Evaluating the hot deformation behavior of a super-austenitic steel through microstructural and neural network analysis

A series of hot compression tests were conducted in the temperature range of 800-1100 °C under the strain rates of 0.001, 0.01, and 0.1 s−1 to assess the flow behavior and microstructure evolution of a super-austenitic stainless steel. The occurrence of dynamic recrystallization has been characterized as the dominant restoration mechanism operating in the investigated range of temperature. This is considered as the main factor affecting the related flow characteristics of the material. To better analyzing the obtained results, an artificial neural network (ANN) model with single hidden layer composed of 20 neurons has been established to simulate the flow behavior of the material. To train the model, a feed-forward back propagation algorithm has been employed. The reliability of the proposed model has been evaluated using standard statistical indices. In addition, the capability of the model has been assessed under the conditions at which the related data were not incorporated in the model. It was found that the developed ANN model employing this algorithm could efficiently track the work hardening and dynamic softening regions of the deforming material.