design and modeling of post-denitrification single-sludge activated sludge processes
TRANSCRIPT
DESIGN AND MODELING OF POST-DENITRIFICATIONSINGLE-SLUDGE ACTIVATED SLUDGE PROCESSES
RICHARD O. MINES, JR.Department of Civil and Environmental Engineering University of South Florida 4202 East Fowler
Avenue, ENB 118 Tampa, Florida 33620, USA
(Received 3 January 1996; accepted 8 July 1996)
Abstract. A step-by-step design approach for sizing oxic/anoxic single- sludge activated sludgesystems is presented. A computer model utilizing a Lotus 1–2–3 spreadsheet was also developed whichsimulates the post- denitrification single-sludge activated sludge process. The biokinetic equationsdeveloped by Lawrence and McCarty are utilized in the model. The model was calibrated using datafrom a full-scale, modified pure oxygen activated sludge process operating in the oxic/anoxic mode.Two- tailed, paired comparison student ‘t-tests’ performed on effluent BOD5, SS, TKN, NH3-N,NO�3 -N, and TN indicated there was no significant difference between the actual effluent values andthat predicted by the model at a 99 percent confidence level.
Key words: nitrification, post-denitrification, single-sludge, steady-state model
1. Introduction
Various models have been proposed for simulating the activated sludge process(McKinney, 1962; Eckenfelder and O’Connor, 1955; and Lawrence and McCarty,1970). Goodman and Englande (1974) have shown that McKinney’s and Ecken-felder’s models are essentially the same. The most widely accepted model in theUnited States is that developed by Lawrence and McCarty (1970). This modelis taught at most universities and is presented in such texts as Reynolds (1982),McGhee (1991), Metcalf and Eddy (1991), Water Environment Federation Manualof Practice (1992), and Viessman and Hammer (1993). With the increasing popu-larity of biological nutrient removal (BNR) systems, other approaches have beenproposed for sizing such systems (Mines et al., 1992; Randall et al., 1992). TheSimulation of Single-Sludge Processes (SSSP) model developed by the Internation-al Association on Water Pollution Research and Control (IAWPRC) task group wasan attempt to try to promote a world-wide accepted, user-friendly model (Dold andMarais, 1986; Hence et al., 1987; and Bidstrup and Grady, 1988). Though this wasa notable idea, the input parameters required for running the model are numerousand at least three of them (inert soluble organic matter -SI , readily biodegradablesubstrate -SS and inert suspended organic matter -XI) are parameters not normallymeasured at domestic wastewater treatment facilities (Hence et al., 1987). Thispaper presents a straight-forward, simplified step-by-step approach to designingpost-denitrification single- sludge activated sludge systems.
Water, Air, and Soil Pollution 100: 79–88, 1997.c 1997 Kluwer Academic Publishers. Printed in the Netherlands.
80 RICHARD O. MINES JR.
2. Design approach
The design approach presented herein, utilizes biokinetic equations developed byLawrence and McCarty (1970) to size the aerobic zone along with a specificdenitrification equation proposed by Burdick et al. (1982) to design the anoxiczone. A similar approach is being promoted for designing pre-denitrification and/orcombined pre- and post-denitrification systems (Nutrient Control, in preparation).
2.1. AEROBIC ZONE CALCULATIONS
First, select minimum weekly temperature, minimum pH, and minimum dissolvedoxygen (DO) level to be maintained. Second, calculate the maximum growth rateof Nitrosomonas using the following equation developed from data presented byKnowles et al. (1965), corrected for pH and DO as presented in the EPA NitrogenControl Manual (Parker et al., 1975):
(�MAX)NS = e(0:0985 T�2:25)�
DOKDO + DO
�[1 � 0:833(7:2 � pH)] (1)
�(MAX)NS = Maximum growth rate of Nitrosomonas, day�1;
T = Mixed liquor temperature, �C;
pH = pH of mixed liquor, standard pH units;
DO = Dissolved oxygen level in aeration basin or oxic zone, mg/L;and
KDO = Half-saturation constant for DO, mg/L; 1.0 mg/L (Bidstrupand Grady, 1988).
Third, calculate the minimum solids retention time (SRT) in days required for a giv-en temperature, DO, pH, and influent ammonia concentration using the followingequation:
1(SRT)MIN
=(�MAX)NS(NH+
4 � N)o
KNS + (NH+4 � N)o
� bNS (2)
(SRT)MIN = Minimum SRT required for nitrification, days;
(NH4+-N)o = Influent ammonium nitrogen concentration, mg/L;
bNS = Endogenous decay coefficient for Nitrosomonas, days�1;
KNS = Half-saturation constant for ammonium nitrogen, mg/L; and
KNS = 100:0151T�1:148 (Knowles et al., 1965).Fourth, calculate the design SRT of the aerobic zone in days, based on a safetyfactor and peaking factor as follows:
(SRT)DESIGN = (SRT)MIN (SF)(PF) (3)
SINGLE-SLUDGE ACTIVATED SLUDGE PROCESSES 81
(SRT)DESIGN = The SRT used in designing the aerobic zone, days;
SF = Safety factor based on uncertainty of performance, usually 1.2to 2.0; and
PF = Peaking factor based on the peak nitrogen loading to averageloading to the treatment unit, usually 1.1 to 1.2.
Fifth, estimate the overall SRT of the biological system in days using the followingequation:
(SRT)OVERALL = (SRT)DESIGN (MF) (4)
(SRT)OVERALL = The SRT of the oxic and anoxic zone, days.The multiplication factor (MF) accounts for the anoxic volume which will increasethe SRT of the system. For most Biological Nutrient Removal (BNR) systems, thisparameter will vary from approximately 1.4 to 2.0. The multiplication factor (MF)is calculated as follows:
MF = 1=(1 � FANOXIC) (5)
FANOXIC = Anoxic zone fraction of total reactor volume.Sixth, for a given set of heterotrophic biokinetic constants, calculate the soluble
effluent BOD5 (Se) in mg/L for the (SRT)OVERALL as follows:
Se =Ks[1 + b(SRT)OVERALL]
[(SRT)OVERALL(Yk � b) � 1](6)
b = Endogenous decay coefficient for organic removal, days�1;
Y = Yield coefficient, g VSS/g BOD5;
k = Maximum substrate (BOD5) utilization rate, days�1; and
KS = Half-saturation constant for BOD5, mg/L.Seventh, for a given set of autotrophic biokinetic constants, calculate the efflu-ent ammonium nitrogen concentration (NH4
+-N)e in mg/L using the followingequation:
(NH+4 � N)e =
KNS[1 + bNS(SRT)DESIGN]
(SRT)DESIGN[YNS kNS � bNS]�1(7)
Eighth, determine the amount of nitrogen to be oxidized (NO) in mg/L as follows:
NO = TKNo � (NH+4 )e � NSYN (8)
TKNo = Influent total Kjeldahl nitrogen, mg/L; and
NSYN = Influent nitrogen used in synthesis of biomass, mg/L.The nitrogen utilized in synthesis (NSYN) neglecting the small amount of nitro-
gen synthesized by the nitrifiers is calculated using the following equation:
NSYN =Y(So � Se)FN
[1 + b(SRT)OVERALL]+ (Xe)FN (9)
82 RICHARD O. MINES JR.
Xe = Effluent VSS concentration, mg/L; and
FN = Fraction of nitrogen in volatile suspended solids (0.05 to 0.12).Ninth, the volume of the aerobic zone (VOXIC) neglecting synthesis by the nitrifiersin cubic meters necessary to achieve nitrification for a given temperature, pH, andDO is calculated as follows:
VOXIC =Q(SRT)DESIGN
X
�Y(So � Se
1 + b(SRT)DESIGN+ XL
�(10)
So = Influent total BOD5 to biological treatment system, mg/L;
X = Mixed liquor suspended solids concentration, mg/L;
Q = Volumetric flowrate into the biological treatment process,m3/day; and
XL = Inert solids in the influent (FSS + nondegradable VSS), mg/L.
2.2. ANOXIC ZONE CALCULATIONS
Estimate the mass of nitrates to be removed in the anoxic zone. First, calculate thenitrate equivalence of dissolved oxygen (NO3
�-N)EQ in the mixed liquor from theaerobic zone in kg/day as follows:
(NO�
3 �N)EQ = (DO)ML
0:35
g NO�
3 � Ng O2
!(Q + RAS)
�1 kg
1000g
�(11)
(DO)ML = Dissolved oxygen concentration in aerobic zone, mg/L, and
RAS = Return activated sludge flow, m3/day.Second, calculate the mass of nitrates (NOR) to be removed in the anoxic zone inkg/day using the following equation:
NOR = Q[NO � (NO3 � N)e ] (1 kg=1000 g) (12)
(NO3� - N)e = Desired effluent nitrate-nitrogen concentration, mg/L.
Third, calculate the total mass of nitrates (TNOR) to be removed in the anoxic zonein kg/day using the following equation:
TNOR = [NOR + (NO�
3 � N)EQ] (1 kg=1000 g) (13)
Fourth, estimate the specific denitrification rate (SDNR2) in days�1 corrected forambient temperature in the post-denitrification, anoxic zone using the followingequation (Burdick et al., 1982):
SDNR2 = 0:12(SRT)�:706OVERALL(1:02)T�20 (14)
Fifth, calculate the volume of the anoxic zone (VANOXIC) in cubic meters asfollows:
VANOXIC =(TNOR) (1000 g=kg)
(X)(SDNR2)(15)
SINGLE-SLUDGE ACTIVATED SLUDGE PROCESSES 83
Sixth, the total volume (VTOTAL) of the biological system in cubic meters can nowbe calculated using the following equation:
VTOTAL = VOXIC + VANOXIC (16)
Seventh, the quantity of sludge (PX) in kg/day that must be wasted from this systemcan be calculated using the following equation which neglects the small amount ofbiomass synthesized by the nitrifiers:
PX =
�Y(So � Se)
1 + b(SRT)OVERALL+ XL
�(Q)(1 kg=1000 g) (17)
Eighth, calculate the overall SRT in days of the biological system as follows:
(SRT )OV ERALL =X(VOXIC + VANOXIC)
PX(1000 g=kg)(18)
If the calculated SRT is not within 5 percent of the assumed (SRT)OVERALL
in Step 5 of the aerobic zone calculations, the entire design procedure must berepeated.
2.3. OXYGEN REQUIREMENTS
Oxygen is required to meet both the carbonaceous and nitrogenous oxygen demands.The total kilograms of oxygen required daily (O2) can be estimated using the fol-lowing equations:
O2 = CBOD + NOD � DOC (19)
CBOD = [Q[(1�1:42Y )(So�Se)] + 1:42(b)(X)VOXIC ] (1 kg=1000 g)(20)
NOD = Q(4:57)(NO)(1 kg=1000 g) (21)
DOC = Q
2:86
g O2
g NO�
3 � N
![NO � (NO�
3 � N)e ] (1 kg=1000 g) (22)
CBOD = Carbonaceous oxygen demand, kg/day;
NOD = Nitrogenous oxygen demand, kg/day; and
DOC = Denitrification oxygen credit, kg/day.
2.4. ALKALINITY REQUIREMENTS
Sufficient alkalinity must be maintained so that the pH does not drop duringnitrification thereby, inhibiting the process. Normally, 50–100 mg/L of alkalinityas CaCO3 are maintained in the effluent from BNR systems. The effluent alkalinity
84 RICHARD O. MINES JR.
Table IBiokinetic Constants Used in Modeling Runs
Heterotrophs Nitrosomonas
Y 0.6 g VSS/g BOD5 0.15 g VSS/g NH4+-N
k 5.0 days�1 3.0 days�1
b 0.06 days�1 0.05 days�1
Ks 60 mg/L BOD5 100:051T�1:148 mg/L NH4+-N
T = temperature �C.
(ALKe) in mg/L as CaCO3 can be calculated as follows from the anoxic/oxicprocess:
ALKe = ALKo � 7:14(NO) + 3:57[NO � (NO�
3 � N)e ] (23)
ALKo = Influent alkalinity, mg/L as CaCO3.
3. Description and validation of model
A spreadsheet model using Lotus 1–2–3 was developed using the equations pre-sented above. Data from the 75,700 m3/day (20 mgd) Main Street WastewaterTreatment Plant in Pensacola, Florida were used in calibrating the model to deter-mine if the design approach presented herein was valid. The major liquid traincomponents include: grit removal, primary clarification, biological treatment, sec-ondary clarification, chlorination, dechlorination, and discharge to Pensacola Bay.The National Pollution Discharge Elimination System (NPDES) permit sets themaximum annual average load discharged to Pensacola Bay at 606 kg/day (1,334ppd) for both five-day biochemical oxygen demand (BOD5) and suspended solids(SS), 455 kg/day (1,000 ppd) for total nitrogen (TN), and 76.4 kg/day (168 ppd)for total phosphorus (TP), respectively. At 75,700 m3/day, this equates to effluentconcentrations of 8 mg/L for BOD5 and SS, 6 mg/L for TN, and 1 mg/L for TP.In the existing configuration the back half of the aeration basin is operated as ananoxic zone resulting in a process capacity of approximately 68,100 m3/day (18mgd) or 80 percent of capacity. Phosphorus removal was accomplished in the pri-maries through alum addition at approximately 110 mg/L to the primary influent.Lime at 13 mg/L was added to the primary effluent to increase the alkalinity forbiological treatment. At times, methanol was added to the anoxic zone to enhancedenitrification. Nonionic polymer at approximately 0.5 mg/L was added to thesecondary influent for enhancing SS removal in the secondaries and to eliminatethe need for effluent filtration.
The equations presented under the design approach section had to be rearrangedslightly since the volume of the oxic and anoxic zones were already known. Parame-ters input into the model included: influent values for BOD5, SS, TN, Q, VSS/SS,
SINGLE-SLUDGE ACTIVATED SLUDGE PROCESSES 85
Tabl
eII
Effl
uent
Com
pari
son
DA
TE
BO
D5
TSS
TK
NN
H3
NO
� 3T
N
1992
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
3/1
6.4
8.3
2.0
8.2
6.5
3.8
5.6
3.0
2.9
3.0
9.4
6.8
3/2
10.5
8.7
2.0
8.6
5.0
4.2
4.5
3.4
3.3
4.3
8.3
8.5
3/3
12.5
8.6
6.8
8.3
6.1
3.9
5.7
3.2
2.7
6.5
8.8
10.4
3/4
11.5
13.8
3.6
16.0
6.8
3.5
6.0
1.2
2.5
3.8
9.3
7.2
3/5
15.0
8.4
68.0
6.9
4.9
0.0
3.9
0.0
2.8
0.0
7.7
0.0
3/6
12.3
4.7
7.6
3.0
5.1
2.9
4.1
2.6
2.6
12.9
7.7
15.9
3/7
8.3
9.3
1.2
8.9
6.6
4.5
5.4
3.6
1.8
1.9
8.4
6.4
3/8
10.3
6.7
3.6
5.7
5.8
3.5
4.8
2.9
2.3
4.1
8.1
7.6
3/9
11.4
7.5
3.6
7.0
3.8
4.8
3.3
4.0
2.7
3.4
6.5
8.2
3/10
14.0
9.7
8.4
9.9
7.2
8.1
5.8
7.3
1.5
0.0
8.7
8.1
3/11
10.0
11.8
19.0
13.3
6.7
6.4
5.1
5.3
2.0
0.2
8.7
6.6
3/12
7.6
11.3
4.8
11.9
6.0
4.1
4.2
2.9
2.5
5.3
8.5
9.4
3/13
8.3
15.2
7.4
17.8
4.9
3.8
3.5
2.2
3.0
1.8
7.9
5.6
3/14
5.4
7.5
2.0
6.6
4.6
10.4
3.3
9.8
2.9
0.1
7.5
10.4
3/15
12.0
9.6
6.4
10.5
6.9
4.0
5.1
3.1
2.1
4.5
9.0
8.5
3/16
7.2
9.0
2.4
8.9
5.9
5.0
4.8
4.1
2.5
3.1
8.4
8.1
3/17
6.6
22.8
3.6
29.2
4.8
4.3
4.5
1.7
2.0
0.0
6.8
4.3
3/18
12.4
16.8
6.0
19.6
4.8
4.1
3.2
2.3
1.1
0.0
5.9
4.1
3/19
7.7
10.2
5.6
10.9
3.5
4.3
2.4
3.3
0.7
0.0
4.2
4.3
3/20
2.9
12.6
2.4
14.3
3.1
4.3
2.2
3.0
3.0
3.2
6.1
7.5
(Tab
leII
cont
inue
son
next
page
)
86 RICHARD O. MINES JR.
Tabl
eII
Con
tinu
ed.
DA
TE
BO
D5
TSS
TK
NN
H3
NO
� 3T
N
1992
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
AC
TU
AL
MO
DE
LA
CT
UA
LM
OD
EL
3/22
9.2
10.6
4.4
11.3
3.5
4.1
2.6
3.1
2.9
3.0
6.4
7.1
3/23
8.3
14.0
3.2
16.0
3.0
4.2
1.9
2.7
1.9
2.6
4.9
6.9
3/24
9.3
10.1
4.0
10.7
3.1
3.4
2.2
2.4
3.3
4.8
6.4
8.2
3/25
12.0
8.4
11.2
8.9
7.6
3.4
5.9
2.6
1.5
3.9
9.1
7.3
3/26
13.8
14.4
5.2
16.2
6.2
4.3
4.6
2.7
2.4
0.0
8.6
4.3
3/27
9.8
10.8
4.2
11.7
5.4
4.1
3.9
3.0
2.4
2.3
7.8
6.4
3/28
3.5
10.3
3.4
11.5
3.2
3.9
1.9
2.9
4.6
3.0
7.8
6.9
3/29
2.9
7.0
2.0
6.7
2.2
2.7
1.1
2.1
5.8
7.9
8.0
10.6
3/30
6.2
6.9
48.0
5.5
2.8
11.3
1.6
10.8
5.0
0.6
7.8
11.9
3/31
8.1
7.8
2.4
7.3
3.7
4.5
2.5
3.9
7.4
5.0
7.4
9.6
All
conc
entr
atio
nsin
mg/
L;i
ncom
plet
eda
tase
ton
3/21
/92
BO
D5
=5-
day
bioc
hem
ical
oxyg
ende
man
dT
SS=
tota
lsus
pend
edso
lids
TK
N=
tota
lKje
ldah
lnitr
ogen
NH
3=
amm
onia
nitr
ogen
NO
3
�
=ni
trat
eni
trog
enT
N=
tota
lnit
roge
n
SINGLE-SLUDGE ACTIVATED SLUDGE PROCESSES 87
and alkalinity; heterotrophic and autotrophic biokinetic constants for Y, k, Ks, andb; mixed liquor values for VSS/SS, DO, sludge volume index (SVI), and tempera-ture; anoxic and oxic volumes; and the overall SRT of the biological system. Thevalues for k and Ks were corrected for temperature variation with a temperaturecorrection coefficient (�) of 1.09 whereas, the endogenous decay coefficients (b)were corrected for temperature with a � of 1.04. Table I lists the biokinetic con-stants utilized in the modeling runs. SS concentrations in the secondary effluentwere estimated by assuming that 98 percent of the mixed liquor suspended solidswhere removed during clarification. Total effluent BOD5 was estimated by addingthe effluent soluble BOD5 predicted from the biokinetic equation to the estimatedBOD5 of the effluent TSS assuming that the theoretical COD of the volatile portionwas 1.42 kg per kg of biomass. The five day BOD was then assumed to be approxi-mately two thirds of the theoretical COD. The return activated sludge recycle ratio(R) was determined from the following equation:
R =
h1�c�
QV
ihQV�
Q(106)(V )(SV I)(X)
i (24)
Primary and secondary effluent data along with SVI values for the month of March1992 were utilized in validating the model. Methanol was not added to the anoxiczone during March 1992. Analyses were conducted on 24-hour composite samplesin accordance to Standard Methods (1985). Table II presents the results fromthe modeling runs. Two tailed, paired comparison student ‘t-tests’ performed oneffluent BOD5, SS, total Kjeldahl nitrogen (TKN), ammonia nitrogen (NH3-N),nitrate nitrogen (NO3
�-N), and total nitrogen (TN) were performed at a 99 percentconfidence level. Results indicated there was no significant difference between theactual effluent parameters and that predicted by the model, thereby validating themodel. The design approach presented herein is a straight-forward procedure thatcan be utilized by engineers to design post-denitrification single-sludge systems.With minor modifications, the approach can be used for sizing pre-denitrificationand combined pre- and post-denitrification activated sludge systems.
4. Summary and conclusions
A step-by-step design approach for sizing post-denitrification single-sludge activat-ed sludge systems was presented. The procedure utilizes the biokinetic equationsdeveloped by Lawrence and McCarty (1970) for the aerobic zone and the specif-ic denitrification rate equation developed by Burdick et al. (1982) for sizing theanoxic zone. The model was validated using data from a full-scale modified pureoxygen process operating in the oxic/anoxic mode. Two tailed, paired comparisonstudent ‘t-tests’ performed on actual versus model effluent parameters for BOD5,SS, TKN, NH3-N, NO3
�-N, and TN indicated there was no significant difference
88 RICHARD O. MINES JR.
between the two at a 99 percent confidence level, thereby validating the model.The design procedure presented herein is not as complex as the SSSP model, andtherefore may be more practical and ‘user friendly’ to practicing design engineers.
References
Bidstrup, S. M. and Grady, C. P. L.: 1988, J. Water Pollution Control Fed., 60, 351.Burdick, C. R., Refling, D. R., and Stensel, H. D.: 1982, J. Water Pollution Control Fed. 54, 1078.Design of Municipal Wastewater Treatment Plants: 1992, Manual of Practice No. 8, Water Environ-
ment Federation, Alexandria, VA, 895.Dold, P. L. and Marais, G. V. R.: 1986, Water Sci. Tech. 18, 63.Eckenfelder, W. W. and O’Connor, D. J.: 1955, Proc. 9th Industrial Waste Conference, Purdue
University, W. Lafayette, IN, Ext. Ser. 89, 39, 2, 512.Goodman, B. L. and Englande, A. J.: 1974, J. Water Pollution Control Fed. 46, 312.Hence, M., Grady, C. P. L., Gujer, W., Marais, G. V. R., and Matsuo, T.: 1987, Water Resources 21,
505.Knowles, G., Downing, A. L., and Barrett, M. J.: 1965, J. General Microbial. 38, 263.Lawrence, A. W., and McCarty, P. L.: 1970, J. SED, ASCE, 96, 757.Metcalf and Eddy: 1991, Wastewater Engineering: Treatment, Disposal, and Reuse. McGraw Hill,
Inc., New York, NY.McGhee, T. J.: 1991, Water Supply and Sewerage. McGraw Hill, New York, New York.McKinney, R. E.: 1962, J. SED, ASCE, 88, 87.Mines, R. O., Smith, D. G., Dahl, B., and Holcomb, S.: 1992, Proc. 65th Annual Conference of the
Water Environment Federation, New Orleans, LA.Nutrient Control (in preparation). Manual of Practice, Water Environment Federation, Alexandria,
VA.Parker, D. S., Stone, R. W., and Stenquist, R. J.: 1975, Process Design Manual for Nitrogen Control.
EPA 625/1–75–007, U.S. E.P.A., Washington, DC.Randall, C. W., Barnard, J. L., and Stensel, H. D.: 1992, Design and Retrofit of Wastewater Treatment
Plants for Biological Nutrient Removal. Technomic Publishing Company, Inc., Lancaster, PA.Reynolds, T. D.: 1982, Unit Operations and Processes in Environmental Engineering. PWS-Kent
Publishing Company, Boston, Massachusetts.Standard Methods for the Examination of Water and Wastewater: 1985, 16th Edition, American
Public Health Association, Washington, D.C.Viessman, W. and Hammer, M. J.: 1993, Water Supply and Pollution Control. Harper Collins College
Publishers, New York, New York.