摘要
针对齿轮故障诊断中的小样本事件,采用了支持向量机(SVM)方法.采集齿轮3种典型故障(断齿、磨损、剥落)和正常状态的振动信号,提取时域指标和能量特征作为SVM输入向量,并采用交叉验证(K-CV)法优化SVM参数,最终得到的故障诊断准确率为100%.结果表明SVM是一种有效的齿轮故障诊断方法.
In this paper,the support vector machine（SVM） classification algorithm has been applied to gear fault diagnosis in small samples based on vibration signals covering four working conditions,i.e.normal,broken teeth,wear and spalling.Firstly,time domain index and energy of the signals were extracted as input of SVM.Then,parameters of SVM were optimized using K-fold cross validation（K-CV） method.Finally,applied the SVM to the fault diagnosis,and the faults were recognized correctly with an accuracy of 100%.The results demonstrate that SVM is an effective method for gear fault diagnosis.
出处
《三峡大学学报（自然科学版）》
CAS
2012年第2期63-65,75,共4页
Journal of China Three Gorges University：Natural Sciences
关键词
支持向量机
交叉验证(K-CV)
故障诊断
support vector machine（SVM）
K-fold cross validation（K-CV）
fault diagnosis
