Prediction of flow stress in Ti–6Al–4V alloy with an equiaxed α + β microstructure by artificial neural networks

Flow stress during hot deformation depends mainly on the strain, strain rate and temperature, and shows a complex and nonlinear relationship with them. A number of semi-empirical models were reported by others to predict the flow stress during hot deformation. This work attempts to develop a back-propagation neural network model to predict the flow stress of Ti–6Al–4V alloy for any given processing conditions. The network was successfully trained across different phase regimes (α + β to β phase) and various deformation domains. This model can predict the mean flow stress within an average error of ∼5.6% from the experimental values, using strain, strain rate and temperature as inputs. This model seems to have an edge over existing constitutive model, like hyperbolic sine equation, and has a great potential to be employed in industries.