Application of neural networks for the prediction of continuous cooling transformation diagrams

In the paper the method of forecasting of Continuous Cooling Transformation (CCT) diagrams for steels by the use of neural networks has been presented. Input data are chemical composition and austenitising temperature. Results of calculation of neural networks consist of temperature of the beginning and the end of transformation in the function of cooling rate, the participation of the structural components and the hardness of steel cooled from austenitising temperature with a fixed rate. Presented quantities enable to draw the CCT diagram. The model presented in the paper enables the analysis of the influence of the chemical composition and austenitising temperature on CCT diagrams. In order to work out the methods, the set of experimental data worked out on the basis of information available in the literature consisting of 400 CCT diagrams made for constructional and machinable steels were used.