AbstractMotivated to reduce the costs incurred by corrosion in material science, this article presents a combined model based on artificial neural networks (ANNs) to predict pitting corrosion status of 316L austenitic stainless steel. This model offers the advantage of automatically determining the pitting corrosion status of the material. In this work, the pitting corrosion status was predicted, with the environmental conditions considered, in addition to the values of the breakdown potential estimated by the model previously, but without having to use polarization tests. The generalization ability of the model was verified by the evaluation using the experimental data obtained from the European project called “Avoiding Catastrophic Corrosion Failure of Stainless Steel”. Receiver operating characteristic space, in addition to area under the curve (AUC) values, was presented to measure the prediction performance of the model. Based on the results (0.994 for AUC, 0.980 for sensitivity, and 0.956 for specificity), it can be concluded that ANNs become an efficient tool to predict pitting corrosion status of austenitic stainless steel automatically using this two-stage procedure approach.
Can energy system modeling benefit from artificial neural networks? Application of two-stage metamodels to reduce computation of security of supply assessments
