Monitoring and Application of Artificial Neural Network Model for Prediction of Organophosphorus Pesticides Residue in Ahvaz Water Treatment Plants

Abstract

Organophosphorus pesticides are the largest and most diverse pesticides. The overuse of pesticides will cause them to remain in the food, water, soil, and air, hazardous to human health. This study was conducted in three seasons to determine organophosphorus pesticide concentration. The experiments were modeled using artificial neural networks. The results showed that parathion, malathion, and diazinon concentrations were significantly different (p<0.05). The most concentrations were observed in Aug, September, and October. The OPPs concentration in water treatment plants' effluents indicated that concentrations of pesticides were below the maximum contaminant level. Base on the results of an artificial neural network, the model performance to be the best prediction for malathion concentration in the WTP (NO.1), with 6 neurons with R2 = 0.887, parathion with 5 neurons and R2 = 0.711, and diazinon with 11 neurons and R2 = 0.714. The finding of ANN modeling for malathion concentration in the WTP (NO.2), with 9 neurons and R2 = 0.713, parathion, one hidden layer with 6 neurons and R2 = 0.71, and parathion with 15 neurons and R2 = 0.674 were showed the best prediction.