Application of neural networks for prediction of critical values of temperatures and time of the supercooled austenite transformations

摘要

The own model is proposed in the paper making it possible to predict the A{sub}(c1), A{sub}(c3), B{sub}s, M{sub}s temperatures and time to begin the bainitic transformation. The original method was used for calculating the anisothermic diagrams of the supercooled austenite transformations of the constructional steels using the artificial neural networks. The set of training data was compiled to carry out this task (400 charges of constructional steels) including their chemical compositions, austenitising temperatures, and the supercooled austenite transformation diagrams during their continuous cooling. The obtained results were compared with results obtained based on those cited in literature and with commonly used empirical relationships, indicating in many cases to the better consistency of calculations made using the new method with the empirical data. Examples of application of the developed model for evaluation of the effect of the alloying elements on the A{sub}(c3) temperature value and time to beginning of the bainitic transformation was presented.