il iii

TTU

O 1151 O1
| and
Operation of :
Activated Sludge
Processes Using
Respirometry

Alan F.Rozich
Anthony F. Gaudy, Jr. —
 Digitized by the Internet Archive
in 2024

https://archive.org/details/designoperationoO00Orozi
 DATE DUE

s
S
Pa
= S|

if
HIGHSMITH = # 45220
Design and

Operation of
Activated Sludge
Processes Using
Respirometry

Alan F. Rozich
Anthony F. Gaudy, Jr.

YY Lewis puBLisHERS
Tennessee Tech Libraiy
Santavilla TN
Library of Congress Cataloging-in-Publication Data

Rozich, Alan F.
Design and operation of activated sludge processes using
respirometry / Alan F. Rozich and Anthony F. Gaudy, Jr.
DD.» -CM.
Includes bibliographical references (p. ) and index
1. Sewage — Purification — Activated sludge
process — Mathematical models. 2. Microbial
respiration — Mathematical models. I. Gaudy, Anthony F.
TD756.R69 1992
628.3'54—dc20 91-34375
ISBN 0-87371-449-0

COPYRIGHT © 1992 by LEWIS PUBLISHERS, INC.

ALL RIGHTS RESERVED

This book represents information obtained from authentic and highly

regarded sources. Reprinted material is quoted with permission, and
sources are indicated. A wide variety of references are listed. Every
reasonable effort has been made to give reliable data and information,
but the authors and the publisher cannot assume responsibility for the
validity of all materials or for the consequences of their use.

Neither this book nor any part may be reproduced or transmitted in

any form or by any means, electronic or mechanical, including
photocopying, microfilming, and recording, or by any information
storage and retrieval system, without permission in writing from the
publisher.

LEWIS PUBLISHERS, INC.

121 South Main Street, Chelsea, Michigan 48118

Printed in the United States of America 1 234567890

 Dedication:
To Gemma, Anton, Adrienne, and Libby
 repos emOy male !

Riera ts
-2 Sis ry aaa

tly Sa o3'a2@a sep rad ae ¥

hs Ps, rng
ea a © sh Ce f
ee 4 he ae
NS % an G2 o%
Ver ( 1 thes gw

' 4 st Ve ou ale Pir 2050 eile 4: aannar oe 3

wy tues Rare mia —4 ) rey 7
7 re whl. gas Sori theg, crt @de Cipher
ciryy vi 7 _ wit
.
seems 14 WS PS fie

‘ad ae Ave . S1ORgghr 4h) 16 7 -_

mes otal miriierae?t ‘4

=
Alan F. Rozich received a BSCE with an emphasis in environmental
engineering from Ohio State University in 1976 and an MS in environ-
mental engineering from Ohio State in 1978. He worked for three years
as a wastewater engineer for the city of Columbus, Ohio. Dr. Rozich
received his Ph D in 1982 from the University of Delaware in environ-
mental engineering. He worked on developing predictive models for acti-
vated sludge systems treating inhibitory or toxic wastes. Since 1983, Dr.
Rozich has worked in various research, development, and consulting
capacities. His efforts focused on developing improved techniques for
designing and operating biological wastewater treatment systems with an
emphasis on toxic, hazardous, and difficult-to-degrade wastes. Much of
this work was directed toward utilizing respirometric methods as a means
to calibrate predictive models for biological treatment systems. This pre-
dictive modeling technique has been applied to relatively diverse biologi-
cal treatment situations as exemplified by a project involving the formu-
lation of modeling strategies for predicting bioremediation rates at
Prince William Sound, Alaska and one involving the formulation of a
modeling approach for predicting landfill gas production rates. Dr.
Rozich has presented and published over 50 technical papers and co-
authored a book on biological waste treatment technology. He is a
licensed professional engineer and also the holder of a process patent in
waste treatment technology.
Anthony F. Gaudy, Jr. received a BSCE from the University of Massa-
chusetts, his MSSE from the Massachusetts Institute of Technology, and
a PhD from the University of Illinois, Urbana. Dr. Gaudy’s career in the
water pollution control field spans nearly 40 years and involves research,
teaching, and consulting to industry and government. He has specialized
in the biological aspects of pollution control because, early on, it
appeared to him that this area comprised the most important and funda-
mental route to preserving the life support system. Staying on this track
has enabled him and his 80 MS and PhD students to engage in what has
amounted to a continuing comprehensive program of investigation into
many aspects of the carbon-oxygen cycle and ways to achieve practical
engineering control of this cycle. The work has sought to stress elucida-
tion of fundamentals and, of more practical use, engineering methodol-
ogy for optimal utilization of these fundamentals. Dr. Gaudy has
authored more than 200 publications, and is H. Rodney Sharp Professor
Emeritus, University of Delaware Department of Civil Engineering. This
text brings together several of his long-term interests in modeling the
microbial growth and respiration processes and their embodiment in a
practical model to predict performance and provide operational control
of the activated sludge process.
 TABLE OF CONTENTS

Preface xi

List of Symbols xvi

1. Fundamentals of Biokinetics for Activated Sludge Systems 1

Introduction 1
Some General Principles 2
Some Important Quantitative Concepts 5
Cell or Sludge Yield 5
Growth Rate and Decay Rate 6
Relationship Between p» and S
(Noninhibitory Waste) 8
Relationship Between » and S
(Inhibitory Waste) 11
Nature of the Growth Curve 15
Relations Between » andS_ 16
Relation Between S, X, and O, 19
Key Concept Summary 21
References and Suggested Additional Reading 23

2. Basic Principles of Bioreactor Modeling 25

Introduction 25
Once-Through System (Chemostat), Noninhibitory
Model 29
Once-Through System (Chemostat), Inhibitory
Model 31
Prediction of Excess Biomass (Excess Sludge) 32

Vii
 Review 33
Effect of Biomass Recycle 33
Key Concept Summary 38
References and Suggested Additional Reading 38

3. Engineering Models for Activated Sludge Systems 39

Introduction 39
Derivation of Predictive Equations 40
Noninhibitory Wastes 40
Inhibitory Wastes 43
Application for Multiple Reactor Systems 47
Critical Point Analysis for Treatment of Inhibitory
Wastes 51
Critical Point Curves 54
Key Concept Summary 58
References and Suggested Additional Reading 59

_ 4. Comparison with Other Approaches and Basic

Applications 61
Introduction 61
Reconciliation with Other Design and Operational
Approaches 62
Impact of Recycle Parameters on System Growth
Rate 66
Determination of Conditions Needed for Minimizing
Sludge Production 67
Computation of Aeration Tank Oxygen Transfer
Requirements 69
Key Concept Summary 74
References and Suggested Additional Reading 75

5. Procedures for Obtaining Biokinetic Constants 77

Introduction 77
Determination of Influent Waste Strength, S; 78
Use of Respirometry to Generate Kinetic Data 82
Conversion of Respirometric Data to Growth or
Substrate Utilization Data 83

Viii
 Example of Obtaining Growth Data from
Respirometric Data 84
Determination of the Biokinetic Constants from
Growth Data 90
Determination of True Cell Yield, Y,, and Decay
Rate, k, 93
Key Concept Summary 95
References and Suggested Additional Reading 95

6. Factors Affecting the Values of the Biokinetic Constants 97

Introduction 97
Specific Growth Rate of the Reactor 98
Waste Composition 105
Toxics 111
Temperature 116
Population Diversity 119
Shock Loads 121
Bioaugmentation 126
Key Concept Summary 127
References and Suggested Additional Reading 128

7. Case Studies and Applications 131

Introduction 131
Case 1: Patapsco Wastewater Treatment Plant,
Baltimore, Maryland 133
Background 133
Model Analysis and Review of Existing Data 134
Determination of Biokinetic Constants 142
Field Evaluation of Model Predictions 150
Operational and Design Recommendations 154
Case 2: Treatment of Impoundment Leachate at a
Superfund Site 156
Background 156
Approach 157
Case 3: Impact of Cobalt on Biological Wastewater
Treatment Plant Performance 161
Background 161
Approach 162
 Case 4: Use of Respirometry for Screening and Ranking
Applications 168
Background 168
Approach 169
Results 170
Key Concept Summary 175
References and Suggested Additional Reading 175

Appendix: Computer Programs 177

Index 183
 PREFACE

Why write a book about respirometry and activated sludge?

This is an obvious question which the reader may ask. Indeed, why
write this book. Many of the reasons have to do with the authors’ experi-
ence concerning the research, development, and application of process
control models for biological treatment. We realized that there was no
one course which integrated the use of respirometry and the application
of process control models for analyzing aerobic biological treatment
systems. Our experience is that the use of process control models which
are calibrated using respirometry represents a rapid, accurate, and cost-
effective technique for generating process information for designing and
operating aerobic biological treatment systems.
Our experience with field applications indicated that process models
which are presented in this book do a reasonably good job of predicting
the behavior of full-scale biological treatment systems. The authors also
realized through feedback from consulting clients that the level of effort
which was required to calibrate the model was a major issue which
prevented its routine use and application. Calibration of the model pri-
marily consists of determining the relationship between biomass growth
rate and substrate or waste concentration. Methods such as the shake
flask technique or the substrate utilization method were adequate for
laboratory needs but unwieldly and resource-intensive for routine appli-
cation. This led us to undertake a major research and development effort
to develop a more cost-effective technique for calibrating the model.
Respirometry or the measurement of oxygen uptake rates of biomass
has long been employed for many applications in the pollution control
field. The basic premise relates to the simple fact that oxygen uptake
rates provide a rapid indicator of microbial activity. In order to use
respirometry for calibrating the model, one must be able to utilize the
oxygen uptake data to compute cell growth or substrate utilization rates.
A major effort which we performed showed that respirometric data can
be employed to determine cell growth rates through transform of the
data into equivalent growth or substrate utilization data. This is accom-

xi
plished by utilizing equations which are derived using the principle of
COD and energy balances for aerobic systems. Several research and field
application efforts validated the use of respirometry as a means to obtain
values of the biokinetic constants for model calibration. The model pre-
dictions for field reactors made using respirometric calibration accu-
rately described the performance of field units. The key concept is that
respirometry is used in these applications to calibrate a model which
predicts the behavior of aerobic biological treatment systems.
The model which is presented in this book has its origins with work
which was presented over 40 years ago. In the early 1950s, Monod and
Novick and Szilard developed the theory of continuous culture. The
significance of this work was to postulate that an engineering control
such as reactor flow rate can be utilized to select the growth rate of a
microbial culture. If the relationship between cell growth rate and sub-
strate concentration is known, then one can obtain a desired effluent
substrate concentration by using the appropriate engineering control
such as flow rate. This concept is known as the “Theory of Continuous
Culture.”
During the 1960s, one of us (AFG) performed extensive work to show
that the theory could be applied for predicting the results of heteroge-
neous microbial systems like those employed in activated sludge treat-
ment processes. This work was necessary because the initial efforts of
Monod, and others, focused on pure culture systems. Further develop-
mental efforts started in the early 1980s extended the predictive modeling
work for substrates which are inhibitory to microbial growth. Inhibitory
kinetics are characteristic of many systems which must handle toxic or
hazardous wastes or materials. In the mid-1980s, we began looking for
kinetic techniques which could calibrate the model (define the relation-
ship between growth rate and substrate concentration) quickly and cost-
effectively. This led to the respirometric methods and the development
and application effort for the technology which is described in this book.
We now have this technology incorporated into a practice which we use
routinely for engineering projects involving biological treatment
systems.

How ts this book set up?

The book contains seven chapters. The first two chapters provide a
review of the fundamentals of modeling aerobic biological treatment
systems. Chapter 3 gives the derivation of predictive equations for acti-

xii
vated sludge systems while Chapter 4 compares and reconciles the model-
ing approach presented in this book with other methods. The methodol-
ogy for using respirometry to obtain biokinetic constants is described in
Chapter 5. Chapter 6 is important because it reviews the various environ-
mental factors which can influence the values of the biokinetic constants
which in turn impact process performance. It needs to be emphasized
that the advent of respirometric techniques such as those described in the
book enable environmental professionals to predict the impact of envi-
ronmental conditions such as temperature, pH, and other factors on
process performance much more quickly and accurately than that which
one can realize with conventional approaches. Chapter 7 presents several
case histories which describe the use of the respirometric technology for
analyzing various design and operational situations.
Each chapter contains Introduction, Key Concept Summary, and Ref-
erences and Suggested Additional Reading sections. The use of an Intro-
duction section is self-explanatory. The Key Concept Summary sections
are designed to give the reader “bullet” summary versions of the impor-
tant technical concepts which are presented in a particular chapter. The
References and Suggested Additional Reading sections are given to pro-
vide those who are interested with more detail and background on the
technical information presented in the text. The authors wish to empha-
size that this book is not intended as a general textbook on biological
treatment processes. Its purpose is to communicate information concern-
ing a specific technical practice area which the authors developed and
utilize. Consequently, only a minimum number of references are given in
order to provide the necessary scientific and engineering back-up.

Who can benefit by using this book?

The main purpose of the book is to inform practitioners in the envi-

ronmental field, especially those involved with biological treatment pro-
cesses, about the utility of respirometrically calibrated models for ana-
lyzing biological systems. One interesting trend which we have observed
is that clients who understand the technology suggest new ways to apply
it for solving their problems or for enhancing operations. We liken this
technology to a personal computer (e.g., a “Mac”) in that once the fun-
damentals are understood, the number of applications is limited only by
the imagination of the user.
At ERM, we have utilized this technology on numerous projects.
Many times, we have been able to save clients both time and money. In

xiii
the environmental business, time is often money, especially when there is
need to fast-track a project in order to meet compliance deadlines. Pro-
jects have ranged from relatively basic treatability work to utilizing the
respirometric approach to design thermophilic aerobic biological systems
to treat a groundwater containing 200,000 mg/L COD.

What is the future of this technology?

As this book goes to publication, we are extending the boundaries and

applications of the technology presented herein. We have completed pro-
jects which took the concepts for the respirometric calibration of acti-
vated sludge models and applied them to other areas. For example, work
was performed that developed a modeling strategy for predicting biore-
mediation rates at oil contaminated beaches which resulted from the oil
spill at Prince William Sound. Other work modified the aerobic model-
ing technique for use in predicting anaerobic methane production rates in
landfills. Other applications which are in the works in our group include
the development of standard protocols for determining the biodegrada-
bility of plastics and plastics substitutes and modifying the techniques
presented herein for performing predictive modeling of composting (aer-
obic and anaerobic) systems.

Who else contributed to this work?

The authors wish to acknowledge some key individuals whose contri-

bution and collaborative efforts were crucial in developing this technol-
ogy. Mr. Richard J. Colvin has been working with us since 1983. He was
instrumental on research and development and consulting projects which
involved the development and application of respirometric methods. Dr.
Elizabeth T. Gaudy ably served as a technical critic, proofreader, and co-
worker. She also ensured that our small consulting operation ran
smoothly.
Messrs. Jerrold Wingeart and Paul H. (“Kip”) Keenan, both of the
City of Baltimore Public Works Department (Kip is now with Wes-
tinghouse), were instrumental in working with the authors to support the
development and implementation of the technology. Their involvement
made our efforts a professionally rewarding experience.
Finally, we wish to thank our technical reviewers for the time they took
to examine and critique the manuscript. Mr. Donald Loftus of Star

Xiv
Enterprise Refinery in Delaware City, Delaware and Dr. Paul J. Usino-
wicz of ERM, Inc., Exton, Pennsylvania made valuable suggestions for
this book. They have our gratitude.

Alan F. Rozich, Ph.D., P.E.

Anthony F. Gaudy, Jr., Ph.D., P.E.

XV
 LIST OF SYMBOLS

chemical oxygen demand (mg COD/L)

initial soluble COD for a batch test (mg
COD/L)
final soluble COD for a batch test (mg
COD/L)
soluble COD at time t (mg COD/L)
soluble COD measured for a given waste
sample (mg COD/L)
total measured COD of a given sample (mg
COD/L)
soluble readily metabolizable COD of waste
sample, COD,-COD, (mg COD/L)
soluble readily metabolizable COD of waste
sample over a given interval of time (mg
COD/L)
the change in the amount of COD
incorporated into the cells (mg COD/L)
coefficient of variation
dilution rate, influent flow rate, F, divided by
aeration tank volume, V (h"')
time in days
wastewater flow rate (mgd)
recycle flow rate (mgd)
time in hours
maintenance energy coefficient or specific
decay rate (h'')
inhibition constant for inhibitory wastes based
on ACOD (mg COD/L)
saturation constant, substrate concentration,
based on ACOD, at which specific growth
rate is one-half the maximum rate (mg/L)
number of samples
cumulative O, uptake concentration at time t
(mg O,/L)
COD per mg of biomass
average COD contained in a mg of biomass
for a particular batch study
primary settling tank
return activated sludge
readily metabolizable COD measured at time t
(mg COD/L)
average readily metabolizable COD
concentration over a given time interval (mg
COD/L)
readily metabolizable effluent COD leaving
plant (mg COD/L)
readily metabolizable final COD measured in
a batch system (mg COD/L)
readily metabolizable influent COD entering
plant (mg COD/L)
readily metabolizable initial COD measured in
a batch system (mg COD/L)
readily metabolizable recycle COD measured
within the plant (mg COD/L)
standard deviation
time (h or d)
hydraulic detention time (h)
reactor volume (million gallons)
largest concentration of biological solids
measured during the substrate removal phase
of a batch test (mg TSS/L)
concentration of biological solids (mg TSS/L
or mg VSS/L)
initial concentration of biological solids in a
batch test (mg TSS/L)
concentration of biological solids in the
recycle flow to the reactor (mg TSS/L)
concentration of biological solids at time t
(mg TSS/L)
average concentration of biological solids over
a given time interval (mg TSS/L)
change in the concentration of biological
solids over a given time interval (mg TSS/L)
cell or sludge yield; mg biomass produced per
mg COD metabolized (mg/mg). In a batch
system the observed yield is the true yield. In
a recycle system, the true sludge yield is
calculated from a maintenance plot.
recycle flow ratio (F,/F)
specific growth rate (h"')
maximum specific growth rate (h"')
 1 FUNDAMENTALS OF
BIOKINETICS FOR
ACTIVATED SLUDGE
SYSTEMS

INTRODUCTION

The principles of the activated sludge process are really not very diffi-
cult to understand. In fact, approximately four decades ago it was
thought by some that we already knew all we were ever going to know
about the process and that further study was pointless. After all, the
saying went, “the bugs eat the waste”; what else did we really need to
know? Today, however, some people take the reverse attitude. They feel
that the activated sludge process has failed to serve the engineering pro-
fession because it is too complex to be understood clearly enough, and
thus controlled precisely enough, to deliver the type of effluent quality
demanded by today’s environment.
Both the oversimplified attitude and the overcomplex attitude give
erroneous impressions about the process; the truth lies somewhere
between the extreme views. One of the authors of this book has enjoyed a
career spanning four decades (mentioned above) in the attempt to
uncover and clarify some of the complexities to simplify engineering of
the process and promote its goal of wastewater purification. In recent
times, the activated sludge process has been increasingly called on to
treat not only readily metabolizable wastes but also those containing
toxicants. There is much evidence that toxic wastes can be successfully
treated provided one adheres to certain fundamental principles. It is the
aim of this text to set down such principles in a quantitative and practical
fashion and to array them in a useful engineering methodology readily

1
2 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

applicable not only to process design but to process operation as well. It

should be apparent that, to possess engineering value, design models
must also possess the analytical properties needed for guiding operations
as well as initial design of the process.
The goal of this chapter is to describe the purification process in terms
of microbial growth, which is the mechanistic basis for the biological
treatment of wastewaters. Following a brief description of the process,
the parameters needed to quantify growth, purification, and oxygen
uptake (i.e., exertion of biochemical oxygen demand, BOD) are given.
The necessary biokinetic parameters are established, and relationships
between growth, purification (i.e., substrate or waste removal), and res-
piration (i.e., O, uptake) are presented. After studying this chapter, the
reader should have a solid grasp of the quantitative concepts required for
understanding the development and calibration of models that describe
the microbial purification of wastewaters containing both toxic and non-
toxic carbon sources.

SOME GENERAL PRINCIPLES

Perhaps the appropriate scientific phraseology for the idea “the bugs
eat the wastes” is as follows: “The heterogeneous microbial population
utilizes waste materials to obtain energy and to grow.” They use the waste
material in the same manner in which we use food materials for energy
and for growth. In order for the microbes to accomplish this, the waste
molecules must be soluble or made soluble so that the process is not
hampered simply by hindered access of the food molecules to the micro-
bial cells. All the primary feeders (bacteria) in the population use soluble
food. The secondary feeders are larger microbes, e.g., protozoa, which
ingest particulate food, mostly bacteria that have grown on the waste
molecules. The primary and secondary feeders, along with whatever par-
ticulate matter is contained in the waste or passes into the aeration tank,
are collectively termed “activated sludge.”
From a chemical standpoint, this sludge consists mainly of several
classes of biochemical compounds that are characteristic of all living
matter (carbohydrates, proteins, lipids, and nucleic acids). Synthesized
by all living systems, these biochemical compounds contain mainly car-
bon, which the cells obtain from the compounds in the waste. Some of
these compounds also contain nitrogen, phosphorus, and many other
elements in lesser amounts. Naturally, all of the elements needed to
synthesize the biochemical compounds must be present in the waste-
 BIOKINETICS OF ACTIVATED SLUDGE 3

water. When the needed compounds are out of balance, those in short
supply must be added. For example, nitrogen and phosphorus are often
necessary additions to some industrial wastewaters, which usually con-
tain such a large supply of carbonaceous material that growth could not
occur in a balanced manner without supplementation of nitrogen and
phosphorus. The aim in the treatment process is to ensure that carbon is
the growth-limiting nutrient and that growth is balanced.
Usually, the chemical composition of the sludge is not of major inter-
est in the activated sludge process, and microbial (sludge) growth is
usually assessed as an increase in the mass of sludge in the system regard-
less of its composition. The reader should realize that balanced growth is
not necessary for growth to occur (that is, growth as measured as an
increase in the mass of material). The mechanism of oxidative assimila-
tion, an important mechanism that sometimes occurs in treatment sys-
tems, is an example of unbalanced growth in which large amounts of
nonnitrogenous carbon compounds such as carbohydrates and lipids
may be synthesized with little or no production of protein or nucleic
acids. Such unbalanced growth can at times provide for the removal of
large amounts of substrate. However, it should be remembered that in
order to maintain the substrate removal capability of an activated sludge,
the biomass that is eventually recycled to the aeration tank should have
the opportunity to synthesize nucleic acid and protein.
Oxygen is supplied to an activated sludge reactor in order to provide
the microorganisms with the means to oxidize a portion of the organic
compounds in the waste material. A portion of the energy released dur-
ing the oxidation process is converted into chemical energy, which per-
mits the organisms to use the remaining portion of the organic carbon as
building blocks to synthesize the particular compounds (carbohydrates,
proteins, lipids, and nucleic acids) needed for their structure and func-
tion. Thus, a wastewater exerts an aerobic biochemical oxygen demand
(BOD). The amount of oxygen used during the process of obtaining
energy and extracting organic carbon from the wastewater for growth is
a measure of the biochemical oxygen demand exerted during the meta-
bolic purification of the wastewater. Thus, the exertion of BOD and the
growth of microorganisms, that is, sludge production, go on concur-
rently and are interrelated processes. The sum or total effect of these
interrelated processes is the removal of the carbonaceous material from
the waste water, i.e., purification of the waste. After the organic carbon
has been either oxidized or taken up by the microorganisms in the growth
process, oxygen uptake will still occur because the newly synthesized
population begins to undergo autodigestion or endogenous respiration.
 4 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

While the biochemical and mechanistic processes are of much interest,

most engineers are concerned more with quantitative description of the
rates of occurrence of these phenomena and the completeness of the
process. This important descriptive exercise can be aided by the study of
Figure 1.1. Let us assume (1) that the batch reactor shown in the figure
has been loaded with the waste containing amounts of organic carbon
compounds, nitrogen compounds, phosphorus compounds, etc., so that
balanced growth occurs, and (2) that the reactor has been seeded with a
small amount of activated sludge that has been thoroughly acclimated to
the carbon compounds contained in the waste. Aeration has begun and
waste constituents in solution are measured, in this case as chemical
oxygen demand (COD). Sludge concentration (X), and O, uptake are
also measured over time. In the figure, COD is indicated as a measure of
the organic matter present in solution. It can be seen that the elimination
of COD proceeds to some lower limit, i.e., there is a residual COD. Since
the biomass seed was well acclimated and conditions of pH, temperature,
nutrition, air supply, and mixing were optimal, the difference between
the initial and the lowest COD (the end of substrate removal phase) is
termed the ACOD and the numerical value of the ACOD is a measure of
the strength of the waste. This term shall be designated as S, in this text
(see Equation 1.1).

S, = ACOD = COD, - COD, (1.1)

Yn °

Xx ”
se Oo Uptake

Ss (as
Vv g
§ AcoD
c Reactor Biomass, X
Batch 8
Reactor 8

Reactor Substrate, S
(COD)

Xx
Time

Figure 1.1. Course of substrate removal, biomass growth and decay,

and oxygen utilization in a once-fed batch reactor under
aerobic conditions (from Gaudy and Gaudy, 1988).
 BIOKINETICS OF ACTIVATED SLUDGE 5

At the end of the substrate removal or purification phase the ACOD

has been either oxidized (i.e., registers as accumulated O, uptake) or
channeled into new biomass. Beyond this phase, called the autodigestive
phase, oxygen uptake increases but biomass decreases. In this phase
some of the primary feeders, bacteria, are used as food by the predator
population, largely the protozoa. The oxygen that is used in this phase
consists primarily of O, used by the protozoa in obtaining energy when
feeding on the bacteria and O, used by the remaining bacteria, which
continue to respire endogenously.
It is of much interest to note here that the 5-day BOD, which is
measured as a single point on the O, uptake curve, usually occurs long
after the substrate removal phase, i.e., well into the autodigestive phase.
Eventually O, uptake becomes asymptotic to some upper limit, the ulti-
mate carbonaceous BOD, L,. This value is always lower than the ACOD
(both values are measured in terms of oxygen). Further exposition of the
relationship between ACOD and L, can be found in the literature (Gaudy
and Gaudy 1988).

SOME IMPORTANT QUANTITATIVE CONCEPTS

In developing quantitative concepts, the substrate will be designated S.

S in practical terms will be considered to be COD or more correctly,
ACOD, but the reader should realize that other analyses could be used as
a measure of substrate; for example, total organic carbon (TOC ) or
ATOC. Biomass or activated sludge will be designated X and O, uptake
simply O,.

Cell or Sludge Yield

One of the most practical quantities of interest is the cell or sludge

yield, Y,. This factor is the ratio of the amount of sludge, or biomass,
produced per unit of substrate removed when the cells are growing rap-
idly. If one takes the biomass readings along the appropriate location on
the curve of Figure 1.1, the cell yield is given by Equation 1.2.

Y, = AX/AS = (X, - X,)/(S, - S) =

(X, - X,)/(COD, - COD,) (1.2)
6 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

At any point during the substrate removal phase shown in Figure 1.1, Y,
remains constant. Of course, a fairly low value will be measured if the
measurement is taken at some point in the autodigestive phase, and
erratic values may be obtained if the measurement is made too early in
the growth phase. We could measure the cell yield well into the endoge-
nous or autodigestive phases, but the value one obtains should not be
confused with Y,, which is often called the true or maximum cell yield,
i.e., the cell yield unaffected by autodigestion. When the cell yield value
can be shown to have been affected by autodigestion, it is designated by
the term Y,. This is a very important parameter and is dealt with later
when we discuss the production of excess sludge in continuous flow
systems.

Growth Rate and Decay Rate

The growth rate of microbial mass dX/dt or AX/At is always

expressed as the specific growth rate, p (i.e., the rate of growth per
average unit of biomass or cells extant during the time interval dt or At)
and is given by Equation 1.3.

pw = (1/X)(dX/dt) (1.3)

Figure 1.2 shows various phases in the microbial growth cycle, which
observers have used to describe the growth and decay of microorga-
nisms. The phases that are of most interest are the logarithmic, increas-
ing phase in growth and the autodigestive phase (mainly the decelerating
portion) because in these phases the specific rates » for growth and
specific decay rate ky for autodigestion are essentially constant. This
offers a convenient manner of quantitative kinetic evaluation as
described below.
The existence of exponential growth in a batch system is easily tested
by plotting the values of X or some marker for X such as optical density
(turbidity) against time on a semilogarithmic scale (see Figure 1.3). Quite
simply, the exponential phase is the straight line portion. The slope, ,, is
constant. During this phase, X at any time can be predicted by integrat-
ing Equation 1.3 as shown in Equation 1.4.

X, = X,exp(ut) (1.4)
The numerical value of » for this system can be evaluated from the
experimental data using Equation 1.5.
 BIOKINETICS OF ACTIVATED SLUDGE 7

(1) (2) (3) (4) , (5) (6)

Lag Li Li Decelerating Autod

oglesen egos5
neon Decreasing oe

x Stationary
a
3
E Accelerating
zs Autodigestion,
Log increasing

Time

Figure 1.2. Distinguishable portions of biomass growth and decay

curve (from Gaudy and Gaudy, 1988).

uw = In(X,/X,)/(t - t,) (1.5)

If one uses as the time interval the time it takes for a doubling of X, i.e.,
the doubling time t,, the specific growth rate p is given by Equation 1.6.

yp = In2/ty = 0.693/t, (1.6)

A plot such as the one shown in Figure 1.3 is extremely useful since it
provides a simple way to determine the existence and the extent of the
logarithmic phase and also shows the beginning of the declining phase of
growth. If there is a lag phase, a plot of the curve also shows when it ends
‘and the logarithmic phase begins. This is very important in handling field
data because the numerical values one obtains for » must be obtained
using only data points within the exponential phase. The more data
points one obtains, the more accurate and useful is the value of yp
obtained.
Similarly, the autodigestive phase is often characterized by a constant
specific rate of decline in biomass. Figure 1.4 depicts the logarithmically
decreasing autodigestion of a biomass, described by Equations 1.7, 1.8,
1.9, and 1.10. In these equations, t,, is the time required for the popula-
tion to decrease by one half.
8 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

2000 is, Maximum X

900
700

500
400
300

200
mg/L
X,
Biomass,
90
70 Declining
=0 Exponential Growth Growth
40
30

20

Time, h

Figure 1.3. Semilogarithmic plot of a biomass growth curve showing

extent of exponential or logarithmic growth (from Gaudy
and Gaudy, 1988).

dX/dt = -k’X (1.7)

X, = X,exp(-k’t) (1.8)

k’ = In(X,/X,)/t (1.9)

k’ = In2/ty; = 0.693/t,,; (1.10)

Relationship Between » and S (Noninhibitory Waste)

As we shall see, the specific growth rate y is a vital system characteris-

tic. It is affected by many physical and chemical environmental factors,
such as temperature, pH, chemical inhibitors, etc. These will be dis-
cussed in more detail in Chapter 6. However, the factor which makes this
 BIOKINETICS OF ACTIVATED SLUDGE 9

10
9
8
7
6
> 5
< 4

E 3
i)
a

1 ee ee ee ey
10 20 30 40
Time, days

Figure 1.4. Decelerating autodigestion (logarithmically decreasing).

Phase 6 of Figure 1.2 (from Gaudy and Gaudy, 1988).

parameter so unique in the description of the behavior of microorga-

nisms toward a specific waste is the fact that its numerical values are
governed mainly by the type of substrate and cells present. This fact
makes » a powerful tool for characterizing waste/activated sludge sys-
tems and predicting effluent quality. Moreover, p is affected by the
concentration of the waste carbon source or substrate. This fact is dis-
cussed below in developing models relating » to the concentration of the
waste. The specific growth rate at which a system is run is by far the most
important factor in determining both the design of the wastewater treat-
ment facilities and the amount of excess sludge to be handled.
If one sets up several experiments like the one shown in Figure 1.3,
results such as shown in Figure 1.5 (part a, b, or c) are generally
obtained. The initial concentration of wastewater is given in parentheses,
and it is seen that the slope, i.e., », is higher for the higher concentrations
of substrate. A plot of » vs. S (see the lower graphs) shows that yu
approaches some upper value as S is increased. The difference between
systems a and bd is that in b a higher concentration of S is required to
make p» approach its maximum value than in a. Examination of the graph
 “(9861 ‘Apnes) pue Apne WosJ) UONEIJUIIUOD 9)BI}SQNS [BHIU! SNSI9A 9}E1 yymMols
JyIIadG (WO}Og) ‘sasayjuaied UL UMOYS SUONBUIDUOD aJBIISGNS [BIIUL JE S9AIND YMOIT) (doy)
ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

-J)BAISQNS JO SUOT}EAJUIIUOD JUIIIJJIP JB PIAIISGO SIAINI YIMOIS YI}Eq JO SadA} SNOLILA JO sajdwiexgq ‘“S*] ansig
bu Os Yui Os “Vou5
000: 008 009 00r OZ 0001 008 009 00r 002 000: 008 009 Or 0%
/ J 10 : 10
10
Z0r Z0r Z0r
z= =
2 60) 2
s
Cx) co 2
v0 v0 70
so'0 S00 soo
g & 10 &
g
ro @ 10 @ x
bad x<
g g g
z0 § zo = z0 §
a a
€0 8
a
£0 co
pe 4 2
v0 = vos <=
10
 BIOKINETICS OF ACTIVATED SLUDGE 11

in part c shows that there is really no significant straight line portion

(i.e., essentially no exponential phase at the concentrations shown). This
is not the general case, and even when this happens, the specific growth
rate varies according to the initial substrate concentration as well as the
changes in S as the organisms grow. A more detailed discussion of these
types of data sets can be found elsewhere (Gaudy and Gaudy 1988) but
the main point to be remembered here is that when one analyzes the data
to determine an equation describing the lower graphs, it is nearly always
found that Equation 1.11 provides a fairly accurate fit of the experimen-
tal data. Equation 1.11 describes a rectangular hyperbola. The symbol
Mmax designates the upper or maximum value of » regardless of how high
S, is made and the symbol K, is a term related to the flatness or sharpness
of the curve as » approaches p,,,;-

econ ees
Misia (1.11)

One can see that the value of K, is numerically equal to the concentra-
tion of S which makes p equal to one-half of y,,,,. This equation is very
familiar in the field of biological kinetics; it is called the Monod equation
after Jacques Monod, who first demonstrated the fit of » to S in accord
with the rectangular hyperbola.
Figure 1.6 provides a numerical example of data description using the
equation. The lower portion shows a linear form of the equation and the
relationships needed to determine the numerical values of K, and p,,,, for
a cell/waste system. Other methods for obtaining numerical values from
experimental data are given in Chapter S.
At times engineers may doubt, when dealing with essentially nonde-
fined sewage and entirely heterogeneous microbial populations, that the
relationship between values of » at different concentrations of S would
follow according to the Monod equation. However, Figure 1.7 should
help allay any such doubts. This figure shows a plot of » values obtained
at varying initial COD concentrations for the soluble portion of a munic-
ipal sewage. The higher COD concentrations were obtained by concen-
trating the municipal sewage.

Relationship Between » and S (Inhibitory Waste)

Often, wastewater contains some substances that inhibit growth as

well as those that stimulate growth. Sometimes, the same waste constitu-
12 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

SLOPE =
Umax

max = 0.6 hr!

Ks= 125 mg/|

So
x 10?

Figure 1.6. Hyperbolic plot of the relationship between specific growth

rate, , and initial substrate concentration, S, (upper
graph), and a straight line plot, S,/n vs. S, of the same data
(lower graph).

ent used for growth stimulation can also inhibit or retard growth. As for
noninhibitory wastes, the specific growth rate depends on waste concen-
tration. For such wastes, the most commonly demonstrated effect of S
on p is One in which yp increases with S up to some concentration beyond
which further increase in S serves only to decrease ». As with nontoxic
wastes, there are several mathematical expressions that have been fitted
to plots of » vs. S, but by far the most commonly observed relationship is
the one expressed by Equation 1.12. We will refer to it as the Haldane
equation because of its similarity in form to the relation found by
Haldane to describe substrate inhibition in some enzyme systems.
 BIOKINETICS OF ACTIVATED SLUDGE 13

(ht)
p

So (mg/! COD)

Figure 1.7. Hyperbolic plot of the relationship between specific growth

rate, », and initial substrate concentration, S or S,, for a
heterogeneous microbial population of sewage origin
growing on a concentrate prepared from the soluble portion
of municipal sewage.

Bos Sieg So1e: |

ie erSite Kost S/Ks Co)

It is emphasized that both Equations 1.11 and 1.12 are selected for
modeling of wastewater processes purely on the practical basis that they
are found to provide rather good fits to experimental results obtained in
systems important to environmental engineers. It is not necessary and is
probably dangerous to assign general theoretical importance to these
equations by bridging between descriptive formulations for enzyme
kinetics and those for microbial growth, other than to realize that micro-
bial growth rate is governed at the molecular level by enzymes. One
should realize, however, that factors governing one enzyme reaction
could not necessarily be expected to govern the overall kinetics of growth
when one considers the complicated array of enzymes embodied in a
living cell. Equations 1.11 and 1.12 are justifiably used in modeling
because (1) they can be shown to fit real wastewater data in the field and
(2) the numerical values of the biokinetic constants they employ have real
quantitative value in predicting activated sludge treatment performance.
14 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Examination of these equations shows that there are two biokinetic con-
stants in Equation 1.11 and an additional one in Equation 1.12. The
constant K;, is termed the inhibitory constant and it expresses the inhibi-
tory nature of the waste to the growing cells. If its numerical value is very
large, then inhibition is minimized and Equation 1.12 is the same as
Equation 1.11.
Figure 1.8 compares the type of plot rendered by Equations 1.11 and
1.12. The difference in the type of behavior of » with increasing S, is
apparent. Increasing concentrations of toxic substrates inhibit (depress)
p. The p» value does not approach some maximum y, not affected by
further increase in S; rather, it goes through some maximum value at a
specific concentration of S. This peak in yp is extremely significant in
assessing the stability and resilience of systems growing on inhibitory
wastes. It is designated the critical specific growth rate, n*. Its numerical
value and the substrate concentration at which it occurs, S*, are easily
determined. Setting the first derivative of Equation 1.12 to zero and
solving for S and yp yields Equations 1.13 and 1.14.

ee ie
e+ 2a a)

Hmax

Haldane

Oia
ee
=

Figure 1.8. Relationship between » and S according to noninhibitory

(Monod) and inhibitory (Haldane) equations for microbial
growth on substrate. u* is the peak, or highest possible,
specific growth rate on the inhibitory substrate and S* is
the substrate concentration at which » = y*. Higher
substrate concentrations cause a decrease in growth rate
(from Rozich, Gaudy, and D’Adamo, 1985)
 BIOKINETICS OF ACTIVATED SLUDGE 15

S* = /KK, (1.14)
These equations yield numerical values of substrate and specific
growth rate beyond which activated sludge systems treating the specific
toxic wastewater in question cannot grow and will fail; that is, the bio-
mass will wash out of the reactor and effluent quality will totally deterio-
rate. How these values are employed in operation and design is detailed
in Chapter 3.

Nature of the Growth Curve

Analysis of the growth curves from which yp is obtained for the various
values of S, deserves special comment. As with nontoxic substrates,
exponential growth (straight line portion of a semilogarithmic plot of X
vs. t) is observed, but growth is generally much slower at all S, concentra-
tions than for the nontoxic substrate and is extremely slow at high S,
values. In addition, at high values of S,, the general shape of the growth
curve after attainment of the exponential phase is distinctly different
than for nontoxic substrates. After the exponential phase, » actually
increases, because as the cells grow, the substrate concentration is
reduced to a less toxic value, permitting a higher specific growth rate.
Figure 1.9 shows growth curves for heterogeneous populations of sew-
age origin on an inhibitory substrate, phenol. Part a of the figure shows
growth rates obtained at S, concentrations from 50 to 500 mg/L phenol
COD, and in part b, concentrations range from 700 to 900 mg/L phenol
COD. The graphs show that all the growth data in this experiment were
obtained at S, concentrations above S*. Note that the growth rates
decreased for increasing S, values. It can also be noted that the curves for
500 to 1000 mg/L show increased slopes after the exponential phase. As
noted above, this is only observed for toxic or inhibitory substrates when
the system is initiated at S, values significantly in excess of S,*. In the
case shown, S* was 30 mg/L phenol COD. When initial substrate con-
centrations below S* are employed, the general shape of the growth
curve is the same as that for a nontoxic substrate. The difference in the
shape of the growth curves on either side of S* can also be demonstrated
by a computational simulation of growth using the Haldane equation.
Such a comparison is made in Figure 1.10 for S, concentrations ranging
from 0.5 S* to 10 S*. It is seen that at S* and below, the growth curve is
that for a typical noninhibitory substrate, i.e., an exponential or appar-
ently exponential phase followed by a decreasing rate; whereas growth at
16 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.40 &
@ 1 =0.150h *1 °
V w=0.146h 71
D p=0.130h 1 =
o » =0.075h -1 °

density)
(optical
X

0 5 10 15 20 25 30 35
Time (h)

Figure 1.942. Growth on phenol at initial concentrations of 50 (@), 100

(V), 200 (CJ), and 500 (°) mg/L. Lines show exponential
phase (from Rozich, Gaudy, and D’Adamo, 1985)

S, concentrations above S* is typical of that of an inhibitory substrate

concentration, i.e., exponential followed by increasing rate and finally
decreasing rate as the carbon source is exhausted. It is important to be
aware of these differences in behavior of the growth curves when devel-
oping and analyzing growth data from which numerical values of the
biokinetic constants are to be determined.

Relations Between » and S

Monod, Nontoxic Wastes

This relationship is defined by the specific growth rate p, which

increases with increased S, values at a decreasing rate of increase, i.e., it
follows the form of a rectangular hyperbola. Equation 1.11 is an empiri-
 BIOKINETICS OF ACTIVATED SLUDGE 17

0.80

0.60

0.40

0.20

0.10
X(optical
density)

0.04

0 10 20 30 40 #50 60 £70
Time (h)

Figure 1.9b. Growth on phenol at intial concentrations of 700 (¥), 900

(M1), and 1000 (©) mg/L. Lines show exponential phase
(from Rozich, Gaudy, and D’Adamo, 1985).

cal one based on observations of experimental results. The equation

holds for heterogeneous populations and mixed carbon sources as well as
for pure cultures and sole carbon sources. The biokinetic constants pax
and K, are dependent upon temperature, pH, etc., as are any chemical
constants, but for given conditions or a range of conditions y,,,, and K,
are characteristic of the waste and acclimated biomass growing on it.

Haldane, Toxic Wastes

This relationship is given in Equation 1.12. It is similar to the Monod

equation but the last term of the denominator accounts for decreasing p
with increasing substrate concentration beyond a critical value of S, i.e.,
S*, which is given by Equation 1.14. The critical specific growth rate p*
that occurs at this substrate concentration is defined by Equation 1.13.
S* and p* are particularly important for design and control of bioreac-
tors in which inhibitory wastes are to be treated because beyond this
18 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.50 (10 S*)

0.30
(10 S*)
Me

X(optical
density)

0.0.

Time (hours)

Figure 1.10. Dependence on initial concentration of the shape of batch

growth curves according to the Haldane relationship
between p and S. y,,,, = 0.34 hr', K, = 32 mg/L, K, = 67
mg/L, Y, = 0.00085 O.D. units/mg substrate/L, S* = 69
mg/L, and initial biomass concentration = 0.025 O.D.
units. Dashed lines represent exponential growth (from
D’Adamo, Rozich, and Gaudy, 1984).

substrate concentration and specific growth rate the system cannot exist
for all practical purposes in most applications.
The Haldane equation becomes the Monod equation when a system
exhibits very high values of K;. Thus, there is a gradation from toxic to
nontoxic wastes. One determines whether a waste exhibits a toxic or a
nontoxic reaction by observing the nature of the experimental plots of
vs. S in accord with Figure 1.8. If it is found that » decreases after
peaking at some value of S, then the system exhibits an inhibitory nature.
Such an observation is important because it determines which type of
equation should be used to design and operate the system. Note the
values of »* and/or S* provide a basis for assessing the severity of the
toxic condition of the system of cells and waste or substrate.
 BIOKINETICS OF ACTIVATED SLUDGE 19

Relation Between S, X, and O,

There is a relationship between substrate removal, i.e., purification,

growth of biomass, and oxidation (O, uptake) implicit in Figure 1.1.
Figure 1.1 tells us that the COD being removed from solution is
accounted for totally by biological oxidation and by growth of biomass.
If the waste components were subject to rapid stripping or were very
easily oxidized by oxygen this condition would not hold. Whether the
waste is strippable or chemically oxidizable simply by aeration is easily
checked; in most cases, chemical oxidation and stripping do not account
for substrate removal. Thus the split of substrate removal between respi-
ration (oxidation) and synthesis (growth) provides a good basis for mak-
ing mass and energy balances on this aerobic biological system, and a
useful methodology to check results and to use respirometry to obtain
needed growth data. At various points along the COD removal curve of
Figure 1.1 the amount of COD removed must be accounted for as the
summation of that used for synthesis (i.e., the amount appearing as
biomass) and the amount that has been oxidized (i.e., the amount that
can be accounted for as. oxygen uptake). These values can be easily
measured and inserted in Equation 1.15.

ACOD, = AO, + AX, (1.15)

The equation states that the amount of COD removed at any time is
equal to the sum of the accumulated O, uptake to that time and the
amount of cells produced. However, one must first express the amount
of cells in terms consistent with the other terms. We already have COD
expressed in terms of oxygen and the AX term can be converted to
equivalent O, by determining the COD of the cells. We can do this either
by direct COD analysis or by determining the unit COD per unit of
biomass, O,. This ratio can change after the substrate removal phase.
The COD of growing cells is often somewhat higher than that for autodi-
gesting cells. During the growth phase this value remains fairly constant,
similar to the cell yield, Y,. Numerical values of O, should be determined
for the wastewater system under study but for purposes of preliminary
analysis, the O, obtained from Dr. Porges’ empirical formula for acti-
vated sludge may be used (Porges, Jascewicz, and Hoover, 1956).

CHENG, 49507 5CO) +"°2H,0"+' NH, (1.16)

O, = (5)(32)/113 = 1.42 mg O,/mg biomass
20 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Equation 1.15 may now be written correctly as:

ACOD, = AO,, + AX,°O, (lad7)

Equation 1.17 can be used to check the accuracy of test data. It is

essentially an energy balance. This relationship between O, uptake,
growth, and purification can also be used in another important way.
Since AX, is related to ACOD, by the cell yield, Y,, Equation 1.17 can be
modified to Equation 1.18.

ACOD, = Y,;O.-ACOD. +,,A03, (1.18)

ACOD, is also defined by Equation 1.19.

ACODs =al KX ays (1.19)

Equation 1.19 is substituted into Equation 1.18 to produce Equation

1.20.

Nr Nor AO Cy ae Gh) (1.20)

With Equation 1.20 we can obtain the growth curve from the O, uptake
curve, provided we know the values of Y, and O,, both of which are
easily determined. This equation is extremely important because it per-
mits one to make use of those parameters to obtain growth curves needed
to evaluate the biokinetic constants y,,,,, K,, and K,, which are employed
in activated sludge models. In contrast to direct measurement of X, by
optical density or gravimetric means, both of which methods require
much labor and expense, respirometric readings of accumulated oxygen
uptake can be obtained rapidly and automatically and converted to X,
values using equation 1-20. Thus the copious amounts of data needed to
define growth curves are easily obtained. The COD remaining in solution
at any time can also be determined through Equation 1.21.

COD, = COD, - AO,,/(1 - 0,Y,) (1.21)

It can be seen that the denominator of the last term of Equation 1.20
combines two system constants and may be represented by a single term,
Ri

R = 1/Y,-0, (1.22)
 BIOKINETICS OF ACTIVATED SLUDGE 21

The term R is defined as the respirometric ratio and has definite physical
significance; it is the milligrams O, consumed per milligram of biomass
produced. Thus the numerical value of R can also be depicted using
Equation 1.23.

Ry 3 AOse/ Sere a) (1.23)

One can determine R using both Equations 1.22 and 1.23 as a check on
the accuracy of the data.
Equations 1.20 and 1.21 can now be written as given below.

SN AOL TR (1.24)
COD, = COD, - AO,,/(R-Y,) (1.25)
After determining the respirometric ratio, R, one can very easily pro-
duce the growth and COD removal curves from the respirometric data.
(Gaudy et al. 1988, 1989, 1990). It should be emphasized that the equa-
tion pertains to the substrate removal phase even though the energy
balance principle upon which these equations were derived is valid in
both the substrate removal and autodigestive phases.
Equations 1.24 and 1.25 provide the engineer with a powerful tool for
obtaining growth and substrate removal rate data which are used to
determine the numerical values of biokinetic parameters for biological
treatment systems. These values are in turn used with the reactor model
to be presented in Chapter 3. However, before proceeding to the descrip-
tion of engineering models for activated sludge reactors, it is most desir-
able to present some basic kinetic reactor engineering principles of con-
tinuous growth, which when combined with the biokinetic principles just
delineated form the basis for the engineering approach to be presented
later.

KEY CONCEPT SUMMARY

It may seem an oversimplification to proclaim that since the beginning

of aerobic biological treatment of wastewaters (especially by the acti-
vated sludge process) the status of its conceptual understanding can be
condensed to the simple definitions and relationships represented in the
25 equations thus far discussed. While space considerations and the
desire to simplify have necessitated avoidance of some important details
22 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

and possible refinements of the concepts, the fact remains that, through
continued research, understanding and use of these concepts are just
coming into engineering practice in the field and the material presented
thus far can be truly represented as “state-of-the-art.”
Since understanding of the foregoing terms and relationships is such a
vital foundation for practical application of material that follows in this
text, it is important to summarize the key terms and related concepts
before proceeding to the next chapter.
The key concepts or definitions contained in this chapter are:

¢ Biomass concentration, X, in units of mg/L.

e Substrate (or waste) concentration, S, limiting nutrient source
(usually organic carbon) in units of mg/L, usually expressed as
biodegradable COD, i.e., ACOD.
e O, uptake, respiration or accumulated oxygen uptake in mg/L, a
measure of the biochemical oxygen demand (BOD) of the sub-
strate or waste.
¢ Specific growth rate, p, in time”', defined according to Equation
1.3. This term is relatable to 0,, F:M, and U (see Chapter 4).
e Specific cell decay rate, ky, in time”.
© Cell yield, Y,, maximum or true cell yield in mg cells/mg S, AX/
AS; the cell yield unaffected by autodigestion, a constant which is
a function of the waste and biomass, defined in Equation 1-2.
¢ Observed cell yield, Y,, the net cell yield, i.e., the cell yield
affected by autodigestion. The amount of autodigestion that
occurs depends on the condition of growth. It is important under
conditions of slow growth in continuous flow reactors (discussed
in Chapter 2).
¢ Monod equation, relates » and S for nontoxic substrates or
wastes (see Equation 1.11).
¢ Haldane equation, relates » and S for toxic or inhibitory wastes
or substrates (see Equation 1.12 ).
¢ Maximum specific growth rate, y,,,,, in time, the highest spe-
cific growth rate obtainable for growth on a nontoxic substrate in
the presence of excess concentrations of that substrate.
¢ Saturation constant, K,, in mg/L. A shaping factor that deter-
mines the sharpness at which a plot of « versus S approaches pax
for growth on nontoxic substrates.
¢ Inhibition constant, K,, in mg/L, a shaping factor that accounts
for the peak and decrease in pw for increasing concentrations of S
 BIOKINETICS OF ACTIVATED SLUDGE 23

in the Haldane equation for growth on toxic or inhibitory

substrates.
¢ Critical substrate concentration, S*, in mg/L. The concentration
of an inhibitory substrate at which the peak in a plot of » versus S
occurs according to the Haldane equation (see Figure 1.8 and
Equation 1.14).
¢ Critical specific growth rate, *, in time™'. The highest or peak
specific growth rate attainable for growth on a toxic or inhibitory
substrate according to the Haldane equation (see Figure 1.8 and
Equation 1.13). Note: Do not confuse y,,,, and u*; for inhibitory
substrates, p,,,, can never be observed experimentally. It must be
determined analytically from growth data obtained at lower spe-
cific growth rates (see Chapter 5).
¢ Relating O, uptake (respiration) to growth and _ substrate
removal.
e Equation 1.20 or Equation 1.24 is used to convert O, uptake
data to cell growth data.
e Equation 1.21 or Equation 1.25 is used to convert O, uptake
data to substrate removal data.
These equations are based on the fundamental energy balance
equation for aerobic metabolism, Equation 1.17. They provide
the key to rapid and easy determination of numerical values of
the biokinetic constants through respirometry.
* O,, oxygen equivalent or COD of cells, mg COD/mg cells.
e R, respiration quotient, O, required to produce a unit of bio-
mass, in mg O,/mg X, determined in accordance with Equations
Le22-and/ord<23:%

REFERENCES AND SUGGESTED ADDITIONAL

READING

D’Adamo, P.C., Rozich, A.F. and Gaudy, A.F. Jr. (1984). “Analysis of
Growth Data with Inhibitory Carbon Sources,” Biotechnology and
Bioengineering, XXVI, pp. 397-402.
Gaudy, A.F. Jr., Rozich, A.F., Moran, N.R., Garniewski, S.T., and
Ekambaram, A. (1988). “Methodology for Utilizing Respirometric
Data to Assess Biodegradation Kinetics.” Proceedings, 42nd Purdue
Industrial Waste Conference, Lewis Publishers, Chelsea, Michigan,
pp. 573-584.
Gaudy, A.F. Jr., Ekambaram, A., and Rozich, A.F. (1989). “A Respiro-
24 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

metric Method for Biokinetic Characterization of Toxic Wastes,” Pro-

ceedings, 43rd Purdue Industrial Waste Conference, Lewis Pub-
lishers, Chelsea, Michigan, pp. 35-44.
Gaudy, A.F. Jr., Ekambaram, A., Rozich, A.F. and Colvin, R.J. (1990).
“Comparison of Respirometric Methods for Determination of
Biokinetic Constants for Toxic and Nontoxic Wastes,” Proceedings,
44th Purdue Industrial Waste Conference, Lewis Publishers, Chelsea,
Michigan, pp. 393-403.
Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental
Engineering. Engineering Press, Inc., San Jose, California, USA.
Porges, N., Jascewicz, L., and Hoover, S. (1956). “Principles of Biologi-
cal Oxidation,” in Biological Treatment of Sewage and Industrial
Wastes, McCabe, B.J. and Eckenfelder, W.W., ed., Reinhold, New
York, pp. 25-48.
Rozich, A.F., Gaudy, A.F. Jr., and D’Adamo, P.C. (1985). “Selection of
Growth Rate Model for Activated Sludges Treating Phenol,” Water
Research, 19, pp. 481-490.
 2 BASIC PRINCIPLES
OF BIOREACTOR
MODELING

INTRODUCTION

The kinetic principles described in Chapter 1 are universally applicable

to the growth of aerobic organotrophic microorganisms regardless of the
reactor in which the growth takes place. It may be a BOD bottle, a river,
a lake, or a pot of soup left standing too long on the stove (or more
contemporarily, in the microwave). However, this is not to say that the
nature of the reactor and how it can be regulated and controlled by an
outside agency (i.e., the design and operations engineers) cannot have a
determining role in the rate and completion of the purification reaction
represented by Equation 1.17.
In fact, it is essential that the reactor be designed to accommodate the
mathematical or kinetic boundaries and assumptions as to experimental
conditions under which the kinetic expressions were derived. Moreover,
certain kinds of reactor systems allow one to exert control of the specific
rate over and above that expressed in either the Monod or Haldane
equations because one must understand that as S controls or determines
specific growth rate, so too does specific growth rate determine S. Thus,
engineering ways and means to control specific growth rate provide pow-
erful tools to control concentration of S in the effluent, which is the
“raison d’etre” for the treatment plant. It is the reason the activated
sludge process has come to such a position of supremacy when the kind
of treatment process to employ is chosen.
There are obviously many different types of reactors in which to
accommodate microbial growth, but in this chapter we shall discuss
important concepts applicable to the type of reactor most apropos to the

25
26 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

activated sludge process, i.e., a continuous flow, completely mixed,

totally fluidized reactor. Two important concepts related to further dis-
cussion of models for activated sludge will be presented:

1. Hydraulic control of specific growth rate, p.

2. Effect of cell recycle on specific growth rate.

The accuracy of the description of substrate removal, growth and

respiration, which was shown in Figure 1.1 for batch growth systems and
subsequently discussed up to this point, remains true for growth in con-
tinuous flow reactors. The difference between the batch and continuous
flow systems is essentially one of hydraulics. This, we shall see, does have
profound effect on the performance of the system. Let us change the
reactor of Figure 1.1 by providing an influent and effluent pump (see
Figure 2.1) so that wastewater can enter and exit from the reactor at the
same rate, i.e., F, (flow in) = F, (flow exiting). Again, vigorous aeration
is provided and the reactor is “completely mixed.” (This latter term sim-
ply means that all of the contents, soluble and particulate, are at the
same concentration in every part of the reactor.) Each component in the
flow is immediately mixed in the reactor volume V and each unit of
outflow is the same concentration X and S as it is in the reactor. Let us
suppose that we set up a feed tank from which a waste at strength S, is
pumped into the reactor at some flow rate F and pumped out.or allowed
to flow out at the same rate F. Let us begin with the reactor filled with the
wastewater at concentration S; as before when we filled the reactor with

Reactor Biomass, X, Xg
fk
Bont
ae ad
on

S
X,
Concentration Reactor Substrate, S, Sg (COD)
Completely Mixed
Continous Flow Reactor

—
Time

Figure 2.1. Course of substrate removal and biomass growth in a

continuously fed, completely mixed reactor.
 PRINCIPLES OF BIOREACTOR MODELING 27

some initial concentration, S,. We will again feed the reactor with a small
concentration of acclimated activated sludge X,, and we will again
sample at various times and develop the substrate and biomass curves
shown in the graph to the right of the reactor. As before, the cells grow,
X increases, and the concentration of substrate in the reactor, S,
decreases. Eventually, the system becomes steady with respect to concen-
tration of S and X. That is, the concentrations remain constant regard-
less of how long we run the reactor. The system has come into a “steady
state” with respect to S and X. The consequence of this result, which is
readily observable in the reactor, can best be shown by analysis of the
growth equation (Equation 1.3) and the hydraulic holding time. The
hydraulic holding time is termed the reactor detention time, t.

ji WAS (2.1)

The reciprocal oft is defined as unit flow rate, D, or dilution rate.

Ds =n FAV) =o lt (2.2)

The term dilution rate is a convenient one, indicating that the incom-
ing wastewater concentration S, is diluted by the factor F/V as it enters
the reactor. It is given in units of time™', i.e., the units for replacement of
hydraulic liquid volume are the same as the units for specific growth rate,
pw. Recall that Equation 1.3 describes the mass rate of increase for batch
growth; this equation is also valid in the continuous flow reactor.

dX/dt = pX (1.3a)
However, in the continuous flow reactor biomass X is also exiting at a
rate given by Equation 2.3.

dxX/dt = -DX (2.3)

When the system attains the steady concentration in S and X, which is

indicated with time in Figure 2.1, dX/dt — 0, which leads to the follow-
ing identity.

dX/dt = ~X - DX
»=D (2.4)
Equation 2.4 is a profoundly useful one. It states that the specific growth
rate is subject to hydraulic control. This is very important; since we can
28 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

exert engineering control over the hydraulics of the system, i.e., D = p

= F/V, we can control the specific growth rate by changing the hydraulic
flow into the reactor. The equation can be refined to account for the fact
that we know from experimental evidence that autodigestion can occur
while the cells are growing. The specific rate of autodigestion, k,, is
usually much slower than the specific growth rate but does manifest a
considerable effect on the prediction of X when the system is made to
grow slowly. It should be remembered that for the activated sludge sys-
tem the specific growth rate is usually held to a very low numerical value.
Thus we shall refine Equation 2.4 by accounting for autodigestion and
introducing the term p,, i.e., the net specific growth rate.

DeSales (2.5)

Equation 2.5 represents a fundamental identity for any completely mixed

continuous flow reactor system.
Although Equation 2.5 tells us much about the reactor system and its
control, it is important to develop reactor equations for predicting X, S,
and another important parameter, the excess cells or sludge produced in
the system. Starting with the simple reactor, i.e., a reactor with no recy-
cle of cells, such as that shown in Figure 2.1, the reactor equations are
developed simply by writing the appropriate mass balances with respect
to biomass X and substrate S.

Mass rate of (+) rate of change (-) rate of change (-) rate of change
change in X due to growth due to decay due to outflow
dS
Pers = Vex - VkyX - FX (2.6)

Mass rate of _ (+) change due (-) change due (-) change due
change inS — to inflow to outflow to growth
dS px
eo
i FS,) FS, gets 3 (2.7)

Note that for each of these mass balances (Equations 2.6 and 2.7), any
biomass in the inflowing line is neglected. Also note that the substrate
concentrations in the reactor and exiting the reactor are the same, i.e., S
= S, in accord with the condition of complete mixing.
Dividing Equation 2.6 by V and setting dX/dt = 0 leads to another
form of the previously given Equation 2.5.

w=D+ky (2.5a)
 PRINCIPLES OF BIOREACTOR MODELING 29

Solving the substrate balance Equation 2.7 when dS/dt = 0 yields Equa-
tion 2.8.

X = Y.DG, - S.)/p (2.8)

ONCE-THROUGH SYSTEM (CHEMOSTAT),

NONINHIBITORY MODEL

The simultaneous solution of Equations 2.5a and 2.8 is facilitated by

substituting equations relating p and S. In the case of the noninhibitory
substrate, the Monod equation is substituted. These predictive equations
are given in Equations 2.9 and 2.10.

SS-Y,DG;-
wpe8.) (2.9)

gre KD ky) (2.10)

Mmax ~ (D ay k4)

The performance of the reactor with respect to biomass concentration

and substrate at various hydraulically controlled growth rates (D values)
is demonstrated in Figure 2.2. The effect of autodigestion on X is felt at
very low D values (high t values) and as the system is run at increased
values of py, the steady state concentration of S increases until it reaches
the value of S; and the biomass is completely washed out of the reactor.
The washout value of D is easily calculated. Combining Equations 1.11
and 2.5a and substituting S, for S provides Equation 2.11.

D’ = p,! = —Hmax Bis sik (2.11)

The numerical values D’ and ,' are termed the critical dilution rate or
critical net specific growth rate when the waste substrate is of a non-
inhibitory nature. The term is comparable to u* or p,* when the waste is
of an inhibitory or toxic nature.
For the case shown in the figure, this washout (i.e. critical) D’ or p,’ is
0.46 hr-!. This calculation tells us that biomass exhibiting the biokinetic
characteristics pmax = 0.50 hr', K, = 75 mg/L, and k, = 0.005 hr''
cannot be made to grow at a net specific growth rate above 0.460 hr".
Prior to approaching this critical ,, value, there is increasing loss of S,
i.e., increasing deterioration of treatment efficiency. The flatness (or
30 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

1000

900

800

700

600

500
mg/L
S,
X,
400

300

200

100

0.1 0.2 03 0.4 05

,, D, hr 4

Figure 2.2. Predicted steady-state biomass and_ substrate

concentrations in a once-through reactor according to
Equations 2.9 and 2.10. y,,,, = 0.5 hr', K, = 75 mg/L,
Y, = 0.6, k, = 0.005 hr', S; = 1,000 mg/L (from
Gaudy and Gaudy, 1988).

sharpness) of the curve predicting effluent quality depends on the value

of K,. Lower K, values will cause the curve to be flatter as the critical net
specific growth rate, y,’ is approached. The figure also shows that the
slower one causes the system to grow, the lower is S,, i.e., the treatment
efficiency is better at lower net growth rates. Also, it is seen that for very
low p, values, X becomes lower because of autodigestion of the biomass,
i.e., » approaches ky.
 PRINCIPLES OF BIOREACTOR MODELING 31

ONCE-THROUGH SYSTEM (CHEMOSTAT), INHIBITORY

MODEL
Substituting the Haldane equation for » in Equations 2.5 and
2.8 provides the following reactor equations for inhibitory
systems.

xX =_ Y:DG;
ae- 5.) (2.9)

22 1/2
Ss az
= onan
Bmax =
1 +?
(:ra DE:
Umax
‘| =
4 k.
K| yer)¢ (2.12)

Figure 2.3 compares the reactor performance in accord with equations

using the Haldane and Monod expressions, i.e., Equation 2.10 vs. Equa-

Haldane Monod

0 0.04 0.08 0.12 0.16 0.20

D (1/hour)

Figure 2.3. Comparison of predicted dilute-out behavior of Monod

and Haldane specific growth rate relationships in a
once-through (chemostat) reactor. p,., = 0.194 hr', K, =
48 mg/L, K; = 62 mg/L, Y, = 1.02 mg/mg, and k, = 0.02
hr! (from Rozich and Gaudy, 1986).
32 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

tion 2.12. The behavior at increasing values of D or yu, (decreasing t) is

radically different for the noninhibitory and inhibitory substrates. It is
seen that the system washes out very abruptly at a much lower substrate
concentration and a different dilution rate for the inhibitory substrate.
There is no sliding up of S, no warning that », may be approaching the
critical specific growth rate, as there is with the nontoxic waste. From the
biochemical constants for this system in which the cells are growing on
phenol, the critical net specific growth rate p,* as calculated using Equa-
tions 1.13 and 2.5 was 0.05 hr-!. One can readily see that for the inhibi-
tory waste there is greater need to control the net specific growth rate and
much greater need to protect the system from sudden changes in S,;
because total failure of the treatment system, i.e., washout, can occur at
much lower waste substrate concentrations and it occurs more abruptly
than with noninhibitory wastes.

PREDICTION OF EXCESS BIOMASS (EXCESS SLUDGE)

Thus far, we have equations that predict S and X for systems when the
reactor is controlled by its operator at various values of net specific
growth rate. We can only predict S and X if we know the numerical
values of the biokinetic constants, p,.,, K,, K; (if the waste is inhibitory),
Y,, and k,. These values are obtained through performance of periodic
laboratory tests, easily facilitated through use of data and Equations
1.20 and 1.21 or 1.24 and 1.25. The methodology is discussed in Chapter
5:
The amount of excess sludge produced, X,,, for the once-through reac-
tor system is readily calculated on the basis of mass per unit time from
Equation 2.13.

a = VX, (2.13)

Another important concept that can be introduced at this point is that

of the time it would take to replace the biomass in the reactor. If one
knew X,, the biomass replacement time ©, could be readily determined
by the following equation.

0, = VX/X, (2.14)
 PRINCIPLES OF BIOREACTOR MODELING 33

This term is familiar to many engineers as the sludge age or mean cell
residence time (MCRT). Examination of Equations 2.13 and 2.14 show
that », and ©, are related as shown in Equation 2.15.

uw, = 1/0, (2.15)

This relationship is discussed more fully in Chapter 3.

REVIEW

In Chapter 1 we introduced the concept of mass and energy balances

based on the partition of substrate into that which is used for respiration
(O, uptake) and that channelled into new biomass. This permits one to
develop a useful quantitative relationship for determining X and/or S
using O, uptake, which is readily obtained in laboratory tests on a waste/
biomass (sludge) system. In Chapter 2, the growth relationships (Monod,
Haldane) have been employed in simple mass balance equations for a
bioreactor (once-through [chemostat] flow system) and predictive equa-
tions for S, X, and X,, were developed; the relationship between the net
specific growth rate «,, and mean cell residence time 0, has been shown.
It has been demonstrated that the », (or ©.) profoundly affects the
performance of the system because this parameter controls the concen-
tration of substrate exiting the reactor.
There is still one last important engineering expedient that must be
delineated. The recycle of biomass to the reactor provides a powerful
tool for controlling reactor behavior because it offers two additional
ways to control the net specific growth rate of the system.

EFFECT OF BIOMASS RECYCLE

A simple once-through reactor system is converted into one capable of

recycling biomass by adding a cell separator after the reactor and making
provisions to recycle concentrated cells. In the case of the activated
sludge process, the cell separator is a sedimentation basin. The essential
difference between the systems is delineated in the schematic diagram of
Figure 2.4.
The reactor equations in X and S for the once-through system are
different from the ones for the recycle reactor simply because the reactor
now has two input lines, the wastewater and the recycle line. This addi-
34 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Once Through Reactor System

SF Vv F,X,Se

Cell Recycle Reactor System

a=FR /F caX GK

Figure 2.4. Comparison of flow diagram for once-through and recycle

reactor systems.

tional input consists of a recycle flow F, which can be controlled at some

fraction (a) of the inflowing wastewater (a = F,/F). This recycle flow
contains a biomass concentration (Xx), which may be controlled at some
ratio c (Cc = X,/X). The flow may contain some amount of substrate,
and for purposes of simplification we may assume that there is essentially
no chemical or biochemical activity in the cell separator or clarifier so
that the substrate concentration in both the effluent and the recycle flow
are the same as in the reactor.
Mass balances for dX/dt and dS/dt are given by Equations 2.16 and
Pay tile

V on = (aF)(cX) + VuX - VkyX - (1 + a)FX (2.16)

t :
inflow growth decay outflow
(recycle)

V =t adc Siete fang =Y; oe a)S, (2.17)

inflow recycle utilization outflow
 PRINCIPLES OF BIOREACTOR MODELING 35

In the example below it is assumed that the waste is of a nontoxic

nature and the Monod relationship between y and S has been employed.
As before, we divide through by V and let dX/dt and dS/dt approach
zero. Solution of Equation 2.16 yields Equation 2.18.

w= DU + a-ac) + ky (2.18)

We see that, as in Equation 2.5, p still falls under hydraulic control; that
is, engineering control, but (D or t) is not the only hydraulic control. The
recycle flow ratio a and the recycle sludge concentration ratio c also exert
controls over the specific growth rate. As for the once-through system,
solving the substrate balance for the steady-state condition and substitut-
ing the Monod relationship for p leads to the following predictive equa-
tions for X and §..

Y,DGS;i = S.) (2319)

en] + a-ac) + k,

K,[DU + a - ac) + ky]

(2.20)
pe eat [D@Lpie = ac) Chaka]

It is seen that Equations 2.9 and 2.10 for the once-through (chemostat)
system differ from the above equations only by the recycle factor (1 +
a—a:*c). We know from previous discussions that S., effluent quality, is
determined by » and from Equation 2.18 that » is determined largely by
the factor D(1 + a - ac). We can see the tremendous effect recycle has
on the effluent quality by examination of Figure 2.5.
For this demonstration the sludge concentration factor (c) of 4 was
selected since this is a concentration ratio easily obtained by simple sedi-
mentation of cells and is commonly found in activated sludge systems.
The most striking conclusion one draws from this figure is that cell
recycle enables one to produce much lower S, values for less reactor
detention time, i.e., less reactor volume than is possible without recy-
cling. Recycle also provides the opportunity to control S, after the plant
is in operation by allowing the operator to select and control the numeri-
cal values of aw and c. The only way D (i.e., t) can change is by a change in
F, since V is fixed once the plant is built. Cyclical as well as long-term
increases in F only serve to worsen effluent quality because an increase in
F obviously increases p. Thus, a and c provide important engineering
controls. These along with the biokinetic constants p,,,,, K,, Kj, Y,, and k,
exert an effect on » and thus on the effluent quality produced. As we
36 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

2200 Cell Recycle System

mg/L
X,S,
Predicted

Once Through System

0 tt} —}—_ fF— fT

0 01 02 03 04 05 06 0708 09 1.0 1.1 12 13 14 1516 1.7 1.8 1.9 20

Dilution Rate D = 1/,, hr *

Figure 2.5. Comparisons of predicted dilute-out behavior for

once-through (chemostat) and cell recycle reactors. p,.,
= 0.5 hr', K, = 75 mg/L, Y, = 0.6, K, = 0.005 hr', a
= 0.25, ¢ = X,/X = 4 (from Gaudy and Gaudy, 1988).

shall learn in a succeeding chapter, the value of S; can also exert a

controlling effect on » and consequently on the quality of the effluent.
Controlling of effluent quality is, of course, the aim of the exercise of
design and operation of the activated sludge process; the engineering and
biokinetic factors are the tools available to accomplish the job of waste-
water purification. Other important aspects, such as provision of ade-
quate mixing and oxygen and separation of biomass from the effluent in
the clarifier, will not be addressed in detail in subsequent chapters of this
COXts
Thus far we have discussed basic kinetic considerations that led to the
model Equations 2.9 and 2.10. One may question the utility of these
reactor equations. Do they have predictive value and can they be used to
design and operate an activated sludge process? These equations, with-
out consideration of ky, were developed many years ago to describe the
growth of various species of microorganisms. Almost as many years ago,
they were employed in experimental laboratory activated sludge pro-
cesses to determine their suitability for mixed cultures and substrate
 PRINCIPLES OF BIOREACTOR MODELING 37

systems. It was found that once-through reactor (chemostat) behavior

was predicted rather well by Equations 2.8 and 2.9 using values of the
biokinetic constants determined from independent batch tests of the sys-
tem. For the recycle system, Equations 2.19 and 2.20 were found to be
acceptable, but they were not readily applicable in view of accepted
engineering practice in the pollution control field. One could easily
employ a, F,/F, as a control because it was easily measured and pumps
are rather easily controlled. Thus, a desired a could be selected and
maintained at that value. However, the concentration factor c is not so
selectable, nor does Xp remain constant. Most clarifier underflow con-
centrations vary somewhat because of changes in sludge settling charac-
teristics. Also, Xp is a variable depending on F and Fx. In practice, it was
found that maintaining c at some selected value was an extremely diffi-
cult task. It was also found that the steady-state condition of X was
subject to some variations because slight changes in Y,, K,, K;, max, and
k, were exaggerated due to the requirement to hold c constant. All these
factors led to the conclusion that c was not a very good engineering
control variable. In order to accommodate the equation to engineering
reality, Xp, rather than c, was used along with a and D, i.e., Equation
2.18 was rewritten:

w= D(I ta-aQ) +k (2.21)

This change does not fundamentally modify the theory of continuous

culture as presented in this chapter, but it does have a fundamental effect
on the form of the final engineering equations. One cannot simply substi-
tute an X,/X term for c in Equations 2.19 and 2.20. It is necessary to
make this substitution in Equation 2.16 and then to proceed with the
simultaneous solution of Equations 2.16 and 2.17. The derivation is
presented in Chapter 3, leading to the development of engineering equa-
tions. There it will be shown that in addition to a, Xp, D (or t), ky, Y,,
Mmax» K,, and K,, the influent substrate concentration S; also influences pu
and thus S and X. The final engineering equations relate all of the
important factors governing the quality of the effluent. They have been
extensively tested and, as will be shown in case studies presented in
Chapter 7, found to be of excellent predictive value for large-scale waste-
water treatment facilities.
38 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

KEY CONCEPT SUMMARY

The key concepts which are contained in this chapter are:

e In addition to the fundamental definitions and relationships

between growth, substrate removal rate, and respiration (O,
uptake) that govern all aerobic microbial growth, presented in
Chapter 1, it is also important to note that the reactor and/or
reactor configuration has a significant impact on growth
kinetics.
e¢ The important reactor concepts are the hydraulic control of u
(Equation 2.4) and the additional hydraulic control of » by bio-
mass recycle (Equation 2.18).
¢ The recycle sludge concentration, Xp, rather than c (Xp/X), is the
preferred control parameter for field systems. Equation 2.21 is
vital to the development of engineering equations, which are
derived and discussed in Chapter 3.

REFERENCES AND SUGGESTED

ADDITIONAL READING

Gaudy, A.F. Jr., Ramanathan, M. and Rao, B.S. (1967). “Kinetic Behav-
ior of Heterogeneous Populations in Completely Mixed Reactors,”
Biotech. Bioeng., 9, pp. 387-411.
Gaudy, A.F. Jr., and Srinivasaragahavan, R. (1974). “Experimental
Studies on a Kinetic Model for Design and Operation of Activated
Sludge Processes,” Biotech. Bioeng., 16, pp. 723-728.
Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental
Engineering, Engineering Press, Inc., San Jose, California.
Ramanathan, M., and Gaudy, A.F. Jr. (1969). “Effect of High Substrate
Concentration and Cell Feedback on Kinetic Behavior of Heteroge-
neous Populations in Completely Mixed Systems,” Biotech. Bioeng.,
11, pp. 207-237.
Rozich, A.F., and Gaudy, A.F. Jr. (1986). “Process Technology for the
Biological Treatment of Toxic Organic Wastes,” Hazardous and
Industrial Waste Testing and Disposal, 6, ASTM STP 933, pp.
319-333.
Srinivasaragahavan, R., and Gaudy, A.F. Jr. (1975). “Operational Per-
formance of an Activated Sludge Process with Constant Sludge Feed-
back,” J. Water Poll. Control Fed., 47, pp. 1946-1960.
 3 ENGINEERING
MODELS FOR ACTIVATED
SLUDGE SYSTEMS

INTRODUCTION

Predictive models for activated sludge systems are derived using the
same methodology detailed in Chapter 2 regarding the theory of continu-
ous culture. This is essentially a reactor engineering approach. The basic
procedure is to write mass balance equations for biomass and substrate
around the reactor. An appropriate expression for relating growth rate to
substrate concentration is then inserted into the mass balance equations,
which then are solved simultaneously to obtain predictors for X and S,
the reactor biomass and substrate concentrations, respectively. This same
procedure can be used to obtain equations for any configuration of an
activated sludge system. In this chapter, we will present the derivation of
predictive equations for an activated sludge system with one completely
mixed reactor. We will also present the derivation of predictive equations
for an activated sludge system with multiple reactors.
It is important to emphasize that the treatment performance of an
activated sludge system is largely controlled via specific growth rate.
Growth rate (and its related environmental counterparts, wc, F:M, etc.) is
in turn determined by engineering controls such as flow rate or detention
time, recycle sludge concentration, and recycle flow rate. The influent
waste concentration also exerts an influence on reactor growth rate. The
activated sludge models are governed by two main sets of constants or
controls: biokinetic and engineering. The biokinetic growth constants
(those found in Equations 1.11 and 1.12) quantify the capability of a
reactor’s biomass to degrade a target waste. The engineering constants
are physical components that the designer or operator can control.

oy
40 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

(Influent waste strength can be controlled with techniques such as equali-

zation.) The model’s purpose is to link the biokinetic constants, which
are largely a function of environmental conditions, with the engineering
constants, which are under the control of the environmental technolo-
gist. If the waste is difficult to degrade or inhibitory (as indicated by the
values of the biokinetic constants), the model will quantify those values
of the engineering parameters which are needed to maintain reliable
process performance for the target treatment condition.
This chapter contains three main sections. The first section deals with
the derivation of the predictive equations for activated sludge systems.
Models are presented for both noninhibitory (Monod function) and
inhibitory (Haldane function) wastes. The second section describes the
derivation of predictive equations for multiple-reactor (tanks-in-series)
systems. The purpose of this section is to demonstrate that the modeling
approach presented herein is flexible and that deriving a model for a
specific system or reactor configuration is simply a matter of following
the methodology given in this chapter. The last section concerns critical
point analysis for treatment of inhibitory wastes. Many toxic and haz-
ardous wastes are characterized by inhibition kinetics. As noted in Chap-
ter 2, wastes that cause an inhibition response from the biomass require
special consideration when designing and operating a bioreactor. This is
primarily due to the potential to exceed initial operating conditions,
which leads to process failures for biological systems treating inhibitory
wastes. Quantitative methods to calculate the operational location of
critical operating points are presented in the third section of this
chapter.

DERIVATION OF PREDICTIVE EQUATIONS

Noninhibitory Wastes

A flow sheet of an activated sludge system with one completely mixed

reactor is given in Figure 3.1. It is assumed that the reactor maintains
aerobic conditions and that recycle sludge concentration, X, has a signif-
icant impact on reactor performance. Xx, can be held constant or can be
controlled at a value selected by the engineer. Although this is rarely
performed in field practice, it needs to be emphasized that this parameter
imparts a substantial impact on reactor growth rate. A highly fluctuating
recycle sludge concentration will lead to process upsets in difficult treat-
 ENGINEERING MODELS 41

Bloreactor
(1+ 0)F,S,X (F - F,,)S,X
Clarifier = :

Recycle Cell
Source

Figure 3.1. Flow diagram used for developing model equations.

ment situations unless other compensatory measures (e.g., long reactor

detention times) are implemented.
The appropriate mass balance equations for biomass, X, and soluble
substrate, S, are given below:

dx
VS = FX, + pXV-K,XV-(1 + a)EX (3.1)

Vvs 2 FS, PieFS,t- (et Ree asia (3.2)

Since in the steady state, dX/dt and dS/dt = 0, an algebraic solution is

possible, and using t = V/F, Equation 3.1 can be solved for pu.

p= Dl +a-ar) + ky (G33)

Also note that since

=o Pathe Ka (3.4)

m= D(lta-aX ipso)
42 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Assuming steady-state conditions, Equation 3.2 can be solved for X:

Y, [DS,
X.. 25 Fi + aDS, - (i + a)DS,]. (3.6)

To obtain a predictive equation for S, Equation 3.6 is substituted into

Equation 3.3:

w = [1 + &)D + ki- HaXp

Y (S, + aS, - (1 + a)S,)
(3.7)

Factoring out p,

ruse aXrY;
= [7 + a)D + k,] (3.8)
i Suictheic Spe d lanka )Ss

To obtain a predictive equation for noninhibitory substrates, substitute

the Monod equation (Equation 1.11) in place of p:

SR aK max Se

et Sekt gS
4 (max
Se XR
S. °K Se7Ki! YY,
=((1 + a)D + k,J[S; + aS8,]
-[(1"+_q)Di+ kal +:0)8, (3.9)

After collecting terms, the predictive equation for S is:

-b + (b? - 4ac)'?
Ss. = eee (3.10)

a= ((() 4°o)D +\kal = idl + @)

b ae mS a Sr a 7)

+ (1 + aD + k,)[(@ + a)K, - (GS, + aS,)]

c = -(S, + aS,)[((1 + a)D + k,IK,

A predictive equation for X, the reactor biomass concentration, is gener-

ated by combining Equations 3.3 and 3.6:
 ENGINEERING MODELS 43

Se Y, (S; + aS, - (1+a)S)D

(11)
He1 Ua rae x Saas

Y, (S; + aSg - (1 +.a)S)

(3.12)
X
Leelee
ky/D

(1 + a + k,/D)X = Y, (S, + aS, - (1 + a@)S) + aX, (3.13)

_ Y, (S + aS, - (1 + a)S) + aXe (3.14)

« 1+a+k,/D

An equation for waste sludge production, X,,, is derived by recogniz-

ing that the amount of sludge that is wasted (or produced) is the differ-
ence between the amount that exits the reactor and the amount that
enters the reactor:

X, = (1 + a)FX - oF X, (3.15)
Equation 3.15 can be rewritten as follows:

eres
Nenaaon aw Xp
sabe (3.16)

X, =| aXp
tia-
SAKV (3.17)

Xb] XV (3.18)
Equation 3.18 will predict the amount of excess sludge that will be pro-
duced by the activated sludge system.

Inhibitory Wastes

Predictive equations for inhibitory substrates such as those that are

often characteristic of toxic or hazardous wastes are derived by substitut-
ing the Haldane equation, in lieu of the Monod equation, into Equation
3.8. After collecting terms, the predictive equation for S, or effluent
quality, is obtained:
44 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

aS} + bS2 + cS. + d = 0

_ (4 a) D+ kl + 2)
a K,

b ier De ea ceed sraK (3.19)

s Heokt a a)

CSL aD, Palell aeaikeadseteaso

ue. (srose 4 = '
d = -[(1 + a)D + k,J[S, + oS, ]K,

Although Equation 3.19 is a cubic, it is easily solved with a numerical

method such as a Newton-Raphson iteration technique. The predictive
equations for X and X,, the reactor biomass concentration and waste
sludge production rate, respectively, are the same for both noninhibitory
and inhibitory wastes.
Equations for both the noninhibitory and inhibitory wastes are sum-
marized in Table 3.1. It is evident in this table that the only difference
between the noninhibitory and inhibitory models is the inhibition con-
stant, K;. It is interesting to note that, as the inhibition constant tends to
go to infinity, which is indicative of a more noninhibitory waste, the
inhibitory equations reduce to the noninhibitory form. Thus, one can
view the inhibitory predictive equations as addressing a more general
case for waste treatment, since they incorporate the impact of waste, or
substrate, inhibition on process performance.
As an example of the application of the predictive equations, consider
Figure 3.2. This figure shows their predictive curves for effluent quality
S and reactor biomass concentration X for an activated sludge reactor
treating phenol. In this case, both the noninhibitory and inhibitory
models were utilized to make predictions of effluent quality. The output
from the modeling procedure is a prediction of effluent quality. The
input to the model consists of the values of the biokinetic constants and
the values of the engineering control parameters, reactor detention time
(or influent flow rate), recycle sludge concentration, and recycle flow
ratio. Influent waste strength, or concentration, is also an input parame-
ter. The curves in Figure 3.2 are generated using the equations from
Table 3.1 and provide predictions of effluent quality in terms of soluble
biodegradable COD for the specified values of the biokinetic and engi-
neering constants and the influent waste strength.
45

}dwo09 Ul § = °¢ :010N
‘W/] = q ‘sulaishs poxtwl Aja}o
MODELS

“AXA = ™X | (gre) “AXA = “X

(81'€)
iif
(Z1€) ane as = x
Yyn+ [Po(o + 1)-Csn +'s)A
ENGINEERING

GPU ee al Yes SyPy+a+D] Cse+'s)- =P

ere
oS Fyos Posty -'sot'slA
[gn +!s)- SH( + DI) (

hae
+'g)
go
h
[Py+a(+1)] + dy

a
DIPA+a@+

Sy[Py + a(o + DP] (sx +'s) -

+
oO
i
shee
Mes
SH@+

(Py+a(™+D] = 49
[@s2

= oH)

1
seuin(o +1)-(
+'s)

=
DI

Ce
Kv
+
-

DI(@+1)
>.
(Py+

art
(+ Samual +(e + i}

ll
Os)
eZ 5
(61'€) +
0=Pt°so+,"S4,’SP
ORY erore
(auepleH) ayB.sqns ALOVIQIyUT
(pouoyy) = Vsqng = —AOJIQuUTUON
FGVL
SUIa}SAS aSPHNIS paywanoy 10J suonenby eHopad 3}8)S-APBIS T°
46 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

1200 Non-Inhibitory 800

600 400 =

©) Se
= dp)

< 1800 1200

1200 800

4
600 400
Inhibitory

T T T IF 7 T i pa
oo 5.0 2.5 1.67 1.25 1.25
Detention Time, hours

Figure 3.2. Comparison of predictive curves for effluent quality

calculated from the Monod and Haldane equations for a
constant X, activated sludge process treating phenol at an
S,; of 1500 mg/L. Values of parameters and biokinetic
constants are: Xz = 6,000 mg/L, a = 0.25, p,,,, = 0.19
hr', K, = 48 mg/L, K, = 62 mg/L, Y, = 1.02, and k, =
0.47 days".

Figure 3.2 also makes an important point regarding the use of the
inhibitory model. It is clear that the inhibitory model for phenol predicts
a sudden decrease in effluent quality once a certain operating condition
is attained. This condition corresponds to the peak of the Haldane equa-
tion as discussed in Chapter 2 and is not predicted with the noninhibitory
model. Extensive testing of the model has shown that the behavior of
activated sludge systems treating inhibitory wastes, i.e., those wastes that
are characterized with Haldane kinetics, is accurately predicted by this
model. It is clear in Figure 3.2 that the Monod-based model predicts a
gradual loss in performance capacity. Consequently, failure to account
 ENGINEERING MODELS 47

for the impact of the inhibition kinetics on the design and operation of
the biological treatment system leads to a gross overestimate of the effec-
tive operating range. The bottom line is that activated sludge systems
that are dealing with inhibitory wastes must be modeled using an algo-
rithm that accounts for the impact of inhibition. The alternative is to
design or operate the system using an extremely conservative approach
(oversized aeration basins) or to risk system failure by exceeding the
system’s critical operating point.

APPLICATION FOR MULTIPLE REACTOR SYSTEMS

Many activated sludge systems are single, completely mixed basins.

However, other systems also use multiple reactors or other different
configurations. The objective of this section is to demonstrate the flexi-
bility of the modeling approach presented herein for addressing different
reactor configurations. That is, if a system is different than complete
mix, one can derive the equations that apply for that system. As an
example, we present the derivation of equations for a multiple-cell acti-
vated sludge system consisting of “m” number of cells as depicted in
Figure 3.3. The approach is the same as that used for the single cell
system except that multiple reactors now must be modeled.
The multi-reactor model is derived by writing mass balance equations
for biomass and substrate concentrations for each cell. For the first cell,
the mass balance equations are:

(1 +0) F,S,X

Settler

Figure 3.3. Flow schematic of multi-cell activated sludge system.

48 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

= “io = OFX, + w(I)X(1)V— - kyX(I)—Vv - (1 + a)FX(1)— G.20)

Vo dS _ Fs, + ofS, + ars) - AP xX @.21

At steady state, the time derivatives equal zero, and Equations 3.20 and
3.21 become:

0 = amDX, = p(1)X(1) - kgX(1) - (1 + @&)mDX(1) (3.22)

0 = mDS, + amD§8, - (1 + a)mDS(1) - OA (3.23)

In Equations 3.22 and 3.23, m represents the number of completely

mixed cells in the aeration tank and V is the total aeration tank volume.
By utilizing equally sized cells, the volume of each cell is V/m. X,,) and
Sq are the biomass and substrate concentrations in reactor 1, respec-
tively; a is the recycle flow ratio; D is the total tank dilution rate time, F/
V; Xp is the recycle sludge concentration; y,) is the specific growth rate in
reactor 1; kg is the specific decay rate; S; is the influent substrate concen-
tration to the tank; S, is the substrate concentration in the recycle line
(usually assumed to equal zero); and Y, is the true cell yield.
Steady-state solutions are obtained by utilizing an appropriate func-
tion for relating » to S and then solving Equations 3.22 and 3.23 for X,)
and S,,. As done previously, the Monod equation is used for noninhibi-
tory wastes and the Haldane equation is employed for toxic or inhibitory
wastes. The steady-state predictive equations for reactor | for both non-
inhibitory and inhibitory wastes are given in Table 3.2.
Mass balance equations for the other cells in the multiple-cell aeration
tank, i.e., reactors 2 through m, are given in Equations 3.24 and 3.25:

¥V oa
dxGj = (1 + a)FXG-1)
:
+ W@)XG)—
Rear "2
- kX: - (1 + a)FX(j) (3.24)
V dS(j
mode = (1 + aFSG-1) - (1 + aFsg)» - _XO
#G) X@) ¥V G25)

It should be noted that Equations 3.24 and 3.25 apply for cells 2
through m provided that X;_,, and S,,_ , are already calculated. The
steady-state versions of these equations are given in Equations 3.26 and
S27:
 gelZ¢ 9}81S-ApbajsPANIIPIIg suo
nen
10; by PIBAHIYaspnis
AouIqIyUTUON S9119§-UJ-S[
UID}[aQ
SAS
SaJB.ISqns (pouoy;)
1O}o
[ evay AJOVQiyUy Sa}JBISGNS
(euepyeR)
lo1
I eay
(Ds= qeF
e &G=) wep
: (&Z|}
'¢)
B® (1)+; q (De=:S £:.
PP)So O=
Lt
se 1) + (ame DE OF
XPul. ® = +p) qua+e C1
a >I

a SSGs
)
+
q = 1)) + aqu+i DOA~ 4
7)) + qui+(o IC+ pa
= 1)
s so
Sy(- o 'S) + ("Ss

= 's)-+ sft
1))Cgn quia(* CPx
2) RANG] 9)
4 af
(%4g70 es
aL

1) + qui+(o IY+ Sy- 's) + (Cg

p = 's)-+ D)C+sn quo
ENGINEERING

“(P+
OX= BR4 go +) HDS Xe o9e|'9 (O=
- yt x STE RT
| D+
+ - Py
ROO
E
;
M ODELS
49
 7°¢
Sa}Bysqng (pouoyA)
ALOJIQIyUTUON A1ojiqryuy sayeysqng (auBpleH)
50

aque
panuyuos
7YSNoIy}
ZYsnoly)
$10}OeOY
S10jDvVIY
W
WI

(Ds= eT (4l7'€)
| OS®
+ (S94
+ (Oso 0=P+ (PLZ'€)

1) + que
+ (4
De GUS BOS eS aa

q = wen 1-Ds}
= ae
+t =q 1)) + qu(o
+ Jes ee
= me
1) + +
aur -
“DCX ((1-Ds

(I-Ds'uy
(I-Ds
"MCX
Gu
au
1))-
=9
-=+91))

ah
=P 1))-+ Guo
+ (I-OSHC}

Gee
axe
aan
=en
|
(q92"€)
(DX
(492°€)
(9x
=©SB
qu+ (0 + Py eae (Os + (0 + Py
[qui sa
ed
4). Se
ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

“XK= I)A+ (w)x(?

- (*x~ (097'¢)
| "x = Dd + (W)xX(°
- (x (997"€)
‘dj, = 1:010N
 ENGINEERING MODELS 51

0-= (1 + a)mDXG-1) + nGj)XG) - k.X(j) - (L-#e)mDXG)G.26)

0 = (na)mBS(ia1) St! +-a)mDSG) - LOX) (3.27)
A solution technique analogous to that presented for the single, com-
pletely mixed reactor system is employed to obtain steady-state predictive
equations for X, and S,). These predictive equations are also given for
noninhibitory and inhibitory wastes in Table 3.2. The equations in this
table are applicable to an activated sludge system with m number of
equal volume reactors. They also cover the case of a single, completely
mixed reactor; i.e., when m = 1, the equations are the same as those
presented in Table 3.1. Finally, a predictive equation is presented for
waste sludge production in the multiple-cell system; this is derived by
performing a mass balance for biomass production in the aeration tank.

CRITICAL POINT ANALYSIS FOR TREATMENT OF

INHIBITORY WASTES

The significance of the peak of the Haldane curve or p* in biological

reactors treating toxic or inhibitory wastes is simply that, once the reac-
tor attains this growth rate, it is subject to sudden effluent deterioration
and washout. This feature of inhibitory waste treatment was demon-
strated analytically in Figure 3.2 using the model equations for an acti-
vated sludge system (single, completely mixed reactor) treating phenol. It
was also noted in Chapter 2 that chemostats treating toxics wash out
once the detention time produces a growth rate corresponding to p* in
the chemostat. These features of inhibitory waste treatment make it
imperative to operate reactors at growth rates that do not approach p*.
It needs to be stressed that the phenomenon of reactor failure at »* has
been demonstrated often in bench-scale pilot plant systems. Figure 3.4
provides a dramatic illustration of the consequences of exceeding p* in
an activated sludge reactor treating the toxic organic phenol. In this
activated sludge system, the influent waste strength S; was 2000 mg/L
phenol, and it can be seen that, once »* was exceeded, effluent quality
rapidly deteriorated and recovery was not possible.
It should be noted that for this activated sludge pilot plant, conditions
for reactor failure were predicted using biokinetic constants determined
from batch tests that used the reactor biomass and the influent phenol
waste. Consequently, it is feasible to formulate operating policies that
52 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

2500

2000

O X, Reactor Biomass
DO Se, Reactor (COD)
mg/|
Indicated,
Analysis V Se, Reactor (Phenol)

0 4P 1TA>~1P 3P 5P OAc 9P

1/29/82 ——>x——— 1/30/82

Time, hours

Figure 3.4. Consequences of exceeding p* in activated sludge reactor

treating phenol.

permit maximum treatment efficiency while avoiding reactor failure

because of operation near p*.
Another way to formulate operating strategies for activated sludge
system treating toxics is to quantify the operating conditions, or critical
operating points, that result in a growth rate of u* in an activated sludge
reactor. Equations that quantify the critical operating point are derived
by considering the activated sludge process flow diagram shown in Fig-
ure 3.1. Mass balances for X and S, the reactor biomass and substrate
concentrations, respectively, are written around the single, completely
mixed reactor.

V ox GFK UNV SERV Le BEX (3.1)

V s ig ost camer sheet eva ue (3.2)

 ENGINEERING MODELS 33

By assuming steady-state conditions and by letting D = F/V, the

nominal dilution rate, Equations 3.28 and 3.29 are obtained.

Oss aD Xu ko dee X=, (1 Hh. a) DX (3.28)

0 = DS, + aDS, - (1 + a)DS - uxY (3.29)

t

Equations 3.28 and 3.29 are manipulated to yield expressions for » and
X.

Xoie adie (3.30)

a
hw=D(+a-

y,
X = 7 (DS, + aDSy - (1 + a)DSJ. (3.6)

An expression for the critical detention time, t*, which produces a

growth rate equal to y* in the activated sludge reactor, is obtained by
combining Equations 3.6, 1.13, 1.14, and 3.30. This expression for t* is
given in Equation 3.31.

aXp — 1

Umax Y,
t* = (1 + a) he eI k, | 6.31)
Ks S; + aSp
- (1 + a) VK,K;
1+ 2\ K,

It should be noted that Equation 3.31 gives a value for the detention
time at which the reactor attains the critical growth rate, p*, i.e., the
detention time at which one can expect rapid effluent deterioration for
an activated sludge system treating an inhibitory waste. Three engineer-
ing control variables, t (or flow rate), a, and Xp, can be selected by the
designer and operator; S,, the influent waste strength, can often be held
relatively constant using equalization techniques. The bottom line is that
by determining the value of the critical growth rate, yu*, via biokinetic
testing using respirometric methods, one can use the model to quantify
the values of the engineering controls needed to avoid operation near the
critical operating point. Design and operational strategies for avoiding
critical point operation can be illustrated with the aid of Equation 3.31
and the use of critical point curves.
54 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Critical Point Curves

A critical point curve is essentially a graphical technique that quanti-

fies the critical operating point, i.e., the values of the engineering con-
stants at which the reactor attains a growth rate equal to p*, the peak
growth rate given by the inhibition function. These curves can be gener-
ated using Equation 3.31. For example, critical detention time t* is plot-
ted versus influent substrate concentration, S,. Consider the graph
depicted in Figure 3.5. The solid line represents the values of t that result
in a growth rate of »* for specified values of the biokinetic constants, the
engineering control variables, and Xx. The cross-hatched area above the
curve comprises a safe operating region where the reactor is not apt to
experience sudden effluent deterioration and washout. Conversely, in the
area below the curve in Figure 3.5, reactor substrate would be higher
than S* because the reactor would be operated on the downward side of
the inhibition function, which “exceeds” »*. Operation well above the
curve is necessary in order to prevent failure.
The use of critical point curves to avoid design or operation near p* for
activated sludge reactors treating inhibitory wastes is demonstrated using

0
~——

si

Figure 3.5. Critical point curve for activated sludge reactor treating an
inhibitory waste. The curve is defined by substitution of
appropriate biokinetic constants and selected values of a
and X, into Eq. 3.31.
 ENGINEERING MODELS 55

the curves presented in Figure 3.6. These curves were constructed by

utilizing values of the biokinetic constants for phenol. The values of
these constants are the same as those used to generate the curves shown
in Figure 3.2. For Figure 3.6, Equation 3.31 was used, and it was
assumed that substrate concentration in the recycle flow is negligible,
i.e., Sp = 0. All curves were generated using a recycle ratio of 0.25; the
recycle sludge concentration X, varied from 6000 to 30,000 mg/L. It can
be seen that increasing X, values decrease the detention time required to
prevent reactor failure due to p* violations. The reason for this trend is
simply that an increase in Xp will reduce specific growth rate in the

Curve No, Xa. mg/l

6000
8000
10000
12000
16.0 15000
20000
Onkhwah—
30000

12.0
72)
=
3
fs
bei 8.0

4.0

0 1000 2000 3000 4000 5000

Si, mg/I

Figure 3.6. Critical point curves generated using Equation 3.31

showing effects of X, on location of critical operating
point. Biokinetic constants are the same as those for
Figure 3.2. a = 0.25 for all curves. (Adapted from Rozich
and Gaudy, 1984).
56 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

reactor (increase 9.), provided the other engineering and biokinetic con-
stants remain relatively unchanged.
It should be also noted that Figure 3.6 depicts various operational
options that can be implemented in order to avoid a reactor washout. As
an example of the utility of critical point curves, consider Figure 3.6 and
an activated sludge reactor operating at an X, of 6000 mg/L, a of 0.25,
S; of 1000 mg/L, and detention time of 10 hr. If it were proposed to
increase the influent waste concentration S,; to 2000 mg/L, or expected
that in time it could reach this concentration, then the system would be
thrust beyond the critical operating point and would inevitably experi-
ence sudden washout. However, a number of options are available that
can be implemented to prevent reactor failure. For example, a design
decision to increase reactor detention time to approximately 13 hr would
avoid operation in the proximity of the critical point. It can be seen in
Figure 3.6 that this change would place the reactor operating point above
the 6000 mg/L X, line, which comprises a safe operating region. A safe
operating condition can also be provided by manipulating the recycle
sludge concentration, Xp, while maintaining reactor detention time at 10
hr. If X, is increased to 10,000 mg/L, then the critical operating point
curve will be shifted to the 10,000-mg/L line. At this X, and t of 10 hr,
the reactor will be in a safe operating region because it will be above the
critical point curve.
It should be noted that the recycle flow ratio a can also be used for
averting critical operating conditions in the aeration tank. Consider the
set of critical point curves depicted in Figure 3.7. These plots illustrate
the effect of a on the critical curve for this reactor system. Recalling the
previous example depicted in Figure 3.6, it can be seen that at an a of
0.25, an Xz Of 6000 mg/L, a detention time of 10 hr, and an §S, of 1000
mg/L, the reactor is in a safe operating region above the curve. However,
if S, changes to 2000 mg/L, the system is placed below the critical curve,
meaning that the system is vulnerable to sudden effluent deterioration.
Figure 3.7 depicts a number of engineering changes that could be imple-
mented to alleviate a washout condition. For example, a change in a
from 0.25 to 0.50 would realize a safe operating condition in the reactor.
Also, provision of more aeration volume, i.e., an increase in reactor
detention time to about 13 hr, would put the system above the critical
point curve. Thus, a change in a, Xp, or t can be used to prevent failure
under the new imposed loading condition.
Once a treatment plant is on line, the only manipulative alternatives
available to the operator are changes in recycle sludge concentration Xz,
and recycle flow ratio a, provided these alternatives have been made
 ENGINEERING MODELS 57

16.0

4.0

0 1000 2000 3000 4000 5000

Si, mg/

Figure 3.7. Critical point curve showing effects of a on location of

critical operating point. X, = 6000 mg/L, a = 0.10, 0.25,
0.50, 1.0. Same biokinetic constants as Figure 3.6.
(Adapted from Rozich and Gaudy, 1984).

available by the designer. There may be a tendency in design to obtain the

pw, or 8, needed to produce the required effluent substrate concentration
by holding the reactor volume, and thus t, at a low value, depending on
the attainment of high X, or a, or both. For example, for an S; of 1000
mg/L, if one wished to employ a 3-hr detention time and an a of 0.25,
the recycle sludge concentration to avoid washout would be 20,000 mg/L
according to the example in Figure 3.6; this would just keep the process
above the critical curve and thickening would probably be required.
Conversely, if one took a somewhat conservative estimate of the concen-
tration of Xz, that could be depended upon due to compaction in the
clarifier, i.e., 6000 mg/L, a would need to be increased above 1.0 to
avoid failure of the system. However, recycle rates above 1.0 impose a
higher solids flux loading to the secondary clarifier, which means that
58 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

underflow concentrations of 6000 mg/L might not even be attainable. It

is thus argued (especially in the treatment of toxic or inhibitory wastes
where the plant needs to be protected against sudden failure, which
occurs as the reactor approaches p*) that it is wise to allow ample deten-
tion time or reactor volume. Such a provision, coupled with provision
for flexibility in recycle pumping capacity and measures for positive
control of Xz, give the operator the ability to keep the process on line.
It should be emphasized that manipulations of Xz, a, and t discussed
here are related to the washout or failure curves for the plant. That is, the
analysis presented herein is related to the critical reactor condition for pu
and S. For the example used here, the critical condition corresponds to
ph S20.07hreyS*x=05 5ane/Ly org =—si2 dl, GA= iss d. Fheseare
much higher values of yz, and S than those for which the treatment plant
would be designed, so that an average plant condition should be well
above the critical curve. However, when one takes into consideration
uncertainty factors such as possible changes in the numerical value of the
biokinetic constants, inflow rate, and waste strength, and the general
lack of control of a and Xx, there is ample reason to incorporate as much
of a safety factor as possible. In design, this usually means the use of
additional hydraulic detention time. This seems especially important for
the case of inhibitory wastes as compared to noninhibitory wastes. In the
latter case, an occasional period of increase in soluble substrate in the
effluent might be tolerated or a fine paid for a permit limit excursion,
but with inhibitory or toxic wastes, the consequences can be total failure
or washout of the system, which would result in an extended period of
permit violation and corresponding fines. The analysis and the critical
curves presented here can be looked upon as a form of “reliability analy-
sis” to prevent total process failure, which would require a shutdown and
restart of the plant.

KEY CONCEPT SUMMARY

The key concepts contained in this chapter are:

¢ The derivation of predictive equations for activated sludge pro-

cesses involves a mass balance approach. Mass balances for bio-
mass and substrate are written around the reactor. An appropri-
ate growth rate expression (Monod or Haldane) is used to relate
to S. The equations are then solved to provide predictive equa-
 ENGINEERING MODELS 59

tions for effluent quality S and reactor biomass concentration

X.
¢ The modeling procedure described in this chapter can be applied
to any reactor configuration to yield configuration-specific pre-
dictive equations.
¢ The operational location of a reactor’s critical point, i.e., the
point at which an activated sludge reactor treating an inhibitory
waste attains the peak growth rate, y*, is quantifiable in terms of
the engineering control parameters and the influent waste con-
centration. The engineering controls are then used to avoid oper-
ation near the critical point and provide input regarding a safe
operating cushion using the process model.

REFERENCES AND SUGGESTED

ADDITIONAL READING

Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental

Engineering. Engineering Press, Inc., San Jose, California.
Rozich, A.F., Gaudy, A.F. Jr., and D’Adamo, P.C. (1983) “Predictive
Model for Treatment of Phenolic Wastes by Activated Sludge.” Water
Research, 17, 1453-1466.
Rozich, A.F., and Gaudy, A.F. Jr. (1984). “Critical Point Analysis for
Toxic Waste Treatment,” J. Environ. Eng. Div., ASCE, 110, pp.
562-572.
Rozich, A.F., and Gaudy, A.F., Jr. (1985). Response of Phenol Accli-
mated Sludge to Quantitative Shock Loading, J. Water Poll. Control
Fed., 57, pp. 795-804.
Rozich, A.F., Gaudy, A.F., Jr., and D’Adamo, P.C. (1985). “Selection
of Growth Rate Model for Activated Sludges Treating Phenol,” Water
Research, 19, pp. 481-490.
Rozich, A.F. and Gaudy, A.F. Jr. (1986) “Process Technology for the
Biological Treatment of Toxic Organic Wastes,” Hazardous and
Industrial Waste Testing and Disposal, 6, ASTM STP 933, pp.
319-333.
 . ;
white sais: > oY see

fi Wz : ee!
neliniersal? 6. mals
. o aetna eten.

‘ewy dena oe Poh suine Sich 6 ™ ve itt

oe
sh kre mad “em

“p= OTe" =yelohom sean pp

Mich hye eed AP wa, dae Bhat hore fos
‘wud be eo ighanty ac (arb
2m wk Ro
2Qomnie fe 7 tou Up =
é are sous 9a ae ah a

? — he;
ahaa Wivir:
eo Wynd. haeog
wee ies init

ong are: ieee

vy THU Vig ie nA é (a

; ~snrtes Innip Rees

3) mise” aa "EY amen" ry + ny

se) ” logelS guitsott Pals, hayeviior, mw) Inbold
2
vert
ge Ot na dontegtrena fay
OM vaoleinsnll "laseeW slenegiOe At AA, buat
el TA MTEO ApS NR
“rel a
to a
= ag antag at
uy) i ol
ae e) 7
TRter ina! typed sre UCT
i vi) taY adding Pie
oni ile atertagh, Mas
eh oF 0° Wr tal GrOt A |re’ auant ct bn:
nei? ge lye: Mmres-inn >
{ Pano” feeFp
as Rp
ew t
fz uO = Tih Have prinylite ge
 4 COMPARISON WITH
OTHER APPROACHES AND
BASIC APPLICATIONS

INTRODUCTION

The purpose of this chapter is to compare and to reconcile the acti-

vated sludge modeling approach presented herein with other more “con-
ventional” methods or parameters often used to analyze these systems.
This chapter also illustrates some basic applications concerning the use
of the modeling approach to understand and to analyze various ancillary
features that are often associated with the design and operation of acti-
vated sludge systems, such as:

e Impact of recycle parameters on system growth rate.

e Determination of conditions needed for minimizing sludge
production.
¢ Computation of oxygen transfer requirements.

The objective of this chapter is to demonstrate to the reader that the

modeling approach presented herein (mostly found in Chapter 3) will
provide analysts of activated sludge reactors with essentially the same
results as the conventional methods. However, because this approach
contains a higher degree of flexibility, it enables environmental technolo-
gists to identify the limits of the system operating envelope more accu-
rately, which enables them to “fine tune” the design and operation of
activated sludge systems with a more structured, quantitative approach.

61
62 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

RECONCILIATION WITH OTHER DESIGN AND

OPERATIONAL APPROACHES

The design and operational methodology presented here represents

methodology that is not often used to analyze activated sludge systems.
Nevertheless, while the approach is different, it is prudent engineering
practice to evaluate how it compares with approaches that employ “food-
to-microorganism” (F:M) ratio, substrate utilization rate U, and mean
cell residence time ©, (which is simply the reciprocal of the net growth
rate »,). F:M and U are computed using Equations 4.1 and 4.2:

F:M = S,/Xt (4.1)

U = (S,- S)/Xt (4.2)
Inspection of these equations shows that these two parameters are
essentially the same, since most designs aim for low effluent substrate, or
waste, concentrations where S, >> S.
Mean cell residence time or ©, is defined by the mass of cells in the
system divided by the rate of cell wasting. It is computed using Equation
ALS:

Geen (4.3)
However, the amount of biomass wastage, X,, is also defined as the
amount in excess of the amount needed for recycle:

xX, = (1 + a)FX - aFX, (3,45)

This equation can be rewritten similar to equation 3.17:

Xw = (1 + a - aXg/X)VX/t (3.17)

From Equations 3.18 and 4.3, it is known that:

Pov a Sows NX = 178, (4.4)

It is thus clear that:

By = X,/VX = (1 + @ - aX,p/X)/t = 1/8, (4.5)

Equation 4.5 makes several important points. First, it analytically dem-

onstrates that ©, and p, are reciprocals of one another; controlling one
 COMPARISONS AND BASIC APPLICATIONS 63

controls the other. Secondly, it shows that there are two ways in which ®,
or pw, can be calculated: (1) a feed forward mode that utilizes the recycle
parameters or (2) a resultant mode using the wasting rate. In a steady-
state situation, both methods will produce the same result. Finally, Equa-
tion 4.5 shows that the recycle parameters, a and Xp, exert a strong
influence over ©, and y,; this becomes extremely important especially
when considering the treatment of toxics or difficult-to-degrade wastes.
By using Equation 3.6 from Chapter 3 and the definition of U as being
the rate of waste utilization (F(S; - (1 + a)S) divided by the system
biomass, VX, it can be shown that:

2 = YU (4.6)
Using Equation 4.6 and Equation 3.4 from Chapter 3, it can also be
shown that:

u, = YU - ky = 1/6, (4.7)
Equations 4.5 and 4.7 quantify the relationship between the engineering
control parameters, recycle sludge flow ratio a and concentration Xz,
and the more traditional parameters of F:M, U, and ®.. It should thus be
obvious that the traditional parameters represent aggregate system indi-
cators that result from the interactions of several factors; the same can be
stated for y,. The alternate approach as presented here isolates the ger-
mane engineering and biokinetic parameters and determines the resulting
aggregate impact on system performance using the process model. Thus,
the engineer can quantify or determine which values of a or Xa, etc., are
needed to meet target u, (or F:M or O,, if those terms are preferable)
values for a specific set of influent conditions and values of the
biokinetic constants.
As a further illustration, consider the curves presented in Figure 4.1.
Figure 4.1 shows the impact of the engineering parameters on resulting
values of F:M, 9,, and p,. Predicted calculations for S were made using
Equation 3.10, while Equations 4.1 and 4.7 were used to compute F:M
and ©,, respectively; values of X, the reactor biomass concentration,
were computed using Equation 3.14. These calculations were made using
an S, of 250 mg/L and values of the biokinetic constants of 0.5 hr', 100
mg/L, 0.60, and 0.04 d', for pmax, K,, Y;, and ky, respectively.
Figure 4.1 shows that there are several possible combinations of a, Xz,
Y,, and t that provide the required value of S. Thus, even though an
engineer may specify one value of F:M or ©,, there are several ways in
64 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

2.0
Ais
<<
2
3}
Q 1.0
o

=
WL

1.0
1.3
0,4
2.5
4.0

Xp = 8,000 mg/L
— — — X= 15,000 mg/L

t hours

Figure 4.1. Effect of X, andtonS (u,,, = 0.5 hr', K, = 100 mg/L, Y,

= 0.6, k, = 0.04 d'). S, = 250 mg/L, a = 0.25.

which one can obtain these target values. It is also interesting to note that
a recycle flow ratio a of 0.25, a recycle sludge concentration of 8,000
mg/L, and a detention time of 8 hr result in a 8, of 5 d. Empirically,
these values are frequently cited as the values for the recycle parameters
and tank detention time that are needed in order to achieve good process
performance for municipal wastes. A key point is that influent waste
concentrations are often greater than 250 mg/L usable COD. When this
occurs, the empirical “rules of thumb” break down, since they are based
on the assumption of relatively low influent concentrations for municipal
wastes.
 COMPARISONS AND BASIC APPLICATIONS 65

Consider Figure 4.2. All engineering parameters and biokinetic con-

stants are the same with the exception of S;, which is now 1000 mg/L
(instead of 250 mg/L as it was in Figure 4.1). The impact of influent
waste strength on the prediction of effluent quality is significant as evi-
denced by the substantial amount of substrate leakage predicted by the
model. Furthermore, the conventional approach of utilizing F:M or 0,
calculations does not provide the ability to determine which combina-
tions of values for the engineering parameters are suitable for insuring
good treatment. That is, the modeling approach presented herein enables
the process analyst to gauge the performance envelope. Stated another
way, the modeling approach allows the engineer to determine what opti-

COD/g
F:M,
X-d
g

Xp = 8,000 mg/L
— — 7— Xp= 15,000 mg/L

t, hours
Figure 4.2. Effect of X, and t on S. Same constants as Figure 4.1. S; =
1000 mg/L and a = 0.25. Compare with Figure 4.1 and
note effect of increased §S,.
66 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

mal combination of the engineering parameters are appropriate for

achieving target F:M or 90, values at higher influent waste
concentrations.

IMPACT OF RECYCLE PARAMETERS ON SYSTEM

GROWTH RATE

The importance of this section is that it points out that the recycle
parameters, a and Xa, have a significant impact on the reactor growth
rate. It will also be demonstrated how Equation 4.5 can be used to
explain qualitatively various aspects of activated sludge system kinetics.
Equation 4.5 is given again below:

py =" + a- aXg/X)/t = 1/0; (4.5)

This equation underscores the advantages of maintaining high recycle

sludge concentrations. If, for example, a plant has a settling or clarifier
problem and the recycle sludge concentration is essentially the same as
the mixed liquor concentration (i.e., Xp, ~ X), Equation 4.5 shows that
u,, (8.) is now heavily dependent on t and that the effect of the recycle
parameters are essentially inconsequential. Many times plants tend to use
high recycle ratios (1 or greater). This practice is not only costly from an
energy (pumping costs) perspective, but it has the other negative effect of
increasing solids flux (and/or surface loading) rates to the secondary
clarifier. Operators and designers should thus be fully cognizant that
both a and Xp, can be used to control growth rate. Control of recycle
sludge concentration can be implemented in the clarifier or can be
achieved by using thickened waste activated sludge as a source to increase
the concentration of recycle sludge. The point is that control of recycle
sludge concentration is not very difficult to engineer, especially in view of
the level of control it imparts for operation of an activated sludge
system.
Bulking is a problem which has impacted many treatment plants.
These situations are the result of changes in the ecology of the biomass,
leading to a sludge with poor settling, compaction, and solids separation
characteristics. Many times bulking is caused by the proliferation of
filamentous microorganisms, which produce a biomass with string-like
growths protruding from the sludge flocs. A filamentous bulking situa-
tion presumably occurs due to changes in environmental conditions (pH,
waste characteristics, etc.) or by naturally occurring fluctuations in the
 COMPARISONS AND BASIC APPLICATIONS 67

system ecology. Equation 4.6 is useful for explaining in a qualitative

sense why the performance of an activated sludge system is apt to deteri-
orate so quickly once this type of settling problem begins. Consider that
the first effect of a bulking problem is a decrease in the underflow, or
recycle, sludge concentration. As X,g drops, Equation 4.6 shows that p,
increases; so does the rate of filament growth, which in turn results in
increased filament proliferation. This intensifies the level of bulking and
presumably causes further drops in X,; further drops in X, simply act to
exacerbate the situation, since growth rate continues to increase. This
leads to suspended solids losses over the secondary clarifier weirs and, if
the situation is severe enough, excessive increases in system growth rate,
as indicated by Equation 4.6, suggests the potential for leakage of solu-
ble substrate (COD).
As discussed in Chapter 3, the treatment of inhibitory wastes mandates
the avoidance of operation near the peak growth rate p»* for the system.
Equation 4.6 indicates that fluctuations in Xg produce fluctuations in
growth rate which, for inhibitory waste treatment situations, can repre-
sent violations of n* with resulting consequences being severe effluent
deterioration and potential system failure. Unless other provisions are
made to provide a growth rate cushion (e.g., by using large tank volumes
to allow for long hydraulic detention times), some means to control
recycle sludge concentration are necessary in order to increase the perfor-
mance envelope of activated sludge systems treating toxic or inhibitory
wastes.

DETERMINATION OF CONDITIONS NEEDED

FOR MINIMIZING SLUDGE PRODUCTION

Extended aeration plants are generally thought of as activated sludge

systems that are designed and operated with the goal of having little or
no excess sludge production (that is, X, = 0). The question, then, is how
to select the engineering control parameters to approach the condition of
zero sludge wastage, or to try to minimize sludge production. One
approach is to simply “waste less and see what happens.” Alternatively,
one can use the modeling approach presented herein and proactively
determine a realistic sludge minimization goal for a given treatment
facility.
First, it is important to recognize that the condition of X, = 0 (zero
sludge wastage) also means that the net growth rate, u,, is also equal to 0.
68 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

This, in turn, means that » = kg since n, = p - kg. Once again, recall

Equation 4.5:

isae(ck aX aXGhXVt 2Hi/0, (4.5)

For p, = 0, the term in parentheses must equal 0. For this to occur, the
requisite reactor biomass concentration X that will prevail under zero
sludge production conditions is determined as follows:

X = Xz (a/(1 + @)) (4.8)

Effluent quality, S, is determined by recognizing that p = ky:

ue = k, —
HmaxS/(K, + S) (4.8a)
Cae k aK /(nax = k,) (4.8b)

Combining Equations 3.6 and 4.5 produces Equation 4.9, which calcu-
lates the detention time t, required to achieve extended aeration condi-
tions for specified values of the engineering parameters a, Xp, and §;:

DAE tentcyite Cled Da eS

ery K" (Aare) ear) Ua)
The values of the biokinetic constants impact the prediction of sludge
minimization conditions by virtue of the fact that the prediction of efflu-
ent quality, S, is given by Equation 2.11. For this application, n, = D =
0. Equation 4.9 is simplified to Equation 4.9a:

orca. SoS Sever Uy ie-( + a) (4.9a)

Be aX,/( + a)
Now consider a treatment plant with a recycle sludge concentration of
10,000 mg/L, a recycle flow ratio of 0.25, and biokinetic constant of
3.15 d', 0.07 d', 105 mg/L, and 0.63 for pax, kg, K,, and Y,, respec-
tively. Table 4.1 presents the calculated required detention times for
achieving extended aeration (u, ~ 0) conditions for a variety of values of
S;, the influent waste strength.
Table 4.1 makes some interesting points. First, it is useful to recall the
old axiom that extended aeration conditions are achieved with a 24-hr
detention time. Consider the fact that “typical” operating parameters for
municipal treatment facilities are the recycle values used in the table
 COMPARISONS AND BASIC APPLICATIONS 69

Table 4.1 Required Detention Time for Achieving Extended Aeration

Conditions in an Activated Sludge System
S; (mg/L COD) t. (hours)*
100 11
250 pe|
500 54
1,000 108
*Computed using Equation 4.9a.

calculations (@ of 0.25 and Xz, of 10,000 mg/L) and that primary efflu-
ents for municipal systems are generally close to 250 mg/L COD. The
model shows that, with these conditions and specified values of the
biokinetic constants, a plant will achieve extended aeration at a detention
time of 27 hr. This value is relatively close to the 24-hr detention time
cited so often.
As S, increases, the required detention time also increases. This occurs
because S; has a profound influence on reactor growth rate. In order to
keep the growth rate at the required value (u = k,) for minimizing sludge
production, it is necessary to increase t substantially. The key point is
that this model analysis enables process analysts to determine the poten-
tial for minimizing sludge production at their facilities. The model pro-
vides the required t and the expected mixed liquor suspended solids
values (X) for various operating conditions (a, Xz, and S,). Thus, facility
operators who desire to minimize sludge production in their systems can
do so using a proactive strategy that enables them to make the assessment
using a structured algorithm as a management tool.

COMPUTATION OF AERATION TANK OXYGEN

TRANSFER REQUIREMENTS

Another useful application of the modeling approach presented here is

with regards to the computation of oxygen transfer requirements for
activated sludge reactors. Many times, design or operation problems
involve the detailed consideration of oxygen transfer. This is not a trivial
consideration; the optimization of the aeration facilities can represent
substantial cost savings since the operation of these systems requires a
relatively large power commitment. The approach presented herein is
targeted for both design and operational applications. That is, this
70 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

approach can be used to size transfer facilities for new systems as well as
to optimize or modify the operation of existing ones.
The modeling methodology for predicting oxygen transfer require-
ments integrates our modeling concepts with the fact that all metabolized
COD in an aerobic reactor is channelled into either oxygen uptake (CO,
evolution) or into cell COD:

ACOD = O, uptake + ACOD,,), G17)

This balance concept, which was presented in detail in Chapter 1, and

the modeling equations presented in Chapter 3, are combined to provide
predictive equations for the required k,, for a reactor.
As an illustration of the application, consider a simple chemostat as
depicted in Figure 4.3. Steady-state mass balances yield predictive equa-
tions for X and S, the effluent substrate and biomass concentrations,
respectively:

(hea) — ¥s. 03, (her Oo)

t. = e ve (4.10)

Ss = —
K, (D+k,) 4.11
Umax ~ (D+k,) ( )

aRa ere
Y,D(; a S)
(4.12)

An additional mass balance is also written for dissolved oxygen concen-

tration, C:

rate of O, rate of O, rate of O, leaving

change transfer reactor

dC (
V ee FC; + (kj, (C.-C)V) - ((F)(C)-O, for metabolic demand)
(4.13)

The oxygen uptake rate for the chemostat is computed using the aerobic
balance principle shown in Equation 4.14.

O, uptake = F(S,-S) - 1.42FX (4.14)

 COMPARISONS AND BASIC APPLICATIONS 71

V, S, X,C

Figure 4.3. Chemostat mass balances for cells, substrate, and oxygen.

It should be noted that, for the chemostat, F(S, - S) represents the

ACOD and 1.42FX represents the ACOD,,,,. Inserting Equation 4.14
into Equation 4.13 results in Equation 4.15:

v & = k,(C,- OV - FC - (FG-S)-1.42XF) (4.15)

Expressions for S and X are given in Equations 4.11 and 4.12, respec-
tively. The steady-state solution for predicting k,, in a chemostat is given
in Equation 4.i6.

ka = (C t.SC=C)
= SL 4a2X) (4.16)

Equation 4.16 is a function of the engineering parameters t and S,, the

biokinetic constants, the reactor dissolved oxygen concentration (DO),
and the saturation DO value. This equation gives the required k,, for
aeration equipment for a specified set of conditions in a chemostat.
The application of this approach for obtaining a predictive equation
for k,, in a completely mixed activated sludge reactor uses Figure 4.4 and
a similar analytical approach. The oxygen uptake rate in a completely
mixed activated sludge system is provided by Equation 4.17:

ACOD ACOD ..i;

O, uptake = (F(S; + aSp) - (1 + a)FS) - ((1 + a)FX - aFX,)1.42)
(4.17)
72 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Aeration Tank

(1 + a) FX, S,C
F, $C;

Vv
aFX, (Recycle sludge)

Figure 4.4. One cell completely mixed activated sludge mass balances
for cells, substrate, and oxygen.

Predictive equations for S and X were derived and are given in Chapter
3; these are Equations 3.10 if the waste is noninhibitory or 3.19 if the
waste is inhibitory (for S) and Equation 3.14 (for X). The oxygen mass
balance is given below:

Vv « = k,(C,-O)V - (1+a)FC (4.18)

- F(S; + aSp - (1+ a)S - 1.42((1 +a)X - aX,))

A steady-state predictive equation for k,, in a completely mixed acti-

vated sludge reactor is given in Equation 4.19:

_ FG +oS,-( + a)S-1.42((1 + a)X-aXg)) + (1+a)C)

(4.19)
Kia (C,-C)V

A situation which is perhaps of more practical importance from both a

design and operational perspective is the prediction of the variation of
required oxygen transfer capacity or k,, in a plug flow reactor. The same
modeling approach employed to obtain the multiple-cell predictive
model in Chapter 3 is utilized to derive the predictive equations for k,, in
a multiple reactor, or plug flow system. The difference here is that we
incorporate the aerobic balance principle, perform an oxygen mass bal-
ance in each cell, and use Figure 4.5. In the multiple-cell system, one will
predict different k,, values for each cell. This enables designers to refine
equipment selection or managers to refine operating policies to conserve
energy expenditures. For the derivation, it is assumed that there are “m”
 COMPARISONS AND BASIC APPLICATIONS 73

Cell M

X(1), S(1), X(2), S(2), +2 X(m)

C1) C42) Cnt

Figure 4.5. Miultiple-cell activated sludge mass balances for cells,

substrate, and oxygen.

number of completely mixed aeration tanks of equal volume. If the total

tank volume is V, then each cell has a volume of V/m.
An expression for the oxygen uptake rate in the first tank in Figure 4.4
is provided in Equation 4.20:

ACOD ACOD arg

(4.20)
O, uptake = (F(S,+ Sg) - (1+a)FS,)) - (1 +.a)FX,) - oFX,)1.42)
The oxygen mass balance in tank | is:

Vaed@ V
m aa = k,(C.-Cy)) m

- (1+a)FC,, - F(S; + eSg-(1 + a)Sqy-1.42((1 + a) XQ) - @XR))(4.21)

At steady state, the predictive equation for k,, in tank 1 is:

ere mF((S, + aSp-(1 + a)S,)-1.42(1 + @)X()-aXp)) + A+ a)C,))

1 oe CR(oo GT can (4.22)
Equations for X,,, and S,, are given in Chapter 3.
For tanks 2 through m, the expression for oxygen uptake rate is:

ACOD BCODun
O, uptake = (1+a@)F((Sg_)-Sg) - (1-42(%q-XGy))) (4.23)
tanks 2
through m
74 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

The oxygen mass balance equation for tanks 2 through m is:

Vera@.
Ca) = Ce
V
pa
m dt ki(Cs-Co) m
- (l+a)FCy - (1 + a)F(SG-1)-Sg-1-42(Xy-XG-1))) (4.24)

At steady state, the predictive equation for k,, for tanks 2 through m in
a multiple-cell or plug flow reactor is given in Equation 4.25.

i 2 m( PPayFisies 2a, Ex
a) ( G-l) (Ccsae G-1)) sere,
Gy) (4.25)

It should be noted that the equations presented in this section can be

used to predict steady-state oxygen uptake requirements in either com-
pletely mixed or multiple-cell reactor configurations. The process analyst
can examine numerous different design or operational scenarios for oxy-
gen transfer requirements by simply changing the various input parame-
ters. These parameters are the recycle variables, a and Xp, the influent
flow rate and waste concentration, F and S,, respectively, and the
biokinetic constants. One could also investigate the optimum number of
reactor cells that can be utilized to minimize the aggregate tank k,,
requirements. This approach provides the activated sludge system
designer, operator, and manager with a structured and relatively straight-
forward algorithm for quantifying and predicting oxygen transfer
requirements. The use of this tool will enable process analysts to refine
the optimization of the oxygen transfer features for activated sludge
system design and operation.

KEY CONCEPT SUMMARY

The key concepts contained in this chapter are:

¢ The approach to activated sludge presented herein is readily rec-

oncilable with other, more conventional approaches, i.e., F:M,
U, and ©... The modeling approach presented in this text differs
in that it is more flexible and that it enables the process analyst to
examine the impact of the individual process components on
system performance (waste strength, recycle parameters, etc.).
This enables the process analyst to define the performance enve-
lope more accurately.
 COMPARISONS AND BASIC APPLICATIONS as

¢ The recycle parameters for an activated sludge system greatly

impact system growth rate and thus significantly influence pro-
cess performance. The approach presented here can be used to
quantify the optimal values of these parameters. This is espe-
cially crucial for systems with high loadings or those that must
handle inhibitory or difficult-to-degrade wastes.
¢ The determination of the specific conditions, i.e., the values of
the recycle parameters, tank detention times, etc., needed to min-
imize waste sludge production can be quantified.
¢ Predictive equations for determining aeration requirements (k,,
values) for activated sludge systems can be formulated with a
relatively simple application of the process model. This enables
process analysts to refine design or operational issues that
involve aeration equipment.

REFERENCES AND SUGGESTED

ADDITIONAL READING

Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental

Engineering, Engineering Press, Inc., San Jose, California.
Gaudy, A.F. Jr., Srinivasaraghavan, R., and Saleh, M. (1977). “Concep-
tual Model for Design and Operation of Activated Sludge Processes.”
J. Environ. Eng. Div., ASCE, 103, pp. 71-84.
 "ey

bieebreae Mingle onda t

di bese 39 pide e196 gut
onan a ST sella
dare seit & Ry, rae
rat

* eritanosh oa euveythrr
2 D- iis ‘ tt

pattandip wma

a.te Nelugnel ad man tities 328 nie Wey)

no eid lbau assem, a Yeteil sly)
a» fi rz aeites:
neal to piety w Cal NCEE Oot ea
y pea
Néiely miad.or tries seaciow ae ation
ee oma Foeoadieiy & oe :
2h WUSEAY eaLiremueeia
Pir vege sha
“se Than yirameetar
arp Sy rim sie nee
ts wait dre cant Voormann. «ga,
Ve dimk ¢ aware. ‘Css — Aloe $e hs
welche tetra te
PRE er ese
‘phaprens se UBT che
ee ers Coretta
ramiramees. Tie lene paeeicaneien a hie Ae zal
me Ovoimirnitliy <fie- ONp yer rensfer iessee f09
coe ‘duds

a r item decian art aneration

y CUNCUDY SUM Ally = ,

VETEte COTA GD de LAS OOPDe eles + = >,

:
porch io anivaled sides timeiiest ter
es inve
" Y dhe tty @orveyrihem! samp
8 ONE THe LOT apynreet Semmentod te
WN La rye WOOre he; mili J rattles thy jeer
Sa ras oF the eal prepilin ang
i i Hive (wane 7 rengih, Ainw
oy. g) i tps ik) he, ) Aahy: « {0 saline Aya. , -
 ) PROCEDURES FOR
OBTAINING BIOKINETIC
CONSTANTS

INTRODUCTION

This chapter will detail analytical and laboratory methodologies that

can be used to generate and then analyze the requisite data needed to
quantify the values of the biokinetic constants. It is the goal of this
chapter to provide a “cookbook” approach that will enable practitioners
to determine the values of these parameters without having to initiate a
major “research” effort. That is, our experience has been that determina-
tion of the constants can be performed on a routine basis without much
difficulty, provided that a reasonably experienced technician is available
for performing the work.
This chapter contains three main sections:

e¢ Determination of Influent Waste Strength, S;

¢ Use of Respirometry to Generate Kinetic Data
e Determination of True Cell Yield, Y,, and Decay Rate, k,

The determination of influent waste strength, S,, is not a trivial issue.

This is stated because waste streams are rarely, if ever, composed primar-
ily of one or only a few components. Consequently, the delineation of
what “S,” is for a biological treatment system can at times be a difficult
task. This chapter will address this issue and present techniques and
strategies that can be employed to quantify influent waste strength.
The biokinetic growth constants, which are contained in the growth
rate expressions which relate » to S (biomass growth rate to substrate or
waste concentration), are the biological parameters having the most

1g
78 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

influence on predicting effluent quality. Also, these parameters are the

ones more likely to exhibit the greatest level of variation with changing
environmental conditions. As previously discussed, there are three
biokinetic growth constants: pa,, K,, and K; (if the waste is inhibitory).
The key data needed to determine the values of the biokinetic growth
constants are growth rate vs. substrate concentration. These data are
generated by analyzing respirometric or oxygen uptake measurements of
the target biomass degrading the target waste; these measurements are
taken in tests that can be completed in as little as eight hours and rarely
go longer than one day. Once the growth rate data are collected, they are
fit to the Monod or Haldane (if the kinetics are inhibitory) growth rate
expressions. This chapter will detail how respirometric tests are set up,
how the data are analyzed to determine growth kinetics, and then how
the growth rate data are analyzed to determine the values of the
biokinetic growth constants.
Finally, this chapter will discuss methods to determine values of true
cell yield, Y,, and decay rate, k,. These constants are more crucial for
predicting quantities of waste sludge production and are not as critical
for predicting effluent quality. This is fortuitous, since the data needed
for determining the true cell yield and decay rate must be generated in
relatively long-term tests using continuous flow reactors.

DETERMINATION OF INFLUENT WASTE STRENGTH, §;,

The definition and explanation of the ACOD test for measurement of

biologically usable COD for a particular biomass was previously dis-
cussed in Chapter 1. Its primary application regarding modeling proce-
dures for designing and operating biological wastewater treatment sys-
tems is aS a means to quantify the influent waste concentration a
particular reactor must handle. It is important to note that the definition
of ACOD comprises both a concept and test. The concept is basically
contained in the fact that any given biological wastewater treatment
system is characterized with a certain amount of inefficiency, which
precludes the total removal of all soluble COD. That is, the soluble
effluent of the biological system will always contain some COD, which
many times is not associated with the influent COD since the effluent
soluble COD is frequently composed of microbial excretion products and
other by-products of biochemical activities. In other situations, a portion
of the influent COD is resistant to biodegradation and passes through the
system. Finally, a biomass may be not be acclimated to a new waste
 OBTAINING BIOKINETIC CONSTANTS 79

stream or component and the new carbon source, although ultimately

biodegradable, may pass through the plant and appear as soluble efflu-
ent COD because the new COD may not represent a readily biodegrada-
ble carbon source for the existing biomass. (For example, consider the
case of the glucose COD not being metabolized by phenol-acclimated
cells, depicted in Figure 6.7). It should be noted that this conceptual
consideration differs somewhat from the ACOD test because the test,
which is described in Chapter 1, assumes that the biomass is well
acclimated.
Examples of actual ACOD test results are given in Figures 5.1, 5.2, and
5.3. The test shown in Figure 5.1 was performed using a population that
was acclimated to the readily biodegradable carbon source, glucose. In
this test, the residual COD, COD,, was quite low, and only traces of
glucose were found. Thus, in this instance, the ACOD essentially equals
the influent COD and the residual is likely attributable to microbial
excretion products, intermediates, and/or components from dead or
dying cells. In this case, S; equals influent COD.
Figure 5.2 represents a quite different situation. This test was per-
formed using an acclimated biomass on a waste with a high lignin con-
tent. It shows that the amount of COD, is quite high, which means that

mg/L
analysis,
Indicated

0 5 10 15 20 25
Time, h

Figure 5.1. ACOD and biomass curves for a sample containing glucose
as carbon source. (Gaudy and Gaudy, 1988).
80 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

2400

2000

fez)Soo
—_

1200

COD,
Soluble
mg/L

Time, h

Figure 5.2. ACOD for a sample of dilute kraft pulp mill digestor
blow-down liquor. (Gaudy and Gaudy, 1988).

there will be a relatively large difference between influent COD and

ACOD (S,). Lignins are not readily metabolized and are not normally
expected to contribute to the organic loading of a biological wastewater
treatment facility.
In Figure 5.3, the results of a ACOD test are presented for a municipal
waste treatment plant that is treating a domestic waste with a relatively
high industrial contribution. The results in Figure 5.3 depict a case that is
between the ones in Figures 5.1 and 5.2. For the test in Figure 5.3, about
70% of the influent soluble COD was readily amenable to biodegrada-
tion, while the tests in Figures 5.1 and 5.2 showed results of 95% and
45%, respectively.
The use of influent COD values for determining S; values for design
and modeling purposes will provide the biological system analyst with a
conservative estimate of the influent waste strength. This is stated
because ACOD will always be less than the influent waste strength
(COD,) and the ACOD provides the estimate of the loading that is ame-
nable to biodegradation by the biomass. S, as defined using ACOD gives
the analyst the actual organic loading that the biomass will degrade.
 OBTAINING BIOKINETIC CONSTANTS 81

250

200

50

Time, hours

Figure 5.3. Example of ACOD test performed on municipal primary

settling tank effluent. ACOD = 131 mg COD/L.

It should be also be pointed out that what is COD, for one reactor may
also be COD, or ACOD for another reactor. For example, consider two
activated sludge reactors in series. (This approach is sometimes utilized
for two-stage nitrification.) The effluent from the first system is the feed
for the second system. The feed for the second system may consist of
COD that passed through the first unit, metabolic intermediates, or
both.
For operational purposes, the periodic determination of the ACOD
can be used as a relatively rapid check on the plant’s performance. That
is, if soluble CODs in the effluent increase, it may be due to an increase
in recalcitrant organics in the influent and not due to any operational
oversights. It should also be noted that quantification of the S; values via
the ACOD procedure enables the plant to optimize operations by deter-
82 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

mining actual loading conditions, which may be different (ideally lower)

than the design values. For instance, an accurate evaluation of S; allows
the plant to determine the actual number of reactors it needs in service,
the level of aeration required, or both. These measures enable the plant
to economize operations via the implementation of energy-saving
practices.

USE OF RESPIROMETRY TO GENERATE KINETIC DATA

The key advantage of using respirometry for collecting kinetic infor-

mation is the fact that data are collected much more efficiently than can
be accomplished with other methods. Additionally, the data can be col-
lected automatically and in greater quantities than is possible with alter-
nate techniques. Previously, kinetic data for calibrating biological treat-
ment models were collected using either the batch growth study (shake
flask) method, substrate utilization technique, or various hybrids of
these procedures to generate growth kinetics which can be analyzed to
determine values of the biokinetic constants. Although the constants
generated using these methods enabled us to obtain reasonable predic-
tions for reactor performance, the work involved with obtaining the
constants and associated analytical difficulties precluded the routine
determination of biokinetic constants. The use of respirometric proce-
dures circumvents many of these problems.
The goal with the respirometric methodology is to obtain data that can
be analyzed to determine the values of the biokinetic growth constants,
Lax» K,, and K, (if the waste is inhibitory); these are the constants con-
tained in the two growth rate relationships, Equations 1.11 and 1.12,
which are utilized to relate » to S. Once we can relate » to S, then we have
calibrated the model.
The respirometric procedure for calibrating the model, i.e., determin-
ing the values of the biokinetic growth constants, involves the generation
of respirometric data which are translated into either biomass growth
data or substrate utilization data. The translated data are then analyzed
to obtain a set of growth rate and substrate (waste) concentration data,
i.e., » and S points, which are then fit to either the Monod or Haldane
function to determine the values of the biokinetic constants ,,,,, K,, and
K,.
 OBTAINING BIOKINETIC CONSTANTS 83

Conversion of Respirometric Data to Growth or Substrate Utilization

Data

The link between oxygen uptake (respiration) and biomass growth or

substrate utilization is based on the assumption that the COD being
removed from solution during metabolism is channelled in varying pro-
portions into the synthesis of new cells and to respiration measurable as
oxygen uptake. This relationship between growth, substrate removal,
and respiration is quantified using Equation 5.1.

ACOD = O, Uptake + ACOD,.:, (5.1)

This equation states that the amount of substrate removal in an aerobic
biological system is accounted for as the amount of COD that has been
incorporated into the cells (ACOD,,,) plus that which has been oxidized,
represented as accumulated oxygen uptake. In the absence of physical
stripping or chemical oxidation of the substrate, Equation 5.1 represents
the complete mass balance for substrate removal in aerobic biological
systems.
Equation 5.1 can be simplified by noting that the amount of substrate
COD that has been channelled to biomass (ACOD,,,,,) can be expressed as
the product of the amount of cells produced (AX) and the unit COD of
the cell mass (O,). This relationship is given in Equation 5.2.

ACOD,,., cells = (AX)(O,) (5.2)

The equation for cell yield is also used to simplify the expression.

Y = AX/ACOD (5.3)

Equation 5.3 is utilized to produce an expression for ACOD.

ACOD = AX/Y (5.4)

Recognizing that AX = X, - Xp, (X, and X) are X at some time t and at

time 0, respectively) and substituting Equations 5.2 and 5.4 into Equa-
tion 5.1, we obtain Equation 5.5.

(X, - X))Y = O, uptake + (X, - X,)O, (5.5)

Rearranging and simplifying Equation 5.5 yields Equation 5.6.

KX, = X, + O, Uptake/(1/Y = O,) (5.6)

84 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Equation 5.6 is used to convert oxygen uptake data into biomass

growth curves, which can then be analyzed to determine growth rates
(u’s) at different initial substrate or waste concentrations (S’s); one
growth curve is generated for each respirometric test unit.
In order to use Equation 5.6 for converting O, uptake data to biomass
data, values for cell yield, Y, and cell COD, O,, must be selected. These
parameters can be determined from actual test data or estimated using
previous information on the particular waste treatment system from past
kinetic tests. To calculate values for Y from test data, use Equation 5.3;
O, values are calculated using Equation 5.7.

O, = (Total COD - COD)/X (3:2)

O, can be calculated during the substrate (COD) removal phase of the

test and then averaged to obtain an average O, value for use in Equation
5.6. According to Porge’s stoichiometric formula for activated sludge
(C;H,NO,) (Gaudy and Gaudy, 1988), values of O, should be close to
1.42 mg COD/mg X; in general, experimentally determined values for O,
are close to this value.

Example of Obtaining Growth Data from Respirometric Data

This example presents the results of a respirometry study performed

for the Patapsco Wastewater Treatment plant (Baltimore, Maryland)
(Rozich and Gaudy, Inc., 1988) using a 24-hour composite sample of
primary effluent as a waste sample (S) and plant return activated sludge
for biomass seed (X). Table 5.1 lists the cumulative O, uptake data that
were collected from five reactor cells at five different initial waste con-
centrations. These data, plotted in Figure 5.4, were collected directly
using an electrolytic respirometer. A separate batch reactor containing an
undiluted waste sample and small biomass inoculum was run to collect
data for calculating Y and O, values; these data and the calculations for
computing Y and QO, are given in Table 5.2. It should be noted that the
data in Table 5.2 can also be used to calculate a ACOD value for this
waste sample.
The biomass growth curves are generated by using Equation 5.6 and
the data in Table 5.1, along with the values of Y and O, computed in
Table 5.2. The results of this analysis yield a computed biomass curve (X
as a function of time) for the different substrate (COD) concentrations in
each flask.
 OBTAINING BIOKINETIC CONSTANTS 85

Table 5.1 Example of Cumulative Oxygen Uptake Data Collected

from a Respirometric Study Performed on 5/12/87 Using a 24-hour
Composite Sample of Primary Effluent
Cumulative Oxygen Uptake (mg O,/L)

Time Percentage Waste Strength

h 20% 40% 60% 80% 100%
0) 0 0 0 0 0
0.2 0 0.6 0 0.6 0
OF 0 1.4 0.9 3.0 ZZ
12, 0 1.6 0.9 5.6 5.0
ie #/ 0.1 1.6 Sat 8.6 8.3
232 0.2 Be ol 12et 11.8
Zo 0.6 1.8 9.2 16.5 16.4
3.2 1s 1.9 12.6 Zit 2173
Sa, 2.3 2 16.1 26.6 29.0
4.2 3.0 ees 19.8 32.2 3574
4.7 3.5 a3 24.0 37.0 41.7
S62 4.3 5.6 26.8 40.7 49.4
5.7. 4.9 6.7 28.7 43.6 54.0
6.2 5.4 T3 30.5 45.8 a ae
6.7 6.0 8.1 30.0 48.0 60.9
The Oxi 8.5 33.4 0:2 63.7
Tel 7.4 9.4 35.4 52.3 66.3
8.2 8.2 1022 36.9 54.3 68.8
8.7 8.9 11.0 38.6 56.2 70.6
9.2 9.7 12.0 40.4 58.1 (8
9.7 10.3 1c 41.2 60.0 Tae7
S, (mg COD/L) 28 55 83 110 138
X, (mg TSS/L) 40 55 80 105 145

The calculated values for X, computed using Equation 5.6, are listed
in Table 5.3, while plots of the data on semilogarithmic paper are pre-
sented in Figure 5.5. The growth rate for each waste concentration is
determined by identifying the exponential growth phase on each curve
and drawing a straight line through the data. The slope of this line is the
specific growth rate. The growth rate is determined using Equation 5.8.

uw = In(X,/X,)/(t, - t)) (5.8)

86 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

80
100%
=©)
—£ 60 80%
g
8a.

Oo
a
40— 60%
@
>
3
e
3 20
40%
20%
0
0 2 4 6 8 10

Time, hours

Figure 5.4. Plot of cumulative oxygen uptake data from Table 5.1.

In Equation 5.8, X, and t, and X, and t, represent the beginning and

end coordinates for the exponential phase, respectively. It should be
noted that depending on if the kinetics adhere to the Monod (noninhibi-
tory) or Haldane (inhibitory) relationship, the analyst can sometimes
expect a different characteristic biomass growth curve that requires some
additional consideration.
Consider the idealized growth curves presented in Figures 5.6 and 5.7.
If the waste is noninhibitory, then one can expect the type of growth
curve depicted in Figure 5.6 and the identification of the exponential
phase and subsequent analysis to compute the growth rate are relatively
straightforward. Such is not always the case for inhibitory wastes. Many
times, inhibitory growth curves resemble the idealized curve shown in
Figure 5.7; the specific shape of the curve is dictated by a combination of
 ‘Q3BIOAR [[BIOAO dY} Ul Popnyour jou o1aM sJUIOd 9so4L» :210N
87

(a]dwes
/*O X% =O 7
+ OWT + CET + OF B= *O10 :sojdures Jo ‘ou

uaatd
Ol =8/(Bril + SrI+ Tel + Sel + 8h

IOJ)
“O
'« / (GOO - GOD IP10L)

“I=&
BIOKINETIC CONSTANTS

*“O Bune[ns[ed
6b'0 = PII/9S = A (A) PIP [199 ssemorg *€
J/dOD 3W pil = STI-6EZ ((GOOV) PeAowel GOD FAANOS “%
T/SSL BUI 9S = pOZ-09Z :[eAOWIAI GOD FQNIOs 01 onp YIMOIS SseUOIg “T
[A Bune[nseD
orl = *O
LZ . = = = = 4 SZ Ci E'S) E11
= +
rp 097 8r0f0 «=E8670 9 SZ SE cs
#67 | SZI
as ZSZ rIZe'0 ISsI€O ¥ SZ Se 9
Srl IPI
Lr p7S 097 rS9f0 68560 6 SZ iss
Srl
co SE €
OBTAINING

I€'l 8ST €8b 8rZ 6IIE0 LSOEO LZ

aS Orz Soleo SsEleo SZ SE gC
Sel O81
pes 9€7 89670 60670 Sz SZ SE TZ Z
Srl SgI
Svs Orz 1€ze"0 ILI€O Ol SZ SE ay |
Or'l 607
pes 9EZ 198770 7O0sz7o 61 SZ SE I
Ze'l €7Z
67S 9IZ O9re'0 90PE'0 62 SZ Se $0
Or'l L@Z
Svs p07 S8ZE0 = PETE 883 SZ SE in Oe 0
«0ST 6£7
38 8 =jaqeT ju jw 3d. Hd 4Y
xe) 1/d00 8u 1/d09 8u T/SSL sul
dod IFeL x WSIAA —-YBIEAA FeVE suNfOA aJduI~g suINjoA dway ou,
sPp93j0dGd00 ‘GOO ajduies
pyoOS + sey sey SpljoS pspusdsng
XO pue Xx Jo UONLUTUIII}IG 10J BIL MEY 7S PAGEL
88 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Table 5.3 Biomass Growth Data Transformed from the Cumulative

Oxygen Uptake Data listed in Table 5.1 using Equation 5.6
Biomass Growth Data (mg TSS/ L)
Percentage Waste Strength
Time
h 20% 40% 60% 80% 100%
0 40.0 55.0 80.0 105.0 145.0
0.2 40.0 55.9 80.0 105.9 145.0
0.7 40.0 STZ 81.4 109.7 148.4
12 40.0 ep 81.4 113.9 152.8
153) 40.2 By os] 84.8 118.4 157.9
Pa. 40.3 57.0 88.0 123.9 163.4
0 ae 40.9 57.8 94.4 130.7 170.6
332 42.3 58.0 99.7 138.9 178.2
3.7 43.6 59:2 105.1 146.5 190.2
4.2 44.7 60.5 110.9 1552 199.8
4.7 45.5 63.3 117.4 1624 ZA0.1
D2 46.7 63.7 121.8 168.5 222.1
S-/ 47.6 65.5 124.8 173.0 229.3
6.2 48.4 66.4 127.6 176.5 234.3
6.7 49.4 67.6 129.9 17S 240.0
gap 50.4 68.3 132.1 183.3 244.4
ies S15 69.7 a2 186.6 248.5
8.2 52.8 70.9 137.6 189.7 252.4
8.7 53.9 dee 140.2 192.7 Pin toe?
9.2 531 oak 143.0 195.7 259.1
9.7 56.1 75.4 144.3 198.6 261.6
S, (mg COD/L) 28 55 83 110 138
X, (mg TSS/L) 40 55 80 105 145
Note: O, = 1.40
NY tres) O49

the level of inhibition, the amount of inocula used (Xj), and the initial
COD (S) that is employed in the particular flask. For the type of curve
depicted in Figure 5.7, the correct exponential phase, which correlates
with the growth rate, is the first exponential phase that is manifested and
not the larger “second” exponential phase. Previous laboratory and ana-
lytical work on the inhibitory substrate phenol showed that analyzing the
 OBTAINING BIOKINETIC CONSTANTS 89

400

200

100
80

60
TSS/L
Xt,
mg

40

0 2 4 6 8 10
Time, Hours

Figure 5.5. Plot of biomass growth data (see Table 5.3) at different
initial waste concentrations for determination of the
specific growth rate, p.

second phase produces a large overestimate of the characteristic growth

rate of the target batch reactor. This in turn will result in producing
erroneously higher estimates of the biokinetic growth constants. Thus,
when analyzing kinetic results for wastes that may be inhibitory, care
should be exercised in determining growth rates. More detail on this
aspect of analyzing growth kinetics in inhibitory systems is given else-
where (D’Adamo, Rozich, and Gaudy, 1984).
90 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Biomass

Time

Figure 5.6. Idealized semilog plot of the increase in biomass

concentration as a function of time for a noninhibitory
waste. The straight-line portion of this curve represents
the exponential growth phase, the slope of which is
defined as the specific growth rate, » (see Equation 1.5).

Determination of the Biokinetic Constants from Growth Data

The growth rate data (u vs. S) obtained via analysis of the respirome-
tric data using Equation 5.6 are fit to either the Monod or the Haldane
equation to determine the values of the biokinetic constants.

Bh = Bmaxo/(K, + S) Monod (1.11)

1 = pmaxS/(K, + S + S?/K)) Haldane (1.12)

 OBTAINING BIOKINETIC CONSTANTS 91

Biomass

Time

Figure 5.7. Idealized semilog plot of the increase in biomass

concentration as a function of time for an inhibitory
waste. The initial straight-line portion of this curve
represents the exponential growth phase, the slope of
which is defined as the specific growth rate, p (see
Equation 1.5).

If w increases with S until it becomes asymptotic to an upper limit, the

test data are described with the Monod equation (noninhibitory). Con-
versely, if the growth data first increase with S, reach a peak, and then
begin to decrease, the Haldane equation (inhibitory) is used to fit the
data.
The growth data can be fit to both the Monod and Haldane equations
using a nonlinear least squares method contained in a computer pro-
gram. Regarding fitting data to the Monod equation, our experience has
been that nonlinear computer methods are both faster than and prefera-
ble to reciprocal plotting techniques. In the case of fitting data to the
Haldane expression, reciprocal plotting techniques are virtually unusa-
ble. It was also determined that because one is fitting data to an expres-
sion with three constants, it is possible to obtain numerous (if not infi-
nite) different combinations of y,,,,, K,, and K, that all provide
reasonable fits to the data. This fitting problem is addressed by incorpo-
rating a bounding procedure around the nonlinear algorithm, which pro-
92 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

vides consistent input for the initial guesses of the values of the con-
stants. As values of the constants are generated from the nonlinear
fitting procedure, they are subjected to a screening test to determine if
the constant values are within the values stipulated in the bounding
procedure. For example, negative values of the constants are automati-
cally rejected. More detailed discussion on this aspect of analyzing inhib-
itory growth data is given elsewhere (D’Adamo, Rozich, and Gaudy,
1984).
A computer program for fitting growth data to the Monod equation is
given in the Appendix; the nonlinear package, “DUNLSF,” can be
obtained from the International Mathematical and Statistical Libraries
(IMSL) in Houston, Texas. Fitting the growth data (u vs. S) in Table 5.3
using the program in the Appendix (values of EPS and NSIG for the
DUNSLF routine equal to 0.0 and 3, respectively) yields values of 0.153
hr' and 55 mg/L COD for p,,,, and K,, respectively, with a residual sum
of squares value (SSQ) of 0.00079. As a check, it is useful to compare a
plot of the fitted curve with the actual growth data as depicted in Figure
5.8. This plot shows that although the Monod equation provided a rea-
sonable description of the data (as also evidenced by the low SSQ value),
the trend of the data can also be interpreted to suggest Haldane kinetics.
Without having the luxury of additional confirmatory data (e.g., growth
rate data at higher COD concentrations would be helpful), it is prudent
to also fit the growth data to the Haldane expression and compare the
results with the Monod fit.
A computer program for fitting growth data to the Haldane equation
is given in the Appendix; it also utilizes the nonlinear fitting subroutine,
DUNLSF, that can be obtained from IMSL. The bounding rules for this
program are given in the Appendix. Fitting the growth data in Table 5.3
to the Haldane equation using the computer program in the Appendix
(EPS value of 0.00001) produces values of 0.272 hr-', 121 mg/L COD,
and 197 mg/L COD for y,,,,, K,, and K,, respectively with an SSQ value
of 0.0007. A comparison of the Haldane and Monod fits to the growth
data listed in Table 5.3 is depicted in Figure 5.9. This figure and the
supporting statistical data indicate that both functions provide a reason-
able description of the growth data. Without additional growth rate data
(at higher COD values), some engineering judgment must be applied to
reconcile this somewhat equivocal situation. Considering the implica-
tions for the prediction of effluent quality, the conservative approach is
to utilize the inhibitory version of the predictive model to make predic-
tions of effluent quality. For analytical completeness, one can use both
sets of constants in the predictive model and compare and contrast the
 OBTAINING BIOKINETIC CONSTANTS 93

0.125

0.100

0.075
(ht)
yw

0.050

0.025 O Growth Data

— Curve Fit (Monod Function)

0 30 60 90 120 150
So, mgCOD/I

Figure 5.8. Plot comparing the experimental growth data listed on

Figure 5.5 to the growth curve generated from applying the
constants obtained from a computer curve fit analysis of
the Monod function to Equation 1.11.

performance envelopes predicted by both the inhibitory and noninhibi-

tory versions.

DETERMINATION OF TRUE CELL YIELD, Y,, AND

DECAY RATE, k,
The complication with determining the values of true cell yield and
decay rate is that these biokinetic constants can only be evaluated using
data collected in relatively long-term pilot plant runs. Fortunately, the
variation of the values of these constants has a relatively small impact on
the prediction of effluent quality. The basic approach is to operate a
biological treatment system at several (at least three) different growth
rates (or 9,’s). The amount of excess sludge produced (or wasted) is
94 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.125

0.100

& 0.075
mT
Oo
0.050

O Growth Data
0.025 — Curve Fit (Monod Function)
~--- Curve Fit (Haldane Function)

0 30 60 90 120 150
So, mgCOD/|

Figure 5.9. Plot comparing the experimental growth data to both the
Monod and Haldane expressions using numerical values of
the biokinetic constants obtained from computer curve fits
of the experimental data listed on Figure 5.5.

noted at each different growth rate or 0.. Once sufficient data have been
collected, a “maintenance plot” is made of the data in order to determine
the values of Y, and k,. This analysis consists of a plot of observed yield,
Y,, versus net growth rate, yu, or O.. Because these analyses are relatively
routine, we will not go into detail in this text but instead provide several
references (Gaudy and Gaudy, 1988; Metcalf and Eddy, Inc., 1979;
Rozich, Gaudy, and D’Adamo, 1983). The information in the references
is sufficient to enable the reader to collect the type of data required and
to the perform the ancillary analyses to compute true cell yield and decay
rate. Rozich, Gaudy, and D’Adamo (1983) provide a relatively straight-
forward example for determining these constants from pilot plant data.
 OBTAINING BIOKINETIC CONSTANTS 95

KEY CONCEPT SUMMARY

The key concepts contained in this chapter are:

¢ The influent concentration of waste or waste strength, S,, is

defined as that fraction of the feed waste that can be biodegraded
by the reactor biomass. This fact is embodied in the ACOD test
and concept. The use of COD or TOC measurements to estimate
S; results in a conservative evaluation of S;, since influent COD or
TOC will always be greater than ACOD or ATOC.
¢ The biokinetic growth constants y,,,,, K,, and K; (if the waste is
inhibitory) are determined by fitting growth data (yu vs. S) to
either the Monod or Haldane equation (noninhibitory or inhibi-
tory, respectively). These growth data are generated using respi-
rometry. The respirometric data must be “translated” into equiv-
alent biomass concentration data before growth rates can be
computed. The biokinetic growth constants have a substantial
impact on the prediction of effluent quality from an activated
sludge system.
e True cell yield Y, and decay rate ky must be determined using
continuous flow tests. These constants primarily impact the pre-
diction of waste sludge production and have a relatively minor
impact on predicting effluent quality from an activated sludge
system.

REFERENCES AND SUGGESTED

ADDITIONAL READING

D’Adamo, P.C., Rozich, A.F., and Gaudy, A.F. Jr. (1984). “Analysis of
Growth Data with Inhibitory Carbon Sources,” Biotechnology and
Bioengineering, XXVI, pp. 397-402.
Gaudy, A.F. Jr., Rozich, A.F., Moran, N.R., Garniewski, S.T., and
Ekambaram, A. (1988). “Methodology for Utilizing Respirometric
Data to Assess Biodegradation Kinetics.” Proceedings, 42nd Purdue
Industrial Waste Conference, Lewis Publishers, Chelsea, Michigan,
pp. 573-584.
Gaudy, A.F. Jr., Ekambaram, A., and Rozich, A.F. (1989). “A Respiro-
metric Method for Biokinetic Characterization of Toxic Wastes,” Pro-
ceedings, 43rd Purdue Industrial Waste Conference, Lewis Pub-
lishers, Chelsea, Michigan, pp. 35-44.
96 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Gaudy, A.F. Jr., Ekambaram, A., Rozich, A.F., and Colvin, R.J.
(1990). “Comparison of Respirometric Methods for Determination of
Biokinetic Constants for Toxic and Nontoxic Wastes,” Proceedings,
44th Purdue Industrial Waste Conference, Lewis Publishers, Chelsea,
Michigan, pp. 393-403.
Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental
Engineering, Engineering Press, Inc., San Jose, California.
Metcalf and Eddy, Inc. (1979). Wastewater Engineering: Treatment, Dis-
posal, Reuse, McGraw-Hill, New York.
Rozich, A.F., Gaudy, A.F., Jr., and D’Adamo, P.C. (1985). “Selection
of Growth Rate Model for Activated Sludges Treating Phenol,” Water
Research, 19, pp. 481-490.
Rozich and Gaudy, Inc. (1988). “Manual-Instructions for Obtaining
Biokinetic Data and Determining Operational Guidelines for Acti-
vated Sludge Systems,” Engineering Report to the City of Baltimore,
MD, Wastewater Facilities Division.
 6 FACTORS AFFECTING
THE VALUES OF THE
BIOKINETIC CONSTANTS

INTRODUCTION

The purpose of this chapter is to review how various environmental

conditions can and do impact the values of the biokinetic constants
which quantify the growth potential which a biomass is capable of on a
target waste. It needs to be emphasized that the system ecology and
consequently the values of the biokinetic constants vary with changes in
environmental conditions, waste characteristics, etc. This should not be
viewed as a weakness with biological treatment technology. Conversely,
with the advent of respirometric techniques that provide a cost-effective
means of biokinetic constant measurement, engineers have methodology
which expedites model calibration. That is, as the ecology changes, it is
practicable to recalibrate the model and quantify the impact of new
environmental conditions on process performance. Engineers can effec-
tively view the values of the biokinetic constants as a measure of the
“acclimation state” of the biomass. Generally speaking, the higher the
maximum growth rate a population achieves on a target waste, the more
well-acclimated it is to that waste. It is also useful to think of the
biokinetic constants as a catalytic characterization of the system ecology.
That is, it is not practicable to isolate and characterize the plethora of
organisms that comprise a given biomass. However, as environmental
control professionals, we are more concerned with the degradative capa-
bility that can be expected by biomass on a given waste. The quantifica-
tion of the degradative capability as provided by the biokinetic constants
is a kinetic means of characterizing system ecology.
The environmental conditions to be reviewed in this chapter include:

97
98 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

e Reactor growth rate: The rate at which a biomass is grown has a

significant impact on the values of the biokinetic constants.
e Waste composition: The composition of wastewater has a huge
effect on the ability of populations to degrade target
components.
e Toxics: The toxic nature of a waste stream or other conditions
can adversely affect the ability of a biomass to degrade wastes.
¢ Temperature: Temperature affects both the values of the con-
stants and the types of wastes that can be treated.
¢ Population diversity: Microbial population diversity affects its
ability to respond to different waste treatment situations.

We will also discuss two topics in biological treatment that are often of
significance: shock loads and bioaugmentation
The topic of shock loads is important because many practitioners tend
to dismiss the applicability of process models for plant design and opera-
tion because the models are derived assuming steady-state conditions
while plants are generally in a dynamic situation. This section will point
out ways that the process models can be employed to analyze potential
shock load situations and develop compensatory strategies for dealing
with these occurrences.
The topic of bioaugmentation is one which will generally solicit a
jaundiced response from all but the most open-minded of environmental
engineers. We will discuss the principles of bioaugmentation, how these
relate to fundamentals of microbial ecology, and potential applications.
Bioaugmentation in theory is a valid technological option. For example,
a familiar application is the use of seed in the BOD test. Bioaugmenta-
tion has received a poor reputation in large part due to questionable
practices of some companies that sell microbial products to wastewater
treatment plants. It is the goal of this section to point out potential
approaches that can be utilized to make this practice more akin to engi-
neering practice than to commercial art.

SPECIFIC GROWTH RATE OF THE REACTOR

As discussed previously, specific growth rate in biological reactors is

equivalent to control parameters such as F:M, 9,., or U. It has been
observed over the years that system growth rate can significantly affect
the type of biomass and consequently the ability of a reactor to handle
certain wastes at certain rates. A simple example that comes to mind is
 VALUES OF BIOKINETIC CONSTANTS 99

nitrification. It is commonly accepted that a mean cell residence time of

8-10 d is needed to achieve nitrification. In terms of the system biomass,
this means that the growth rates have to be slow enough to enable nitrify-
ing bacteria to persist in the system at a level large enough to convert the
bulk of the ammonia or total Kjeldahl nitrogen (TKN) to nitrate.
A simple guideline to remember is that the faster or slower one grows a
microbial population, the higher or lower, respectively, the resulting
maximum growth rate will usually be. This consequently means that the
values of the biokinetic constants that influence maximum growth rate
will also be affected.
As an illustration of this point, consider Figure 6.1. This figure depicts
the actual and predicted values for effluent quality for heterogeneous
populations growing on glucose in a chemostat. During the course of this
work, several steady-state runs were made at different flow rates or

S, carbohydrate

a [o)oO
|

Xmg/l
S,

altel iF
0 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
D,h"

Figure 6.1. Dilute-out curves in S and X for increasing values of D

(reciprocal of t) for heterogeneous microbial populations
growing on glucose. S; = 3000 mg/L. Dashed lines show
dilute-out pattern calculated from Eqs. 2.9 and 2.10 using
experimentally determined values of y,,,,, K,, and Y, (ky
assumed to be 0). (Gaudy and Gaudy, 1988).
100 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

dilution rates. That is, the growth rate in the chemostat was adjusted by
changing the flow rate or dilution rate. The flow rate was held constant
until a steady-state condition (little variation in the observed values of
reactor biomass and substrate concentration) was achieved. An example
of steady-state data from this work is depicted in Figure 6.2. During each
steady state, biomass was harvested from the reactor and used as inocula
in separate batch growth tests. As illustrated in Figure 6.1, the predicted
and observed values for effluent substrate concentration deviate from
one another at higher dilution or reactor growth rates. This deviation
also been observed for pure culture studies and has been attributed to the
selection of subspecies which have higher y,,,, values.

S,
Si,
Xmg/l

0 2 4 6 8 10 12 14 16
Time, days

Figure 6.2. “Steady-state” values of X and §S, for feed values S, shown
in a once-through reactor operating at D = 0.33 bh".
(Gaudy and Gaudy, 1988).
 VALUES OF BIOKINETIC CONSTANTS 101

For the work described in Figure 6.1, evidence suggested that the faster
reactor growth rates selected organisms which had higher y,,,, values.
The corollary argument is that the faster system growth rates washed out
slower-growing species. This effect is analogous to activated sludge nitri-
fication in that lower ©, values (faster reactor growth rates) wash out
nitrifying bacteria, since these organisms cannot maintain sufficient
growth rates to persist in the system.
A more dramatic example of the impact of reactor growth rate on the
values of the biokinetic constants is found in some of the work which our
group performed regarding the biodegradation kinetics of toxic or inhib-
itory components, specifically phenol. During these efforts, a number of
different reactor types and configurations were employed in studying the
biodegradation kinetics of phenol. A chemostat, an activated sludge
process, and a two-stage continuous culture system were all used. This
variety of reactor types enabled us to utilize a wide range of reactor
growth rates for growing heterogeneous populations on phenol.
A comparison of the growth kinetics of cells utilizing phenol as a sole
carbon source in a chemostat and in an activated sludge reactor is pre-
sented in Figure 6.3. The activated sludge system was operated at various
steady states over a net growth rate range of 0.005 to 0.032 hr"! (0, range
of 1.3 to 7.7 d) while the chemostat was run over a growth rate range of
0.014 to 0.054 hr"! (©, range of 0.80 to 3.0 d). Figure 6.3 shows that, as
one would expect, the cells grown in the chemostat were capable of
achieving higher growth rates, which was reflected in higher values of the
maximum growth rate (u* for inhibitory systems). That is, as the growth
rate of the reactor was increased, it increased the value of the maximum
growth rate, reflected in changes of the biokinetic constants.
A more profound impact of reactor growth rate on the values of the
biokinetic constants was realized by growing heterogeneous populations
on phenol using the two-stage continuous culture system depicted in
Figure 6.4 (Colvin and Rozich, 1986). The initial intent for using this
reactor system was to develop a continuous flow technique to quantify
the growth kinetics of inhibitory wastes. Since a chemostat treating an
inhibitory waste will wash out once the flow rate produces a growth rate
equal to the p* of the biomass, it is not feasible to collect continuous flow
data on the right side of the inhibition curve using a chemostat alone. In
a two-stage continuous culture system (shown in Figure 6.4), the flow of
biomass from the first reactor enables biomass to persist in the second
reactor and avoid a washout. This enables one to collect steady-state
growth data on an inhibitory substrate at relatively high steady-state
102 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.15

0.10

Chemostat

~~~___ Activated Sludge

On ee, —.
— es,—
—
——.,
———

100 200 300 400 500

S, mg/L phenol

Figure 6.3. Comparison of phenol growth kinetics from biomass grown

in a chemostat (y,,,, = 0.19 hr-', K, = 7.9 mg/L, and K, =
139 mg/L, Colvin and Rozich, 1986) and in an activated
sludge system (p,,,, = 0.194 hr', K, = 48 mg/L, and K, =
62 mg/L, Rozich and Gaudy, 1985).

REACTOR 1 REACTOR 2

Figure 6.4. Schematic of two-stage continuous culture system. (Colvin

and Rozich, 1986)
 VALUES OF BIOKINETIC CONSTANTS 103

substrate concentrations which is not possible in a chemostat because of

washout due to exceeding the p* criterion.
Heterogeneous populations grown on phenol in the second stage of the
two-stage continuous culture system responded with greatly enhanced
growth rates, as illustrated in Figure 6.5. This figure shows that the
maximum growth rate (u*) of the cells grown in the second stage of this
system was approximately four times higher than that of the rate exhib-
ited by the organisms in the first stage. The technique of forcing higher
growth rates in the second stage of this system selected a biomass with
greatly enhanced growth kinetics, evidenced by the contrast in values of
the biokinetic constants.
For additional comparative purposes, consider Figure 6.6. This figure
simply plots the data in Figures 6.3 and 6.5 on one graph. The purpose of
this figure is to illustrate the range of potential growth rates and values of
the biokinetic constants that are possible for heterogeneous populations
metabolizing phenol. The changes in the values of the biokinetic con-
stants were achieved by selecting different microbial populations using a
variety of reactor configurations to manipulate system growth rate.

0 R2 Cells
° R1 Cells

0 200 400 600 800

S, mg phenol/L
Figure 6.5. Plot of average growth rates detected in batch growth
studies fitted to the Haldane Equation. Resulting values
fOr pmax» K,, and K, were 0.19 hrs", 7.9 mg/L and 1.39
mg/L, respectively, for the R1 Cells and 1.07 hr', 79
mg/L, and 172 mg/L, respectively, for the R2 Cells.
(Colvin and Rozich, 1986).
104 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.50

= Hi Rate Cells

0.40
i
a Rey

rs
oy
~
SS
0.30 ae
r ta
ne ~sAN.

=
=
0.20

0.10
Chemostat

— Activated Sludge
ee
eS eS cee

0 100 200 300 400 500

S, mg/L phenol

Figure 6.6. Comparison of kinetics of cells grown on phenol in a

chemostat (Figure 6.3) and in activated sludge (Figure 6.3)
with kinetics of cells grown in a high-rate system (u,,.. =
1.07 hr', K, = 79 mg/L, and K, = 172 mg/L, Colvin and
Rozich, 1986).

As aside note, it should be emphasized that the high growth rates were
achieved on the inhibitory or “toxic” organic phenol by simply using
heterogeneous populations and a relatively novel but readily usable cul-
turing technique. The point is that developing biological wastewater sys-
tems to degrade high-strength organic waste streams rapidly does not
require the use of “mutant” bacteria or genetically engineered organisms.
By employing the principles of reactor engineering and the knowledge
 VALUES OF BIOKINETIC CONSTANTS 105

that the range of growth kinetics of natural microbial systems is usually

larger than we suspect, environmental engineers and scientists can devise
biological waste treatment systems that are capable of a much greater
performance envelope than what is currently thought to be achievable.

WASTE COMPOSITION

Waste composition is a factor that one would intuitively expect to exert

an effect on the values of the biokinetic constants and consequently on
the performance of an activated sludge system. Specifically, this section
refers more to variations in biodegradable carbon source than to the
presence of materials, such as heavy metals, that may adversely affect the
ability of the biomass to degrade wastes; the impact of substances such as
heavy metals on the values of the biokinetic constants will be discussed in
the “Toxics” section. If one is concerned about the potential impact of
waste composition changes on the performance of a biological treatment
system, it is a simple task to evaluate the potential changes in the values
of the biokinetic constants and to use the model to predict the impacts on
plant performance.
An important point about a target wastewater is that the effects of the
composition are not readily predictable. For example, an old axiom in
the environmental field was that a biodegradable inhibitory waste stream
such as phenol was easier to handle with the presence of a more degrad-
able carbon source such as glucose. From a reactor engineering perspec-
tive, this implies that the presence of the alternate carbon source gives the
biomass a stronger capacity to metabolize the phenol; this particular case
lends itself as a good illustrative example regarding the difficulty of
predicting the impact of waste composition on biomass capability to
degrade target components.
Our group became interested in the situation presented in the preced-
ing paragraph. That is, is it appropriate and/or beneficial to mix an
easily biodegradable carbon source with a difficult or inhibitory one in
order to enhance the treatability of the inhibitory carbon source? This
work (Rozich and Colvin, 1986) was a corollary to other efforts con-
ducted regarding the delineation of the biodegradation kinetics of phenol
and other toxic or inhibitory compounds.
For this work, two different types of biomass were cultivated in bench-
scale reactors: one acclimated to phenol as a sole carbon source and one
acclimated to a mixed carbon source consisting of phenol and a more
easily degradable compound, glucose. An initial set of tests were con-
106 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

ducted to assess the influence of glucose on the ability of the cells solely
acclimated to phenol to degrade phenol. Representative results are
shown in Figures 6.7 and 6.8. In each of these tests, two concurrent batch
reactors were run using cells acclimated only to phenol; one reactor
received only phenol while the other reactor received a mixed waste
consisting of phenol and glucose. Interestingly enough, these results
show that the glucose, which is representative of an easily degradable
waste, hinders phenol removal. This is stated because the results showed
that phenol removal rates slowed down in the reactors that received the
glucose. Additionally, and somewhat surprising, the data show that the
cells acclimated only to phenol preferentially removed phenol while not
metabolizing the glucose. These data, along with other work in our
laboratories, suggested that adding a more easily biodegradable carbon
source to a toxic waste can actually decrease the ability of the biomass to
degrade the target waste. These statements apply, of course, to a biomass
that degrades a toxic organic waste (e.g., phenols and other aromatics)
when using it as a sole source of carbon and energy and not to co-
metabolic situations.
Other tests were performed using the biomass which was acclimated to
a mixture of phenol and glucose. Typical results are shown in Figures 6.9
and 6.10. These tests showed, as one would expect, that the biomass
preferentially removed glucose. That is, with the shift in acclimating

Yegr

Roby Fi) 19 soty 28

o Phenol COD (Central)

© Phenol COD
¥v Glucose COD (Anth)
9 Glucose COD (Enzy.)

Analysis
Indicated,
mg/L

TIME, hours

Figure 6.7. Mixed phenol/glucose substrate removal by phenol

acclimated cells. Initial biomass concentration = 150
mg/L. Glucose assessed utilizing both anthrone and
enzymatic methods. (Rozich and Colvin, 1986).
 VALUES OF BIOKINETIC CONSTANTS 107

_ Glucose Added

v v
ole Fv eo Bvig

oe Phenol COD (Central)

@ Phenol COD
wv Glucose COD (Anth)
© Glucose COD (Enzy.)

mg/L
Indicated,
Analysis
TIME, hours

Figure 6.8. Mixed phenoi/glucose substrate removal by phenol

acclimated cells. Glucose fed to mixed waste reactor 1.5 h
after cells were actively utilizing phenol. Initial biomass
concentration = 150 mg/L. Glucose assessed utilizing
both anthrone and enzymatic methods. (Rozich and
Colvin, 1986).

conditions, a change was realized in the acclimation state of the biomass

that resulted in the preferential use of the more easily degradable waste.
However, these results and others did not show that the biomass exhib-

© Phenol (Control)
4 Glucose (Cont. Enzy.)
O Phenol
9 Glucose (Enzy.)

SUBSTRATE
COD,
mg/L

TIME, hours

Figure 6.9. Mixed phenol/glucose substrate removal by phenol/glucose

acclimated cells subjected to nonproliferating conditions.
Initial biomass concentration = 400 mg/L. (Rozich and
Colvin, 1986).
108 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

oe Phenol (Control)
a Phenol
3 c Glucose (Cont. Enzy.)
¥ Glucose (Enzy.)

as8

SUBSTRATE
COD,
mg/L
So

TIME, hours

Figure 6.10. Mixed phenol/glucose’ substrate removal by

phenol/glucose acclimated cells. Initial biomass
concentration = 300 mg/L. (Rozich and Colvin, 1986).

ited an increase in the values of the biokinetic constants on phenol.

Conversely, this line of work in our laboratory indicated that popula-
tions acclimated to both phenol and glucose had a decreased ability to
degrade phenol as indicated by the values of the biokinetic constants.
The above statements are reinforced by the comparative kinetic
results, which are shown in Figure 6.11. This figure compares the kinetic
characteristics of phenol on organisms that were cultivated in a two-stage
continuous culture system; the lower kinetic response was demonstrated
by the cells accclimated to a mixture of phenol and glucose while the
greater growth potential on phenol was realized by the biomass grown
using phenol as a sole source of carbon. The bottom line is that the
addition of a more easily degradable carbon source is not necessarily
going to make a toxic waste more degradable, unless, of course, there is a
co-metabolic requirement for the alternate carbon source to achieve
biodegradation of the target toxic organic. In point of fact, the addition
of a more easily degradable carbon source may actually lower the kinetic
ability, as quantified by the values of the biokinetic constants, of a
biomass to degrade certain difficult-to-degrade or toxic wastes.
Another reason why waste composition effects are not readily predict-
able is due to the potential effect of metabolic control mechanisms
 VALUES OF BIOKINETIC CONSTANTS 109

O PHENOL
0.50 © PHENOL/GLUCOSE

0 200 400 600 800

S mg/L phenol

Figure 6.11. Comparison of phenol batch growth rates for cells

cultured using phenol as a sole carbon source and using
the mixed phenol/glucose waste. (Rozich and Colvin,
unpublished data).

(Gaudy and Gaudy, 1980). Metabolic control mechanisms occur when a

microbial system will preferentially degrade one carbon source over
another. In one aspect of these mechanisms, organisms repress the manu-
facturing of enzymes needed for degrading a target carbon source
because of the presence of an alternate source. This is termed repression
because the cells attenuate the synthesis of enzyme required for metabo-
lizing the one carbon source. In another aspect, the presence of an alter-
nate carbon source results in a lowering of enzyme activity. This is
termed catabolite inhibition when the alternate carbon source inhibits the
activity of the enzymes needed for degrading the target carbon source
(Gaudy and Gaudy, 1980).
The effects of metabolic control mechanisms in batch systems are
illustrated in Figure 6.12. Cells were subjected to resting conditions
wherein the nitrogen source was withheld to prevent net synthesis of
protein. This figure shows that rapid blockage of the removal of both
 []25 “DU0d
JB -4SOYS 110
BuIpeo]
= OSE 1/8W
(7) [80L d0Oo
(G) ss0dn|y
dO)
| (auoryqu
| e)
(©) louuuew
dO)
| (ayeporsad)
(a) JOyIUUeWJO1UOD
asoon[yPappe
ye $°Z SINOY

(v7) [ROL GOO

\ 1210) Od ) (a @soonidO)
y
\ | (N+9) (auo1yjue)
NM (0) 191!1G1009
§
||| Pai ) |
(4)
(@1epou
| sad)
JOJIQIOS JO1]UOD
Ea Qo (21epotad)

\
BENENSNEEE pepe)
assoon|yPappe
BS je ¢"¢ SINOY

S
r=)
SENSE

Gj
°
ina

J/8wW — (GOO se) PayeoIpul uoTeurUua}aq

v
So
i=)
4
10}1q1
\ 05
JoyluueW
ey)
ae
ebdg| (NNa?
JO1]U09
aE
7/8wW (Od se) payesipur UO! PUIUID}0G
fa

0
(4 € v S 2 L

0
Bae 6
‘out
y
‘ouny
Y
ainsi‘Z[°9 aseypo
JO lg [eAOUI
Jo II [O}tuUQJ9])
vU pu [OJIGI0
St)S (JY Aq ‘asoonysyoysem paysaf
Sv &ur Sis
ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

asop OUI ay) WNIpaUT BuLNp Yy)MOI3

UO ay) sesns ‘S[Oyoo]e ‘<pney) ‘14jow
pus
Woy ‘Apney ‘(p96
 VALUES OF BIOKINETIC CONSTANTS 111

mannitol and sorbitol was achieved after injection of glucose into the
medium, which demonstrates the attenuation of enzyme function due to
the presence of glucose. Mannitol and sorbitol utilization did not resume
until the glucose was metabolized.
The impact of metabolic control mechanisms may be most relevant
regarding qualitative shock loads (i.e., situations involving a sudden,
radical change in the composition of the influent wastewater). As an
example, consider Figure 6.13. This figure depicts the result of changing
the influent waste stream from glycerol to a combination of glycerol and
glucose and the impact of changing a waste stream from sorbitol to a
mixture of sorbitol and glucose. The results show that in addition to
leaking glycerol and sorbitol in the respective systems, both reactors also
leaked a significant quantity of product (nonsubstrate) COD. These
results suggest that a sudden and significant change in waste quality will
cause leakage of both influent waste COD and product COD.

TOXICS

Many times the presence of substances such as heavy metals, non-

biodegradable organics, salts, etc. can by themselves or in conjunction
with other materials produce a toxic effect on the ability of a biomass to
degrade target wastes. This effect differs from the previously discussed
phenomenon of substrate inhibition in that the toxic material in this case
is not being biodegraded, but it exerts a detrimental effect on the ability
of the biomass to metabolize the target waste. As one should expect, this
toxic effect, which presumably will depress system performance, is pre-
dictable if the impact of the toxic materials on the values of the
biokinetic constants is quantified.
As a generic example, consider Figure 6.14. This figure depicts two
idealized curves showing biomass growth kinetics, i.e., » as a function of
waste or substrate concentration. The curve showing the higher growth
rates represents the kinetic ability of the biomass without the presence of
the toxic material; this represents the baseline case. The other curve
quantifies the degradative capability of the biomass in the presence of a
toxic material and shows, as would be expected, lower growth rate poten-
tial. Once these data are generated, the values of the biokinetic constants
are inserted into the process model and the impact of the toxic material
on the performance or the operating envelope of the biological waste
treatment system is quantified.
 [eoIdojo1g
Spilos

112

[Ov
==

a
1/3 "paredipul uoljeulmiajag

7 /8w ‘pareotpur uon eulwajaq

ee ii

£
B-
yf
Goo
CleeVal

ry
Dal

asooni9\
os I

0 b
iL

8 a 91 0z 74 87
owiy
y ‘owny
y
BNsIy
"9 “E[ YIOYSPEO] asuOdsal
JO SNOIUIZ019}9Y
[vIGo1DIWI SUOHENdOd SUIMOISJapuN Apea}s9)8}s SUONIPUOD
C38= 4/1
JOLId
0} & adUEYD
UT SuIMO[JUI *2}849Sqns(3J2]) QW‘Y ay) paey UONISOdWOD
sem pasuvyd
Woy QOS T/SuI
JO [O419d4[S
0} 00S
T/3uJo jo1994[3
snd QOS, T/3wW Jo ‘asoonis
ayy paysep‘aul paddy eso0dn|s‘GOD sjuasasdas
ay) asoonjsTOD JEY) PINOM
JALY Usaq JUaSeId
JI JI 319M jou ‘pazijoqujow
paddy [01294[8QD SI BY) UOIL1}UIIUOD
JO [0499A4]3
GOD JEY) P[NOM PABY U99q
JUaSoId
JI ay) WSI[oquJaW
JO [01994[3
py paddojs
Je ay) aUIT)JO SuLJa)sIUIUIpE
94) “YOYS GYUSIY)FY ‘YO ay) pray UOHISOdUIOD
ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

sem pasuvyo
Woly QOOL /SW JT Jo [OJIGIOS
0} YYYZ T/BW JO ‘[0}IGIOS
PUL KOOL /3UJ ISOIN[S
SBM PIJIaLUT APJIIIIP
OFUT IY} 10}9891
su & Snjs ‘asop ay], paysepsaul] pajaqe| pardde asoon{s pus pat|dde [OJIGIOS JUISIId9J
9Y) SUOTBIQUIIUOD
JO JY} OA) SPUNOAUIOD
jEY} P[NOM BABY UVIq jUBSIId
JI Ady) peYy jou usaq *pazijoqej}auJIJOWOY)
pue ‘Apne ‘(9961
 VALUES OF BIOKINETIC CONSTANTS 113

BASELINE

TOXICS

S, mg/L
Waste Concentration

Figure 6.14. Idealized growth curves illustrating the impact of toxicants

on growth kinetics.

As specific examples, we will show the results of the impact of toxic

effects for three activated sludge systems:

¢ A domestic waste treatment system that was affected by an

aggregate impact of several industrial discharges.
e An industrial waste treatment facility that was handling an easily
biodegradable organic waste but that had to contend with the
effects of the heavy metal cobalt.
e An oxygen-activated sludge facility that was plagued by carbon
dioxide inhibition due to CO, accumulation in the reactor
headspace.

Figure 6.15 compares the kinetics of the biomass in a municipal waste

treatment plant receiving a relatively heavy industrial waste input with
that of a system with a light industrial contribution. For some time, the
staff of the former plant had been complaining that the plant was being
besieged by toxic inputs. However, when kinetic analyses were per-
formed, they indicated that the growth kinetics adhered to the “conven-
114 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.50

Stillwater Plant Kinetics

0.40

0.30

hr"
u,
0.20

0.10

0 100 200 300 400 500

S, mg/L COD

Figure 6.15. Comparison of kinetics for a municipal plant receiving

little industrial waste (y,,,, = 0.50 hr-', K, = 63 mg/L)
(Stillwater) and one receiving a heavy industrial waste
contribution (Patapsco) (u,,, = 0.18 hr', K, = 176
mg/L).

tional” or Monod-type kinetics. This made it somewhat troublesome to

argue that the plant was in a difficult treatment situation that was not
representative of a “typical” domestic waste. The comparison in Figure
6.15 of the values of the biokinetic constants of the former plant with
those of the plant receiving a light industrial input quantified the differ-
ence in biodegradative characteristics, enabling us to make a convincing
case that this treatment situation was not comparable to a typical domes-
 VALUES OF BIOKINETIC CONSTANTS 115

tic waste and that the design and operation of the plant warranted special
consideration. Subsequent process analyses using the model and the val-
ues of the biokinetic constants that were generated as part of this work
reinforced this point and defined system design and operating require-
ments (Rozich and Gaudy, Inc., 1986).
The case of the cobalt was more straightforward. For this work (Con-
stable et al., 1991), a biomass was acclimated to a synthetic waste mix-
ture that simulated the organic composition of the influent waste stream.
Two comparative biokinetic tests were performed using the acclimated
biomass; one test utilized the organic waste mixture and the other test
employed the organic waste mix, which was amended with | mg/L of
cobalt. The results are depicted in Figure 6.16. This figure shows that the
presence of the cobalt cut the biodegradative potential of the biomass by
almost a factor of two. As was the case with the previous example,
follow-on process analyses using the values of the biokinetic constants
and the process model quantified the impact of the presence of the cobalt
on both plant performance and capacity.
A case of CO, inhibition presents an interesting example of a toxic
effect inhibiting biomass ability to degrade waste. The efficient transfer
of oxygen in oxygen-activated sludge systems is due to the fact the oxy-
gen is delivered as a relatively pure stream whereas in conventional sys-

0.07

ti _No Cobalt
0.05

_1 mg/L Cobalt

_10 mg/L Cobalt

(1/hours)
Growth
rate

09 02040608 112141618 2
(Thousands)
Synthetic waste conc. (mg COD/L)

Figure 6.16. A series of respirometer studies were performed to

quantify the impact of cobalt on the operation of a
biotreatment process (Rozich and Coivin, 1990).
116 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

tems air is the carrier gas. For process economy reasons (generation of
pure oxygen is expensive), activated sludge reactors in these systems are
covered to maintain a relatively high oxygen concentration in the carrier
gas in order to maintain efficient oxygen transfer kinetics. Covering the
reactors also means that CO, concentrations can accumulate in the head-
space above the liquid level of the reactors, which results in relatively
high CO, concentrations in the mixed liquor. Work performed at the
Patapsco Wastewater Treatment Plant in Baltimore, Maryland (Martin,
1988) demonstrated the adverse effects of CO, inhibiton on the kinetic
capability of the biomass to remove COD. Figure 6.17 shows the results
of tests performed to measure the influence of the percent of carbon
dioxide saturation on inhibiting COD removal by the plant biomass.
Although the data were somewhat scattered, they did indicate a clear
impact of CO, on COD removal. Other ancillary tests were performed to
validate that the inhibition was due in large part to the elevated levels of
CO, in the mixed liquor and not attributable to depressed pH, which will
also result from CO, accumulation.
Various toxic materials can by themselves or in synergy with other
components depress the ability of system biomass to degrade target
wastes. This effect can be quantified via comparative evaluation of the
biokinetic constants. This information can then be used in conjunction
with the process model to provide quantitative data for making manage-
ment decisions regarding the removal or pretreatment of the toxic mate-
rials and potential cost/benefits for the purpose of enhancing the perfor-
mance of the biological waste treatment system.

TEMPERATURE

Temperature is well-recognized by microbial kineticists as having a

significant effect on microbial growth kinetics. Nitrification is a good
example of temperature effects. Nitrifying organisms have a relatively
small temperature range with 10°C being the low end and approximately
45°C defining the high end. Consequently, operators are concerned when
they must achieve winter nitrification. In contrast, one of the advantages
of the autothermal aerobic digestion (ATAD) process is that the high
operating temperatures suppress nitrification, which means that these
systems have a lower aeration demand than conventional aerobic
digestion.
Practically speaking, the basic approach for quantifying the impacts
of temperature on the treatment capability of a system is to delineate the
 VALUES OF BIOKINETIC CONSTANTS 117

REMOVAL
COD
OF
INHIBITION
PERCENT

PERCENT CARBON DIOXIDE

Figure 6.17. Inhibitory effect of increasing carbon dioxide

concentration on removal of chemical oxygen demand
by activated sludge (Martin, 1988).

effect of temperature on the values of the biokinetic constants. A good

case in point is some work which was performed on the biodegradation
of a landfill leachate (Rozich and Colvin, 1988). The engineers on the
project were concerned with the potential performance of the system
under winter conditions. The landfill was located in Canada where low
temperatures (~ 10°C) are commonplace. The methodology employed
consisted of evaluating the biokinetic constants at normal temperatures
of 20°C and comparing these results with a kinetic evaluation performed
at 10°C. (Most respirometers have the feature of temperature control
and obtaining kinetics at different temperatures is relatively easy.) The
118 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

results of the comparative kinetic evaluations are shown in Figure 6.18

and indicate, as one would expect, that the lower temperatures exert a
significant impact on the values of the biokinetic constants; these data
are then used to predict the effect of lower temperatures on system
performance.
Finally, some comments are in order regarding the application of
autothermal systems for high-strength (=>30,000 mg/L COD) waste
treatment. Autothermal aerobic digestion systems (or ATAD) are self-
heating aerobic biological treatment processes which operate at high
temperature ranges (45°C to 70°C). Needless to say, microbial degrada-
tion rates at these temperatures will be quite high, and higher tempera-
tures will also result in lower sludge production because cell yields drop
with increased temperatures.
Conventional engineering strategy tends to lend itself toward utilizing

0.06

0.05 rT.

hr”
u,
So So(a)

0.01 7°¢

0 100 200 300 400 500

S, mg/L COD

Figure 6.18. Impact of temperature on kinetic response of activated

sludge acclimated to landfill leachate.
 VALUES OF BIOKINETIC CONSTANTS 119

anaerobic systems for high-strength waste applications. If methane utili-

zation is economical, then this seems to be the prudent choice. However,
it should be pointed out that autothermal aerobic systems can be mod-
eled with the same approach as that used for aerobic systems, which
Operate at conventional temperatures. (That is, operate a bench-scale
pilot reactor, evaluate the biokinetic constants utilizing respirometry, and
use the model to perform the value engineering analyses of the system.)
We suggest that environmental technologists consider the autothermal
aerobic process as a potentially viable alternative because work in our
own laboratories (Gaudy, Rozich, and Garniewski, 1987) indicated the
potential for improved treatment for a high-strength landfill leachate
when utilizing an autothermal thermophilic system. The key point is that
the predictive tools are available for engineering high-temperature aero-
bic treatment processes. Other work (Rozich, A.F., Clay (nee Gar-
niewski), S.G., and Colvin, R.J., 1991) showed that a bench-scale ther-
mophilic aerobic system operating at 50°C demonstrated excellent COD
removal when treating a high-strength ground water containing 200,000
mg/L COD. Ancillary respirometric tests were performed at 50°C to
quantify kinetics and define the process operating envelope of the sys-
tem. Thus, ATAD systems may warrant more serious consideration for
treatment applications that have traditionally been exclusively reserved
for anaerobic technology.

POPULATION DIVERSITY

Population diversity refers to the diversity of the ecology that makes

up a biomass. A pure culture consisting of one species of micoorganism
is essentially devoid of diversity. In contrast, an activated sludge system
with a long ©, will not only have a plethora of single-cell bacteria, fungi,
protozoa, etc., but it will also have multicelled creatures, which are
representative of the highest trophic level in the system ecology. The level
of diversity exerts an influence on the types of waste a biological system
can handle and the rates at which the wastes can be processed; this, of
course, is quantified via the biokinetic constants. In general, it is prob-
ably safe to state that the more specialized, or less diverse, a biomass is,
the greater the growth kinetics will be on a target waste or compound;
that is, the biomass will have an enhanced capability to degrade a partic-
ular waste stream or compound. As a corollary, it is also probably safe to
state that as populations grow more diverse, they gain an enhanced
120 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

capability to deal with a greater diversity of target compounds in a waste

stream. Examples will be discussed below.
In a previous section of this chapter, we discussed the use of a two-
stage continuous culture system depicted in Figure 6.4 to study the kinet-
ics of heterogeneous populations metabolizing phenol. It is interesting to
note that workers (Jones, 1973) used this same reactor system to evaluate
the inhibition kinetics of phenol of a pure culture of bacteria. This group
did not report enhanced phenol degradation rates in the second stage of
the two-stage reactor. That is, the cells growing in the second stage of the
system shown in Figure 6.4 at high phenol concentrations demonstrated
inhibited growth kinetics and exhibited lower growth rates than the
organisms in the first stage of the system. This contrasts sharply to the
results observed for heterogeneous populations which were cultivated on
phenol in the two-stage system; that is, the heterogeneous, or more
diverse, biomass exhibited substantially enhanced growth kinetics in the
second stage of the reactor system depicted in Figure 6.4.
Figure 6.19 compares the phenol growth kinetics of the pure culture
and those of the heterogeneous cells cultivated in the second stage of the
two-stage system. This figure quantifies the differences in growth kinet-
ics between the pure culture and the heterogeneous biomass. It is evident
that using a more diverse heterogeneous population for the two-stage
continuous culture work produced a biomass with enhanced growth
kinetics on phenol, due in large part to the fact that the use of a more
diverse culture provided a larger pool of microbial species with which to
select cells that exhibited enhanced phenol degradation kinetics. As pre-
viously stated, less diverse populations are less likely to be capable of
handling relatively diverse waste streams. Consider the data presented in
Figures 6.7 and 6.8; these figures show the response of a biomass that
was acclimated to phenol as a sole source of carbon to a mixed waste
consisting of phenol and glucose. This biomass can be considered as
relatively specialized and lacking in diversity, especially when compared
to the populations grown in, for example, a domestic wastewater treat-
ment facility which has a waste stream with a large variety of carbon
sources. These figures show that, although the added carbon source,
glucose, is easily degradable, the biomass ignored and preferentially
metabolized the phenol. If this were a full-scale plant, the glucose would
show up in the plant effluent as BOD and the facility would likely
experience a permit violation. Thus, the more specialized phenol popula-
tions were not capable of handling a more diverse waste stream consist-
ing of phenol and the relatively easily degradable compound, glucose.
 VALUES OF BIOKINETIC CONSTANTS 121

0.60

0.50
Ven nO SAO Cells, Heterogeneous Population
/ ~~._ (Colvin and Rozich, 1986)
0.40 / ee
4 | oe

£ 0.30 / 2 RG
=?

Two-Stage Cells, Pure Culture

(Jones, 1973)

0 100 200 300 400 500

S, mg/L COD

Figure 6.19. Comparison of phenol growth kinetics in a two-stage

system of heterogeneous populations (u,,,, = 1.07 hr', K,
= 79 mg/L and K, = 172 mg/L, Colvin and Rozich,
1986) and a pure culture (u,,,, = 0.29 hr', K, = 1 mg/L,
and K; = 110 mg/L, Jones, 1973).

SHOCK LOADS

The topic of shock loads conjures up feelings of dread for any waste-
water treatment plant manager or operator, especially those involved
with biological wastewater treatment facilities. The purpose of this sec-
tion is to provide some discussion regarding the key features of shock
loads as they relate to biological wastewater treatment operations. We
will then discuss ways in which the modeling approach described in this
book can be applied to formulate proactive strategies that can attenuate
process upset due to a shock loading situation.
It is first prudent to define what we mean when we say “shock load.”
There are generally three different types of shock load:
122 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

e qualitative
¢ quantitative
© toxic

A qualitative shock load consists of a radical compositional change in the

nature of the target compounds in the influent waste stream. For
example, the results depicted in Figures 6.7 and 6.8 are an example of a
qualitative shock load to the phenol-acclimated cells since this biomass
was challenged with a mixed phenol/glucose waste. A quantitative shock
load entails a quantitative mass loading increase in the target waste com-
ponents in the influent waste stream; this increase can be caused hydrau-
lically by an increase in waste flow rate or by an increase in the concen-
tration of the target waste components. A toxic shock load involves the
sudden discharge of substances (e.g., heavy metals, high salt concentra-
tions, etc.) that can depress the ability of the biomass to degrade the
target wastes.
The potential response of the biological system to all of the above
situations can be assessed by evaluating the impact of the new waste
condition on the values of the biokinetic constants. Once the constants
are defined, they are inserted along with appropriate loading and opera-
tional information into the process model to predict the system capacity
or performance envelope that can be expected for the new conditions.
A common criticism of model application is that the model presented
herein is derived assuming steady-state conditions, while a shock load
situation is a dynamic event requiring a model to account for the
dynamic conditions. This criticism is valid in part, because the model is
in fact derived using a steady-state assumption, but this does not invali-
date the application of the model for predicting ultimate system response
to a new operating condition or for formulating strategies vis-a-vis modi-
fications of operational parameters to buttress the plant against potential
upset due to a shock load event. The key point to recognize is that a
shock load event entails going from one steady state to another; the pre-
shock condition represents the “old” steady state while the postshock
condition represents the new steady state. The new set of biokinetic
constants defines a “worst case” scenario for the biomass at the new
condition since the biomass has not had a chance to acclimate to the
waste change. Using the biokinetic constants, the model, and the new
loading condition, we can predict system response performance at the
new steady state.
It should be noted that a substantial amount of investigational work
(Gaudy and Gaudy, 1980) has shown that a biomass frequently demon-
 VALUES OF BIOKINETIC CONSTANTS 123

strates an extra assimilative capacity during a shock load situation. That

is, extensive controlled shock load tests have shown that systems do not
immediately leak soluble substrate during the transient state between the
old and new steady-state conditions. An example is given in Figure 6.20.
This figure shows a bench-scale activated sludge reactor that was treating
a waste stream consisting of glucose as the sole carbon source and sub-
jected to a sixfold quantitative shock load in influent COD concentration
(S; went from 500 to 3000 mg/L). It is evident in this figure that although
there was effluent deterioration due to suspended solids in the clarifier
effluent (an effect which we are not able to predict), little leakage of
COD occurred; the data suggested that the extra capacity was due to an
oxidative assimilation mechanism that enabled the cells to take in and
store the excess carbon during the shock load condition. This mechanism
attenuated soluble COD leakage and provided an extra “buffer” not
predicted by the model.
Conclusions regarding “extra” assimilative capacity may be questioned
because the test results shown in Figure 6.20 were generated using the
relatively biodegradable carbon source, glucose. Figure 6.21 shows the
results for a bench-scale activated sludge process that was treating the
toxic organic phenol and was subjected to a twofold quantitative shock
load (S, went from 1000 mg/L to 2000 mg/L phenol) (Rozich and Gaudy,
1985). As indicated by the data, the system exhibited little leakage of
soluble COD or phenol; however, 7 d after initiation of the shock load
condition, the system realized a catastrophic failure manifested by mas-
sive leakage of both soluble COD and phenol (800 mg/L). The steady-
state version of the model accurately predicted this failure. We attributed
the observed extra assimilation capacity, which “bought time” before the
failure, to oxidative assimilation; other work in our laboratories (Rozich
and Lowe, 1985) showed that microorganisms can remove phenol from
solution via this mechanism.
The key point with the examples in Figures 6.20 and 6.21 is that work
to date indicates that cells will frequently exhibit an extra sorptive capac-
ity with regard to substrate or waste removal during the transient in a
shock load situation. This means that model predictions for process
performance in the new steady state will be conservative with respect to
soluble COD leakage; that is, the model predictions can be utilized to
“bound” expected performance results for the aftermath of a shock load
or, from a proactive posture, be used to formulate strategies to attenuate
the likelihood of permit violation due to COD leakage. It should be
noted that the model presented herein or any other model is not capable
of predicting the impact of a shock load to biomass settlability and the
124 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

=
-
2
oN
26cal
x
ee
=F3s
N
2 on S&S

~
dedi = ie)
et
Ss 8
a
i
ll “es
a
xE
— 7=—
2%
WwW
aS
ss
ls
3
3 =m
:s| a
oe
e £nMn

ar asS
gs £:
s g&
€
=
osioe)
Eo
Oo =

i?a) mo
-—

S
aw
Sg,
so
ss
— OS
DA tm
—
.—
£3
Ss
iro
Sy

pi
i]
=)
s =
Ss 3
9 =

Se
vT
me
£ ¢

a ee)
oO

6.20.
Figure
 SOLLSIHSLOVYVHO LNANISS3

X "Sexe(joueud)
(doo) 's *"='s4a00)'s yOu o00z' s +7/6u oo0t 's

R
01
raoeeees

8
(Ud) Ss

V6u's “*s *°X

°
90
SOLLSIHSLOVYVHS YOLOVSY

g 8
T6u “*s
SSVWOI

V6u'x
SOLLSIHSLOVYVHS
VALUES OF BIOKINETIC CONSTANTS

pow “x
‘SLL sAep
aunsiy "179 SSO]JO SsEUIOIG
pu 3JBIJSqNS
WOIJJY) 10}dI1
| YIM 49)Je Suldjdde
& days asvasour
ul jousyd
125

UOHeQUIDU09
Wosy NOOT
0} ONT /3UI“J YdIZ0Y)
puv ‘Apne “(S86I
126 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

subsequent potential for suspended solids losses in the secondary clari-

fier. If settlability problems persist, it is recommended that alternate
strategies (e.g., the use of chemical or flocculant aids or the use of
selector technology) be evaluated to address this situation.

BIOAUGMENTATION

Bioaugmentation generally refers to the use of microbial additives for

the stated goal of enhancing the performance of biological wastewater
treatment systems. Operationally, it can be thought of as any situation
where microbes are introduced to enhance or speed up a biological reac-
tion. The BOD test is a good example, since seed from a primary clarifier
effluent is added to a BOD bottle to insure that sufficient oxygen deple-
tion occurs in a 5-d period.
The efficacy of bioaugmentation has been shown to be valid under
controlled conditions. For example, Edgehill and Finn (1983) demon-
strated that a bench-scale activated sludge system that was dosed with a
small side stream of organisms preacclimated to pentachlorophenol
(PCP) could take a PCP shock without leaking PCP. A duplicate unit
that did not receive the organisms responded to the shock situation by
leaking a substantial portion of the PCP. This laboratory demonstration
clearly showed that the concept of bioaugmentation is tenable.
Assuming that the concept of bioaugmentation is valid, the next step is
to determine if a product (“bugs” from a vendor) will achieve any more
on a particular waste stream or component than the biomass which is
indigenous to a particular reactor. This is a performance question and
can be answered by determining the kinetics of both the indigenous
reactor biomass and the product on the target waste or stream; the
respirometric methodologies and ancillary analyses described in Chapter
5 can be used to quantify the y,,,,, K,, and K; (inhibitory situation) of
each biomass. Care must be exercised in performing these tests to insure
that each the initial biomass concentrations are quantified. For example,
if one were to test a product with an inordinately high concentration of
cells, the sample would show superior kinetics based on respirometry
alone. However, once the data are translated into equivalent biomass
concentrations and the corresponding growth rates are calculated, a true
kinetic comparison can be made.
The key point is that, in some cases, bioaugmentaton may be a useful
operational tool for difficult streams or as backup to existing operational
protocols. The first question is whether the product organisms are any
 VALUES OF BIOKINETIC CONSTANTS 127

better at degrading the target streams than the existing biomass; this
question can be answered with a comparative respirometric test and
ancillary kinetic analyses. The next question is economical. Is it economi-
cal to use the bioaugmentation product vs. using some other design or
operational technique to achieve the target treatment goals?

KEY CONCEPT SUMMARY

The key concepts contained in this chapter are:

e The values of the biokinetic constants are all influenced by vari-

ous environmental and operational conditions. The key point is
that this variation can be easily quantified using the respirometric
techniques described in Chapter 5. The values of the biokinetic
constants and the engeineering models presented in Chapter 3
enable the process analyst to quantify the performance envelope
for a biological treatment system for a target environmental or
operatonal condition.
¢ The growth rate of a biological reactor (or the equivalent F:M or
6.) imparts a significant impact on the values of the biokinetic
constants. Faster-growing systems will have a tendency to select
for populations with higher y,,,, or u* (if the waste is inhibitory)
values; the opposite also holds for slower-growing systems.
Although a system may be ultimately capable of attaining a
higher yp... Or “* value, it is important to realize that reactor
performance is dictated by kinetic capability of the organisms
which are currently inhabiting the reactor.
¢ The composition of the waste stream plays a significant role in
determining the kinetic capability of a biomass to handle certain
components.
¢ Toxics such as heavy metals, organics, or high concentrations of
salt impact reactor performance by depressing kinetic capability.
Temperature also exerts a significant effect on the values of the
biokinetic constants.
¢ Greater diversity in a microbial population gives it the capability
to degrade a greater variety of waste components and to respond
to changing waste treatment situations.
128 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

REFERENCES AND SUGGESTED

ADDITIONAL READING

Colvin, R.J., and Rozich, A.F. (1986). “Phenol Growth Kinetics of Het-
erogeneous Populations in a Two-Stage Continuous Culture System,”
J. Water Poll. Control Fed., 58, pp. 326-332.
Constable, S.W., Rozich, A.F., DeHaas, R., and Colvin, R.J. (1991).
“Respirometric Investigation of Activated Sludge Bioinhibition by
Cobalt/Manganese Catalyst” Presented, 46th Annual Industrial
Waste Conference, Purdue University, West Lafayette, IN, May
1991.
Edgehill, R.V., and Finn, R.K. (1983). “Isolation, Characterization, and
Growth Kinetics of Bacteria Metabolizing Pentachlorophenol,” Eur.
J. Appl. Microbiol. Biotechnol., 16, pp. 179-189.
Gaudy, A.F. Jr., and Gaudy, E.T. (1980). Microbiology for Environmen-
tal Scientists and Engineers, McGraw-Hill, New York.
Gaudy, A.F. Jr., and Gaudy, E.T. (1988). Elements of Bioenvironmental
Engineering. Engineering Press, Inc., San Jose, California.
Gaudy, A.F., Jr., Komolrit, K., and Gaudy, E.T. (1964). “Sequential
Substrate Removal in Response to Qualitative Shock Loading of Acti-
vated Sludge Systems.” Appl. Microbiol., 12, pp. 280-286.
Gaudy, A.F. Jr., Rozich, A.F., and Garniewski, S.T. (1987). “Biological
Treatment of Concentrated Landfill Leachate,” Proceedings, 4/st
Annual Purdue Industrial Waste Conference. Lewis Publishers, Inc.,
Chelsea, Michigan, pp. 627-638.
Jones, G.L. (1973). “Substrate Inhibition of Bacterium NCIB 8250 by
Phenol,” J. Gen. Microbiol., 74, pp. 139-149.
Komolrit, K., and Gaudy, A.F., Jr. (1966). “Biochemical Response of
Continuous-Flow Activated Sludge Processes to Qualitative Shock
Loadings.” J. Water Poll. Control Fed., 38, pp. 85-101.
Martin, J.K. (1988). “Inhibition of Respiration in Activated Sludge by
High Carbon Dioxide Concentration—A Laboratory Study,” Public
Technology, Inc., Washington, D.C.
Ramanathan, M., and Gaudy, A.F. Jr. (1969). “Effect of High Substrate
Concentration and Cell Feedback on Kinetic Behavior of Heteroge-
neous Populations in Completely Mixed Systems,” Biotech. Bioeng.,
11, pp. 207-237.
Rozich, A.F., and Colvin, R.J. (1986). “Effects of Glucose on Phenol
Biodegradation by Heterogeneous Populations,” Biotechnology and
Bioengineering, XX VII, pp. 965-971.
Rozich, A.F., and Colvin, R.J. (1988). Unpublished results.
 VALUES OF BIOKINETIC CONSTANTS 129

Rozich, A.F., and Colvin, R.J. (1990). “Formulating Strategies for Acti-
vated Sludge Systems,” Water Engineering & Management, 137, 10,
pp. 39-41.
Rozich, A.F., Clay (nee Garniewski), S.G., and Colvin, R.J. (1991).
Unpublished results.
Rozich and Gaudy, Inc., (1986). “Determination of the Numerical Values
of the Biokinetic Constants and Implications to the Design of the
Expanded Facilities for the Patapsco Wastewater Treatment Plant,”
Engineering Report to the City of Baltimore, MD, Wastewater Facili-
ties Division.
Rozich, A.F., and Gaudy, A.F., Jr. (1985). Response of Phenol Accli-
mated Sludge to Quantitative Shock Loading. J. Water Poll. Control
Fed., 57, pp. 795-804.
Rozich, A.F., Gaudy, A.F. Jr., and D’Adamo, P.C. (1983). Predictive
Model for Treatment of Phenolic Wastes by Activated Sludge, Water
Research, 17, pp. 1453-1466.
Rozich, A.F., Gaudy, A.F., Jr., and D’Adamo, P.C. (1985). “Selection
of Growth Rate Model for Activated Sludges Treating Phenol,” Water
Research, 19, pp. 481-490.
Rozich, A.F., and Lowe, W.L. (1984). “Oxidative Assimilation Treat-
ment of a Nitrogen-Deficient Toxic Waste,” Biotechnology and Bioen-
gineering, XXVI, pp. 613-619.
 ard ee
ane
; at a
=
’
i cold
ee = 7

Wexe Confereny, Puntne \

keeps seein saionee ‘anaMtESa
ry ni angbait
oatanu

-~ enh Radice ot
ato nde .

Paine
wey ’
ideinian ae
arta AB

eS
Geage, A. bez :
vedere A Cre a on
Anead! Pirtes jena: tegerre.
Lily, (oer qaqey pP- at po
Sinn Keke art Sanaa Jepdacene)
__ heat” 4.Gre. Atevebes *.18 fete . 8
Your; Kron Daas At ho CR
i swinvew (< 2Aeainbiod Swipe Pranmee
“wig.” iver Pali. Comat Fed ae, “
Marge 0 (1988. "inipaan OF arene
+ hgiy Cav late aatda Conceriniiig al’ a
Techheugy. Ine, Water 12)
s vuttvet., (46, at Cainbs. owl, te-abase-
care cramin and Coll Feotaae) yo ater
=gxvit) FY er ¢ in Somperg!) saa aaid
wm.
Riukh, 4.7 > a Coe &.,}, (Gi, “efit af
Himdemmmgseokl by) Dlecrypeneml we
Suis Sp inter ud, NRW, 7D. * rm ane...
oes, AF, ex Coltas tly i. SOD ea
 a CASE STUDIES AND
APPLICATIONS

INTRODUCTION

The purpose of this chapter is to illustrate with actual case studies the
application of the process analysis techniques that use respirometry for
calibration. It is important to remember that the utilization of respirome-
trically calibrated activated sludge models are usable for both opera-
tional and design purposes. The reader should recognize that this tech-
nology is applicable for both design and operation.
It is especially useful for operational applications because the collec-
tion of respirometric data often takes less than a day; the completion of
ancillary modeling analyses requires two hours or less. A plant upset or
production needs often pressure management to make decisions rapidly.
The performance of a full-blown treatability test using the “conven-
tional” approach generally takes a minimum of a month to complete.
The alternative approach using respirometry rapidly identifies a suitable
strategy for the operation of the biological treatment facility.
For design applications, respirometric techniques and the associated
process modeling do not replace the conventional treatability study; this
methodology augments the quantity and quality of data obtained during
these efforts. In some cases, experience has shown that these techniques
enable one to “fast-track” projects, curtailing both the associated time
and the effort required for the conventional approach. When a treatabil-
ity study is warranted, it entails at the very least the operation of one or
more bench-scale reactors. With this level of effort already invested in a
project for reactor maintenance, etc., the generation of batch respirome-
tric data is not a significant work addition to a project. During the course
of a treatability study, the periodic determination of the biokinetic con-

131
132 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

stants enables the designer to assess both the range and the variation of
the values of these parameters. One can also evaluate the time needed for
acclimation. With this information and the use of the model, it is feasible
to perform a relatively thorough value engineering analysis. A process
analyst performs a value engineering analysis by generating a series of
predictive curves for effluent quality. Different values of the biokinetic
constants generate different predictive curves.
This chapter will present four case studies that are representative
examples of applying the respirometric techniques and associated model-
ing methodologies. Three of the case studies presented in this chapter
relate the application and use of the model. Another case study involves
the application of respirometry for screening applications. Each case
study emphasizes a different type of application and the associated meth-
odology involved with performing the process analysis.
The first case study concerns the use of the respirometric methodol-
ogies to perform a relatively lengthy design and operational analysis.
This effort involved a municipal treatment facility that encountered dif-
ficulties caused by the inhibitory nature of the influent wastewater. The
techniques described herein were used to analyze the treatment situation,
devise a concept design for a facility’s expansion, and perform a verifica-
tion analysis of the model. Data were collected to compare model predic-
tions and actual values for effluent quality.
Another case study involves the use of these analytical methodologies
to determine startup criteria for an activated sludge facility located at a
Superfund Site. The facility had to treat wastewaters impounded in sev-
eral lagoons at the site. This case study is a good example of the ability of
this technology to enable a project manager to fast-track a process star-
tup, design, or operational modification. Acute time and budgetary con-
straints characterized this effort. The respirometric methodology and
associated modeling protocol provided accurate information for the
startup and operations. The actual operating data for the facility verified
the integrity of the respirometric approach.
The third case study involves a biological treatment facility located in a
chemical manufacturing plant. Discharges of the heavy metal cobalt to
the activated sludge process concerned operations staff. The methodol-
ogy for process analysis described in this book analyzed this treatment
situation and determined the impact of cobalt on process performance.
A model analysis for process performance used the data and quantified
the impact of cobalt on the activated sludge facility.
A fourth case study will describe the use of respirometry to screen, or
rank, the biodegradability of various waste products at a specialty chemi-
 CASE STUDIES AND APPLICATIONS 133

cal manufacturing facility. The wastewater treatment facility faced the

possibility of relatively tight COD limits for effluent quality. The objec-
tive of the work presented herein was to “rank” the various waste prod-
ucts produced at the manufacturing facility. These data provide guidance
to determine which wastes to remove from the influent waste stream to
the biological treatment system. Alternatively, the difficult wastes can
receive special consideration for treatment in order to prevent excess
pass-through of COD in the plant effluent.

CASE 1: PATAPSCO WASTEWATER TREATMENT PLANT,

BALTIMORE, MARYLAND

Background

The Patapsco Wastewater Treatment Plant in Baltimore, Maryland

represents a unique treatment situation in that it is a municipal treatment
plant receiving a relatively heavy loading of industrial effluents from
chemical manufacturers. Additionally, because of concerns over space
constraints, the activated sludge system was designed as an oxygen acti-
vated sludge facility to conserve land utilization; this means that the
design nominal hydraulic detention time (excluding the impact of the
recycle stream) is 2 hr. Because of the need to remove phosphorus, the
plant was modified to operate using an anaerobic selector as a means to
foster a biomass with the capability to remove phosphorus biologically.
This process modification means that an aerobic treatment system that
already has difficulty with process performance will be further burdened
with having one-fourth of the aeration tank (the size of the anaerobic
selector section) removed from aerobic treatment capability.
The consulting effort involved both an operations and a design analy-
sis; a design analysis was included because the plant was to undergo an
expansion in order to accommodate an increased hydraulic load. Plant
operations personnel desired to have the issue of influent waste toxicity
addressed as part of both the operations and design analyses. The overall
effort consisted of four main tasks:

¢ Model analysis and review of existing data.

¢ Determination of biokinetic constants.
¢ Field evaluation of model predictions.
¢ Concept design and operational recommendations.
134 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Each of these tasks will be reviewed in detail below.

Model Analysis and Review of Existing Data

This task essentially consisted of a “forensic engineering” effort; the

plant had experienced a number of upsets which consisted of gross leak-
age of COD and five-day BOD (BOD,); these were attributed to “toxic
events” during which large quantities of toxics were presumably dis-
charged to the plant. These occurrences were identified via interviews
with plant staff and review of appropriate operating records. The data
review showed that the plant did experience large, temporary increases in
influent COD (approximately 900 mg/L) associated with these events.
It was decided to utilize the model presented in Chapter 3 to analyze
the biological system in an attempt to identify any trends with regard to
key operational parameters, the influent loading condition observed dur-
ing the toxic events, and the occurrences of effluent COD leakage. At
this point in the effort, there was no biokinetic data available, so it was
decided to analyze the activated sludge process using the model and to
use literature values for the biokinetic constants. Since the key parame-
ters that determine effluent quality are the biokinetic growth constants,
and since plant personnel were fairly convinced that the upsets were due
to the input of toxics, the assumption was made that the waste was
inhibitory, and the model analysis was performed using a wide range of
values for pmax, K,, and K,; the selected ranges for these constants were
0.05 to 0.35 hr', 50 to 100 mg/L, and 100 to 300 mg/L, respectively.
Since the yield and decay constants, Y, and k,, have a relatively small
impact on prediction of effluent quality, one set of values, 0.50 mg/mg
and 0.01 hr"! respectively, were utilized throughout the analysis.
Another feature that was thought to have an influence over the treat-
ment situation at Patapsco was the fact that there was little or no control
over the return sludge flow rate. That is, the return sludge flow rate, Fp,
was essentially constant and a, the return flow ratio, was thus allowed to
vary. It was determined that this aspect warranted investigation. The
return flow issue was investigated by modifying the model presented in
Chapter 3, which assumed a constant recycle flow rate in lieu of a con-
stant recycle flow ratio; the “constant Fp” model is produced by substi-
tuting F,/F for a. Comparative process performance predictions using
both versions of the model were made for the Patapsco biological treat-
ment system.
Other values for the model analysis were obtained via review of plant
 CASE STUDIES AND APPLICATIONS 135

operating records or through interviews with plant staff. Since the origi-
nal plant design specified a 2.0 hr nominal detention time in the aeration
tanks, the model analysis evaluated process performance in the range of
0 to 5.0 hr. The impact of influent waste concentration, S,, was examined
over a range of 250 to 1000 mg/L COD. At this point in time, primary
influent COD values were around 500 mg/L while shock loads of up to
1000 mg/L were known to occur periodically. The recycle sludge concen-
tration, Xz, was reported to attain concentrations as high as 20,000
mg/L, but a more conservative value of 15,000 mg/L was used for the
model analysis. Table 7.1 summarizes the values of all the engineering
and biological parameters that were used in the model analysis.
Predictive curves for effluent quality are presented in Figures 7.1, 7.2,
7.3, and 7.4. When the constant a model was used, a value of 0.25 was
used for a while a value of 5.0 million gallons per day (MGD) was used
for Fz when the constant recycle flow version of the model was used.
Both values were selected after interviews with plant staff. Since the
reactors have a volume of 1.9 million gallons (MG), an additional
abscissa scale is used that presents predictive results as a function of flow
rate per reactor (MGD per reactor). Results were presented in this man-
ner so that the required number of reactors that are needed to treat a
given influent flow rate can be easily calculated.
The predictive curves that were produced as a result of the model
analysis make two important points regarding this treatment situation.

Table 7.1 Values of Engineering and Biokinetic Constants Used for

Model Analysis
Biokinetic Constant Value or Range
Geel 0.05 to 0.35
Kote COD 50 to 100
K,, mg/L COD 100 to 300
Y,, mg/mg 0.5
Kaen 0.01
Engineering Constants
S, mg/L COD 250 to 1,000
t, hours 0 to 5.0
F,, MGD 5.0
DOE FA 15,000
V, MG L.9
136 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

500 "| —— F, =5.0 MGD

Oe 1025

S, = 1000 mg/L

S, = 500 mg/L

S, = 250 mg/L
100

—) —_ — (<=

0 9.1 18.2 27.4 36.5 45.6

F, MGD / Reactor

Figure 7.1. Predictive curves for effluent quality showing effect of

flow rate and influent waste concentration, S,, on effluent
quality, S, as measured by soluble metabolizable COD, for
a 1.9 MG reactor. Values of constants used for the
computations were: Xz = 15,000 mg/L; S, = 30 mg/L;
Bmax = 9.20 hr'; K, = 100 mg/L; K, = 300 mg/L; Y, =
0.50 mg/mg; k, = 0.01 hr“

First, it is interesting to scrutinize these curves in light of the toxic events,

i.e., incidences of increased influent and/or effluent toxicity, that were
reported by plant staff. In one case, the plant was subjected to a high-
strength (approximately 1,000 mg/L COD), high-toxicity shock. A scan
 CASE STUDIES AND APPLICATIONS 137

S,
COD
mg/L

F, MGD / Reactor

Figure 7.2. Predictive curves for effluent quality showing effect of

flow rate and influent waste concentration, S,, on effluent
quality, S, as measured by soluble metabolizable COD, for
a 1.9 MG reactor. Values of constants used for the
computations were: Xp = 15,000 mg/L; S, = 30 mg/L;
Lmax = 0.10 hr'; K, = 100 mg/L; K; = 300 mg/L; Y, =
0.50 mg/mg; k, = 0.01 hr'. Note difference in y,,,, values
compared to Figure 7.1.

of the predictive curves in Figures 7.2 and 7.4 shows that a flow rate of
about 16 MGD per reactor at an influent COD concentration of 1,000
mg/L results in a fair amount of effluent deterioration. It should be
138 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

F, =5.0 MGD
500 —- — — a =0.25

S$;= 1000 mg/L

400
S$; = 500 mg/L
Q
Oo
© 300
=
o
E
yu) 200

100

F, MGD / Reactor

Figure 7.3. Predictive curves for effluent quality showing effect of

flow rate and influent waste concentration, S,, on effluent
quality, S, as measured by soluble metabolizable COD, for
a 1.9 MG reactor. Values of constants used for the
computations were: Xz = 15,000 mg/L; S, = 30 mg/L;
max = 0.05 hr'; K, = 100 mg/L; K, = 300 mg/L; Y, =
0.50 mg/mg; k, = 0.01 hr-'. Compare with Figures 7.1 and
Vike
 CASE STUDIES AND APPLICATIONS 139

FR = 5.0 MGD
500 = SS BES

$j = 1000 mg/L
400

°
O 300 S$; = 500 mg/L

=
oD)
£
w 200

100

Sj = 250 mg/L

0 0.2 0.4 0.6 0.8 1.0

D,h™

F, MGD / Reactor

Figure 7.4. Predictive curves for effluent quality showing effect of

flow rate and influent waste concentration, S;, on effluent
quality, S, as measured by soluble metabolizable COD, for
a 1.9 MG reactor. Values of constants used for the
computations were: Xp = 15,000 mg/L; S, = 30 mg/L;
Pmax = 0.10 hr-!; K, = 50 mg/L; K, = 100 mg/L; Y, = 0.50
mg/mg; k, = 0.01 hr'. Compare with Figures 7.1, 7.2,
and 7.3.

noted that although both model analyses in these figures employ a py, Of
0.10 hr~!, the deterioration predicted in Figure 7.4 is worse because these
model predictions were made with smaller K, and K; values. Data showed
140 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

that the plant did in fact experience substantial leakage of toxic material
during this incident. Thus, it is reasonable to state that the model yielded
a reasonably good qualitative prediction (with regard to COD/toxicity
leakage) during this time period.
The other important point that the model analysis makes regarding
this treatment situation concerns the impact of the recycle parameters on
reactor performance. The predictive curves show that the constant F,
operating mode is more sensitive to process upset than the constant
recycle flow ratio mode (constant a). This point is readily understood if
one examines the equation for net reactor growth rate for the activated
sludge reactor.

pw, = (1 + @ + aX,/X)/t Ci)

As previously discussed, the constant Fp, model has a@ defined as F,/F.

This means that as detention time, t, decreases and flow rate, F,
increases, the recycle flow ratio, a, is decreasing since Fx, is held con-
stant. Reactor growth rate computations using equation 7.1 show that
the impact of decreasing a values is to increase yu, in the reactor; conse-
quently, increases in flow rate increase reactor growth rate by decreasing
detention time, t, and by decreasing a in the constant F, mode. In
contrast, flow rate changes in the constant a mode only impact growth
rate by virtue of affecting t. This is reflected in the prediction of less
COD leakage for the constant a mode. The plant eventually imple-
mented a recycle flow system that was “paced” to the influent flow rate
(ironically, this operating change was implemented for other reasons);
this in effect maintained a constant a or recycle flow ratio. The key point
is that the maintenance of a constant a provides protection against efflu-
ent deterioration because it enables the operator to impart more control
over reactor growth rate. As pointed out in the predictive curves, this is
especially the case when a plant is under stress or shock load conditions
such as those experienced by Patapsco.
Equation 7.1 also makes an important point regarding the importance
of the recycle sludge concentration X, in maintaining reactor growth
rates in the reactors. Since this system has such a low t (2 hr for
Patapsco, while “conventional” municipal systems have an 8-hr deten-
tion time), the bulk of the control over growth rate has to come from the
recycle parameters, recycle flow ratio, or recycle flow rate. It was previ-
ously noted that a value of 15,000 mg/L was used for Xx in the modeling
analyses. Other model analyses, such as that depicted in Figure 7.5,
showed that drops in Xx to 10,000 mg/L could adversely impact process
 CASE STUDIES AND APPLICATIONS 141

50 ———_ XR = 15,000 mg/L

— — — XR=10,000 mg/L
si=600 //
/ /

QUALITY,
EFFLUENT
ACOD
mg/L

0 0.2 0.4 0.6 0.8 1.0

DILUTION RATE, h”’

e@) 5.0 2.5 1.7 1.3 1.0

DETENTION TIME, h

0 9.1 18.2 27.3 36.5 45.6

FLOWRATE, MGD / REACTOR

Figure 7.5. Dilute-out curve showing the effects of X, and S; on flow

capacity of a 1.9 MG activated sludge system. S, represents
different permit limits for effluent quality. Other
parameters used in this analysis were y,,,, = 0.108 hr"', K,
= 34 mg/L ACOD, Y, = 0.4, k, = 0.005 hr’, and a =
0.25.

performance; a review of operating data showed that fluctuations of Xz

to this level occurred routinely. Poor performance is attributed to the
fact that X, heavily impacts reactor growth rate; this performance prob-
142 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

lem caused by fluctuating Xz is especially exacerbated when the reactor

detention time is on the low end, as is the case with Patapsco. Increases
in S,, the influent waste concentration, serve to stress this situation fur-
ther, as depicted by the various predictive curves. Control of reactor
growth rate is heavily influenced by detention time and the recycle
parameters. In the case of Patapsco, the low detention times force a
heavy reliance on the recycle parameters to maintain the necessary
growth rates that will enable the system to prevent leakage of COD.

Determination of Biokinetic Constants

The biokinetic constants were determined using two different meth-

ods: the batch growth study, or shake flask method and the respirometric
method. The initial effort for this work focused on utilizing the shake
flask method because the respirometric technique had not yet been fully
developed. The shake flask method consists of inoculating waste samples
at different COD concentrations with the target biomass; the change in
biomass concentration over time is then determined using optical density
measurements. Once these data are collected, they are analyzed to obtain
a set of » and S points which are in turn analyzed to determine the values
of the biokinetic constants, y,,,,, K,, and K; (if the waste is inhibitory).
The application of the shake flask technique is usually straightforward
when employing synthetic wastes such as those utilized in research
efforts. It is quite another matter when the method is applied to a “real”
waste stream such as that found at the Patapsco plant.
During the course of work, we noted several interferences that greatly
hindered the application of the shake flask technique. One interference
was the fact that often the soluble COD of the waste that was available as
a substrate source for the biomass was very low; this means that only
marginal biomass production can be expected during a test, which
greatly hinders the accurate calculation of a growth rate because there
will be such a small change in biomass concentration. The only practical
countermeasure is to concentrate the waste via an evaporative technique;
this was done by evaporating the waste under reduced pressure at rela-
tively low temperatures (55 to 60°C) and adding distilled water to consti-
tute a more concentrated waste; if solids precipitation occurred, the
modified waste mixture was filtered. Not only is this procedure labor-
intensive, it may compromise the test results if the nature of the substrate
is modified as a result of the concentrating method.
Other interferences included the existence of relatively high levels of
 CASE STUDIES AND APPLICATIONS 143

suspended solids and coloring of the waste, which interfered with

absorbance measurements. It was observed that the waste was occasion-
ally colored and the level of coloring was pH-dependent and would
change during the course of a test. These interferences made interpreta-
tion of the shake flask data time-consuming because analysis of the
growth data would often produce equivocal results. Conversely, when we
started applying the respirometric method described in Chapter 5, i.e.,
collecting respirometric data, “translating” it to simulated biomass
growth curves, and then calculating growth rates from these curves, we
found that the greater majority of our kinetic analyses were extremely
straightforward. In point of fact, and employing a popular contempo-
rary expression, we found the analyses of the curves that resulted from
the respirometric analyses to be “no brainers.”
The results of the initial biokinetic determination efforts for the
Patapsco plant are given in Tables 7.2 and 7.3. Table 7.2 presents the
values of the biokinetic constants determined using the shake flask
method, while Table 7.3 gives the values for the constants that resulted

Table 7.2 Biokinetic Constants from Batch Growth Studies with

Patapsco Waste
INHIBITORY RESPONSE

Bmax K, K,
Date hr! mg/L mg/L SSQ*
3/6/86 0.149 6 193 0.43(10°3)
3/13/86 Orza2 47 89 0.20(10°7)
3/18/86 0.296 23 80 0.30(10°)
4/16/86 0.279 51 1497 0.47(10°’)
4/22/86 0.182 63 191 0.35(10-7)
5/20/86 0.364 71 706 0.38(10°)

NONINHIBITORY RESPONSE
Pmax K,

Date hr! mg/L SSQ

3/25/86 0.207 83 0.69(10°7)
3/27/86 O2217 47 0.79(10°)
5/1/86 0.248 34 0.31(10°)
5/27/86 0.180 176 0.10(10)
*Sum of squared residuals resulting from data fit.
 144
IIqQUL¢°7 A1wUIWINS
Jo YIMOID SIHIUTYSUISE)
BY) INA[OIIIIY JIIWOMASIY
dNauryjoIg sjuBjsuoD
ad¥}UIIIIg
JO ISVAA YISUINS xeu Sy 'y Oss
aed COLsiCiODsiSSs—iCiSCSZC“‘<i‘éiTCS:C‘
«608 «6S8) 0OL 4 8m “STS 7/do0
3u 1/400
L8/LI/€ 4 (-4) €90°0 r80°0 8010 £110 1110
ogdOD3) (1/d09 OL 66 IZI 8rl CEI O) LEC ZLI = L070 €-d
°s 3) (1/d09 61 SE 9¢ CL 6 Lvl'0 Sz = 096°0b-J

L8/¥7/€ 7 (4) ¢S0'0 650°0 8r70°0 790°0 €50'0

ogOD31) (1/09 89 ZOI v0 OLI L6I 0900 6 = 9€1'0€-d
°s 3u) (1/d09 Sz os CL 00! SZI LS0°0 I = Lrl‘0€-d

L8/LO/¥ 7 (,-u) Z10°0 610°0 910°0 1700 1Z0'0

ogO0D31) (1/d0D Lv OL 7 IZ €SI 1€0°0 89 = O11'0b-d
°s 3u) (1/dOO 91 (a3 8P >9 08 $z70'0 9I = Orl'0b-d

L8/¢1/¥ 7 (,-4) 9£0°0 €90°0 €90'0 $L0°0 ZL0°0

ogO0O31) (1/d09 iss 66 IZI 19] I8I 1710 801 = 098°0b-d
°s 3) (1/dO0 IZ tp v9 $8 901 960°0 67 = 0060 b-A

L8/LT/¥ 4 (-) 190°0 090°0 8L0'0 780°0 060'0

ogOD3) (1/dO9 6£ LL (4) Sol OZI €11'0 tp = LS7'0€-d
°s 3) (1/dO9 L v1 77 67 9€ 660°0 9 a OL10€-d
ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY
 L8/7Z1/S 4 (,-y) £50°0 090°0 O CII
ogago8w) (1/409 19 660°0
901 8S]
°s 3u) (1/dO9D 87 067 €LZ7°0
gs €8 069°0£-d
EI
8 cL70
L8/LO/L 4 (,-y) 0020£-d
ogga8u) (1/409 9010
°s

veo'0
8u1)

180°0
(1/dOD 8€I vol'O

060°0
66
0960 b-A

6c 91
08

89 ce
ILI'0

8p
+L8/L0/L 7 (,-u) SCO) Cael
L100 6£0°0
ogao3w) (1/409 €€ 6r0°0
L9
°s 3u) (1/409 €bI £60°0

790°0
LI be breo €-d
98 180°0 9€

cOl cs
L8/6
+ 0/6 7 (,-4u) Ivo €-d
61€'0 I8E°0 LrEO SL70
ogago3w) (1/409
LI S070 IZE0
Sp £9
SOT

8C
°s 3w) (1/409S Silas OP IYO IT
el 61 Orl0I-d

8
I€ bE $09°0 €
Ajaiewxo
%TZSjo
iddy
ay), CC Ori‘I-d
Areuttidsprjos a1am Parouias
+ Syy paas U2¥e] wo
WO1J24) aisem a[duies a10Jady)
q Isa} SBM
O/y
CASE STUDIES AND APP

‘ssa00id “PeiBIS
145 LICATIONS
146 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

from the respirometric technique described in Chapter S. It is interesting

to note that the values of the constants in both tables are in approxi-
mately in the same range. However, except for a few instances, the
constants were characteristic of Monod kinetics and not Haldane. That
is, except for a few tests that showed inhibitory characteristics, the data
indicated that the inhibitory effect thought to be associated with the
influent toxicity problem was not present. This presented somewhat of a
problem because all other indicators suggested that Patapsco warranted
different consideration than one would give to a “typical” municipal
treatment facility. The question, then, is how to demonstrate and quan-
tify this aspect of the treatment situation.
The question of the severity of the Patapsco waste and why it should
not be considered typical was addressed by comparing the biokinetic
results of the Patapsco plant with those obtained for a municipality that
did not receive a substantial industrial flow. Peil and Gaudy (1971)
evaluated the biokinetic constants for the Stillwater, Oklahoma Waste-
water Treatment Plant; this facility received little industrial input, and it
was reasonable to state that facility was representative of a typical
domestic waste treatment system. The biokinetic constants for this facil-
ity were measured as 0.50 hr' and 63 mg/L for p,,,, and K,, respectively
(Monod kinetics); Figure 7.6 compares the “worst case” constants (lowest
max» Highest K,) with the Stillwater constants. Although this figure
clearly shows that the Patapsco system, while not inhibited, still never-
theless represents a more difficult treatment scenario than the “typical”
situation, it is not clear how this will impact the design of the expanded
facilities. This aspect is addressed by comparing predictive curves for
effluent quality for both the Stillwater constants and the Patapsco
constants.
Figures 7.7 and 7.8 depict predictive curves, which were generated
using the model equations given in Chapter 3, the appropriate values of
the biokinetic constants, and the values of the engineering constants that
were used in the model analyses in Figures 7.1 through 7.4. Figure 7.7
compares the predictive curves generated using an average kinetic
response for the Patapsco plant with the predictive curves produced
using the Stillwater constants. It is interesting to note that three of the
curves predict satisfactory effluent quality over a wide range of detention
times; the curve that predicts some potential for effluent deterioration is
the one generated using the average Patapsco constants and an influent
waste strength, S,, of 1000 mg/L.
It needs to be emphasized that the only difference between the
Patapsco and typical curves is with regard to the values of the biokinetic
 CASE STUDIES AND APPLICATIONS 147

0.50

Stillwater Plant Kinetics

0.40

0.30

s
r=
0.20

Patapsco Plant Kinetics __

0.10 ae

0 100 200 300 400 500

S, mg/L COD

Figure 7.6. Comparison of kinetics for a municipal plant receiving little

industrial waste (y,,,, = 0.50 h', K, = 63 mg/L) and one
receiving a heavy industrial waste contribution (y,,,, = 0.18
hr', K, = mg/L).

constants; that is, any difference in predictions for process performance

is solely attributable to the kinetic differences of the two biomass sys-
tems. It is also interesting to note that for a wide range of flow rates, the
typical system shows little tendency for effluent deterioration, even at an
S; of 1000 mg/L and a detention time of 1.0 hr (at these flow rates, the
system would likely fail due to overloading of the secondary clarifier, but
not experience a biochemical failure). This result qualitatively agrees
with the claims of manufacturers of pure oxygen activated sludge treat-
148 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

a S, = 1000 mg/L

— — — S, =500 mg/L
200

a © 150
D
.
wo
100 “Typical” constants

F, MGD / Reactor

Figure 7.7. Predicted profile for effluent quality, S, as metabolizable

COD for a 1.9 MG reactor. Values of the constants used for
computations were: y,,,, = 0.253 h', K, = 90 mg/L, Y, =
0.34, ky = 0.13 d', Xp = 15000 mg/L, a = 0.25.
“Typical” profile predictions for normal domestic waste
generated using p,,,, = 0.50 h' and K, = 63 mg/L (Peil
and Gaudy, 1971).

ment systems. That is, it has been stated that these systems can be oper-
ated on municipal wastes in high-flow rate, low-detention time situations
and deliver acceptable effluent quality (provided that sufficient second-
ary clarifier capacity is available). If one accepts that the Stillwater con-
 CASE STUDIES AND APPLICATIONS 149

250
S, = 1000 mg/L

— — — §; =500 mg/L
200

S,
mg/L "Typical" constants

0 9.1 18.2 27.3 36.5 45.6

F, MGD / Reactor

Figure 7.8. Predicted profile for effluent quality, S, a 1.9 MG reactor

for the critical noninhibitory response (low y,,,, , high K,).
Values of the constants used for computations were: y,,,, =
0.18 h', K, = 176 mg/L, Y, = 0.40, k, = 0.12 d', X, =
15,000 mg/L, a = 0.25. Typical profile predictions for
“normal” domestic waste generated using p,,,, = 0.50 h'
and K, = 63 mg/L (Peil and Gaudy, 1971).

stants typify the kinetics of municipal activated sludge systems, then the
model analysis results provide a reactor engineering argument that sup-
ports the high-rate performance contention of proponents of oxygen-
activated sludge systems.
150 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

When the values of the biokinetic constants are much lower than those
of a typical system, Figure 7.8 shows that the situation is more critical.
This figure shows that the Patapasco system will have a tendency for
substrate leakage at S; values of both 500 and 1000 mg/L COD, in
contrast with the typical system (Stillwater), which shows little tendency
for COD leakage over a wide flow rate range. This figure also demon-
strates that the biomass kinetics do not have to be inhibitory (Haldane
equation) to put a system at risk to experience effluent deterioration
caused by COD leakage. In the case of Patapsco, the difficulty with the
waste is that the kinetics are low in comparison to a typical municipal
waste. Most design bases for municipal systems are based on empirical
information. Since the kinetic characteristics of the plant are atypical for
a municipal system, then the design basis is also atypical as defined by
the values of the biokinetic constants and the associated modeling analy-
ses. The overall situation was further exacerbated by the short detention
time in the reactors and the periodic high-strength shock loads which
were referred to as “toxic events.”

Field Evaluation of Model Predictions

In this phase of the work, an effort was initiated to validate the pre-
dictability of the model given in Chapter 3 and the utility of calibrating it
(i.e., determining the values of the biokinetic constants) using the
respirometric methods described in Chapter 5. A sampling program was
undertaken to obtain primary effluent (influent to the activated sludge
system) and secondary effluent samples along with samples of activated
sludge biomass which were taken from the recycle sludge line. The influ-
ent and effluent samples were analyzed for soluble BOD, in order to
obtain a measurement of S, and S, respectively, as ACOD (BOD, can be
converted into ACOD); the ultimate goal of the exercise was to determine
how well the model predicted S (i.e., effluent quality) as measured as
soluble ACOD or BOD,.
Using respirometry, nine sets of biokinetic constants were determined
from April through September, 1989. On each day that biomass were
harvested for testing, influent and effluent samples and appropriate
analyses were also taken and performed as previously described. Addi-
tionally, the values of the engineering parameters (recycle sludge concen-
tration, Xz, recycle flow ratio, a, and primary effluent flow rate, F) were
also recorded on the particular sampling day.
The values of the biokinetic constants obtained as part of this effort
 CASE STUDIES AND APPLICATIONS 131

are given in Table 7.4; this table shows that only three of the sets of
constants were inhibitory. Table 7.5 provides the list of engineering
parameters, influent and effluent data, and model predictions. The
model predictions were made using the values of the biokinetic constants
in Table 7.4, the values of the engineering parameters given in Table 7.5,

Table 7.4 Values of the Biokinetic Constants for the Patapsco Plant
Biomass

P-max K, K, Y,
Date h! mg/L ACOD mg/L ACOD mg/mg
4/25/89 0.109 15 2 0.41
4/27/89 0.090 15 - 0.41
7/24/89 0.064 28 108 0.29
7/31/89 0.027 ify — 0.29
8/9/89 0.053 28 108 0.31
8/17/89 0.036 16 - 0.27,
8/29/89 0.155 1.4 204 0.55
9/5/90 0.090 1 - 0.46
9/11/89 0.037 15 - 0.27
Note: ky = 0.005 d"' was assumed for each study.

Table 7.5 Values of Engineering Parameters and Actual and Predicted

Values for Effluent Quality for the Patapsco Plant
Soluble
Effluent
ACOD
Xr S; F mg/L
Date mg/L VSS a mg/L ACOD MGD Actual Pred.
4/25/89 14,035 0.37 161 45.7 5 1
4/27/89 9,485 0.40 167 45.3 11 me
7/24/89 14,600 0.38 143 55.4 6 3
7/31/89 13,000 0.34 159 13.9 5 9
8/9/89 14,130 0.44 115 44.0 8 2
8/16/89 10,730 0.44 V2) 44.6 3 2
8/29/89 9,310 0.37 169 41.5 5 Gal
9/5/89 8,900 0.35 121 40.0 3 0.1
9/11/89 10,360 0.37 192 44.4 6 4
152 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

and the model equations provided in Chapter 3. Table 7.5 shows that
predicted and actual values for effluent quality were relatively reason-
able, especially considering the complex nature of trying to predict efflu-
ent quality in a large (70 MGD) operating facility. The difficulty encoun-
tered in trying to predict low values of a given parameter should also be
emphasized. This also supports the conclusion that the model calibrated
with biokinetic constants determined via analyses of respirometric data
provided reasonable predictions of plant effluent quality.
Given the favorable predictions provided by the model and the sup-
porting positive operational data, one may question the utility of per-
forming the modeling exercise. The answer involves the need to predict
the operating envelope (those operating conditions that could put the
plant at risk of effluent deterioration and subsequential permit viola-
tion). Figures 7.9 and 7.10 illustrate the utility of the modeling approach
in predicting plant capacity for various waste strengths (S,). The predic-
tive curves in Figure 7.9 were constructed utilizing the kinetic constants
and engineering parameters that were relevant for 4/27/89, while those
curves for Figure 7.10 were generated using the appropriate information
from 8/17/89. The figures show that, at the actual S; values measured for
each day (167 mg/L and 121 mg/L for 4/27/89 and 8/17/89, respec-
tively), the plant has a wide operating range before experiencing prob-
lems; as previously discussed, the actual field data showed that the plant
had little trouble in delivering excellent effluent quality at these S, values.
(In this case, plant capacity is more likely to be limited by the capacity of
the secondary clarifiers as influent flow rate increases and detention time

EFFLUENT
STANOARD

COD/)
CONC.
WASTE
EFFLUENT
(mg

REACTOR FLOWRATE (MGD)

Figure 7.9. Predictive curves for effluent quality for the Patapsco plant
(4/27/89) from Colvin et al. (1991).
 CASE STUDIES AND APPLICATIONS 153

EFFLUENT
STANDARD

COD/L)
CONC.
WASTE
EFFLUENT
(mg

0 20 40 60 80 100
REACTOR FLOWRATE (MGD)

Figure 7.10. Predictive curves for effluent quality for the Patapsco
plant (8/17/89) from Colvin et al. (1991).

decreases.) Both figures indicate that as S; increases to 600 mg/L COD,

the reactors are susceptible to effluent deterioration via COD leakage. If
one utilizes 15 mg/L as effluent criteria (see Colvin et al. (1991) for a
detailed explanation), the plant capacity at an S, of 600 mg/L is 55 MGD
for 4/27/89 and 38 MGD for 8/17/89.
It should be noted that the primary difference in conditions between
the predictions in Figures 7.9 and 7.10 is the difference in values between
the values of the biokinetic constant, y,,,,. The engineering parameters
used for the predictions are relatively close in value. The point is that the
model predictions are sensitive to the values of the biokinetic constants
determined using analyses of respirometric data via the methodology
described in Chapter 5. The biokinetic constants, in turn, are sensitive to
various environmental and plant conditions (e.g., waste quality, temper-
ature, pH, etc.) that affect the ecology of the activated sludge biomass;
these aspects were discussed more fully in Chapter 6. Thus, as conditions
change that affect the biomass, and consequently plant capacity, the
modeling procedure (when updated with respirometric calibration to
determine new values of the biokinetic constants) has the ability to quan-
tify the impact of environmental changes on plant capacity.
154 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Operational and Design Recommendations

The key question to be answered at this point concerns the recommen-

dations made to the City of Baltimore that resulted from the analytical
and modeling efforts. Two primary recommendations were made:

e Additional aeration volume to meet the expanded treatment

needs.
e Additional level of control and operational flexibility for the
biological system in view of the difficult nature of the waste.

Based on the design analysis as exemplified by the predictive curves for

effluent quality given in Figures 7.7 and 7.8, it was decided that the 1.9
MG reactor could safely handle a flow rate of approximately 18.2 MGD.
It was also noted that each reactor will only utilize 75% of its volume for
aerobic treatment since the first quarter of the reactor was to be held
anaerobic in order to provide a selector mechanism to promote biological
phosphorus removal. Using these numbers and a design flow of 87.5
MGD, one computes a need for 6.4 reactors (87.5/(18.2°0.75)); this cal-
culation suggests the need for three additional reactors because the plant
already has four. This suggestion created a problem with some of the
design engineering staff because the original recommendation called for
only one additional reactor. However, this recommendation was based
on the assumption that the Patapsco waste was a conventional municipal
waste. Clearly, the kinetic results (e.g., refer again to Figure 7.6, which
compares the Patapsco kinetic results to those of a more “conventional”
municipal waste) and associated modeling analyses showed that this was
not the case. However, an increase from one to three reactors was still a
bit hard to swallow for the design engineering staff. We then suggested
that, if the plant were given some level of operational flexibility and
control for the recycle parameters (recycle sludge concentration X, and
recycle flow ratio a) then two additional 1.9 MG reactors would be
acceptable because control of X, would enable the plant to “buy” more
capacity by virtue of the fact that this control technique would enable the
plant to increase its operating envelope. (It should be noted that as part
of the retrofit effort to implement biological phosphorus removal, the
plant had control over recycle flow rate which enabled it to maintain a
constant a.) Thus, if conditions changed, X, could be increased to
increase capacity. It should be stressed that the design already called for a
short detention time; this situation was further exacerbated by the con-
version of 25% of the activated sludge reactor volume to an anaerobic
 CASE STUDIES AND APPLICATIONS 155

zone. Given the difficult nature of the waste and the other consider-
ations, the recommendation of at least some form of Xx control or the
provision of some means to bolster recycle sludge concentration should
be viewed as essentially a basic need for this system.
Figure 7.11 shows an example of the application of recycle sludge
concentration control for an activated sludge plant. (The example shown
is not for Patapsco because, at the time of the preparation of this book,
other considerations precluded the incorporation of all of the receommen-
dations into the final design.) The concept design shown in Figure 7.11 is
for an industrial waste treatment system that must handle high-strength
organic wastes, which are characterized by inhibition. One key feature
for the design is the use of thickeners to provide additional thickening
capability for the return activated sludge, X,; it may also be possible to
realize this goal by simply utilizing polymer dosing to the mixed liquor to
increase underflow sludge concentration. Another feature is the use of
dosing tanks to store the thickened return sludge and to maintain X, at a
constant concentration. It should be noted that these tanks will also
function to store excess biomass in case the system takes an inadvertent
toxic shock load or “hit.” The excess biomass storage capacity also
enables plant personnel to bring up an out-of-service aeration tank rap-
idly in case there is an unforeseen shift in production schedules that
causes an increase in plant loading. If there is a change in the quality or
quantity of the waste that causes a change in the values of the biokinetic
constants, then plant personnel have the flexibility to reevaluate the

EFFLUENT

Figure 7.11. Concept design for activated sludge system using control
of Xp.
156 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

constants (respirometrically, of course), insert these values and the pro-

jected loading information into the model, and then use the model to
determine the recycle values needed to maintain effluent quality; the
plant design given in Figure 7.11 will then enable the operators to “dial-
in” the appropriate recycle values which are needed to keep the plant in
compliance.

CASE 2: STARTUP AND OPERATION OF

BIOLOGICAL TREATMENT FACILITY FOR
IMPOUNDMENT LEACHATES AND WASTEWATERS
AT A SUPERFUND SITE

Background

A potentially responsible party (PRP) was charged with having to

implement the treatment of impoundment leachates at a Superfund site.
The leachates were primarily contaminated with organics produced at the
manufacturing facility that used to be operated at the site. The manufac-
turing facility associated with the site produced rayon, polyester, and
polypropylene fibers; however, the primary product was rayon and fila-
ment fiber.
As a result of the production of these fibers, several waste products
were generated and disposed in surface impoundments. These wastes
included viscose liquid and solids, primary metal precipitates, waste acti-
vated sludge from the plant wastewater treatment system, and coal com-
bustion flyash from the facility boiler room. Most of the liquid and solid
wastes were conveyed to surface impoundments and landfill areas.
Eleven earthen impoundments contained viscose, which is a waste mate-
rial which is generated from the backwashing of viscose filters used to
remove undissolved cellulose particles and waste associated with the
fiber-making process. The other basins primarily held inorganic wastes,
waste activated sludge and flyash.
The large impoundments were used to collect stormwater and other
runoff waters. The regulatory agency that was monitoring site cleanup
activities noticed that the impoundments were essentially filled to capac-
ity. It was also noted that a heavy rainstorm could cause the impound-
ment which collected leachates, wastewaters, and stormwaters to spill
over into river adjacent to the site, which was used for recreational
activities and fishing. The situation prompted the oversite agency to
 CASE STUDIES AND APPLICATIONS 157

compel the PRP to begin treatment of the impoundment liquors in order

to keep the liquid level in these waste basins at a manageable level that
would not present a threat to the river.
Although analyses of the leachates indicated a fairly high BOD,, the
agency recommended the application of physical/chemical treatment
because of the perception that this represented a reliable technological
choice. The recommended treatment flow scheme consisted of chemical
precipitation, pressure sand filtration, and granular activated carbon.
Operating experience showed that despite best efforts, this treatment
scheme was unable to meet applicable or relevent and appropriate
requirement (ARAR) limits for BOD,. The oversite agency then required
the implementation of biological treatment in order to meet the BOD
removal requirements. The PRP decided to restart and use an existing
but idle wastewater treatment facility at the site. This facility consisted of
primary treatment followed by an activated sludge system. The initial
treatment strategy called for pH adjustment (raise the pH) of the
impoundment liquor prior to primary treatment to drop out solids that
might interfere with the biological treatment process. After primary
treatment, the pH was then lowered using acid and the liquor fed to the
activated sludge system.

Approach

An initial investigation showed that the activated sludge equipment

was in good condition. However, the PRP group was faced with a rigor-
ous compliance deadline and there was little time to perform any sort of
involved treatability study to verify the efficacy of biological treatment.
This also meant that the PRP would not be able to utilize conventional
treatability methods to determine the operating envelope or to optimize
operational parameters, which prompted the PRP to employ respirome-
try testing and associated modeling analyses to determine the plant per-
formance envelope for startup and operations.
The plan for implementing biological treatment of the impoundment
liquors involved the use of biomass from a local publicly owned treat-
ment works (POTW) to seed the aeration basins. A series of tests were
performed using the methodology described in Chapter 5 to determine
the values of the biokinetic constants. A modeling analysis was then
performed using the values of the constants, the projected flow rate for
the waste F, a range of values for the influent waste strength S;, and the
predictive model presented in Chapter 3.
158 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

A graph which provides the initial kinetic results for the impoundment
wastes is given in Figure 7.12. Biomass and waste from the site were
shipped via an overnite express shipping service to a laboratory for
respirometric analyses. The kinetic results and associated modeling
analyses were presented to the site wastewater treatment manager the
next afternoon (turn-around time of 30 hr). The kinetic results for the
startup condition were 0.095 hr1, 62 mg/L, and 351 mg/L for p,,,x, K,,
and K,, respectively. Prior to startup, the wastewater treatment manage-
ment staff wanted to operated the facility based on a target value for
mixed liquor of 4,000 mg/L. The modeling analyses compute the value
for mixed liquor suspended solids (MLSS) that one can expect. This
aspect was especially significant for this facility because the facility was
designed for a flow rate of 3500 gallons per minute (gpm) and the flow
rate for the impoundment leachate was projected at 400 gpm with a
strength of 400 mg/L COD; this translated to a nominal hydraulic deten-
tion time of 64 hr. (That is, the activated sludge system will operate in an
underloaded condition.)

GROWTH
(1/hours)
RATE

0 200 400
WASTE CONCENTRATION (mg COD/L)
4A STARTUP

Figure 7.12. Initial growth kinetics of biomass treating Superfund

leachate.
 CASE STUDIES AND APPLICATIONS 159

Table 7.6 shows the predicted results for the activated sludge system.
Values for the engineering parameters were selected based on best avail-
able information for the system; a was 1.0, Xz was 1,000 mg/L, S,; was
400 mg/L COD, and the tank volume, V, was 1.53 MG. The computa-
tions given in Table 7.6 predict that the facility will not have a problem
meeting effluent requirements (15 mg/L based on COD) over a large
operating range. This is due to the large detention time that is available
because the tank is underloaded; large detention times place less empha-
sis on the recycle parameters to keep low system growth rates or high 0,
values (refer to Equation 4.6). The predictions also show that the reactor
can expect to have a mixed liquor volatile suspended solids (MLVSS)
value of approximately 600 mg/L; this translates to a projected MLSS
value of 860 mg/L assuming a volatile fraction of 70%. This value
contrasts to the MLSS value of 4,000 mg/L considered to be reasonable
by the facility management staff. The difference can be reconciled by
noting the F:M and ©, values that were predicted by the model analysis.
(It should be noted that the model presented in Chapter 3 is easily modi-
fied to present results using parameters for activated sludge that are more
familiar to process analysts; a more detailed discussion concerning this
topic is given in Chapter 4.) At detention times greater than 83 hr, the
model predicts that the system is already beyond the point of extended
aeration; that is, X,, the predicted mass of waste sludge, is negative. This
means that the system cannot support the imposed recycle condition of
1,000 mg/L for Xz. Consequently, it is inappropriate to expect that the
system can support the higher recycle values (approximately 10,000

Table 7.6 Predictive Model Analyses for StartUp of Superfund

Leachate Treatment Facility
t S x MCRT F:M dS
(h) mg/L COD mg/L VSS d mg COD/mgX/d__iIb/d
125 2 540 oo 0.14 <0
100 2 550 oo 0.17 <0
83 3 562 63 0.20 bi3
63 4 573 20 0.27 365
45 6 582 10 0.36 748
26 11 592 4 0.62 1770
Note: Xz = 1,000 mg/L, V = 1.53 MG, S, = 400 mg/L COD, pas
= 0.095 hr!, Ks = 62 mg/L, K, = 351 mg/L, Y, = 0.55, and
k, = 0.002 hr.
160 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

mg/L) that one would expect at the design flow rate of 1600 gpm. At a
detention time of 63 hr, Table 7.6 shows that the reactor will operate at a
6. of 20 d and an F:M of 0.27 kg COD/kg MLVSS. Although conven-
tional wisdom suggests that these values are adequate to meet target
treatment goals, they had to be determined using the “unconventional”
modeling approach presented in Chapter 3. That is, the idea of operating
the system at an MLSS value of 4,000 mg/L is irrelevant for this treat-
ment situation.
The reactor started up with no problem and the field results indicated
that the predictive modeling results were very accurate. Subsequent
respirometric tests were performed to refine operations. One set of tests
focussed on optimization of feed pH. Other respirometric tests were
employed to evaluate the benefits of pretreating the liquor in the pri-
maries to remove suspended solids. A good example of the application of
respirometric techniques for process optimization is provided in Figure
7.13. The impact of pH on process kinetics (u versus S) is depicted in this
figure. Raising the pH prior to primary treatment and then bringing it

0.06

(1/hours)
RATE
GROWTH

~2b0 abo.
0 + ‘7 ie

WASTE CONCENTRATION (mg COD/L)

ia) pH=7.0 + pH=8.1 ° pH=9.0

Figure 7.13. Impact of pH on growth kinetics of biomass treating

Superfund leachate.
 CASE STUDIES AND APPLICATIONS 161

back to neutrality afterwards, presumably to enhance biotreatment oper-

ations, resulted in substantial chemical usage. The kinetic analyses in
Figure 7.13 show that adjusting the pH of the reactor feed close to
neutrality is actually detrimental to process kinetics and hinders biomass
performance. Using less acid and keeping pH values closer to those of
the primary effluent saves on chemical usage and enhances biomass
performance. This also serves to improve the process performance enve-
lope of the activated sludge system.

CASE 3: IMPACT OF COBALT ON BIOLOGICAL

WASTEWATER TREATMENT PLANT PERFORMANCE

Background

An organic chemicals manufacturing plant consisting of integrated

synthetic fibers and polymer intermediates production facilities dis-
charged wastewaters containing cobalt. The source of the cobalt was
likely attributable to a cobalt/manganese organic acid production cata-
lyst. The wastewater treatment facility consists of an activated sludge
system operated in two stages. The wastewater for the polymer interme-
diate production facilities is discharged after equalization into the first
stage activated sludge system. This system typically removes about 95%
of the organic acid COD. The effluent from the first system, along with
wastewater from the synthetic fiber manufacturing unit, are discharged
to the second-stage activated sludge system.
Production process modifications were made to comply with a new air
emissions permit. Soon after these productions changes were made,
operations staff noticed a reduction in the treatment efficiency of the
first-stage activated sludge unit. The first stage typically treats 0.5 MGD
of a mildly acidic (pH about 4-5) wastewater which has a high-strength
COD of 10,000 to 15,000 mg/L. Analysis of wastewater feed samples for
the first stage were performed to search for potential inhibitory agents.
The results indicated elevated levels of cobalt and manganese. A review
of the production units showed that the source of the cobalt was likely
the cobalt/manganese catalyst. The major source of catalyst discharge to
the first-stage activated sludge system was believed to be a solids slurry
that was temporarily diverted to wastewater treatment in order to main-
tain production while process modifications to production units were
made in order to comply with the previously mentioned air permit
162 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

requirements. The concentrations of soluble cobalt in the equalization

basins were measured at levels from 10 to 40 mg/L after the slurry was
introduced to the treatment system.
The general inhibitory effects of metals on activated sludge biomass
were well-known. However, specific data regarding the impact of cobalt
were not. A testing regimen for kinetic analysis using respirometry was
devised in order to quantify the impact of cobalt on plant operations.
The conceptual plan was relatively straightforward. The kinetics of acti-
vated sludge biomass degrading the first-stage wastewater were evaluated
at various cobalt concentrations. The testing methodology is described in
Chapter 5, while Chapter 6 discusses conceptual aspects of the impact of
toxicants (in this case, cobalt) on growth kinetics (specifically, refer to
Figure 6.14). The kinetic effects are quantified by determining the effects
of cobalt on p,,,., K,, and K, (if substrate inhibition is prevalent in the
base case, i.e., wastewater free of toxicants). The constants are then
inserted into the process model presented in Chapter 3 to quantify the
impact of the cobalt on the performance envelope of the first-stage acti-
vated sludge system.

Approach

The goal of the kinetic evaluation effort was to evaluate the detrimen-
tal impact of cobalt on the performance of the first-stage activated
sludge system. Chemical analysis of the first-stage biomass showed that
samples had elevated concentrations of cobalt and manganese. Since the
goal of the work was to evaluate the impact of cobalt on activated sludge
performance, it was decided that the kinetic testing effort required a
cobalt-free biomass. A biomass without cobalt contamination but that
was acclimated to the organic composition in the first-stage wastewater
was needed to develop the “baseline” case. This biomass was developed
by obtaining a sample of first-stage biomass that contained cobalt. The
biomass was then acclimated to a synthetic wastewater that simulated the
first-stage wastewater but did not contain cobalt. The “de-cobalted” bio-
mass was grown in bench-scale activated sludge units with internal recy-
cle (i.e., “Eckenfelder units”). Sludge in the reactor was analyzed on a
weekly basis for cobalt concentration. Kinetic testing of the biomass
began once the cobalt concentration in the biomass was less than 1 ppm.
A corollary effort involved the assessment of the effectiveness of a
pretreatment step to remove cobalt from the influent of the first-stage
system. Plant wastewater was pretreated to remove cobalt. Kinetic tests
 CASE STUDIES AND APPLICATIONS 163

were then performed on raw and pretreated wastewaters using the bio-
mass which was free of cobalt and cultivated in the bench-scale unit.
The numerical results for the biokinetic constants for the cobalt testing
are listed in Table 7.7. Comparative biokinetic growth curves are pre-
sented in Figures 7.14 and 7.15. Figure 7.14 illustrates the impact of
increasing concentrations of cobalt on the biodegradation kinetics of the
seed biomass treating the synthetic plant feed. Figure 7.15 shows a com-

Table 7.7 Kinetic Results for Cobalt Wastewater Analysis

Pemax K, K;
Wastewater h' mg/LCOD mg/L COD
Synthetic Feed 0.129 296 970
Synthetic Feed + 1 mg/L Cobalt 0.061 154 685
Synthetic Feed + 10 mg/L Cobalt 0.028 287 1249
Normal Plant Feed 0.068 251 693
Pretreated Plant Feed 0.099 248 1241

0.06 + = =
(d —

0.05 4 8
g 8

°Q =|
& G

= 0.04 b
<
: |
= ()
is 0.03
=
fe)
[4
oO
0.02

0.01

02 0.4 06 08 I 12 14 1'6 8 2
(Thousands)
SYNTHETIC WASTE CONC. (mg COD/L)
Oo NO COBALT + 1 mg/L COBALT © 10 mg/L COBALT

Figure 7.14. Impact of cobalt concentration on biomass kinetics.

164 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

0.06 +

0.05 |

e 0.04
J
fo}

&
A
ft 0.03
ro
e |
Z% 0.02

0.01

0 el a ee
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
(Thousands)
WASTE CONC. (mg COD/L)
ie) NORMAL PLANT FEED + PRETREAT PLANT FEED

Figure 7.15. Impact of pretreatment on kinetics of biomass treating

cobalt-containing wastewaters.

parison of the biomass kinetics on the normal and pretreated waste-

waters. It should be noted that the base case of no cobalt in Figure 7.14
and the pretreated case in Figure 7.15 compare quite favorably. The
biomass used for these two tests was the same, but the feeds were differ-
ent. The test in Figure 7.14 used the synthetic feed while the one in Figure
7.15 employed the pretreated plant wastewater. The fact that the kinetic
results are relatively close suggests that the synthetic feed did a good job
of mimicking the plant wastewater.
Figure 7.14 makes a number of points. It demonstrates and quantifies
the impact of increasing concentrations of cobalt on the biodegradative
capability of the biomass. Cobalt clearly has a significant impact on the
biomass to degrade the target waste. The biokinetic constants which are
listed in Table 7.7 were employed in a process analysis of the first-stage
activated sludge system. The process analysis was performed using the
predictive equations presented in Chapter 3. Predictive curves for efflu-
ent quality are presented in Figures 7.16 and 7.17. The engineering and
influent parameters used for these analyses are a recycle sludge concen-
 CASE STUDIES AND APPLICATIONS 165

130

120

a 110
va)
{e) 100
Oo
oD
e. 90

9 80
8 70
5
<< 60
>
&z 50

5 40
ta
ic
(ts) 30

26

10

0 T T qe gy = seat
20 60 1 140 180 20 5 260

REACTOR DETENTION TIME (hours)

+ 10000 ° 15000 4 20000

Figure 7.16. Predictive curves for effluent quality at different S; values

for biomass growing on synthetic wastewater (no cobalt).

tration, (Xx), of 6,000 mg/L, a recycle ratio (a) of 1.0, and influent COD
concentrations (S,s) of 10,000, 15,000, and 20,000 mg/L.
Figure 7.16 shows that the plant has substantial capacity for treating
the synthetic waste. At an S; of 10,000 mg/L COD, it can process down
to a detention time of 40 hr without significant deterioration in effluent
quality. Since the waste is inhibitory and the influent strength is high,
critical point violations are likely for this waste treatment situation. (For
more discussion, please refer to the inhibitory kinetics sections of Chap-
ters 1, 2, and 3.) At 15,000 and 20,000 mg/L influent COD, the limiting
detention times are approximately 70 and 110 hr, respectively. Figure
7.17 shows that the impact of 1 mg/L of cobalt is to reduce treatment
capacity at an S; of 10,000 mg/L from a detention time of 40 hr to 70 hr.
There is virtually no capacity at 15,000 mg/L COD. Other process analy-
ses of the 10 mg/L cobalt situation showed that there was essentially little
processing capability by the biomass. It is recognized that the seed bio-
mass employed in these tests was relatively unacclimated to the presence
of cobalt and that a better kinetic response would have been obtained
166 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

130 4
120 |

a= 110
Q
ro) 100 —
Oo

& 90 4
> 80
g= 70 |
a< 60 4

E so -
5 40 4
Fe
i) 30 4

20 +
°

10 4
0 | T T tT T i eo | Se ae ea Shan. ta omer
20 60 100 140 180 220 260
REACTOR DETENTION TIME (hours)
+ 10000 ° 15000

Figure 7.17. Impact of 1 mg/L cobalt on operating envelope of

activated sludge system. Compare with Figure 7.16.

with a biomass more acclimated to cobalt. However, the analysis showed

that eliminating cobalt from the waste stream substantially enhances the
performance envelope of the plant. Evaluation of the impact of acclimat-
ing the biomass to cobalt by examining “intermediate” levels of cobalt
contamination of the biomass as judged to be too time-consuming for
the purposes of this work.
Figures 7.18 and 7.19 show favorable results regarding the impact of
pretreating the first-stage feed. The biokinetic constants used to generate
these curves were the normal plant feed and the pretreated plant feed for
Figures 7.18 and 7.19, respectively. The engineering constants were the
same as those used in Figures 7.16 and 7.17. A comparison of Figures
7.18 and 7.19 shows that the effect of pretreatment is to extend the
performance envelope of the plant. Figure 7.19 shows that the effect of
pretreatment is to increase the plant capacity to a detention time of 30 hr
at an influent COD of 10,000 mg/L. The limiting detention times for
15,000 mg/L and 20,000 mg/L are 60 and 90 hr, respectively. The normal
plant feed has limiting detention times of 80 and 160 hr for S; values of
 CASE STUDIES AND APPLICATIONS 167

~ —)

EFFLUENT
CONC.
WASTE
COD/L)
(mg

220 260
REACTOR DETENTION TIME (hours)
+ 10000 ° 15000

Figure 7.18. Operating envelope of activated sludge system treating

cobalt-containing wastewater.

10,000 mg/L and 15,C00 mg/L, respectively. The analyses also showed
that at 20,000 mg/L COD, the plant is only able to provide marginal
treatment at best for the untreated waste.
This example provides a good case study of how to use respirometric
testing and associated modeling analyses to predict and to quantify the
impacts of upstream process changes (in this case, pretreatment to
remove cobalt) on the performance capacity of the plant. This enables
plant operations staff and management to make informed economic
decisions regarding operational modifications instead of having to rely
on “plant folklore” or less structured analytical approaches.
168 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

140 4 —
130 4
120 4
a 110 4
8) 100 +
oO

g 90 -|
S 80 4
is)
ns 70 +
< 60 4

5a
& 50
40 4
4
:
w 30

20 4 ~s
lo 4 i
0 Se ee a eee T T
20 60 100 140 180 220 260
REACTOR DETENTION TIME (hours)
+ 10000 ° 15000 a 20000

Figure 7.19. Operating envelope of activated sludge system treating

pretreated (cobalt precipitated out) wastewater. Compare
with Figure 7.18.

CASE 4: USE OF RESPIROMETRY FOR SCREENING AND

RANKING APPLICATIONS

Background

A large international corporation was investigating the biodegradabil-

ity of five organic chemical products that were soon to be manufactured
at an overseas location. The overseas plant was facing the possibility of
meeting strict effluent requirements for COD; consequently, it was
important to identify any troublesome or recalcitrant products which
may pass through the biological treatment process and contribute to
effluent COD. A corollary effort involved devising a ranking procedure;
both the screening and ranking procedures involved the interpretation of
respirometric data.
 CASE STUDIES AND APPLICATIONS 169

Approach

The general testing approach involved the operation of a bench-scale

activated sludge unit that supplied seed for separate batch testing using
respirometry. A schematic outlining this procedure is given in Figure
7.20. A waste stream containing equal-COD portions of the target com-
ponents was fed to the bench-scale activated sludge reactor. The bench-
scale unit operated at a long mean cell residence time in order to simulate
extended aeration conditions that would be utilized by the field system.
After an acclimation period of three weeks, biomass was harvested from
the bench-scale reactor and used in separate batch respirometry tests to
identify the biodegradability of the individual component wastes.
A total of four respirometry tests were performed using a six-unit
electrolytic respirometer. Each test was performed using six 1-L flasks,
with each flask containing the desired amount of waste sample, nutri-
ents, and biomass seed for each test. Cumulative oxygen uptake mea-
surements were taken for a period of 48 hr. The waste samples added to

General Approach
Feed synthetic waste mixture
to bench reactor

Evaluate biodegradability of
biomass on individual products

Develop biodegradability ranking

Figure 7.20. General approach for organic screening procedure using

respirometry.
170 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

each flask were taken from stock solutions made from concentrated
waste samples. Variations in initial COD values in the test flasks were
attributed largely to the emulsified nature of the wastes. This points out
the utility of respirometry for evaluating biodegradation in this waste
treatment situation. Relying on COD measurements alone leads to
ambiguous results. The respirometric data augments conventional treata-
bility data and allow the analyst to obtain a clearer interpretation of the
biodegradation data.

Results

The results of the batch respirometry tests are depicted in Figures 7.21
through 7.24. Analysis of the data indicated that approximately half of
the batch test results represent good COD depletion and O, uptake bal-
ances. The COD depletion is accounted for in the mass of oxygen utilized
and the increase in cell mass. This balance analysis provided an addi-

ENDOGENOUS
WASTE A
D WASTE 8B
WASTE C
WASTE D
WASTE &
xPoond

(mg/L)
UPTAKE
OXYGEN
CUMULATIVE

TIME (hours)

Figure 7.21. Case Study 4 Respirometric Test #1: Individual

components plus endogenous seed reactor (Colvin et al.,
1991).
 CASE STUDIES AND APPLICATIONS 171

ee
al
o
o
O WASTE B . er on
E 50 © WASTE C ao
wy 4 WASTE D :
x4 XWASIEE
a.
=)
40 Ss

FE
i 30 o—_O—O—_0—o— 09 9
ze 0 WY
3s a 0 o 7, y, , Y, 9, Y,
wi =20 ee
=

eS : 16 0000-000
5 10 VE es ae
5 a
0 10 20 30 40 50

TIME (hours)

Figure 7.22. Case Study 4 Respirometric Test #2: Individual

components plus “All Waste” reactor (Colvin et al.,
1991).

tional check on the biodegradation data and aided in eliminating confus-

ing results.
The results of the first respirometry test are presented in Figure 7.21.
After a review of the data, it was determined that Waste A was most
biodegradable, while Wastes B and D were problematic. The oxygen
uptake results for Waste C were judged to be erroneous and inordinately
high because the COD depletion did not match O, uptake. A check of the
equipment indicated that the respirometer had a faulty seal that resulted
in an overproduction of oxygen in the Waste C flask. The balance analy-
sis enabled us to sort out the analytical confusion quickly.
The results of the other tests supported the results presented in Figure
7.21. Wastes B and D were problematic; Waste A tested relatively biode-
gradable with the other wastes remaining somewhere in the middle.
These results provided a qualitative ranking of the wastes. The client
requested that we provide a more quantitative interpretation of the data.
The ultimate goal is to formulate a methodology to to quantify the
biodegradation ranking of each waste component and to develop an
algorithm for estimating the effluent soluble COD contribution of each
172 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

a fl0 Vv ALL WASTES - E

=a c WASTE A
3 O WASTE B
= 9 © WASTE C .
Ww 4 WASTE D
<t 80 X WASTE £ A
Ooi
a 60
a) 50 z : . , ¥
(S) 40 y, SS
WwW ———
= 30 Xxfr
S20 Vu é
= 70 ae gt
0 10 20 30 40 50

TIME (hours)

Figure 7.23. Case Study 4 Respirometric Test #3: Individual

components plus “All Waste” reactor (Colvin et al.,
1991).

waste. The objective is to predict the impact of the individual wastes on

effluent quality as measured by COD. The approach utilized the data in
the batch test to rank the components by the product of ACOD (COD
removed) divided by the initial COD used in the test. This gives the
fraction of initial COD removed. ACOD was computed using the
respirometric data according to the following expression:

ACOD = O, uptake/(1 - O,Y,) (7.2)

Table 7.8 lists the ACOD/initial COD values for the five wastes tested.
The information in Table 7.8 is used to develop the biodegradability
ranking in Table 7.9. In order to predict the impact of the influent waste
mixture on effluent COD characteristics, the batch results listed in Table
7.9 need to be “scaled” to the bench-scale system. This is stated because
bench-scale reactors that simulate full-scale systems generally have
greater COD removal than batch reactors. The bench scale-reactor for
this study achieved a 73% removal of COD. The ranking procedure for
individual wastes is formulated by assuming that each waste is character-
ized by its own COD removal ratio. This removal efficiency is assumed
 CASE STUDIES AND APPLICATIONS 173

v ENDO GENOUS
O REACTOR SOLUBLE EFFLUENT
ay #0
S
cw]
ts
y 3
3A tiel
fe) e&
WASTE D
<< x WASTE E
(ate 0)
J ,
a 25
ee
= 20 O
zs UO
=
w 15
O
=
AN
= —— =
2
CO yy —— GH
oS
5 a
$ : v,
v, o 0D

Re
v, V 7, v,

OC
e _
= 5

i si
20 30
0 10
TIME (hours)

Figure 7.24. Case Study 4 Respirometric Test #4: Components B, C, D,

and E: effluent BOD; and endogenous reactor (Colvin et
al., 1991).

Table 7.8 ACOD/Initial COD Values Calculated for Each Waste from
the First Three Batch Respirometry Tests (From Colvin et al., 1991).
Test No. Waste A Waste B Waste C Waste D Waste E
1 ——+ Or322 0.361* 0373 O13
pe 0.561 0.420 07332 Ozte9 0.378
3 0.635* 0.220 O10 0.790* 0.532
Avg. 0.598 0.321 0.401 0.434 0.541
+ ACOD/Initial COD Value was greater than 1.0.
* Used actual final ACOD for this computation.
Note: The results of Test 4 are not included because Waste A was not
analyzed in this test.
174 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Table 7.9 Summary of ACOD/Initial COD Data and Biodegradability

(From Colvin et al., 1991)
Waste Average ACOD/Initial COD*
A 0.598
E 0.541
D 0.434
C 0.401
B 0.321
*Test 4 not included in calculations.

to be additive which results in an aggregate ACOD for the waste

mixture:

ACOD, te ACOD, ele — ACOD Waste Mixture (7.3)

Adding up the average ACOD/initial COD numbers from Table 7.7

and dividing by 5 (the bench-scale reactor was fed a mixture consisting of
five equal-COD contributions of the five waste products) yields 0.459.
The bench-scale unit had a COD removal efficiency of 73%. The scale
factor is calculated as 0.73/0.459 which equals 1.59. The adjusted
ACOD/ initial COD values for the wastes are given in Table 7.10. The
information in Table 7.10 is utilized by inserting each removal factor into
a formula for predicting the impact of waste stream composition on
effluent COD. It is recognized that this approach is subject to more
rigorous evaluation. It does, however, represent a reasonable starting
point for utilizing bench-scale respirometric data for predicting the
impact of different waste components on effluent COD.

Table 7.10 Adjusted ACOD/Initial COD Factors for Each Waste

(From Colvin et al., 1991)
Waste Adjusted Factor
A 0.951
E 0.860
D 0.690
(a 0.638
B 0.510
Combined Waste 0.730
 CASE STUDIES AND APPLICATIONS 175

KEY CONCEPT SUMMARY

The key concepts contained in this chapter are:

The process modeling methodology described herein is suitable

for both design and operational application. Case 1 presented an
example of a design application while Cases 2 and 3 are examples
of operational analyses.
A process analysis is performed by utilizing the predictive equa-
tions presented in Chapter 3. If a system does not conform to
complete/mix or cells-in-series, then the same derivation proce-
dure presented in Chapter 3 can be used to derive equations
which are appropriate for the particular system configuration.
The modeling procedure described in Chapter 3 can be applied to
any reactor configuration to yield configuration-specific predic-
tive equations.
The model equations are calibrated using the respirometric tech-
niques given in Chapter 5. Various issues concerning the acquisi-
tion of kinetic data and the impact of environmental conditions
on the values of the biokinetic constants are discussed in Chap-
ters 5 and 6.
A model analysis identifies the values of the engineering control
parameters wich are needed to prevent plant upset. Alternatively,
the analysis quantifies the operating envelope for a given set of
influent conditions and values of the biokinetic constants.
Respirometric methods can also be applied in a screening mode
to evaluate the relative potential of wastes to pass through a
biological treatment system. These applications are described in
Case 4.

REFERENCES AND SUGGESTED ADDITIONAL READING

Colvin, R.J., Rozich, A.F., Gaudy, A.F. Jr., and Martin, J. (1991).
Application of a Process Control Model Calibrated with Respirome-
try to Predict Full-Scale Activated Sludge Performance. Proceedings,
45th Purdue Industrial Waste Conference, Lewis Publishers, Chelsea,
Michigan, pp. 501-508.
Colvin, R.J., Rozich, A.F., Hough, B.J., and Gaudy, A.F. Jr. (1991).
“Use of Respirometry to Evaluate Biodegradability of Emulsified Spe-
cialty Chemical Products” (with R.J. Colvin, B. Hough, and A.F.
176 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Gaudy, Jr.) Proceedings, 45th Purdue Industrial Waste Conference,

Lewis Publishers, Chelsea, Michigan, pp. 477-486.
Constable, S.W., Rozich, A.F., DeHaas, R., and Colvin, R.J. (1991).
“Respirometric Investigation of Activated Sludge Bioinhibition by
Cobalt/Manganese Catalyst,” Presented, 46th Annual Industrial
Waste Conference, Purdue University, West Lafayette, IN, May,
1991.
Gaudy, A.F. Jr., Rozich, A.F., and Lowe, W.L. (1987). “Design Criteria
for Treatment of Combined Waste,” Presented, Annual Conference
of the Environ. Eng. Div., ASCE, Orlando, FL.
Peil, K.M., and Gaudy, A.F. Jr. (1971). “Kinetic Constants for Aerobic
Growth of Microbial Populations Selected with Various Single Com-
pounds and with Municipal Wastes as Substrates.” Appl. Microbiol.,
21, pp. 253-256.
Rozich, A.F., and Colvin, R.J. (1990). “Formulating Strategies for Acti-
vated Sludge Systems,” Water Engineering & Management, 137, 10,
pp. 39-41.
Rozich, A.F., Usinowicz, P.J., and Colvin, R.J. (1991). Operation of a
Superfund Site Biological Wastewater Treatment Plant Using Predic-
tive Respirometric Analyses. Presented, 64th Annual WPCF Confer-
ence, Toronto, Canada.
Rozich and Gaudy, Inc., (1986). “Development of a Process Control
Plan for Managing Waste Inhibition at the Patapsco Wastewater
Treatment Plant,” Engineering Report to the City of Baltimore, MD,
Wastewater Facilities Division.
Rozich and Gaudy, Inc., (1986). “Determination of the Numerical Values
of the Biokinetic Constants and Implications to the Design of the
Expanded Facilities for the Patapsco Wastewater Treatment Plant,”
Engineering Report to the City of Baltimore, MD, Wastewater Facili-
ties Division.
 Appendix: Computer Programs

PART I: LISTING OF THE COMPUTER PROGRAM USED

TO FIT GROWTH DATA TO THE MONOD FUNCTION TO
DETERMINE THE BIOKINETIC CONSTANTS u,,,, AND K,

THIS PROGRAM UTILIZES THE CURVE FITTING ROUTINE

“DUNLSF” FOR FITTING BATCH GROWTH DATA TO THE
MONOD FUNCTION.
IMPLICIT REAL *8 (A-H, O-Z)
EXTERNAL PARAB
INTEGER M,N,LDFJAC,IPARAM(6),NSIG
DIMENSION XGUESS(2),XSCALE(2), FSCALE(100)
DIMENSION RPARAM(7), X(2), FVEC(100), FJAC(100,2),
F(100)
COMMON /ULF/Y(100),V(100)
DATA FSCALE/100*1.0D0/,XSCALE/2*1.0D0/
N=2
140 FORMAT(/ ’ ENTER NUMBER OF DATA POINTS, I.’)
150 FORMAT (/ ’ ENTER, THE EXPERIMENTAL SO AND U
DATA, PLEASE.’)
[35 FORMAT (/ ’ ONE DATA SET PER LINE PRINTING SO
FIRST.)
160 FORMAT (/’ ENTER INITIAL GUESSES FOR PARAMETERS
“UMAX” AND “KS” 1, PLEASE/”)
162 FORMAT (/ ’ THE INITIAL GUESS FOR UMAX IS YOUR
LARGEST EXPERIMENTAL GROWTH RATE (U).’)
164 FORMAT (/ ’ THE INITIAL GUESS FOR KS IS YOUR
SMALLEST INITIAL SUBSTRATE CONCENTRATION
(SO).’)
165 FORMAT (/’ ENTER VALUES FOR EPS AND NSIG (START
WITH EPS =0.0 AND NSIG= 3).’)
170 FORMAT (/5X,’ UMAX (1/H) = ’,F10.3)
173 FOPRMAT (/5X,’ KS (MG COD/L) = ’,F10.1)
Wg
178 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

rip FORMAT (/5X,’ SSQ =’,F10.6)

180 FORMAT (/? DO YOU WISH TO CHANGE YOUR INITIAL
GUESSES? YES =1 AND NO=2’?
185 FORMAT (/? DO YOU WISH TO TRY ANOTHER SET OF
DATA? YES=1 AND NO =2’
190 FORMAT (/” NICE TALKING TO YOU.’)
10 WRITE (6,140)
READ(5,*) M
LDFJAC-100
WRITE(6, 150)
WRITE(6, 155)
DO 15,1=1,M
15 READ(5,*) V(1), (1)
17 CONTINUE
WRITE(6, 160)
WRITE(6, 162)
WRITE(6, 164)
READ(5,*) XGUESS(1),XGUESS(2)
CALL DU4LSF(IPARAM,RPARAM)
WRITE(6, 165)
READ(5,*) EPS,NSIG
RPARAM(4) = EPS
IPARAM(2) = NSIG
IPARAM(3) = 1000
IPARAM(4) = 1000
CALL DUNLSF FPARAB,M,N,XGUESS,XSCALE,
FSCALE, IPARAM,RPARAM,X,FVEC,FJAC,LDFJAC)
WRITE (6,170) X(1)
WRITE (6,173) X(2)
DO 20, K=1,M
SSQ = FVEC(K)**2 + SSQ
20 CONTINUE
WRITE(6,175) SSQ
WROTE(6, 180)
READ(5,*) ANS
IF (ANS.EQ.1) GO TO 17
WRITE(6, 185)
READ(5,*) ANS2
IF (ANS2.EQ.1) GO TO 10
WRITE(6, 190)
STOP
 APPENDIX 179

END
SUBROUTINE PARAB(M,N,X,F)
IMPLICIT REAL*8 (A-H,O-Z)
INTEGER M,N
DIMENSION X(2), F(100)
COMMON /UFL/Y(100),V(100)
DO 5, 1=1,M
F(I) = Y(I) - (X()*V())/(X2)
+ VD)
RETURN
END

PART II: LISTING OF THE COMPUTER PROGRAM USED

TO FIT GROWTH DATA TO THE HALDANE FUNCTION
TO DETERMINE THE BIOKINETIC CONSTANTS 4u,,,,, K,
AND K;

“NONLIN FORTRAN” CURVE FITTING PROGRAM

FOT FITTING BATCH GROWTH DATA TO
THE HALDANE GROWTH FUNCTION
IMPLICIT REAL*8 (A-H,O-Z)
EXTERNAL FUNCI
INTEGER N,M,IPARAM(6), LDFJAC
DIMENSION XGUESS(3),XSCALE(3), FSCALE(100),
RPARAM(7)
DIMENSION X(3),FVEC(100),FJAC(100,3),F(100)
COMMON/BLOCK1/XEXPT(100), YEXPT(100)
DATA FSCALE/100*1.0D0/,XSCALE/3*1.0d0/
DATA IN,IOUT /5,5/
CALL ERSET(4,-1,0)
WRITE(6, 10)
FORMAT(’ ENTER THE NUMBER OF DATA POINTS,
PLEASE.’)
READ(5,*) M
WRITE(6,15)
FORMAT(? ENTER THE LARGEST EXPERIMENTAL
GROWTH RATE, PLEASBP’)
READ(5,*) USTAR
WRITE(6,20)
180 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

20 FORMAT(’ ENTER THE EXPERIMENTAL SO AND U DATA,

PLEASE.’)
WRITE (6,25)
25 FORMAT(’ ONE SET OF DATA PER LINE STARTING WITH
SO.’”)
DO 301=1,M
30 READ(5,*) XEXPT(I), YEXPT(I)
32 WRITE(6,35)
35 FORMAT(’ ENTER A VALUE FOR “EPS” (0.1 TO 0.00000001),
PLEASE’)
READ(5,*) EPS
N=3
LDFJAC = 100
SSQ =0.0
INPUT GUESS MATRIX: USTAR < UMAX < 10*USTAR
1.0 wexcaeKS2 ee 0ens000
LO. << KE = Waste.
UMAXH = 10*USTAR
UMAXL=USTAR
DO 500 I=1,8
XGUESS(1) = UMAXL*(1.28571438*FLOAT(I)-0.28571438)
DO 600 J=1,12
XGUESS(2) = 27.181818*FLOAT(J)-26.181818
DO 700 K=1,15
XGUESS(3) = 107.071429* FLOAT(K)-106.071429
CALL DU4LSF(IPARAM,RPARAM)
IPARAM(I)=1
IPARAM(2) =3
IPARAM(3) = 20000
IPARAM(4) = 20000
RPARAM(1)=0.0
RPARAM(4) = EPS
CALL DUNLSF(FUNC1,M,N,XGUESS,SCALE,FSCALE,
IPARAM,RPARAM,X,FVEC,FJAC,LDFJAC)
IF((IPARAM(3).EQ.20000).OR.(IPARAM(4).EQ.20000)) GOTO
79
IF((X(1).GT.UMAXL).AND.(X(1).LT. UMAXH)) GOTO 40
GOTO 700
40 IF((X(2).GT.1).AND.(X(2).LT.300)) GOTO 50
GOTO 700
50 IF((X(3).GT.1). AND.(X(3).LT.1500)) GOTO 60
 APPENDIX 181

700 CONTINUE
600 CONTINUE
500 CONTINUE
79 WRITE(6,80)
80 FORMAT(//,’ CURVE FIT HAS FAILED! TRY AGAIN.’
GOTO 82
60 DO 70 L=1,M
SSQ = FVEC(L)**2 + SSQ
70 CONTINUE
GOTO 90
82 WRITE(6,85)
85 FORMAT(’;) WOULD YOU LIKE TO TRY A NEW “EPS”
VALUE? YES=1 AND NO =2’)
READ(5,*) ANS
IF(ANS .EQ. 1) GOTO 32
GOTO 155
90 WRITE(IOUT, 100) X(1)
100 FORMAT(/5X,’ UMAX (1/HOURS) =’,F10.4)
WRITE(IOUT, 110) X(2)
110 FORMAT(/5X,’ KS (MG COD/L) =’,F10.4)
WRITE(IOUT, 120) X(3)
120 FORMAT(/5X,’ KI (MG COD/L) =’,F10.4)
WRITE(IOUT, 130) SSQ
130 FORMAT(/5X, SSQ =’,F10.5)
GOTO 82
155 WRITE(6, 160)
160 FORMAT(’’? WOULD YOU LIKE TO TRY ANOTHER SET OF
DATA? YES=1 AND NO =2?
READ(S5,*) ANS2
IF(ANS2 .EQ. 1) GOTO 5
STOP
END
SUBROUTINE FUNCI (M,N,X,F)
IMPLICIT REAL*8 (A-H,O-Z)
DIMENSION F(100),X(3)
COMMON/BLOCK1/XEXPT(100), YEXPT(100)
DO 800 I=1,M
F(I) = YEXPT(I)-(X(1)/(1 + (X(2)/XEXPT(I)) + (XEXPT(I)/
X(3))))
800 CONTINUE
RETURN
182 ACTIVATED SLUDGE PROCESSES USING RESPIROMET
RY

END
Bounding Rules for “Nonlin Fortran”
Coefficient Allowable Range
pes OS Be = 10%"
K, 1< K, s 300
K; 1< K; < 1500
 INDEX

Acclimation state, 97, 107 computer programs in

Activated sludge, defined, 2 determination of, 92, 177-182
Aeration, 4 cost-effective means of
Aeration tank oxygen transfer measurement of, 97f
requirements, 69-74 critical point analysis in, 52, 54,
Assimilative capacity, 123 55, 56, 58
ATAD. See Autothermal aerobic engineering models and, 63, 65
digestion estimates of, 89
Autodigestion, 3, 5, 6, 7, 28, 29, 30 from growth data, 90-93
Autothermal aerobic digestion influent waste strength and, 77,
(ATAD) process, 117, 118, 119 78-82, 95
measurement of, 97f
Bacteria, 2, 5, 99, 104, 117, 120. See oxygen transfer and, 71, 74
also specific types for Patapsco Wastewater
Batch reactors, 4, 6, 26, 89, 106, 109 Treatment Plant, 142-150
Batch tests, 52 population diversity and, 119-120
Bench-scale reactors, 51, 105 respirometry for generation of,
Bioaugmentation, 98, 126-127 82-93
Biochemical oxygen demand (BOD), shock loads and, 98, 121-126
Dole 208 126 sludge minimization and, 68
five-day, 134 specific growth rate and, 98-105
limits for, 157 temperature and, 117-119
measurement of, 22 toxics and, 111-117, 127
at Patapsco Wastewater Treatment true cell yield and, 93-94, 95
Plant, 134, 150 waste composition and, 105-111
soluble, 150 Biokinetics, 1-23
Biodegradation kinetics, 101 cell yield and, 5-6
Biokinetic constants, 35, 39, 40, 44, constants in. See Biokinetic
77-95, 97-127 constants
as acclimation state measurement, decay rate and, 6-8
97 general principles of, 2-5
bioaugmentation and, 98, 126-127 growth rate and, 6-8
as catalytic characterization of quantitative concepts in, 5-21
system ecology, 97 sludge yield and, 5-6

183
184 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

Biokinetic testing, 53 delta, 4, 5, 20, 70

Biomass biokinetic constants and, 78-82
acclimation state of, 97, 107 defined, 78
concentration of, 22, 44, 47, 48, initial COD ratio to, 172, 174
BY) at Patapsco Wastewater
excess, 31-33, 155 Treatment Plant, 150
exiting of, 27 depletion of, 170, 171
growth of, 19, 88 effluent, 168
mass balance for, 39, 51, 52 effluent quality and, 133
oxygen uptake and growth of, 83 effluent requirements for, 168
production of, 51 glucose, 79
recycle of, 33-37 high concentrations of, 92
replacement time for, 32 influent, 123, 166
separation of effluent from, 36 initial, 11, 88, 172, 174
storage capacity of, 155 leakage of, 67, 111, 123, 134, 140,
Bioreactor. See Reactor 142, 153
BOD. See Biochemical oxygen organic acid, 161
demand at Patapsco Wastewater Treatment
Bulking, 66 Plant, 134, 135, 137, 140, 142,
153
Carbohydrates, 2, 3. See also specific phenol, 15
types product, 111
Carbon, 2, 3, 79 remaining, 20
biodegradable, 105 removal of, 19, 21, 78, 83, 116,
concentration of, 9 WG 172
exhaustion of, 16 residual, 79
mixture of sources of, 105, 106, resistance of to biodegradation, 78
108 soluble, 78, 142, 171
total organic, 5 substrate, 83, 84
Carbon compounds, 3. See also at Superfund site, 158, 159
specific types total removal of, 78
Carbon dioxide, 113, 115, 116-117 usable, 64
Catabolite inhibition, 109 waste, I11
Cell decay rate, 6-8, 22 Cobalt, 113, 115, 161-167
Cell recycle, 26 COD. See Chemical oxygen demand
Cell residence time, 33, 62, 99 Completely mixed reactors, 26, 39,
Cellulose, 156 40, 51, 52, 71
Cell yield, 5-6, 22, 93-94, 95 Computer programs, 92, 177-182
Chemical oxygen demand (COD), 4, Continuous culture system, 37, 39,
23, 169 108
biodegradable, 44 Continuous flow reactors, 26, 27
carbon dioxide impact on, 117 Critical detention time, 53, 54
cell, 70 Critical dilution rate, 29
cobalt and, 165, 166, 167 Critical growth rate, 53
 INDEX 185

Critical net specific gravity rate, 29 Engineering constants, 39, 40, 44, 54,
Critical operating points, 52, 53, 54, 56, 65
56 Engineering models, 39-59, 61-75
Critical point analysis, 51-58, 59 biokinetic constants and. See under
Critical point curves, 54-58 Biokinetic constants
Critical reactor condition, 58 critical point analysis in, 51-58
Critical specific growth rate, 14, 17, for multiple reactors, 47-51
Pee iil oxygen transfer requirements and,
Critical substrate concentration, 69-74
22-23 predictive equations for, 40-46
Cumulative oxygen uptake, 84, 85, reconciliation with other methods,
169 62-66
recycle parameters and, 66-67
sludge production minimization
Decay rate, 6-8, 22, 95
and, 67-69
Declining phase of growth, 7
Enzymes, 13, 109. See also specific
Degradative capability, 97
types
Delta COD, 5, 70
Equalization methods, 40, 53
biokinetic constants and, 78-82 Excess biomass, 31-33, 155
defined, 4, 20, 78
Excess sludge, 9, 31-33, 43, 93
initial COD ratio to, 172, 174
at Patapsco Wastewater Treatment
Flow rate, 39, 53, 134, 137
Plant, 150
Food-to-microorganisms ratio, 62,
Detention time, 27, 39, 41, 44, 56,
65, 66
58, 69
critical, 53, 54
hydraulic, 53, 67 Glucose, 79, 99, 105, 106, 108, 111,
low, 148 120
Dilution rate, 27, 29, 31 Glycerol, 111
Dissolved oxygen, 70, 71 Growth, 2, 23, 40
Doubling time, 7 balanced, 3
DUNLSF, 92, 177-179 biokinetic constants for. See
Biokinetic constants
curve for, 15-16
Effluent declining phase of, 7
COD of, 168 equations for, 27
deterioration of, 51, 53, 54, 56 exponential, 6, 15
quality of, 9, 30, 36, 44, 46, 78, inhibition of, 12
99 at molecular level, 13
COD limits for, 133 net, 67
prediction of, 93, 99 oxygen uptake and, 20, 83
separation of biomass from, 36 peak rate of, 67
substrate in, 57 quantification of, 2
washout of, 51, 54 rate of, 6-8
Energy balance, 20, 21, 33 recycle parameters and, 61, 66-67
186 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

respirometric data and, 84-89 Kjehdahl nitrogen, 99

specific rate of. See Specific
growth rate Landfill leachate, 117
stimulation of, 12 Lignins, 80
substrate concentration and rate Lipids, 2, 3. See also specific types
of, 39, 78
substrate removal and, 83
unbalanced, 3 Maintenance plot, 94
Manganese, 161
Mannitol, 111
Haldane curve, 51
Mass balance, 33, 34, 51, 52, 59, 70,
Haldane equation, 12, 15, 17-18, 18,
73
25, 31, 43, 46, 48, 59. See also
Mass balance equations, 39, 41, 47,
Inhibitory waste; Toxic waste 48, 74
biokinetic constants and, 78, 82, MCRT. See Mean cell residence time
90, 91, 92 Mean cell residence time, 33, 62, 99
computer programs and, 179-182
Metabolic control mechanisms, 109,
defined, 22
111
growth data fitted to, 92
Metals, 111, 113, 127, 162. See also
Haldane kinetics, 46
specific types
Haldane relationship, 86
Methane, 118-119
Haldane waste. See Inhibitory waste;
Microbial growth. See Growth
Toxic waste
Mixed liquor suspended solids
Heavy metals, 111, 113, 127. See also
(MLSS), 158, 159, 160
specific types
Mixed liquor volatile suspended
Hydraulic control of specific growth
solids (MLVSS), 159, 160
TALC WZO MA od COS
MLSS. See Mixed liquor suspended
Hydraulic detention times, 53, 67
solids
Hydraulic holding time, 27
MLVSS. See Mixed liquor volatile
Hydraulics, 26, 28
suspended solids
Models. See also specific types
Industrial discharges, 113 engineering. See Engineering
Inflow rate, 58 models
Influent, 39, 44, 95, 123 predictive. See Engineering models
Influent waste strength, 40, 44, 53, reactor. See Reactor modeling
77, 78-82, 95 Monod equation, 11, 17, 18, 25, 29,
Inhibition constant, 22, 44 31, 42, 46, 48, 59. See also
Inhibitory constant, 14 Noninhibitory waste
Inhibitory models, 31 biokinetic constants and, 78, 82,
Inhibitory waste, 17-18, 31, 48, 58. 90, 91, 92, 114
See also Haldane equation; computer programs and, 177-179
Toxic waste defined, 22
critical point analysis for, 551-58 growth data fitted to, 92
predictive equations for, 43-46 toxics and, 114
specific growth rate in, 11-15 Monod relationship, 35, 86
 INDEX 187

Monod waste. See Noninhibitory Oxygen-activated sludge systems, 113,

waste; Nontoxic waste 115
Multiple-cell reactors, 39, 47-51, 72, Oxygen transfer, 115, 116
74 capacity for, 72
Mutant bacteria, 104 requirements for, 61, 69-74
Oxygen uptake, 4, 5, 19, 19-21, 22,
Net specific gravity rate, 29 Pee AAI
Newton-Raphson iteration method, accumulated, 5
44 balance in, 170
Nitrification, 99, 117 biomass growth and, 83
Nitrifying bacteria, 99, 117 conversion of data on, 84
Nitrogen, 2, 3, 99, 109, 117 cumulative, 84, 85, 169
Noninhibitory models, 29-30 description of, 26
Noninhibitory waste, 29, 31, 48, 58. growth and, 20, 83
See also Monod equation; increase in, 5
Nontoxic waste quantification of, 2
predictive equations for, 40-43 quotient for, 23
specific growth rate in, 8-11 rate of, 73
Nonnitrogenous carbon compounds, steady-state, 74
3. See also specific types substrate removal and, 83
Nontoxic waste, 16-17. See also substrate utilization and, 83
Noninhibitory waste; specific
types Patapsco Wastewater Treatment
Nucleic acids, 2, 3. See also specific Plant, Baltimore, 84, 116,
types 133-156
background of, 133-134
Once-through systems, 29-31, 33, 37 biokinetic constants for, 142-150
Operating conditions, 52 design recommendations for,
Organic carbon, 5 154-156
Oxidation. See Oxygen uptake evaluation of model predictions
Oxidative assimilation, 3 for, 150-153
Oxygen, 5, 36 model analysis for, 134-142
biochemical demand for. See operational recommendations for,
Biochemical oxygen demand 154-156
(BOD) review of data on, 134-142
chemical demand for. See Chemical PCP. See Pentachlorophenol
oxygen demand (COD) Pentachlorophenol (PCP), 126
dissolved, 70, 71 Phenol, 15, 52, 88, 105
high concentration of, 116 biodegradation kinetics of, 101
mass balance of, 73, 74 biokinetic constants for, 55
supply of, 3 COD of, 15
transfer of. See Oxygen transfer degradation of, 108
uptake of. See Oxygen uptake glucose mixed with, 105, 106, 108
188 ACTIVATED SLUDGE PROCESSES USING RESPIROMETRY

heterogeneous populations grown specific growth rate of. See

on, 103 Specific growth rate
high growth rates for, 104 totally fluidized, 26
inhibitory model for, 46 two-stage, 108, 120
metabolism of, 120 volume of, 57
removal of, 106 Recycle, 61, 66-67
Phosphorus, 2, 3, 154 biomass, 33-37
Plug flow reactors, 72 flow rate of, 39
Population diversity, 119-120 flow ratio of, 44, 56, 57, 64, 151
Porge’s formula, 19, 84 pumping capacity of, 58
Predators, 5 rates of, 58
Predictive models. See Engineering variables in, 74
models Recycle sludge, 39, 40, 44, 56, 57, 64,
Primary feeders, 2, 5 155
Proteins, 2, 3, 111. See also specific Regression, 109
types Reliability analysis, 58
Protozoa, 2, 5
Replacement time for biomass, 32
Respiration. See Oxygen uptake
Pumping capacity, 58
Respirometric ratio, 20, 21
Purification, 2, 3, 5, 19, 20

Salts, 111
Reactor modeling, 25-38. See also
Saturation constant, 22
Reactors
Screening and ranking of
biomass recycle and, 33-37
applications, 168-174
excess sludge and, 31-33
Secondary feeders, 2
inhibitory, 31
Shake flask method, 142
noninhibitory, 29-30
Shock loads, 98, 111, 121-126, 155
once-through, 29-31, 33, 37 Sludge
Reactors. See also specific types age of, 33
batch, 4, 6, 26, 89, 106, 109 concentration of, 4, 35
bench-scale, 51, 105 excess, 9, 31-33, 43, 93
carbon dioxide accumulation in, minimization of production of, 61,
113, 116-117 67-69
completely mixed, 26, 39, 40, 51, production of, 43, 44, 61, 67-69
by, Ah| recycle, 39, 40, 44, 56, 57, 64, 155
continuous flow, 26, 27 settling characteristics of, 37
detention time for, 27, 39, 41, 56 yield of, 5-6
failure of, 51, 52, 55, 56, 58 Solids, 126, 158, 159, 160
growth rate for. See Growth Solids flux loading, 58
modeling of. See Reactor Sorbitol, 111
modeling Specific cell decay rate, 22
multiple-cell, 39, 47-51, 72, 74 Specific gravity rate, 29
once-through, 29-31, 33, 37 Specific growth rate, 7, 22, 27, 39
plug flow, 72 biokinetic constants and, 98-105
 INDEX 189

cell recycle and, 26 leakage of, 65, 67

Ciiticalemde hjneoseesi mass balance for, 39, 52
hydraulic control of, 26, 27, 28, noninhibitory. See Noninhibitory
Ds BS). Shes waste
in inhibitory waste, 11-15 removal of, 3, 5, 19-21, 23, 26, 83
maximum, 22 specific growth rate relationship to,
8-18, 25
in noninhibitory waste, 8-11
in inhibitory waste, 11-15
in nontoxic waste, 16-17
in noninhibitory waste, 8-11
reduction in, 56
in nontoxic waste, 16-17
substrate relationship to, 8-18, 25 in toxic waste, 17-18
in inhibitory waste, 11-15 utilization of, 62, 83-84
in noninhibitory waste, 8-11 Superfund site case study, 156-161
in nontoxic waste, 16-17 Suspended solids, 126, 158, 159, 160
in toxic waste, 17-18
Steady-state conditions, 35, 37, 41, Temperature, 117-119
AV OS. silis Sika Why 12? TKN. See Total Kjehdahl nitrogen
Stoichiometry, 84 TOC. See Total organic carbon
Substrate. See also specific types Total Kjehdahl nitrogen (TKN), 99
Totally fluidized reactors, 26
balance of, 29
Total organic carbon (TOC), 5
COD of, 83, 84
Toxics, 111-117, 127. See also
concentrations of, 9, 14, 22, 47,
specific types
48
Toxic waste, 17-18, 48. See also
biomass growth kinetics as Haldane; Inhibitory waste
function of, 111 Treatment efficiency, 52
critical, 22-23 True cell yield, 93-94, 95
growth rate and, 39, 78 Two-stage reactors, 108, 120
mass balance for, 52
effluent, 57 Volatile suspended solids, 159, 160
inhibitory. See Inhibitory waste
inhibition of, 12 Washout, 51, 54, 56, 57
| loen ll loon(oon
 Also available from Lewis Publishers:

Toxic Substances in Municipal Wastewater: A Guidance Manual for

Negotiating Permits
This timely new book presents praetieal information about the effective
implementation of water-quality-based toxies control for wastewater
treatment plants. It addresses five main subject areas, including
approaches to toxics control, legal issues, toxicity identification and
reduction evaluations, whole effluent toxicity testing, and chemical
_ specific limits. Case examples from actual experiences illustrate
workable solutions for implementing new water-quality-based permit
requirements.
The book provides a useful reference for wastewater treatment
personnel, industrial wastewater treatment professionals, government
agency regulators, environmental consultants, and environmental
attorneys.

ee

EES”
Cee
eae ewe:

| L449
. ISBN 0-87371-449-0