Application of neural networks in generating processing map for hot working

An important parameter in the mechanical working of materials is called workability, which is the relative ease with which a metal can be shaped through plastic deformation without the formation of any defect. Workability can be evaluated by means of processing maps, constructed from experimentally generated flow stress variation with respect to strain, strain rate and temperature. The present work demonstrates the use of neural network in generating processing maps for hot working processes. A neural network model was trained and tested for predicting the flow stress by taking data available in the literature for 99.99% pure aluminum. It was found that the trained neural network could predict the flow stress for unseen data quite reliably. At strain of 0.4, power dissipation and instability maps were constructed, utilizing the flow stress prediction by neural network. Superimposition of these maps provided processing maps at 0.4 strain, which was similar to that available in the literature. This established the potential of applying neural network, which is more robust technique than conventional method, for generating the processing map.