Scholarlink

<p>The acceleration vibration signals acquired by the Micro-Electro-Mechanical System (MEMS) sensor may be vulnerable to noise and interference due to the integration requirements of the Electric Drive System (EDS). A new fault feature frequency band extraction process, named Particle Swarm Optimization-Spectral Gini Index (PSO-SGI), was proposed. The spectral Gini index (SGI) was utilized as the fitness function in an iterative optimization process to ascertain the center frequency of the optimal frequency band. The simulation signals demonstrate a comparative analysis of multiple indicators to characterize the features of bearing faults and assess the influence of bandwidth. A bench test of EDS with bearing fault was designed, and vibration signals were collected by both Integrated Electronics Piezo-Electric (IEPE) and MEMS sensors. Compared with Fast Kurtogram, Protrugram, and Autogram, the proposed method has the accuracy and universality of fault feature identification for different typical faults under low quality signals.</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 Fault Diagnosis in Electric Drive Systems Utilizing an Optimal Band Extraction Method for Low-Quality Signals

  • 期刊论文

  • 发表日期：

    2025年4月8日

查看全文
  • Cong Yue,
  • Jingwen Kang,
  • Guoqiang Tong,
  • Ping Cheng

摘要

The acceleration vibration signals acquired by the Micro-Electro-Mechanical System (MEMS) sensor may be vulnerable to noise and interference due to the integration requirements of the Electric Drive System (EDS). A new fault feature frequency band extraction process, named Particle Swarm Optimization-Spectral Gini Index (PSO-SGI), was proposed. The spectral Gini index (SGI) was utilized as the fitness function in an iterative optimization process to ascertain the center frequency of the optimal frequency band. The simulation signals demonstrate a comparative analysis of multiple indicators to characterize the features of bearing faults and assess the influence of bandwidth. A bench test of EDS with bearing fault was designed, and vibration signals were collected by both Integrated Electronics Piezo-Electric (IEPE) and MEMS sensors. Compared with Fast Kurtogram, Protrugram, and Autogram, the proposed method has the accuracy and universality of fault feature identification for different typical faults under low quality signals.