Chinese Journal of Oceanology and Limnology http://dx.doi.org/10.1007/s00343-015-4108-8
Evaluation and improvement of wastewater treatment plant performance using BioWin* OLEYIBLO Oloche James1, **, CAO Jiashun (操家顺)1, 2, FENG Qian (冯骞)1, WANG Gan (王淦)3, XUE Zhaoxia (薛朝霞)1, FANG Fang (方芳)1 1
College of Environment, Hohai University, Nanjing 210098, China
2
National Engineering Research Center of Water Resource Efficient Utilization and Engineering Safety, Hohai University,
3
Anhui Guozhen Environmental Protection Sci. and Tech. Co. Ltd., Hefei 230088, China
Nanjing 210098, China
Received May 13, 2014; accepted in principle Jul. 4, 2014; accepted for publication Jul. 16, 2014 © Chinese Society for Oceanology and Limnology, Science Press, and Springer-Verlag Berlin Heidelberg 2015
Abstract In this study, the activated sludge model implemented in the BioWin® software was validated against full-scale wastewater treatment plant data. Only two stoichiometric parameters (Yp/acetic and the heterotrophic yield (YH)) required calibration. The value 0.42 was used for Yp/acetic in this study, while the default value of the BioWin® software is 0.49, making it comparable with the default values of the corresponding parameter (yield of phosphorus release to substrate uptake (YPO )) used in ASM2, ASM2d, and ASM3P, respectively. Three scenarios were evaluated to improve the performance of the wastewater treatment plant, the possibility of wasting sludge from either the aeration tank or the secondary clarifier, the construction of a new oxidation ditch, and the construction of an equalization tank. The results suggest that construction of a new oxidation ditch or an equalization tank for the wastewater treatment plant is not necessary. However, sludge should be wasted from the aeration tank during wet weather to reduce the solids loading of the clarifiers and avoid effluent violations. Therefore, it is recommended that the design of wastewater treatment plants (WWTPs) should include flexibility to operate the plants in various modes. This is helpful in selection of the appropriate operating mode when necessary, resulting in substantial reductions in operating costs. 4
Keyword: activated sludge; aeration tank; equalization tank; secondary clarifier; wet weather
1 INTRODUCTION The use of mathematical models for WWTP design, process optimization, operator training, and developing control strategies has become a standard engineering tool in the last decade (Hulsbeek et al., 2002; Petersen et al., 2002; Rieger et al., 2008). This is because substantial resources have already been invested in existing wastewater treatment infrastructures. The need to maximize the use of these previous investments by upgrading existing plants to be able to handle increased loads, incorporate nutrient removal capability, and meet stricter effluent demand requires a system prototype in the form of a mathematical model. Mathematical models play service role, advice, and analysis roles (Russel et al., 2002). In general, biological nutrient removal is the most cost-effective
method of treatment; however, such techniques result in complex process configurations, making systems that employ them difficult to operate (Vanrolleghem, 2001; Puig et al., 2008). The various interactions that occur among process units and the fact that the biological potential is taken to its limit makes nutrient removal plants rather susceptible to external disturbances or erroneous manipulations of control variables (Vanrolleghem, 2001). Thus, the increased complexity resulting from integration of these processes requires the introduction of models, advanced instrumentation and control systems. * Supported by the College of Scientific Innovation Significant Cultivation Fund Financing Projects (No. 708047) and the Key Special Program for Pollution Control (No. 2012ZX07101-003) ** Corresponding author:
[email protected]; evangjamesa1@ yahoo.com
CHIN. J. OCEANOL. LIMNOL.,
Optimization of an existing facility in terms of capital or operational improvements is the most cost effective method of achieving compliance. Mathematical simulations have provided impetus to this approach by increasing motivation for improving existing plant performance. In particular, standardized mathematic models implemented in several simulator platforms are used in the design, troubleshooting, optimization, control and research of activated sludge systems (Gernaey et al., 2004; Ferrer et al., 2008). This study was conducted to optimize existing wastewater treatment plant (WWTP) facilities to ensure consistent biological nutrient removal (BNR) in compliance with the designed effluent concentrations. This study was also intended to identify reliable operational and capital improvements to enable future expansion that will accommodate increased flows during wet weather conditions while minimizing the cost.
2 MATERIAL AND METHOD 2.1 Description of WWTP The WWTP investigated in this study is located in Hefei, Anhui Province, People’s Republic of China. The plant is operated and maintained by Anhui Guozhen Environmental Protection Sci-Tech Co. Ltd. (GZEP). The system consists of four-train oxidation ditches (ODs), each of which is preceded by an anaerobic and anoxic zone. The total volume of the ODs is 27 498 m3, whereas the volumes of the anaerobic and anoxic tanks are 4 150 m3 and 7 922 m3, respectively. Each OD operates in alternating anoxicaerobic modes, and two secondary clarifiers are connected to each OD train. Each clarifier has a diameter of 47 m (excluding the area occupied by the centre well) and a depth of 4.6 m, giving a total surface area of 1 657 m2. The anaerobic zone serves as the medium for biological phosphorus removal. Nitrified recycling is accomplished by a return from the oxidation ditch to the anoxic zone. The plant was designed to treat 200 000 m3/d of wastewater with the following influent concentrations (mg/L): BOD=180, COD=450, NH3-N=30, SS=200, TP=6, and TN=45, whereas the effluent concentrations (mg/L) are: BOD≤10, COD≤50, NH3-N≤8, TN≤15, TP≤0.5, and SS≤10. At present, the design capacity is exceeded at times; therefore, expansions to enable treatment of an additional 50 000 (m3/d) have been proposed. These expansions will require construction of an oxidation ditch train, necessitating the use of mathematical
simulations to evaluate different scenarios for its implementation. 2.2 Data collection and evaluation of historical data Operational and performance data from 2010 to 2012 were evaluated (Fig.1a–j) to assess plant operations and determine the treatment efficiency of the existing process. Moreover, data reliability was evaluated and possible systematic errors were identified by the application of mass balances (Meijer et al., 2002; Langergraber et al., 2004; Puig et al., 2008; Thomann, 2008). Additionally, the unit processes and an additional sampling campaign aimed at specific compounds and key parameters of interest were evaluated based on the given data. The WWTP was intensively sampled for 5 and 7 days (May 14–18, and November 1–7, 2012) for proper influent wastewater characterization to enable model calibration and validation, respectively. Samples were collected every 2 h from the influent, raw influent with recycled influent, secondary effluent and final effluent for 5 and 7 days, respectively. The influent characterization was determined according to the STOWA method (STOWA, 1996; Meijer et al., 2001). Specifically, the Chemical Oxygen Demand (COD) of non-filtered samples (total COD) and samples filtered through 0.45-μm and 1.2-μm diameter pores (COD influent soluble, filtered, flocculated), Biochemical Oxygen Demand (BOD) of non-filtered and filtered samples, volatile fatty acids (VFA), ammonium nitrogen (NH4-N), nitrate nitrogen (NO3-N), total Kjeldahl nitrogen (TKN), total phosphorus (TP), ortho-phosphate (PO34ˉ), total suspended solids (TSS) and volatile suspended solids (VSS) were measured. Experiments were performed in the GZEP laboratory in accordance with the standard methods for the examination of water and wastewater (APHA, 1998). The wastewater characteristics and operating parameters of the WWTP at the time this work was performed are given in Tables 1 and 2. 2.3 Model calibration and validation The simulation was performed using the BioWin® software version v.3.0 (EnviroSim Associates Ltd., Canada). This model uses the integrated activated sludge/anaerobic digestion (AS/AD) model, which is referred to as the BioWin General Model. The BioWin® AS/AD model is an extended version of previously described models (Barker and Dold, 1997;
a
350
b
2×105
300
1.5×105
250
1×105
200 Monthly average Q
5×104
Influent COD
150 100
Jan-2010
c
Jul-2011
Jul-2012 45
d
40
100 90
35
80
30
70
25
Effluent COD
Influent BOD
60
20
50 Jan-2010 140
Jan-2011
Jan-2012
Jan-2010
e
Jul-2011
Jul-2012 60
f
120
BOD/COD
100
Influent TP (mg/L)
40 30
60
20
COD/TP
40
10
20 Jan-2010
Jan-2011
Jan-2012
Jan-2010
g
Jul-2011
h Influent TP
5
Jul-2012
Effluent TP
1.2 1
4
0.8 3
0.6
2
0.4 0.2
1 Jan-2010
Influent TN & NH3-N (mg/L)
50
80
6
50
Effluent COD (mg/L)
110
Jan-2012
Influent BOD/TP (mg BOD/mg P)
120
Jan-2011
Effluent TP (mg/L)
Influent COD/TP (mg COD/mg P)
Monthly average infl BOD (mg/L)
Jan-2010
Monthly average infl CODt (mg/L)
2.5×105
Jan-2011
Jan-2012
Jan-2010 Date (year), 2010–1012
i
Jul-2011
Jul-2012
20
j TN NH3-N
40
15
30 10 20 10
Influent TN
5
Influent NH3-N
0
0 Jan-2012
May-2012
Sep-2012
Jan-2012 Date (months), 2012
Jul-2012
Effluent TN & NH3-N (mg/L)
Monthly average infl Q (m3/d)
OLEYIBLO et al.: Wastewater treatment plant improvement using BioWin
Nov-2012
Fig.1 WWTP influent/effluent concentrations for 3 years
Liwarska et al., 2010; Liwarska et al, 2013) with regard to AS and AD. Additionally, BioWin® has three models to evaluate settling and separation processes, a flux based model, an ideal separation model, and a point separation model (BioWin User Manual, 2008).
The ideal separation model was used in this study. Under the hydrodynamic conditions of the investigated WWTP, the oxidation ditch process approaches a completely mixed reactor with a relatively short plug flow time; thus, complete mixing was assumed for the
CHIN. J. OCEANOL. LIMNOL., Table 1 Operational data and wastewater characteristics of the WWTP, May, 2012
Table 2 Operational data and wastewater characteristics of the WWTP, Nov. 2012
Parameter
Unit
Mean
SD
Minimum
Maximum
Parameter
Unit
Mean
SD
Minimum
Maximum
Flow rate
m /d
218 618
9 069
206 419
238 779
Flow rate
m /d
195 018
5022
190 340
204 533
3
3
SRT
d
17
0.8
15
20.3
SRT
d
16
0.6
14.7
18
RAS flow
m3/d
97 628
3 881
28 889
107 451
RAS flow
m3/d
78 025
2 060
76 184
107 451
Temperature
°C
22.4
0.85
21.4
23.8
Temperature
°C
13.7
1.84
11
17
pH
-
7.3
-
-
7.3
pH
-
7.3
-
-
7.3
TSS
mg/L
153
11.1
138
165
TSS
mg/L
207
49.9
146.8
235.7
VSS
mg/L
103
10.82
90
120
VSS
mg/L
153.2
48.5
89
198
ISS
mg/L
50
10
36
52
ISS
mg/L
53.7
5
45
60
BOD5
mg/L
101
12.9
80.2
116
BOD5
mg/L
139
29
90
168
fBOD5
mg/L
49.3
6.3
39.3
56.8
fBOD5
mg/L
65.7
19
26
79
CODtot
mg/L
253
19.24
224
276
CODtot
mg/L
252
68
158
343
fCOD
mg/L
143
11
132
161
fCOD
mg/L
119
35
79
173.6
ffCOD
mg/L
76
6
73
83
ffCOD
mg/L
67
21.2
42
92.8
VFA
mg/L
30.3
8.62
19
41
VFA
mg/L
40.6
16.6
23
86
TP
mg/L
2.43
0.36
2
3
TP
mg/L
5.4
1.6
2.9
8.2
SPO4
mg/L
1.73
0.25
1.43
2.14
SPO4
mg/L
4.2
2
2.2
6.1
TN
mg/L
35.5
1.26
33.8
37.4
TN
mg/L
34.5
4
29
40
TNi-sol
mg/L
31.2
1.1
29.8
32.9
TNi-sol
mg/L
31.7
1.9
26.2
35.4
N-NH3
mg/L
26.8
1.36
25.3
29.1
N-NH3
mg/L
29.6
3.96
25.3
33.5
N-NO3
mg/L
0.25
0.1
0
0.32
N-NO3
mg/L
0.2
0.18
0.05
0.59
Table 3 Wastewater COD fractions Name
Description
Model default
Calculated
Fbs
Fraction of total influent COD that is soluble and readily biodegradable, including acetate
0.15
0.2 (plant data)
Fac
Fraction of readily biodegradable material that is VFA or fermentation product-acetate (essential for bio-P removal)
0.16
0.19 (plant data)
Fus
Fraction of total influent COD that is soluble non-biodegradable
0.05
0.052 (plant data)
Fup
Fraction of total influent COD that is particulate non-biodegradable
0.13
Default
Fxsp
Fraction of slowly biodegradable influent COD that is particulate and colloidal
0.75
Default
hydraulic conditions in the bioreactor of the studied WWTP. In this study, one of the treatment trains was configured and modelled using the BioWin® software v.3.0. This is because the four treatment trains receive the same influent wastewater, both in quantity and concentrations, and all clarifiers are of the same size. The choice to model one treatment train is in agreement with a study conducted by Hulsbeek et al. (2002), who attributed considerable deviation from default model parameter values to errors caused by lack of proper representation of the system being studied in the simulator. Model calibration was conducted by fitting a model
to a set of data or information collected from the fullscale WWTP under study (Petersen et al., 2002). As such, an accurate estimate of the influent wastewater to describe site-specific wastewater compositions and process characteristics is essential to model performance and reliable predictions. The analytical results of the specific raw wastewater COD fractions determined according to Mamais et al. (1993) are shown in Table 3, whereas the general wastewater characteristics are given in Tables 1 and 2. The BioWin® configuration of the WWTP is shown Fig.2. The data used for the steady-state model calibration were averaged daily composite data collected every 2 hours for 1 month in May of 2012. The average
OLEYIBLO et al.: Wastewater treatment plant improvement using BioWin
Influent
Anaerobic
Anoxic OD aerobic1 OD aerobic2
OD aerobic4 OD anoxic2 OD aerobic3 OD anoxic1
Effluent
WAS
Fig.2 BioWin schematic flow diagram of the Hefei WWTP biological reactor Table 4 Statistical values of the model predictive ability Variable
TN
TP
COD
MLSS
J
0.94
0.83
0.98
1.2
Table 5 ARD results for calibration and validation periods Month
TP
TN
COD
May, 2012, Calibration
6.12
9.44
4.35
November, 2012, Validation
5.2
4.85
3.92
temperature (17.6°C) for the entire period was used for simulation of the steady-state data. The data were evaluated statistically and the confidence intervals were calculated by a Student’s t-test at the 98% confidence level. Calibration targets were set for parameters of interests as follows: effluent TP=10%, TN=10%, COD=10%, NO3-N=10%, and NH3-N=20% relative percent difference (RPD) between model and plant data. The average data were then input into the model. The data collected during the 5 day intensive sampling campaign between May 14 and 18, 2012 shown in Table 1 were used for the dynamic simulation (referring to model calibration), and the dynamic temperature data were used for the simulation. The steady-state model was validated using average daily composite data collected every 2 hours for 1 month in November of 2012. The data collected during 7 days from November 1 to 7, 2012 were used for dynamic simulation (Table 2). Dynamic input influent wastewater characteristics were used for dynamic simulations for both model calibration and validation. It should be noted that the data in Table 1 and 2 are averages of the dynamic input influent wastewater characteristics used for both dynamic simulations. The model performance was further assessed by determining its predictive quality and stability between the calibration and validation datasets based on the Janus coefficient, J (Sin et al., 2008; WEF,
2010). The value of J varies between 0 and ∞. The higher J is above 1, the poorer the predictive ability of the model with respect to that variable is, while J will be approximately equal to 1 if the predictive ability of the model remains more or less constant outside the calibration period. Additionally, the average relative difference (ARD) between the observed and simulated values was used to further evaluate the model performance. The statistical evaluation results are given in Table 4 and 5, respectively. The Janus coefficient was calculated as follows: 1 m ni 2 C n i Cmod i 1 ob , J2 m 2 1 n i i Cob Cmod n i 1
(1)
where J=the Janus coefficient, n=the number of values in the calibration data set, m=the number of i = the value values in the validation data set, Cob i C observed at time i, and mod = the modelled value at time i. ARD was determined by Eq.2: ARD
1 N
| obi mod i | 100% , obi i 1 N
(2)
where ARD=the average relative deviation, N=the number of observations, obi=the value observed at time i, modi=the modelled value at time i.
3 RESULT AND DISCUSSION The dynamic results for the model calibration and validation are shown in Fig.3a–d and Fig.4a–d, respectively, and the results of the statistical evaluations are given in Tables 4 and 5, respectively. The goal of steady state calibration is to obtain average effluent concentrations in the same order of magnitude with measured data, and the correct solids balance to enable good initial conditions for the dynamic simulations. For steady-state simulations,
CHIN. J. OCEANOL. LIMNOL., 0.8 Effluent TP (mg/L)
b
Observed TP Simulated TP
0.6
Observed TN Simulated TN
12 10
0.4 8 0.2
6 14
15
16
17
18 14
15
16
17
18
c
d
Observed MLSS Simulated MLSS
4600
Observed COD Simulated COD
Effluent COD (mg/L)
40
4800 MLSS (mg/L)
Effluent TN (mg/L)
14 a
35
4400
30
4200
25 20
4000 14
15
16
17
18 14
15
16
17
18
Date (day), May 14–18, 2012
Fig.3 Dynamic state graphs for model calibration 0.8
11
Simulated TP Observed TP
10 9 8
0.4
7 6
0.2 1 5600 MLSS (mg/L)
Effluent TN (mg/L)
0.6
12
b
Simulated TP Observed TP
2
3
4
5
6
7
1
c
2
3
4
5
6
7 40
d Simulated TP Observed TP
Simulated TP Observed TP
5400
35
5200
30
5000
25
4800
20
4600
Effluent COD (mg/L)
Effluent TP (mg/L)
a
15 1
2
3
4
5
6
7
1
2
3
4
5
6
7
Date (day), November 1–7, 2012
Fig.4 Dynamic state graphs for model validation
the average influent concentrations for each measured constituent in Tables 1 and 2 were input into the model with the calculated COD fractions in Table 3 and the simulations were performed with the default values of kinetic and stoichiometric parameters of the BioWin® AS model. There were no significant differences between the measured and simulated TP, TN, COD, BOD, NH3-N, and mixed liquor suspended solids (MLSS); consequently, no parameter was calibrated for the steady state simulation apart from the calculated COD fractions shown in Table 3. This underscores the combined influence and importance
of proper wastewater characterization (e.g., COD fractions) and data reconciliation (e.g., SRT) in model prediction. Dynamic calibration was performed using data collected during a 5-day sampling campaign. The results of the simulation from the dynamic calibration revealed differences between measured and simulated effluent TP and MLSS concentrations in the reactor, necessitating calibration of two stoichiometric parameters. Specifically, the value of the heterotrophic yield (YH) was increased from 0.666 to 0.672 to fit the MLSS concentration in the reactor. YH is one of the
Unit
Default value
Calibrated value
YH
mgCOD/mgCOD
0.666
0.672
Yp/acetic
MgP/mgCOD
0.49
0.42
Table 7 Results of simulation of scenarios with 30% increases in flow to the WWTP Variable
S1
S2
S3
S4
TP
0.48
0.6
0.48
0.48
TN
9.7
9.1
9
6.6
COD
29.8
31
30.9
27.4
NH3-N
1.3
1
1.1
0.47
NO3-N
6.8
6
6
4
TSS
9.4
12.5
11.7
9.7
parameters responsible for long-term behaviour (Liwarska et al., 2010). The value of YH was within the values found in the literature, which range from 0.5 to 0.74 (Henze et al., 2002; Liwarska and Biernacki, 2010; Liwarska et al., 2013), indicating that the value used in this study is acceptable. To capture the dynamic trend in phosphorus concentration in the effluent, one parameter associated with PAO (Yp/acetic) was calibrated. Specifically, the value of (Yp/ acetic) was reduced from 0.49 to 0.42. It should be noted that Yp/acetic used in the BioWin® model is equivalent to the yield of phosphorus release to substrate uptake (YPO4), which is used in ASM2, ASM2d, and ASM3P. The default values are 0.4 for ASM2 and ASM2d and 0.35 for ASM3P; therefore, these findings show that the value of 0.42 used in this calibration is comparable and in the range of Yp/acetic used in ASMs (Rieger et al., 2001; Henze et al., 2002). Additionally, the influent composition of the investigated WWTP was similar to the wastewater concentrations used during development of ASM2. It should be noted that BioWin® is an extension of ASM2 (Barker and Dold, 1997; Henze et al., 2002). Consequently, the need for calibration of only two parameters in this study is in agreement with the results of other studies (Meijer et al., 2001; Petersen et al., 2002). The results in the present study confirmed that the investigated WWTP is properly configured hydraulically in the simulator, and that the correct primary data (e.g., SRT, recycle flow rate) were input in the simulator used (Hulsbeek et al., 2002). As shown in Table 4, the predictive ability of the model outside its calibration is quite good, indicating model stability and consistency for predicting all of
TN effluent (mg/L)
Parameter
COD effluent (mg/L)
Table 6 Calibrated and default values of the parameters
TP effluent (mg/L)
OLEYIBLO et al.: Wastewater treatment plant improvement using BioWin
BioWin Chart
0.6 0.45 0.3 0.15 03/11/2012
10/11/2012 17/11/2012 24/11/2012 Date (days) Eff. TP observed Eff. TP simulated BioWin Chart
12 8 4 03/11/2012
10/11/2012 17/11/2012 24/11/2012 Date (days) Eff. TN observed Eff. TN simulated BioWin Chart
30 20 03/11/2012
10/11/2012 17/11/2012 24/11/2012 Date (days) Eff. COD observed Eff. COD simulated
Fig.5 Dynamic simulation for the entire month of November, 2012
the variables with reasonable accuracy. Although the J was slightly greater than 1 (1.2) for the MLSS (Table 4), this value can be considered acceptable given the dynamic nature of the influent compositions, which can affect the MLSS. As shown in Table 5, the ARD evaluation based on application of activated sludge models showed a wide range of model deviations considered acceptable, with relative errors ranging from 10% to 40% (Melcer et al., 2003; Sin et al., 2008). The ARD results for all variables were below 10% for both the calibration and validation period. The model default and calibrated parameter values are given in Table 6. To evaluate the long term predictive ability of the model before its application for analysis of additional scenarios, the dynamic data used to validate the model were applied to simulate the entire month of November, 2012 (Fig.5). The simulation was performed for an additional 8 months (May, 2012–December, 2012) using averages of the sum of each constituent from the dynamic data as the input values (Fig.6). Figures 5 and 6 show the long term predictive ability of the validated model outside its calibration period. The model was able to simulate the WWTP trend with reasonable accuracy and was therefore used to investigate three scenarios. The scenario results shown in Table 7 are based on 30% increased flows and the operating mode of the WWTP. The decision to use 30% increased flows was
CHIN. J. OCEANOL. LIMNOL.,
COD effluent (mg/L)
TN effluent (mg/L)
TP effluent (mg/L)
BioWin Chart 0.6 0.45
Eff. TP observed
0.3
Eff. TP simulated
0.15 11/05/2012 10/06/2012 10/07/2012 09/08/2012 08/09/2012 08/10/2012 07/11/2012 07/12/2012 Date (days) BioWin Chart 16 12
Eff. TN observed
8
Eff. TN simulated
4 11/05/2012 10/06/2012 10/07/2012 09/08/2012 08/09/2012 08/10/2012 07/11/2012 07/12/2012 Date (days) BioWin Chart 40
Eff. COD observed EFF. COD simulated
20 11/05/2012
10/07/2012
08/09/2012 Date (days)
07/11/2012
Fig.6 Simulation results for May, 2012–November, 2012
based on the fact that it is slightly higher than the proposed expansions, and a little higher than the maximum flows recorded in 3 years of historical data available for the investigated WWTP. In scenario S1, one clarifier is assumed to be functioning and sludge is wasted from the aeration tank. It is also assumed that one of the clarifiers is taken out for maintenance. In scenario S2, an equalization tank is installed, the two clarifiers are functioning and sludge waste is from the clarifiers. Scenario S3, represents normal operation, with both clarifiers functioning, and sludge waste from the clarifiers. In scenario S4, one new oxidation ditch train is added to the existing oxidation ditch trains, the two clarifiers are functioning and sludge is wasted from the clarifiers. As shown in Table 7, the results for S4 are better for all cases except TP, where S1 and S3 have the same value (0.48) as S4. Conversely, S2 has the worst TP, COD, and TSS effluent concentrations. However, the differences in the effluent concentrations are negligible when compared with the total cost of building a new oxidation ditch train or an equalization tank. Moreover, there is no effluent violation in S1. S1 also requires only minor modifications (floating control valve) to the sludge conveyance pipes, which can be operated only when necessary. Consequently, S1 is recommended for implementation in the WWTP. This is because wasting sludge from the aeration tank:
1) reduces the risk of clarifier failure due to increasing solids loading during storm events; 2) reduces the risk of secondary phosphorus release in the clarifier due to rising sludge blanket; 3) facilitates maintenance of consistent volumetric sludge age control when wasting from the aeration tank.
4 CONCLUSION The calibrated BioWin® model was successfully used for scenario analysis, and the results revealed that wasting sludge from the aeration tank during wet weather conditions was preferred over the construction of either an equalization tank or a new oxidation ditch train. However, it is recommended that the flexibility to operate in various modes be incorporated into the design of WWTPs to aid in the selection of an appropriate operating mode when necessary, and possibly reduce or eliminate the cost of building additional component(s) in times of extreme conditions.
5 ACKNOWLEDGMENT The authors are grateful to Hohai University for supporting this work. We also thank the staff of Anhui Guozhen Environmental Protection Sci-Tech Co. Ltd., the editors, and three anonymous reviewers for their valuable suggestions.
OLEYIBLO et al.: Wastewater treatment plant improvement using BioWin References APHA. 1998. Standard Methods for the Examination of Water and Wastewater, 20th edn, American Public Health, Washington DC, USA. Barker P S, Dold P L. 1997. General model for biological nutrients removal activated-sludge systems: model presentation. Water Environ. Res., 69(5): 969-984, http:// dx.doi.org/10.2175/106143097X125669 EnviroSim Associates Ltd. 2008. User Manual for BioWin v.3.0, EnviroSim Associates Ltd., Hamilton, Canada. Ferrer J, Seco A, Serralta J, Ribes J, Manga J, Asensi E, Morenilla J J, Llavador F. 2008. DESASS: a software tool for designing, simulating and optimising WWTPs. Environ Modell Softw., 23(1): 19-26, http://dx.doi.org/ 10.1016/j.envsoft.2007.04.005 Gernaey K V, Loosdrecht M C M, Henze M, Lind M, Jorgensen S B. 2004. Activated sludge wastewater treatment plant modelling and simulation: state of the art. Environ. Modell. Softw., 19(9): 763-783. http://dx.doi. org/10.1016/j.envsoft.2003.03.005. Henze M, Guje W, Mino T, Matsuo T, Loosdrech M V. 2002. Activated sludge models ASM1, ASM2, ASM2d and ASM3. IWA Scientific and Technical Report No.9, London, UK. p.45-102 Hulsbeek J J W, Kruit J, Roeleveld P J, van Loosdrecht M C M. 2002. A practical protocol for dynamic modelling of activated sludge systems. Water Sci. Technol., 45(6): 127136. Langergraber G, Rieger L, Winkler S, Alex J, Wiese J, Owerdieck C, Ahnert A, Simon J, Maurer M. 2004. A guideline for simulation studies of wastewater treatment plants. Water Sci. Technol., 50(7): 131-138. Liwarska-Bizukojc E, Biernacki R. 2010. Identification of the most sensitive parameters in the activated sludge model implemented in BioWin software. Bioresour. Technol., 101(19): 7 278-7 285, http://dx.doi.org/10.1016/j.biortech. 2010.04.065. Liwarska-Bizukojc E, Olejnik D, Biernacki R, Ledakowicz S. 2013. Improving the operation of a full-scale wastewater treatment Plant with use of a complex activated sludge model. Environ. Prot. Eng., 39: 659-670, http://dx.doi. org/10.5277/EPE130114. Mamais D, Jenkins D, Pitt P. 1993. A rapid physical-chemical method for the determination for readily biodegradable soluble COD in municipal Wastewater. Water Res., 27(1): 195-197, http://dx.doi.org/10.1016/0043-1354(93)90211-y. Meijer S C F, van der Spoel H, Susanti S, Heijnen J J, van Loosdrecht M C M. 2002. Error diagnostics and data reconciliation for activated sludge modelling using mass
balances. Water Sci. and Technol., 45(6): 145-156. Meijer SCF, van Loosdrecht MCM, Heijnen J J. 2001. Metabolic modelling of full-scale biological nitrogen and phosphorus removing WWTP’s. Water Res., 35(11): 2 711-2 723, http://dx.doi.org/10.1016/S0043-1354(00) 00567-4. Melcer H, Dold P L, Jones R M, Bye C M, Takacs I, Stensel H D, Wilson A W, Sun P, Bury S. 2003. Methods for wastewater characterization in activated sludge modelling. Report No. 99-WWF-3 Water Environment Research Foundation (WERF), Alexandria, VA. USA. Petersen B, Gernaey K V, Henze M, Vanrolleghem P A. 2002. Evaluation of an ASM1 model calibration procedure on a municipal-industrial wastewater treatment Plant. J. Hydroinform., 4: 15-38. Puig S, van Loosdrecht M C M, Colprim J, Meijer S C F. 2008. Data evaluation of full-scale wastewater treatment plants by mass balance. Water Res., 42(18): 4 645-4 655, http:// dx.doi.org/10.1016/j.watres.2008.08.009. Rieger L, Vanrolleghem P A, Takács I, Johnson B R. 2008. Wastewater Treatment Modelling, Quo Vadis? Water 21, IWA Publishing, London, UK. p.59-60. Rieger L, Koch G, Kuhni M, Gujer W, Siegrist H. 2001. The EAWAG BIO-P Module for activated sludge model No.3. Water Res., 35(16): 3 887-3 903. Russel B M, Henriksen J P, Jorgensen S B, Gani R. 2002. Integration of design and control through model analysis. Comput. Chem. Eng., 26(2): 213-225, http://dx.doi.org/ 10.1016/S0098-1354(01)00742-6. Sin G, De Pauw D J W, Weijers S, Vanrolleghem P A. 2008. An efficient approach to automate the manual trial and error calibration of activated sludge models. Biotechnol. Bioeng., 100(3): 516-528, http://dx.doi.org/10.1002/ bit.21769. STOWA. 1996. Methods for influent characterization: inventory and guidelines. Report STOWA 96-08. Utrecht, The Netherlands. (in Dutch) Thomann M. 2008. Quality evaluation methods for wastewater treatment plant data. Water Sci. Technol., 57(10): 1 6011 609, http://dx.doi.org/10.2166/wst.2008.151. Vanrolleghem P A. 2001. Model Simulation for Improved Operation and Control of Wastewater Treatment Plants. BIOMATH. Department for Applied Mathematics, Biometrics and Process Control Ghent University, Coupure Links 653, B-9000 Gent, Belgium. http:// biomath.rug.ac.be/~peter. Water Environment Federation (WEF). 2010. Nutrient Removal—A Manual of Practice No. 34, McGraw-Hill, New York, Water Environment Federation, Alexandra, VA22314-1994 USA.
