基于SVM和BP神经网络的滚动轴承故障诊断  

Rolling Bearing Fault Diagnosis Based on Support Vector Machine and BP Neural Network

作　　者：刘然[1] 傅攀[1] 

机构地区：[1]西南交通大学机械工程学院,四川成都610031

出　　处：《机械工程与自动化》2014年第6期132-134,138,共4页Mechanical Engineering & Automation

基　　金：中央高校基本科研业务费专项基金资助项目(SWJTU12CX039)

摘　　要：为更好地对滚动轴承进行状态监测和故障诊断,采集3种不同状态下的滚动轴承振动信号,根据振动信号特点提取其时域和频域的相关特征,然后分别利用SVM(支持向量机)和BP神经网络进行模式识别。不断减少每种状态下训练样本集的个数,利用2种不同的方法进行模式识别。当每种状态下的样本个数为3个时,支持向量机仍然能准确地将测试样本进行分类,而BP神经网络完全无法识别。实验结果表明,支持向量机比BP神经网络更适合于小样本的故障诊断。In order to make rolling bearing condition monitoring and fault diagnosis better ,three kinds of different states vibration signals of rolling bearing were collected , the features of different states＇ vibration signals in time and frequency domain were extracted .The support vector machine（SVM） and BP neural network was used to conduct the pattern recognition .In the case of decreasing the number of training samples in each state ,two methods were applied to pattern recognition .When the number of samples in each state was 3 ,the SVM was still able to classify the test samples accurately ,but the BP neural network could’t identify completely .Experimental results show that SVM is more suitable for small samples in fault diagnosis than BP neural network .

关 键 词：滚动轴承 BP神经网络 故障诊断 SVM 

分 类 号：TH133.33[机械工程—机械制造及自动化] TH165.3