Experimental and theoretical analysis of the UV/H2O2 advanced oxidation processes treating aromatic hydrocarbons and MTBE from contaminated synthetic wastewaters

The aim of the present study is experimental and theoretical investigation of treatment of contaminated wastewater with aromatic hydrocarbons and MTBE by UV/H2O2 process. An artificial neural network (ANN) approach is developed for prediction of degradation extent of pollutants as a function of initial concentration of H2O2 and pollutants, pH, solution temperature, reaction time and UV intensity. The proposed ANN model is developed using 1007 experimental dataset which are collected from 15 different references as well as our 497 experimental datasets. Our experimental investigation showed that 3 UV lights illumination (18 W), initial concentration of 0.421, 0.724, 1.11, 1.34 and 1.736 (g/L) for H2O2 and acidic pH of 3.1 is optimal condition for BTEX degradation. After 180 min of starting of the UV/H2O2 process, 90% of 550 mg/L and 98% of 210 mg/L BTEX has been removed. The performed statistical analysis confirmed a multi-layer perceptron neural network (MLPNN) with only one hidden layer composed of fifteen neurons is the best ANN architecture for modeling of the considered task. The developed MLP model shows an absolute average relative deviation (AARD) of 10.26 and a mean square error (MSE) of 5 x 10(-4) for estimating all of the experimental dataset. The obtained results confirm that the proposed MLPNN model is an applicable and feasible tool to predict the degradation extent of aromatics hydrocarbon and MTBE with high accuracy. (C) 2014 Elsevier Ltd. All rights reserved.