Scholarlink

<p>Early detection of ball-bearing faults in squirrel-cage induction motors is crucial for minimizing industrial downtime and maintenance costs. This study proposes a novel vibration-based method to diagnose three common ball-bearing faults—ball, cage, and outer ring defects. The method integrates variational mode decomposition (VMD), envelope analysis, and power spectral density (PSD) to enhance fault detection. VMD decomposes the vibration signal into distinct modes, isolating fault-related frequencies and reducing noise. An adaptive fusion strategy determines the number of modes (K) based on center frequency separation, ensuring precise isolation of fault components and mitigating mode-mixing issues. Envelope analysis extracts subtle fault signatures, while PSD identifies fault types based on characteristic frequencies. Experimental results from a 2-hp induction motor under no-load, half-load, and full-load conditions demonstrate that the method effectively detects all fault types, with significant amplitude differences in PSD spectra, particularly at low loads. Compared to Hilbert-Huang Transform and standalone envelope methods, this approach offers superior sensitivity to incipient faults, making it a reliable solution for industrial applications.</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

Ball-Bearing Fault Diagnosis of Induction Motors Using a Fusion of Variational Mode Decomposition, Envelope and Power Spectral Density Based on Vibration Signal

  • 期刊论文

  • 发表日期：

    2025年12月18日

查看全文
  • Gholam Reza Agah,
  • Akbar Rahideh,
  • Hosein Khodadadzadeh,
  • Vahid Moradzadeh Tehrani,
  • Mostafa Shahnazari,
  • Shahin Hedayati Kia

摘要

Early detection of ball-bearing faults in squirrel-cage induction motors is crucial for minimizing industrial downtime and maintenance costs. This study proposes a novel vibration-based method to diagnose three common ball-bearing faults—ball, cage, and outer ring defects. The method integrates variational mode decomposition (VMD), envelope analysis, and power spectral density (PSD) to enhance fault detection. VMD decomposes the vibration signal into distinct modes, isolating fault-related frequencies and reducing noise. An adaptive fusion strategy determines the number of modes (K) based on center frequency separation, ensuring precise isolation of fault components and mitigating mode-mixing issues. Envelope analysis extracts subtle fault signatures, while PSD identifies fault types based on characteristic frequencies. Experimental results from a 2-hp induction motor under no-load, half-load, and full-load conditions demonstrate that the method effectively detects all fault types, with significant amplitude differences in PSD spectra, particularly at low loads. Compared to Hilbert-Huang Transform and standalone envelope methods, this approach offers superior sensitivity to incipient faults, making it a reliable solution for industrial applications.