Research on Bearing Fault Diagnosis of Industrial Robot Based on CEEMDAN-SVM Algorithm
The transmission efficiency of industrial robots cannot be separated from bearings.In order to improve the recognition ability of bearing vibration signals,a bearing fault diagnosis method based on adaptive noise complete integrated Empirical Mode decomposition(CEEMDAN)-support vector machine(SVM)algorithm was designed for industrial robots.The vibration signal was decomposed by CEEMDAN,the IMF component containing the key information was screened based on the correlation coefficient,and the bearing fault type of industrial robot was identified by SVM.The research results show that the correlation coefficients of IMF1-IMF5 are all above 0.2,and the classification accuracy is 100％,which indicates that the algorithm can effectively detect the fault categories of industrial robot bearings.Compared with EMD and PSO methods,the classification accuracy of CEEMDAN algorithm in this paper is obviously the highest,and signal decomposition is helpful to EMDE algorithm,which verifies the effectiveness of signal decomposition.The proposed algorithm has considerable accuracy and stability in fault identification,and can be applied to other transmission fields.
NETL
NSTL
万方数据
