Optimizing removal of antiretroviral drugs from tertiary wastewater using chlorination and AI-based prediction with response surface methodology

Abstract

Chemical and pharmaceutical chemicals found in water sources create substantial risks to human health and the environment. The presence of pharmaceutical contaminants in water can cause antibiotic resistance development, toxicity to aquatic organisms, and endocrine disruption. Hence, the elimination of chemicals and other contaminants from wastewater prior to its release is a burgeoning concern in the domains of engineering and science. The use of treatment technologies in wastewater treatment plants can remove pharmaceutical contaminants through the oxidation process. However, many traditional wastewater treatment plants lack the advanced monitoring tools required to detect low concentrations of pharmaceuticals. Without the ability to detect these compounds, it's challenging to treat them effectively. The goal of this study was to use Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) algorithms to model and improve how Nevirapine and Efavirenz break down in different chlorination conditions. The RSM analysis revealed statistically significant models (F-values: Nevirapine, pH-t: 108.15, T-t: 76.55, ICC-t: 110.84), indicating a strong correlation between operational parameters (pH, temperature, and initial chlorine concentration) and degradation behavior. The ANN model accurately predicted the degradation of both Nevirapine and Efavirenz under various chlorination conditions, as confirmed by analyzing actual-predicted graphs, residual plots, and Mean Squared Error (MSE) values. The ANN model using ICC-t achieved the highest MOD value of 31.31 % for Nevirapine. The ANN model based on ICC-t yielded a maximum MOD value of 16.06 % for Efavirenz. These findings provide valuable insights into optimizing chlorination processes for better removal of these pharmaceutical contaminants from water.