Scholarlink

<p>To enhance the accuracy of bearing fault diagnosis under high-noise conditions, this study proposes an adaptive optimization approach for tuning the parameters of maximum second-order cyclostationary blind deconvolution (CYCBD). The envelope of the autocorrelation function (ACFE) is proposed to estimate the fault frequency, but the effect of ACFE is poor in noisy environments. To achieve better results, the vibration signal is decomposed via complete ensemble empirical mode decomposition (CEEMDAN). By selecting the IMFs reconstruction signal with a strong correlation with the original signal, the fault correlation features are enhanced, while the noise was effectively suppressed. A two-strategy improved grey wolf optimization algorithm (IGWO) is introduced, and the envelope spectrum characterization energy ratio (ESCER) is proposed, which is used as a fitness function for the adaptive optimization of the CYCBD input parameters N and α. This method dynamically adjusts the decomposition process to ensure the accuracy of fault feature extraction and reduce the influence of frequency interference. The bearing fault simulation signal, XJTU-SY bearing dataset, and experimental data are used to verify the effectiveness and robustness of the method in identifying bearing faults under complex noise conditions.</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

Application of parameters adaptive maximum second-order cyclostationary blind deconvolution in fault identification of rolling bearings

  • 期刊论文

  • 发表日期：

    2025年6月27日

查看全文
  • Meng-Meng Dou,
  • Yi Zhang,
  • Fei-Hu Sun,
  • Hua-Xi Zhou,
  • Hou-Chang Pei

摘要

To enhance the accuracy of bearing fault diagnosis under high-noise conditions, this study proposes an adaptive optimization approach for tuning the parameters of maximum second-order cyclostationary blind deconvolution (CYCBD). The envelope of the autocorrelation function (ACFE) is proposed to estimate the fault frequency, but the effect of ACFE is poor in noisy environments. To achieve better results, the vibration signal is decomposed via complete ensemble empirical mode decomposition (CEEMDAN). By selecting the IMFs reconstruction signal with a strong correlation with the original signal, the fault correlation features are enhanced, while the noise was effectively suppressed. A two-strategy improved grey wolf optimization algorithm (IGWO) is introduced, and the envelope spectrum characterization energy ratio (ESCER) is proposed, which is used as a fitness function for the adaptive optimization of the CYCBD input parameters N and α. This method dynamically adjusts the decomposition process to ensure the accuracy of fault feature extraction and reduce the influence of frequency interference. The bearing fault simulation signal, XJTU-SY bearing dataset, and experimental data are used to verify the effectiveness and robustness of the method in identifying bearing faults under complex noise conditions.