PSO-Driven Optimization of Dielectric Barrier Discharge Ozone Generation Parameters for Water Treatment Applications

Abstract

This investigation presents a comprehensive optimization framework for ozone generation in dielectric barrier discharge (DBD) systems applied to water treatment processes. The study employs a Venturi injection configuration to systematically analyze the influence of three critical operational parameters: applied voltage (6-8 kV), oxygen flow rate (3-5 L/min), and water flow rate (8-16 L/min) on dissolved ozone concentration in treated water. Experimental design methodology utilizing MODDE 5.0 software enabled the development of a second-order polynomial mathematical model correlating input variables with ozone concentration response. The derived model demonstrated excellent statistical validity with coefficient of determination (R²) of 99.2% and predictive capability (Q²) of 90.3%, confirming robust model accuracy and reliability.Parameter optimization was achieved through implementation of Particle Swarm Optimization (PSO) algorithm, a bio-inspired metaheuristic approach that mimics collective intelligence behavior observed in natural swarms. The PSO methodology employed a population of 50 particles over 100 iterations to minimize prediction errors between theoretical model outputs and experimental measurements. Convergence analysis revealed rapid optimization performance, achieving optimal parameter extraction within 90 iterations. Comparative analysis between PSO-optimized parameters and conventional MODDE 5.0 results demonstrated superior performance of the proposed approach. Optimal operating conditions were identified as: applied voltage of 7.95 kV, oxygen flow rate of 3.38 L/min, and water flow rate of 8.27 L/min. The PSO-based optimization yielded enhanced ozone generation efficiency through reduced water flow rates and optimized voltage levels, confirming the voltage parameter as the dominant factor in ozone production.