Artificial Neural Network Modeling of Fenton-Based Advanced Oxidation Processes for Recycling of Textile Wastewater

Abstract

In this study, Fenton-based advanced oxidation processes (homogeneous and heterogeneous Fenton/photoFenton), were applied for advanced treatment of real textile wastewater. Fe2O3-sepiolite was used as a catalyst in the heterogeneous Fenton/photo-Fenton processes. Among the processes studied, heterogeneous photoFenton process showed highest colour and TOC removal. At the optimum conditions, colour and TOC removal efficiencies were achieved as 98% and 69% for the homogeneous Fenton process, 100% and 71% for the homogeneous photo-Fenton process, 92% and 63% for the heterogeneous Fenton process and 98% and 83% for the heterogeneous photo-Fenton process, respectively. The reusability of the catalyst in heterogenous Fenton/photoFenton processes was also investigated. The catalyst showed better reusability performance in photo-Fenton process compared to Fenton process in terms of colour and TOC removal. Artificial neural network (ANN) was used to simulate and predict the performance of the Fenton-based processes. The results predicted by the ANN are very close to the experimental data with the correlation coefficients (R2) of 0.9847 and 0.9752 for homogeneous and heterogeneous Fenton processes, respectively. The catalyst dose was the most effective parameter for homogeneous Fenton process with an importance of 47.5%, while the contact time was the most effective parameter with 40.37% for heterogeneous Fenton process.