Scholarlink

<p>In this work, we employ a hierarchical signal processing framework to introduce a novel method for identifying failures in rotating equipment. The methodology begins with segmenting vibration signals into sub-signals from experimental setups that include various fault conditions, such as bearing and gear problems. Each sub-signal is analyzed using the Hilbert-Huang transform (HHT) to extract time-frequency information. The signal is then reconstructed using the non-zero elements of the HHT signal. From the reconstructed signal, informative indicators are calculated to capture fault features. Singular value decomposition (SVD) is applied for data reduction to manage the high-dimensional feature space while preserving critical fault-related information. The effectiveness of the proposed method is demonstrated through experimental validation using data from a specialized test bench at the Signal and Industrial Process Analysis Laboratory (LASPI), achieving a fault diagnosis accuracy of 99.9% with supervised learning methods.</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

Enhanced diagnosis of bearing and gear faults using Hilbert-Huang transform, singular value decomposition, and supervised learning methods

  • 期刊论文

  • 发表日期：

    2025年6月18日

查看全文
  • Abdel Wahhab Lourari,
  • Bilal El Yousfi,
  • Ahmed Bouzar Essaidi

摘要

In this work, we employ a hierarchical signal processing framework to introduce a novel method for identifying failures in rotating equipment. The methodology begins with segmenting vibration signals into sub-signals from experimental setups that include various fault conditions, such as bearing and gear problems. Each sub-signal is analyzed using the Hilbert-Huang transform (HHT) to extract time-frequency information. The signal is then reconstructed using the non-zero elements of the HHT signal. From the reconstructed signal, informative indicators are calculated to capture fault features. Singular value decomposition (SVD) is applied for data reduction to manage the high-dimensional feature space while preserving critical fault-related information. The effectiveness of the proposed method is demonstrated through experimental validation using data from a specialized test bench at the Signal and Industrial Process Analysis Laboratory (LASPI), achieving a fault diagnosis accuracy of 99.9% with supervised learning methods.