Scholarlink

<p>Thin-walled aerospace components are extensively utilized for their lightweight properties and superior performance characteristics. However, these components are susceptible to chatter during milling operations, which significantly impairs machining quality. To address this challenge, this study develops a cutting force prediction model that accounts for the structural characteristics of discrete-edge end mills. The model incorporates a chip-split groove influence factor, achieving a 23% improvement in prediction accuracy compared to conventional models when evaluating end mills with chip-split grooves. Building upon this foundation, a physics-based milling stability prediction model was established. Furthermore, the study presents an innovative integration of deep learning and physical modeling, employing Bayesian methods to enable stability prediction with limited sample data. The optimal milling parameters are obtained using the whale optimization algorithm (WOA), with the desired stability domain as a constraint and both material removal rate and tool life as optimization objectives. This research demonstrates that the proposed modeling methodology and optimization strategy offer an effective approach for enhancing the stability and accuracy of thin-walled aerospace component milling operations.</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

Prediction method for milling stability of discrete-edge end mills based on physical model and deep learning

  • 期刊论文

  • 发表日期：

    2025年8月16日

查看全文
  • Xiangfu Fu,
  • Enyi Chen,
  • Minli Zheng,
  • Shuo Li,
  • Chenglong Wang

摘要

Thin-walled aerospace components are extensively utilized for their lightweight properties and superior performance characteristics. However, these components are susceptible to chatter during milling operations, which significantly impairs machining quality. To address this challenge, this study develops a cutting force prediction model that accounts for the structural characteristics of discrete-edge end mills. The model incorporates a chip-split groove influence factor, achieving a 23% improvement in prediction accuracy compared to conventional models when evaluating end mills with chip-split grooves. Building upon this foundation, a physics-based milling stability prediction model was established. Furthermore, the study presents an innovative integration of deep learning and physical modeling, employing Bayesian methods to enable stability prediction with limited sample data. The optimal milling parameters are obtained using the whale optimization algorithm (WOA), with the desired stability domain as a constraint and both material removal rate and tool life as optimization objectives. This research demonstrates that the proposed modeling methodology and optimization strategy offer an effective approach for enhancing the stability and accuracy of thin-walled aerospace component milling operations.