Scholarlink

<p>Electrochemical machining (ECM) is an effective manufacturing method for machining hard-to-cut materials, especially for complex profiles components, such as the blisk of aero engines. Given the interplay of multiple physical fields during the ECM forming process, the governing mechanism underlying the forming process are intricate. Consequently, the design and subsequent correction of the tool cathode is the key to achieving precision manufacturing of the blisks. Manual correction of tool cathodes is a laborious and time-intensive process, which complicates the development of a quantifiable correction model. Therefore, this paper proposes a machine-learning-assisted approach to correct tool cathodes, and establishes a tool cathodes correction model using the support vector regression (SVR) algorithm. Within this model, the profile deviation of the workpiece is used as the input layer of the SVR model, and the correction amount on the tool cathode surface is used as the output layer. Subsequently, the corrected tool cathode was employed in the blisk ECM experiment. The machined workpiece had a maximum profile deviation of 0.0116&#xa0;mm, and a smaller accuracy variance of the blade control line is obtained, which is 0.0302, which is 56.92% lower than the manually corrected one. The experimental results indicated that the application of a machine-learning-assisted correction method for tool cathodes enhances both the machining accuracy and repeatability in blisk ECM. Hence, it indicates that machine learning extensive potential for the refinement of tool cathode correction in blisk ECM, improving both cathode correction quantifiability and machining accuracy.</p> Graphical Abstract <p></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 Intelligent Assisted Correction of Tool Cathode for Blisk Electrochemical Machining

  • 期刊论文

  • 发表日期：

    2025年3月1日

查看全文
  • Zhiliang Xu,
  • Zhengyang Xu,
  • Fangge Chen,
  • Jiangwei Lu,
  • Zhenyu Shen,
  • Shili Wang,
  • Liang Cao

摘要

Electrochemical machining (ECM) is an effective manufacturing method for machining hard-to-cut materials, especially for complex profiles components, such as the blisk of aero engines. Given the interplay of multiple physical fields during the ECM forming process, the governing mechanism underlying the forming process are intricate. Consequently, the design and subsequent correction of the tool cathode is the key to achieving precision manufacturing of the blisks. Manual correction of tool cathodes is a laborious and time-intensive process, which complicates the development of a quantifiable correction model. Therefore, this paper proposes a machine-learning-assisted approach to correct tool cathodes, and establishes a tool cathodes correction model using the support vector regression (SVR) algorithm. Within this model, the profile deviation of the workpiece is used as the input layer of the SVR model, and the correction amount on the tool cathode surface is used as the output layer. Subsequently, the corrected tool cathode was employed in the blisk ECM experiment. The machined workpiece had a maximum profile deviation of 0.0116 mm, and a smaller accuracy variance of the blade control line is obtained, which is 0.0302, which is 56.92% lower than the manually corrected one. The experimental results indicated that the application of a machine-learning-assisted correction method for tool cathodes enhances both the machining accuracy and repeatability in blisk ECM. Hence, it indicates that machine learning extensive potential for the refinement of tool cathode correction in blisk ECM, improving both cathode correction quantifiability and machining accuracy.

Graphical Abstract