Dynamic optimization of a wastewater treatment process for sustainable operation using multi-objective genetic algorithm and non-dominated sorting cuckoo search algorithm

Abstract

Wastewater treatment plants (WWTP) face challenges due to nonlinearity and influent oscillations. An efficient optimal control system is necessary to improve effluent water quality without increasing operating costs. This research proposes a novel dynamic optimization control system using a non-dominated sorting-based multi-objective cuckoo search optimization algorithm (NSMOCS). Benchmark Simulation Model 1 (BSM1) was used as the WWTP model. Two sets of objectives were analyzed: the first set consists of the effluent quality index (EQI) and operating cost index (OCI), and the second comprises the reduction in nitrogen and ammonia concentrations at the WWTP. A Pareto front was obtained by evaluating these objectives using a multi-objective genetic algorithm (MOGA). Contribution analysis was used to find the individual weight values of these objectives. Furthermore, two sets of multi-objective cost functions were devised using weighted sum analysis. By minimizing the above-mentioned cost functions with NSMOCS algorithm, the best tuning parameters for the proportional integral (PI) controller were found. The simulation results with this proposed control strategy showed that the pollution units per day and the operating cost index were reduced during dry, rain, and storm weather conditions, with reductions of 1.3 %, 0.77 %, 0.91 %, and 0.9 %, 0.85 %, and 0.63 %, respectively.