Catalytic Wet Air Oxidation and Neural Network Modeling of High Concentration of Phenol Compounds in Wastewater

Abstract

Wastewater and other harmful compounds can be produced from different industrial activities and may contains toxic compounds like phenols. At high phenol concentrations, the carcinogenic effects are more likely and the environmental problems are of high concerns. In this study, the removal of a very high phenol concentration from refinery wastewaters by Catalytic Wet Air Oxidation was investigated and compared with the predicted results by neural network model. Alumina-supported platinum catalyst was prepared, characterized, and examined in a trickle bed reactor. A back propagation neural network model constructed in three layers was used to validate the experimental results. The first layer had five inputs that represent different operating conditions of feed acidity, weight hourly space velocity, pressure, time, and temperature; the second hidden layer had 18 neurons; and the third one was represented by one output, namely, phenol conversion percentage. The experimental results showed that the very high phenol concentration was successfully converted to a safe concentration and to acceptable concentrations of intermediates. The predicted phenol conversion of the neural network model shows the least error compared with the experimental test.