Prediction of Lead Corrosion Behavior Using Feed-Forward Artificial Neural Network

The Feed-Forward Artificial Neural Networks (FFANNs) were used to predict the corrosion behavior of lead. A 3-9-2 network was adopted to train the networks and predict the lead corrosion behavior. The descriptors (input) were obtained using experimental methods. Solution concentration, pH and passive time were selected as the ANN input to predict the corrosion current and potential. To this end 80 samples were selected. The criterion of TSE was 0.004. It was found that the FFANNs could be used to predict the corrosion of lead.