Prediction of the flow behaviour of AISI 321 austenitic stainless steel during dynamic recovery using sine hyperbolic constitutive equation and artificial neural network

In the present investigation, AISI 321 austenitic stainless steel was subjected to hot compression deformation at temperatures of 800, 850, and 900°C and strain rates in the range of 0.001-1 s-1. Results showed that in all the predefined conditions, the microstructure of the samples consisted of elongated austenite grains which indicates the occurrence of dynamic recovery. Furthermore, the hot flow curves obtained at this temperature range showed a trend similar to the dynamic recovery flow type behavior. In these curves, the flow stress increased with increasing applied strain and then tended to a constant value. The saturation stress at 800˚C increases from about 200 to 280 MPa with increasing strain rate from 0.001 to 1 s-1. For the same strain rate changes, the saturation stress increases from 150 to 240 MPa, and 90 to 200 MPa, at deformation temperature of 850 and 900˚C, respectively. The flow curves obtained from the hot compression tests were modelled using the sine hyperbolic constitutive equation and alternatively an artificial neural network. The results showed that the artificial neural network better predicts the flow behavior of AISI 321 austenitic stainless steel during dynamic recovery compared to the sine hyperbolic constitutive equation.