Scholarlink

<p>AA6110 aluminum alloy in different aging conditions at 160&#xa0;°C for 0, 1, 12, and 168&#xa0;h was addressed under tensile loading. The microstructures of the alloy were characterized using a transmission electron microscope and a differential scanning calorimeter. Deformation, as well as strain hardening behavior, was investigated using Kocks-Mecking analysis. Flow stresses in the hardening regime were predicted using various constitutive models, including Ludwik, Hollomon, Swift, Ludwigson, and Voce. All material constants and parameters of the constitutive models were determined using the metaheuristic technique known as particle swarm optimization (PSO). A novel constitutive model was introduced, which predicts flow stresses covering hardening and softening regimes. Machine learning techniques, specifically adaptive neuro-fuzzy inference systems (ANFIS) with true strain and aging time as input features, were performed to predict the flow stresses of AA6110 aluminum alloy under various aging conditions. Three hyperparameters of ANFIS, i.e., number of memberships, type of membership function, and output function, were optimized using grid search and PSO. Finally, an optimized ANFIS was used to establish a flow stress prediction diagram based on the true strain and aging time of AA6110 aluminum 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

Effects of Aging Treatment on Strain Hardening Behavior of AA6110 Aluminum Alloy and its Flow Stress Modeling with Particle Swarm Optimization

  • 期刊论文

  • 发表日期：

    2025年5月20日

查看全文
  • Bhavin Bhatrasupong,
  • Oranicha Theerakiat,
  • Patiphan Juijerm

摘要

AA6110 aluminum alloy in different aging conditions at 160 °C for 0, 1, 12, and 168 h was addressed under tensile loading. The microstructures of the alloy were characterized using a transmission electron microscope and a differential scanning calorimeter. Deformation, as well as strain hardening behavior, was investigated using Kocks-Mecking analysis. Flow stresses in the hardening regime were predicted using various constitutive models, including Ludwik, Hollomon, Swift, Ludwigson, and Voce. All material constants and parameters of the constitutive models were determined using the metaheuristic technique known as particle swarm optimization (PSO). A novel constitutive model was introduced, which predicts flow stresses covering hardening and softening regimes. Machine learning techniques, specifically adaptive neuro-fuzzy inference systems (ANFIS) with true strain and aging time as input features, were performed to predict the flow stresses of AA6110 aluminum alloy under various aging conditions. Three hyperparameters of ANFIS, i.e., number of memberships, type of membership function, and output function, were optimized using grid search and PSO. Finally, an optimized ANFIS was used to establish a flow stress prediction diagram based on the true strain and aging time of AA6110 aluminum alloy.