WASTEWATER TREATMENT

D ON A L D W. SUI MDSTROM
and
HERBERT E. KLEI
Department of Chemical Engineering
The University of Connecticut
\
PRENTICE-HALL, INC., Englewood Cliffs, NJ. 07632
Library of Congress Cataloging in Publication Data
Sundstrom, Donal d Wi l l i am, 1931-
Wastcwater treatment.
Includes bibliographical references and index.
1. SewagePurification. I. Kl ei , Herbert E. ,
1935- joint author. II. Ti tl e.
TD745.S85 628'.3 78-13058
ISBN 0-13-945832-8
Editorial production supervision
and interior design by: JAMES M. CHEGE
Cover design by: EDSAL ENTERPRISES
Manufacturing buyer: GORDON OSBOURNE
1979 by Prentice-Hall, Inc., Englewood Cliffs, N.J. 07632
\
All rights reserved. No part of this book
may be reproduced in any form or
by any means without permission in
writing from the publisher.
Printed in the United States of America
1 0 9 8 7 6 5 4 3 2 1
PRENTICE-HALL INTERS
PRENTICE-HALL I
PRENTICE-HALL <
PRENTICE-HALL OF?l NDi A: P$^ArgLmm
PRENTICE-HALL OF JAPAN, JNO:,- 'Xpkyo: Aaca>\ . j .n.
PRENTICE-HALL OF SOUTHEAST ASIA PTE. LTD. , Singapore
WHITEHALL BOOKS LIMITED, Wellington, New Zealand
CONTENTS
Pr ef ace xv
SECT I ON I : WA S T E WA T E R CHARACT ERI ST I CS
A N D T R E A T ME N T PROCESSES
1. Wa s t e wa t e r Char act er i z at i on 3
1 - 1 . Water supply and consumption 3
1 - 2. Water quality 5
1 - 3. Wastewater components and analysis 10
Physical characteristics 11
Chemical characteristics 13
Biological characteristics 19
vii
viii Contents
2. Wast ewat er T r e a t me n t Pr ocesses
2- 1. Classification of processes 28
Pre- and primary treatment 29
Secondary treatment 29
Tertiary treatment 31
2-2. Physical-chemical treatment 32
2-3. Sludge disposal 33
2-4. Industrial wastes 37
SECTI ON I I . BI OLOGI CAL PROCESSES
3. Aer at i on and Ma s s Tr ansf er
3- 1. Aeration methods 42
3-2. Limiting resistances for mass transfer 43
3-3. Gas-liquid interface 44
Two-film theory 44
Correlations for mass transfer coefficients
Penetration and surface renewal theories
^ 3-4. Liquid mixing 56
Axial dispersion description and tracer analysis
Mixing with mechanical agitation 63
3-5. Mass transfer near the biomass 66
Liquid-solid resistance 66
Diffusion within the biomass 67
4. Bi ol ogi cal Mechani sms and Ki net i cs
4 - 1 . Describing chemical reactions 75
4- 2. Enzymes as biological catalysts 77
4- 3. Mechanisms in biological reactions 78
4-4. Growth rates in batch reactors 82
4-5. Biological reaction kinetics 84
Michaelis-Menten model 84
Other kinetic models 88
Contents
4- 6. Kinetic constants from batch reactor studies
4- 7. Kinetic constants from f l ow reactor studies
4- 8. Temperature, pH and other effects 98
Mo d e l s f or Bi ol ogi cal React or s
5 - 1 . Basis for models 107
5- 2. Batch reactor 108
5- 3. Wel l - mi xed reactors wi thout recycle 110
5- 4. Wel l - mi xed reactors wi t h recycle 113
5- 5. Stability wi t h recycle solids 116
5- 6. Pl ug- f l ow reactor 118
5- 7. Fl ow reactor wi t h axial mixing 722
5- 8. Wel l - mi xed reactors in series 124
5- 9. Comparison of reactors 128
5- 10. Fixed-film reactors 129.
5 - 1 1 . Trickling-filter models 131
5- 12. Rotating biological contactor 136
5- 13. River analysis 137
Et i ol ogi cal Wa s t e T r e a t me n t Pr ocesses
6 - 1 . Steady-state design equations 745
6- 2. Biological design parameters 145
Kinetic constants 145
Sfudge age 146
Sludge volume index 147
Process loading factors (F/M ratio) 149
Sludge production 149
Oxygen requirements 152
6- 3. Activated sludge processes 154
Process descriptions 154
Nitrification-denitrification 162
Design procedure 164
6- 4. Anaerobic digestion 168
Chemical parameters 169
Process parameters 170
X
Contents
6- 5. Fixed-film biological systems 772
Trickling filters 172
Rotating biological contactor 174
SECTI ON I I I . PHYSI CAL PROCESSES-
7. Se di me nt a t i on, Thi ckeni ng, and Fl ot at i on
7 - 1 . Classifications of settling 188
7- 2. Settling of discrete, particles 188
7- 3. Ideal settling basins 190
7- 4. Tube settlers 755
7- 5. Flocculent suspensions 196
7- 6. Gravity thickening 755
7- 7. Continuous thickeners 200
7- 8. Sedimentation practice 207
. 7 - 9 . Frotation 209
7- 10. Dissolved-air flotation process 209
7 - 1 1 . Design of air flotation units 27 7
7- 12. Flotation practice 275
8. Fi l t r at i on and Ce nt r i f uga t i on
8 - 1 . Types of filters 219
8- 2. Theory of granular filters 223
8- 3. Granular filter practice 227
8- 4. Theory of vacuum filtration 229
8- 5. Vacuum filtration practice 233
8- 6. Centrifugation 234
8- 7. Theory of centrifugation 235
8- 8. Centrifuge practice 237
Contents
9. Ads or pt i on 247
9 - 1 . Nature of adsorbent 241
9- 2. Nature of adsorbate 242
9- 3. Adsorption equilibria 243
9- 4. Equilibrium batch adsorption 249
9- 5. Rates of adsorption 253
9- 6. Continuous fl ow adsorbers 254
9- 7. Adsorption column design 256
9- 8. Adsorption practice 269
9- 9. Regeneration 270
10. Me mb r a n e Separ at i on Pr ocesses
1 0 - 1 . Membrane processes 275
10- 2. Reverse osmosis 2 7 7
Membrane structure and rejection mechanism
Osmotic pressure 279
Transport models and flux equations 285
Concentration polarization 286
10- 3. Ultrafiltration 289
10- 4. Electrodialysis 292
SECT I ON I V. CHE MI CAL PROCESSES
1 1 . Che mi c a l Equi l i br i a i n Aqueous Syst ems 301
1 1 - 1 . Chemical equilibrium 302
11 - 2 . Equilibrium concentration 303
11 - 3. Electrolyte solutions 305
11 - 4. Equilibrium calculations 307
11 - 5. Acids and bases 308
11 - 6. Carbonate systems 310
11 - 7. Dissolved carbonates 311
274
277
Contents
11 - 8. Carbonate concentration diagrams 314
11 - 9. Contact wi t h C 0
2
gas 317
11- 10. Contact wi t h solid calcium carbonate 318
1 1 - 1 1 . Carbonate equivalence points 319
11- 12. Alkalinity 320
11- 13. Neutralization 322
11- 14. Solubility 323
11- 15. Phosphorus equilibria 325
11- 16. Water softening by precipitation 327
11- 17. Gas stripping 331
Coagul at i on
1 2 - 1 . Properties of suspended solids 335
12- 2. Destabilization mechanisms 338
12- 3. Destabilization chemicals 341
12- 4. Flocculation model 348
12- 5. Applications to wastewater 350
I on Exchange
1 3 - 1 . Materials 356
13- 2. Reactions 358
13- 3. Capacity and selectivity 359
13- 4. Equilibria and kinetics 360
13- 5. Fixed-bed design 361
13- 6. Applications 363
Oxi dat i on and Di si nf ect i on
1 4 - 1 . Oxidation-reduction reactions
14- 2. Redox equilibria 370
14- 3. Iron and manganese removal
14- 4. Cyanide conversion 374
368
373
Contents
xi i i
14- 5. Disinfection 375
14- 6. Kinetics of disinfection 376
14- 7. Chlorine 377
14- 8. Chlorine practice 380
14- 9. Ozone 382
SECTI ON V. SOLI DS T R E A T ME N T
A N D S Y S T E MS ANAL YSI S
15. Sl udge Di sposal 387
15- 1. Sludge characteristics 387
15-2. Sludge treatment processes 389
15- 3. Thermal processes 392
15- 4. Ultimate disposal 396
16. Syst ems Anal ysi s 398
16- 1. Optimization techniques 398
1V2. Cost analysis 400
16- 3. Process control 406
Appendi ces
A. Notation 421
B. Atomic weights 427
C. Physical properties of water 428
D. Solubility of oxygen in water 429
E. MPN index for bacterial concentrations 430
F. Conversion factors 431
G. Chemical kinetics 433
I ndex 435
5
MODELS FOR BIOLOGICAL REACTORS
Biological reactors involve a variety of geometries and hydraulic regimes.
Batch or serrft-batch reactors are often used for laboratory studies, for
anaerobi c digestion, and for manufact ure of pharmaceuticals. Flow reactors
a^^^mr n^ni y^ employed for aerobic treatment of municipal and industrial
wastes.
To model a biological process, we need information on the stoichiometry
and kinetics of the reactions, and on t he hydraulic regime of the system. The
stoichiometry of a reaction relates the quantities of reactants consumed, such
as substrates, to the quantities of product s formed, such as microorganisms.
The hydraulic regime refers to the pat t erns of flow into and out of the pro-
cess, and the mixing and distribution of fluids and solids within the reactor.
The influent and effluent for the process are described in terms of the time
variation of flow rates and concentrations of species. Any recirculation of
biological solids must also be considered in the analysis.
In flow reactors, the two extremes in mixing are represented by well-
stirred and plug-flow reactors. Intermediate degrees of mixing are often des-
cribed by well-stirred reactors in series or by plug-flow reactors with axial
dispersion. Mor e complex mixing models can be devised but their use may
not be justified because of limitations in knowledge of the system.
In this chapter, we will develop mathematical models for several types of
106
%
Sec. 5-1 Basis for Models 107
3GICAL REACTORS
geometries and hydraulic regimes,
t used for laboratory studies, for
e of pharmaceuticals. Fl ow react ors
atment of municipal and industrial
ed information on the stoichiometry
hydraulic regime of the system. The
antities of reactants consumed, such
ts formed, such as mi croorgani sms.
Tis of flow into and out of t he pr o-
fluids and solids within the react or,
are described in terms of t he time
ns of species. Any recirculation of
n the analysis.
n mixing are represented by well-
ite degrees of mixing are often des-
>r by plug-flow reactors wi t h axial
5 can be devised but their use may
knowledge of the system,
matical models for several types of
biological reactors. The overall model of the reactor is obtained by combining
the equations for the hydraulic regime and the kinetics of the reactions.
5-1 BASIS FOR MODE L S
A number of assumptions and approximations are used in deriving the models
for this chapter.
1. The chemical kinetics of the substrate and biomass reactions are
described by the Monod model including endogenous respiration.
The Monod model is adequate for many steady-state processes but is
often in error for rapidly changing processes.
2. The substrate is the growth-limiting substance and all other nutrients
are present in excess.
/ ty The kinetic constants are independent of concentrations or the degree
of conversion.
4. The yield coefficient (biomass formed/substrate consumed) is constant
and independent of the age of the microorganisms. In practice, the
yield coefficient depends upon the nature of the substrate and process
conditions. .
5. The concentration off active biological solids is proport i onal to a
readily measurable paramet er such as mi xedl i quor volatile suspended
solids. ~~
6. The rates of the biological reactions at e controlled either by chemical
kinetics or by diffusional effects.
7. The contents of the reactor are isothermal. Since biological processes
have a heat of reaction and feed conditions may change, t emperat ure
variations are possible. In most wastewater t reat ment reactors, short-
term temperature changes are usually small.
8. Physical properties of the fluid are constant. If the average values are
used, little error is introduced.
9. The transport of oxygen and substrate t hrough the fluid is rapid
relative to the rate of reaction so t hat concentration gradients in the
bulk liquid are negligible.
10. If a thickener is used to concentrate biological solids for recycle, the
reaction of substrate in the thickener is negligible.
Any or all of these assumptions can be eliminated or modified if necessary
for a particular process. If the complexity of the model is increased appre-
108
Models for Biological Reactors Chap. 5 \
ciably, solution of the equations may be difficult. In many cases, knowledge
of the biological process is not adequat e to justify a more sophisticated
analysis.
5- 2 BAT CH REACTOR
For a reaction following the Monod kinetic model, the material balances on
substrate and biomass in a bat ch reactor ar e:
k
0
XS dS
"dt
s
Y(K
m
+ S)
dX _ k
B
XS , y
dt - (K
m
+ S)
K
'
X
(5-1)
(5-2)
In the previous chapter, these equations were integrated for several special
cases and then applied to obtain the kinetic constants from experimental
dat a. Once the kinetic constants are known, they can bejused to design other
bat ch reactors operating with the same type of substrate and microorganisms.
Time
FIGURE 5-1 Schematic diagram of a batch reactor.
If no simplifying assumptions can be made, Eqs. (5-1) and (5-2) are best
solved by numerical procedures. For the common case where endogenous
respiration can be neglected, Eqs. (5-1) and (5-2) can be integrated to give
In S = In [> + Y(S - S) |J] + ( ^ ^ J
5
* ) In [:
x + Y(s - sy
k
0
t(X + YS)
YK
In X = k
0
t + In X
(5-3)
Gr
0
+ YS)
l n
[ ( z
5
) {YS +
S
X - x)~\
( 5 _ 4 )
where S and X" are initial concentrations and S and X are concentrations
at time t.
Sec. 5-2 Batch Reactor
109
Since Eqs. (5-3) and (5-4) are implicit in S and X, they must be solved by
trial and error. Wi t h known kinetic constants, we can use these equations to
design a batch reactor by calculating S and X at various intervals of time.
EXAMPLE 5-1
An organic waste is inoculated with heterogeneous microorganisms in
an agitated batch reactor. After 15 min reaction time, a sample is found to
contain 182mgCOD/ and 198 mg biomass/^. From previous studies, the
kinetic constants for this wast ej mdcyj ^^ = 0.5 hr"
1
, k
d
=0. 01 hr*"
1
,
K
m
= 75 mg COD/, Y = 0.6 mg biomass/mg COD. Estimate the COD and h
biomass concentration after 90 min operation. / '
SOLUTION:
The magnitudes of K
m
and S are comparable and X is not large enough to {
be taken as a constant. Therefore, the approximate solutions for the batch
reactor material balances given in Chapt ej ^c^mot ge used. If the endogenous I
respiration term is small compared with the growfrt term, we can apply Eqs.
(5-3) and (5-4).
Substituting values of S and Xa t / = 15 min, Eqs. (5-3) and (5-4) are
In (182) = I n[ s + 0.6(5 - 182)~]
X -f 0.65 . [X* 0.6(5 - 182)1 _ 0.5(H)(* + 0.65)
0.6(75) L X J 0.6(75)
In (198) = 0.5(Jg) + In X*
i n f Y
l 9 8
v
a 6 5
Yl
L\* A0. 65
0
- +X - 198/J
0.6(75) . T/198V 0.65
X + 0.65
The equations can be solved by trial and error for the initial concentrations of
5 and X
5 = 210 mg COD/^
X = 181 mg/ /
After 90 min residence time,
In 5 = In [210 + 0.6(210 - S) j gj ]
181 0.6(210)
+
0.6(75)
0.5(fg)(181 4-0. 6 x 210)
0.6(75)
. H81 -f 0.6(210 - 5)1
L 181 J
In * == 0 . 5 ^ ) - fi n 181
0.6(75)
181 +0.6(210)
0.6(75) . [X( 0 . 6 x 2 1 0 \
LI8IVO.6 x 210 + 181 - XJ
110
Models for Biological Reactors Chap. 5 \
Solving by trial and error,
S = 40 mg COD/ /
X = 282 mg//
To check on the assumption of negligible endogenous respiration, the magni-
tudes of the growth rate and death rate terms at 90 min can be compared.

#
k
0
XS 0.5(282X40)
Growth rate - xf~+-
=
75 + 40
= 49 mg/;(hr)
Death rate = k
d
X = 0.01(282)
= 2.8 mg/;(hr)
The endogenous rate is. about 5J/
0
of the growth rate of biomass at 90 min
and a smaller percentage at lower times. Neglecting the endogenous term is
thus a reasonable approximation.
5- 3 WE L L - MI X E D REACTORS
WI T H O U T RECYCLE
In a well-mixed flow reactor, the composition is uniform t hroughout the
reactor. Thus^the exit stream from this type of reactor has the same composi -
tion as the fluid within the reactor. The mixing action must be sufficient to
disperse the incoming feed rapidly t hroughout the reactor.
A schematic diagram of a well-mixed biological reactor of volume V is
shown in Fig. 5-2. The stream arriving at the process has a flow rate Q, a
Q
V Q
S.X
s.x
FIGURE 5-2 Schematic diagram of well-mixed
reactor without recycle.
substrate concentration S, and a biomass concentration X. A material
balance around the reactor states
Accumulation = Input Out put + Format i on by reaction
Sec. 5-3 Well-Mixed Reactors without Recycle
111
For the substrate and biomass, t he material balances are
K ^ = e 5 -QS+ r,V
V^= QX -QX + r
x
V
(5-5)
(5-6)
If the rate of reaction follows the MondJcmeUcjnj3del,
r =
k

S X
Y(K
m
+ S)
r
k
0
SX 1 y
r
' - ^ T T ^
k d X
Substituting these rate expressions, the material balances become
~di ~
Q S
- Y ( K
m
+ S)
= QX - QX + x
S
*
V
s
- k
d
XV
(s-io)
+
<vr
Equations (5-9) and (5-10) are the general material balances for an
unsteady-state well-mixed reactor obeying Monod kinetics. Certain special
cases are frequently met in practice. If the reactor operates at sjte^dy-state
conditions (constant 5, X and 2), the material balances are
0OO qo
S
_ ^J^XV_ _
Y(K
m
+ S)
k
Q
SXV
K
m
+ S
2^ - QX + * ,
K
0
- k
d
XV = 0
( 5- H)
, (5-12)
If the concentration of cells in t he feed stream is negligible, the biomass
balance is
. -QX+j^-k
d
XY=0 (5-13)
If the reactor operates in the logarithmic growth phase where endogenous
respiration is small, the biomass balance is simply
I .
no y _i_ kpSXV n
(5-14)
/
112
Models for Biological Reactors Chap. 5^
In a particular application, the appropriate substrate and biomass bal -
ances can be solved simultaneously for the unknown quantities. For example,
if feed conditions and kinetic constants are known, the material balances
can be solved to give the volume_of_the reactor required to reduce t he
effluent^u^slrate' cor7cen^ to a desired level.
The matelTalTalaiices are often written in terms of residence time, which
is defined as 6 = V/Q. The residence time is the average time t hat a fluid
element spends in tKe reactor. For a well-mixed reactor, there is a distribu-
tloiToT residence times around the average value. The steady-state Eqs. (5-11)
and (5-12) in terms of residence time are
S
X* X
k
0
SXO
K
m
+ S
k
d
xe = o
(5-15)
(5-16)
Ifj j oj t ai crooj gani sms enter with t he feed, there is a critical flow rat e at
which microorganisms are washed out of t he reactor faster t han they are
generale<rby~ t he r eact i omJJnder washout. conditions, the concentration of
biomassTh~feTeacfbr(props to zerojfrnd there is no conversion of substrate.
Setting X" = 0 in Eq. (5-16) and solving for 9 gives
9
1
(5-17)
[k
0
SKK
m
+ S)]-k
d
The critical resMence time for washout occurs when 5 becomes equal to S".
1
0 . =
[k
0
Sl(K
m
+ S")) - k
d
(5-18)
Thus, there will be no conversion of.substrate for residence times equal to or
less t han 9
W
. If the endogenous reaction is negligible, the washout criterion is
given by ^
k
0
S"
(5-19)
For stable reactor operation at residence times greater than 9
wy
Eq. (5-17)
can be rearranged to
(1 + kfi)K
m
k
0
9 - l - k<9
(5-20)
Since S does not appear in this equation, the effluent substrate concentra-
tion does not depend upon the magnitude o f j h e e n t ^ ^ con-
centration. Wi t h a given value of residence time, the conversion of substrate
increases with increasing influent concentration to maintain a constant
7 tq*J4
l
/cQ
Li*
Sec. 5-4 Well-Mixed Reactors with Recycle
113
output substrate concentration. This response represents an inherent "self-
control" by the reactor since changesTn feed concentration do not affect the
out put j ubst r at e concejUratirjr^The^m for this self-control action is
an increase in biological solids concentration t hat is sufficient to handle the
higher loading of substrate. ^
c
/ M7 ?
5-4 WE L L - MI X E D REACT ORS
WI T H RECYCLE
Microorganisms formed in a biological process are frequently fed back to the
entrance of the reactor. Since the reactions are autocatalytic in nature, the
performance of the process can be modified by recycle of biological solids.
In a properly operated biological process, t he main purpose of recycle -is to
increase the concentration of biomass in the reactor. The addition of a recycle
stream also dilutes the concentration of entering substrate and decreases the
residence time of fluid elements in the reactor.
A schematic di agram of a biological reactor with recycle of biomass is
shown in Fig. 5-3. The effluent from the well-mixed reactor is settled in a
Reactor Q+ Q
r
Clarifier
S,X
0
<
V.S.X
s.x
Clarifier
K
\ \
o
r
.s,x
r Q
W
.X,
FIGURE 5-3 Schematic diagram of well-mixed reactor with recycle.
clarifier and a port i on of the concentrated sludge is returned to the reactor
with flow rate Q
r
and concentration X
r
. If the reaction of substrate in the
clarifier is negligible, t he recycle stream will contain the same substrate con-
centration as the effluent from the reactor.
The material balances a r oundj he reactor include terms for the addi t i on
of substrate and biomass with the recycle st ream:
Accumulation = Feed input + Recycle input Out put
+ Format i on by reaction
V = QS + Q
r
S - (2 + Q
r
)S -
Y
*
X
+
S )
V ^ = QX + Q
r
X
r
- (Q
Q
+ Q
r
)X +
S
*
V
S )
- k
d
XV
(5-21)
(5-22)
1
Models for Biological Reactors Chap. 5
x
Defining a recycle ratio R = Q
r
/Q and a residence time based on fresh feed
as 9 = V/Q
9
the material balances become
. f - i ^ - ^ - i d f e s )
( 5
-
2 3 )
^ = + * - (1 + + - (5-24)
The residence time based on fresh feed is a constant for a specified flow rate
of entering feed. The true residence time, given by Vj(Q + Q
r
\ changes with
the recycle ratio of the process. If the recycle rat i o is used as a control variable,
the residence time based on total flow will be a variable instead of a constant.
For a steady-state reactor, the material balances are
S
-
S
-YMTS) =
(5
-
25)
X +RX,- (1 + R)X + x
S
*
d
s
- k
d
X9 = 0 (5-26)
If t he biomass entering with fresh feed is negligible, the material balance on
biological solids is
RX
r
- (1 + R)X + x
S
*
0
s
- k
d
X9 = 0 ' (5-27)
If t he reactor is also operating in the growt h region where endogenous respira-
tion is uni mport ant , the material balance on biomass is
RX
r
- (1 + R)X + X^ = 0 (5-28)
Biological reactors generally produce an excess of biological solids that
must be removed from the system during steady-state operation. The quantity
of excess biomass formed is equal to the net growth of biomass in the reactor.
Net biomass = *1 - k
d
XV (5-29)
Thi s excess biomass is usually wasted from the separator t hat produces
the concentrated solids for recycle. In a waste treatment plant, these solids
are sent to some type of sludge disposal process.
Two operational parameters are widely used in the design and operation
of biological treatment systems. The process loading factor is the mass of
1
Reactors
Chap. 5
esidence time based on fresh feed
(5-23)
(5-24)
i constant for a specified flow rat e
iven by Vj(Q + Qr), changes with
e ratio is used as a control variable,
>e a variable instead of a const ant .
I balances are
I k
0
SX9 _ o
' ~ Y{K
m
+ S)
- k
d
X9 = o
(5-25)
(5-26) c
B
SX9
negligible, the material balance, on
Sec. 5-4 Well-Mixed Reactors with Recycle
115
substrate consumed over a finite time period (usually a day) by the mass of
microorganisms in the react or:
~ k
0
SXV i
u
lY(K
m
+ S)j
~ XV
k
a
S
Y(K
m
+ S)
(5-30)
The quantity U is also referred to as substrate removal velocity or food to
microorganism rat i o.
The other common paramet er is the solids residence time. The true mean
solids residence time, #
m
, is the ratio of the amount of solids in the system to
the sum of the rates of biomass synthesis and solids input with fresh feed. If
the solids are assumed to be largely in the reactor and the reactor is taken to
be well mixed, t hen the true mean solids residence time for the conventional
activated sludge process of Fig. 5-3 becomes
9
m
=
XV
XQ -{- k
0
XSV
(5-31)
(K
m
+ S)
At steady-state conditions, t he denomi nat or of Eq. (5-31) equals the rate at
which biomass is lost from the system by outflow and endogenous respiration,
'^L-k
d
x9 = o
(5-27)
h region where endogenous respira-
on biomass is
k
0
SX9 _
K
m
+ S
^ = 0
(5-28)
an excess of biological solids t hat
steady-state operation. The quant i t y
,et growth of biomass in the reactor.
.-k
d
XV
(5-29)
9
m
=
XV
Q
w
X
r
+ Q
e
X, + k
d
XV
(5-32)
where: Q
w
flow rate of waste from recycle line
Q

= overflow rate from clarifier

X
e
= biomass concentration in overflow from clarifier
Generally the mean cell-residence time has not been widely adopt ed since
the value of k
d
is often not known for the influent wastewater. Instead, the
term sludge age has been adopt ed, which is defined as the ratio of biomass in
the reactor to the net rate of biomass generation:
vx
K
m
+ S
kpSXV _
k x y
~ k
0
S - k^K
m
+ S)
(5-33)
K
m
+ S
_ from the separator t hat produces
waste treatment plant, these solids
process.
ly used in the design and operat i on
ocess loading factor is the mass of
The processjoadjng factor is simply related to sludge age t hrough the kinetic
parameters Yand k
d
:
= YU - k
d
(5-34)
V 5 fy
M
116
Models for Biological Reactors Chap. 5 >>
Equations (5-30) and (5-33) can be solved for S to give
5
__ YUK
m
_ K
m
+ KJcA
* - k
Q
- YU ~ k
0
9
e
- kj9
c
- 1
P
"
j : > ;
If the kinetic constants are known, specifying any one of the quantities S,
U or 9
e
determines the other t wo.
At steady-state conditions, the net rate of biomass generation is equal to
the rat e at which biomass flows out of the system. If biomass is removed by
wasting from t he recycle line and by losses in the clarifier overflow, t he
sludge age is given by
$
'
=
Q.X!+Q.X.
(5
"
36)
Thus, sludgejige can be controlled by the rate of wasting of biomass. Since
the sludge age omits the endogenous term from the denominator, sludge age
will be greater t han mean solids residence time. However, the ease of applying
Eq. (5-36) to treatment pl ant operations encourages the use of sludge age.
The sludge age has an i mport ant effect on the settling characteristics of the
bi omass and will be discussed further in the next chapter.
5- 5 STABI LI TY WI T H RECYCLE SOLI DS
In designing a biological reactor, the concentration of biomass in" the
recycle stream, is needed. This' concentration depends upon the specific type
of separator and its operating characteristics. Thus, the reactor and separator
should be considered together in the design of a biological process with recycle.
To illustrate the effect of process variables on reactor performance, t wo
idealized models of the separator are often used. In one model, the separator
is assumed to give a constant ratio of out put to input solids concentration,
i.e.,
f} = -ji = constant
This type of behavior might be approximated by a sedimentation vessel. In
the other model, the separator is assumed to provide a recycle stream with
a constant biomass concentration, i.e., X
R
constant. A relatively constant
biomass concentration might be achieved with a centrifuge.
If the concentration ratio across the separator is assumed constant,
bi omass Eq. (5-28) becomes
^ ( / ? - l ) - l + ^ ^ _ = 0
or
KJl + R BR)
Sec. 5-5 Stability with Recycle Solids
The steady-state biomass concentration can then be calculated from Eq.
(5-25):
Y
_ ( S -S)(K
m
+ S)Y
k
0
Sd
(5-38)
At washout conditions, the biomass concentration drops to zero and no con-
version of substrate occurs in the reactor. The critical fresh-feed residence
time for washout is obtained by substituting S = S in Eq. (5-37):
0 w g
( * . + S)(l + R - PR)
( 5
.
3 9 )
According to this model of the separator, there is no conversion of substrate
for fresh-feed residence times equal to or less t han 6
W
.
If the concentration of biomass in the recycle stream is assumed to be a
constant, the solution for substrate concentration involves a quadratic equa-
tioji. For example, combining material balance equations (5-25) and (5-28)
gives
S ^ - 6 A / y - J g (5. 40)
where: a = 1 + R k
0
9
b = k
0
9 (s + + (K
m
- + R)
c = -SK
m
(l+R)
Once S is calculate^!. JTcan be found from Eq. (5-38). *
With constant recycle solids concentration, complete washout of the
reactor is not possible. The presence of cells in the entering stream ensures
that cells will also exist in the reactor. As long as some biological solids are
present in the reactor, some finite conversion of substrate will occur.
A comparison of reactor performance for the two separator models is
shown in Fig. 5- 4
2
. The values for the kinetic constants and recycle ratio are
fairly typical for activated sludge reactors. For the constant-ratio model, a
value of 4 was selected for the ratio of out put to input concentrations across
the separator. The effluent substrate and biomass concentrations from the
reactor (curves A) were calculated from Eqs. (5-37) and (5-38). For the con-
stant-concentration model, the recycle solids concentration was taken as
10,000 mg/C. In this case, Eqs. (5-38) and (5-40) were used to calculate the
substrate and biomass concentrations for the reactor (curves B). As pre-
dicted by Eq. (5-39) for the constant-ratio model (curves A), washout occurs
at a fresh-feed residence time of 0.58 hours. Since washout is not possible at
constant X
r
, the conversion of substrate is finite at all residence times (curves
Models for Biological Reactors Chap. 5
2500
r
2000 -
- 1500
1000
500
A
Biomass
/B
_
*
0
= 0.5 hr"
1
K
m
= 75 mg/
Y =0.6
_
\B \B
Substrate COD
i -4 1
Substrate COD
i -4
0.5 1.0 1.5
Residence time (hr)
2.0
FIGURE 5-4 Comparison of reactor performance for two separator
models.
2
Curves A at constant concentration of 10.000 malt. Curves
B at a constant concentration ratio of X
r
/X - 4.
B). A separator giving a high stable concentration of recycle solids is desirable
since reactor performance is improved during periods of high volumetric
flow rates.
\
5- 6 P L UG- F L OW REACTOR
In a plug-flow reactor, the velocity is constant at any given cross-section and
no mixing of fluid elements occurs longitudinally along the flow pat h. Thus,
all elements of fluid have the same residence time in the reactor. Since com-
position of the fluid varies from position to position along the reactor, the
material balances must be made on a differential element of fluid.
A schematic diagram of a plug-flow reactor with recycle is shown in
Fig. 5-5. For the differential element of volume dV, the steady-state material
balance is
I nput Out put + Format i on by reaction = 0
The terms for the substrate are
I nput = (2 + Q
r
)S = + R)S
Out put = (2 + Q
r
)(S + dS)= Q\\ + R)(S + dS)
Format i on by reaction ==
k
0
XS
Y(K
m
+ S)
dV
Sec. 5-6
Plug-Flow Reactor
119
q+ A .
s,x S.. X;
~Ws~7,
1
X + dX
^*-dZr
o
r
.sx
r
FIGURE 5-5 Schematic diagram of a plug-flow reactor.
The steady-state material balance on substrate is
k
0
XS
G(l + R) dS +
dV=0
Y(K
m
+ sy
Similarly, the steady-state material balance on biomass is
( 5^1)
+ R) dx - [Y^S ~
KDX
)
DV
=
(5_42)
The material balances can be expressed as a function of axial position, Z,
by the relation
dV == A dZ
where A is the cross-sectional area of the reactor
+R)dS+
Y{
zf+
S)
A
d z
= c
5
-
43
)
e(l +R)dX- ( +
s
- k
4
x}AdZ = 0 (5-44)
The boundary conditions for these equations are obtained by material
balances around the entrance where fresh feed mixes with recycle.
QS + Q
r
S, = (2 + Q
r
) S,
QX + Q
r
X, = (2 + Q
r
) X,
where S, and X, are the concentrations of substrate and biomass in t he total
fluid entering the reactor. After dividing by Q and rearranging, the boundary
conditions become
S -
5
+
R S
- at 7 - 0
l+R
a t Z
~
Z t = Z
l + R
K a t Z = 0
(5-45)
(5-46)
120
Models for Biological Reactors Chap. 5
Equat i ons (5-43) to (5-46) form a set of nonlinear differential equat i ons
t hat are difficult to solve analytically. For this general case, the equat i ons
are more readily solved by computer techniques. The equations can be
integrated directly, however, in certain special cases.
For example, if the amount of biomass formed by the reaction is small
relative to the amount entering the reactor, then the concentration of biomass
is nearly constant along the length of the reactor. Denoting the average con-
centration of biomass in the reactor as X
ay
the substrate material balance is
+R)dS + y^MA dZ = 0 (5-47)
Int egrat i ng with X
a
as a constant gives
( S , - * ) + * In f =
Y
*!*-f
R)
(5-48)
The small change in biomass along the reactor is approximated by
(X - X,) = Y(S, - S) - Q
0
ff'+
Z
R)
Substituting for (S
{
S) from Eq. (5-48) gives
x < * - X.) = ^ -
}
- K
m
Y I n - (5-49)
The validity of the assumption of nearly constant biomass concentration
can now be checked by compari ng the inlet concentration with the outlet
concent rat i on predicted by Eq. (5-49).
Plug-flow reactors are generally operated with recycle of microorganisms.
In t he absence of recycle, the fresh feed is the only source of biomass. Since
the concent rat i on of biomass in fresh feed is often very low, the reaction rate
at the ent rance of the reactor would also be low. Wi t hout longitudinal mixing
in an ideal plug-flow reactor, there is no feedback mechanism for biomass.
Thus, the reaction rate would remain low for a substantial length of the
reactor. By recycle of biomass, the reaction rate is increased and the length
of reactor for a given conversion is decreased.
As in the case of a well-mixed reactor, washout cannot occur if there is
a fixed concent rat i on of biomass in the stream entering the reactor. Washout
is possible, however, if the fresh feed contains no biomass and if the separator
produces a const ant ratio of out put to input solids concentration (j5). In this
case, t hen, t he biomass concentration in the recycle approaches zero as the
Sec. 5-6 Plug-Flow Reactor 121
biomass concent rat i on leaving the reactor goes to zero. The critical fresh-feed
residence time at washout for constant /? is given by
3
e
=
(i+rxv + kj
l n
i R
( 5
.
5 0 )
At steady state, t here is no conversion of substrate for fresh-feed residence
times equal to or less t han 6
W
. Wi t h no recycle (R = 0), 6
W
is infinite and t he
reactor washes out for all residence times.
EXAMPLE 5-2
Fresh feed enters an activated sludge plant at a flow rate of 0.088 jn
3
/sec
(2 mgd) and with a substrate concentration of 300 mg/C. Sludge is recycled to
the reactor from a separator at a flow rate of 0.013 m
3
/sec (0.3 mgd) and a
biomass concentration of 6000 mg/(. If the conversion of substrate is 95 %,
determine the residence time and volume for:
(a) a well-mixed reactor
(b) a plug-flow reactor
The kinetic constants are: k
Q
= 0.4 hr "
1
; K
m
= 75 mg/; Y = 0.6; k
d
~ 0
SOLUTION:
(a) For a well-mixed reactor, the substrate balance is
\
The biomass balance (A'
0
= 0) is
RX
r
- (1 -f- R)X + | = 0 (5-28)
R = 0.3/2 = 0.15; X
r
= 6000 mg/; S = 300 mg/t; S == 0.05(300) = 15
mg/^. Substituting in Eqs. (5-25) and (5-28),
300 - 15 -
0 A
^
x d
- o
3 0 0 1 5
0.6(75 - f 15) ~
0
0.15(6000) - (1 + 0. 15)* + 7 5
( 1
^
g
= 0
X = 930 mg/e
Q = 2.76 hr
V = 0QO
=
2.76 X 0.088 X 3600
= 874 m
3
(2.3 x 10
5
gal)
Solving,
122
Models for Biological Reactors Chap. 5
(b) For a plug-flow reactor, the substrate balance is
G(l +R)dS+
y{
K*+
S
)
DV
=
(5_43)
The biomass balance is
ed + R)dX- K
m
X
+
S
s
dV = 0 (5-44)
At V = 0,
X / = = X
1
( 5 _ 4 6 )
Substituting in the above equations:
o 4
(0.088 x 3600X1 + 0.15) dS +
Q 6
(
7
5 - j . 5) ^
F
=
0
(0.088 x 3600X1 + 0.15) dX - r/K = 0
^ -
3
^
+
U 5
5 ( 1 5 ) = = 2 6 3 m
^
Since the amount o{ biomass formed in the reactor is fairly large relative to
the biomass entering, the equations were solved numerically. Small increments
in Kwere assumed and 5 and A"were determined. The required reactor volume
was found at S = 15 mg/.
X = 930 mg/{
V = 288 m
3
(76,000 gal)
9 = 0.91 hr
Equations (5-48) and (5-49) will give a reasonable approximation of the
results.
The plug-flow reactor is about \ the size of the well-mixed reactor.
5- 7 FLOW REACTOR WI T H AXI AL MI X I NG
Plug flow represents an idealized flow pattern in which all fluid elements have
identical residence times. In actual reactors, some degree of mixing occurs in
the axial direction of the reactor. For example, activated sludge reactors are
often designed as long tanks with aeration of the liquid along the l engt hof
Chap. 5 Sec. 5-7 Flow Reactor with Axial Mixing
123
(5-43)
(5-44)
(5-45)
(5-46)
5 dV = 0
git

tit-
or is fairly large relative to
merically. Small increments
"he required reactor volume
ible approximation of the
ie well-mixed reactor.
I NG
which all fluid elements have
degree of mixing occurs in
activated sludge reactors ar e
he liquid along the length-of
the t ank. The turbulence created by this aeration process causes fluid mixing in
the flow direction.
The plug-flow model can be modified to account for axial mixing by adding
a dispersion term. In a dispersion-flow model, the rate of axial mixing is
assumed to be proport i onal to the concentration gradient of the diffusing
component in the reactor. For the differential element of Fig. 5-5, the dif-
fusion terms for substrate ar e:
I nput by diffusion = D
X
A
dS
dZ
Out put b y diffusion = ~[d,A j | + ^(p
x
A ^ dZ^
where D
g
is the eddy diffusivity or axial dispersion coefficient for axial
mixing with units of l engt h
2
/ t i me. If the eddy diffusivity is assumed constant,
the net rate of longitudinal dispersion is given by
Net diffusion = -D
X
A ^ dZ
By addi ng this term to the plug-flow material balance, Eq. (5-43) becomes
Q%l+R)dS-D
2
A^
2
dZ +
k
0
XS
dZ*~ ' Y(K
m
+ S)
A dZ = 0 (5-51)
In a similar manner, when an axial dispersion term is added to Eq. (5-44),
the material balance for biomass becomes
e ( l + R) dX - D
Z
A ^ d Z - ( ^ p ^ - k
d
x) AdZ = 0 (5-52)
Several different boundary conditions have been suggested for this type
of second-order equation. The boundary conditions proposed by Danckwerts
4
are used frequently. At the entrance where Z = 0,
(2 + Qr)S
t
= (2 + Q
r
)S,.o -
(2 + Q
r
)x, = (2 + Q
r
)x,-o - D,(4g) A
(5-53)
At the exit where Z = L,
(5-54)
124
Models for Biological Reactors Chap. 5 ^
If Eq. (5-51) and (5-52) are multiplied by t he length of the react or L
9
they can be rearranged to give
dS_-e
k

X S
= 0 (5-55)
\uL)d(ZILy d(Z(L) Y(K
m
+ S)
V K

D > }
0(l 4- R)
where: u = ^
v
. mean fluid velocity
6 = = mean residence time
u
. *
The combination uL/D
2
is a dimensionless group known as the Peclet number,
which can be used to characterize the degree of mixing in the vessel. For a
plug-flow reactor with no axial mixing, the eddy diffusivity is zero so t hat
t he^ecl et number is infinite. The diffusion terms can then be dropped from the
material balances. Thus, a plug-flow reactor is approximated by the dis-
persion model with a very large Peclet number. In a well-mixed reactor, the
eddy diffusivity is infinite and the Peclet number becomes zero. Therefore,
the dispersion model with a very low Peclet number approximates t he con-
ditions in a well-mixed reactor.
Because of the complexity of the nonlinear differential equations, an
analytical solution is not available even for the case of constant biomass
concentration in th^ reactor. The equat i ons are best solved by numerical
met hods implemented on a digital comput er.
5- 8 WE L L - MI X E D REACTORS I N SERI ES
A sequence of well-mixed reactors can give higher conversion t han a single
well-mixed reactor with the same total volume. Also, the model for well-
mixed reactors in series is useful in simulating the performance of certain
designs of activated sludge reactors.
A schematic diagram of N well-mixed reactors in series is shown in Fig.
5-6. The concentrated biomass from the separator is recycled to the first
reactor. For the /?th reactor in the sequence, the material balances on sub-
strate and biomass are given by
v
n
^ = e(i + - e(i + R)s
a
- Y{f
nX
+s)
(5
"
57)
= eo + ~ eo + w + g g f f i ^ -
k
^v
n
* (5-58)
Reactors Chap. 5
y the length of the reactor L,
- Q k
a
XS
0
(5-55)
= (5-56)
locity
roup known as the Peclet number,
ee of mixing in the vessel. For a
e eddy diffusivity is zero so t hat
*ms can then be dropped from t he
tor is approximated by the dis-
iber. In a well-mixed reactor, t he
lumber becomes zero. Therefore,
et number approximates the con-
llinear differential equations, an
for the case of constant bi omass
>ns are best solved by numerical
er.
ERIES
re higher conversion t han a single
olume. Also, the model for well-
lating the performance of certain
reactors in series is shown in Fig.
separator is recycled to the first
ice, the material balances on sub-
126
Models for Biological Reactors Chap. 5 *
Defining the residence time on the basis of fresh-feed flow rate, the material
balances become
-dT--ffr
{S
-
1 S
">-Y(K
m
+ S
n
)
( 5
"
5 9 )
3 V^ii-l ^n/ "T "P j FT
K
d
A
n
dX,_
(5-60)
where: <?=g5
At steady-state conditions, the material balances are
- *
+
. + c f + o - <t t % -
0 ( 5
-
6 2 )
The i nput concentrations of substrate and biomass to the first reactor are
given by material balance equations (5-45) and (5-46).
v
_ x +
Washout of the reactors can occur if no biomass enters in the fresh feed
(X = 0) and if the clarifier operates with a constant ratio of output to input
biomass concentration (fi = X
r
/X
N
). For N equal-sized well-mixed reactors
in series, the critical residence time for washout of the nth reactor is
3
V*c TTco (5-63)
K
m
+ 5
K d
For a given overall conversion of substrate, the total volume of the
reactor system decreases as the number of reactors in series is increased.
Thus, several small reactors in series give better performance t han one large
reactor with the same total volume. Since the cost of several small reactors is
greater t han the cost of one large reactor, there is usually an economic opti-
mum for the number of well-mixed reactors in series.
Sec. 5-8
WeII-Mixed Reactors in Series
127
In the limiting case where the number of reactors in series becomes infi-
nite, the conversion in each reactor is infinitesimal and the performance
characteristics of the reactor system are identical to those for a plug-flow
reactor. Thus, a finite number of well-mixed reactors in series can be used to
simulate intermediate degrees of mixing between a single well-stirred reactor
and a plug-flow reactor.
EXAMPLE 5-3
Determine the residence time and reactor volume if the feed of Example
5-2 is treated in 3 equal-sized well-mixed reactors in series.
SOLUTION
Equations (5-61) and (5-62) with X
0
= 0 and k
d
0 can be applied to
each reactor in series.
Reactor 1:
/c c* koS\X\0\
n
^
i , }
Y(K
m
+ STxi + r) -
0
rxr v \ _i kpS\X\6\
A
Reactor 2:
/ R * C \ k
o
S
2
X
2
0
2
_
n
Ki>
\ ^ Y(.K
m
+S
1
){\+R)-
0
(Y koS
2
X
2
0
2
_
n
x {X, x
2
) +
{ K m
-
S i ) { x + R )
- o
Reactor 3:
/ c c\ kpS-jXiB-j N
(v V
,
* _I _ kySiXiOi
n
(Xz - +
{ K m
+ SMI + R) ~
Input concentrations to first reactor:
S a
= 1J = 300 +0- 1*13)
= 2 6 3 m g J (
For equal volumes, 0
X
= 0
2
= # 3 . Since S
3
is specified as 15 mg/, the
unknowns are S
u
X
u
S
2t
X
2
, X
2f
and 0. The six simultaneous equations can
be solved algebraically but the expressions become very complicated.
Instead, the equations were solved by trial and error. Values for 0 were
128
Models for Biological Reactors Chap. 5
v
assumed until .S
3
= 15 mg/. The results were:
0 , = 0
2
= 0
2
= 0.42 hr
0 t o u i = 3 x 0.42 - 1.26 hr
V
x
= V
2
= K
3
= 132 m
3
(35,000 gal)
Ktoui = 396 m
3
(105,000 gal)
As expected, the 3 well-stirred reactors in series are intermediate between
plug-flow and well-stirred.
5- 9 C O MP A R I S O N OF REACT ORS
The well-mixed and plug-flow- models represent the two extremes in flow
behavior for a biological reactor. The well-mixed model assumes t hat the
influent stream is dispersed instantaneously and uniformly t hroughout the
contents of t he vessel. The plug-flow model assumes that no mixing occurs
along t he flow pat h and t hat all fluid elements have identical residence
times.
For a well-mixed reactor, the substrate concentration in the vessel is the
same as the substrate concentration in the effluent. Thus, the fresh feed is
immediately dispersed into an environment of lower concentration. In a
plug-flow reactor, the substrate concentration decreases continuously along
the length of the\vessel. If other conditions are t he same, a higher substrate
concentration gives a higher rate of reaction. As a result, a plug-flow reactor
generally produces a higher conversion of substrate in a given volume t han
a well-mixed reactor.
When the biomass concentration entering the reactor with fresh feed or
recycle is very low, the plug-flow reactor may require a larger volume than
a well-mixed reactor. If little biomass is present in the influent to a plug-flow
reactor, the rate of reaction is very low in the entrance region but gradually
increases as biomass is generated by the reaction. If no biomass enters a
plug-flow reactor, no biological reaction can occur and the reactor washes
out. On the other hand, the influent to a well-stirred reactor is mixed with
vessel fluid containing biomass so that the reaction can be sustained even in
the absence of biomass in the feed stream.
The well-mixed reactor is generally more stable than a plug-flow reactor
in response to toxic and shock loadings. For example, if a concentrated pulse
of a toxic substance enters a plug-flow reactor, the concentration remains high
as it moves along the reactor and into the clarifier. Because of the high con-
centration, the toxic substance may destroy an appreciable quantity of the
biomass in the system and cause a long-term upset in reactor performance.
Sec. 5-10 Fixed-Film Reactors
129
With a well-mixed reactor, the pulse of toxic material is dispersed rapidly
t hroughout t he vessel and its concentration level is reduced. The metabolic
processes of t he microorganisms may be only slightly affected by the toxic
substance at the lower concentration level. In general, a well-mixed reactor
gives a more uniform effluent under varying feed conditions.
Anot her advant age of a well-mixed reactor is the uniformity of oxygen
consumption rates by the microorganisms. Since concentrations and reaction
rates are constant t hroughout the vessel, the oxygen upt ake rates are also the
same everywhere in the reactor. Wi t h a plug-flow reactor, the oxygen demand
is usually greatest in the inlet region where the substrate concentration is
high. At the exit where the substrate concentration is low, the oxygen con-
sumption rate is also generally low. Thus, the aerators for a plug-flow reactor
should be designed to provide more oxygen transfer in the inlet region.
The well-mixed and plug-flow reactors are idealized models that are
difficult to achieve in large-scale biological reactors. In actual mixed reactors,
short-circuiting of fluid and stagnant zones may occur. Some portion of the
fluid may flow from the inlet to the outlet with a short residence time because
of incomplete mixing with the bulk of the reactor fluid. Also, some regions
of the reactor may be relatively stagnant so that fluid elements entering these
zones have long residence times.
In axial-flow reactors, aeration of the fluid causes longitudinal mixing
and a distribution of residence times. Thus, long biological reactors with
aeration are often better simulated by an axial dispersion model or a tanks
in series model. If aeration is fairly uniform along the length of the reactor,
the axialMispersion model may approxi mat e the hydraulic regime. If discrete
aerators give several local zones of intense mixing along the reactor, the
model for well-mixed tanks in series may be preferable. Tracer techniques
are useful in establishing an appropri at e hydraulic model for a biological
reactor.
5- 10 F I XED- F I L M REACT ORS
Fixed-film reactors consist of a solid surface with an attached layer of
biomass. Substrate and oxygen diffuse into the layer and react to form pro-
ducts and addi t i onal biomass. The most common types of fixed-film reactors
are the trickling filter and the rotating biological contactor.
A trickling filter is a bed packed with rocks or plastic structures. Waste-
water is distributed over the top of the packing and allowed to trickle through
the bed. Recycle of a port i on of the effluent liquid is often practiced. A
rotating biological contactor consists of many plastic discs attached to a
central drive shaft. The discs are parallel to each other with intermediate
spacing to permit movement of fluid between them. The discs are partially
130
Models for Biological Reactors Chap. 5 \
submerged in the liquid so that they are exposed alternately to substrate in
the wastewater and oxygen in the air during rot at i on.
Mi croorgani sms grow on the surface of the packing or discs and meta-
bolize t he organic substrate from the wastewater. As the microorganisms
multiply, the thickness of the biological layer increases. At some thickness,
the diffusing oxygen is consumed by reaction before it penetrates t hrough
the entire layer. As a result, bacteria near the solid surface are in an anaerobi c
envi ronment (Fig. 5-7). In addition, most of the substrate is consumed before
reaching t he packing so that anaerobic bacteria adjacent to the surface enter
the endogenous phase of growth. These bacteria have less ability to adhere
to the solid support and are washed off by shear forces. A new layer of
microorganisms then starts to grow at this location in the bed.
Solid
packing
Biological
film
Liquid
layer
Organics
Oxygen
Air
Products
Anaerobic Aerobic
Flow
or
FIGURE 5-7 Schematic diagram of phases in a trickling filter.
Most of the biological conversion occurs in the aerobic portion of the
film, the depth of which depends upon the rate of reaction, the diffusivity of
oxygen and the liquid flow rate. Oxygen is supplied to the film by diffusion
t hrough the liquid layer. The oxygen in the liquid either enters with the
wastewater or is transferred from the air. Concentration gradients of bot h
oxygen and substrate exist in the liquid and biomass layers. Under steady-
state conditions, the* mass fluxes of substrate and oxygen across the liquid
biomass interface equal the corresponding rates of substrate and oxygen
consumpt i on by reaction in the biological layer.
The design of fixed-film reactors from fundamental theory is difficult
because of the multiple phases involved and the flow patterns in the units.
Differential equations can be written to describe the processes occurring in
Sec. 5-11
TrickUng-Filter Models
131
the liquid and biomass regions. If the velocity profiles, reaction kinetics,
diffusivitics, and boundary conditions are known, these differential equations
can be solved for substrate conversion.' Since the equations are difficult to
solve and conditions are often not known accurately, assumptions are usually
made to provide working models.
5- 11 TRI CKLI NG- FI LTER MOD E L S
Many theoretical and empirical models have been proposed to describe the
performance of trickling filters
3
-
6
. A simplified reaction rate model can be
developed by considering each cross-section as a pseudo-homogeneous reactor
with no gradient in substrate across the be d
7
. The reaction rate within the
film is assumed to be reaction-rate controlled and to proceed at a rate
determined by the substrate concentration in the liquid at that cross-section.
For t he differential element, dZ, in Fig. 5-8, a material balance on substrate
gives for g= Q(l + R),
or
QS - Q(S + dS) + r
s
V
b
= 0
-QdS + r
s
V
b
=0
(5-64)
Q
\
RQ
(1 + R)CF\ S,
S + dS
z=o
dZ
z = z
a
Q
FIGURE 5-8 Schematic diagram of trickling
filter for development of models.
132
Models for Biological Reactors Chap. 5
If the reaction follows Monod kinetics, the rate is given by
- k
0
X
b
S
s
- Y(K
m
+ S)
The vol ume of active biomass where reaction occurs is
V
b
= aSA dZ (5-65)
where: a = area of packing per unit volume of reactor
5 = thickness of active film
A = cross-sectional area.
Substituting in t he material balance, Eq. (5-64) becomes
QdS+ Y^
a
+
%
d Z = 0
<
5
"
66
>
At substrate concentrations above about 500 mgj( COD, the thickness of
active film is essentially c ons t a nt
8
.
Integration from S
t
at Z = 0 to S
e
at Z = Z
t
yields
K
m
In ^ + (S
t
- S,) =
k

a
X>
Z
< (5-67)
To simplify this equation for application, t he quantities k
0y
Y, X
by
and 8
can be combined into a single paramet er
P
=
hA* (
5
_
68)
Eq. (5-67) can therefore be written as
In
=
S
*
S
* _ PaAZ
e
(5-69)
l n
S
t
K
m
(QK
m
)
With a given packed bed and flow rate, the quantities K
m
and P should be
constants. For this model, Eq. (5-69) shows that a plot of In (S
e
/S
t
) versus
S
t
S
e
should give a straight line with a slope of \/K
m
and an intercept of
PaAZJQK
m
at SJS
t
'= 1. Thus, t he paramet ers P and K
m
can be evaluated
from l aborat ory or pilot plant dat a in which S
t
is varied at a fixed flow rate.
P and K
m
are not true physical constants but depend upon liquid flow rate.
With the paramet ers P and K
m
determined for a given packing, Eq. 5-69
can be used to predict effluent conditions for various liquid flow rates, bed
depths and influent concentrations.
Sec. 5-11 TrickUng-Filter Models 133
At substrate concentrations below about 300 mg!C COD, the thickness
of the active film increases almost linearly with substrate concentration
8
,
S =SS
where 5 is a proport i onal i t y const ant . In this case, the material balance
equat i on becomes
Q dS +
k
f(f
2
l
S
s) dZ = 0 (5-70)
The integrated form of Eq. (5-70) is
"f c- Ms K) - ^ <
5
-
7,)
where: P' =
k

5
*
Xb
(5-72)
The parameters P' and K
m
in this model should be evaluated from a plot of
ln(SjS
t
) versus (l/S
e
- l/S
t
).
The rate of substrate removal may also be determined by diffusional
effects in the liquid phase or in t he biomass film. If the rate is controlled by
mass transfer of substrate at the liquid-biomass interface, a material balance
on a differential section of packing is
\
QdS= -k
L
aA(S - S
f
) dZ (5-73)
where k
L
is a liquid-phase mass transfer coefficient and S
f
is the concentra-
tion of substrate at the interface between liquid and biomass film. Since S
f
approaches zero for mass transfer control, Eq. (5-73) can be integrated to
give
l n | = - ^ f (5-74)
For turbulent liquid flow, k
L
is proport i onal to Q/A:
where m is usually between 0.5 and 0.7. The mass transfer model can then be
expressed as
where n is generally between 0.3 and 0.5.
134
Models for Biological Reactors Chap. 5
Many empirical models have been developed for the design of trickling
filters. As a particular example, Eckenfelder
9
correlated dat a by the following
expression
where: Q
a
= liquid flow rate per unit area (Q/A)
Z = bed depth
K and n = constants for the specific packing
The form of this empirical model is the same as the mass transfer model
given by Eq.-(5-75).
When effluent liquid is recycled, substitution of S
t
from Eq. (5-45) gives
The constants K and n can be determined by operating a trickling filter at
several different liquid loadings and measuring substrate concentrations at
several depths. With no recycle, Eq. (5-76) suggests t hat a plot of\nSJS
t
vs. Z
should give a straight line for each liquid loading. Since the absolute values
of the slopes of these lines are K/Q", a. plot of In [slope] versus In Q
a
should
also yield a straight line with slope n and intercept AT as illustrated in Example
Values of K and n for a variety of packings are tabulated in Chapter 6.
Wi t h known values for these constants, Eq. (5-76) and (5-77) can be used to
estimate the performance of a trickling filter.
EXAMPLE 5-4
A trickling filter with a depth of 8 ft is operating with a hydraulic loading
of 0.18 gpm/ft
2
(0.12 /m
2
-sec), a recycle ratio of 0.7 and an influent substrate
concentration of 150 mg/. Based on the following laboratory data, determine
the effluent substrate concentration from the trickling filter.
The laboratory unit has 10-ft depth with sampling taps at 2.5-ft intervals.
Substrate concentrations were measured for hydraulic loadings of 0.1,0.2, and
0.3 gpm/ft
2
(0.068, 0.136, 0.204 /m
2
-sec) with no recycle.
(5-76)
(5-77)
5-4.
In [slope] = In K n In Q
a
Fraction substrate remaining (S
e
/S
t
)
Depth (ft) >, - 0.1 >, - 0.2 0.3
2.5
5
7.5
10
0.61
0.37
0.23
0.14
0.70
0.49
0.34
0.24
0.74
0.55
0.41
0.30
Sec. 5-11
TrickUng-Filter Models
1
SOLUTION:
Eckenfelder's trickling-filter model will be used:
Plot log (SjSt) vs. Z for each Q
a
as shown in Fig. 5-9. The slope of each
line, -K/Qz, is equal to A(ln SJS,)/AZ.
1.0
0.5
CO
0.2
0.1
10 2 4 6 . 5
Z( ft )
FIGURE 5-9 Fraction of substrate remaining vs. 6*pth in trickling filter.
Plot log.(AT/22) vs. log Q
a
as shown in Fig. 5-10. The slope is n and
intercept (at Q
a
= 1) is log K.
n = 0.45
K = 0.07
For the trickling filter with recycle,
S + RS
t
^ 150 4- 0. 75,
A |
1 + 1 +0 . 7
. S
e
KZ
1 n
5,(1.7) 0.07(8)
150 + 0.75, (0. 18)
0
-
4 5
.S, = 30 mg/Y
136
Models for Biological Reactors Chap. 5 \
0.3 -
0.051 i | |
0.1 0.2 0.5 1.0
FIGURE 5-10 Evaluation of constants n and K for
Example 5-4.
5- 12 ROT AT I NG BI OLOGI CAL CONT ACT OR
The fixed-film model for reaction-rate control can be readily adapted to
rot at i ng biological discs. Assuming t hat the liquid phase in contact with the
discs is well-mixed and t hat no mass transfer limitations exist, a material
balance on the substrate is
QS
t
- QS

+ r,V
b
= 0 (5-78)
The vol ume of active biomass on t he discs is
V
b
=aV5 (5-79)
where: a = biofilm surface area per uni t volume of liquid
V = volume of liquid.
For a reaction obeying Monod kinetics, the material balance becomes
QS, - QS, -
Y
^ * f
s
f V 5 = 0 (5-80)
Substituting the paramet er P = k
0
5X
b
f Y in Eq. (5-80) gives
| - , 'Pad <
5
-
8l
>
^ K
m
+ S.
where: 0 V/Q, the liquid phase residence time.
.Sec. 5-13 River Analysis
If diffusion of substrate from the liquid to the liquid-biofilm interface
limits the rate, the material balance is given by
By introducing liquid residence time and letting S
f
at the interface approach
zero, Eq. (5-82) becomes
For N stages in series with the same liquid residence time for each stage, the
substrate removal is
If additional substrate removal occurs by suspended solids in the liquid,
anot her consumption term should be included in the material balance. In
most cases, however, removal by attached biomass is much greater than by
suspended biomass.
Since LaMot t a
9
has found t hat external diffusion is rate-limiting for
liquid velocities below about 0.8 m/sec, rates in many rot at i ng biological
contactors may be controlled by mass transfer.
5- 13 RI VER ANAL YSI S
The effluent from a wastewater treatment plant is usually discharged into a
body of water so t hat remaining impurities will be diluted below harmful
concentrations. The remaining carbon and nutrient compounds in the effluent
will be consumed by the microorganisms in the river and will exert an oxygen
demand upon the river. The amount of self-purification t hat can occur in the
river will depend upon the river flow rate, its oxygen content and its reaera-
tion capacity. If the BOD of a wastewater is above the reaeration capacity
of the river, the oxygen level will dr op as the water flows downst ream from
the wastewater discharge site. Eventually the oxygen content of the stream
may be reduced far enough to kill fish and other marine life. When the reaera-
tion capacity again exceeds the microorganism demand rate, the oxygen
level in the stream will rise and approach its initial level. This "oxygen sag
curve" in the river is shown in Fig. 5-11.
The river may be analyzed as a plug-flow reactor with first order kinetics
since the waste concentrations are usually less t han the K
m
o? the waste
substrate.
1 1 , 1 2
. If an oxygen balance is made on a differential distance of
QS
i
-QS

-k
L
a{S,-S
f
)V = Q (5-82)
(5-84)
1 3 8
Models for Biological Reactors Chap. 5
Distance downstream from point of discharge, Z
FIGURE 5-11 Oxygen sag curve in rivers.
the river, we arrive at
k
B
SX d
dt YK.
+ P,-R,-B, (5-85)
where:
h
u
Pr
K
B,
S
mass transfer coefficient for surface reaeration and pre-
dicted by Problem 3-3.
depth of the river
river velocity
rate of photosynthesis
rate of algae respiration
rate of river bot t om respiration
: waste concentration in river
Maki ng a balance on the waste in the same differential section, we have
dS dS k
0
SX
8
, S
^_
=
_ ^ _ _ (5-86)
Chap. 5
Problems
For steady state conditions, Equation (5-86) becomes
S/S
0
= e-
k(
*
/u)
(5-87)
where: S = BOD at some point z downstream
S
0
= BOD at point of discharge after mixing with the river
k X
k = ^ = modified rate constant
Therefore, introducing Eq. (5-87) into (5-85), neglecting P
ry
R
r
and B and
solving for steady state, the oxygen concentration downstream becomes
(C* - C) = (C* - C
Q
)e-
k
^
/hu
+
k
k
^
h
k h
[ e -
k l / u
- e~
k
^
lhu
\ (5-88)
Equation (5-88) is the Streeter-Phelps equation used in river analysis. The
minimum or critical oxygen concentration, C
c>
occurs at some distance, z
c
,
from the discharge point where the rate of oxygen utilization equals the
rate of reaeration.
(C* - C
c
) == M**-*^" (5-89)
Differentiating Eq. (5-88) and setting it equal to zero, we have for the critical
distance,
, - - ^ - & [ ' -
< c ,
- % g ' -
w
i <
5
-'>
Distances for z
c
are commonly several kilometers downstream from the point
of wastewater discharge into the river. Levels of the critical oxygen con-
centration should be maintained above about 4mg/t to avoid harmful
effects on desirable aquatic species.
P R O B L E M S
5-1. A well-mixed activated sludge reactor is operated with a flow rate of 1000 gpm
(63 /sec), an influent substrate concentration of 400 mg COD/( and an effluent
substrate concentration of 60 mg COD/. The Monod kinetic constants for the
wastes are:
k
Q
= 0.4 hr"
1
K
m
= 100 mg/
k
d
= 0.01 hr"
1
Y = 0.5 mg/mg COD
140 Models for Biological Reactors Chap. 5
Calculate the following quantities:
(a) Effluent sludge concentration
(b) Reactor volume
(c) Residence time in hr
(d) Total sludge in reactor
(e) Loading in kg COD/day/kg solids
5-2. A well-mixed activated sludge reactor using recycle is operated with a flow rate
of 1 mgd (0.044 m
3
/sec) and a substrate feed concentration of 500 mg COD//.
The recycle ratio is 0.5 and the recycle sludge concentration is 6000 mg//. The
kinetic constants are the same as in Problem 5-1. If the residence time (based
on fresh feed) is 3 hr, calculate the following quantities:
(a) Conversion of substrate
(b) Effluent sludge concentration
(c) Sludge age in days
(d) Sludge wastage per day
(e) Loading in kg COD/day/kg solids
5-3. Fresh feed enters a plug-flow activated sludge reactor at a flow rate of 2 mgd
(0.088 m
3
/sec) and with a substrate concentration of 300 mg//. Sludge is
recycled from a thickener to the entrance of the reactor with a biomass con-
centration of 6000 mg//. The volume of the plug-flow reactor is 40,000 gal
(151 m
3
) . Determine the recycle ratio needed to give a substrate conversion of
95%. The kinetic constants for the reaction are:
A-o = 0.3 hr"
1
Y = 0.5 mg/mg COD
tf
m
= 75mg/ / k
d
~0
\
5-4. A waste stream enters an 800,000 gal (3030 m
3
) well-mixed activated sludge
reactor at a flow rate of 5 mgd (0.22 m
3
/sec) and a waste concentration of
200 mg/ / . Recycle sludge from a separator also enters at a flow rate of 1.5 mgd
(0.066 m
3
/sec) and a biomass concentration of 8000 mg//. To increase the
capacity of the plant, a second well-mixed reactor of 400,000 gal (1515 m
3
) is
placed in series with the first. Should this reactor be placed before or after the
original reactor ? What increase in waste flow rate can be handled if the conver-
sion in the series reactor is the same as in the original single reactor? The
kinetic constants for the reaction are the same as in Problem 5-3.
5-5. A waste stream containing 500 mg/ / of substrate is fed to a well-mixed
activated sludge reactor operating to produce the maximum net growth rate of
biomass (mass of biomass formed per unit time and volume). Determine the
maximum cellular growth rate based on the following kinetic constants.
k
0
=0.3hx-
1
Y= 0.6
K
m
= 75 mg/ / k
d
= 0.05 h r
1
5-6. A wastewater containing 300 mg/ / is treated in two trickling filters in series to
give an effluent concentration of 20 mg/ / . Clarified liquid from each filter is
Chap. 5
References
eactors
Chap. 5 1
141
recycle is operated with a flow rate
concentration of 500 mg COD/ / ,
e concentration is 6000 mg/ / . The
n 5-1. If the residence time (based
g quantities:
dge reactor at a flow rate of 2 mgd
icentration of 300 mg/ / . Sludge is
of the reactor with a biomass con-
the plug-flow reactor is 40,000 gal
ed to give a substrate conversion of
n are:
0.5 mg/mg COD
0
30 m
3
) well-mixed activated sludge
sec) and a waste concentration of
also enters at a flow rate of 1.5 mgd
ion of 8000 mg//. To increase the
reactor of 400,000 gal (1515 m
3
) is
reactor be placed before or after the
ow rate can be handled if the conver-
in the original single reactor? The
same as in Problem 5-3.
substrate is fed to a well-mixed
uce the maximum net growth rate of
it time and volume). Determine the
the following kinetic constants.
y = o.6
k
d
= 0.05 h r
1
sated in two trickling filters in series to
>//. Clarified liquid from each filter is
recycled to that filter at a recycle ratio of 0.5. Each bed is packed to a depth of
6 ft with rocks of about 2.5 in. diameter. Find the area of the trickling filters
needed to handle a flow rate of 2 mgd (0.088 m
3
/sec).
The constants for the Eckenfelder trickling filter model are:
K = 0.645, n = 3.80, a = 27.6 ft
2
/ft
3
, Q
a
in mgd/acre
5-7. The wastewater from a city is discharged into a river flowing at 10 m/min,
with a mean depth of 2.5 meters, and with a river bed drop of 1 meter/1000
meters.
The river is completely mixed in the vertical direction and has a waste
concentration after the wastewater injection of 150 mg BOD/ / and a dissolved
oxygen concentration of 8.0 ppm. The constants for the removal of the waste
are,
k
0
- 0.5 h r
1
, K
m
= 100 mg BOD/ / , X = constant = 20
mg
^
L S S
Y - 0 6
m g M L S S
r* o -)
Y
~ *
6
me BOH
C
=
9 2
PPm
mg BOD
If the only oxygen source is reaeration and the only sink is the biological
oxidation of carbonaceous matter, estimate the minimum dissolved oxygen
concentration along the river. How far along the river after the wastewater
injection will the minimum oxygen concentration occur?
R E F E R E N C E S
1. LAWRENCE, A. W., and P. L. MCCARTY, "Unified Basis for Biological Treat-
ment Design and Operation," /. San. Eng. Div., ASCE, 96 (SA3), 757 (1970).
2. RAMANATHAN, M., and A. F. GAUDY, "Steady State Model for Activated Sludge
with Constant Recycle Concentration," Biotech, and Bioeng., 13, 125 (1971).
3. FAN, L. T., et al, "Effect of Mixing on the Washout and Steady State Perfor-
mance of Continuous Cultures," Biotech, and Bioeng., 12, 1019 (1970).
4. DANCKWERTS, P. V., "Continuous Flow Systems,".Chem. Eng. Sci., 2, 1 (1953).
5. CLARK, J. W., W. VIESSMAN, and M. J. HAMMER. Water Supply and Pollution
Control, 3rd. ed., Ch. 11. International Textbook, 1977.
6. ROBERTS, J. , "Towards a Better Understanding of High Rate Biological Film
Flow Reactor Theory," Water Res., 7, 1561 (1973).
7. KORNEGAY, B. H., and J. F. ANDREWS, "Application of the Continuous Culture
Theory to the Trickling Filter Process," Proc. 24th Industrial Waste Conf.,
Purdue, 1398 (1969).
142 Models for Biological Reactors Chap. 5 \
8. HARRIS, N . P. , and G. S. HANSFORD, " A Study of Substrate Removal in a
Microbial Film Reactor," Water Res., 10, 935 (1976).
9. ECKENFELDER, W. W., Industrial Water Pollution Control, Ch. 13, McGraw-Hill,
1966.
10. LAMOTTA, E. J. , "External Mass Transfer in a Biological Film Reactor,"
Biotech, and Bioeng., 18, 1359 (1976).
11. METCALF and EDDY, Wastewater Engineering, Ch. 15, McGraw-Hill, 1972.
12. THOMANN, R. V., Systems Analysis and Water Quality Management, Ch. 5,
McGraw-Hill, 1972.
\