Scholarlink

<p>Post-weld heat treatment (PWHT), a thermal processing technique, is generally used to improve metallurgical and mechanical properties of the weld. However, this process is time-consuming and resource-intensive, resulting in costlier end-products. The objective of this study is to substitute PWHT with an approach aided by Artificial Intelligence (AI) tools to make welding economical. Bonobo Optimizer (BO), an intelligent optimization tool, could outperform Particle Swarm Optimizer (PSO) to serve this purpose, as it showed a better convergence rate. The governing parameters of BO were configured to be adaptive; consequently, it became an intelligent optimization tool. BO could closely predict the input (process) parameters of welding that could give rise to the mechanical properties of the weld, namely yield strength, ultimate tensile strength, and elongation that could have been yielded through the PWHT. Thus, the novelty of this study deals with testing the applicability of AI tools to suggest the appropriate input/process parameters so that the enhanced properties of the weld could be achieved even by not adopting the costly process of PWHT.</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

Study on AI-based approaches to replace costly heat-treatment process for obtaining enhanced mechanical properties of the weld

  • 期刊论文

  • 发表日期：

    2025年9月27日

查看全文
  • Santosh Kumar Gupta,
  • Bitan Pratihar,
  • Dilip Kumar Pratihar,
  • Partha Saha

摘要

Post-weld heat treatment (PWHT), a thermal processing technique, is generally used to improve metallurgical and mechanical properties of the weld. However, this process is time-consuming and resource-intensive, resulting in costlier end-products. The objective of this study is to substitute PWHT with an approach aided by Artificial Intelligence (AI) tools to make welding economical. Bonobo Optimizer (BO), an intelligent optimization tool, could outperform Particle Swarm Optimizer (PSO) to serve this purpose, as it showed a better convergence rate. The governing parameters of BO were configured to be adaptive; consequently, it became an intelligent optimization tool. BO could closely predict the input (process) parameters of welding that could give rise to the mechanical properties of the weld, namely yield strength, ultimate tensile strength, and elongation that could have been yielded through the PWHT. Thus, the novelty of this study deals with testing the applicability of AI tools to suggest the appropriate input/process parameters so that the enhanced properties of the weld could be achieved even by not adopting the costly process of PWHT.