摘要
识别局部放电的类型对变压器状态评估十分重要。文中构造了四种变压器局部放电实物模型,从放电信号中提取18个统计特征量,使用基于变量预测模型的模式识别方法(Variable Predictive Model based Class Discriminate method,VPMCD)完成局部放电信号的分类。对比实验结果表明,VPMCD方法在识别率和计算效率均高于BP神经网络。
Partial discharge pattern recognition is very important for state evaluation of transformer. In this paper, 18 types of statistical characteristics are extracted from PD signals which are collected from four kinds of typical partial discharge models in laboratory, and partial discharge patterns are recognized by variable predictive model based class discriminate method (VPMCD). Comparative analysis results demonstrate that VPMCD algorithm gains more recogni- tion rate and better computational efficiency than BP neural network.
出处
《电测与仪表》
北大核心
2017年第8期47-51,共5页
Electrical Measurement & Instrumentation
关键词
变量预测模型
变压器
局部放电
模式识别
variable predictive model, transformer, partial discharge, pattern recognition