基于小波分析诊断和支持向量机的避雷器在线监测数据故障分析

Fault Analysis of MOA Online Monitoring Data Based on Wavelet Analysis Diagnosis and Support Vector Machine

摘要:

金属氧化物避雷器（MOA）的健康状况对变电站中线路和设备有着巨大影响，对MOA在线监测数据进行故障诊断能够有效减少因事故引起的负荷损失。文章以基于Penalized软阀值的小波包对某变电站的MOA在线数据进行滤波，从统计角度得到滤波后的故障信号特征，采用支持向量机对信号进行分类，能够在MOA完全失效前诊断出故障出现时间点，为计划停电检修提供依据。

Abstract:

The health of metal oxide surge arresters (MOA) has a huge impact on the lines and equipment in substations. The fault diagnosis of MOA online monitoring data can effectively reduce the load loss caused by the accident. The wavelet packet based on Penalized soft threshold is used to filter the MOA online data of a substation, and the signals are classified by support vector machines. It can diagnose the time point of failure before the MOA completely fails, which can provide a basis for the planned power outages.