Scholarlink

<p>This study addresses the issue of insufficient prediction accuracy caused by environmental noise in tool wear prediction. It proposes a prediction model integrating signal processing, optimization algorithms, and deep learning. First, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) combined with wavelet threshold denoising (WTD) is employed to decompose and denoise the original signal. Afterward, multi-sensor data feature extraction and fusion are performed using multi-domain feature extraction and Kalman filtering to generate more robust fused data. Finally, the dung beetle optimizer (DBO) algorithm is used to optimize the hyperparameters of the gated recurrent unit (GRU) network, thereby establishing an efficient neural network prediction model. Experimental results demonstrate that the GRU network model, enhanced by the DBO algorithm, significantly outperforms traditional deep learning models in tool wear prediction accuracy, offering a reliable reference for tool replacement in actual metal-cutting processes.</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

Investigation on tool wear prediction model based on IW-DBO-GRU

  • 期刊论文

  • 发表日期：

    2025年9月1日

查看全文
  • Hua Huang,
  • Min Nie,
  • Peidong Yang,
  • Zhiye Lu,
  • Yi jun Xia

摘要

This study addresses the issue of insufficient prediction accuracy caused by environmental noise in tool wear prediction. It proposes a prediction model integrating signal processing, optimization algorithms, and deep learning. First, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) combined with wavelet threshold denoising (WTD) is employed to decompose and denoise the original signal. Afterward, multi-sensor data feature extraction and fusion are performed using multi-domain feature extraction and Kalman filtering to generate more robust fused data. Finally, the dung beetle optimizer (DBO) algorithm is used to optimize the hyperparameters of the gated recurrent unit (GRU) network, thereby establishing an efficient neural network prediction model. Experimental results demonstrate that the GRU network model, enhanced by the DBO algorithm, significantly outperforms traditional deep learning models in tool wear prediction accuracy, offering a reliable reference for tool replacement in actual metal-cutting processes.