基于主成分分析和支持向量机分类模型的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on Classification Model of Principal Component Analysis and Support Vector Machine

为了提高滚动轴承故障诊断的准确率,提出一种基于主成分分析(principal component analysis,PCA)和支持向量机(support Vector machine,SVM)模型的滚动轴承故障诊断方法。通过比较不同方法计算的标准差和拉依达准则对数据进行误差分析,利用MATLAB软件中的PCA函数对数据进行主成分分析,将8个原始变量降维成3个综合变量,分别从降维前和降维后的输入属性数据中随机选取70%的数据作为训练集来建立SVM分类模型和PCA-SVM分类模型,而把剩余的30%作为测试集来对模型的性能进行仿真测试。MATLAB仿真测试的结果表明,PCA-SVM模型的分类效果更好,其预测正确率对于绝大多数故障诊断来说是可以接受的,可以作为一种故障诊断的评价标准。

In order to improve the accuracy of rolling bearing fault diagnosis,Machine learning demonstrates its absolute advantage in equipment fault diagnosis with its excellent learning ability.Therefore,a fault diagnosis study of rolling bearing based on principal component analysis(PCA)and support vector machine(SVM)model was proposed.First,error of the data was analyzed by comparing the standard deviations that calculated by different formulas in the Pauta Criterion.Secondly,principal component analysis on the data was conducted using the PCA function in MATLAB software,from which eight original variables were reduced to three integrated variables.Finally,70% of the data from the input attribute data before and after the dimension reduction were randomly selected as the training set,with which the SVM classification model and the PCA-SVM classification model were established,and the remaining 30% was used as the test set of the models.Results of MATLAB simulation test show that the SVM model established by data without principal component analysis has a better classification effect,and its prediction accuracy is acceptable for most fault diagnosis;thus,this method can be used as evaluation criteria for fault diagnosis.