Amin Ahmadpour - Faculty of Chemical, Gas and Petroleum Engineering, Semnan University, Semnan, Iran
Ali Haghighi Asl - Faculty of Chemical, Gas and Petroleum Engineering, Semnan University, Semnan, Iran
Narges Fallah - Chemical Engineering Department, Amirkabir University of Technology, Tehran, Iran

In this study, the photocatalytic method was used for treating the spent caustic in the wastewater of Olefin units used in petrochemical industries which contain large amounts of total dissolved solids (TDS). By using the synthetic photocatalyst of suspended titanium dioxide and measuring the chemical oxygen demand (COD) which was reduced in the photocatalyst (lbc) process, the values of COD were modeled and evaluated by means of the Box-Behnken (BBD) and the artificial neural network (ANN) using experimental tests in a double-cylindrical-shell photo reactor. According to the applied calculations, it was found that the artificial neural network was a more suitable method than the experimental design in modeling and forecasting the amount of COD removal. The modeling employed in this research showed that increasing the concentration of the photocatalyst in a state of neutral pH enhanced the COD removal up to the optimal amount of 1.31 g/L without restrictions and 2 g/L with restrictions at the rate of 81% and 79%, respectively. In addition, the study of the parameter effects including oxidizer amount, aeration rate, pH, and the amount of loaded catalyst indicated that all factors except pH had a positive effect on the model; furthermore, if the interactions were neglected, the COD removal efficiency would increase by increasing each of these factors (except pH). In addition, there was no interaction between the aeration and the concentration of the photocatalyst, and the acidic pH was more suitable at low concentrations of the photocatalyst. Besides that, by increasing the pH, the efficiency of removal was reduced when the oxidant was at its low level. The results showed that photolysis and adsorption adoptions had a very small effect on the efficiency of the removal of COD compared to the photocatalyst adoptions, and it was insignificant. In addition, the photocatalytic method had an acceptable capacity for removing the phenol in the wastewater sample, whereas it was inefficient in reducing the sulfide solution in the wastewater

Artificial neural networks (ANN) ،Design of experiment (DOE)،Photocatalytic wastewater treatment،Spent caustic wastewater،Titanium dioxide