Computational and statistical modeling for parameters optimization of electrochemical decontamination of synozol red dye wastewater

Abstract

In this study, computational and statistical models were applied to optimize the inherent parameters of an electrochemical decontamination of synozol red. The effect of various experimental variables such as current density, initial pH and concentration of electrolyte on degradation were assessed at Ti/RuO0·3TiO0·7O2 anode. Response surface methodology (RSM) based central composite design was applied to investigate interdependency of studied variables and train an artificial neural network (ANN) to envisage the experimental training data. The presence of fifteen neurons proved to have optimum performance based on maximum R2, mean absolute error, absolute average deviation and minimum mean square error. In comparison to RSM and empirical kinetics models, better prediction and interpretation of the experimental results were observed by ANN model. The sensitive analysis revealed the comparative significance of experimental variables are pH = 61.03%>current density = 17.29%>molar concentration of NaCl = 12.7%>time = 8.98%. The optimized process parameters obtained from genetic algorithm showed 98.6% discolorization of dye at pH 2.95, current density = 5.95 mA cm−2, NaCl of 0.075 M in 29.83 min of electrolysis. The obtained results revealed that the use of statistical and computational modeling is an adequate approach to optimize the process variables of electrochemical treatment.