基于WA-ABC-LSSVM的电气故障诊断  

Electrical Fault Diagnosis Based on WA-ABC-LSSVM

作　　者：史增芳[1] 姜岩蕾[1] 

机构地区：[1]河南工业职业技术学院机电工程学院,河南南阳473009

出　　处：《电子器件》2016年第3期728-731,共4页Chinese Journal of Electron Devices

基　　金：河南省科技攻关重点项目(122102210497)

摘　　要：为了改善当前电气故障诊断的效果,提出一种基于小波消噪和人工蜂群优化最小二乘支持向量机的电气故障诊断方法(WA-ABC-LSSVM)。首先收集电气状态信息,并采用小波变换对其进行去噪处理,消除噪声的干扰,然后提取电气状态中的特征,并且进行归一化处理,最后采用训练样本对最小二乘支持向量机进行训练,采用人工蜂群算法优化最小二乘支持向量机参数,建立电气故障诊断分类器。仿真实验结果表明,本文方法可以较好描述电气系统的工作状态,诊断性能要明显优于其它的电气故障诊断方法。In order to improve the effect of current electrical fault diagnosis,a method of electrical fault diagnosis is presented based on artificial bee colony optimizing least squares support vector machine. Firstly,the electrical sta-tus information is collected,and wavelet transform is used to eliminate the noise,and then extract the features of the electrical state is extracted,and the normalized processing is carried out,finally,the training samples are used to train the least squares support vector machine which artificial bee colony algorithm is used to optimize the parame-ters to establish electrical fault diagnosis classifier. Simulation results show that the proposed method can better de-scribe the working state of the electrical system,and the diagnostic performance is better than the other methods of electric fault diagnosis.

关 键 词：电气系统 故障分类 小波去噪 状态特征 故障分类器 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]