Scholarlink

<p>Ultrasonic vibration assisted friction stir welding (UVAFSW) is a promising technique for joining aluminum alloys, offering enhanced mechanical properties and improved joint quality over conventional FSW. However, optimizing the process parameters for optimal results remains a complex challenge. This study presents a Q-learning-based approach to optimize the process parameters of UVAFSW for AA2014-T651 alloy. By leveraging the reinforcement learning capabilities of Q-learning, the algorithm efficiently explores the vast parameter space. It identifies the optimal combination of rotation speed of the tool is 1032 rpm, and the traverse speed of the tool is 61 mm/min. Experimental validation of the optimized parameters demonstrates significant improvements in tensile strength (TS) of 471.33 MPa, yield strength (YS) of 398.55 MPa, elongation (%EL) of 11.42 %, impact strength (IS) of 8.70 J, and microhardness (H) of 151.54 HV (Vickers) of the welded joints. The Q-learning approach offers a robust and efficient method for optimizing UVAFSW, enabling the production of high-quality welded joints with enhanced mechanical properties.</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

Q-learning-based optimization of process parameters for ultrasonic vibration assisted friction stir welding of AA2014-T651

  • 期刊论文

  • 发表日期：

    2026年3月10日

查看全文
  • Borigorla Venu,
  • Jyothhi Yarlagaddaa,
  • Ramakrishna Malkapuram,
  • Perumalla Janaki Ramulu,
  • L. Suvarna Raju

摘要

Ultrasonic vibration assisted friction stir welding (UVAFSW) is a promising technique for joining aluminum alloys, offering enhanced mechanical properties and improved joint quality over conventional FSW. However, optimizing the process parameters for optimal results remains a complex challenge. This study presents a Q-learning-based approach to optimize the process parameters of UVAFSW for AA2014-T651 alloy. By leveraging the reinforcement learning capabilities of Q-learning, the algorithm efficiently explores the vast parameter space. It identifies the optimal combination of rotation speed of the tool is 1032 rpm, and the traverse speed of the tool is 61 mm/min. Experimental validation of the optimized parameters demonstrates significant improvements in tensile strength (TS) of 471.33 MPa, yield strength (YS) of 398.55 MPa, elongation (%EL) of 11.42 %, impact strength (IS) of 8.70 J, and microhardness (H) of 151.54 HV (Vickers) of the welded joints. The Q-learning approach offers a robust and efficient method for optimizing UVAFSW, enabling the production of high-quality welded joints with enhanced mechanical properties.