Comparing the Capability of Phenomenological (Johnson-Cook and Arrhenius-Type) and Artificial Neural Network Models in Predicting the Hot Deformation Behavior of Additively Manufactured 316L Stainless Steel

The high temperature flow behavior of additively manufactured 316L stainless steel was investigated in this study by hot compression tests at the temperatures of 973, 1073, 1173 and 1273 K and strain rates of 0.001-0.1 s-1. Constitutive models consisting of Johnson-Cook and Arrhenius-type were employed. The results indicated that the Arrhenius-type constitutive equation had higher accuracy than the Johnson-Cook model, but these constitutive models could not predict (i) the strength levels at all temperatures and strain rates, and (ii) the flow hardening/softening behavior, accurately. Therefore, an artificial neural network with a feed-forward back propagation learning algorithm has been established to predict the high temperature flow behavior of additively manufactured 316L stainless steel. This model includes three layers namely the input layer, the hidden layer (with 20 neurons), and the output layer. The input data consisted of true strain (ε), strain rate ( ), and deformation temperature (T) while the predicted flow stress (σ) was the output data. In order to evaluate the performance of employed models, standard statistical parameters such as the average absolute relative error (AARE), root mean square error (RMSE) and correlation coefficients (R) were used. The results showed that the artificial neural network model was more accurate than the constitutive equations in predicting the high temperature flow behavior of additively manufactured 316L stainless steel.