Prediction of martensite start temperature by neural network analysis

Abstract
Commercial steels are nowadays sophisticated alloys formed by a large number of alloying elements. The martensite start ( M_s) temperature of such steels is of vital engineering importance, and its prediction through models allows us to enhance the design and development of industrial products. In the present work, M_s temperature dependence on chemical composition has been examined by neural network analysis. Neural networks represent powerful methods of non-linear regression modelling. The network is a mathematical function which is fitted to experimental data. The influence of alloying elements such as C, Mn, Si, Cr, Ni, Mo, V, Co, W, Al, Nb, Cu, B and N on M_s temperature was analysed. Finally, a new empirical equation for M_s temperature was derived based on the neural network results.