D. Brdjanovic*, M. Mithaiwala** , M.S. Moussa*** , G. Amy* and M.C.M. van Loosdrecht**** *Department of Urban Water and Sanitation, UNESCO-IHE Institute for Water Education, Westvest 7, PO Box 3015, 2061 DA Delft, The Netherlands (E-mail:
[email protected]) **Drainage Department, Surat Municipal Corporation, Muglisara, Surat Gujarat 395003, India (Email:
[email protected]) ***Civil Engineering Department, Faculty of Engineering Mataria, Helwan University, Egypt (Email:
[email protected]) ****Department of Biochemical Engineering, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands (Email:
[email protected] ) Abstract This paper presents results of a novel application of coupling the Activated Sludge Model No. 3 (ASM3) and the Anaerobic Digestion Model No.1 (ADM1) to assess a tropical wastewater treatment plant in a developing country (Surat, India). In general, the coupled model was very capable of predicting current plant operation. The model proved to be a useful tool in investigating various scenarios for optimising treatment performance under present conditions and examination of upgrade options to meet stricter and upcoming effluent discharge criteria regarding N removal. It appears that use of plant-wide modelling of wastewater treatment plants is a promising approach towards addressing often complex interactions within the plant itself. It can also create an enabling environment for the implementations of the novel side processes for treatment of nutrient-rich, side-streams (reject water) from sludge treatment. Keywords ADM1; ASM3; developing countries; modelling; tropical conditions; upgrade; wastewater treatment plant
Water Science & Technology Vol 56 No 7 pp 21–31 Q IWA Publishing 2007
Use of modelling for optimization and upgrade of a tropical wastewater treatment plant in a developing country
Introduction
The vast majority of wastewater treatment facilities in developing countries are designed to accomplish the removal of suspended solids (SS) and biodegradable organic matter (BCOD) only. In areas where the treatment plant effluent is reused (e.g. for irrigation), no legal requirements are usually imposed regarding nitrogen and phosphorus removal. However, increasing numbers of developing countries are currently strengthening their environmental regulations, which in many cases implicitly means an upgrade of their existing treatment facilities or construction of new plants equipped with technologies for nutrient removal. It is noteworthy that, under tropical conditions, high wastewater temperature enhances kinetics of biological conversion processes. Combined with the fact that some wastewater treatment plants have available sufficient reactor volume and aeration capacity, although originally designed for suspended solids and organic matter removal only, such plants can exhibit partial or even full nitrogen removal (nitrification and denitrification). In addition to strengthening effluent discharge standards, increasing attention is being paid to the management of wastewater treatment residuals (sludge). Also in developing countries, especially in those with a tropical climate, larger activated sludge plants usually include anaerobic sludge digestion. In general, the overall performance of a treatment plant depends on the efficiency of both wastewater and sludge treatment, where their interaction often plays an important role. Therefore, it is interesting from a practical doi: 10.2166/wst.2007.675
21
D. Brdjanovic et al.
as well as a modelling perspective to further investigate interactions between wastewater treatment and sludge treatment, such as (a) effects of changing return sludge flow or/and sludge wasting rate, (b) impact of plant upgrade by nitrogen removal facilities on anaerobic sludge digestion, and (c) influence on high strength return flows from sludge treatment on the activated sludge system. As the use of mathematical models for simulation of wastewater treatment plant operations is becoming increasingly popular, a promising approach to investigate such interactions is the application of plant-wide modelling by coupling activated sludge and anaerobic digestion models. So far, the successful practical application of such approach, especially involving plants operating under tropical conditions in developing countries, has not yet been reported. In this study, the wastewater treatment plant Anjana of the city of Surat in India was subjected to modelling by using a combination of Activated Sludge Model No.3 – ASM3 (Gujer et al., 1999) and Anaerobic Digestion Model No.1 – ADM1 (Batstone et al., 2002). The main goals of this modelling study were to (1) optimise plant performance regarding the effluent quality and biogas production, (2) integrate ASM3 and ADM1, (3) carry out the application of such a combined model on a full-scale installation operating under tropical conditions and in a developing country, and (4) the modelling-aided investigation of operational, technological and infrastructural requirements to meet new effluent requirements concerning nitrogen removal. Since the first three goals have already been addressed by Brdjanovic et al., (submitted), this paper presents findings of the assessment of the usefulness of the coupled model for the evaluation of different scenarios for the optimisation of treatment performance under present conditions, and for upgrade options to meet stricter and upcoming effluent discharge criteria.
Materials and methods
22
The plant Anjana is situated on the bank of the River Tapti in the city of Surat in India, 22 km inland from the Arabian Sea in the southern part of Gujarat State. The plant is the oldest of the six municipal wastewater treatment plants of Surat. The local climate is characterised by hot summers (ambient temperatures from 37 to 44 8C) and mild winters (10 to 16 8C). The raw sewage temperature is in the range of 23 to 35 8C with an annual average of approximately 30 8C. The average annual rainfall of 1.143 mm is unevenly distributed over the year due to monsoon seasons. The sewerage network in the city gravitating to Anjana plant is a separate sanitary sewer, covering an area of 10 km2 and currently servicing approximately 200,000 people and the associated industry (mostly dyeing, printing and weaving industries). The sanitary sewage is transported to the treatment plant by four pumping stations, while storm water is collected and taken away by a separate system and discharged untreated. The plant was constructed in 1958 with a design capacity of 25,000 m3/d with only primary treatment, with the settled sewage used for irrigation. In 1981 the plant capacity was extended to 76,000 m3/d. Meanwhile, due to rapid expansion and the urbanisation of agricultural land, the local authorities decided to dispose of the plant effluent (settled sewage) to nearby Koyli Creek, which ultimately discharges into the nearby Arabian Sea. To meet the effluent disposal requirements regarding organic matter, the plant was upgraded in 1996 by biological sewage treatment facilities (activated sludge system, Figure 1) with a capacity of 82,500 m3/d, and by facilities for anaerobic sludge digestion. The local effluent disposal standards at present are: BOD5 # 20 mg/L, SS # 30 mg/L and COD # 100 mg/L (GPCB, 1999). The expected limit for total nitrogen concentration in the effluent, to be implemented in the near future, is 10 mg/L, while introduction of effluent legislation regarding phosphorus is currently under consideration.
D. Brdjanovic et al.
Figure 1 Wastewater treatment plant in Anjana
Sewage treatment is based on a conventional configuration containing screens (two), grit chambers (two), primary settlers (two), biological tanks with surface aeration (12), and final clarifiers (three). Most of the settled sludge is returned to the beginning of the aeration tank while the desired sludge retention time (SRT) is controlled by wasting excess sludge from final clarifiers (Figure 2). Sludge treatment consists of a conventional anaerobic sludge digestion process comprising sludge thickeners (three), digesters alternately fed by primary sludge and thickened waste activated sludge (three), a biogas collection tank (one), and sludge drying beds. Supernatants from thickeners and digesters, as well as the filtrate from sludge drying beds, is returned to the beginning of the plant and mixed with raw sewage. The plant is
Figure 2 Simplified plant process schematic
23
D. Brdjanovic et al.
also equipped with an H2S scrubber. The biogas-based power plant (capacity 0.5 MW) was installed in 2004 and includes CH4 and H2S analysers. Currently, one secondary clarifier serves as a stand-by unit, sludge digesters are operated as low-rate units without mixing arrangements (despite provision of agitators for sludge mixing) and the biogas plant is generating an electric power of 0.2 MW. The main reason for this is the fact that the plant current hydraulic load is approximately 50% of its design capacity. During the investigation period the weather conditions were favourable (dry weather with sewage and air temperature about 30 8C). The routine monitoring of plant performance includes three locations only and six parameters. An additional 2-week sampling programme was carried out and included nine parameters for wastewater characterisation and two for sludge characterisation at existing sampling locations (Table 1). The sampling points, the sampling frequency and the choice of the parameters were governed by the specific requirements of the mathematical models applied to this plant and by the content of existing records. The analyses were performed according to Standard Methods for Examination of Water and Wastewater (APHA, 1996). The information obtained through the sampling program was combined with the data routinely collected by the staff of the plant. In addition, the influent flow records were used together with the internal flow rates obtained from the records and on-site calibration of the pumps. The methane content of the biogas was measured by an on-line analyser. For evaluation of different upgrading alternatives regarding optimisation of nitrogen removal and methane generation under present and future conditions, the combined ASM3-ADM1 model was used. Information concerning the modelling approach and methodology, wastewater and sludge characterisation, model development, coupling, calibration and validation do not fall under the scope of this paper and are presented elsewhere (Brdjanovic et al., submitted).
Results
Besides an assessment of the current situation (average hydraulic load of 40,000 m3/d), two scenarios were identified by the local sewage corporation based on medium and high flow increase prognosis for the year 2010 (60,000 m3/d) and 2020 (80,000 m3/d), respectively. For the current situation (base case) as well as for each of the two scenarios, a number of actions and adjustments were identified and they were subject to simulation using the combined model (Table 2). The approach was to first check how the existing Table 1 Averaged data obtained from the sampling program Raw sewage
Load P.E. Flow m3/d COD tot mg/L COD filt mg/L BOD5 mg/L TKN mg/L NH þ 4 mg/L NO 2 3 mg/L NO 2 2 mg/L TSS mg/L VSS mg/L T 8C DO mg/L pH 24
270,000 38,140 853 381 43.2 21.9 1.7 0.1 781 550 29.5 0.87 7.2
Settled sewage
Final effluent
357 156 136 37.2 21.8 1.9 0.1 142 104 29.5 1.81 7.2
78 55 17 18.8 14.7 1.8 0.7 29 16 27.7 6.03 7.6
Table 2 Selection of applied actions and measures regarding the current situation and two development scenarios for Anjana plant A. Base case: Current situation (Q 5 40,000 m3/d)
B. Scenario 1: Medium flow increase prognosis (Q 5 60,000 m3/d)
C. Scenario 2: High flow increase prognosis (Q 5 80,000 m3/d)
A.1. Alteration in return activated sludge flow (RAS) and waste activated sludge flow (WAS) A.1.1. Existing situation A.1.2. Increased RAS from current 50 to 75% expressed as percentage of the influent flow A.1.3. Increased RAS from 50 to 100% A.1.4. RAS at 50% and reduced WAS by 13 A.1.5. Increased RAS to 50 to 75% and reduced WAS by A.1.6. Increased RAS to 50 to 75% and reduced WAS by
B.1. Increased COD concentration to biological tank B.1.1. Increased COD load by 13
C.1. Increasing primary sludge withdrawal rate to optimize methane production C.1.1. Increased primary sludge withdrawal rate by100% C.1.2. Increased recirculation from 50 to 100%, increased primary sludge withdrawal rate by 100% and reduced WAS by 13
A.2. Reduction in sludge withdrawal from the digester to optimize the methane production A.2.1. Increased RAS to 50 to 75%, reduced WAS by 13 and reduced sludge withdrawal from the digester by 25%
B.2. Increasing primary sludge withdrawal rate to optimize methane production B.2.1. Increased primary sludge withdrawal rate by 50% B.2.2. Increased primary sludge withdrawal rate by 50% and reduced WAS by 13 B.2.3. Increased primary sludge withdrawal rate by 50%, reduced WAS by 13, and reduced sludge withdrawal from the digester by 50%
C.2. Adjustment of DO level in the biological tank from current 0.20 to 0.80 mgO2/L C.2.1. Increased DO level in the biological tank from current 0.20 to 0.80 mgO2/L
A.3. Adjustment of DO level in the entire biological tank from current 0.3 to 1.0 mgO2/L A.3.1. Reduced WAS by 13 A.3.2.Reduced WAS by 13 and one digester shut down A.3.3.Reduced WAS by 13 and two digesters shut down
B.3. Adjustment of DO level in the biological tank from current 0.15 to 0.90 mgO2/L B.3.1. Increased primary sludge withdrawal rate by 50%, and reduced WAS by 13
C.3. Creation of anoxic phase for N-removal upgrade in the last quarter of exiting biological tank C.3.1. Creation of anoxic zone in the last quarter of the existing biological tank C.3.2. Creation of anoxic zone in the last quarter of the existing biological tank, increased COD load by 13, and increased primary sludge withdrawal rate by 65%
1 3 2 3
D. Brdjanovic et al.
25
D. Brdjanovic et al.
26
Table 2 (continued) A. Base case: Current situation (Q 5 40,000 m3/d)
B. Scenario 1: Medium flow increase prognosis (Q 5 60,000 m3/d)
C. Scenario 2: High flow increase prognosis (Q 5 80,000 m3/d)
A.4. Adjusting DO level in the second half of biological tank from current 0.3 to 1.0 mgO2/L A.4.1. Partial DO increase A.4.2. Partial DO increase and reduced excess sludge by 13
B.4. Creation of anoxic phase for N-removal upgrade in the last quarter of exiting biological tank B.4.1. Increased primary sludge withdrawal rate by 50%, and reduced WAS by 13 B.4.2. Increased recirculation from 50 to 100%, increased primary sludge withdrawal rate by 50%, and reduced WAS by 13
C.4. Introduction of additional anoxic volume C.4.1. Introduction of additional anoxic volume after existing biological tank C.4.2. Introduction of additional anoxic volume after existing biological tank, reduced RAS from 100% to 50% C.4.3. Introduction of additional anoxic volume after existing biological tank, increased COD concentration by 25%, and increased primary sludge withdrawal rate by 60% C.4.4. Introduction of additional anoxic volume before existing biological tank and increased internal recirculation 200%
A.5. Creation of anoxic phase for N-removal upgrade within the existing biological tank A.5.1. Partial DO increase, reduced WAS by 50% and creation of anoxic zone in the last quarter of the existing biological tank A.5.2. Partial DO increase, reduced WAS by 50%, creation of anoxic zone in the first quarter of the existing biological tank, and internal recirculation of 300%
B.5. Creation of anoxic phase for N-removal upgrade in the last quarter of exiting aeration tank with increased COD concentration B.5.1. Increased RAS from 50 to 100%, increased primary sludge withdrawal rate by 50%, and reduced WAS by 13, at current COD concentration B.5.2. Increased RAS from 50 to 100%, increased primary sludge withdrawal rate by 25%, reduced WAS by 13 and increased COD concentration by 13 B.5.3. Increased RAS from 50 to 100%, increased primary sludge withdrawal rate by 10%, reduced WAS by 13 and increased COD concentration by 50% B.6. Creation of anoxic phase for N-removal upgrade in the first quarter of exiting biological tank B.6.1. Increased RAS from 50 to 100%, increased primary sludge withdrawal rate by 50%, reduced WAS by 13 and increased internal recirculation 250%
plant operation could be improved with relatively simple operational adjustments (alternations of flow rates of internal streams, adjustment of oxygenation, creation of anoxic zones, etc.), and then to examine more substantial measures towards the plant upgrade by construction of additional treatment units. The results of simulations are presented in Table 3.
The coupled model was very capable of predicting the existing situation (A.1.1). For example, the predicted effluent ammonia and nitrate concentrations were 14.8 and 1.8 mg N/L, respectively, against average measured concentrations of 14.7 and 1.8 mg N/L. The same conclusion can be drawn based on a comparison of predicted and measured methane production. Simulation results showed that by simple manipulation of return activated sludge (RAS) and waste activated sludge (WAS) flow rates, better N removal could be achieved (A.1.6: 8.7 g NH4-N/m3 and 2.9 g NO3-N/m3) with only a marginal effect on methane production (reduction from 3,097 to 3,013 m3/d). Increased aeration resulted in improved nitrification (A.3.1: 2.7 g NH4-N/m3) while overall N-removal deteriorated (22.9 g NO3-N/m3). A partial increase in aeration in the first half of the biological volume resulted in improved N removal (A.4.2: 3.8 g NH4-N/m3 and 12.8 g NO3-N/m3). Conversion of one quarter (6.562 m3) of the biological volume from aerobic to anoxic resulted in improved N removal efficiency, especially in the case of pre-denitrification (A.5.2.: 3.6 g NH4-N/m3 and 6.1 gNO3-N/m3, methane production unchanged). According to coupled model predictions, these adjustments did not affect methane production significantly (variations within 3%). However, taking one or two digesters out of operation resulted in a decrease in methane production by 4 and 15%, respectively.
D. Brdjanovic et al.
A. Base case: Current situation (Q 5 40,000 m3/d)
B. Scenario 1: Medium flow increase prognosis (Q 5 60,000 m3/d)
The simulation with increased COD load to the activated sludge system by one third, suggested that the existing plant configuration will result in poor N removal efficiency (B.1.1: 31 g NH4-N/m3) and increased methane production (22%). Assuming that the currently oversized primary settling tank could retain its present efficiency regarding COD removal, further simulations with alternations in RAS and WAS flow rates only, indicated that overall N removal efficiency could be approximately equal to that of the present situation, while methane production could be increased by 47% (B.2). An increase of aeration over the entire biological tank improved nitrification, as expected; however, denitrification was absent (B.3.1). Introduction of post-denitrification within the existing biological volume (last quarter), improved overall N removal (B.5.1), however, the new N removal target of total N # 10 mg/L was not achieved assuming the present efficiency of primary clarifiers. Further simulations indicated that increased COD load to the activated sludge system would improve denitrification and satisfactory N removal could be achieved (B.5.2 and B.5.3). The comparative option with pre-denitrification appeared to be the better solution: the results showed similar N removal efficiency (effluent ammonia and nitrate concentration of 4.6 and 5.3 g N/m3, respectively); however, methane production was increased by 15 to 24%. C. Scenario 2: High flow increase prognosis (Q 5 80,000 m3/d)
In general, analogous results were obtained in comparison with scenario 1, however, different combinations of adjustments of RAS, WAS, increase in DO, etc., within the existing plant configuration, still could not bring the total N effluent concentration below 10 mg/L. Simulations indicated that with the existing efficiency of primary settlers, introduction of post-denitrification within the existing volume would not bring total N
27
D. Brdjanovic et al.
28
Table 3 Summary of simulations of selected actions and measures regarding the current situation and two development scenarios for Anjana plant Parameter a)
COD sol b)
NHþ 4
NO2 3
DO c)
O2-Net
SRT d)
MLSS
COD sludge
CH 4
Gasgen
CH 4gen.
g/m 3
gN/m 3
gN/m 3
gO2/m 3
gO2/m 3
d
g/m 3
kg/m 3
%
m 3/d
m 3/d
141 127 114 152 134 142 134 201 201 201 173 181 170 171
7.3 8.6 9.6 10.4 12.2 21.1 12.2 10.3 10.3 10.3 7.3 10.3 12.9 13
1419 1539 1598 1761 1885 2536 1885 1726 1726 1726 1425 1726 1983 2003
63.9 63.1 62.5 62.2 61.4 59.2 78.6 62.1 62.4 64.6 63.7 62.1 61.2 61.2
67.7 67.9 68.0 66.2 66.3 64.8 66.3 66.5 66.7 67.4 68.1 66.5 65.5 65.5
4572 4505 4457 4670 4602 4654 4602 4639 4458 4015 4540 4639 4696 4699
3097 3058 3031 3090 3051 3013 3051 3084 2974 2705 3090 3084 3074 3077
168 128 138 138 200 160 123 123 150 162 128
7.8 7.3 10.5 10.5 10.4 10.2 14.1 12.5 13.5 13.9 15.0
3612 2050 2519 2519 2342 2329 2509 2509 3320 3690 2790
77.8 98.9 96.7 123.8 96.0 96.1 93.5 93.49 78.56 72.49 92.3
65.6 66.0 66.4 66.4 66.7 66.6 66.8 66.8 67.1 67.3 67.0
5778 7062 6862 6862 6778 6801 6585 6585 5606 5198 6483
3792 4663 4557 4557 4522 4526 4400 4401 3763 3498 4344 Table 3 (continued )
Scenario
A. Base case: Current situation (Q ¼ 40,000 m3/d) A.1.1. 49 14.5 1.8 0.30 A.1.2. 49 13.4 1.9 0.30 A.1.3. 49 12.9 1.9 0.29 A.1.4. 49 11.6 2.4 0.29 A.1.5. 49 11.0 2.4 0.28 A.1.6. 49 8.7 2.9 0.27 A.2.1. 49 11.0 2.4 0.28 A.3.1. 49 2.7 22.9 1.00 A.3.2. 49 2.7 22.9 1.00 A.3.3. 49 2.7 22.9 1.00 A.4.1. 49 5.0 12.3 0.26/1.00 A.4.2. 49 3.8 12.8 0.25/1.00 A.5.1. 49 7.0 7.7 0.95/Anox A.5.2. 49 3.6 6.1 Anox/0.95 B. Scenario 1: Medium flow increase prognosis (Q ¼ 60,000 m3/d) B.1.1. 71 31.0 0.02 0.15 B.2.1. 49 20.5 0.3 0.23 B.2.2. 49 17.5 0.2 0.20 B.2.3. 49 17.5 0.2 0.20 B.3.1. 49 3.1 21.0 0.90 B.4.1. 49 9.7 4.2 0.81/Anox B.4.2. 49 8.5 5.0 0.80 B.5.1. 49 8.5 5.0 0.80 B.5.2. 65 8.3 1.2 0.69 B.5.3. 71 8.1 0.6 0.64 B.6.1. 49 4.6 5.3 Anox/0.78
Table 3 (continued) Parameter a)
COD sol b)
NHþ 4
NO2 3
DO c)
O2-Net
SRT d)
MLSS
COD sludge
CH 4
Gasgen
CH 4gen.
g/m 3
gN/m 3
gN/m 3
gO2/m 3
gO2/m 3
d
g/m 3
kg/m 3
%
m 3/d
m 3/d
112 92 145 114 141 129 167 204 172
6.6 12.5 11.9 11.8 13.0 14.5 11.0 11.8 11.0
2700 3428 2969 2977 4010 2800 2688 3545 2710
65.3 65.9 66.2 66.1 66.2 66.2 65.9 66.0 65.9
9336 8741 8610 8634 7377 8587 8858 7660 8863
6096 5762 5699 5704 4885 5681 5839 5052 5840
Scenario
C. Scenario 2: High flow increase prognosis (Q ¼ 80,000 m3/d) C.1.1. 49 26.4 0.1 0.20 C.1.2. 49 23.3 0.1 0.14 C.2.1. 49 4.0 18.6 0.80 C.3.1. 49 11.1 2.2 0.68/Anox C.3.2. 65 10.9 0.3 0.53/Anox C.4.1. 49 7.3 7.8 0.80 C.4.2. 49 8.0 7.3 0.80 C.4.3. 65 8.2 2.3 0.80 C.4.4. 49 4.7 6.5 Anox þ 0.80
132.9 80.5 79.6 79.7 66.7 79.4 81.5 69.0 81.6
a) Water quality parameters are related to plant effluent b) In all simulations COD content of the settled sewage was used. Settled sewage COD concentration obtained from the sampling programme (357 mg/L) was used in most of cases assuming that the efficiency of primary settling tanks remain unchanged as result of existing spare capacity within these units. In a few cases (B.1.1., B.5.2., B.5.3., C.3.2. and C.4.3.) the process performance was checked with comparatively higher COD concentrations in settled sewage (33 and 50% increase). c) Since aeration capacity was designed for the organic load associated with influent flow rate of 82,500 m3/d, the need for aeration supplement in both base case and future scenarios still need to be verified. d) SRT was calculated taking into account biomass present in the biological tanks, final settlers, plant effluent and excess sludge stream.
D. Brdjanovic et al.
29
D. Brdjanovic et al. 30
concentration in the effluent to a satisfactory level (C.3.1). Reduced settling efficiency would help N removal (C.3.2: 10.9 g NH4-N/m3 and 0.3 g NO3-N/m3) on the expense of methane production (reduction by 17%). The best option appeared to be construction of an additional anoxic volume (6.562 m3) in front of the existing aeration tanks which showed the best balance between the effluent quality requirements and methane production (C.4.4: 4.7 g NH4-N/m3 and 6.5 g NO3-N/m3, marginal influence on methane production). Discussion
Under the present load and current regulations in place, the performance of wastewater treatment plant Anjana is, in general, assessed to be satisfactory. Although the plant was originally designed for organic matter removal only, the availability of spare capacity (plant is currently approximately 50% oversized), combined with high sewage temperature, allow for partial nitrification and denitrification. However, partial N removal led to higher oxygen demand, resulting in lower DO levels in the aeration tanks. The operating sludge retention time (SRT) was calculated to be at the low end of that required for nitrification (approximately 7 days). The decreased WAS flow rate (increase in SRT) generally showed improvement in both nitrification and denitrification, while the increased RAS flow rate did not result in major improvement of the plant performance. In addition, the likely presence of “dead” zones within the aeration tank and availability of COD allowed for limited denitrification. Concerning the existing situation, it was shown that limited improvement in N removal efficiency could be achieved by manipulation of WAS and by step-wise increasing the DO level in the aeration tanks. However, future effluent discharge standards seem to be best achievable by introduction of pre-denitrification by creation of an anoxic zone within the existing biological tanks. It is important to note that the surface aeration (turbines) capacity to support N removal still needs to be assessed. Given that the system is designed for a COD load of twice that currently taking place, it is likely that a certain increase in DO concentration can be accommodated as well. Further investigation into the current way of operating aerators might bring more possibilities for increasing air input into the biological tanks. Requirements for additional N removal are, as expected, conflicting with the goal to maximise energy recovery by methane production. Since upgrade of the system for N removal requires longer SRT and results in less WAS flow, this will consequently lead to a lowering of methane production. Within the investigated range of sludge age from 7 to 21 days, methane production dropped by a maximum of 14%, which was considered acceptable. It was also shown that, due to spare capacity, taking one digester out of operation will have no impact on anaerobic sludge digestion and taking two digesters out of operation will reduce methane production for only 15%, due to present overcapacity. In both future scenarios, the model predicted satisfactory COD removal. This was to be expected due to the fact that the plant extension from 1996 provided for hydraulic capacity of the activated sludge system of 82,500 m3/d. The investigation of the performance of the activated sludge plant under most unfavourable conditions (increased flow, reduced efficiency of primary settlers) still resulted in satisfactory COD removal, however, nitrification did not occur. This strongly suggested that the efficiency of the primary clarifier should be carefully considered. Similar to the present case, an upgrade of the plant for N removal required the introduction of pre-denitrification. Also, the other major findings are analogous to those of the present case. For the maximum plant capacity desired, the targeted N removal was only possible to achieve by construction of an additional anoxic tank in front of the the existing aerobic tanks. The additional aeration capacity required for required N removal in the future still has to be investigated
in more detail. Overall, the coupled model indicated that the upgrade of plant for N removal would have a marginal negative influence on methane production. Conclusions
D. Brdjanovic et al.
In general, the coupled model was very capable of predicting current plant operation. The model proved to be a useful tool in investigating various scenarios for optimising treatment performance under present conditions, and examination of upgrade options to meet stricter and upcoming effluent discharge criteria. Although modelling aspects were not elaborated in this paper, it is important to report that (i) the Dutch wastewater characterization (STOWA) procedure proved capable of successful application under conditions of a tropical developing country such as in Surat India, (ii) the applied procedure for primary and secondary sludge fractionation for ADM1 appears to be sufficient for obtaining good simulation results, (iii) application of a simple calibration procedure resulted in adjustment of only five default values, three for ASM3 and two for ADM1, (iv) the need for standardisation of parameters in both models became obvious, and (v) tropical conditions clearly favour process performances. It appears that the use of plant-wide modelling for optimisation and upgrade of wastewater treatment plants is a promising approach towards integrated modelling. It can also create an enabling environment for the implementation of novel side processes for the treatment of nutrient-rich, side-streams (reject water) from the sludge treatment (such as SHARON, ANAMMOX or BABE). It is noteworthy that the Surat Municipal Corporation, based on the results of this research study, is currently finalising the model-aided design documentation for plant upgrade for N removal (pre-denitrification, see C.4.4.). The construction works at Anjana plant are planned to be completed in the summer of 2007.
References APHA, AWWA, WEF (1996). Standard Methods for the Examination of Water and Wastewater, 19th edn, American Public Health Association, Washington DC. Batstone, D.J., Keller, J., Angelidaki, R.I., Kalyuzhnyi, S.V., Pavlostathis, S., Rozzi, A., Sanders, W.T.M., Siegrist, H. and Vavilin, V.A. (2002). Anaerobic Digestion Model No. 1, IWA Publishing, London, ISBN: 1 900222 78 7. Brdjanovic, D., Moussa, S.M., Mithaiwala, M., Amy, G. and Van Loosdrecht, M.C.M. Plant wide modeling of a tropical wastewater treatment plant in developing country. Water Research (submitted). GPCB (1999). GPCB Norms, Gujarat Pollution Control Board, Gandhinagar (India). Gujer, W., Henze, M., Mino, T. and van Loosdrecht, M. (1999). Activated Sludge Model No. 3. Water Sci. Technol., 39(1), 183 – 193.
31
