Advancing wastewater reuse: AI-driven insights into ozone-based organic pollutant reduction

Abstract

The high concentrations of organic pollutants in treated wastewater (TWW) pose significant environmental challenges, especially in agricultural irrigation. This study analyzed TWW samples collected from irrigation zones to evaluate organic pollutant levels, mainly focusing at the Biochemical Oxygen Demand (BOD). The efficacy of ozone in reducing these pollutants was investigated through an ozone concentration of 0.83 mg/L in TWW. Samples were exposed to ozone for varying durations, namely 0, 15, and 25 min, with real-time monitoring of key water quality parameters using a HANNA multiprobe, including temperature, pH, electrical conductivity (EC), resistivity, total dissolved solids (TDS), salinity, and oxidation–reduction potential (ORP). The ORP levels, initially 155–165 mV in untreated samples, rose significantly post-ozonation to 208 and 218.30 mV for 15 and 25 min of treatment, respectively while reducing the elevated BOD levels. Subsequently, the study further integrated Artificial Intelligence (AI) assisted Machine Learning (ML) for accurate ORP prediction. Several models such as ANFIS-M1, ANFIS-M2, SVR-M1, SVR-M2, RLR-M1, and RLR-M2 (α, β, γ, Δ, ε, ζ variations) were developed and evaluated. Based on the performance evaluation metrics, ANFIS-M1 consistently demonstrated superior performance, while SVR-M2 and RLR-M2 exhibited greater variability and generally lower accuracy. It is therefore believed that this research provides insights into managing TWW for irrigation and advances AI’s role in sustainable water management practices.