In order to make up for the deficiencies in the selection of feature quantities and intelligent algorithm models in the early identification of bearing collar internal plunge grinding chatter,a method based on energy en-tropy and mean square frequency combines with the artificial bee colony(ABC)optimization support vector ma-chine(SVM).Firstly,the acoustic emission and vibration signals of the inner circle of the bearing collar during the tangential grinding process are extracted by the feature extraction method,and four feature parameters of en-ergy entropy,time domain parameters,energy occupation ratio and mean square frequency are extracted.Sec-ondly,the four feature parameters are composed two by two into a feature vector for SVM analysis and the re-sults are analysed using accuracy,Kappa coefficient and confusion matrix.Finally,the mean square frequency and energy entropy are imported into different intelligent algorithms for monitoring and analysis.The results show that the feature vectors composed of mean square frequency and energy entropy have the best effect on the identification of early chattering in the inner circle plunge grinding of bearing rings,and the combination of ABC-SVM can achieve 100%identification,which provides an effective method for online monitoring of early chatter in the inner circle plunge grinding of bearing rings.
NETL
NSTL
万方数据
