Application of artificial neural network for prediction of dyes photo-degradation efficiency in UV/TiO2 condition based on structural information

Wastewaters contaminated with dyes remained from the textile, paper, and other industries are source of environmental problems, especially in the third world countries. Synthetic dyes with wide variety of chemical structures and compositions of inorganic and organic compounds are one of the main groups of pollutants in wastewater [1, 2]. In this study, QSPR model based on Bayesian Regularized artificial neural network (BR-ANN) was developed for the modelling and accurate prediction of dyes degradation efficiency based on their chemical structures [3-5]. For this purpose, the photo-catalytic degradation of various dyes (31 dyes) have been investigated in TiO2/UV suspensions. Under constant condition of including amount of photo-catalyst (0.0100g), sample volume (50 ml) and pH of 7 the degradation efficiency of each dye was measured 15 minute under UV irradiation. Dyes structures were converted to the molecular descriptors using Dragon software and the most significant descriptors were selected and used as inputs in ANN modelling. After obtaining the best BR-ANN model, the prediction ability of the proposed model was evaluated by prediction of degradation efficiency of test set. The regression coefficient (R2) for the test set was 0.85. The results obtained, showed the superior prediction ability of the proposed model in the prediction of degradation efficiency.