基于CEEMDAN样本熵与FWA-SVM的高压断路器机械故障诊断

针对振动信号判别断路器机械故障过程受干扰影响的特征提取问题,提出一种自适应白噪声完整集合经验模态分解(CEEMDAN)与样本熵相结合的故障特征提取方法。通过CEEMDAN提取若干反映断路器操动过程机械状态信息的本征模态函数(IMF)分量,依据各IMF相关系数与能量分布,将前7阶IMF分量进行小波包软阈值去噪,计算其样本熵作为特征量,最后采用基于免疫浓度思想的烟花算法(FWA)优化支持向量机(SVM)分类器,对断路器不同运行状态进行分类识别。实验结果表明:基于CEEMDAN样本熵特征对于信号干扰不敏感,FWA-SVM诊断方法对于高压断路器分闸操动过程故障辨识效果良好。

Mechanical fault diagnosis of high voltage circuit breaker based on CEEMDAN sample entropy and FWA-SVM

Aiming at the problem that feature extraction is easy to be affected by jamming signal in the process of mechanical fault identification based on vibration signal of circuit breaker, a fault feature extraction method that combines CEEMDAN(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) and sample entropy is proposed. Several IMF(Intrinsic Mode Function) components that reflect the mechanical state information of operating process are extracted by CEEMDAN. The top 7th order components are selected based on the energy distribution and correlation coefficients and denoised by wavelet packet soft threshold, and their sample entropies are calculated as the feature quantities. The FWA(FireWorks Algorithm) based on immune concentration is used to optimize the support vector machine classifier to identify the different operating states of the circuit breaker. The experimental results show that the features based on CEEMDAN sample entropy extraction are not sensitive to signal interference, and the FWA-SVM diagnosis method has a good effect on fault identification of high voltage circuit breakers.