Scholarlink

<p>The rising application of high-strength steels, especially dual-phase (DP) steels, for vehicle lightweighting is a result of the automobile industry’s pursuit of Improved fuel efficiency. These materials are highly valued for their exceptional strength and formability due to the combination of ferrite and martensite. A crucial factor influencing its formability is the plastic anisotropy, which is measured by the Lankford constant (r-value). As conventional methods of determining r-value are time-consuming and resource-intensive, this study used machine learning to create a predictive model. Six supervised regression models were trained using a dataset of 356 DP steel samples, each of which was distinguished by its distinct chemical composition and annealing temperature. With a predictive score of 85.93%, the gradient boosting regressor was the most successful of them. Notably, the study extends beyond prediction by revealing the relationships between annealing temperature and chemical composition, thereby providing a better understanding of their combined influence on plastic anisotropy using SHAP plots. This work will help in the development of more efficient material design and processing strategies for the automotive sector.</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

Influence of Alloying Elements and Annealing Temperature on Lankford Constant of High Strength Dual Phase Steels Using Supervised Regression Models

  • 期刊论文

  • 发表日期：

    2026年1月24日

查看全文
  • Avijit Pal,
  • Rajan Verma,
  • Pritam Mandal,
  • Samjukta Sinha,
  • Prabhat Das,
  • Snehanshu Pal,
  • Manojit Ghosh

摘要

The rising application of high-strength steels, especially dual-phase (DP) steels, for vehicle lightweighting is a result of the automobile industry’s pursuit of Improved fuel efficiency. These materials are highly valued for their exceptional strength and formability due to the combination of ferrite and martensite. A crucial factor influencing its formability is the plastic anisotropy, which is measured by the Lankford constant (r-value). As conventional methods of determining r-value are time-consuming and resource-intensive, this study used machine learning to create a predictive model. Six supervised regression models were trained using a dataset of 356 DP steel samples, each of which was distinguished by its distinct chemical composition and annealing temperature. With a predictive score of 85.93%, the gradient boosting regressor was the most successful of them. Notably, the study extends beyond prediction by revealing the relationships between annealing temperature and chemical composition, thereby providing a better understanding of their combined influence on plastic anisotropy using SHAP plots. This work will help in the development of more efficient material design and processing strategies for the automotive sector.