Scholarlink

<p>Traditional machining methods face significant challenges when working with titanium beta alloy Ti-15&#xa0;V-3Cr-3Al-3Sn (Ti-15–3) due to its high hardness and low thermal conductivity. Electrical Discharge Machining (EDM), a non-contact thermal-erosion technique, offers an effective alternative for high-precision machining of this alloy. Nevertheless, determining optimal EDM parameters for key performance indicators such as Material Removal Rate (MRR), Tool Wear Rate (TWR), and Surface Roughness (SR) is difficult due to the complex, non-linear relationships among process variables. To address this, the present study applies Artificial Intelligence (AI) models whose hyperparameters are tuned using a hybrid metaheuristic optimization approach that combines Genetic Algorithm and Particle Swarm Optimization (GAPSO). Three AI models, Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF), were optimized and evaluated. Experimental findings reveal that model effectiveness is inherently metric-dependent. For MRR, GAPSO-optimized SVM achieves the strongest test-set generalization (R<sup>2</sup> = 0.8649), a substantial improvement over unoptimized SVM (R<sup>2</sup> = 0.7124). For TWR, both PSO and GAPSO-optimized SVM jointly achieve the highest accuracy (R<sup>2</sup> = 0.8420), compared to R<sup>2</sup> = 0.4483 for the unoptimized baseline, confirming that swarm-based optimization is highly effective for this output. For Surface Roughness, the unoptimized Decision Tree delivers the highest predictive accuracy (R<sup>2</sup> = 0.9024), outperforming all optimized variants, indicating that SR exhibits threshold-structured behavior better captured by rule-based partitioning than by kernel hyperparameter tuning. Overall, the proposed GAPSO-AI framework significantly improves the predictive accuracy of EDM performance, providing a reliable decision-support tool for process optimization and surface quality enhancement during machining of the Ti-15–3 alloy.</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

Artificial intelligence based prediction and optimization of electrical discharge machining performance characteristics for titanium alloy using hybrid and single metaheuristic approaches

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月22日

查看全文
  • Md. Ashikur Rahman Khan,
  • Fatema Jannat Dihan,
  • Ishtiaq Ahammad,
  • Md. Ahnaf Sad Khan,
  • Md. Mustafizur Rahman

摘要

Traditional machining methods face significant challenges when working with titanium beta alloy Ti-15 V-3Cr-3Al-3Sn (Ti-15–3) due to its high hardness and low thermal conductivity. Electrical Discharge Machining (EDM), a non-contact thermal-erosion technique, offers an effective alternative for high-precision machining of this alloy. Nevertheless, determining optimal EDM parameters for key performance indicators such as Material Removal Rate (MRR), Tool Wear Rate (TWR), and Surface Roughness (SR) is difficult due to the complex, non-linear relationships among process variables. To address this, the present study applies Artificial Intelligence (AI) models whose hyperparameters are tuned using a hybrid metaheuristic optimization approach that combines Genetic Algorithm and Particle Swarm Optimization (GAPSO). Three AI models, Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF), were optimized and evaluated. Experimental findings reveal that model effectiveness is inherently metric-dependent. For MRR, GAPSO-optimized SVM achieves the strongest test-set generalization (R2 = 0.8649), a substantial improvement over unoptimized SVM (R2 = 0.7124). For TWR, both PSO and GAPSO-optimized SVM jointly achieve the highest accuracy (R2 = 0.8420), compared to R2 = 0.4483 for the unoptimized baseline, confirming that swarm-based optimization is highly effective for this output. For Surface Roughness, the unoptimized Decision Tree delivers the highest predictive accuracy (R2 = 0.9024), outperforming all optimized variants, indicating that SR exhibits threshold-structured behavior better captured by rule-based partitioning than by kernel hyperparameter tuning. Overall, the proposed GAPSO-AI framework significantly improves the predictive accuracy of EDM performance, providing a reliable decision-support tool for process optimization and surface quality enhancement during machining of the Ti-15–3 alloy.