United States
                   Environmental Protection
                   Agency	
Risk Reduction Engineering
Laboratory
Cincinnati OH 45268
                   Research and Development
 EPA/600/S2-88/065 Apr. 1989
&ER&         Project Summary
                   Oxygen  Utilization  in  Activated
                   Sludge Plants:  Simulation  and
                   Model  Calibration
                   C. Robert Baillod
                   The objective of this study is to
                  apply recent advances in activated
                  sludge  process modeling to the
                  simulation of oxygen utilization rates
                  in  full-scale  activated sludge
                  treatment plants. This is done by
                  calibrating  the  International
                  Association  for Water  Pollution
                  Research and  Control (IAWPRC)
                  Model  and  associated  SSSP
                  (Simulation  of  Single-Sludge
                  Processes for Carbon Oxidation,
                  Nitrification,  and Denitrification)
                  micro-computer   software  to
                  operating data at  six full-scale
                  activated  sludge treatment plants.
                  Field data were used to calibrate the
                  key  biological parameters contained
                  in the model  so that the oxygen
                  utilization rates, dissolved oxygen
                  concentrations, mixed liquor volatile
                  suspended solids concentrations,
                  and  process performance simulated
                  by  the  model  matched  the
                  corresponding quantities observed in
                  the treatment plants.
                   The results showed that the model
                  and  associated software  package
                  provide a useful capability to analyze,
                  simulate,  and  predict  oxygen
                  utilization  rates. It was possible to
                  obtain  reasonable  agreement
                  between the measured and simulated
                  values  of oxygen  uptake  rate,
                  dissolved oxygen concentration, and
                  other process parameters at most of
                  the  plants studied.  The key  model
                  parameters were the heterotrophic
                  yield coefficient, heterotrophic decay
constant,  and autotrophic maximal
specific growth rate constant
  This information is of value to
engineers in the cost-effective design
and  operation  of wastewater
treatment  systems  because it
provides a data base of applicable
stoichiometric  and kinetic model
parameters  that the  engineer  can
use, with  appropriate judgment, to
similate and predict the behavior of
oxygen  transfer  systems  in
wastewater treatment
  This  Project  Summary  was
developed by EPA's Risk Reduction
Engineering Laboratory, Cincinnati,
OH, to announce key findings of the
research project that is  fully
documented in a separate report of
the same title  (see Project Report
ordering information at back).

Introduction
  The activated  sludge process is the
most widely used method for secondary
wastewater treatment  in  the United
States, and its popularity is increasing.
Providing oxygen to the active organisms
through  aeration is  the most energy
intensive  aspect of  activated sludge
process operation and consumes 60%to
80% of the total energy requirements in
wastewater treatment.  Moreover, the
performance of the  biological treatment
system is intimately linked to the proper
design and operation of  the aeration
system. Thus, it  is important that the
oxygen demands of the biological system
and the oxygen transfer capability of the

-------
aeration  equipment  be accurately
predicted. Over the past few  years,
significant progress  has been made in
measuring  and modeling the oxygen
transfer capability of aeration  equipment
under clean  water  and   process
conditions. This study is concerned with
improvements in the techniques used to
simulate and predict oxygen utilization in
activated  sludge wastewater  treatment
systems.
   The objective of this study  is to apply
recent advances  in activated sludge
process  modeling (Grady et  al.,  1986)
(Bidstrup and Grady,  1987) to  the
simulation of oxygen uptake rates in
full-scale  activated sludge  treatment
plants. This  is  done by calibrating the
International Association for  Water
Pollution  Research   and  Control
(IAWPRC) Model and associated  SSSP
software  to  operating data at six full-
scale activated sludge treatment plants.

Models Applied to Oxygen
Utilization in Activated Sludge
   Applying  predictive models that  can
mimic the performance of  biological
waste treatment systems can  aid in their
design and operation. Because biological
waste treatment systems are  inherently
complex, the model must be somewhat
simpler than the actual system. Over the
past several  years, various conceptual
models  ranging from single  substrate,
single biomass aerobic models  to
multiple  substrate, multiple biomass
anaerobic models have  been  applied to
biological waste treatment.

Conventional Approach to
Heterotrophic Oxygen
Utilization
   The conventional approach  represents
an activated sludge treatment system by
a  single  completely mixed reactor  and
separator  and applies  overall process
oxygen, substrate, and biomass balances
to  estimate  carbonaceous  oxygen
demand. The substrates, processes, and
pathways involved in  this  model are
illustrated in Figure  1.
   This   model  considers  only  two
processes:   aerobic  growth  of
heterotrophic  biomass and  oxidative
decay of heterotrophic  biomass.  The
decay process is  termed "oxidative"
because it uses oxygen. It is important to
distinguish this decay process from the
hydrolytic decay process incorporated
into the IAWPRC Model. The hydrolytic
process uses no oxygen, and  because of
this,  the  magnitude of  the  oxidative
decay coefficient is roughly  38%  of the
hydrolytic decay coefficient. The  model
assumes that each unit of biomass that
undergoes decay is  partially oxidized
and  partially converted to paniculate
products, which are resistant to further
decay. The observed fraction of biomass
converted   to  resistant  particulate
products, XD, has been reported  to  be
0.2.
   The conventional approach  can  be
applied to  a wastewater treatment plant
by  viewing the  aeration  tank  as  a
completely  mixed  reactor.  Material
balance equations for biomass, substrate,
oxygen,  particulate products, and inert
particulates  can be written over the plant
and  combined  to  express  the
carbonaceous oxygen utilization rate,  Rc
as,
 Rc  =
                                  (1)
  / 1 + b  0  - ¥„ - f Yub 9  \
  / 	c  c	H    o  H c  c \
  \           i + b e            /
                  c  c
where:

Rc        = mass of oxygen required
           per unit time for the carbon
           oxidation processes, M/t

Q        = flow  rate  entering  and
           leaving the process, L3/t

Srjl,Sc2  = total carbonaceous oxygen
           demand  (ultimate BOD) of
           the  flows  entering  and
           leaving the process, M/L3
bc        = conventional decay coeffi-
           cient for  oxidative decay of
           biomass, 1/t
YH       = yield  coefficient, mass of
           volatile suspended solids
           COD produced  per  unit
           mass  of carbonaceous
           oxygen demand use

©c       = solids  retention time  (SRT)
           at steady state  conditions,
           defined as the biomass in
           inventory divided by  the
           rate of biomass wasting, t

f0        = fraction of biomass yielding
           nonbiodegradable
           particulate products upon
           oxidative decay,  generally
           taken as 0.2

   The decay  coefficient, bc,  and  yield
coefficient,   YH,  are  biological
parameters characteristic of the biomass
itself,  whereas the solids retention tin
Qc, is the key parameter that controls t
process  operation  and performanc
Typical values of solids retention tir
range from  3 to 12  days.  Numerk
values of  Y depend on the units in whi
biomass  and substrate are expresse
Typical values for Y range from 0.45
0.7 mass  of VSS COD  produced per u
mass  of  feed  COD  utilized. Typic
values of  the oxidative  decay coefficie
bc, range  from 0.04 to 0.4 per day.

Conventional Approach for
Nitrogenous Oxygen Utilization
   The overall  nitrogenous  oxygi
utilization  rate  can  be  calculated  basi
on the stoichiometric requirements that
mg of ammonia  or  Kjeldahl  nitrogi
requires  4.57  mg  of  oxygen   f
conversion to nitrate, and denitrificatii
of 1  mg  of nitrate nitrogen provides ,
equivalent oxygen  credit of  2.86  m
Therefore,  the  nitrogenous  oxyg<
demand, Rn, can be expressed as:
where:

NTN  = the amount of  total nitrog<
       converted  to nitrate,  mass/tim
       determined by  subtracting tl
       total Kjeldahl  nitrogen (TKN)
       the  effluent from  the  tot
       nitrogen available for  nitrificatic
       The amount of nitrogen availab
       for nitrification 
-------
  Substrate
 Organics. Sc
 Oxygen, S0
Decay
  Products, Xp
                   Pathway
                                         Process
                                     Aerobic
                                     Heterotrophic
                                     Biomass Growth
                                        Heterotrophic
                                       Biomass Decay
                                      /Negative Growth)
                                                             Mediated by
                                                              Biomass XB
                                                            Biomass XB
Figure 1.    Substrates, pathways, and processes included in the conventional carbonaceous
           oxygen utilization model.
TKN and nitrate concentrations in the
effluent. It does  not provide  any
capability to simulate  or predict these
concentrations  based on wastewater or
process parameters.

IAWPRC Model
   In 1986,  the  IAWPRC task  group  on
"Mathematical Modeling for Design and
Operation  of  Biological  Wastewater
Treatment"  developed a  model
applicable to an activated sludge system
performing  simultaneous   carbon
oxidation, nitrification, and denitrification.
This IAWPRC  Model  is described  by
Gradyetal. (1986).
   The  substrates,  pathways,  and
processes  included  in the  IAWPRC
Model  are  described  in Figure  2.
Comparison  of  Figure 2 with  Figure 1
shows the comprehensive nature of the
IAWPRC Model.  The model  contains
nine  substrates  inolved   in  eight
processes  mediated by  two  active
biomass fractions. This model is  more
complex  than  the  conventional
carbonaceous  model  because
autotrophic  nitrification  and other
nitrogenous pathways  are included and
because two  classes  of carbonaceous
substrate  are recognized:   readily
degradable  (soluble)  organics and slowly
degradable (particulate) organics.


Solution of the IAWPRC Model
Using the SSSP Software
Package
  Bidstrup and Grady  (1987) developed
the SSSP computer  software  package
based  on the  IAWPRC  Model.  This
 oftware, written for the IBM  Personal
Computer  or compatible machines
                                       performs steady  state  and  dynamic
                                       simulations of activated sludge systems
                                       based on the IAWPRC Model. Both the
                                       program  and user's manual may  be
                                       copied freely. The program is versatile in
                                       that it allows the  user  to  define the
                                       system configuration by using up to nine
                                       completely  mixed  reactors  in  series.
                                       Additional flexibility  is provided  by the
                                       capability to  define  the influent addition,
                                       return sludge recycle, and mixed liquor
                                       recirculation  flow diagrams between the
                                       various reactors.
                                         The program  contains  "default"
                                       values of the stoichiometric and kinetic
                                       parameters   required by  the  IAWPRC
                                       Model.  These  serve as reasonable
                                       starting  points for  simulations  and
                                       calibration. One valuable feature of the
                                       program is  that it  allows  the user to
                                       specify either reactcr dissolved  oxygen
                                       concentration or the  reactor  volumetric
                                       mass transfer coefficient, KLB. The mode
                                       in which  the  K|_a  value  is specified is
                                       particularly   useful  for  calibrations in
                                       which the reactor K(_a is  known. When
                                       K|_a is specified, however, the program
                                       uses identical values for  the  dissolved
                                       oxygen  half-saturation  constants for
                                       autotrophs and heterotrophs, and where
                                       different  values for these parameters are
                                       in the input  file, they are averaged to
                                       select the value used. For this study, the
                                       default values of the DO  half-saturation
                                       constants were set at 0.1  mg/L, which is
                                       the   normal  default value  for  the
                                       heterotrophic organisms.


                                       Activated Sludge Plants
                                       Studied
                                         Six municipally owned  and operated
                                       wastewater treatment plants  located in
Wisconsin and  Michigan  were selected
for  calibration  of  the  oxygen  uptake
model. Table 1  summarizes the  process
and flow characteristics of the plants. A
wide range of plant sizes, configurations,
industrial contributions, and  SRT were
represented. The percent of plant studied
indicates the portion of the plant  flow
passing  through  the aeration  basins
monitored during  the field studies. The
seventh column gives  the  length/width
ratios of the aeration basins, and these
values are related to the basin residence
time distribution.

  All six of the  plants were concurrently
being studied as part of the ASCE-EPA
Fine Bubble  Diffused Aeration  Design
Manual Project, and because  of  this,
background  information  on process
configuration and  operation was  readily
available. Moreover, frequent  off-gas
measurements of oxygen  transfer were
being made  at  five of the six plants as
part  of the  ASCE-EPA project.  These
measurements were particularly valuable
because  they  produced accurate
estimates of the process volumetric mass
transfer coefficients for oxygen (K(_a).

  The  24-hr studies  conducted  at each
of the  plants were designed to  acquire
information necessary  to  calibrate the
IAWPRC Model. The  steady state model
solution, based  upon flow  weighted
average conditions, was used to calibrate
key parameters, and these parameters
were later used in the dynamic  solution
to simulate diurnal variations in  oxygen
uptake rate  (OUR).  By comparing the
simulated and measured diurnal  ranges,
the  usefulness  of the dynamic  solution
for design purposes could be assessed.

-------
Table 1. Plants Included in Field Study







cs
c
SF
L/W

Plant
Portage Lake
Green Bay
Madison
Monroe
Jones Island
South Shore
= Contact Stabilization
= Conventional
= Step Feed
= Length/Width
Average
Flow
m3/day
8,700
182,000
151,000
8,300
288,000
371,000

Annual
BOD-5
mg/L
150
375
170
418
300
162

- Percent of
Plant
Studied
50%
33%
13%
33%
4%
5%

Process
SRT Days
10.6
3.1
16.4
8.4
2.8
4.3

Configuration
CS
CS
C
SF
C
SF

Aer., UW
1.5
3.4
26
8.2
15
6.5

Industrial Contribution
Flow %
<5%
30%
6%
17%
11%
6%

BODS %
<5%
50%
15%
50%
38%
18%

Oxygen Uptake Rate
Measurement
  OUR was measured in  two ways:  by
the conventional  batch  BOD bottle
technique and  by  off-gas  measurement.
The BOD bottle technique required that a
mixed  liquor sample  be  rapidly
withdrawn, aerated, and transferred to a
BOD bottle. The DO concentration in the
bottle was monitored, and the OUR was
calculated from the rate of DO decrease.
Although, this  method itself  gives  an
accurate and  precise  in-vitro measure
of the oxygen utilization rate that occurs
in the BOD bottle, this may not represent
the in-situ  OUR occurring  in the reactor
from which the sample was  withdrawn.
Two conditions, oxygen  limitation and
soluble substrate  limitation,  can cause
the OUR measured by the  BOD bottle
method  to differ significantly  from  the
in-situ value.
  Oxygen  limitation  arises when the  in-
situ DO is near zero; this  causes the  in-
situ OUR to be limited by  the availability
of oxygen.  A  sample  subjected to the
BOD  bottle method is exposed to high
DO concentrations  and will respire at a
higher rate. Consequently, when  in-situ
DO concentrations are near  zero,  the
OUR  indicated  by the  BOD bottle
method will be  its potential maximal
value, which will be greater than the  in-
situ  OUR.  Substrate  limitation arises
when  the  in-situ  exogenous  substrate
concentration  is near  zero; this causes
the OUR  of  a withdrawn  sample  to
decrease between the time of withdrawal
and the time  at which the  BOD bottle
OUR  is  measured.  To  avoid  these
limitations, the BOD bottle OUR values
measured  on samples taken from
regions where  the DO  was less than  1.5
mg/L  were not used in model calibration,
and  more weight was given to  bottle
OUR  values  measured  under
endogenous conditions.
  Off-gas measurements made at  five
of the six plants were used in the gas-
side  oxygen  balance to estimate  the
process water volumetric mass transfer
coefficient (K|_a).   These coefficients
were then incorporated  into  the  liquid-
side  oxygen  balance in the IAWPRC
Model to relate the DO concentration to
the OUR. Therefore, when the model  was
calibrated to  simulate  a DO value in
agreement with the measured DO  value,
the simulated  OUR agreed with the OUR
measured by  the off-gas measurement.
Theoretically,   this  enabled the  in-situ
OUR to be   determined  from  DO
measurements and served as a check on
the BOD bottle OUR.  However, even
though the off-gas  measurements were
made simultaneously  with  the  24-hr
studies at two of the plants, changes in
wastewater characteristics cause  the
alpha factor to change continuously  and
make it impossible  to  determine  the
process water K|_a  with  a precision
greater than 10%.


Model Calibration
  Calibration  of the  IAWPRC   Model
required that  the  wastewater feed
components required by the model be
determined from the plant data, that the
kinetic and stoichiometric  parameters
required by the model be estimated, and
that  criteria for calibrating the model to
the plant data be established.  Thirteen
components   are  considered in  the
IAWPRC Model and most of them may
be  present  in the feed  wastewater.
During the 24-hr plant studies,  samples
of the wastewater fed to and samples of
clarified  effluent from the activated
sludge processes were analyzed for total
COD,  soluble COD,  ammonia, total
organic nitrogen, and nitrate. These d
were used to establish the concentrat
of feed components for model calibrate
  A sensitivity analysis of the bionru
parameters showed  that OUR was m
sensitive  to the  heterotrophic  a
autotrophic yield and decay coefficier
Overall, the heterotrophic  yield, YH, a
decay, bn,  showed the most impact
the simulated values of OUR,  DO, a
MLVSS. The autotrophic yield, Ya, dec;
ba,  and maximal specific growth  ra
Ham. had the most impact on simulal
values of ammonia and nitrate. Values
the more   sensitive  parameters  we
determined by calibrating  the model
the data. Parameters to which the moi
predictions were insensitive were set
the adjusted  default values.  In genei
these  values  were  determined
applying a temperature correction to 1
default kinetic parameters given  in  I
SSSP program.
  This study was primarily  concern
with the oxygen utilization  aspects of I
model. Because the  IAWPRC moc
simulated many other system respons
(e.g.,  MLVSS, DO, nitrate) in addition
OUR,  an effort was made to calibrate 1
model so that it simulated MLVSS, C
nitrate, and  ammonia  as  well as  OL
Primary   emphasis was  placed
matching  the OUR and DO.  Less
weight was placed  on   matching  t
MLVSS and nitrogen concentrations.
  The model parameters were estimat
by calibrating the model to the avera
plant  data  measured during  the  24
studies. The  process flow diagram w
modeled as realistically as possible
using  a combination  of completely mix
reactors and perfect clarifier/thickeners
steady state simulation was performed
using  the  SSSP software package w
default values  of all parameters.
second  steady  state simulation  w

-------
                      Substrate
Pathway
                                                                 Process
Mediated by
                     Soluble
                     Organic Ss
                     Oxygen So
                     Ammonia SNH
                     Alkalinity SAik
                     Nitrate SNO
                     Soluble
                     Organic N SND
                     Paniculate
                     Organic N XND   <-
                      Paniculate
                      Organic Xs
                      Decay
                      Products XE
                     Aerobic
                     Heterotrophic
                     Biomass Growth
                                                            Anoxic
                                                            Heterotrophic
                                                            Biomass Growth
                      Autotrophic
                      Biomass Growth
                                                              Hydrolysis and
                                                              A mmonification
                                                              (Non-Growth)
                      Heterotrophic
                      Biomass Decay
                      (Negative Growth)
                                                             Autotrophic
                                                             Biomass Decay
                                                             (Negative Growth)
Heterotrophic
Biomass XBH
                                           Heterotrophic
                                           Biomass XBH
Autotrophic
Biomass XB*
                                           Heterotrophic
                                           Biomass XBH
Heterotrophic
Biomass XBH
                                           Autotrophic
                                           Biomass, XBA
                     Figure 2.    Substrates, pathways, and processes included in the IAWPRC Model.
 performed  by  using  temperature
 adjusted  default  values  for the  kinetic
 parameters contained in  the model. The
 model-simulated values  of  OUR,  DO,
 MLVSS, nitrate, and ammonia, based on
 the  temperature  adjusted  default
 parameters,  were then compared  with
 the measured  average values, and  the
 model  was calibrated  to the  measured
 data by  adjusting  selected key  model
 parameters.  In all  but one case, where
 the hydrolysis rate constant was  also
 changed, these adjustments were limited
 to  the  heterotrophic yield and  decay
 parameters and the autotrophic maximal
 specific growth rate.   In  addition,  the
, autotrophic half saturation constant for
 DO was made  equal to the default value
   of  the  heterotrophic  half  saturation
   constant for DO (0.1 mg/L).
     Dynamic  simulations were  also
   performed with the use  of  the SSSP
   software package with the calibrated
   parameters determined  from  the steady
   state simulation.  This  gave simulated
   diurnal  profiles for OUR and  DO. These
   were then compared with the measured
   profiles to judge how  usefully the model
   simulates  reasonable variations  in
   oxygen utilization.


   Results
     The   average  measured   values
   reasonably agreed with  the steady state
 simulated values for OUR, DO, MLVSS,
 nitrate,  and  ammonia  at  most  of  the
 plants studied.  Only  two of three  key
 parameters  were calibrated to achieve
 this agreement. The other 17 parameters
 were  set at  their temperature  adjusted
 default  values.    Notable  lack  of
 agreement between  the simulated  and
 measured values was evidenced at the
 Jones Island Plant where the simulated
 and  measured  effluent  ammonia
 concentrations did not agree well, and at
 the  South   Shore  Plant  where  the
 simulated and measured MLVSS values
 did not agree well.
  Table 2 summarizes the values of the
 key parameters that resulted from model
 calibration.

-------
Total Process Average Oxygen
Utilization Rates

  An estimate of the total daily process
average oxygen  utilization  rate can  be
made by multiplying the estimated OUR
values  by  the  corresponding reactor
volumes.  In  the  absence  of direct
measurements of OUR values for all tank
regions, the best estimates of  OUR  for
the various  reactor  regions  are given  by
the SSSP  simulation  based on the
calibrated parameters. These  estimates
are consistent with the measured OUR,
DO,  off-gas, and  solids data  used  to
calibrate  the  model.  An example
calculation for the portion of the Madison
plant studied is shown in Table 3.
  Values of the  best estimates of the
total  process average oxygen utilization
rates calculated in this manner for all  six
plants are shown in column 2 of Table 4.
These quantities  apply to that portion of
the plant included  in the 24-hr studies.
Consequently,  the magnitudes of the
utilization rates  reflect both the size of
the plant and the fraction of flow treated
by the portion of the plant included in the
study. Thus, the utilization rates reported
for the Green  Bay  plant where the
portion under study treated 33%  of the
plant flow are  much greater than the
rates reported for the larger Jones Island
plant where the  portion  under  study
treated only 4% of the plant flow.
  A similar, more conventional estimate
may  be made by  viewing the process
aeration tank  as a single completely
mixed  reactor  and  applying  the
conventional model given by Equation 1
to calculate the  carbonaceous oxygen
utilization rate, Rc. The  values of the
heterotrophic yield  coefficient, YH, used
in the IAWPRC (SSSP) and conventional
models are identical.  However, the
appropriate  value of the oxidative  decay
coefficient, bc, for use in the conventional
model  was calculated as  38% of the
IAWPRC hydrolytic decay  coefficient, b.
The  total oxygen utilization rate is then
calculated according to Equation 2 after
determining the  nitrogenous oxygen
demand from Equation 3.
  Column 3 of Table 4 summarizes the
total  process average oxygen utilization
rates calculated for all six plants  using
the conventional  model. For most of the
plants,  the agreement  between the
IAWPRC and conventional models was
excellent as  the estimates generally
agreed  within  10%. The  exceptions to
this  were  the two Milwaukee plants,
Jones Island and South Shore, where the
conventional  approach  estimated  an
oxygen utilization rate 15% greater than
the IAWPRC approach.
  It can  be concluded that, for given
values   of  the  yield  and   decay
coefficients, the two approaches produce
estimates of the average process oxygen
utilization rate, which agree reasonably
well. In this study, the IAWPRC model
was  calibrated  to plant  data  and,
therefore, was  judged to give the better
estimates  of the  process  oxygen
utilization rates. There is some indication,
based on the  Milwaukee plants, that the
conventional   approach  may  slightly
overestimate the oxygen utilization rate.
  These results  show  that the
conventional approach  is useful but has
several disadvantages  when  compared
with  the  IAWPRC  Model  and  SSSP
software package. Both approaches need
applicable values  of the heterotrophic
yield  and  decay  coefficients. The
conventional model, however,  does not
distinguish  between slowly and readily
degradable  substrates and cannot easily
be  adapted to simulate  spatial and
temporal  variations of OUR.  Moreover,
the conventional  approach can  neither
simulate  nor predict nitrogenous oxygen
demand.  It  can  only  calculate
nitrogenous oxygen  demand based  on
observed influent  and  effluent  nitrogen
concentrations. On the other hand, it was
observed in this study that the  IAWPRC
Model needs only  one  additional
calibrated  parameter,  the  maximal
autotrophic  specific  growth  rate,  to
simulate the entire process. In almost all
cases,  default  values of all but two or
three  parameters  (heterotrophic yield,
heterotrophic  decay,  and autotrophic
maximal  specific  growth  rate) were
adequate in model calibration.

Conclusions and
Recommendations
  Based  on the results of this  study, it
can be concluded that:

  I.The  IAWPRC  Model  and  related
    SSSP  microcomputer software
    package provides a useful capability
    to analyze and simulate  the average
    oxygen utilization  rate in  municipal
    wastewater treatment plants.

  2. It was possible to obtain reasonable
    agreement between the  average
    measured  and  steady  state
    simulated values  of  OUR,  DO
    concentration,  MLVSS,  effluent
    ammonia, and effluent   nitrate
    concentrations at most of the plants
    studied.   This agreement was
    achieved  by adjusting or calibrating
    only 2  of 3  key model  parameti
    while keeping the other 17  mo<
    parameters at their adjusted  defe
    values.  The  key model  parameti
    were  the  heterotrophic  yie
    coefficient,  heterotrophic  dec
    constant,  and autotrophic  maxin
    specific growth rate constant.

  3. When provided with a realistic val
    of the process water volumetric mz
    transfer  coefficient  (Ki.a),  t
    calibrated model and software w<
    normally able to simulate the spal
    and  temporal  range  of  C
    concentrations and oxygen utilizati
    rates observed  in  the operati
    treatment plants. Instances in whi
    lack of agreement between simulat
    and observed values occurred coi
    possibly  be explained  by caus
    other than  model and/or  softwe
    inadequacy.

  4.Under  conditions  of  low  C
    concentration  (below 1  to  2  mg/
    in-vitro OUR values indicated  I
    the  traditional BOD  bottle meth'
    tended  to be greater than the in-s
    OUR values determined from o
    gas analyses.  These  differenc
    were most  evident in the inl
    sections of long aeration tanks whe
    OUR values  were  high  and  C
    values were low.

  S.When  the  IAWPRC  Model  ai
    conventional  approaches  to estimj
    average process oxygen  utilizatii
    rates  were  compared,  the  tv
    approaches agreed reasonably w
    for given values of  the yield ai
    decay coefficients. In this study,  tl
    IAWPRC Model was calibrated to tl
    data and, therefore,  was judged
    give  the  better  estimates  of tl
    process utilization rates. Moreovt
    the IAWPRC Model was judged to I
    advantageous because of its abili
    to  simulate  nitrogenous  oxyg<
    demand.

  This  study  relied  on adjusting
calibrating key parameters to show th
the IAWPRC Model  could be applied
analyze, simulate, and predict oxyg<
utilization  and  other  activated  sludc
process characteristics. Such informatk
is  valuable  because  it provides  a da
base  of  key parameters  from  whi<
engineers  can  draw information  f
process design and  operation.  Howevt
this data base  should be  expanded
include information from more than  tl
six plants  of  this study.   Moreover,
would  be  of  particular  value
demonstrate  the  applicability

-------
Table 2. Summary of Model Calibration of Key Parameters"
Plant
Default (20°C)
Madison
Monroe
Portage Lake
Jones Island
South Shore
Green Bay
Heterotrophic Yield
Coefficient,
g COD/g COD
0.67
0.57
d
0.69
d
0.46
0.57
Heterotrophic Decay
Constant, 1/day
0.62
d"
0.80
1.10
0.35
0.50
d
Autotrophic Maximal
Specific Growth
Rate, 1/day
0.65
0.34
0.34
d
0.62
d
0.45
"All other parameters were set at temperature adjusted default values.
"Signifies temperature adjusted default value.
Table 3. Calculation of Total Process Average Oxygen Utilization Rate for Madison Plant
        Based on Calibrated IAWPRC Model
Reactor Volume, m3
1 1,733
2 4,008
3 3,827
4 1,914

OUR, g/m3/day
468.7
541.1
404.8
341.3
Total
Utilization Rate,
kg/day
812.3
2,168.7
1,549.2
653.2
5,1 83.4 kg/day
Table 4. Comparison of Total Process Average Oxygen Utilization Rates Estimated by the
        IAWPRC and Conventional Models

                             Oxygen Utilization Rate, kg/day
Plant
Madison
Monroe
Green Bay
Portage Lake
Jones Island
South Shore
IAWPRC
Model Total
5,183
822
29,350
593
2,544
3,079
Conventional Model
Total
5,001
845
30,600
633
2,925
3,532
*c
3,797
587
30,448
592
2,803
2,352
«n
7,204
264
752
47
722
7,780
Percent
Difference
+ 3.5
-2.9
-4.3
-6.8
-75.
-75.
techniques for key parameter estimation
based on direct measurements.
  Accordingly, it  is recommended that
additional  in-depth studies applying the
IAWPRC Model  and associated  SSSP
software to well controlled, full-scale
wastewater  treatment  plants  be
conducted. It is also recommended that
design  engineers  be  encouraged  to
apply the IAWPRC Model and associated
SSSP  software  to activated  sludge
process design. Although other improved
models and software might be developed
in the future, this model and software are
available now and are easily usable  by
any  design  engineer. With  the aid  of
engineering judgment and the parameter
data  base given  in this report, this
approach will facilitate more  economical
and  effective design and  operation  of
wastewater treatment plants.
References
Bidstrup, S.M., and Grady, C.P.L, Jr., "A
    User's  Manual for SSSP," Clemson
    University, Clemson,  South Carolina,
    1987.

Grady, C.P.L., Jr., Gujer,  W., Henze, M.,
    Marais,  G.v.R.,  and  Matsuo,  T.,  "A
    Model for Single Sludge Wastewater
    Treatment  Systems," Wat.  Sci.
    Tech.,  75:47-61, 1986.

  The full  report  was  submitted  in
fulfillment  of  CR813162-01-2  by
Michigan Technological University under
the sponsorship  of the  U.S.  Environ-
mental Protection Agency.

-------
  C. Robert  Baillod is with Michigan Technological University, Houghton,  Ml
  49931.
  James A. Heldman is the EPA Project Officer (see below).
  The  complete  report, entitled "Oxygen Utilization in Activated Sludge Plants:
    Simulation and  Model Calibration," (Order No. PB 89-125 9671'AS; Cost:
    $T5.95, subject to change) will be available only from:
           National Technical Information Service
           5285 Port Royal Road
           Springfield, VA 22161
           Telephone:  703-487-4650
  The EPA Project Officer can be contacted at:
           Risk Reduction Engineering Laboratory
           U.S. Environmental Protection Agency
           Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
      BULK RATE
POSTAGE & FEES P/
         EPA
  PERMIT No. G-35
Official Business
Penalty for Private Use $300

EPA/600/S2-88/065
 0000329
                                            PS
                                              POCTIOH
                               CHICAGO

-------