Scholarlink

<p>This study investigates the fatigue crack growth behavior of 316L stainless steel weldments through a combined approach of fracture mechanics evaluation and machine learning-based data classification. Fatigue crack growth rate (d<i>a</i>/d<i>N</i>) versus stress intensity factor range (Δ<i>K</i>) analyses were conducted on base metal (BM), heat-affected zone (HAZ), and weld center (WC) specimens, both with and without post-weld heat treatment (PWHT). The results revealed that the weld center exhibited the lowest crack growth rates, indicating superior fatigue resistance compared with the HAZ and BM. Machine learning models, including K-Nearest Neighbors (KNN) and Support Vector Machine (SVM), successfully classified fatigue behavior across different weld regions, achieving up to 98% accuracy for WC PWHT and 91–96% for other conditions. Overall, the weld center without PWHT had the best resistance to fatigue, followed by the weld center with PWHT, the HAZ without PWHT, the HAZ with PWHT, and finally the base metal, which had the worst resistance to crack growth. The integration of experimental and ML-based evaluation provides an effective framework for assessing structural integrity in stainless steel weldments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integration of Fracture Mechanics and Artificial Intelligence for Crack Growth Analysis in 316L Austenitic Stainless Steel Weldments

  • 期刊论文

  • 发表日期：

    2025年12月10日

查看全文
  • Hermawan Agus Suhartono,
  • Yunan Prawoto,
  • Kirman Kirman,
  • Franky Melky,
  • Harris Zenal,
  • Yudi Irawadi,
  • Budi Prasetyo,
  • Yusuf Afandi,
  • Erinna Dyah Atsari,
  • Puguh Triwinanto

摘要

This study investigates the fatigue crack growth behavior of 316L stainless steel weldments through a combined approach of fracture mechanics evaluation and machine learning-based data classification. Fatigue crack growth rate (da/dN) versus stress intensity factor range (ΔK) analyses were conducted on base metal (BM), heat-affected zone (HAZ), and weld center (WC) specimens, both with and without post-weld heat treatment (PWHT). The results revealed that the weld center exhibited the lowest crack growth rates, indicating superior fatigue resistance compared with the HAZ and BM. Machine learning models, including K-Nearest Neighbors (KNN) and Support Vector Machine (SVM), successfully classified fatigue behavior across different weld regions, achieving up to 98% accuracy for WC PWHT and 91–96% for other conditions. Overall, the weld center without PWHT had the best resistance to fatigue, followed by the weld center with PWHT, the HAZ without PWHT, the HAZ with PWHT, and finally the base metal, which had the worst resistance to crack growth. The integration of experimental and ML-based evaluation provides an effective framework for assessing structural integrity in stainless steel weldments.