Research On Fault Diagnosis Method Of Rolling Bearing Based On Feature Enhancement And MPE

The normal operation of rotating machinery and equipment is directly related to whether industrial production can be availably realized.Due to the harsh working environment,rotating machinery is prone to defects or failures.Therefore,the research on fault diagnosis technology is vital and can enhance industrial production efficiency.And at the same time assure production security.If we can perceive the failure of rotating machinery equipment in due course and supplant damaged parts in the early stage through the homologous diagnosis technology method,accidents can often be evaded.Therefore,this paper studies the fault diagnosis methods of rotating machinery and equipment.Taking the rolling bearing of representative familar parts in rotating machinery as the prime research object,the fault diagnosis experimental platform is built,and the methods are verified by collecting the vibration data of the platform.The collected vibration signals are researched on noise reduction,fault feature extraction and pattern recognition technology,and the effective features that can characterize the disparate states of rolling bearings are extracted and fault classification is realized.The main details of this article are as follows:(1)In this paper,a rolling bearing fault feature enhancement method based on variational mode decomposition(VMD)and minimum entropy deconvolution(MED)is proposed.The vibration signal is decomposed by using the VMD to complementally realize the signal reconstruction of the effective component.On this basis,the reconstructed signal is deconvoluted by MED to enhance the shock characteristics of the rolling bearing fault vibration signal.After the above analysis,the envelope spectrum of the noise reduction signal is extracted,which can efficaciously extract the rolling bearing characteristic frequency and high-order harmonics.(2)In this paper,the VMD and multi-scale permutation entropy(MPE)algorithm are used to analyze the vibration signal effectively.VMD is performed on the vibration signal of the rolling bearing to obtain the effective components containing information of different frequency bands,and the permutation entropy of the effective components of the VMD is calculated to realize feature extraction.In view of the fact that the single scale permutation entropy of the effective components of the VMD signals among multiple samples has a large characteristic difference,this paper proposes the idea of quantifying its characteristics by using the mean value of the effective component MPE,and the obtained feature vector reduces the characteristic value difference between different samples of the same kind of signals.(3)The unreasonable parameter setting of MPE will cause large errors in the calculation consequence.Aiming at the phenomenon that MPE is readily affected by the two parameters of embedding dimension and time delay,an improved MPE is proposed.Rolling bearing fault feature extraction method optimizes the selection of embedding dimension and time delay of MPE,which realizes the potent extraction of rolling bearing signal fault features.(4)In this paper,a fault diagnosis method based on improved multi-scale permutation entropy and kernel extreme learning machine(KELM)is proposed.The optimized MPE of some samples is input into KELM as training samples for training,and the optimized MPE of some samples is input into the training model as test samples for classification,so as to effectively determine the common fault types of rolling bearing.Compared with other fault feature extraction and pattern recognition methods,the proposed method has higher fault diagnosis rate.