Scholarlink

<p>Currently, the diagnosis of motorized spindle faults primarily focuses on analyzing the current vibration patterns but often neglects forecasting future vibration tendencies. This study introduces a novel vibration trend prediction model for high-speed motorized spindles to anticipate and mitigate potential vibration issues. Acknowledging the noise challenges inherent in data acquisition, this research integrates variational modal decomposition with LMS adaptive filtering and employs the state-of-the-art northern goshawk optimization algorithm for enhancement. The efficiency of this signal processing technique is experimentally validated. In terms of predicting vibration patterns, long short-term memory network architecture is utilized. Moreover, the prediction model is further optimized using the northern goshawk optimization algorithm. Comparative experiments confirm that this optimized model surpasses the standard LSTM model in accurately predicting the future vibration trends of high-speed motorized spindles. By interpreting these predictive results, it enables the proactive determination of the spindle’s vibration trend, thereby aligning with the requirements of predictive maintenance strategies.</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

Vibration trend forecasting of motorized spindle on the basis of signal processing and deep learning

  • 期刊论文

  • 发表日期：

    2025年4月2日

查看全文
  • Ye Dai,
  • Xiao Liu,
  • Jian Pang,
  • YiNing Li,
  • YanJun Lv,
  • YongNing Niu

摘要

Currently, the diagnosis of motorized spindle faults primarily focuses on analyzing the current vibration patterns but often neglects forecasting future vibration tendencies. This study introduces a novel vibration trend prediction model for high-speed motorized spindles to anticipate and mitigate potential vibration issues. Acknowledging the noise challenges inherent in data acquisition, this research integrates variational modal decomposition with LMS adaptive filtering and employs the state-of-the-art northern goshawk optimization algorithm for enhancement. The efficiency of this signal processing technique is experimentally validated. In terms of predicting vibration patterns, long short-term memory network architecture is utilized. Moreover, the prediction model is further optimized using the northern goshawk optimization algorithm. Comparative experiments confirm that this optimized model surpasses the standard LSTM model in accurately predicting the future vibration trends of high-speed motorized spindles. By interpreting these predictive results, it enables the proactive determination of the spindle’s vibration trend, thereby aligning with the requirements of predictive maintenance strategies.