基于CEEMDAN与样本熵的高压断路器机械故障诊断

依据高压断路器振动信号特性,提出一种自适应白噪声完整经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)与样本熵相结合的高压断路器故障特征提取方法。首先利用CEEMDAN将分闸振动信号分解成一系列内禀模态函数(intrinsic mode function,IMF),然后利用相关系数法与归一化能量筛选包含信号主要特征信息的前7阶IMF分量,求取其样本熵作为特征量,最后采用粒子群算法(particle swarm optimization,PSO)优化支持向量机(support vector machine,SVM)分类器,对断路器不同故障类型进行分类识别。实验结果表明该特征提取方法能准确提取振动信号特征量,输入PSO-SVM诊断高压断路器故障能取得良好的效果。

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

According to the vibration signal characteristics of high voltage circuit breaker, a fault characterization method for high voltage circuit breaker based on CEEMDAN and sample entropy is proposed. Firstly, the switching vibration signal is decomposed into a series of IMF by CEEMDAN. Then the first seven order IMF components containing the main feature information of the signal are screened by correlation coefficient method and normalized energy and calculate the sample entropy as the feature. Finally, the particle swarm optimization (PSO) is used to optimize the support vector machine(SVM) to identify the different operating states of the circuit breaker. The experimental results show that the feature extraction method can accurately extract the vibration signal features, and input PSO-SVM to diagnose the fault of high voltage circuit breaker can achieve good results.