Abstract:In this study, artificial neural network (ANN) was used to model the hot deformation behavior of 7075 aluminum alloy during compression test, in the strain rate range of 0.0003-1 s−1 and temperature range of 200-450 °C. The inputs of the model were temperature, strain rate, and strain, while the output of the model was the flow stress. The feed-forward back-propagation network with two hidden layers was built and successfully trained at different deformation domains by Levenberg-Marquardt training algorithm. Comparative analysis of the results obtained from the hyperbolic sine, the power law constitutive equations, and the ANN shows that the newly developed ANN model has a better performance in predicting the hot deformation behavior of 7075 aluminum alloy.

The Use of ANN to Predict the Hot Deformation Behavior of AA7075 at Low Strain Rates