Scholarlink

<p>This study introduces a model based on advanced machine learning techniques to estimate the mechanical and microstructural properties of the Al6wt.%Si-2.5wt.%Cu alloy under various processing conditions, including thixoforming and T6 heat treatment. The alloy's microstructure was analyzed using optical and electron microscopy. Tensile strength, hardness, and fatigue resistance tests adhering to technical standards were assessed. The findings confirm that thixoforming—particularly the holding time at 583&#xa0;°C (corresponding to a 60% liquid fraction)—along with the T6 heat treatment, has a substantial impact on the silicon shape factor, grain size, hardness, tensile strength, and fatigue resistance of the alloy. The machine learning model used processing conditions and heat treatment as input variables to predict critical properties such as grain size, shape factor of the silicon particle, hardness, ultimate tensile strength, elongation, yield strength, and fatigue strength. Among the models tested, the random forest algorithm provided the most accurate predictions, with a coefficient of determination of 87% and a mean logarithmic error of 0.002 on the test dataset. The machine learning models effectively predicted mechanical and microstructural characteristics based on processing conditions, optimizing the performance of the thixoformed process of Al-Si-Cu alloys.</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

Machine Learning Algorithms Model Applied to Predict Mechanical Properties of Thixoformed Al-6wt.%Si-2.5wt.%Cu Alloy

  • 期刊论文

  • 发表日期：

    2025年4月24日

查看全文
  • Wendel Leme Beil,
  • Bárbara Dora Ross Veitía,
  • Hipolito Domingo Carvajal Fals,
  • Eugenio José Zoqui

摘要

This study introduces a model based on advanced machine learning techniques to estimate the mechanical and microstructural properties of the Al6wt.%Si-2.5wt.%Cu alloy under various processing conditions, including thixoforming and T6 heat treatment. The alloy's microstructure was analyzed using optical and electron microscopy. Tensile strength, hardness, and fatigue resistance tests adhering to technical standards were assessed. The findings confirm that thixoforming—particularly the holding time at 583 °C (corresponding to a 60% liquid fraction)—along with the T6 heat treatment, has a substantial impact on the silicon shape factor, grain size, hardness, tensile strength, and fatigue resistance of the alloy. The machine learning model used processing conditions and heat treatment as input variables to predict critical properties such as grain size, shape factor of the silicon particle, hardness, ultimate tensile strength, elongation, yield strength, and fatigue strength. Among the models tested, the random forest algorithm provided the most accurate predictions, with a coefficient of determination of 87% and a mean logarithmic error of 0.002 on the test dataset. The machine learning models effectively predicted mechanical and microstructural characteristics based on processing conditions, optimizing the performance of the thixoformed process of Al-Si-Cu alloys.