Scholarlink

<p>This study proposes an integrated optimization strategy for preparing HEA coatings via laser cladding, aiming to maximize coating hardness while minimizing dilution rate. CoCrFeNiSi HEA coatings were fabricated using a Taguchi orthogonal design with laser power, scanning speed, and powder feed rate as variables. A support vector regression (SVR) model was built to capture the nonlinear relationship between process parameters and coating properties. To improve prediction accuracy, hyperparameters were optimized using both k-fold cross-validation and the NSGA-II. The NSGA-II-optimized SVR model achieved markedly lower mean squared errors for hardness (79.31% reduction) and penetration depth (90% reduction), confirming its superior global search capability. Based on this model, NSGA-II multi-objective optimization identified a Pareto frontier reflecting the trade-off between hardness and dilution rate. The optimal parameter set: laser power 1300 W, scanning speed 3&#xa0;mm&#xa0;s<sup>–1</sup>, and powder feed rate 1.3&#xa0;r&#xa0;min<sup>–1</sup>—was experimentally validated. Compared with initial conditions, the optimized coatings exhibited a significant reduction in dilution rate (72.38% to 18.00%), aspect ratio (12.81 to 2.34), and achieved high microhardness. These results demonstrate the effectiveness of the NSGA-II-SVR approach in guiding the design of high-performance HEA coatings.</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

Cooperative Optimization of Laser Cladding Process and Performance Validation for HEA Coatings Using the SVR Model and NSGA-II Algorithm

  • 期刊论文

  • 发表日期：

    2026年4月20日

查看全文
  • Xuejin Zhao,
  • Yongjun Shi,
  • Quan Li,
  • Kaijun Fan,
  • Shuyao Wang,
  • Ying Li,
  • XinYu Yan,
  • Mingjun Tian,
  • Menglu Xi,
  • Yangshu Liu

摘要

This study proposes an integrated optimization strategy for preparing HEA coatings via laser cladding, aiming to maximize coating hardness while minimizing dilution rate. CoCrFeNiSi HEA coatings were fabricated using a Taguchi orthogonal design with laser power, scanning speed, and powder feed rate as variables. A support vector regression (SVR) model was built to capture the nonlinear relationship between process parameters and coating properties. To improve prediction accuracy, hyperparameters were optimized using both k-fold cross-validation and the NSGA-II. The NSGA-II-optimized SVR model achieved markedly lower mean squared errors for hardness (79.31% reduction) and penetration depth (90% reduction), confirming its superior global search capability. Based on this model, NSGA-II multi-objective optimization identified a Pareto frontier reflecting the trade-off between hardness and dilution rate. The optimal parameter set: laser power 1300 W, scanning speed 3 mm s–1, and powder feed rate 1.3 r min–1—was experimentally validated. Compared with initial conditions, the optimized coatings exhibited a significant reduction in dilution rate (72.38% to 18.00%), aspect ratio (12.81 to 2.34), and achieved high microhardness. These results demonstrate the effectiveness of the NSGA-II-SVR approach in guiding the design of high-performance HEA coatings.