Model predictive control (MPC) and Fuzzy controllers are designed in a two-level hierarchical supervisory control framework for control of activated sludge-based wastewater treatment plants (WWTP) in order to efficiently remove nitrogen and phosphorus. Benchmark simulation model No.3 with a bio-phosphorus (ASM3bioP) module is used as a working platform. The hierarchical control framework is used to alter the dissolved oxygen in the seventh reactor (DO7 ) to control ammonia. Lower-level PI, MPC, and Fuzzy are used to control the nitrate levels in the fourth reactor (SNO4 ) by manipulating internal recycle (Qintr ) and DO7 in the seventh tank by manipulating mass transfer coefficient (KL a7 ). MPC and Fuzzy are designed in the supervisory layer to alter the DO7 set-point based on the ammonia composition in the seventh reactor (NH7 ). From the analysis, it is observed that the effluent quality is improved with a decrease in ammonia, TN, and TP. Though a little difference was observed in the cost for all the control strategies, a trade-off is maintained between cost and percentage improvement of effluent quality. MPC-MPC combination showed significant removal in ammonia and better effluent quality when compared to other control strategies. PRACTITIONER POINTS: Developed novel strategies in hierarchical configurations for better nutrient removal with optimal costs in an A2 O process. Lower level control strategies deals with dissolved oxygen in last aeration tank and nitrate in fourth anoxic tank (PI/MPC) Higher level control strategy deals with ammonia in the last aeration tank (MPC/Fuzzy). Average and violations of nutrient removal, economy and overall effluent quality for three weather conditions (Dry, Rain and Strom) are studied. A trade-off is observed between EQI and OCI.
Keywords: Benchmark simulation model (ASM3BioP); effluent quality; fuzzy control; model predictive control; supervisory control.
© 2021 Water Environment Federation.