This study shows the results of the experimental investigation of refinery wastewater treatment by the electro-Fenton process. The experiments were designed using Taguchi design approach with an orthogonal array (OA) of L16 runs. Four process variables
(current density (CD), temperature (T), Fe2+ concentration (Fe2+), and time (t)) at four different levels were considered for the present design. Regression analysis was performed to predict a correlation for the response function (chemical oxygen demand (COD) reduction efficiency) of the treatment process by the electro-Fenton technique. Analysis of variance (ANOVA) was carried out to verify the significant variables that control COD reduction. Moreover, a linear model analysis was implemented for the signal to noise (S/N) ratios with "larger the better" and for means. Based on S/N ratios and means responses, the operating conditions for optimum COD reduction were: CD = 8 mA/cm2, T = 60°C, Fe2+ = 0.4 mM, and t = 6 h, at which 87.35% of COD was reduced. ANOVA shows that the operating temperature had the most significant impact on COD reduction efficiency. The R-square value for the predicted COD reduction correlation was 90.00%. Considering ranks based on delta statistics, the relative magnitude of effects of the process variables were: T, Fe2+, t, and CD.