Abstract
A membrane bioreactor by sequentially alternating the inflow and by applying a two-stage coagulation control based on pre-coagulation was evaluated in terms of phosphorus removal efficiency and cost-savings. The MBR consisted of two identical alternative reaction tanks, followed by aerobic, anoxic and membrane tanks, where the wastewater and the internal return sludge alternatively flowed into each alternative reaction tank at every 2 hours. In the batch-operated alternative reaction tank, the initial concentration of nitrate rapidly decreased from 2.3 to 0.4 mg/L for only 20 minutes after stopping the inflow, followed by substantial release of phosphorus up to 4 mg/L under anaerobic condition. Jar test showed that the minimum alum doses to reduce the initial $PO_4$-P below 0.2 mg/L were 2 and 9 mol-Al/mol-P in the wastewater and the activated sludge from the membrane tank, respectively. It implies that a pre-coagulation in influent is more cost-efficient for phosphorus removal than the coagulation in the bioreactor. On the result of NUR test, there were little difference in terms of denitrification rate and contents of readily biodegradable COD between raw wastewater and pre-coagulated wastewater. When adding alum into the aerobic tank, alum doses above 26 mg/L as $Al_2O_3$ caused inhibitory effects on ammonia oxidation. Using the two-stage coagulation control based on pre-coagulation, the P concentration in the MBR effluent was kept below 0.2 mg/L with the alum of 2.7 mg/L as $Al_2O_3$, which was much lower than 5.1~7.4 mg/L as $Al_2O_3$ required for typical wastewater treatment plants. During the long-term operation of MBR, there was no change of the TMP increase rate before and after alum addition.
