Application of artificial neural networks for prediction of photocatalytic reactor.

In this paper, forecasting of kinetic constant and efficiency of photocatalytic process of TiO2 nano powder immobilized on light expanded clay aggregates (LECA) was investigated. Synthetic phenolic wastewater, which is toxic and not easily biodegradable, was selected as the pollutant. The efficiency of the process in various operation conditions, including initial phenol concentration, pH, TiO2 concentration, retention time, and UV lamp intensity, was then measured. The TiO2 nano powder was immobilized on LECA using slurry and sol-gel methods. Kinetics of photocatalytic reactions has been proposed to follow the Langmuir-Hinshelwood model in different initial phenol concentration and pH. Several steps of training and testing of the models were used to determine the appropriate architecture of the artificial neural network models (ANNs). The ANN-based models were found to provide an efficient and robust tool in predicting photocatalytic reactor efficiency and kinetic constant for treating phenolic compounds.