Research On Bearing Fault Diagnosis System Based On Variational Mode Decomposition

Rolling bearings are one of the most widely used components in rotating machinery.Therefore,monitoring the working state of rolling bearings and ensuring their normal operation have important economic and practical significance for ensuring the safety of mechanical systems.Therefore,this paper takes rolling bearing as the research object,starts with the processing of bearing vibration signal,and carries out a series of research work on the feature extraction and failure mode classification of rolling bearing vibration signal.The main contents are as follows:First,on the basis of theoretical analysis and engineering application,the research background and practical significance of this topic are systematically elaborated.The research status and research progress of the application of rolling bearing's structure,vibration mechanism,diagnosis method,and methods used in the extraction of fault features of rolling bearings are presented.The development trend is more comprehensively elaborated.Second,in order to overcome the defects of modal aliasing and under-enveloping in the empirical mode decomposition(EMD)method,based on a review of traditional signal processing techniques and modern signal processing techniques,the variational modal decomposition is introduced.(Variational Mode Decomposition,VMD)method,and through simulation and comparison analysis,it is proved that the VMD algorithm can complete the frequency separation well,and is superior to the EMD method in anti-modal aliasing and impact feature detection.Thirdly,for the traditional support vector machines(SVMs),the method of cross-validation is used to obtain the optimal parameters of the grid.Although this method can obtain the global optimal solution,if you want to find the best parameters in a larger range,it will be very Time-consuming,there is a defect that the network convergence is slow and inefficient;and the generation of populations in a random way for the traditional particle swarm optimization algorithm leads to uneven population distribution,so that the convergence speed of the algorithm in the numerical optimization part is low and it is easy to fall into local optimum.This paper proposes a chaotic function replacement particle swarm algorithm to generate initial populations in a random manner.Chaos particle swarm optimization algorithm is used to replace the method of using grid optimization initialization parameters in traditional support vector machines.Using the VMD multi-scale entropy mean-bias value as a characteristic parameter as a network input for classification test,the classification results show the effectiveness of this method in the classification of failure modes of rolling bearings.Finally,this paper developed a Lab VIEW-based rolling bearing fault diagnosis platform,which can achieve the three main functions of acquisition and storage of rolling bearing vibration signals,feature extraction and failure mode classification.This platform can notonly collect rolling bearings for QPZZ-II test platform.The vibration signal is used for fault diagnosis,and has a good man-machine interface.The test results prove that the platform has good practical value.