Modeling and Optimization of Acid Blue 193 Removal by UV and Peroxydisulfate Process

Abstract

In this study, the advanced oxidation process for dye removal from textile wastewater treatment was investigated by means of an experimental setup in which the effect of several parameters on dye removal efficiency [peroxydisulfate concentration, ultraviolet (UV) irradiation, temperature, dye concentration, and time] was examined. In order to predict the removal efficiency, two types of artificial neural networks were used: an adaptive neuro-fuzzy inference system (ANFIS) and an artificial neural network determined with differential evolution called hybrid self-adaptive differential evolution with neural networks (hSADE-NN). After the successful development of ANFIS, its ability to predict test data was checked. Also, a series of models of the process was determined with hSADE-NN. Comparison of the two approaches indicates that both methods provide good results, the average absolute relative error for hSADE-NN being 3.61% and that for ANFIS 5.18%. After that, a process optimization was performed, the scope being to determine the conditions for maximum dye removal efficiency under various constraints, considered as a means to reduce resources consumed.