The use of artificial neural network (ANN) for modeling of diesel contaminated soil remediation by composting process

In this study two models for remediation of diesel contaminated soil by composting process were used: mathematical modeling and artificial neural network (ANN) modeling. The mathematical model was solved iteratively and validated with experimental data. Then, a three-layer back propagation ANN was trained, tested and validated to predict the decomposition of diesel in contaminated soil according to 3600 data sets which were obtained from mathematical model. The Best neural network result has been obtained with one hidden layer network, with 14 neurons. “tansig” for hidden layers and “purelin” for the output layer gave the best performance compared to other activation functions. The ANN architecture contains six inputs. Diesel decomposition percent is the only output of ANN. ANN predicted results are very close to the target data. The high correlation coefficient, 0.9995, between the network prediction and the corresponding data proves that ANN modeling is a satisfactory method for remediation process.