model-based optimum design of sequencing batch reactors for cod ...

WEFTEC®.06

MODEL-BASED OPTIMUM DESIGN OF SEQUENCING BATCH REACTORS FOR COD AND NITROGEN REMOVAL FROM A SLAUGHTERHOUSE WASTEWATER Hank Andres, Zhi-rong Hu, Spencer Snowling, Oliver Schraa Hydromantis, Inc. 1685 Main St. W., Suite 302, Hamilton, Ontario, Canada, L8S1G5 ABSTRACT A dynamic model of the activated sludge process was used to analyze and optimize the operation of an SBR treating slaughterhouse wastewater. The existing treatment cycle (duration of fill, aeration, mix, decanting and wasting periods) was found to be inadequate for meeting effluent requirements under a number of different loading scenarios. Modelling analysis indicated that the aeration phase was too long and the settling phase too short. Simulation of a new SBR cycle operation, in which the superfluous time in the aeration phase was distributed to the settling phase and a new anoxic phase, confirmed that the unit could meet the stringent effluent requirements. Using an iterative approach, optimal cycle settings were determined for each of the loading and temperature scenarios investigated. KEYWORDS SBR, slaughterhouse, model, simulation, process optimization, nitrogen, COD INTRODUCTION The Sequencing Batch Reactor (SBR) is an activated sludge process designed to accommodate both biological reactions and solid–liquid separation in a time sequence in the same tank. Currently, sequencing batch reactor (SBR) technology is a well-promoted and tested alternative, which has a relatively low cost and small footprint. The SBR process offers flexibility of operation, where the sequence of successive phases can be adjusted to create the required combination of the growth conditions for different groups of microorganisms to remove different contaminants from wastewater, i.e.: • • •

aerobic for COD removal only, aerobic/anoxic for COD/nitrogen removal and aerobic/anoxic/anaerobic for COD/nitrogen/phosphorus removal.

The number of biological processes and components, together with the complexity of SBR hydraulics, can make it very difficult to evaluate and optimize the performance solely based on experience and steady state analysis. A dynamic mathematical model is an extremely useful tool for analyzing complex processes. Mathematical simulation models provide quantitative descriptions of the dynamic behavior of the system, providing predictions of the system response and performance under various operating conditions. From these predictions, design and operational parameters can be identified and optimized to maximize system performance (Hu 2001). Copyright ©2006 Water Environment Foundation. All Rights Reserved

5100

WEFTEC®.06

For this purpose, a number of mathematical activated sludge models have been developed over the last two decades. These models fall into two general categories: 1) those that model COD removal/nitrification/denitrification (e.g. UCTOLD, Dold et al., 1991; ASM1, Henze et al., 1987), and 2) those that model COD removal/nitrification/denitrification/biological P removal (e.g. UCTPHO, Wentzel et al., 1992; ASM2d, Henze et al., 1999). These models have been used extensively over the past decade to analyze and optimize activated sludge plants of various types. This paper will focus on the case study of one SBR plant where model-based analysis was used to optimize process operation for optimal treatment. MODELING METHODOLOGY A case study is presented in this paper to demonstrate how a simulation model can be used to evaluate and optimize the performance for COD and nitrogen removal of a SBR plant designed using a conventional experience-based approach. For this purpose, a simulation model for a specific SBR plant was first developed by using the IWA ASM1 model (Henze et al., 1987), and implemented in GPS-X™ simulator (Hydromantis, 2003). The developed simulation model was then used to evaluate and optimize the SBR plant by simulating the SBR performance under various operating conditions against the design criteria. SBR Plant The SBR plant being studied was originally analyzed by the University of the Basque Country (Unai Iriarte, 2003), and treats a wastewater stream from a commercial slaughterhouse operation. The influent characteristics for the SBR plant are shown in Table 1. Table 2 summarizes the physical design parameters of the treatment plant. Table 1: SBR Influent Characteristics Influent Parameter BOD5 TSS TKN NO3 Total Average Influent Flow Total Average Peak Flow

Value 147 mg/L 98 mg/L 34 mgN/L 0 mgN/L 30,000 m3/d 49,000 m3/d

(7.9 MGD) (12.9 MGD)

Table 2: SBR Physical Design Parameters Design Parameter Trains Volume of Each Reactor Total Reactor Volume Water Level

Value 4 5,700 m3 (1.5 MGal) 22,800 m3 (6.0 MGal) 5.13m – 6.5m (low/high water levels)

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5101

WEFTEC®.06

The original cycle settings are shown in the center of Table 3. The design objectives for effluent and maximum MLSS in the reactor for the SBR plant are shown in Table 4. Table 3: Cycle Settings for the Designed and Optimal SBR Process SBR Phases

Phase length (minutes)

filling and mixing filling and aeration Aeration only Mix only (anoxic) Settling Decant Desludge Total cycle length

Filling

45 45 135 55 60 20 360

Table 4: Desired Effluent Quality and MLSS Parameter

Design Objectives (mg/L)

Effluent CBOD5 15 Effluent TSS 20 Effluent NH4 3 Effluent NO3 15, 10, 5 * Maximum MLSS 3000 * Effluent standard under 100C (winter), 130C (average) and 200C (summer) conditions.

The SBR plant model was developed by using the SBR unit process object in the GPS-X simulation platform. The hydraulic operation combines the CSTR and secondary clarifier together. The SBR model is sectioned into a number of layers, each with its own volume, the sum of which will produce the total volume. The layout of the SBR plant model used in this project includes an influent object, a four-way splitter (to direct the flow to each of the SBR trains), four SBR units in parallel, and a holding tank to collect the decant effluent from the SBRs (see Figure 1). It operates under four 6-hour cycles per day. The ASM1 model was used for biological processes and the BOD-based influent model for influent characterization. Default values for all kinetic and stoichiometric parameters in the ASM1 model (Henze et al., 1987) were used for the simulations.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5102

WEFTEC®.06

Figure 1: The layout of the SBR plant model RESULTS AND DISCUSSION The model was used to evaluate whether the SBR plant would meet the effluent quality and the reactor mixed liquor suspended solids (MLSS) requirement (less than 3000 mg/L) with the original cycle settings. This evaluation was conducted at 3 different temperatures (10oC/50 oF, 13oC/55 oF, and 20oC/68 oF) for both the “average” (30,000 m3/d, 7.9 MGD), and “peak” (49,000 m3/d, 12.9 MGD) flow rates. In cases where the performance requirements were not met, the model was used to adjust and optimize the cycle setting to meet the effluent and reactor MLSS requirements. Evaluation of Original SBR Design Cycles Six different simulation scenarios were evaluated using the original SBR cycle settings. The influent characteristics (TSS, TKN, BOD5, etc.) were kept the same for all simulations. The simulation results are shown in Table 5, alongside the design objectives.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5103

WEFTEC®.06

Table 5: Predicted Effluent Quality and MLSS for the Original SBR Design Max. MLSS mg/L Design objectives 15 20 3 15 10 5 3000 Scenario 1 T =10 0C; Average Flow 4.1 24.6 0.04 12.6 2400 Scenario 2 T =10 0C; Peak Flow 7.6 34.1 0.06 12.4 1800 Scenario 3 T =13 0C; Average Flow 3.8 23.9 0.04 12.7 2300 Scenario 4 T =13 0C; Peak Flow 7.2 33.6 0.05 12.4 1700 Scenario 5 T =20 0C; Average Flow 3.4 23.7 0.04 12.5 2100 Scenario 6 T =20 0C; Peak Flow 6.3 32.9 0.05 12.6 1600 * Effluent standard under 100C/50 oF, (winter), 130C/55 oF (average) and 200C/68 oF (summer) conditions, respectively. CBOD5 mg/L

Parameters

TSS mg/L

NH4 mg/L

NO3 mg/L*

The simulations based on the original SBR cycle settings showed that the maximum reactor MLSS concentration (measured during the aeration phase) was below 3000 mg/L for all six simulation scenarios, but that the effluent quality did not always meet the design objectives. At a temperature of 10oC, the effluent requirements were met for BOD5, ammonia, and NO3, but the effluent TSS concentrations exceeded the design objectives. At temperatures of 13oC and 20oC, the effluent requirements were met for BOD5 and ammonia, but not for TSS or NO3. SBR Cycle Optimization An extensive simulation study was performed to optimize the SBR cycle settings, with the goal of meeting all design objectives. This was carried out as a sensitivity analysis, by increasing and decreasing the duration of each phase of the SBR cycle, and observing the resultant impact on effluent quality. The sensitivity study indicated that the higher-than-desired effluent TSS concentrations were due to insufficient settling time during one SBR operational cycle. The sensitivity analysis also found that the total aeration time of 180 minutes was longer than necessary, given the loading of COD and nitrogen to the reactor. To illustrate this, the oxygen uptake rate (OUR) was plotted versus time for the duration of the aeration phase. Figure 2 shows the results for the average influent flow and 20oC scenario. The graph indicates that the OUR decreases substantially after the first 50 minutes of the total 180 minutes of aeration time. This information was used to optimize the length of the aeration phase.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5104

WEFTEC®.06

Oxygen Uptake Rate During Aeration Phase (Scenario 3) 90

3

OUR (gO2/m /h)

80 70 60 50 40 30 20 10 0 0

25

50

75

100

Time (minutes)

Figure 2: Oxygen uptake rate during aeration Phase –Original cycle setting Lastly, it was determined that an anoxic phase could be added following the aeration phase to promote denitrification, thereby reducing the effluent NO3 concentration. By implementing changes based on information from the sensitivity analysis, the cycle settings were adjusted to meet the effluent requirements in all six scenarios. The optimized cycle settings are shown in Table 6. The simulation results with the optimal cycle settings are shown in Table 7, alongside the design objectives. All design objectives are met under all scenarios with the exception of the effluent NO3 at peak loading and 20ºC/68 ºF. Table 6: Cycle Settings for the Designed and Optimal SBR Process SBR Phases Filling

filling and mixing filling and aeration Aeration only Mix only (anoxic) Settling Decant Desludge Total cycle length

Phase length (minutes) optimal cycle setting for each scenario Originally Designed cycle settings for all scenarios 1 2 3 4 5 6 45 30 30 30 30 30 0 45 60 60 50 60 60 70 135 40 20 0 60 50 0 30 50 110 0 20 90 55 120 120 90 160 150 150 60 60 60 60 30 30 30 20 20 20 20 20 20 20 360 360 360 360 360 360 360

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5105

WEFTEC®.06

Table 7: Predicted Effluent Quality and MLSS for the Optimized SBR Designs Max MLSS mg/L Design objectives 15 20 3 15 10 5 3000 0 Scenario 1 T =10 C; Average Flow 3.0 15.1 0.7 7.2 2400 Scenario 2 T =10 0C; Peak Flow 4.9 17.7 1.7 10.9 1800 Scenario 3 T =13 0C; Average Flow 3.2 14.0 1.0 7.1 2300 Scenario 4 T =13 0C; Peak Flow 5.3 18.4 2.2 10.0 1700 Scenario 5 T =20 0C; Average Flow 3.4 17.0 2.8 3.8 2100 Scenario 6 T =20 0C; Peak Flow 5.1 18.1 3.0 5.9 1600 * Effluent standard under 100C/50 oF, (winter), 130C/55 oF (average) and 200C/68 oF (summer) conditions, respectively. CBOD5 mg/L

Parameters

TSS mg/L

NH4 mg/L

NO3* mg/L

These results are based upon a detailed influent characterization, and use the default values of stoichiometric/kinetic parameters in the ASM1 model (Henze et al., 1987). In the future, additional simulation effort may be undertaken to further optimize the modified SBR cycle settings. CONCLUSIONS This case study illustrates the benefits of simulation for evaluation and optimization of the operation of an SBR treating a high strength industrial wastewater. The operational improvements indicated by the simulations were as follows: • high effluent TSS concentrations in the baseline operation were due to inadequate settling time in the secondary clarifiers, • the aeration stage in the baseline operation was far longer than necessary, consuming excess energy with no additional benefit, • a reduction in the time for the aeration stage could be used to provide an anoxic stage to promote denitrification, and • the SBR treatment cycle could be optimized to meet the stringent effluent criteria for average and peak flow conditions, as well as winter, spring and summer wastewater temperatures, with the exception of the effluent NO3 at peak loading and 20ºC/68 ºF. ACKNOWLEDGEMENTS The authors would like to acknowledge the help and assistance of Unai Iriarte and the Chemical Engineering Department of the University of the Basque Country, Spain. REFERENCES Dold P.L., Wentzel M.C., Billing A.E., Ekama G.A. and Marais G.v.R. (1991) Activated sludge simulation programs: Version 1.0 Nitrification and nitrification/denitrification systems. Pub. by Water Research Commission, P/Bag X03, Gezina 0031, South Africa. Henze M., Grady C.P.L. (Jr), Gujer W., Marais G.v.R and Matsuo T. (1987) Activated Sludge Model No.1. IAWQ Scientific and Technical Report No.1, IAWQ, London. 33pp. Copyright ©2006 Water Environment Foundation. All Rights Reserved

5106

WEFTEC®.06

Henze M., Gujer W., Mino T., Matsuo T., Wentzel M.C., Marais G.v.R. and Van Loosdrecht (1999) Activated sludge model No. 2d. Wat. Sci. Tech., 39 (1), 165-182. Hydromantis Consulting Engineers, GPS-X Version 4.1.2 (June 2003) Hu Zhi-rong (2001). External nitrification biological nutrient removal activated sludge systems – development and modelling, Ph.D thesis, Dept. of Civil Engineering, University of Cape Town, South Africa. Unai Iriarte, (2003) Personal communication Wentzel M.C., Ekama G.A., Marais G.v.R (1992) Process and modelling of nitrification denitrification biological excess phosphorus removal systems - a review. Wat. Sci. Tech., 25 (6), 59 - 82.

Copyright ©2006 Water Environment Foundation. All Rights Reserved

5107