摘要
GIS断路器的声纹信号通常具有多种频率成分及复杂的非线性特性,采用直接分析的方式往往只能捕捉到其中的部分特征,从而降低了故障诊断的效率。因此,提出基于声纹特征EMD提取的GIS断路器机械故障诊断方法。采用EMD算法对GIS断路器声纹信号进行处理,将复杂信号分解为多个IMF分量。基于分解后的声纹信号,利用均方根能量和线性判别分析提取关键特征,并通过Relief-F算法筛选出最具分类效能的特征集合。依据提取的声纹信号特征,确立边缘诊断特征阈值,并引入分类算法模型评估测试样本与已知故障样本的匹配程度,实现对GIS断路器机械故障的诊断。实验结果表明,该研究方法能够在模拟的多种GIS断路器机械故障类型中准确匹配实际诊断类型,并且在不同故障严重程度下均展现出较快的诊断速度,显著提升了诊断效率。
The voiceprint signal of GIS circuit breaker usually has multiple frequency components and complex nonlinear characteristics.The direct analysis can only capture some of the features,thereby reducing the efficiency of fault diagnosis.This paper proposes a GIS circuit breaker mechanical fault diagnosis method based on EMD extraction of voiceprint features.The GIS circuit breaker voiceprint signal is processed by EMD algorithm,and the complex signal is decomposed into multiple IMF components.Then RMS energy and linear discriminant analysis are used to extract key features,and the feature set with the most efficient classification is selected by the Relief-F algorithm.Accordingly,the threshold of edge diagnosis features is established,and a classification algorithm model is introduced to evaluate the matching degree between the test samples and the known fault samples,so as to realize the mechanical fault diagnosis of GIS circuit breaker.The experimental results show that the proposed method can accurately match the actual diagnosis types among various simulated types of GIS circuit breaker mechanical fault. It also shows fast diagnosis speed under different fault severity, which significantly improves the diagnosis efficiency.
作者
郭斌
刁文正
吴晓瑞
GUO Bin;DIAO Wenzheng;WU Xiaorui(SDEE Hitachi High-Voltage Switchgear Co.,Ltd.,Jinan 250103,China)
出处
《机械》
2025年第8期69-73,80,共6页
Machinery
关键词
EMD分解
声纹特征
特征提取
GIS断路器
故障诊断
EMD decomposition
voiceprint features
feature extraction
GIS circuit breaker
fault diagnosis
