Degradation of 4-Chloro phenol in aqueous media thru UV/Persulfate method by Artificial Neural Network and full factorial design method

Abstract

ABSTRACT In this project, the degradation of 4-Chloro phenol (4-CP) in aqueous environment by ultraviolet/persulphate (UV/PS) process was investigated in a batch photo-reactor. The full factorial design (FFD) and artificial neural networks (ANN) were used to investigate the influence of experimental variables comprising initial concentration of persulfate and 4-CP, and pH on the removal of 4-CP. The optimal conditions were achieved at 60 mM of persulphate(PS) and 0.5 mM of 4-CP and pH of 10. At this condition, the removal of 4-CP was 90.3% (experimental), and the predicted quantity by FFD and ANN approaches were 91.24 and 90.37%, respectively, and the removal of COD was 54.3% after 60 min of reaction. Also, the ANN model was better than FFD and the root mean square error (RMSE) of ANN was lower than FFD model (0.6017AAN ˂ 0.6832FFD). The ANN needs larger sets of data and computational time. A high correlation coefficient (R2 ANN = 0.9987, R2 FFD = 0. 9983) was attained by an assessment between the results of experimental and model. The average percentage error for ANN and FFD were 0.188 and 0.545, respectively, representing the benefit of ANN in taking the nonlinear performance of the system.