Scholarlink

The three-phase voltage source inverters are the most popular power converters, and they are used in multiple industrial applications such as electric vehicles, HVDC transmission lines, UPS, etc.. The reliability of power electronic equipment is very significant, and to guarantee the stability of a three-phase VSI system, it is essential to identify and diagnose the fault as soon as possible. This article uses a Machine learning(ML) based fault diagnosis technique for 3-phase VSI. The ML technique is used in this work due to its merits, such as high accuracy, less computation time, and reduced complexity. This work uses feature selection techniques, like the 3D Plot method and DQ0 transform. The output current signals of the three-phase voltage source inverter are chosen as the input fault characteristic signals. Mean and phase angles are selected as features to diagnose the faulty switch. The trained features are utilized in an ML technique to categorize the faults. The results show that the 3D plot-ML approach diagnoses the faults of VSI in less time than the DQ0 transform-based ML technique.

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

Fault Diagnosis in Three-Phase Voltage Source Inverter Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • V. Sudha,
  • V. Gomathi,
  • A. Kannabiran,
  • R. Valarmathi

摘要

The three-phase voltage source inverters are the most popular power converters, and they are used in multiple industrial applications such as electric vehicles, HVDC transmission lines, UPS, etc.. The reliability of power electronic equipment is very significant, and to guarantee the stability of a three-phase VSI system, it is essential to identify and diagnose the fault as soon as possible. This article uses a Machine learning(ML) based fault diagnosis technique for 3-phase VSI. The ML technique is used in this work due to its merits, such as high accuracy, less computation time, and reduced complexity. This work uses feature selection techniques, like the 3D Plot method and DQ0 transform. The output current signals of the three-phase voltage source inverter are chosen as the input fault characteristic signals. Mean and phase angles are selected as features to diagnose the faulty switch. The trained features are utilized in an ML technique to categorize the faults. The results show that the 3D plot-ML approach diagnoses the faults of VSI in less time than the DQ0 transform-based ML technique.