Scholarlink

<p>In milling operation, the analysis of power signals are quite important because power signals provide the immediate information of the process that allows the operator to monitor the performance and make necessary modifications in real time. In addition, the power signals are correlated with the quality of machined surface because high power may result in tool breakage that leads to poor surface finish. Therefore, this work firstly investigates the analysis of power signals and its impact on surface characteristics during the milling of additively manufactured aluminium alloys under different conditions. The milling experiments were performed under different cooling conditions and the power signals were calculated from the main cutting forces. Then, the Swin transformer based deep learning model was used on the calculated power signals, and the performance was compared with different models to validate the accuracy of Swin transformer model. Further, the relation between cutter angle rotation and power signals were analyzed under different conditions. According to the surface analysis, the higher power ranges observed during dry machining have a major impact on surface quality and lead to rough milling surfaces. The outcomes also demonstrated that the Swin transformer shows better performance with higher accuracy and Kappa metrics with 0.0001 learning rate.</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

Deep Learning-Driven Models for Analyzing Numerical Control Power Signals and Surface Characteristics in Sustainable Milling of Additively Manufactured Al-Si Alloys

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月9日

查看全文
  • Munish Kumar Gupta,
  • Nimel Sworna Ross,
  • C. Sherin Shibi,
  • Roman Chudy,
  • Piotr Nieslony,
  • Mehmet Erdi Korkmaz,
  • Jolanta B. Krolczyk,
  • Grzgeorz M. Krolczyk

摘要

In milling operation, the analysis of power signals are quite important because power signals provide the immediate information of the process that allows the operator to monitor the performance and make necessary modifications in real time. In addition, the power signals are correlated with the quality of machined surface because high power may result in tool breakage that leads to poor surface finish. Therefore, this work firstly investigates the analysis of power signals and its impact on surface characteristics during the milling of additively manufactured aluminium alloys under different conditions. The milling experiments were performed under different cooling conditions and the power signals were calculated from the main cutting forces. Then, the Swin transformer based deep learning model was used on the calculated power signals, and the performance was compared with different models to validate the accuracy of Swin transformer model. Further, the relation between cutter angle rotation and power signals were analyzed under different conditions. According to the surface analysis, the higher power ranges observed during dry machining have a major impact on surface quality and lead to rough milling surfaces. The outcomes also demonstrated that the Swin transformer shows better performance with higher accuracy and Kappa metrics with 0.0001 learning rate.