The time domainindicatorsof vibration signalof rolling bearing have a strong correlation and more redundant information. So a bearing fault diagnosis method based on Principal Component Analysis ( PCA) and Support Vector Machine ( SVM) is proposed. Firstly, vibration signalsare measuredby acceleration sen-sor on the fault simulation test bench, and then the principal component are extracted by the method of PCA based on the basic time-domain features of vibration signals. Finally, bearing fault diagnosis is achieved by usingSVM. Experiment results show that the recognition rate is beyond 90% for four bearing conditions, the method based on PCA and SVM is very effective for rolling bearing fault diagnosis.%针对滚动轴承振动信号的时域指标之间存在很强的关联性,冗余信息较多,采用主成分分析结合支持向量机实现了滚动轴承故障的准确诊断. 首先在故障模拟试验台测量振动信号,然后提取振动信号的12个时域特征,对12个基本时域特征进行主成分分析,提取累计贡献率≥95的特征值信息作为主成分. 最后将提取的精简特征作为支持向量机的输入,实现对不同轴承故障的分类识别.实验结果证明针对四种轴承状态,识别率达到90%,提出的结合PCA-SVM是一种有效的滚动轴承故障诊断方法.
展开▼
