carmen's and 
 
 Kozeny's equation, 
 Ki = term in bracket in Equation 20, 
 L = length of porous bed, (L), 
 N = number of aggregates per unit 
 
 weight of solids, ( p ), 
 
 n = exponent in Richardson's and 
 Zaki's equation, 
 Ap = pressure difference, (F/L 2 ), 
 u = superficial upward velocity, 
 
 (L/T), 
 V = independent settling velocity 
 of an individual aggregate, 
 (L/T), _ 
 V c = suspension interface subsidence 
 
 velocity, (L/T), 
 V c = interface subsidence velocity 
 
 with stirring, (L/T), 
 V„ = interface subsidence velocity 
 
 without stirring, (L/T), 
 V = volume (L 3 ), 
 P = density, (FT 2 /L 4 ), 
 $ = volumetric concentration, 
 n = absolute viscosity, (FT/L 2 ), 
 
 and 
 e = suspension porosity. 
 
 Subscript Notations 
 a = aggregate, 
 k = solids, 
 
 m — spherical particle, 
 s = suspension of activated sludge, 
 
 and 
 w = liquid. 
 
 References 
 
 1. Wold, M. J., "Computer Simulation of Floe 
 
 Formation in a Colloidal Suspension." 
 Jour. Colloid Sti., 18, 684 (1963). 
 
 2. Michaels, A. S., and Bolger, J. C, "The 
 
 Plastic Flow Behavior of Flocculated 
 Kaolin Suspensions." Ind. Eng. Chem. 
 Fundamentals, 1, 153 (1962). 
 
 3. Richardson, J. F., and Zaki, W. N., "Sedi- 
 
 mentation and Fluidization. Part I." 
 Trans. Inst. Chem. Engr., 32, 35 (1954). 
 
 85 
 
JOURNAL WPCF 
 
 May 1969 
 
 4. Carmen, P. C, "Fluid Flow Through 
 
 Granular Beds." Trans. Inst. Chem. 
 Engr., 15, 150 (1937). 
 
 5. Stokes, G. G., "On the Theories of Internal 
 
 Friction of Fluids in Motion, and of the 
 Equilibrium and Motion of Elastic 
 Solids." Trans. Cambridge Phil. Soc, 
 8, 287 (1845). 
 
 6. Maude, A. D., and Whitmore, R. L., "A 
 
 Generalized Theory of Sedimentation." 
 Brit. Jour. Appl. Phys., 9, 477 (1958). 
 
 7. Krone, R. B., "Discussion of Evaluation 
 
 of Activated Sludge Thickening Theories 
 by Richard I. Dick and Benjamin B. 
 Ewing," Jour. San. Eng. Div., Amer. 
 Soc. Civil Engr., 94, SA3, 554 (1968). 
 
 8. Kynch, G. J., "A Theory of Sedimenta- 
 
 tion." Trans. Faraday Soc, 48, 166 
 (1952). 
 
 9. Dick, R. I., and Ewing, B. B., "Evaluation 
 
 of Activated Sludge Thickening The- 
 ories." Jour. San. Eng. Div., Proc. 
 Amer. Soc. Civil Engr., 93, SA4, 9 (1967). 
 
 10. Talmage, W. P., and Fitch, E. B., "De- 
 
 termining Thickener Unit Areas." Ind. 
 Eng. Chem., 47, 38 (1955). 
 
 11. Shannon, P. T., Dehaas, R. D., Stroupe, 
 
 E. P., and Tory, E. M., "Batch and 
 
 Continuous Thickening." Ind. Eng. 
 Chem. Fundamentals, 3, 250 (1964). 
 
 12. Shannon, P. T., and Tory, E. M., "Settling 
 
 of Slurries." Ind. Eng. Chem., 57, 18 
 (1965). 
 
 13. Hassett, N. J., "Design and Operation of 
 
 Continuous Thickeners." Ind. Chemist, 
 34, 489 (1958). 
 
 14. Fitch, E. B., "Sedimentation Process 
 
 Fundamentals." Trans. Amer. Inst. 
 Mining Eng., 223, 129 (1962). 
 
 15. Coulson, M. M., "The Flow of Fluids 
 
 Through Granular Beds: Effect of 
 Particle Shape and Voids in Streamline 
 Flow." Trans. Inst. Chem. Engr., 27, 
 237 (1949). 
 
 16. Scott, K. J., "Mathematical Models of 
 
 Mechanism of Thickening." Ind. Eng. 
 Chem. Fundamentals, 5, 109 (1966). 
 
 Vesilind, P. A., "The Influence of Stirring 
 in the Thickening of Biological Sludge." 
 Doctoral Thesis, Univ. North Carolina, 
 Chapel Hill, North Carolina (1968). 
 
 Hasset, N. J., "Concentrations in a Contin- 
 uous Thickener." Ind. Chemist, 40, 29 
 (1964). 
 19. Shannon, P. T., and Tory, E. M., "The 
 Analysis of Continuous Thickening." 
 Soc. Mining Engr. 235, 375 (1966). 
 
 17 
 
 18 
 
 86 
 
APPENDIX V 
 
 DISTRIBUTION OF COMPRESSIVE FORCES 
 IN SUBSIDING SLUDGE MASSES 
 
 by 
 
 Richard I. Dick 
 and 
 Byong S. Shin 
 
 Summary of Preliminary Results 
 of Work in Progress 
 
 87 
 
DISTRIBUTION OF COMPRESSIVE FORCES 
 IN SUBSIDING SLUDGE MASSES 
 
 In a suspension settling at uniform velocity, the effective weight of 
 the particles is equalled by the forces resisting their subsidence. In a 
 concentrated suspension, these forces may arise from fluid drag and from 
 interpart icle contacts with adjacent particles. In order to obtain under- 
 standing of the fundamental factors which control subsidence of sludges it 
 is of interest to know the relative significance of the two forms of support 
 under different conditions. The i nterparticle forces produce compressive 
 stresses in the sludge which can result in elimination of water by compres- 
 sion. The fluid drag forces give rise to excess hydrostatic pressure in 
 the pore water of the settling suspension. 
 
 Experimentally, the work was conducted by observing profiles of excess 
 hydrostatic pressure and solids concentration during batch sedimentation 
 tests with sludge from a water softening plant. Tests were conducted in a 
 3.5 in. diameter, 4-ft deep laboratory settling column. The solids profiles 
 gave a measure of the effective weight of suspended solids above any point 
 at any time and comparison of this value with the measured excess hydrostatic 
 pressure at the same location permitted computation of the amount of solids 
 supported by i nterpart i cle forces and hence of the compressive stress within 
 the supporting sludge. Figure 1 shows typical data arising from use of this 
 procedure. It is seen, as expected, that very low compressive stresses exist 
 at the beginning of sedimentation tests but that appreciable stresses are 
 achieved after a period of time. The technique also permits computation of 
 the permeability of the suspension at any point or time. Typical results 
 are shown in Figure 2. As shown by the figure, permeability has been found 
 to be principally a function of concentration, regardless of when or where 
 
 88 
 
that concentration occurs. Note that the suspension permeability changed 
 dramatically with solids concentration. A four-fold increase in concentra- 
 tion resulted in a reduction in permeability by a factor of about 1 80 . 
 
 Although the compressive yield strength in confined conditions cannot 
 be computed from measured shear stress on theoretical basis, it is inter- 
 esting to compare the relative magnitude of the two values. Figure 3 shows 
 the measured compressive yield strength from sedimentation tests as a func- 
 tion of yield stress as measured in a rotational viscometer. The figure 
 illustrates that laboratory measurement of the yield strengths of suspen- 
 sions could potentially serve to permit evaluation of their potential thick- 
 ening characteristics. 
 
 Although the work on measurement of the relative significance of fluid 
 drag and i nterpart icle forces is still underway, analyses to date indicate 
 that, as expected, the final concentration of solids which can be achieved 
 depends upon the compressibility of a suspension while the rate at which 
 this concentration is reached depends also on suspension permeability. An 
 increase in sludge depth increases final solids concentration because it 
 increases the compressive force on the sludge mass. Low permeability has 
 been shown to have a double effect on thickening. It not only retards the 
 rate of upward egression of water but also reduces the amount of compressive 
 force for sludge consolidation. 
 
 The analysis has permitted evaluation of the effect of stirring in 
 laboratory settling columns. Stirring significantly increases the compres- 
 sibility of a suspension but reduces the permeability and thus retards 
 escape of the water displaced from the subsiding sludge mass. 
 
 89 
 
HEIGHT, ft 
 
 p 
 
 f\> 
 
 
 CO 
 
 o 
 
 H — 
 
 m co 
 za cz 
 
 > 
 
 H O 
 
 3 -n 
 
 r* t* 
 > m 
 
 c 
 
 O CO 
 
 O H 
 
 m 70 
 
 H 
 
 m o* 
 *. o 
 
 3 
 5* 
 
 00 
 
 o 
 
 o 
 o 
 
 O 
 
 
 - -pr — r 
 
 ■«^i 1 1 — *-~4^.i 1 1 — ~7~ 
 
 
 
 ^^^-^ ^^^-^o / 
 
 ~ 
 
 
 \^ \° / 
 
 - 
 
 
 \ / A • 
 
 _\ 
 
 
 ]S / / - 8 - 
 
 
 \ \° 
 \ \ >> 
 
 / / / z c 
 
 / / / m <f 
 
 
 
 
 
 / / / ^J. 
 
 
 
 / / / "n r 
 
 -* 
 
 V? \ 
 
 V V* \ 
 
 / / / m ^ " 
 
 / / / 
 
 -\ 
 
 V? \ 
 
 I J / / 
 
 
 
 
 / / 
 
 
 \ \ p \ 
 
 
 / / 
 
 
 
 
 / / ^ Q z P £ ~ 
 
 / / z O co c 
 
 / / "O J 2 « z 
 
 
 \p \ \ 
 
 
 / / j* m jj 
 
 
 
 
 
 
 
 
 
 
 // -° £ (/>*> 
 
 -\ 
 
 
 
 // 3 £ - ^ - 
 r $ z *> 
 / m 2 
 
 
 \— • \ \ 
 
 
 (/> o 
 
 _ 
 
 \r»» \ ' 
 
 \ / 
 
 
 
 
 
 \ " mm 
 
 - 
 
 ll 1 
 
 V K - 
 
 1 / 
 
 / 
 ii i i i i i _i 1 
 
 90 
 
100 
 50h 
 
 20 
 
 JO 
 
 5 
 
 5 2 
 
 m 
 < I 
 
 S 0.5 
 a 
 
 0.2 h 
 0.1 
 0.05 
 
 0.02 
 
 
 1 1 
 
 1 
 
 
 SYMBOL LAYER 
 
 
 
 
 
 
 
 • 
 
 | 
 
 
 
 - 
 
 
 
 
 + 
 
 2 
 
 
 - 
 
 D« 
 
 
 
 
 D 
 
 3 
 
 
 
 c$ 
 
 
 
 
 X 
 
 4 
 
 
 
 v\ 
 
 
 
 
 A 
 
 7 
 
 ■ 
 
 5 
 6 
 
 7 
 
 
 — 
 
 
 
 
 
 O 
 
 O 
 
 A 
 
 8 
 9 
 IO 
 
 
 
 
 
 X. a 
 
 • 
 X 
 
 3 
 
 
 II 
 12 
 
 
 
 
 
 
 
 
 
 + 
 
 □ 
 
 
 
 
 
 
 
 X ^| 
 
 X 
 
 
 
 
 
 
 
 3 
 
 
 
 oV 
 
 oX 
 
 (1 > 
 
 IO 
 
 
 - 
 
 
 
 
 
 
 
 
 — 
 
 — 
 
 RUN 2 
 
 C - lOOg/f 
 
 
 
 
 
 
 — 
 
 
 NO STIRRING 
 
 1 1 
 
 1 
 
 
 1 
 
 I 
 
 
 
 0.10 
 
 0.15 0.20 0.25 0.30 0.35 
 SOLIDS CONCENTRATION, gr/cu cm 
 
 0.40 
 
 FIGURE 2. CHANGES IN PERMEABILITY OF WATER TREATMENT PLANT SLUDGE 
 AS CONCENTRATION CHANGES DURING SEDIMENTATION 
 
 91 
 
01 * wo bs/ s3uXp ' H1DN3U1S 
 C113IA 3AISS3UdWOD 03NldNOD 
 
 92 
 
APPENDIX VI 
 
 INFLUENCE OF BIOLOGICAL VARIABLES 
 ON THE PHYSICAL PROPERTIES OF ACTIVATED SLUDGE 
 
 by 
 
 Richard I . Di ck 
 Sajal K. Chakrabarti 
 
 and 
 Gloria L. McCutcheon 
 
 Prepubl ication Manuscript 
 
 93 
 
INFLUENCE OF BIOLOGICAL VARIABLES 
 ON RHEOLOGICAL PROPERTIES OF ACTIVATED SLUDGE 
 
 Richard I. Dick, Sajal K. Chakrabarti, and Gloria L. McCutcheon 
 
 INTRODUCTION 
 
 The effectiveness and cost of waste treatment by the activated sludge 
 process depends, in large measure, on the physical properties of the acti- 
 vated sludge formed. Sludges with unfavorable physical properties settle 
 poorly. This creates difficulty in obtaining concentrated sludge for return 
 to the aeration tank which, in turn, may impair the quality of the effluent 
 and increase the volume of sludge which must be wasted from the process. 
 Similarly, the cost of treatment and ultimate disposal of waste activated 
 sludge is increased if the sludge has adverse physical properties. Such 
 sludges are difficult to concentrate for economical digestion, transport, 
 or combustion, require higher doses of conditioning chemicals, and cause a 
 reduction in the capacity of sludge dewatering equipment. 
 
 In spite of the importance of the physical properties of activated 
 sludge, reliable techniques for measuring sludge properties are lacking. 
 The sludge volume index has long been used as the only measure of the phys- 
 ical properties of activated sludge. While the SVI of an activated sludge 
 is easy to determine and may be useful for routine treatment plant control, 
 it is not an adequate description of the physical nature of a sludge. The 
 basic nature of the sludge volume index was investigated by Dick and 
 Vesilind (1969) who reported that the SVI was influenced in varying and 
 unpredictable degrees by various physical properties of sludges and that it 
 is seriously influenced by anomolies associated with the small cylinder 
 used in the laboratory test. 
 
 9* 
 
Use of basic Theological measurements has previously been suggested 
 (Dick and Ewing, 1 967) as a more fruitful method for monitoring sludge 
 quality and relating it to process performance. Wood (1970) found that 
 rheological measurements were the best means for relating activated sludge 
 properties to its flotation behavior, and Geinopolos and Katz (196*0 re- 
 ported that the capacity of a collector for a flotation unit was related 
 to the rheology of the sludge being collected. Dick (1969b) related devi- 
 ations in the settling behavior of activated sludge from conventional 
 theory to the rheological properties of the suspension, and Colin (1970) 
 used rheological measurements to assess the conditioning effect of poly- 
 electrolytes . 
 
 Realizing that the behavior of activated sludge in treatment processes 
 is influenced by its rheological nature, it is necessary to have information 
 on the influence of variables in the activated sludge process on the rheology 
 of the sludge produced. The purpose of this paper is to present some pre- 
 liminary results of this type. 
 
 RHEOLOGY 
 
 Rheology is the study of the response of a material to an applied stress. 
 Most pure liquids afford a simple illustration of one type of rheological 
 behavior. They exhibit the well known Newtonian type of behavior in which 
 the velocity gradient is directly proportional to the applied stress - the 
 proportionality constant relating the two is called viscosity. Hookeian 
 or elastic solids afford another simple illustration. In this case, dis- 
 tortion is proportional to applied stress and the constant relating the 
 two is the modulus of elasticity. Between the extreme and ideal types of 
 behavior described by Hooke and Newton many more complex types of behavior 
 are possible. The reader is referred to sources such as Scott Blair ( 1 969) 
 
 95 
 
and Reiner (I969) for a complete discussion of types of Theological behavior. 
 
 If in a suspension such as activated sludge a continuous network of 
 particles is formed, then one might expect that the material would behave as 
 a solid until the internal structure was destroyed and that when greater 
 stresses were applied the material would flow somewhat as a liquid. Indeed 
 this type of rheological behavior, called plastic behavior, has been found 
 with activated sludge (Dick and Ewing, 1 967) - As originally described by 
 Bingham (1922), the behavior of a plastic material may be described as 
 
 where t is shearing stress, t is the yield stress, n is the coefficient 
 of rigidity or plastic viscosity, and -7— is the velocity gradient or rate 
 of shear. Determination of the values of t and n. affords basic measures 
 
 y 
 
 of the strength of the solid phase of the suspension and of the flow charac- 
 teristics of the two phase system. It should be noted that it sometimes is 
 difficult experimentally to distinguish between plastic materials and mater- 
 ials exhibiting a related rheological property called pseudoplast i ci ty . It 
 is thus appropriate to consider the value of x in equation (l) to be an 
 apparent value realizing that the material may not fit Binghams ideal model 
 exactly. 
 
 It might also be anticipated that activated sludge would display another 
 rheological property, thixotropy, and indeed such behavior is commonly found. 
 Thixotropy is a time dependent rheological characteristic which occurs when 
 breakdown of a suspension occurs as a function of time as well as of shear 
 rate. Thus, if a suspension of activated sludge is continuously sheared at 
 a constant rate, the resulting shearing stress will normally diminish with 
 time approaching some equilibrium value at which the rate of breakdown of 
 
 96 
 
the sludge particles is equal to their rate of reformation. 
 
 VISCOMETRY 
 
 As suggested by equation (l), the rheological properties of yield 
 strength and plastic viscosity may be determined by measuring shearing 
 stress and velocity gradient at several flow conditions. This can be done 
 conveniently in a coaxial cylinder viscometer. In such a viscometer, a 
 sludge sample is placed in the annular space between two coaxial cylinders 
 and one cylinder is rotated relative to the other to produce velocity 
 gradients and hence shearing forces. The velocity gradient is deduced by 
 observing the relative angular velocity, and shearing stress is computed 
 from measuring torque on one of the cylinders. The basis of such calcula- 
 tions has been discussed previously (Dick and Ewing, 1967) and is described 
 in detail by Van Wazer et^ aj_. (1963). 
 
 Unique requirements of rotational viscometers used for observing the 
 rheology of activated sludge were discussed by Dick and Ewing (1967) and 
 the instrument used in this work was a refinement of the one described by 
 them. Figure 1 shows the geometry of the cylinders. The outer cylinder 
 was rotated by means of a continuously variable speed drive and the result- 
 ing torque on the inner cylinder was calculated by observing the deflection 
 of the calibrated wire by which it was suspended. 
 
 Basic procedures for operating the viscometer were as described by 
 Dick and Ewing (1967), except that activated sludge samples were introduced 
 into the viscometer through the hollow drive shaft at the bottom of the 
 outer cy 1 inder. 
 
 EXPERIMENTAL PLAN AND PROCEDURES 
 
 In order to observe the influence of biological variables on the physi- 
 cal properties of activated sludge, the rheological behavior of activated 
 
 97 
 
sludges cultivated under various controlled conditions was examined. Con- 
 tinuous flow activated sludge units such as shown in Figure 2 were used. 
 Solids recirculation was accomplished by aspiration of settled sludge under 
 the baffles separating the aeration and sedimentation compartments. Feed 
 was stored in a k°C refrigerator and, for each desired loading condition, 
 was pumped at a constant rate to the continuous flow treatment system. The 
 mixed liquor suspended solids concentration was maintained at the desired 
 level by daily wasting of sludge. In the studies of the influence of organic 
 loading intensity on physical behavior of sludges, no rheological observa- 
 tions were made until "steady-state" was reached as indicated by a constant 
 degree of COD removal and the absence of any trend in the amount of sludge 
 synthes is . 
 
 Studies on the effect of extended periods of endogenous conditions on 
 sludge characteristics were conducted by discontinuing feed and removing the 
 baffle from the activated sludge system. Studies of the changes in physical 
 properties of sludges which occurred during the feeding cycle were conducted 
 by discontinuing the feed, removing the baffle from the activated sludge 
 unit, and operating it as a batch feed system operating on a 24-hr cycle. 
 
 Preliminary studies showed that activated sludge grown in laboratory 
 units using chemically defined substrates bore little physical resemblance 
 to real activated sludge. Hence the laboratory unit was seeded with activated 
 sludge from the Urbana-Champai gn , Illinois Sanitary District Main Treatment 
 Plant and primary sedimentation tank effluent from the same plant was used 
 as feed. Unfortunately, it was not possible to achieve high organic loadings 
 using this waste because the unit became hydraul i cal ly overloaded. Hence it 
 was necessary to change to a more concentrated synthetic substrate for those 
 studies identified as "high-loading" experiments. To simulate waste of the 
 
 98 
 
nature commonly encountered, an artificial waste consisting of dry dog food 
 (Gaines Meal), toilet tissue (Scott Soft Weave), urea, and inorganic constitu- 
 ents was used. The synthetic waste was the same as used by Hunter e_t al . 
 (1966) except that the toilet tissue concentration was 580 mg/1 . The dog 
 food and toilet tissue were ground and shredded and mixed with a small quan- 
 tity of water in a blender for 5 min prior to addition to the synthetic 
 waste. The suspended solids content of the artificial waste was approximately 
 3000 mg/1, the BOD was 2200 mg/1, and the COD was 2400 mg/1. 
 
 COD, BOD and SVI measurements were conducted according to Standard 
 Methods (1965). Suspended solids were measured using the glass fiber filter 
 method described by Gratteau and Dick (1968). Dehydrogenase enzyme activity 
 was measured using the method proposed by Ford e_t_ a_l_. ( 1 966) although the 
 reader is referred to a recent paper on the subject by Patterson et al . 
 (1969) which appeared after this work was completed. 
 
 Equipment was available to permit 3 or k continuous flow activated 
 sludge units to be operated simultaneously. To cover the desired range of 
 loading factors it was necessary to make several different runs and a new 
 sample of seed sludge was used at the commencement of each individual run. 
 In the low loading studies (Chakrabart i , 1 968) , run A consisted of units 
 operating at 0.08, 0.12, and 0.16 lb/BOD/day/lb MLSS and run B included 
 units operating at 0.20, O.38, 0.49, and 0.61 lb/BOD/day/lb MLSS. In the 
 high loading study (Caban, 1 969) , three runs were used to span the desired 
 range of loading factors; run C - 0.15, 0.30, and 0.45 lb/BOD/day/lb MLSS; 
 run D - 0.60, 0.90, and 1.4 lb/BOD/day/lb/MLSS ; and run E - 2.0, 3.0, and 
 4.0 lb/BOD/day/lb MLSS. 
 
 To permit comparison at the same concentration of the yield strength 
 and plastic viscosity of sludges produced under different organic loading 
 
 99 
 
conditions, it was necessary to make rheological investigations over a 
 range of suspended solids concentrations. To achieve a range of concentra- 
 tions, mixed liquor was thickened or diluted with effluent. Four different 
 suspended solids concentrations within the range of about 1000 or 1500 mg/1 
 to 3500 mg/1 were normally examined at each organic loading intensity. 
 Then, by plotting yield strength or plastic viscosity as a function of con- 
 centration for each loading, values could be selected at any desired concen- 
 tration for purposes of comparison. 
 
 RESULTS AND DISCUSSION 
 Effect of Organic Loading Intensity 
 
 The yield strength over a range of suspended solids concentrations of 
 sludges cultivated at various organic loading intensities in the low loading 
 study are shown in Figure 3. Similar data from the high loading study are 
 shown in Figure k . It is seen that in both cases, the yield strength varied 
 with concentration according to the relationship 
 
 T y - je kC (2) 
 
 where j and k are constants for a particular sludge. This is the same re- 
 lationship reported from earlier studies with activated sludge (Dick and 
 Ewing, 1967) • 
 
 In Figures 3 and k, it may be seen that each individual run tended to 
 produce its own family of curves. This becomes more apparent when the yield 
 strength of activated sludge at various concentrations is plotted as a func- 
 tion of loading intensity as in Figure 5. It is seen that similar trends 
 were indicated in runs A and B but the relationship between loading and yield 
 strength was not continuous between the two runs. This can be attributed 
 to the fact that different samples of seed sludge were used to start the two 
 
 100 
 
units. The difference in the physical characteristics of the organisms 
 comprising the activated sludge persisted in spite of the fact that the 
 original seed had been essentially lost due to sludge wastage before the 
 Theological observations were made. The same phenomena was observed in 
 the high loading studies although the scale of the plot of loading versus 
 yield strength in Figure 6 disguises the discontinuities between individual 
 runs . 
 
 This variation in physical properties of sludges attributable to dif- 
 ferences in the nature of organisms present can also be seen in Figure 7. 
 There, the plastic viscosity of sludges produced in the high loading study 
 is shown as a function of concentration. The upper line represents a reason- 
 able fit for all data obtained from runs C and E and shows that the plastic 
 viscosity varied as a function of suspended solids concentration but was not 
 highly dependent on the organic loading. However, data from the family of 
 curves produced by run D form their own separate line indicated by the 
 lower curve in Figure 7- For a given sludge, the plastic viscosity varied 
 with suspended solids concentration and was not a function of loading. 
 However, as indicated by the great difference between the two curves, 
 plastic viscosity is highly dependent upon the nature of the organisms 
 which comprise the sludge which develops under any particular loading con- 
 dition. 
 
 The morphologic studies required to interpret the specific cause for 
 the change in rheological characteristics from run to run were not made. 
 However, recent studies by Wood (1970) confirm that changes in sludge 
 rheology under given conditions are attributable to population shifts amongst 
 the organisms comprising the sludge. 
 
 It is interesting to note that performance of the activated sludge units 
 
 101 
 
in terms of BOD removal was quite uniform and did not reflect the differences 
 in the physical properties of the organisms which predominated during the 
 various runs. Also, the dehydrogenase enzyme activity varied continuously in 
 a predictable manner as illustrated in Figure 8 and was not influenced by 
 the physical differences in the organisms. 
 
 The standard measurement of sludge physical properties, the sludge 
 volume index, was not sufficiently sensitive to detect the basic differences 
 in the sludge organisms present during the preliminary runs. Sludge volume 
 index measurements obtained during the high loading studies are shown in 
 Figure 9- They have been recorded at equal suspended solids concentrations 
 so that differences normally attributable to measurement variations in sus- 
 pended solids concentrations have been eliminated. 
 Changes in Physical Properties during Aerobic Digestion 
 
 During aerobic digestion, yield strength of activated sludge changed 
 both because of the change in the physical nature of the solids and because 
 of the decrease in suspended solids concentration. To separate these two 
 effects yield strengths during aerobic digestion were expressed as a percent- 
 age of the yield strength of a control sludge where the control sludge was 
 defined as a sample of equal concentration and loading intensity which had 
 not been exposed to endogenous conditions. Results from two different 
 sludges are shown in Figure 10. It is seen that pronounced changes in the 
 physical nature of the sludge solids occurred during aerobic digestion and 
 after six days the sludge solids had only 10-15 percent of their initial 
 yield strength. As might have been anticipated, the sludge produced at the 
 higher organic loading intensity was able to sustain itself under endogenous 
 conditions without pronounced change in physical properties for a day while 
 the sludge grown under leaner conditions changed from the onset of endogenous 
 
 102 
 
condi t ions . 
 
 Plastic viscosity also decreased during aerobic digestion. However, the 
 change was much less pronounced than the change in yield strength. 
 Changes in Physical Properties Following Feeding 
 
 To determine the physical response of activated sludge to feeding, the 
 rheological properties of a sludge acclimated to a 2^-hr feeding cycle was 
 monitored. Typical results are shown in Figure 11. It is seen that dramatic 
 changes occurred, particularly in yield strength. The wide scatter in yield 
 strength data during the first hour was caused by the fact that the sludge 
 sample was changing appreciably during the period of inspection in the vis- 
 cometer. Also, in an attempt to accurately define the early portion of the 
 yield strength curve, samples were collected during several successive feeding 
 cycles and all of those data are included on the same curve. The concentra- 
 tion of suspended solids in the batch yield changed somewhat with time because 
 of synthesis and endogenous respiration; however, during the run for which 
 data are shown in Figure 10, suspended solids did not fall below 1300 nor 
 above 1700 mg/1 and this change in solids content is inadequate to explain 
 the orders of magnitude change in yield strength. Rather, this change was 
 caused by a pronounced change in the physical nature of the activated sludge 
 solids in response to feeding. 
 
 SUMMARY AND CONCLUSIONS 
 
 Although the design, performance, and cost of activated sludge waste 
 treatment plants is highly dependent upon the physical properties of the 
 activated sludge produced, little is known about the effect of process vari- 
 ables on physical behavior of the sludge. The only parameter commonly used 
 to express the physical nature of activated sludge, the sludge volume index, 
 is influenced by many separate sludge properties and by laboratory test 
 
 103 
 
conditions and hence does not provide sufficient and i nterpretable informa- 
 tion on the physical properties of sludge. In this study, more fundamental 
 measures of the properties of suspensions, yield strength and plastic viscos- 
 ity have been used. These parameters have more direct influence on the per- 
 formance of sludge treatment processes such as thickening and flotation than 
 does the more conventional SVI value. Results of preliminary studies on the 
 influence of biological variables on yield strength and plastic viscosity 
 are presented here. 
 
 A yield strength of a particular activated sludge depends on its con- 
 centration and the organic loading intensity. Yield strength increases 
 exponentially with suspended solids concentration and also increases signifi- 
 cantly and continuously as the organic loading intensity is increased. 
 However, yield strength is also influenced by the nature of the particular 
 organisms which comprise the sludge, and two sludges of equal concentration 
 and organic loading grown on the same waste and giving the same biological 
 performance may display different physical properties if they are comprised 
 of different flora. , 
 
 The plastic viscosity of a sludge of given biological composition is 
 dependent only on the suspended solids concentration and is not markedly 
 influenced by a change in organic loading. However, again shifts in biolog- 
 ical population may produce significant changes in plastic viscosity without 
 accompanying changes in solids concentration, organic loading, or biological 
 process performance. 
 
 Yield strength decreases significantly during aerobic digestion. Reduc- 
 tions of 85-90 percent were observed within six days. 
 
 Remarkable changes in the yield strength of activated sludge occur follow- 
 ing contact with substrate. In the 2k hr batch feeding studies reported here, 
 
 104 
 
the yield strength of sludge increased by as much as two orders of magnitude 
 within the first 15 to 20 min following feeding, and then declinsed. 
 
 These changes in sludge rheology are caused by changes in the basic 
 physical nature of the sludge. While they must influence the performance of 
 the biological process and of waste sludge treatment processes, for the most 
 part, they remain undetected by standard measures of the physical character- 
 istics of sludge. 
 
 ACKNOWLEDGEMENTS 
 This work was supported in part by Research Grant 17070 DJR from the 
 Federal Water Quality Administration and by an undergraduate research grant 
 from the National Science Foundation. 
 
 REFERENCES 
 
 Bingham, E. C, "Fluidity and Plasticity," McGraw-Hill Book Co., New York 
 (1922). 
 
 Caban, G. L., "Physical Characteristics of Activated Sludge under High 
 Loading Conditions," Undergraduate Special Problem, Department of 
 Civil Engineering, University of Illinois, Urbana (1969). 
 
 Chakrabarti, S. K. , "Changes in Some Physical Properties of Activated Sludge 
 
 under Different Biological Conditions," Thesis submitted in partial 
 
 fulfillment for the degree of Master of Science, University of Illinois; 
 Urbana, 65 pp. (1968) . 
 
 Colin, F. , "Application de Techniques Rheologiques a 1 'etude des Boues 
 
 Residuaires," la Tribune du CEBEDEAU , 23_, 6317, 178-187 (April 1970). 
 
 Dick, R. I., "Thickening Characteristics of Activated Sludge," In Advances 
 
 in Water Pollution Research , Proceedings of Fourth International Confer- 
 ence on Water Pollution Research, Prague, 1969, 625-6^2 (1970). 
 
 Dick, R. I., and Ewing, B. B., "The Rheology of Activated Sludge," Journal 
 Water Pollution Control Federation , 39, z t, 5^3-560 (1967). 
 
 Dick, R. I., and Ves i 1 i nd , P. A., "The Sludge Volume Index - What Is It?" 
 Journal Water Pollution Control Federation , 4l_, 7, 1285-1291 (1969). 
 
 105 
 
Ford, D. L., Eckenfelder, W. W. , and Yang, T., "Dehydrogenase Enzyme as a 
 Parameter of Activated Sludge Act ivi t ies ." Proceed i ngs 21st Industria l 
 Waste Conference , Purdue Univ., Eng. Ext. Ser. 121 , 53^5^3 (1966) . 
 
 Geinopolis, A., and Katz, W. J., "A Study of a Rotating Cylinder Sludge 
 Collector in the Dissolved Air Flotation Process," Journal Water 
 Pollution Control Federation , 36, 6, 712 (June 196VJT 
 
 Gratteau, J. C, and Dick, R. I., "Activated Sludge Suspended Solids 
 Determinations," Water and Sewage Work s, 1 15 , 10, 468-^72 ( 1 968) . 
 
 Hunter, J. V., Genetelli, E. J., and Gilwood, M. E., "Temperature and 
 Retention Time Relationships in the Activated Sludge Process," 
 Proceedings 21st Industrial Waste Conference , Purdue Univ., Eng. Ext. 
 Ser. 121, 953-963 (1966). 
 
 Patterson, J. W. , Brezonik, P. L. and Putnam, H. D. , "Sludge Activity 
 
 Parameters and Their Application to Toxicity Measurements and Activated 
 Sludge," P roceedings 24th Industrial Waste Conference , Purdue Univ., 
 Eng. Ext. Ser. 135, 1 27~ 1 5^ (1969) ■ 
 
 Reiner, M. , "Deformation, Strain, and Flow," H. K. Lewis and Co., Ltd., 
 London, 3^7 pp. (1969) . 
 
 Scott Blair, G. W. , "Elementary Rheology," Academic Press, New York, 158 pp. 
 (1969). 
 
 Standard Methods for the Examination of Water and Waste Wat er, 12th 
 Ed., Amer. Pub. Health Assoc, New York, 769 pp. (1965). 
 
 Van Wazer, J. R., Lyons, J. W. , Kim, K. Y., and Colwell, R. E., "Viscosity 
 and Flow Measurement, " I nterscience, New York (1963). 
 
 Wood, R. F. , "The Effect of Sludge Characteristics upon the Flotation of 
 Bulked Activated Sludge," Thesis submitted in partial fulfillment of 
 the requirements for the degree of Doctor of Philosophy, University of 
 Illinois, Urbana, 1^3 pp. (1970) . 
 
 106 
 
- Oil Daaplag 
 
 Driving Shaft 
 Bearing 
 
 FIGURE 1. OQAXIAL CTLIMDCR VISCOMTER 
 
 107 
 
Influent F««d Lin* 
 
 Froa 
 R«frlg*ratox 
 
 Potltlv* 
 
 DisplacMMnt 
 Poap 
 
 -* — Air Supply 
 
 ttevabl* B«ff U 
 
 Overflow 
 
 — A*r*tor 
 
 FIGURE 2 SCHEMATIC DIAGRAM CP COITINUOUS FLOW 
 
 ACTIVATED SLUDGE UNIT 
 
 108 
 
in 
 
 o 
 
 z 
 
 QJ 
 QC 
 
 \n 
 
 Q 
 
 -J 
 UJ 
 
 > 
 
 Q 
 
 NUMBERS INDICATE 
 LOADING IN 
 IbBOD/DAY/lbMLSS 
 
 1000 2000 3000 4000 5000 
 
 SUSPENDED SOLIDS CONCENTRATION, 
 
 mg/* 
 
 FIGURE 3 
 
 VARIATION OF YIELD STRENGTH WITH SUSPENDED SOLIDS 
 CONCENTRATION - LOW LOADING STUDIES 
 
 109 
 

 
 (!) O.IS 
 
 
 
 (2) 0.30 
 
 too 
 
 — 
 
 (3) 0.U5 
 
 
 — 
 
 (U) 0.60 
 
 ftn 
 
 — 
 
 (5) 0.90 
 
 
 ™~ — 
 
 (6) \.k 
 
 (7) 2.0 
 
 (8) 3.0 
 
 so 
 
 
 (1) k.O 
 
 20 
 
 10 
 
 y 5 
 
 Lo*dln 9 f»ctors(lb iO0/<J«y/1b MISS) 
 
 500 1000 1500 2000 
 
 Suspended solids concent rat ton, 1*9/1 
 
 :'00 
 
 3000 
 
 FIGURE k. 
 
 VARIATION OF YIELD STRENGTH \<MTH SUSPENDED SOLIDS 
 CONCENTRATION - HIGH LOADING STUDIES 
 
 ltO 
 
>- CO 
 
 •z. o 
 
 Ol xj»/qi 'H19N3dlS QH3IA 
 
Susf*n4*d solid* concentration 
 
 (1) - 2000 *,/! 
 C) - 2500 rng/1 
 (3) - 3000 ng/1 
 
 2.0 3.0 
 
 Coatflnf , t» MD/H KISS 
 
 5.0 
 
 FIGURE 6. INCREASE IN YIELD STRENGTH WITH INCREASED ORGANIC LOADING - 
 
 HIGH LOADING STUDIES 
 
 112 
 
100 
 80 
 
 50 
 
 20 
 
 Q V 
 
 Loading factors (lb MO/day/lb MISS) 
 
 - 0.15 
 
 - 0.30 
 
 - 0.l»5 
 • 0.60 
 
 - 0.90 
 
 - ).k 
 
 - 2.0 
 D - 3.0 
 O - k.O 
 
 500 1000 1900 2000 
 
 Suspended solids concentration, mg/1 
 
 2500 
 
 3000 
 
 FIGURE 7. 
 
 CHANGES IN PLASTIC VISCOSITY DUE TO CHANGES IN SUSPENDED 
 SOLIDS CONCENTRATION - HIGH LOADING STUDIES 
 
 H3 
 
1.0 
 
 
 
 1 
 
 1 
 
 1 
 
 1 
 
 
 0.8 
 
 
 
 
 
 
 ■*" o~~ 
 
 — 
 
 0.6 
 
 
 
 
 s* 
 
 
 
 — 
 
 0.U 
 
 
 ofi 
 
 
 
 
 
 — 
 
 0.2 
 
 
 
 
 
 
 
 " ™ 
 
 0.0 
 
 
 
 1 
 
 1 
 
 1 
 
 1 
 
 
 FIGURE 
 
 2 3 * 
 
 Loadings, lb 100/ lb MISS 
 
 DEHYDROGENASE ENZYME ACTIVITY AT DIFFERENT ORGANIC 
 LOADINGS - HIGH LOADING STUDIES 
 
 114 
 
200 
 
 1 1 1 
 
 Suspsndad solids concentrations, mq/) 
 O - 1500 «♦*/» 
 
 O - 2500 mq/) 
 V - 3000 mq/\ 
 
 J L 
 
 t 2 3 % 
 
 Loading factor, lb lOD/day/lb MISS 
 
 FIGURE 9. CHANGE 114 SLUDGE VOLUME INDEX WITH ORGANIC LOADING 
 HIGH LOADING STUDIES 
 
 115 
 
lOaiNOD dO 1N3D U3d 'H±S)N3ti±S (TI3IA 
 
 116 
 
9 0I Xj»j/o»»-qt <Xai«oo«iA oia»»id 
 
 117 
 
APPENDIX VI I 
 
 THICKENING 
 
 by 
 Richard I. Dick 
 
 Reproduced from 
 
 Advances in Water Quality Improvement - 
 Physical and Chemical Processes 
 Edited by E. F. Gloyna and W. W. Eckenfelder, Jr 
 University of Texas Press, Pages 358-369 
 1970 
 
 118 
 
THICKENING 
 Richard I. Dick 
 
 University of Illinois, Urbana 
 
 Physical, biological, and chemical methods of waste treatment share the common goal 
 of concentrating pollutional material into a small portion of the total waste flow. This 
 permits return of the bulk of the flow to the environment, but leaves behind a trouble- 
 some residue. Most often, the materials removed from waste flows remain in the form of 
 suspended solids. Typically, the solids are light and flocculent and cannot be consoli- 
 dated readily into high concentrations for economical treatment and disposal. 
 
 The most economical means of reducing sludge volume is by thickening. To illustrate, 
 thickening of a waste sludge from 3 percent to 9 percent solids by weight results in a 
 three fold reduction in sludge volume. To accomplish an additional three fold volume 
 reduction (to 27 percent solids) would require a dewatering process such as vacuum fil- 
 tration. Burd (1) has indicated that the normal annual cost of gravity thickening is $1.50 
 to $5 per ton of dry solids, while the cost of vacuum filtration averages about $15 per ton 
 of dry solids and ranges as high as $50 per ton. 
 
 Concentration of sludges by gravity thickening and flotation is considered here. It 
 should be noted that the concepts discussed apply to the concentration of sludge as it oc- 
 curs in any settling tank or flotation unit and are not restricted to basins called "thicken- 
 ers" (5). 
 
 GRAVITY THICKENING 
 
 Analysis of the performance of full-scale continuous thickeners requires: (a) knowl- 
 edge of the basic settling characteristics of the suspension being thickened, and (6) 
 an understanding of the manner in which these basic settling properties govern the per- 
 formance of a continuous thickener. Theoretical means are available to relate settling 
 properties to continuous thickening. However, less is known about how to obtain data 
 which reveal settling behavior to be expected in full-scale settling tanks. These two items 
 are considered in the sections which follow. 
 
 Relationship between Settling Properties and Continuous Thickening 
 
 To predict the performance of a full-scale continuous thickener, itis necessary to know 
 the settling velocity of the sludge being thickened at all concentrations which could occur 
 in the thickener. Discussion of the problem of obtaining these data is reserved for the 
 section which follows. 'Here, it is assumed that data indicative of actual settling behavior 
 in a full-scale thickener are available, and their relationship to the performance of the 
 thickener is explored. 
 
 It is necessary to know the settling velocity of the sludge at each concentration which 
 could exist in a thickener because, for a given operating condition, the settling velocity 
 of sludge at some limiting concentration determines the capacity of the suspension to 
 pass solids to the bottom of the tank. To illustrate, consider first a batch thickener with 
 no underflow. The downward flux of solids, G B , at any point in the thickener is 
 
 G B = c t v { 
 
 119 
 
Thickening 
 
 SUSPENDED SOLIDS CONCENTRATION 
 
 Figure 195. Typical batch flux 
 
 where c f is the concentration of sludge at that point, and v t is the settling velocity cor- 
 responding to that concentration. A typical batch flux curve is shown in Figure 195. Note 
 that the flux is zero when Cj = and at extremely high concentrations where v, ap- 
 proaches zero, and that G B reaches a maximum at some intermediate concentration. 
 
 In a continuous thickener, withdrawal of solids from the bottom of the tank increases 
 the rate of downward transport of solids. If solids and downward flow are considered to 
 be uniformly distributed radially, the added flux, G u , due to underflow, is: 
 
 G u 
 
 (2) 
 
 where u is the downward velocity of the suspension relative to the thickener walls due 
 to sludge removal. In a thickener with cross-sectional area A, removal of sludge at volu- 
 metric underflow rate Q u gives: 
 
 -* 
 
 (3) 
 
 Because the total rate of solids throughput in a continuous thickener is Q u c„, where c u 
 is the concentration of the underflow, the solids flux in a continuous thickener may be 
 written as : 
 
 120 
 
Sludge Separations 
 
 TOTAL 
 FLUX 
 
 SUSPENDED SOLIDS CONCENTRATION 
 
 Figure 196. Total flux in a continuous thickener. 
 
 r _Qu£u 
 
 Lr c — —j-=C u U 
 
 (4) 
 
 Q u and u thus are related inversely to the underflow concentration, c u . 
 
 When a continuous thickener of area, A, is operated at some underflow velocity, u 
 (corresponding to a desired underflow concentration, c u ), the total possible flux, G c , for 
 any concentration of the suspension, c„ is the sum of Equations (1) and (2): 
 
 G c — CjVi + CjU 
 
 (5) 
 
 Figure 196 shows a plot of the total possible flux for solids concentrations which could 
 occur in a continuous thickener, and illustrates relative contribution of gravity subsid- 
 ence and underflow fo solids flux. Note that the possible solids flux passes through some 
 minimum value, G L , between the feed concentration, c F , and the underflow concentra- 
 tion, c u . The suspension at concentration c L thus limits the rate at which solids can reach 
 the tank bottom and becomes a bottleneck in operation of the continuous thickener, for 
 if solids are withdrawn at the desired underflow concentration, flux through the settling 
 tank cannot exceed G L . The required area for a thickener is, hence 
 
 121 
 
Thickening 
 
 A ^ (6) 
 
 where Q F and c F are the influent flow rate and solids concentration. If an area smaller 
 than QfC F IG L were to be provided, solids in the amount QpCy — AG L would be lost over 
 the effluent weir, for they could not be transmitted to the bottom of the thickener at a 
 rate greater than G L . To stop the loss of solids over the weir a compromise could be 
 made in the desired underflow concentration so as to increase the bulk downward veloc- 
 ity, u, to the extent that the limiting solids flux would be increased to equal the actual 
 solids loading rate. 
 
 While a total flux plot such as shown in Figure 196 is an instructive means of eval- 
 uating the behavior of a suspension in a thickener, the method is not convenient. This is 
 because a separate plot (with a different slope on the underflow flux line) is needed to 
 consider each alternate thickener area or underflow concentration. 
 
 An alternate, and more convenient, means of thickener analysis involves use of a 
 batch flux curve. From Equation (4): 
 
 G c 
 
 " = 7 (7) 
 
 Hence the scope of a line connecting any point on the ordinate (G) axis of a flux plot 
 with any point on the abscissa (c) gives the underflow velocity in a thickener with the cor- 
 responding flux and underflow concentration. From Equation (5), and Figure 197, it may 
 be seen that if this operating line intersects the batch flux curve (c^) the intercept of 
 the line on the ordinate axis will be (Cjf, + c,u) or the solids handling capacity of the sus- 
 pension at concentration c t when it is being thickened to c u , the intercept on the abscissa. 
 When the operating line is tangent to the flux plot, as in Figure 197, the intercept, G c , is 
 the maximum possible loading on the thickener, and the point of tangency is at c L , the 
 concentration of the suspension with the minimum capacity for transmitting solids under 
 the operating conditions described by the operating line. Intersections of the operating 
 line with the rising portion of the batch flux curve are. not of concern if the concentration, 
 c,-, corresponding to the point of intersection, is less than the feed concentration (or the 
 concentration to which the feed is diluted at the inlet. 1 
 
 By use of a flux plot such as shown in Figure 197, a variety of alternate thickener de- 
 signs can be evaluated readily by selecting various possible underflow concentrations, 
 drawing tangents from the c u values to the flux curve, and observing the corresponding 
 maximum flux curve, and observing the corresponding maximum flux values from the 
 intersections with the ordinate axis. The thickener area required to produce various al- 
 ternate thickened sludge concentrations may then be determined from Equation (5). 
 
 The flux plot may be used to describe performance of existing thickeners as well. The 
 influence of a change in the volumetric feed rate or a change in the solids concentration 
 of the feed may be quickly evaluated by plotting the resulting solids loading on the or- 
 
 ■A more common method of showing the utility of a batch flux curve (16) is based on the work of Kynch 
 (10). Kynch showed that in batch sedimentation, a slight concentration discontinuity is propagated upward 
 at a velocity equal to —dG/dc. Such a discontinuity would form at the top of a layer of concentration c L because 
 of its inability to transmit solids at a rate faster than the limiting solids flux. In a continuous thickener, the 
 rate limiting layer must be prevented from being propagated upward to avoid loss of solids over the effluent 
 weir. This can be done by withdrawing solids at a rate sufficient to give a bulk downward velocity, u, equal 
 to the rate of upward propagation of the layer in a batch test, —dGldc. This rate is given by the slope of the 
 tangent to flux curve as illustrated in Figure 199. The alternate derivation presented here is equivalent in 
 utility to that developed by Kynch. 
 
 122 
 
Sludge Separations 
 
 dinate axis, drawing a tangent to the flux curve, and observing the resulting underflow 
 concentration. 
 
 Illustrative Problems 
 
 1. Determine the required area for a thickener which is to concentrate 0.3 mgd of 
 waste-activated sludge from an initial concentration of 7,000 mg/1 to 2 percent solids by 
 weight. Settling characteristics .of the activated sludge, which has excellent physical 
 properties, are shown in Figure 198. 
 
 Solution: Figure 199 shows a solids flux plot obtained by multiplying each settling ve- 
 locity in Figure 198 by its corresponding concentration. Operating line 1 is drawn as a tan- 
 gent to the flux plot from the desired underflow concentration, 20,000 mg/1, and indi- 
 cates that the maximum solids flux through the thickener is 13 lb/ft 2 /day. The required 
 area is, hence 
 
 Am Q& 
 
 (0.3) (7,000) (8.34) 
 13 
 
 1,350 ft 2 
 
 2. Following construction of the thickener as prescribed by the calculations above, the 
 
 
 sG L = 
 
 c L v L + c L u 
 
 
 X 
 
 3 
 
 
 ulS. \ 
 
 
 C/) 
 Q 
 
 -J 
 O 
 C/) 
 
 
 . i 
 
 « 
 
 Up Ul_ u u 
 
 SUSPENDED SOLIDS CONCENTRATION 
 
 Figure 197. Use of flux curve to identify limiting conditions. 
 
 123 
 
Thickening 
 
 5000 10,000 15,00023,000 
 CONCENTRATION, c,, mq/2 
 
 Figure 198. Settling properties of sludge considered in illustrative problem. 
 
 flow of waste-activated sludge increases to 0.4 mgd, and the feed sludge concentra- 
 tion increases to 8,000 mg/1. Physical characteristics of sludge solids are unaltered. What 
 will be the consequence of continuing to withdraw sludge at 2 percent solids? 
 
 Solution: Solids are now being applied at 19.8 lb/ft £ /day. But under the operating 
 conditions, only 13 lb/ft 2 /day can reach the bottom of the tank. Thus 6.8 lb/ft 2 /day or 
 9,200 lb/day of solids will be lost over the effluent weir. 
 
 3. How can operation of the thickener be modified to avoid loss of solids? 
 
 Solution: The rate of removal of sludge from the thickener must be increased to in- 
 crease the solids flux. Line 2 in Figure 199 shows the modified operating conditions. 
 It is drawn as a tangent to the flux plot from the required solids flux, 19.8 lb/ft 2 /day, 
 and shows that the underflow solids concentration must be decreased to about 17,500 
 mg/1 to accommodate the increased loading on the thickener. Note that the sludge which 
 limits the capacity of the thickener now has a concentration of about 12,500 mg/1. 
 
 Alternatively, and more laboriously, the modified conditions can be determined by 
 use of a plot of Equation (5) as shown in Figure 200. The original operating conditions are 
 shown as solid lines and illustrate the relative contribution of underflow and subsidence 
 
 }2k 
 
Sludge Separations 
 
 5000 10,000 15,00020,000 
 
 CONCENTRATION, c., mg/J? 
 
 Figure 199. Operating conditions considered in illustrative problem. 
 
 to the solids handling capacity of the suspension. Note that the minimum solids flux at 
 14,500 mg/1 agrees with the value obtained before. To increase the capacity of the thick- 
 ener to receive solids, the underflow velocity (the slope of the underflow flux curve) 
 must be increased as illustrated by the dashed lines in Figure 200. By trial-and-error, it 
 is found, as before, that when the underflow rate is increased to 183,000 gpd (correspond- 
 ing to an underflow concentration of 17,500 mg/1) the possible flux through the thickener 
 is equal to the applied load, 19.8 lb/ft 2 /day. 
 
 Measurement of Settling Properties 
 
 The preceding analysis of continuous thickening was founded on the assumption that 
 the settling velocity of any concentration of sludge as it might exist in the full-scale con- 
 tinuous thickener was known. It remains to show how these data may be obtained. 
 
 The settling velocity of various concentrations of sludge as displayed in laboratory 
 batch settling columns can be determined readily. A transparent cylinder is filled with a 
 uniformly dispersed sample of sludge of known concentration and a plot of the solids- 
 
 125 
 
Thickening 
 
 liquid interface position as a function of time is obtained. As illustrated in Figure 201, 
 the initial slope of this curve, following a period of reflocculation, is the settling velocity 
 of the sludge at its initial concentration under the laboratory hatch-settling conditions. 
 
 Few opportunities are available to compare laboratory batch-settling results directly 
 with the real settling properties of sludge in a full-scale thickener. However, sufficient 
 evidence is available from laboratory data obtained under various conditions to realize 
 that these data can be seriously misleading unless care is exercised in obtaining and 
 interpreting them. 
 
 One obvious factor which can cause laboratory settling behavior to differ from sedi- 
 mentation in a full-scale thickener is that the container boundary is more prominent in 
 laboratory vessels. In order to make effective use of laboratory data, one must use a 
 laboratory column large enough to reduce the wall effect significantly and understand 
 the effect of the diameter chosen on the results obtained. Unfortunately, because of the 
 complexity of the effect of the wall on sedimentation and because the effect is highly de- 
 pendent on the type of suspension being observed, one can find convincing evidence in 
 the literature to justify any column size selected. 
 
 60 
 
 o 
 
 o 
 
 X 
 
 3 
 
 d 
 
 if) 
 
 MODIFIED TOTAL FLUX 
 
 MODIFIED 
 UNDERFLOW^/ 
 
 j 
 
 BATCH 
 FLUX 
 
 UNDERFLOW FLUX 
 
 I l 
 
 5000 10,000 15,00020,000 
 
 CONCENTRATION, c., mg/J 
 
 Figure 200. Relative significance of underflow and subsidence for the operating conditions con- 
 sidered in illustrative problem. 
 
 126 
 
Sludge Separation* 
 
 TIME 
 
 Figure 201. Typical batch settling curve. 
 
 Data for activated sludge presented by Vesilind (17) illustrate the complex effect 
 of the cylinder wall. The smooth path along the wall provides a less tortuous means of 
 escape for displaced liquid and thus tends to increase the rate of subsidence. This is 
 counteracted by a tendency for arching of the sludge solids across the column, which 
 reduces the observed subsidence rate. At relatively low concentrations, the hastening 
 of subsidence because of flow of displaced liquid along the wall is the predominant ef- 
 fect, and sludge tends to settle faster in the laboratory than in the field. However, at 
 higher concentrations the arching mechanism predominates, and sludges tend to settle 
 more slowly in the laboratory than in the field. 
 
 The depth to which sludge is placed in laboratory settling columns can also influence 
 results. The Kynch theory (10), on which much thickening work is based, supposes that 
 settling velocity is independent of depth. While the theory is a valuable description of 
 ideal settling behavior, it does not describe completely the actual settling behavior of the 
 real suspensions encountered in waste treatment (5). With these sludges, greater depths 
 are accompanied by higher settling velocities. While the effect of depth is particularly 
 significant at high solids concentrations, there is no abrupt change to compression con- 
 dition. Rather, an element of compression exists even at dilute concentrations. 
 
 In the preceding section and in the illustrative problem, the customary consideration 
 
 127 
 
Thickening 
 
 of the required volume or depth of the thickener as descrihed hy Roberts (1 1) was absent. 
 This is because the settling velocities used were taken as those which actually occur in 
 the thickener. That is, some "effective" depth in the thickener was assumed and this 
 depth was used when the laboratory settling data were collected. In this way the deter- 
 mination of thickener area and depth are interrelated. Should it be desired to consider 
 alternate effective depths, alternate flux curves corresponding to different initial depths 
 could be evaluated. 
 
 It is, therefore, necessary to conduct laboratory settling tests at initial depths which 
 give compression conditions comparable to the effective compression the corresponding 
 concentration layer would experience in a full-scale thickener. Unfortunately the "ef- 
 fective" depth of a particular concentration layer in a full-scale thickener cannot be 
 determined directly. This is because the weight of the overlying, more dilute solids, 
 transmitted downward, influences the "effective" depth of a particular concentration in 
 a continuous thickener as does the weight and structural rigidity of the more concen- 
 trated sludge below. Fortunately, sludge depth, unlike thickener area, is an operational 
 variable. If settling tests for determination of required thickener area are conducted at 
 depths conservatively representative of probable average field conditions, some flexibil- 
 ity will remain in operation of the thickener. Rarely are thickeners operated under steady 
 conditions. Indeed, the most rational approach to establishing the depth of a thickener 
 may be to consider the volume required to store the excess sludge which accumulates 
 when the feed rate exceeds the solids-handling capacity of the thickener. 
 
 Mixing is another factor which influences settling as observed in laboratory batch 
 tests. Traditionally, it has been thought that slow-speed stirring is necessary in laboratory 
 settling tests to duplicate the mixing conditions caused by rake action and turbulence in 
 full-scale thickeners. Recent work by Vesilind (18) suggests that stirring serves to elim- 
 inate the artificial conditions imposed by small laboratory settling columns rather than 
 to duplicate the stirring conditions in full-scale thickeners. Vesilind tried a variety of 
 stirrer configurations and speeds in settling columns of various diameters. He found that 
 slow-speed stirring dramatically hastened settling in small columns, but that the bene- 
 ficial effect diminished with increasing column diameter, and that no difference be- 
 tween stirred and unstirred conditions could be detected in 36-in diameter columns. 
 Vesilind's conclusion was that while slow-speed stirring is essential in laboratory settling 
 tests, its purpose is not to simulate the pickets or rakes in full-scale thickeners, but to 
 cause the same agglomeration of solids to occur in small containers that occurs naturally 
 in large containers. 
 
 While the ordinary effect of slow stirring of sludges from waste treatment is to in- 
 crease settling rates, Scott has reported that slow stirring reduces the subsidence rate 
 of flocculated pyrophylite (13), calcium carbonate (14), and flocculated silica (15). Scott 
 has interpreted the cause of this decrease to be the disruption of channels which form 
 between sludge particles to permit more rapid egress of displaced water. He consid- 
 ered the channels to be a laboratory anomaly. Their presence in field-scale thickeners is 
 prevented by rake and slurry movement. 
 
 Thus, although laboratory batch-settling tests are easy to run, their interpretation is 
 difficult. Observed results depend on the conditions under which the tests are conducted 
 and on the properties of the sludge tested. The difficulties in understanding laboratory 
 sedimentation and in applying results to the design or control of thickeners have led to 
 several recent studies of the fundamental behavior of thickening suspensions. The rheo- 
 logical behavior of activated sludge has been investigated (4) and related to the thick- 
 ening characteristics of the suspension (3). Scott (13) and (14) has described three settling 
 regimes for concentrated suspensions. At dilute concentrations, particles settle as in- 
 
 128 
 
Sludge Separations 
 
 dividual units as in an "ideal" slurry. At high concentrations, subsidence is brought 
 about by mechanical compression. At intermediate concentrations, channel formation 
 gives rise to abnormally high settling velocities. The formation of channels within set- 
 tling slurries has also been studied by Dell and Kaynar (2). Scott (12) has described a 
 change from interaggregate to intra-aggregate How during thickening, and Javaheri and 
 Dick have computed the changes in size, number, and specific gravity of activated- 
 sludge aggregates as thickening takes place. 
 
 To summarize, the major .factors which affect laboratory data are diameter, depth, 
 and stirring conditions. I have used column diameters of 3.5 in or more, depths of 
 about 3 ft, and mixer-tip speeds of slightly less than 10 in/min for work with activated 
 sludge. However, it is not advisable to establish standard test conditions because of dif- 
 ferences in the physical nature of various types of sludges. 
 
 FLOTATION 
 
 Recent developments in flotation thickening have largely been in application of the 
 process. Little has been done to advance basic understanding of the process or to de- 
 velop rational methods for design and control of this method of sludge thickening. 
 
 Because of the difficulty in thickening activated sludge to high concentrations by 
 gravity thickening, most recent applications of dissolved air-pressure flotation in waste 
 treatment have been for concentrating activated sludge. Katz and Geinopolos (9) and 
 Jones (8) have reviewed the performance of activated-sludge flotation equipment at twen- 
 ty plants. Float solids averaged 4.6 percent at plants which do not condition with poly- 
 mers, while solids capture averaged slightly less than 90 percent. Plants which used 
 polymers captured an average of 98.6 percent of the activated-sludge solids and pro- 
 duced an average thickened sludge with 5.8 percent solids. 
 
 At present, design and control of flotation processes is an empirical art involving cor- 
 relations of air to solids ratios and rise rates with float solids concentrations and clari- 
 fication efficiencies. Other than the difference in the direction of separation, flotation 
 differs from gravity thickening only in that flotation systems include provisions for 
 rendering the solids floatable. This introduces design and operational variables which are 
 not involved in gravity thickening. While the basic utility of the solids flux curve (Figure 
 197) would seem to extend to flotation thickening, it has not been used in evaluating the 
 effect of operational variables such as recycle ratio, feed concentration, pressure, poly- 
 mer concentration, and float concentration. Procedures have been developed for con- 
 ducting laboratory flotation tests (6), but the relationship of flotation characteristics in 
 laboratory units to those in full-scale equipment has not been reported. It is likely that 
 the limitations of laboratory flotation data are less severe than for laboratory settling 
 data, because of the faster separation rates involved. 
 
 SUMMARY 
 
 Thickening is an important part of waste-treatment systems because it accomplishes 
 major sludge volume reductions at a low cost. Thickening may be accomplished in grav- 
 ity or flotation thickeners. Recent fundamental advances in thickening have related 
 mostly to gravity thickening. 
 
 The theoretical thickening behavior of sludges is relatively well understood, and con- 
 venient techniques are available for using knowledge of basic suspension settling prop- 
 erties for thickener design and for rational control of thickener operation. Determination 
 of the required area and volume for thickening are interrelated. It is proposed that area 
 
 129 
 
Thickening 
 
 determination should be based on settling characteristics of the suspension while volume 
 determination normally should be based on the need for storing sludge during periods 
 when the solids feed rate exceeds the capacity of the thickener for passing solids. 
 
 The major need in thickening technology at present is to learn more about the manner 
 in which sludges subside in laboratory and full-scale thickeners so as to permit collection 
 of settling data which give more accurate determination of design and operational re- 
 quirements. It is thought, however, that thickening technology has advanced to the de- 
 gree that sizing and operation of thickeners may be done on a rational basis. 
 
 REFERENCES 
 
 (1) Burd, R. S. A Study of Sludge Handling and Disposal. Federal Water Pollution Control 
 
 Administration, Publication No. WP-20-4(1968). 
 
 (2) Dell, C. C. and M. B. Kaynar. "Channelling in Flocculated Suspensions," Filtration and Sep- 
 
 aration, 5, 323 (1968). 
 
 (3) Dick, R. I. "Thickening Characteristics of Activated Sludge," Advances in Water Pollution 
 
 Research, Proceedings of Fourth International Conference on Water Pollution Research 
 held in Prague, Czechoslovakia (September 1968). 
 
 (4) Dick, R. I. and B. B. Ewing, "The Rheology of Activated Sludge," Journal, Water Pollution 
 
 Control Federation, 39, 543 (1967). 
 
 (5) . "Evaluation of Activated Sludge Thickening Theories," Journal, Sanitary Engineer- 
 ing Division, American Society of Civil Engineers, 93 (SA4), 9 (1967). 
 
 (6) Eckenfelder, W. W., Jr. Industrial Water Pollution Control New York, N. Y.: McGraw-Hill, 
 
 Inc. (1966). 
 
 (7) Javaheri, A. R., and R. I. Dick. "Aggregate Size Variations During Thickening of Activated 
 
 Sludge," Journal, Water Pollution Control Federation 41 (2) R 197 (1969). 
 
 (8) Jones, W. H. "Developments with Pressurized Flotation," Paper presented at 38th Annual 
 
 Conference of the Water Pollution Control Federation, Atlantic City, New Jersey (1965). 
 
 (9) Katz, W. J., and A. Geinopolos. "Sludge Thickening by Dissolved-Air Flotation," Journal, 
 
 Water Pollution Control Federation, 39, 946 (1967). 
 
 (10) Kynch, G. J. "A Theory of Sedimentation," Transactions, Faraday Society, 48, 166 (1952). 
 
 (11) Roberts, E. J. "Thickening: Art or Science?" Mining Engineering, 1, 61 (1949). 
 
 (12) Scott, K. J. "Mathematical Models of Mechanisms of Thickening," Industrial and Engineer- 
 
 ing Chemistry Fundamentals, 5, 109 ( 1966). 
 
 (13) . "Theory of Thickening: Factors Affecting Settling Rate of Solids in Flocculated 
 
 Pulps," Transactions, Institution of Mining and Metallurgy, 77, C85 (1968). 
 
 (14) . "Thickening of Calcium Carbonate Slurries: Comparison of Data with Results for 
 
 Rigid Spheres," Industrial and Engineering Chemistry Fundamentals, 7, 484 (1968). 
 
 (15) . "Experimental Study of Continuous Thickening of a Flocculated Silica Slurry," 
 
 Industrial and Engineering Chemistry Fundamentals, 7, 582 (1968). 
 
 (16) Shannon, P. T., R. D. Dehaas, E. P. Stroupe, and E. M. Tory. "Batch and Continuous Thick- 
 
 ening," Industrial and Engineering Chemistry Fundamentals, 3, 250 (1964). 
 
 (17) Vesilind, P. A. "Discussion of 'Evaluation of Activated Sludge Thickening Theories' by R. 
 
 I. Dick and B. B. Ewing," Journal, Sanitary Engineering Division American Society of 
 Civil Engineers, 94 (SA-1), 185 (1968). 
 
 (18) . "The Influence of Stirring in the Thickening of Biological Sludge," Ph.D. Thesis, 
 
 University of North Carolina, Chapel Hill (1968). 
 
 130 
 
APPENDIX VI II 
 
 ROLE OF ACTIVATED SLUDGE FINAL SETTLING TANKS 
 
 by 
 Richard I . Dick 
 
 Reproduced from 
 
 Journal of the Sanitary Engineering Division 
 Proceedings of the American Society of Civil Engineers 
 Volume 96, No. SA2 , Pages ^23-^36 
 April, 1970 
 
 131 
 
7231 April, 1970 
 
 Journal of the 
 SANITARY ENGINEERING DIVISION 
 
 Proceedings of the American Society of Civil Engineers 
 
 ROLE OF ACTIVATED SLUDGE FINAL SETTLING TANKS 
 By Richard I. Dick, 1 M. ASCE 
 
 INTRODUCTION 
 
 The characteristic feature of the activated sludge process is that biolog- 
 ical solids are separated from treated effluent for recycle back into the 
 process. Efficient performance requires that the separation be done effec- 
 tively because the solids are needed to sustain the process and because solids 
 which escape separation impair the quality of the effluent. In addition, it is 
 important that the solids be consolidated into a concentrated suspension be- 
 fore being recycled because the performance of the overall process depends 
 on the number of active microorganisms which can be crowded into each unit 
 of aeration tank volume. 
 
 Unfortunately, the light, flocculent, biological solids formed in the ac- 
 tivated sludge process usually do not settle and compact well. Hence, the 
 effluent quality which can be achieved by activated sludge treatment as well 
 as the size and cost of the required facilities often is controlled by the set- 
 tling characteristics of the sludge. 
 
 Final settling tanks thus must be designed to accomplish high degrees of 
 recovery and consolidation of activated sludge solids. Regrettably, attempts 
 to design activated sludge waste treatment plants on a rational basis normally 
 have been limited to consideration of biological aspects of the process. Little 
 attention has been given to rational analysis of the solids separation phase. 
 
 In this paper, the need for more thorough consideration of the final settling 
 tank in design of activated sludge plants is noted, and the fundamental bases 
 for rational design of the tank are described. An example is presented to 
 
 Note.— Discussion open until September 1, 1970. To extend the closing date one 
 month, a written request must be filed with the Executive Secretary, ASCE. This paper 
 is part of the copyrighted Journal of the Sanitary Engineering Division, Proceedings of 
 the American Society of Civil Engineers, Vol. 96, No. SA2, April, 1970. Manuscript 
 was submitted for review for possible publication on September 19, 1969. 
 
 i Assoc. Prof, of Sanitary Engrg., Dept. of Civ. Engrg., Univ. of Illinois, Urbana, 111. 
 
 132 
 
April, 1970 
 
 SA 2 
 
 show how failure to consider the solids loading on the final settling tank can 
 lead to unsatisfactory performance of the entire activated sludge process. 
 
 ROLE OF FINAL SETTLING TANK 
 
 Like any settling basin, the final settling tank in the activated sludge pro- 
 cess has two functions: (1) The production of an effluent which is relatively 
 free of settleable solids; and (2) the production of an underflow which con- 
 tains, in high concentration, the solids which have been settled in the tank. 
 Both functions, clarification and thickening, must be considered in design if 
 the basin is to satisfactorily accomplish both of its tasks. 
 
 The importance of the clarification function of final settling tanks is illus- 
 trated in Fig. 1. As shown by the figure, the total BOD of the effluent from an 
 activated sludge treatment plant is the dissolved BOD plus the BOD of the 
 suspended solids. While the BOD equivalent of activated sludge solids de- 
 pends on the treatment plant loading (8) 2 and other factors, each milligram 
 
 Effluent Suspended Solids Concentration , mg 
 
 FIG. 1.— EFFECT OF CLARIFICATION ON QUALITY OF EFFLUENT FROM ACTI- 
 VATED SLUDGE PLANT 
 
 of solids lost over the weir of the final settling tank commonly increases the 
 effluent BOD by about 0.6 mg. Thus, if, as in Fig. 1, the dissolved effluent 
 BOD is 10 mg per 1, then with an effluent suspended solids concentration of 
 about 20 mg per 1, over half of the total BOD leaving an activated sludge plant 
 is in the form of solids which were not removed in the sedimentation process. 
 The further deterioration of effluent quality which can be caused by still 
 less effective solids removal is apparent from Fig. 1. 
 
 The consequence of inadequate performance of the other function of final 
 settling tanks, thickening, is more complex. Improper design for the thickening 
 function of final tanks can lead to direct deterioration of effluent quality by 
 loss of suspended solids just as improper design of the clarification function 
 can. In this case solids are lost because the basin lacks adequate capacity 
 for conveying solids to the bottom of the tank. The mechanism responsible 
 for loss of solids due to inadequate attention to thickening is described in 
 
 2 Numerals in parentheses refer to corresponding items in the Appendix I.— 
 References. 
 
 133 
 
SA 2 
 
 ACTIVATED SLUDGE 
 
 more detail in a later section. In addition, improper thickening in the final 
 settling tank can lead to disruption of the treatment process by altering con- 
 ditions in the biological phase of the process because of inefficient return of 
 solids. This effect can be illustrated by considering that the basic design 
 parameter in the activated sludge process is the number of pounds of BOD 
 applied per pound of activated sludge solids under air, i.e., the organic loading 
 intensity. The amount of BOD applied is established by the characteristics of 
 the waste and the waste flow. The mass of activated sludge to which this load 
 is applied depends on the volume of the aeration tank and the mixed liquor 
 suspended solids concentration, (MLSS). With an existing plant, the aeration 
 tank volume is fixed and the MLSS depends on the degree of sludge consoli- 
 dation which is accomplished in the final settling tank. Fig. 2 illustrates this 
 dependence of the MLSS on the underflow concentration from the final settling 
 
 
 
 Q + R 
 
 
 
 
 Q= WASTE FLOW 
 
 , , c = MLSS j 
 
 Q*R 
 
 , 1 Q 
 
 c - ' 
 
 
 AERATION 
 TANK 
 
 C = Co 
 
 FINAL SETTLING C»*0 
 mNK \ 
 
 
 
 
 R+ W 
 
 
 
 
 
 
 C = C u 
 
 
 
 R = RECYCLE 
 
 
 
 
 c = c u 
 (Q*R)MLSS = Q x ♦ R x c u 
 
 MLSS = qTfT c " 
 
 FIG. 2. -DEPENDENCE OF MIXED LIQUOR SUSPENDED SOLIDS CONCENTRATION 
 ON FINAL SETTLING TANK UNDERFLOW CONCENTRATION 
 
 tank, c u . If the suspended solids content of the waste flow and synthesis of 
 solids in the aeration tank are ignored then 
 
 R 
 
 MLSS = 
 
 (1) 
 
 in which Q and R are the waste flow rate and the rate of recycle. As illus- 
 trated by Eq. 1, failure to design the final settling tank to give the desired 
 underflow concentration will lead to a decrease in the MLSS and hence to 
 an increase in the organic loading intensity. The increase in organic loading 
 intensity may adversely alter the flocculating and settling characteristics of 
 the sludge (7) and make it even more difficult to achieve the desired underflow 
 concentration. Hence, the problem becomes compounded. In this case, what 
 may appear to be a problem related to the biological aspects of the activated 
 sludge process actually stems from inadequate design of the solids separation 
 phase of the process. 
 
 From the preceding analysis it may be seen that the size and cost of the 
 aeration tank can be reduced by adequate design of the final settling tank. 
 That is, if the final tank can be designed to assure that solids will be re- 
 turned in sufficiently high concentration, the required amount of solids under 
 aeration can be contained in a smaller aeration tank. Also, since the cost of 
 virtually all methods of waste sludge treatment and disposal depends on the 
 concentration of solids in the sludge, proper thickening in the final settling 
 
 134 
 
April, 1970 SA 2 
 
 tank can reduce the cost of dealing with the waste sludge which is identified 
 as W in Fig. 2. 
 
 DESIGN OF FINAL SETTLING TANKS 
 
 To assure that a final settling tank satisfactorily accomplishes both of its 
 functions, both clarification and thickening must be considered in design. If 
 the size of the tank required to give the desired degree of clarification ex- 
 ceeds the size required to give the desired degree of thickening, then the 
 clarification function governs sizing of the final settling tank. Conversely, if 
 the thickening requirement is greater, thickening governs. Conventional de- 
 sign procedures have considered only the clarification function, and the re- 
 quirements for thickening are normally overlooked. Because the concepts 
 relating to design requirements for the clarification function are well es- 
 tablished they will be given only brief attention here. More attention will be 
 given to the less familair concepts relating to design for the thickening 
 function. 
 
 For proper clarification, sufficient surface area must be provided in the 
 final settling tank so that the hydraulic loading per unit area does not exceed 
 the settling velocity of the slowest settling material which is to be completely 
 removed. The settling velocity of interest is often taken as the zone settling 
 velocity of the activated sludge at the MLSS concentration (5), although it is 
 likely that consideration of the subsidence of individual floe particles found in 
 the upper regions of final settling tanks would impose more limiting con- 
 ditions. If it is desired to take advantage of the flocculant nature of the acti- 
 vated sludge solids to accomplish removal of the dilute solids in the upper 
 part of final settling tanks, then sufficient volume must be provided in the 
 clarification section of the tank to permit particle collisions to occur (4). 
 
 To assure that final settling tanks satisfactorily accomplish their thickening 
 function, sufficient area must be provided so that solids are applied at a rate 
 which is not in excess of the rate at which solids are able to reach the bottom 
 of the tank. The biological solids are transported to the bottom of the final 
 settling tank by two mechanisms. One mechanism is their subsidence under 
 the influence of gravity, and the other is the bulk downward transport due to 
 sludge withdrawal from the bottom of the settling tank. The total rate at which 
 solids of concentration c i pass downward in the final settling tank is 
 
 G = c i v i + c { u (2) 
 
 in which G is the solids flux, commonly in pounds per square foot per day, v { 
 is the settling velocity of the activated sludge solids at concentration c z -, and 
 u is the average downward velocity caused by removal of sludge from the 
 bottom of the settling tank at concentration c u . 
 
 The amount of solids transported by gravity subsidence, c i v i , depends on 
 the settling characteristics of the sludge, whereas the amount transported by 
 sludge withdrawal, CjU, is a variable controlled within limits by the operator 
 of the activated sludge process. The operator can vary the value of it by 
 varying the rate of removal of thickened sludge from the bottom of the tank. 
 Assuming essentially complete solids capture, the total mass of solids, M, to 
 be thickened in a final settling tank per unit time is 
 
 M = Q n c n = Q,.c y (3) 
 
 135 
 
SA 2 ACTIVATED SLUDGE 
 
 in which Q u and c u are the flow rate and solids concentration of sludge re- 
 moved from the tank; and Q and c are the influent flow rate (including re- 
 cycle) and solids concentration. The average bulk downward velocity due to 
 solids removal, n, is then - 
 
 "-*-& w 
 
 Thus, for a fixed solids loading, M, on a final settling tank with area, A, the 
 rate of bulk downward movement, u, is related inversely to the concentration 
 at which sludge is withdrawn from the tank. The goal of proper solids control 
 in the activated sludge process might be thought of as getting the solids to the 
 bottom of the final settling tank by maximizing the c i v i term in Eq. 2 through 
 improvement of the settling characteristics of the sludge solids and by 
 minimizing the Cjti term by removing underflow at the highest possible 
 concentration. 
 
 In the final settling tank, opportunity is afforded for activated sludge to 
 exist in all concentrations from that of the mixed liquor to that of the under- 
 flow. In fact, if dilution with previously clarified effluent occurs at the inlet, 
 concentrations below the MLSS concentration will exist. Hence, c i in Eq. 2 
 can assume all values from less than c to c u . With a particular activated 
 sludge, the rate of gravity subsidence, v if depends primarily on the solids 
 concentration, c e -. Characteristically, v i varies with c z - in such a fashion that 
 the flux rate, G, in Eq. 2 passes through a minimum for some limiting con- 
 centration of activated sludge solids. This minimum value of possible solids 
 flux is the limiting solids handling capacity, G L , of the suspension and is the 
 basis for determining the area required in a final settling tank to accomplish 
 the desired degree of thickening. This required area is 
 
 A = -?- = ?o£a (5) 
 
 For example, if a final settling tank is to receive 60,000 lb per day of solids 
 and the limiting solids handling capacity, G L , as given by the minimum value 
 of Gfor concentrations expected to occur within the settling tank, is 20 lb per 
 sq ft per day, then 3,000 sq ft of final settling tank area must be provided to 
 assure that all solids will be transported to the bottom of the tank. 
 
 It ordinarily has been considered that in addition to the area required for 
 thickening, a certain volume, or depth, also must be provided. The volume 
 requirement commonly has been felt to be associated with the need for time 
 to accomplish compression of the sludge solids. Analysis of the volume re- 
 quirement for thickening has not been developed on a rational basis, but sev- 
 eral empirical methods have been proposed (1,6). Fortunately, volume is not 
 considered to be the primary factor controlling thickening (11). Eq. 2 is a 
 general equation which applies to all concentrations of sludge. There would 
 seem to be no need for a separate set of concepts to deal with the high con- 
 centrations associated with compression of sludges as long as all concen- 
 trations up to the underflow concentration are considered in selecting the 
 limiting solids flux. The volume required for thickening is, then, that which 
 provides sufficient depth to assure that settling velocities which occur in the 
 final settling tank are the same as those used in Eq. 2 for determining the 
 required area. In addition, it is necessary to provide sufficient volume to 
 assure that thickened sludge will be transported to the point of sludge with- 
 
 136 
 
April, 1970 SA 2 
 
 drawal without being mixed with more dilute solids and to allow for accumu- 
 lation of solids during any period when the rate at which solids enter the 
 thickener exceeds the rate of removal of thickened solids. 
 
 It is commonly held that the volume of sludge retained in the final settling 
 tank must be minimized to assure quick return of the sludge to avoid anaer- 
 obiosis and accompanying deterioration of sludge quality. While this seems 
 reasonable, few data are available to support the merit of the practice. In- 
 deed, Whurmann (16) reported that the properties of activated sludge were 
 not influenced by exposure to long periods of anaerobiosis, and workers at 
 the British Water Pollution Research Laboratory (15) have shown that re- 
 tention of activated sludge solids under anaerobic conditions for periods of 
 24 hr did not adversely influence the sludge. Indeed, in addition to the pos- 
 sible merits of retention of sludge in final settling tanks for thickening, it 
 has been proposed that holding sludge under oxygen deficient conditions can 
 be used advantageously to control growth of unwanted organisms (9). Pos- 
 sible adverse effects of long term retention of activated sludge solids are 
 flotation due to denitrification (10) and release of phosphates previously 
 removed by the activated sludge solids (12). 
 
 ILLUSTRATION 
 
 The role of thickening in the activated sludge process and concepts re- 
 garding the design of settling tanks to accomplish their thickening functions 
 can be best illustrated by example. The data used in this illustration are for 
 an actual activated sludge from a municipal contact stabilization plant with 
 an average organic loading of about 0.5 lb BOD per day per lb sludge solids. 
 Activated sludge thickening is a problem at the plant; often sludge can not be 
 withdrawn from the final tank at concentrations greater than 2,500 mg per 1 to 
 3,500 mg per 1. This sludge was purposely selected for use in the illustration 
 to emphasize that thickening concepts are involved even when solids con- 
 centrations are much less than 1% by weight. Fig. 3 shows the settling char- 
 acteristics of the sludge solids as established by laboratory settling tests. 
 Procedures for collection of these data are considered briefly in a later 
 section. 
 
 For purposes of illustration, presume that a 1-MGD plant has been de- 
 signed, and that the final settling tank was conventionally sized on the basis 
 of an overflow rate of 800 gpdpersq ft, as commonly might be done. Thus the 
 final tank would have a surface area of 1,250 sq ft and be 40 ft in diam. Pre- 
 sume further that the waste entering the activated sludge process has a BOD 
 of 200 mg per 1, that the biological portion of the design was based on an or- 
 ganic loading intensity of 0.5 lb BOD per day per lb MLSS, and that it was 
 conservatively estimated that 2,000 mg per 1 of mixed liquor suspended solids 
 would be maintained. The aeration tank volume hence would have to be 0.2 mg. 
 
 The assumed final settling tank design is typical of conventional practice 
 with attention being given to only the clarification function of the basin. Con- 
 sider now the ability of the basin designed in this fashion to carry out its 
 thickening function. From inspection of Fig. 3, it is tentatively concluded that, 
 for this sludge, 6,000 mg per 1 is a reasonable goal for the suspended solids 
 concentration of the return sludge. Using this value for c u , the required re- 
 cycle rate, R, from Eq. 1 is 50%, or 0.5 MGD, if 2,000 mg per 1 MLSS are to 
 
 137 
 
SA 2 
 
 ACTIVATED SLUDGE 
 
 be maintained. The total solids loading on the final tank produced by the flow 
 of 1.5 MGD at 2,000 mg per 1 is, from Eq. 3, 25,000 lb per day or 20 lb per sq 
 ft per day. Ignoring the waste sludge flow, W, the average bulk downward 
 
 O 2000 4000 6000 
 
 SUSPENDED SOLIDS CONC.mg// 
 
 4000 6000 
 SUSPENDED SOLIDS CONC.mg// 
 
 FIG. 3.-SETTLING CHARACTERISTICS 
 OF ACTIVATED SLUDGE CONSIDERED 
 IN ILLUSTRATIVE PROBLEM 
 
 FIG. 4.-SOLIDS TRANSPORT DUE TO 
 SEDIMENTATION 
 
 - FLUX FROM 
 SLUDGE 
 WITHDRAWAL 
 
 2000 4000 6000 
 SUSPENDED SOLIDS CONC.,mg/£ 
 
 2000 4000 6000 
 
 SUSPENDED SOLIDS C0NC.,mg/j2 
 
 FIG. 5.— SOLIDS TRANSPORT DUE TO 
 SLUDGE WITHDRAWAL WITH 50% 
 RECYCLE 
 
 FIG. 6. -TOTAL POSSIBLE SOLIDS 
 TRANSPORT WITH 50% RECYCLE 
 
 velocity, u, produced by removal of 0.5 MGD from the bottom of the tank, 
 from Eq. 4, is 0.037 ft per min. 
 
 Will all of the solids be able to reach the bottom of the tank? The possible 
 solids transport for each concentration of the activated sludge due to gravity 
 
 138 
 
April, 1970 SA 2 
 
 subsidence alone is shown in Fig. 4. The curve represents the product of the 
 settling velocity, r,-, from Fig. 3 and the associated concentration, Cj. The 
 additional solids flux, c^u, produced by the 0.037 ft per min bulk downward 
 velocity is shown in Fig. 5. Fig. 6 shows the total possible solids flux for 
 various suspended solids concentrations with the particular activated sludge 
 and operating conditions described. The upper curve in Fig. 6 represents the 
 sum of the curves in Figs. 4 and 5 as given by Eq. 2. It is seen that the solids 
 handling capacity of this activated sludge under these operating conditions 
 passes through a minimum limiting value of about 13 lb per sq ft per day at 
 a suspended solids concentration of about 2,500 mg per 1. The lower solids 
 handling capacities on the rising portion of the curve at concentrations below 
 about 400 mg per 1 are not of concern unless all of the solids are diluted 
 with clarified effluent to a concentration below this value at the inlet of the 
 tank. The shape of the curve in Fig. 6 is typical for activated sludge and other 
 suspensions. The magnitude of the limiting solids flux and of the concen- 
 tration corresponding to the limiting value change depending on the nature of 
 the suspension and the operating conditions of the settling basin. 
 
 It is apparent that this activated sludge process is not going to perform as 
 planned. To sustain the 2,000 mg per 1 MLSS concentration with a recycle 
 rate of 50%, 20 lb per sq ft per day of activated sludge solids must be trans- 
 mitted to the bottom of the settling basin. However, it is not physically pos- 
 sible for more than 13 lb per sq ft per day of the solids to pass through a 
 layer with a concentration of 2,500 mg per 1. Because in passing from the 
 MLSS concentration to the underflow concentration solids must pass through a 
 concentration of 2,500 mg per 1, this concentration becomes a bottleneck 
 limiting the performance of the thickening portion of the final settling tank, 
 and no more than 13 lb per sq ft of tank area will reach the bottom. 
 
 If solids are applied to a rate greater than 13 lb per sq ft per day they 
 will accumulate above the 2,500 mg per 1 layer. If the solids loading is sus- 
 tained at a level greater than 13 lb per sq ft per day for a long period of time, 
 the excess solids will propogate upward and be lost over the effluent weir. 
 The total mass of solids, M, which can pass through the 1,250 sq ft tank is 
 16,250 lb per day, and from Eq. 3, the underflow concentration with the 0.5 
 MGD recycle rate will be 3900 mg perl, not 6,000 mg per 1 as planned. From 
 Eq. 1 it is seen that the MLSS concentration can be maintained at only 1,300 
 mg per 1 and, hence, the organic loading intensity for the process will be 
 increased to 0.77 lb BOD per day per lb sludge solids. This increase of more 
 than 50% in the intended loading could lead to further deterioration of the 
 settling characteristics of the sludge solids. 
 
 While alternate operating conditions are possible, none can produce the 
 desired operating condition with 2,000 mg per 1MLSS and a 50% recycle rate. 
 For example, the solids flux due to bulk downward transport could be in- 
 creased by increasing the recycle rate. Fig. 7 shows the solids handling 
 capacity of the activated sludge with a 70% recycle rate with a resulting 
 underflow concentration of 3,500 mg per 1 (Eq. 3). It is seen that the limiting 
 solids handling capacity is increased to 16.5 lb per sq ft per day because of 
 the greater bulk downward velocity; however, the MLSS concentration will be 
 only 1,450 mg per 1, the organic loading intensity will still be greater than 
 planned (0.69 lb BOD per day per lb solids), and the actual retention time in 
 the aeration tank will be reduced appreciably. The possible change in the 
 settling characteristics of the sludge solids brought about by different operating 
 
 139 
 
SA 2 
 
 ACTIVATED SLUDGE 
 
 conditions has not been considered in estimating the alternate mode of 
 operation. 
 
 For this plant, requirements for the thickening function should have 
 governed sizing of the final clarifier. The tank should have been sized so that 
 the 25,000 lb per day of solids could have been transmitted to the bottom. 
 The example illustrates that when final clarifiers for an activated sludge 
 process are designed without consideration for their thickening function, 
 conditions may be created which make proper performance of the process 
 impossible. Symptoms of improper design for the thickening function are the 
 occasional excessive loss of solids over the final settling tank weir and the 
 inability to withdraw sludge from the final settling tank in high enough con- 
 centration to maintain the desired mixed liquor suspended solids concentration 
 in the aeration tank. It will be recognized that the symptoms are common in 
 the operation of the activated sludge process. However, they are not commonly 
 thought of as being caused by the limited capacity of the activated sludge 
 suspension for transmitting solids to the bottom of the final clarifier. 
 
 ALTERNATE METHODS OF ANALYSIS 
 
 While the flux plots for a continuous thickener (Figs. 6 and 7) were useful 
 in the preceding analysis of a final settling tank because they showed the rel- 
 
 23 
 
 24- 
 
 TOTAL FLUX 
 
 2000 4000 6000 
 
 SUSPENDED SOLIDS CONC.mg/2 
 
 FIG. 7. -TOTAL POSSIBLE SOLIDS TRANSPORT WITH 70% RECYCLE 
 
 ative contributions of gravity subsidence and bulk transport, more conve- 
 nient means of analysis are available. One method involves use of the approach 
 described over 50 yr ago by Coe and Clevenger (2), and the other is use of 
 batch flux plots as described by Yoshioka,et al. (17) and others. These meth- 
 ods are described in the paragraphs which follow. 
 
 The solids loading on a final settling tank, M, is equivalent to the actual 
 
 hO 
 
April, 1970 
 
 SA 2 
 
 solids flux times the area of the basin or GA. Hence, from Eq. 4 
 
 G 
 u = — 
 
 (6) 
 
 Substituting this value for u in the basic equation for solids flux in a settling 
 basin (Eq. 2) gives 
 
 G = 
 
 (7) 
 
 Eq. 7 is the same as the expression developed in a different manner by Coe 
 and Clevenger. The required size of a thickener can be determined from Eq. 
 7 by substituting various values of c i and v i to determine the limiting solids 
 handling capacity for alternative values of the underflow concentration, c u . 
 Fig. 8 shows plots of Eq. 7 for the data used in the illustrative problem when 
 
 FIG. 8. -USE OF COE AND CLE- 
 VENGER METHOD (2) FOR IDEN- 
 TIFYING LIMITING SOLIDS HAN- 
 DLING CAPACITY 
 
 FIG. 9. -BATCH FLUX PLOT 
 METHOD OF ANALYZING THICK- 
 ENING PERFORMANCE OF FINAL 
 SETTLING TANK 
 
 return sludge is removed at 6,000 mg per 1 and 3,900 mg per 1. It is seen that 
 if sludge were removed at 6,000 mg per 1, capacity of the tank would be lim- 
 ited to 7 lb per sq ft per day. By adjusting the underflow rate to 50% recycle 
 to give a thickened sludge concentration of 3,900 mg per 1, the capacity is in- 
 creased to 13 lb per sq ft per day. This value agrees with the result obtained 
 by use of the continuous thickener flux plot in the preceding illustrative 
 problem. 
 
 Because G in Eq. 6 was taken as the actual flux through the settling basin, 
 Eq. 7 is valid only for the concentrations of sludge which exist in the settling 
 tank. That is, the plot of G for c u = 3,900 mg per 1 in Fig. 8 does not coincide 
 with the upper curve in Fig. 6. Eq. 7 and the curves in Fig. 8 are valid at the 
 concentration which limits the solids handling capacity of the thickener and 
 hence can be used for sizing the area in a final settling tank required for 
 thickening. The equation is not valid for other concentrations, however.. This 
 
 \k\ 
 
SA 2 
 
 ACTIVATED SLUDGE 
 
 shortcoming of the Coe and Clevenger approach limits its usefulness in ana- 
 lyzing the performance of an existing thickener. 
 
 Another means of analysis of design or operation of a thickener is much 
 more convenient. This method involves use of a batch flux curve such as 
 shown in Fig. 4. From Eq. 6, the slope of a line connecting a flux, G, on the 
 ordinate axis of a batch flux plot with the corresponding underflow concen- 
 tration, c u , is the bulk downward velocity, u. Hence, as illustrated by Fig. 9 
 in conjunction with Eq. 2, the design or operation of a final settling tank can 
 be evaluated directly from a batch flux plot. The solids flux represented by 
 the distance from the origin to G B , c L v L , is the solids transport due to sub- 
 sidence. The additional flux, c L u, due to the bulk downward transport of the 
 
 '0 2000 4000 6000 
 
 SUSPENDED SOLIDS CONC.mg/fc 
 
 FIG. 10. -USE OF BATCH FLUX PLOT TO READILY DETERMINE ALTERNATE OP- 
 ERATING CONDITIONS CONSIDERED IN ILLUSTRATIVE PROBLEM 
 
 final settling tank contents when sludge is removed at concentration c u is 
 represented by the distance between G L and G B in Fig. 9. 
 
 The convenience of the batch flux curve method is illustrated in Fig. 10 
 where the three alternate operating conditions previously calculated are ob- 
 tained simply by striking three tangents, operating lines, to the flux curve. 
 As before, it is seen that when thickened activated sludge is withdrawn at 
 3,900 mg per 1, capacity of the thickener is limited to 13 lb per sq ft per day 
 by the ability of the sludge at 2,500 mg per 1 to transmit solids under the 
 operating conditions. Similarly, the limiting capacity when sludge is with- 
 drawn at 3,500 mg per 1 is seen to be 16.5 lb per sq ft per day in agreement 
 
 \kl 
 
April, 1970 SA 2 
 
 with Fig. 7, and 7 lb per sq ft per day when sludge is withdrawn at 6,000 mg 
 per 1 as previously shown in Fig. 8. 
 
 The three methods for analyzing the thickening function of final settling 
 tanks, the component method illustrated by Fig. 6, the Coe and Clevenger 
 method shown in Fig. 8, and_ the batch flux plot method demonstrated in Fig. 
 10, do not differ fundamentally, and all give the same results. They have been 
 presented in order of increasing convenience for routine use and in order of 
 increasing complexity for acquiring insight into the concepts involved in de- 
 sign of the thickening function of settling basins. These methods, especially 
 the batch flux plot technique, are equally useful in analysis of performance of 
 an existing final tank. The effect of operational changes can be anticipated by 
 drawing various operating lines on a curve such as Fig. 10. 
 
 COLLECTION AND ANALYSIS OF SETTLING DATA 
 
 The data needed for analysis of design or operation of final settling tanks 
 are those describing the relationship between suspended solids concentration, 
 c z -, and settling velocity, w z -. From these data a reasonable return sludge sus- 
 pended solids concentration and the limiting solids handling capacity can be 
 estimated. The required data are obtained by multiple batch settling tests, 
 using various initial concentrations of the activated sludge. The slope of the 
 straight line portion of the batch settling curve is taken as the settling ve- 
 locity of the sludge at its initial concentration. 
 
 While it is not the purpose of this paper to consider techniques for gath- 
 ering meaningful sedimentation data, it would be remiss not to mention com- 
 mon pitfalls in collecting the data. The turbulence level created in dispersing 
 the solids at the beginning of the settling test can influence results because 
 nonhomogeneous conditions can be created while the turbulence subsides (3). 
 Observed settling velocities are influenced by the initial depth of sludge in the 
 batch settling column (3). Settling tests must be conducted with sludge depths 
 representative of the effective depth which is expected to occur in the final 
 tank. The presence of the wall in laboratory settling columns can have a pro- 
 found effect on settling characteristics. With a given sludge, the wall effect 
 can be either to retard or to hasten subsidence, depending on the solids con- 
 centration (14). Slow mixing is extremely important in laboratory settling 
 tests. Indications are that the role of mixing is to minimize the wall effect 
 and not primarily, as has been presumed, to duplicate the rake action in full 
 scale settling basins (13). 
 
 The state-of-the-art is not such that precise conditions for laboratory 
 settling tests which will give results representative of field performance can 
 be specified. The data used in the previous example were obtained from lab- 
 oratory settling columns 3.5 in. in diam with initial sludge depths of 3.5 ft 
 with a stirring rate of 1 rpm. Initial distribution of the solids in the settling 
 column was obtained by pumping the sludge into the column at a controlled 
 rate. 
 
 SUMMARY 
 
 Conventional procedures for design of final settling tanks in the activated 
 sludge process consider only the clarification function of the tanks. Failure 
 
 H3 
 
SA 2 ACTIVATED SLUDGE 
 
 to consider the thickening function of final settling tanks can lead to unneces- 
 sarily high capital costs, operational difficulties, and impairment to effluent 
 quality. 
 
 Design for the thickening function of final settling tanks is based on the 
 limiting solids handling capacity of the sludge. If a final settling tank is de- 
 signed without regard for the limited capacity of the suspension for trans- 
 mitting solids, a severe restriction may be imposed on the rate at which 
 suspended solids can be passed through the settling basin. The result is a 
 reduction in the amount of suspended solids which can be maintained in the 
 aeration tank, and possibly a loss of suspended solids over the effluent weir. 
 Because the solids handling capacity of a suspension is determined both by 
 the settling characteristics of the sludge and by the way the final settling 
 tank is operated, the capacity of an existing tank for transmitting solids can 
 be increased by increasing the rate of sludge removal. However, this opera- 
 tional change reduces the suspended solids concentration in the sludge with- 
 drawn from the tank. 
 
 Three techniques for analyzing the thickening performance of a final 
 settling tank for design or operational purposes have been illustrated. The 
 batch solids flux plot is the most convenient of the methods. 
 
 ACKNOWLEDGMENTS 
 
 This work was supported in part by Research Grant WP 01011 from the 
 Federal Water Pollution Control Administration. The settling data used in the 
 illustration were obtained by A. R. Javaheri. 
 
 APPENDIX I. -REFERENCES 
 
 1. Behn, V. C, and Liebman, J. C, "Analysis of Thickener Operation," Journal oj the Sanitary 
 Engineering Division, ASCE, Vol. 89, No. SA3, Proc. Paper 3535, June, 1963, pp. 1-15. 
 
 2. Coe, H. S., and Clevenger, G. H., "Methods for Determining the Capacities of Slime Settling 
 Tanks," Transactions. American Institute of Mining Engineers, Vol. 55, 19 16, pp. 356-384. 
 
 3. Dick, R. I., and Ewing, B. B„ "Evaluation of Activated Sludge Thickening Theories," Journal of 
 the Sanitary Engineering Division, ASCE, Vol. 93, No. SA4, Proc. Paper 5367, August, 1967 
 pp. 9-29. 
 
 4. Dick, R. I., discussion of "Settling of Activated Sludge in Horizontal Tanks," by C. N. H. Fish 
 erstrom, E. Isgard, and I. Larsen, Journal oj the Sanitary Engineering Division, ASCE, Vol. 93 
 No. SA6, Proc. Paper 5623, December, 1967, pp. 271-273. 
 
 5. Eckenfelder, W. W., Jr., and Melbinger, N., "Settling and Compaction Characteristics of 
 logical Sludges," Journal Water Pollution Control Federation, Vol. 29, No. 10, Oct., 1957, pp 
 1114-1122. 
 
 6. Edde, H. J., and Eckenfelder, W. W., Jr., "Theoretical Concepts of Gravity Sludge Thickening 
 Scaling-Up Laboratory Units to Prototype Design," Journal Water Pollution Control Federa- 
 tion. Vol.40, No. 8, Aug., 1968, pp. 1486-1498. 
 
 7. Ford, D. L., and Eckenfelder, W. W., Jr., "Effect of Process Variables on Sludge Floe Form 
 tion and Settling Characteristics," Journal Water Pollution Control Federation, Vol. 39, No. 1 
 Nov., 1967, pp. 1850-1859. 
 
 8. McCarty, P. L., and Broderson, C. F., "Theory of Extended Aeration Activated Sludge," Jour 
 
 \kk 
 
April, 1970 SA 2 
 
 nal Water Pollution Control Federation, Vol. 34, No. 1 1, Nov., 1962, pp. 1095 1 103. 
 9. Okun, D. A., discussion of "Effect of Oxygen Tension in Biochemical Reactions in Sewage Puri- 
 fication Plants," by K. Wuhrmann, Advances in Biological Waste Treatment, W. W. Eckenfeld- 
 er, Jr., and J. McCabe, eds„ Pergamon Press, New York, 1963, pp. 38-39. 
 
 10. Sawyer, C. N., and Bradney, L., "Rising of Activated Sludge in Final Settling Tanks," Sewage 
 Works Journal. Vol. 17, No. 6, Nov., 1945, pp. 1191-1209. 
 
 1 1 . Shannon, P. T., and Tory, E. ML, "The Analysis of Continuous Thickening," Society for Mining 
 Engineers of AIME Transactions. Vol. 235, No. 12, Dec, 1966, pp. 375 382. 
 
 12. Shapiro, J., Levin, G. V., and Humberto, Z. G., "Anoxically Induced Release of Phosphate in 
 Waste Water Treatment," Journal Water Pollution Control Federation. Vol. 39, No. 1 1, Nov., 
 1967, pp. 1810-1818. 
 
 13. Vesilind, P. A., "The Influence of Stirring in the Thickening of Biological Sludge," thesis pre- 
 sented to the University of North Carolina at Chapel Hill, North Carolina, in 1968 in partial ful- 
 fillment of the requirements for the degree of Doctor of Philosophy. 
 
 14. Vesilind, P. A., discussion of "Evaluation of Activated Sludge Thickening Theories," by R. I. 
 Dick and B. B. Ewing, Journal of the Sanitary Engineering Division. ASCE, Vol. 94, No. SA I, 
 Proc. Paper 5773, February, 1968, pp. 185-190. 
 
 15. Water Pollution Research Board, Water Pollution Research 1962. Department of Scientific and 
 Industrial Research, London, England, 1962. 
 
 16. Wuhrmann, K., "Effect of Oxygen Tension in Biochemical Reactions in Sewage Purification 
 Plants," Advances in Biological Waste Treatment. W. W. Eckenfelder, Jr., and J. McCabe, eds., 
 Pergamon Press, New York, 1963, pp. 27-38. 
 
 17. Yoshioka, N., et al., "Continuous Thickening of Homogeneous Flocculated Slurries, Chemical 
 Engineering. Tokyo, Vol. 21, No. 2, 1957, pp. 66-74. 
 
 APPENDIX II. -NOTATION 
 
 The following symbols are used in this paper: 
 
 A = area of settling basin, L 2 ; 
 
 c = concentration of suspended solids, M/L 3 ; 
 
 c i = suspended solids concentration in layer i, M/L 3 ; 
 
 c L = suspended solids concentration of layer with minimum capacity for 
 transmitting solids, M/L 3 ; 
 
 c = suspended solids concentration of flow entering final settling tank, M/L 3 ; 
 
 c u = underflow suspended solids concentration from final settling tank, M/L 3 ; 
 
 G = solids flux, M/L 2 /T; 
 G L = limiting solids handling capacity, M/L 2 /T; 
 
 M = total suspended solids loading on final settling tank, M/T; 
 
 Q = rate of flow of waste entering activated sludge process, L 3 /T; 
 Q = rate of flow of waste entering final settling tank, L 3 /T; 
 Q u = rate of sludge removal from settling basin, L 3 /T; 
 
 R = rate of recycle of activated sludge, L 3 /T; 
 
 u = average bulk downward velocity in settling tank caused by sludge re- 
 moval, L/T; 
 
 i>j = settling velocity of sludge with suspended solids concentration Cj, L/T; 
 v L = settling velocity of layer with limiting solids handling capacity, L/T; and 
 
 W = rate of flow of waste activated sludge, L 3 /T. 
 
 1*5