Scholarlink

<p>Advanced optimization techniques are vital for engineering applications that require high-strength and corrosion-resistant welds in cladding processes. Traditional methods generally rely on empirical trial-and-error or isolated optimization that often overlook multi-objective trade-offs between tensile strength and corrosion resistance. This study proposes a Hybrid Taguchi-ANOVA-Grey Wolf Optimizer (GWO) Framework, enhanced by Random Forest (RF) surrogate modeling, for multi-objective optimization of TIG, MIG, and CMT cladding parameters. The Taguchi L9 Orthogonal Array Design minimizes experimental runs at the same time capture the main factor effects. One-way ANOVA uses an Ordinary Least Squares model to identify statistically significant parameters mainly current, torch speed, and heat input responsible for ultimate tensile strength (UTS) and corrosion rate while RF regression models, trained on an expanded dataset, serve as surrogates to capture non-linear interactions also provided feature importance rankings. These surrogates were integrated into a GWO, configured with domain-specific constraints, to optimize a weighted objective function maximizing UTS (540–570&#xa0;MPa) and minimizing corrosion rate (0.21–0.34&#xa0;mm/year) within 50 iterations. This hybrid framework offers cost-effective, interpretable, and scalable optimization which resulted in enhanced cladding performance with improved fidelity for industrial applications.</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

Design of an improved model using Taguchi design ANOVA and Grey Wolf optimization for multi-objective cladding process optimization

  • 期刊论文

  • 发表日期：

    2025年10月31日

查看全文
  • Samrat Kavishwar,
  • Vinod Bhaiswar

摘要

Advanced optimization techniques are vital for engineering applications that require high-strength and corrosion-resistant welds in cladding processes. Traditional methods generally rely on empirical trial-and-error or isolated optimization that often overlook multi-objective trade-offs between tensile strength and corrosion resistance. This study proposes a Hybrid Taguchi-ANOVA-Grey Wolf Optimizer (GWO) Framework, enhanced by Random Forest (RF) surrogate modeling, for multi-objective optimization of TIG, MIG, and CMT cladding parameters. The Taguchi L9 Orthogonal Array Design minimizes experimental runs at the same time capture the main factor effects. One-way ANOVA uses an Ordinary Least Squares model to identify statistically significant parameters mainly current, torch speed, and heat input responsible for ultimate tensile strength (UTS) and corrosion rate while RF regression models, trained on an expanded dataset, serve as surrogates to capture non-linear interactions also provided feature importance rankings. These surrogates were integrated into a GWO, configured with domain-specific constraints, to optimize a weighted objective function maximizing UTS (540–570 MPa) and minimizing corrosion rate (0.21–0.34 mm/year) within 50 iterations. This hybrid framework offers cost-effective, interpretable, and scalable optimization which resulted in enhanced cladding performance with improved fidelity for industrial applications.