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

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

Analyzing the structural integrity of ferritic steel structures subjected to large temperature variations requires the collection of the fracture toughness (<i>K<sub>J</sub></i><sub>c</sub>) of ferritic steels in the ductile-to-brittle transition region. Consequen...

Full description

Saved in:
Bibliographic Details
Main Authors: Kenichi Ishihara, Hayato Kitagawa, Yoichi Takagishi, Toshiyuki Meshii
Format: article
Language:EN
Published: MDPI AG 2021
Subjects:
fracture toughness
machine learning
artificial neural network
predictor
yield stress
tensile strength
Mining engineering. Metallurgy
TN1-997
Online Access:https://doaj.org/article/6877efa5a65d4c2f999a2270a37b7b39
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Analyzing the structural integrity of ferritic steel structures subjected to large temperature variations requires the collection of the fracture toughness (<i>K<sub>J</sub></i><sub>c</sub>) of ferritic steels in the ductile-to-brittle transition region. Consequently, predicting <i>K<sub>J</sub></i><sub>c</sub> from minimal testing has been of interest for a long time. In this study, a Windows-ready <i>K<sub>J</sub></i><sub>c</sub> predictor based on tensile properties (specifically, yield stress σ<sub>YSRT</sub> and tensile strength <i>σ</i><sub>BRT</sub> at room temperature (RT) and <i>σ</i><sub>YS</sub> at <i>K<sub>J</sub></i><sub>c</sub> prediction temperature) was developed by applying an artificial neural network (ANN) to 531 <i>K<sub>J</sub></i><sub>c</sub> data points. If the <i>σ</i><sub>YS</sub> temperature dependence can be adequately described using the Zerilli–Armstrong <i>σ</i><sub>YS</sub> master curve (MC), the necessary data for <i>K<sub>J</sub></i><sub>c</sub> prediction are reduced to <i>σ</i><sub>YSRT</sub> and <i>σ</i><sub>BRT</sub>. The developed <i>K<sub>J</sub></i><sub>c</sub> predictor successfully predicted <i>K<sub>J</sub></i><sub>c</sub> under arbitrary conditions. Compared with the existing ASTM E1921 <i>K<sub>J</sub></i><sub>c</sub> MC, the developed <i>K<sub>J</sub></i><sub>c</sub> predictor was especially effective in cases where <i>σ</i><sub>B</sub>/<i>σ</i><sub>YS</sub> of the material was larger than that of RPV steel.

Similar Items