Artificial neural network modeling to evaluate and predict the mechanical strength of duplex stainless steel during casting

摘要

This paper presents the modeling of tensile properties of cast duplex stainless steel using the artificial neural network model, exclusively developed for this work. For this research, melts of varying chemical composition were poured, heat treated and tested for the tensile properties. The artificial neural network model was developed using the composition as input and tensile properties as the targets. The prediction performances of the models were evaluated by the mean absolute error (MAE), and the model with less MAE was considered for predicting the properties. Multilayer feed forward back propagation models with two hidden layers were implemented to predict the tensile properties of cast duplex stainless steel. The ANN model developed and validated shows a reliable correlation between chemical compositions and tensile properties.