Scholarlink

This paper presents a diagnostic methodology predicated on the demagnetization defects of permanent magnet synchronous linear motors (PMLSM). Firstly, a three-dimensional finite element model of the PMSLM is established to carry out demagnetization fault analysis, and the upper, middle, and lower stray magnetic field signals are extracted as the demagnetization fault signals of the motor’s permanent magnet array. The one-dimensional stray demagnetization signal is transformed into a two-dimensional image using the relative position matrix conversion method, which improves the analysis of fault features. Next, the EfficientNetV2 deep learning classification model is employed for fault classification, enabling precise diagnosis and identification of demagnetization faults. Finally, simulation comparison experiments confirm the superiority and reliability of this approach.

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

Demagnetization Fault Diagnosis of PMSLM Based on Relative Position Matrix Transformation and EfficientNetV2 Classification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Xianhong Wu,
  • Juncai Song,
  • Siliang Lu

摘要

This paper presents a diagnostic methodology predicated on the demagnetization defects of permanent magnet synchronous linear motors (PMLSM). Firstly, a three-dimensional finite element model of the PMSLM is established to carry out demagnetization fault analysis, and the upper, middle, and lower stray magnetic field signals are extracted as the demagnetization fault signals of the motor’s permanent magnet array. The one-dimensional stray demagnetization signal is transformed into a two-dimensional image using the relative position matrix conversion method, which improves the analysis of fault features. Next, the EfficientNetV2 deep learning classification model is employed for fault classification, enabling precise diagnosis and identification of demagnetization faults. Finally, simulation comparison experiments confirm the superiority and reliability of this approach.