Scholarlink

<p>By virtue of their unique mechanical properties, Carbon Fiber Reinforced Polymer (CFRP)/Ti-6Al-4&#xa0;V stacks have found extensive application in the aerospace industry. However, the significant disparity in machinability between these two materials leads to exacerbated tool wear, which severely compromises both machining quality and efficiency. To enable precise and intelligent monitoring of tool wear, this paper introduces a novel method for tool state identification and prediction based on multi-modal sensing signals and deep learning. The proposed method effectively integrates a Temporal Convolutional Network (TCN), a Bidirectional Stacked Gated Recurrent Unit (Bi-stacked GRU), and a self-attention mechanism to extract the spatio-temporal correlation features from vibration, acoustic emission, and current signals. Experimental results on a custom-built dataset from the gyroscopic milling of these stacks demonstrate that the model achieves an average state identification accuracy of 98%, significantly outperforming current state-of-the-art methods. Furthermore, to achieve precise quantitative prediction of wear, we developed a deep neural network model based on a Stacked Bidirectional Long Short-Term Memory network (Stacked Bi-LSTM), which effectively establishes a non-linear mapping between the multi-channel sensor time-series data and the tool wear values. The model demonstrates exceptional predictive performance on our dataset, achieving a coefficient of determination exceeding 0.986. Moreover, ablation studies were conducted to validate the contribution of each key component. The intelligent monitoring method proposed in this study offers an effective and reliable solution to the persistent challenge of tool wear in machining CFRP/Ti-6Al-4&#xa0;V stacks. This research holds significant theoretical value and practical potential for advancing the manufacturing precision and intelligence level of aerospace components.</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

Tool Wear State Identification and Prediction in Gyroscopic Milling of CFRP/Ti-6Al-4V Laminated Material

  • 期刊论文

  • 发表日期：

    2026年2月25日

查看全文
  • Lijia Liu,
  • Wenjie Kang,
  • Quanwen Fu

摘要

By virtue of their unique mechanical properties, Carbon Fiber Reinforced Polymer (CFRP)/Ti-6Al-4 V stacks have found extensive application in the aerospace industry. However, the significant disparity in machinability between these two materials leads to exacerbated tool wear, which severely compromises both machining quality and efficiency. To enable precise and intelligent monitoring of tool wear, this paper introduces a novel method for tool state identification and prediction based on multi-modal sensing signals and deep learning. The proposed method effectively integrates a Temporal Convolutional Network (TCN), a Bidirectional Stacked Gated Recurrent Unit (Bi-stacked GRU), and a self-attention mechanism to extract the spatio-temporal correlation features from vibration, acoustic emission, and current signals. Experimental results on a custom-built dataset from the gyroscopic milling of these stacks demonstrate that the model achieves an average state identification accuracy of 98%, significantly outperforming current state-of-the-art methods. Furthermore, to achieve precise quantitative prediction of wear, we developed a deep neural network model based on a Stacked Bidirectional Long Short-Term Memory network (Stacked Bi-LSTM), which effectively establishes a non-linear mapping between the multi-channel sensor time-series data and the tool wear values. The model demonstrates exceptional predictive performance on our dataset, achieving a coefficient of determination exceeding 0.986. Moreover, ablation studies were conducted to validate the contribution of each key component. The intelligent monitoring method proposed in this study offers an effective and reliable solution to the persistent challenge of tool wear in machining CFRP/Ti-6Al-4 V stacks. This research holds significant theoretical value and practical potential for advancing the manufacturing precision and intelligence level of aerospace components.