Bench-scale studies were undertaken to explore the relative efficiency of three different process configurations for enhanced biological phosphorus (EBPR) removal; the University of Cape Town (UCT) process, the Sammamish Biological Nutrient Removal (SmBNR) process, and the University of Washington Membrane Biological Nutrient Removal (UW-MBNR) process. In this study, they all had the same anaerobic, anoxic, and aerobic volumes. The UCT process had the highest phosphorus (P) removal efficiency when the amount of P removed per unit of COD consumed was considered. It was found that addition of alum to the anaerobic zone of the process resulted in improved P removal. Critical to EBPR performance was the finding that phosphorus uptake rates are slower than anaerobic release rates and that the uptake rate is slower for lower reactor P concentrations. This encourages consideration of staged reactors to achieve lower effluent P concentrations. For all the systems, significant quantities of phosphorus accumulating organisms (PAO) and tetrad-forming glycogen accumulating organisms (GAO) were observed. Simulations using the commercial BioWin™ software overpredicted measured P removal. This may have been the result of GAO growth in the bench-scale reactors, which is not modeled in BioWin™. The BioWin™ models were calibrated by reducing the influent readily biodegradeable COD in the model. Although the BioWin™ models confirmed that the UCT process had the highest P removal efficiency, it was found in simulations of a full-scale plant with lower relative nutrient concentrations than were used in the bench-scale tests that both the UCT and SmBNR process configurations resulted in near complete P removal.
The combination of membrane bioreactor (MBR) technology and enhanced biological phosphorus removal (EBPR) for wastewater treatment offers the opportunity for nearly complete removal of effluent phosphorus (P) by combining removal of soluble phosphate by biological uptake with nearly complete elimination of particulates by membrane filtration. A potential disadvantage is that MBR systems have a relatively long solids retention time (SRT). This produces less excess biomass production, which affects EBPR efficiency as the phosphorus is removed via the wasted excess biomass. EBPR-MBR systems will, in most cases, also result in nitrate production and removal and, depending on the process configuration, portions of the biodegradable COD (bCOD) that can be used to promote EBPR will be consumed by biological denitrification. There is a need to consider different process configurations that address both nitrate removal and EBPR efficiency in MBR systems. This was a key goal of the research reported here.
It is well established that ability to achieve a low effluent soluble phosphorus concentration from EBPR systems is related to the influent bCOD:P ratio. When this ratio is low enough to limit EBPR, chemical addition is often used to provide additional phosphorus removal. In EBPR-MBR systems, chemical addition may be in a primary treatment step or within the MBR process. An additional aspect of the study was to compare the performance of two UCT systems operated in parallel with and without alum addition to the anaerobic contact zone.
Following the laboratory work a biological process model was calibrated to the bench-scale data and then used to evaluate the performance of various configurations for a full-scale plant design.
In this study, three different process configurations for both nitrogen removal and EBPR were tested and compared in bench-scale MBR reactors using a synthetic wastewater. The processes are illustrated in Figure 1, which shows the commonly used University of Cape Town (UCT) process and two other process configurations unique to this study, termed the Sammamish biological nutrient removal (SmBNR) process and University of Washington membrane BNR (UW-MBNR) configuration. The reactor sequence is anaerobic-anoxic-aerobic for the UCT process, anoxic-anaerobic-aerobic for the SmBNR process, and anoxic-anaerobic-aerobic for the UW-MBNR process. The latter two processes eliminate a recycle line, and for the UW-MBNR process, the influent is fed to the anaerobic zone rather than the anoxic zone.
A disadvantage of the UCT process is the need for an additional recycle stream and that the concentration of the mixed liquor in the anaerobic zone is only about 80 percent of that for the other zones. A disadvantage of the SmBNR process is that a large portion of the influent bCOD will be used by denitrifying organisms, leaving less available for EBPR, and potentially a higher effluent soluble phosphorus concentration.
A good measureof the efficiency of the EBPR processes is the amount of phosphorus removed relative to the amount of bCOD applied to the system. This could be defined as a bCOD-applied: P-removed ratio (bC/P ratio). EBPR systems that operate with higher bC/P need more influent degradable COD to achieve the same effluent P concentration versus a system that can operate with a lower bC/P ratio, and are thus less efficient. Factors that result in higher bC/P ratios include longer SRT and consumption of degradable COD by denitrifying organisms or glygocen accumulating organisms (GAO) instead of phosphorus accumulating organisms (PAO). Thus, this study also included monitoring and evaluating nitrate removal and bCOD utilization in the anaerobic and anoxic reactors, as well as microscopic surveillance of tetrad forming organisms (TFO).