Optimization of municipal wastewater biological nutrient removal using ASM2d

Activated sludge model No. 2d (ASM2d) was calibrated then used to determine optimal recycle rates and basin retention times, with respect to final N, soluble P (sP) or N plus sP concentrations for Modified Bardenpho (MB), Modified University of Cape Town (MUCT), and anaerobic–anoxic–oxic (A²O) biological nutrient removal processes. Simulations were conducted at two temperatures and three primary effluent (PE) concentrations (COD, TSS, TKN, sP). All BNR processes were shown to be capable of achieving an effluent sP concentration below 1 mg/L at all conditions when effluent N concentration was neglected. None of the processes were capable of producing an effluent with N concentrations below 5 mg/L at high PE concentration. The MB and MUCT processes were both successful in achieving a combined sP and N removal below 1 mg/L and 5 mg/L for low PE concentrations at 10 and 20 ºC. Only the MB process, at 20 ºC with medium PE concentrations, was found to achieve an effluent below 1 mg/L and 5 mg/L, respectively for sP and N. Recycle from the anoxic basin of the MUCT process had an insignificant effect on N and sP removals. All input variables to the MB and A²O process proved to be somewhat significant and it is recommended that they be kept within future experimental designs. Dynamic tests with real Ottawa plant influent flow data indicated that the MB process was more robust with respect to TN, sP, and CBOD₅ removal than the MUCT process.Key words: activated sludge, biological nutrient removal, simulation, model, ASM2d.