Prediction of Hot Deformation Behavior of AA2030 Alloy Using Artificial Neural Networks

Hot deformation behavior of materials is completely complex. Flow stress during hot deformation depends mainly on the strain, strain rate and temperature, and shows a complex and nonlinear relationship with them. Hot deformation characteristics of a AA2030 alloy were investigated by hot compression test over the temperature range from 350-500 ℃ and strain rate range from 0.005 to 0.5 s. Based on these experimental results, a feed-forward back propagation artificial neural network (ANN) model was developed to predict the flow behaviors of AA2030 alloy during hot deformation. The inputs of the neural network were deformation temperature, log strain rate and strain whereas flow stress was the output. This developed network consisted of one hidden layer containing 12 neurons with a tan-sigmoid activation function and Levenberg–Marquardt training algorithm. A very good correlation between experimental and predicted results has been obtained. The results show that the developed artificial neural network model presented an excellent capability to predict the flow stress level, and also the hardening and dynamic softening behavior.