Modeling of low nickel-chromium steels by using backpropagation neural networks

Steel is the most important and versatile engineering alloy. It finds numerous applications and stands second to cement in its consumption in the world. The properties of steel can be controlled and varied over a wide range by changing the heat treatment and the alloying elements. The relation between the properties and the alloying elements of the alloy is, however, very complex and non-linear in nature. This paper presents the development of a back-propagation (BP) neural network model for prediction of the mechanical properties, and the sensitivity of each element on the properties for a given combination of composition and heat treatment variables for low nickel-chromium steels. The simulation results show that the average output-prediction error by BP network is less than 5 % of the prediction range in more than 95 % of the cases, which is quite acceptable for all metallurgical purposes.