Performance Modeling of the Fenton Process Used as a Single Unit for Treating Raw Textile Effluent

This study investigates the direct application of the Fenton Oxidation Process (FOP) to untreated textile wastewater, specifically from a woven fabric production facility. Under optimized conditions (pH 3, 0.7 g/L Fe+2, 2 mM H2O2), the process achieved significant removal efficiencies: 81% Chemical Oxygen Demand (COD), 80% Suspended Solids (SS), and 93% color removal. Artificial Neural Networks (ANN) and NARX-ANN were utilized in Matlab R2020a to model FOP performance, employing Levenberg-Marquardt (trainlm) and Scaled Conjugate Gradient (trainscg) algorithms. With a 9-20-3 network topology, the ANN model demonstrated superior predictive capability, achieving an R2 of 0.9843.