The prediction of mechanical behavior for steel wires and cord materials using neural networks

Abstract

The tensile strength of the steel wire material is required to be sufficiently high for better performance. Steel with a high cleanness will prevent problems during drawing and the heat treatment. The studies show that among many defects the most important ones are the non-metallic inclusions and undesirable phases encountered during improper heat treatment. Especially different non-metallic inclusions will play an important role during crack propagation due to their weak matrix bond. In this study typical wire and cord failures due to non-metallic inclusions are examined. A generalized regression neural network was developed to predict the tensile strength as a function of experimental conditions. The predicted values of the tensile strength estimated by neural network are found to be in good agreement with the actual values from the experiments.