基于主成分分析-支持向量机优化模型的断路器故障诊断方法研究

Research on Circuit Breaker Fault Diagnosis Method Based on Principal Component Analysis⁃support Vector Machine Optimization Model

为了准确实现电磁斥力机构真空断路器的故障诊断,针对电磁斥力机构真空断路器分闸振动信号,在正常状态、单一故障状态、多个故障共存状态下,通过小波包分解、希尔伯特-黄变换(HHT)提取振动信号能量熵向量,结合特征提取方法原理分析了两种方法提取特征量的有效性,提出了主成分分析(PCA)—支持向量机(SVM)优化模型进行故障诊断,并验证了网格搜索算法(GSA)、粒子群算法(PSO)、遗传算法(GA)3种参数寻优算法的性能。经实验测试,PCA-SVM优化模型解决了因样本特征信息存在噪声和冗余而引起的SVM识别准确率下降的问题,有效提升了测试样本的识别准确率和模型效率,具有较好的诊断效果。

In order to accurately realize the fault diagnosis of electromagnetic repulsive mechanism vacuum circuit breaker,according to the breaking vibration signal of the electromagnetic repulsion mechanism vacuum circuit breaker,the energy entropy vector of the vibration signal is extracted by wavelet packet decomposition and Hilbert⁃Huang transform(HHT)in the normal state,single fault state and multiple fault coexistence states,combining the principle of feature extraction method,the effectiveness of the two methods for extracting feature quantities is ana⁃lyzed.A principal component analysis(PCA)⁃support vector machine(SVM)optimization model is proposed for fault diagnosis.The performance of three parameter optimization algorithms,namely grid search algorithm(GSA),particle swarm optimization(PSO)and genetic algorithm(GA),is verified.Through experimental tests,the PCA⁃SVM optimization model solves the problem of SVM recognition accuracy caused by noise and redundancy of sample feature information,which effectively improves the recognition accuracy and model efficiency of test samples,and has a good diagnostic effect.