DEVELOPMENT OF THE AquaMB PROCESS Ken Mikkelson, Ph.D. Ed Lang Lloyd Johnson, P.E. Aqua Aerobic Systems, Inc. Aqua-Aerobic Systems, Inc. 6306 N. Alpine Road Rockford, IL 61111 Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 1
DEVELOPMENT OF THE AquaMB PROCESS Introduction The use of membranes for wastewater treatment is growing rapidly. As the demand for reclaimed water grows, the need for higher quality effluent will be more apparent. In addition, more robust membrane materials and improved manufacturing methods are resulting in lower costs of ownership. In February 2000, Aqua-Aerobic Systems, Inc. initiated development of a Membrane Bioreactor (MBR) system. The objective of the program was to develop a system that offered superior performance at the lowest life cycle cost compared to existing technologies. After three years of testing, the objective has been met with the AquaMB Process (Multiple Barrier Membrane System). Following is the development history, prototype testing results and evolution of the AquaMB Process experience. Prototype Development Conditions Prototype testing of various biological membrane treatment systems was conducted at the Rockton, Illinois wastewater treatment facility. The Rockton WWTP is a typical flow through activated sludge system that processes approximately 350,000 gpd of municipal wastewater with a limited industrial component. Operation of the prototype systems at this site provided an adequate source of typical raw wastewater and permitted the return of treated flow streams and waste sludge streams to the headworks of the system. Testing operations were initiated in February 2000 and concluded in January 2003. Phase I testing concentrated on the evaluation of sequencing batch reactor (SBR) technology combined with side stream cross flow hollow fiber microfiltration membranes. The hollow fiber membrane system was designed to process a high solids concentration flow stream that was directed to the inside of the hollow fiber with permeate collection on the exterior of the hollow fiber in the module, a modified MBR approach. Phase I testing was initiated in February 2000 and concluded in August 2001. Phase II testing was also conducted with SBR technology combined with side stream cross flow hollow fiber microfiltration. During Phase II operations, an outside-in hollow fiber microfiltration system was evaluated. The microfiltration membrane again processed a high solids concentration flow stream but the stream was directed to the outside of the hollow fiber with permeate collection on the interior of the hollow fiber, a modified MBR approach. Phase II testing was initiated in September 2001 and concluded in July 2002. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 2
Phase III testing was based upon the AquaMB Process. Clarified effluent was introduced to a cloth media filtration system. The cloth media filtration system effluent was processed by the side stream cross flow hollow fiber microfiltration membrane system. In this case, a low solids concentration flow stream was directed to the outside of the hollow fiber with permeate collection on the interior of the hollow fiber. Phase III testing was initiated in September, 2002 and concluded in January 2003. Phase I of the pilot testing program was designed to document the effluent quality produced by the particular system that was evaluated. Phases II and III focused upon optimizing the microfiltration component of the system. Each system was evaluated to determine the maximum flux rate that was feasible for long term operation. This objective dictated the evaluation of different modes of operation including variations in the production cycle duration; the optional use of a continuous or intermittent air scrub either in the module side or base, or both; variation of air flow rates during air scrub sequences; backpulse sequences in the production cycle and/or flux maintenance cycle; implementation of a variety of flux maintenance actions that included air scrub reverse filtration and reverse filtration with chemical addition at various durations and flow rates and enhanced flux maintenance procedures at elevated temperature conditions. Various clean in place (C.I.P.) procedures were implemented and the required frequency and effectiveness of such procedures was evaluated with respect to flux recovery, membrane integrity and system and chemical cost requirements. Phase I Development Phase I testing documented the performance of the sequencing batch reactor (SBR) system and the microfiltration system that processed a high solids concentration flow stream based upon an inside-out flow condition. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 3
The SBR system was operated at two distinct conditions. Condition A was based on dual reactors configured to operate with a 2.0 hour treatment cycle. Biological treatment objectives were based upon carbonaceous reduction and nitrification. Condition B was based on dual reactors configured to operate with a 2.0 hour treatment cycle comprised of a Mixed Fill phase, a React Fill phase and a React discharge phase. The aeration was cycled to create aerobic/anoxic conditions during the 2.0 hour cycle to achieve nitrification/denitrification as well as carbonaceous reduction. In addition, liquid alum was injected directly to the reactors to accomplish total phosphorus reduction. Solids retention was accomplished by the membrane system for both conditions. A summary of the design conditions and performance follows: Prototype Testing Design Conditions Condition A Condition B Treatment Objectives Carbonaceous Reduction Yes Yes Nitrification Yes Yes Denitrification No Yes Phosphorous Reduction No Yes Operating Parameters MLSS (mg/l) 10600 11590 Hydraulic Retention Time (Hours) 7.06 10.36 Sludge Age (Days) 30 37 # of Treatment Cycles / Day 12 12 Hours / Cycle 2 2 React / Fill 1 0.75 Mixed Fill N/A 0.25 React / Discharge 1 1 Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 4
Prototype Testing Performance Condition A Condition B Parameter Influent Effluent Influent Effluent BOD 5 179 mg/l <2 mg/l 181 mg/l <2 mg/l TSS 137 mg/l <2 mg/l 164 mg/l <2 mg/l COD 512 mg/l <20 mg/l 460 mg/l <20 mg/l TKN 38.6 mg/l 1.05 mg/l 36.6 mg/l 1.32 mg/l NH 3 -N 26.1 mg/l 0.17 mg/l 23.7 mg/l 0.56 mg/l NO 3 -N N/A N/A 5.1 mg/l 1.63 mg/l NO 2 -N N/A N/A 0.027 mg/l 0.03 mg/l T-N N/A N/A 41.73 mg/l 2.98 mg/l T-P N/A N/A 5.64 mg/l 0.07 mg/l Turbidity --- 0.32 NTU --- 0.19 NTU Average performance data for the system under Condition A was at an instantaneous flux rate of 25.5 GFD. Condition B was at an instantaneous flux rate of 24.3 GFD. Phase I operations confirmed that sustained membrane flux rates of 25 GFD are feasible with specific flux maintenance procedures. The sequencing batch reactor operated successfully at 10-12,000 mg/l of MLSS in a simple two or three phase configuration that did not utilize a React or Settle phase. The React discharge phase was proven to be a reliable mode of operation that provided high quality effluent. Efficient removal of COD, BOD5, total nitrogen and total phosphorus was successfully demonstrated. The process operating strategy utilized in Phase I operations resulted in a U.S. patent. Fine screening of the mixed liquor stream exiting the SBR and entering the membrane system was determined to be a necessity with the inside-out flow pattern through the hollow fibers employed in Phase I. The flow volume that required screening was based upon the system flow plus the excess cross flow that was required for the module and recycled to the SBR units. Membrane cleaning sequences were efficient during Phase I operations and can easily be automated. The various membrane modules utilized during Phase I testing provided 23.5 to 48 square feet of surface area per module based upon a maximum module length of 1.0 meter. The pore size ranged from 0.2 microns (nominal) to a membrane with a Molecular Weight Cutoff (MWCO) of 500,000. This available module configuration exaggerated the membrane cost factor in full-scale systems. The 25 GFD flux rate was equivalent to or greater than other membrane alternatives but the module housing, while limited to the 1.0 meter length, did not result in an overall cost effective solution. The necessity of fine screening and the cost effectiveness aspect of the 1.0 meter membrane module were the primary driving forces that resulted in a reconfiguration of the system that was evaluated during Phase II testing. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 5
Phase II Development Phase II testing was focused upon the operating characteristics of the micofiltration membrane system. The performance requirements of the SBR system were documented during Phase I testing, and the reactors were operated during Phase II to generate the high solids concentration feed stream to the membrane system. The microfiltration system was designed to operate with a high solids concentration feed stream with an outside-in flow pattern. The system featured membrane modules that were approximately 2.0 meters in length and provided 538.2 square feet of filtration area. The pore size ranged from 0.17 to 0.20 microns (nominal). The SBR system was operated to provide a feed stream MLSS concentration of approximately 8-10,000 mg/l. The membrane system was configured to operate with a goal of achieving a sustainable time averaged flux rate of 15-20 GFD. Fine screening was maintained in the system. Rather than screening the mixed liquor exiting the SBR system, the fine screen was moved to the inlet to the SBR unit. At the SBR inlet, the screened volume was reduced to the design flow rather than the design flow plus any cross flow requirement of the membrane system. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 6
The system performance was observed and documented under a number of operational variables. The objective in each condition was to maximize the flux rate and minimize the cycle to cycle flux maintenance requirement, the enhanced flux maintenance requirement and the chemical cleaning requirement. The major operational variables included the following items: 1. Variation of the feed flow rate and the resultant cross flow rate based upon a target filtrate flow rate. 2. Variation in the retentate or back pressure in the system. 3. Incorporation of air scrub techniques during the production cycle. a. Air injection to the module on a continuous or intermittent basis. b. Configuration of devices to distribute air flow during side injection. c. Base air entry to the module either in-line or through air distribution ports directed to the normal liquid entry porting. 4. Backpulsing techniques during the production cycle and/or flux maintenance sequence. 5. Flux maintenance sequence composition and techniques. 6. Production cycle duration. 7. Enhanced flux maintenance techniques and frequency. 8. Clean in place techniques and frequency. 9. Fiber density in the microfiltration module. 10. Fiber porosity and type in the microfiltration module. As a result of Phase II testing, it was determined that the 2.0 meter module with 538.2 square feet of filtration area could process the high MLSS flow stream with the following operational characteristics: Production Cycle Air Scrub/Reverse Filtration Reverse Flow Chemical Cleaning (NaOCI) Enhanced Flux Maintenance Clean-In-Place Duration 12.0 minutes 60 seconds 30 seconds 30 minutes (every 24 hours) 2-6 hours (every 3-4 days) Operating Values 17.0 GFD 6 SCFM, 20 gpm 20 gpm 10-20 mg/l *Backpulse is incorporated in production and flux maintenance cycles. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 7
Phases I and II Development Conclusion The research efforts of Phases I and II provided beneficial data for comparison to existing membrane technologies available in the market place. In addition to solving the technical issues associated with the direct application of high mixed liquor solids concentrations to the membrane module, it was important to have an attractive market value with reasonable life cycle costs. During the course of this study, the reference capital costs for alternative membrane bioreactor configurations dropped significantly. Although the targeted installed system value for the Aqua MBR had been achieved, it did not meet the primary objective of lowest life cycle cost based upon new market conditions. The primary objective of the program at this stage of development shifted to an economical solution using technical knowledge gained from the work in Phases I and II. Three key areas were targeted: 1. Reduce the system market value and life cycle costs. 2. Improve the membrane integrity/reliability. 3. Eliminate or reduce the need for supporting operations like fine screening. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 8
Phase III was implemented to focus on these three targets and established a reliable system design that not only met but also bettered the current market value for membrane bioreactors. The key cost component in the system is the membrane surface area required to perform the final filtration step. Utilizing the existing technical knowledge base and recognizing the merits of a well-treated flow stream prior to the membrane, the membrane surface area could be significantly reduced. This decision not only solved the market value issue but also solved the membrane integrity/reliability issue as well as eliminated the need for fine screening. Phase III Development Phase III testing was based upon the concept of the AquaMB Process. The AquaMB Process is an integrated approach comprised of a modified activated sludge system, cloth media filtration and a microfiltration membrane system. The rationale of Phase III testing was based upon configuring a biological treatment system that incorporated a Settle phase and produced a clarified effluent. The Aqua cloth media filtration system processed the clarified effluent and produced a feed stream to the microfiltration membrane system that provided the potential for operation at flux rates that exceed those of higher solids concentration feed streams. The biological treatment process characteristics are primarily dictated by nitrogen removal requirements and the minimum operating temperature. Fine screening is not required prior to the activated sludge or membrane system. Prior investigations documented that the nature and concentration of suspended solids in the feed stream are primary factors with respect to the operational flux rate of membrane systems. In essence, the achievable flux rate will vary with the concentration of the solids present in the feed stream. This relationship is depicted in Figure No. 1. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 9
Figure 1. Solids Loading vs. Membrane Performance Flux Rate GFD AquaMB Process Traditional SubmergedMBRs Feed Stream Suspended Solids Concentration mg/l. Prior evaluations at the University of California - Davis, in conjunction with Title 22 applications in California and pilot studies at Rockton, Illinois, indicated that an activated sludge effluent polished by a cloth media filtration system would provide a consistent, high-quality feed stream that was low in turbidity and suspended solids concentration. Figures No. 2 and 3 depict this condition. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 10
Figure 2. 1 2 Effluent turbidity, NTU 1 0 8 6 4 1 2 3 4 a 4 b N e e d le F e lt P ile 2 0 0 5 1 0 1 5 2 0 2 5 In flu e n t tu rb id ity, N T U? Deep-bed, continuous backwash upflow mono-medium filters Shallow depth, automatic backwash mono, dual and multi-medium downward flow filters Deep-bed, mono-medium downward and/or upward filters? Shallow-depth, mono-medium filters? Shallow-depth, dual medium filters? Cloth Media Disk Filter (needlefelt media)? Cloth Media Disk Filter (pile cloth media) Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 11
Figure 3. Rockton, Illinois - CMF Evaluation Filtration Rate: 3.0 gpm/ft2 3 2.5 2 Effluent 1.5 Suspended solids (mg/l) 1 0.5 0 0 10 20 30 40 50 60 70 80 Influent A high quality feed stream allows the membrane system to be operated at elevated flux rate values while producing the required final water quality as demanded by reuse criteria. An indication of this operating condition is demonstrated in Figure No. 4 as a 12 hour tracing that depicts a typical period of operation with the feed stream turbidity averaging about 1.0 NTU. The tracing starts at approximately 8.0 hours after an enhanced flux maintenance cycle. The filtrate flow rate is at 21.5 gpm, which is equivalent to an instantaneous flux rate of 57.5 GFD. The transmembrane pressure (TMP) at the end of each production cycle increases from approximately 17 psig to 22.0 psig throughout the 12.0 hour tracing period or approximately 0.42 psig/hour. This rate of TMP increase equates to a 10.0 psig increase over a 24 hour period of operation that matches the target level of operation. The enhanced flux maintenance cycle is designed to restore the degree of cleanliness to very nearly the base line condition with respect to the initial TMP value. As the initial TMP value increases over approximately a 30 day period of operation the necessity of a CIP cycle becomes evident to the system operator. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 12
Operation of the membrane system at elevated flux rates results in a reduction of the surface area requirements and the system cost. Testing has indicated the membrane system can be maintained with a reduced level of flux maintenance, enhanced flux maintenance and CIP procedures. Monthly rather than multiple weekly CIP operations are planned based upon studies at the Rockton test site. Figure 4. Operation of the membrane system at elevated flux rates reduces the membrane surface area requirement, the potential cost for membrane replacement and the overall life cycle cost. The reduced flux maintenance and CIP requirements prolongs Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 13
membrane life and reduces the cost of flux maintenance and CIP chemical cleaning requirements. Phase III Development Conclusion The ability to test high and low solids loadings in a variety of arrangements provided Aqua-Aerobic Systems, Inc. with a diverse knowledge of membrane bioreactor systems. Based on Phase III testing, Aqua was better able to make the educated decision to advance the design concept of the AquaMB Process to the commercial marketplace. Technical integrity having been established, the AquaMB Process was determined to possess such benefits as treatment objective design flexibility, optional discharge points, reduced membrane surface area and lower capital and life cycle costs. Test market respondents viewed these benefits as highly desirable. Development Summary The design of the AquaMB Process is based on three years of development and testing. During that time, much was learned about the advantages and disadvantages of designing and operating MBR systems utilizing direct membrane filtration of high level mixed liquor. This experience, as well as the experience of others, was considered in the decision to design a system that would provide the greatest value to customers. The AquaMB Process provides the lowest capital and life cycle costs when compared to traditional multi-unit processes or MBR systems. Value is not only measured in terms of cost. The AquaMB Process provides value through flexible biological process design, consistently high quality membrane feed and optional points of discharge. The AquaMB Process is the clear choice for advanced wastewater treatment. Copyright 2003 Aqua-Aerobic Systems, Inc. All rights reserved. 14