Modelling electro-fenton process using nature-inspired hybrid techniques in antibiotics’ removal from aqueous solutions

Abstract Antibiotics are commonly used to treat and prevent bacterial illnesses in both humans and animals worldwide. The average life expectancy of living things has improved significantly due to antibiotics. Antibiotics are highly stable in mild environments, making it difficult for traditional wastewater treatment facilities to remove them effectively. In the electro-Fenton (EF) process, in the presence of bipotential iron produced from cathode electrodes, hydrogen peroxide is affected and strong hydroxyl radicals are produced. It can oxidize various pollutants into harmless components. Amoxicillin (AMX), a widely prescribed beta-lactam antibiotic (65% of the global market) with high environmental persistence, was selected due to its resistance to conventional wastewater treatments. To predict how the EF process will behave when treating AMX antibiotics, this study evaluates the efficacy of machine learning algorithms based on artificial neural networks (ANN), such as Multilayer Perceptron (MLP) hybridized with vortex search (VS), shuffled complex evolution (SCE), and backtracking search algorithm (BSA). Four input variables were employed for the analysis: pH, EDTA (mM), initial concentration (mg/l), and current intensity (mA). The coefficients of determination (R 2 ) for the BSA-MLP, SCE-MLP, and VS-MLP models were 0.9197, 0.9372, and 0.8894 in the testing dataset, and 0.98, 1.0, and 0.9998 in the training dataset, respectively. These results demonstrate that the SCE-MLP model outperforms the other models in terms of accuracy, prediction stability, and generalization capacity under real-world conditions. This novel approach establishes a robust framework for predicting antibiotic elimination, with significant implications for addressing environmental challenges. The findings advance our understanding of wastewater treatment optimization and enhance predictive capabilities for antibiotic removal.