Study On Fault Diagnosis Method Of Rolling Bearing Based On EEMD And SVM

Rolling bearing is one of the most widely used general mechanical components in all kinds of rotating machinery,whether its running state is normal or not often directly affects the operation of the whole mechanical equipment.Therefore,a further study of rolling bearing fault diagnosis technology is not only very important for the safe operation of the whole machine,but also is of great significance for effectively avoiding the happening of the accident.Based on the classification model of support vector machine,a fault diagnosis method of rolling bearing is designed,which is based on ensemble empirical mode decomposition(EEMD),genetic algorithm(GA)and support vector machine(SVM),this method improves the classification accuracy of rolling bearing fault diagnosis to some extent.In the feature extraction section of the rolling bearing,the ensemble empirical mode decomposition and threshold noise reduction technique were used to suppress the noise and mode aliasing in the feature extraction process.On the basis of the original,this paper proposes an improved ensemble empirical mode decomposition method.This method extracts more useful information,which is helpful to improve the accuracy of rolling bearing fault diagnosis.In the feature value selection part of rolling bearing,the four feature values of energy entropy,sample entropy,permutation entropy and kurtosis are compared by cluster graph.Finally,the energy entropy with the best clustering effect is selected as the input vector of the rolling bearing fault diagnosis model.In the state recognition section of rolling bearing,genetic algorithm and particle swarm optimization algorithm are used to optimize the parameters of support vector machine,and two fault classification models of rolling bearing are constructed.By comparing the two classification models,the genetic algorithm with higher fitness is selected to optimize the parameter of SVM.In this paper,taking the vibration signal of rolling bearing as the research object,a classification model is constructed by combining the improved integrated empirical mode decomposition method with optimized support vector machine(SVM).By selecting the eigenvalues with good clustering as the input vector of the model,it is verified that this method has high accuracy for fault diagnosis of rolling bearings.