Developing a model of metal structure formation in the heat-affected zone of high-strength pipe steels

Abstract

The article presents the results of developing a model that employs artificial neural networks (ANNs) to predict the structural-phase composition of the weld-affected zone (WAZ) metal in high-strength steels used to produce pipes of K60–K70 strength classes. The model consists of four sub-blocks that sequentially predict parameters determining the final structural-phase composition of the WAZ metal, such as average austenite grain diameter, critical temperatures of austenite decomposition, and both qualitative and quantitative structural-phase compositions. Each sub-block utilizes ANNs that have been developed, trained, and stored as functions in the MATLAB software environment.