基于最小二乘支持向量机的故障电机诊断方法  

Diagnosis of vibration fault for asynchronous motors based on LS-SVM

作　　者：许允之[1] 方磊[1] 谭风雷[1] 戚慧[1] 葛垚[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221116

出　　处：《实验技术与管理》2012年第5期35-37,共3页Experimental Technology and Management

摘　　要：由于从电机的频谱无法区别出故障电机,因而使用CZT变换(线性调频Z变换)分析采集到的电机数据,判断出电机是否有故障,对电机进行了分类;把采集到的数据分类后训练最小二乘向量机,再把相同维数的数据送入训练好的最小二乘向量机进行判断,最终得出用最小二乘向量机进行电机的故障诊断的准确性,从而说明了用最小二乘向量机进行故障诊断的可行性和可靠性。Because the failure of motor can not be distinguished from the spectrum, the analysis of CZT-transform into motor data is started with, it can figure out whether the motor is at fault or not. The motor has been classified. This article describes the classification training after collected data least squares vector machines, then the dimension data into the same training good judgment of least squares vector machines. Finally, this article points out the feasibility and realiability of this diagnosis by using the LS-SVM.

关 键 词：最小二乘支持向量机 CZT变换 训练结果 准确率 

分 类 号：TM307[电气工程—电机] TP18[自动化与计算机技术—控制理论与控制工程]