Application of an Artificial Neural Network to Develop Fracture Toughness Predictor of Ferritic Steels Based on Tensile Test Results

Abstract

Analyzing the structural integrity of ferritic steel structures subjected to large temperature variations requires the collection of the fracture toughness (KJc) of ferritic steels in the ductile-to-brittle transition region. Consequently, predicting KJc from minimal testing has been of interest for a long time. In this study, a Windows-ready KJc predictor based on tensile properties (specifically, yield stress σYSRT and tensile strength σBRT at room temperature (RT) and σYS at KJc prediction temperature) was developed by applying an artificial neural network (ANN) to 531 KJc data points. If the σYS temperature dependence can be adequately described using the Zerilli–Armstrong σYS master curve (MC), the necessary data for KJc prediction are reduced to σYSRT and σBRT. The developed KJc predictor successfully predicted KJc under arbitrary conditions. Compared with the existing ASTM E1921 KJc MC, the developed KJc predictor was especially effective in cases where σB/σYS of the material was larger than that of RPV steel.

Keywords

• fracture toughness
• machine learning
• artificial neural network
• predictor
• yield stress
• tensile strength