Scholarlink

<p>This paper presents a highly efficient method for the rapid detection and localization of open circuit and short-circuit faults in a single-phase nine-level switched-capacitor multilevel inverter, incorporating fault reconfiguration capabilities. The proposed system leverages data-driven techniques that utilize the inverter voltage signal as raw input data, processed through variable mode decomposition (VMD) for extracting critical information and features from the voltage signal. VMD decomposes the voltage signal into three intrinsic mode functions (IMFs-1 to IMFs-3), with IMF-1 providing essential information for fault analysis due to its high sensitivity to disturbances. IMF-1 is subjected to singular value decomposition to extract singular values, which are then transformed into an analytic signal and fed into a machine learning-based fuzzy inference system (FIS). The FIS accurately detects and localizes faulty switches and triggers the reconfiguration process by activating redundant switches or alternative paths, ensuring that output power is preserved and capacitor voltages remain balanced both pre- and post-fault. The proposed method undergoes a rigorous validation process, including comprehensive and competitive analysis, supported by both extensive experimental and simulation results, which demonstrate the robustness and effectiveness of the proposed topology.</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

A data-driven method for fault detection and localization in reconfigurable switched-capacitor inverter

  • 期刊论文

  • 发表日期：

    2025年4月21日

查看全文
  • Vikram Singh,
  • Anamika Yadav,
  • Shubhrata Gupta

摘要

This paper presents a highly efficient method for the rapid detection and localization of open circuit and short-circuit faults in a single-phase nine-level switched-capacitor multilevel inverter, incorporating fault reconfiguration capabilities. The proposed system leverages data-driven techniques that utilize the inverter voltage signal as raw input data, processed through variable mode decomposition (VMD) for extracting critical information and features from the voltage signal. VMD decomposes the voltage signal into three intrinsic mode functions (IMFs-1 to IMFs-3), with IMF-1 providing essential information for fault analysis due to its high sensitivity to disturbances. IMF-1 is subjected to singular value decomposition to extract singular values, which are then transformed into an analytic signal and fed into a machine learning-based fuzzy inference system (FIS). The FIS accurately detects and localizes faulty switches and triggers the reconfiguration process by activating redundant switches or alternative paths, ensuring that output power is preserved and capacitor voltages remain balanced both pre- and post-fault. The proposed method undergoes a rigorous validation process, including comprehensive and competitive analysis, supported by both extensive experimental and simulation results, which demonstrate the robustness and effectiveness of the proposed topology.