论文部分内容阅读
针对传统的K均值聚类方法无法解决故障特征数据维数高,在故障样本交叠严重时多分类性能较差的问题,在电路故障特征数据预处理阶段,提出了一种类互均衡核聚类预处理方法,不仅克服了传统方法中分配不均或漏分问题,而且解决了特征数据维数高带来的奇异性问题,有效地提高了故障样本交叠时的多分类聚类性能.在此基础上,设计了一种用于模糊支持向量机的核密度函数,实现多故障的分类.将该方法应用于国际标准电路中的CTSV(continuous-time state-variable)滤波器电路故障诊断中.结果表明,该方法能突出不同故障的特性,具有很好的故障识别率.
Traditional K-means clustering method can not solve the problem of high dimension of fault feature data and poor multi-classifying performance when fault samples overlap seriously. In the stage of preprocessing of circuit fault feature data, The preprocessing method not only overcomes the problem of uneven distribution or leakage in traditional methods, but also solves the singularity problem caused by the high dimensionality of feature data and effectively improves the performance of multi-class clustering when fault samples overlap. Based on this, a kernel density function for fuzzy support vector machine is designed to realize the classification of multiple faults.This method is applied to the fault diagnosis of CTSV (Continuous-time state-variable) filter circuit in international standard circuits The results show that this method can highlight the characteristics of different faults and has a good fault recognition rate.