| 注册
首页|期刊导航|电子器件|基于激光扫描的模块化多电平变流器故障检测方法

基于激光扫描的模块化多电平变流器故障检测方法

郑欣 易满成 钟泽权 陈威 祁倬锐 江辉煌

电子器件2025,Vol.48Issue(1):136-141,6.
电子器件2025,Vol.48Issue(1):136-141,6.DOI:10.3969/j.issn.1005-9490.2025.01.022

基于激光扫描的模块化多电平变流器故障检测方法

Fault Detection Method of Modular Multilevel Converter Based on Laser Scanning

郑欣 1易满成 2钟泽权 2陈威 1祁倬锐 1江辉煌1

作者信息

  • 1. 广东电网有限责任公司广州南沙供电局,广东 广州 511458
  • 2. 广东电网有限责任公司广州供电局,广东 广州 510620
  • 折叠

摘要

Abstract

Aiming at the problem that the loss caused by the fault of modular multilevel converter induced by the serial connection of multiple submodules increases significantly,the fault detection method of modular multilevel converter based on laser scanning is stud-ied.According to the operation characteristics of the modular multilevel converter,the laser scanning technology is used to collect the fault distributed big data of the modular multilevel converter,such as short circuit data,open circuit data,valve trigger data,and so on,through lasers,imaging lenses,detectors,etc.The statistical characteristics of converter fault laser scanning output are obtained through adaptive restructuring of converter fault laser scanning data.The feature is input into the fault detection model based on RBF neural net-work,and the model realizes the fault detection of modular multilevel converter through learning and training process.The historical fault causes of modular multilevel converter are referred,and the fault categories of modular multilevel converter are outputed.The experi-mental results show that the fault detection accuracy of this method is higher than 95%under the condition of single fault to multiple faults,which can ensure its stable operation.

关键词

激光扫描/模块化/多电平/变流器/故障检测/神经网络

Key words

laser scanning/modularization/multilevel/converter/fault detection/neural network

分类

动力与电气工程

引用本文复制引用

郑欣,易满成,钟泽权,陈威,祁倬锐,江辉煌..基于激光扫描的模块化多电平变流器故障检测方法[J].电子器件,2025,48(1):136-141,6.

基金项目

中国南方电网科技项目+基于5G的配网数字化作业工作量智慧监管系统+086400KK52200003(GZHKJXM2020049) (GZHKJXM2020049)

电子器件

1005-9490

访问量0
|
下载量0
段落导航相关论文