Scholarlink

<p>Wheelset bearings are core components of train bogies, playing a crucial role in load transmission and motion conversion. Their operating status directly affects train safety and transmission device performance. To address the difficulty in accurately extracting early fault features, a fault diagnosis method based on the fusion of Complementary Ensemble Empirical Mode Decomposition (CEEMD) and Resonance Sparse Decomposition (RSSD) optimized by the Crown Porcupine Optimizer (CPO) is proposed. First, the original signal is preprocessed by RSSD and decomposed into high-resonance periodic and low-resonance impact signals. Second, the low-resonance component containing impact signals is decomposed by CEEMD to enhance impact fault features. Third, Intrinsic Mode Functions (IMFs) are screened and reconstructed, and envelope demodulation is performed on the reconstructed signal to extract fault features. Then, CPO is used for global optimization of the subjective parameters of CEEMD and RSSD. Finally, the method is verified by MATLAB-simulated signals and the CWRU bearing dataset. Results show that compared with CPO-CEEMD, CPO-RSSD and SSA-CEEMD-RSSD, the proposed method increases the average energy ratio of impact components by 16.31%, improves the signal-to-noise ratio by 2.3–3 times, and achieves a maximum fault recognition accuracy of 99.17%. It can effectively filter noise, highlight periodic fault features, and has good adaptability and robustness.</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

Fault diagnosis of train running gear wheelset bearings based on CPO-optimized CEEMD-RSSD

  • 期刊论文

  • 发表日期：

    2026年7月2日

查看全文
  • Jiandong Qiu,
  • Meng Li,
  • Minan Tang,
  • Dingwang Zhang,
  • Shutong Liu,
  • Jiaolong Wang,
  • Guangcan Lei,
  • Shusheng Xu

摘要

Wheelset bearings are core components of train bogies, playing a crucial role in load transmission and motion conversion. Their operating status directly affects train safety and transmission device performance. To address the difficulty in accurately extracting early fault features, a fault diagnosis method based on the fusion of Complementary Ensemble Empirical Mode Decomposition (CEEMD) and Resonance Sparse Decomposition (RSSD) optimized by the Crown Porcupine Optimizer (CPO) is proposed. First, the original signal is preprocessed by RSSD and decomposed into high-resonance periodic and low-resonance impact signals. Second, the low-resonance component containing impact signals is decomposed by CEEMD to enhance impact fault features. Third, Intrinsic Mode Functions (IMFs) are screened and reconstructed, and envelope demodulation is performed on the reconstructed signal to extract fault features. Then, CPO is used for global optimization of the subjective parameters of CEEMD and RSSD. Finally, the method is verified by MATLAB-simulated signals and the CWRU bearing dataset. Results show that compared with CPO-CEEMD, CPO-RSSD and SSA-CEEMD-RSSD, the proposed method increases the average energy ratio of impact components by 16.31%, improves the signal-to-noise ratio by 2.3–3 times, and achieves a maximum fault recognition accuracy of 99.17%. It can effectively filter noise, highlight periodic fault features, and has good adaptability and robustness.