Study Of Fault Feature Extraction For Rolling Bearing Based On Vibration Signal Analysis

Rolling bearing is one of the important parts of mechanical equipment.Once the fault occurs,it will directly affect the normal operation of mechanical equipment,so it is very important to monitor and diagnose the fault of rolling bearing.According to the related research data,nearly 30% of the faults in rotating machinery are caused by the damage of rolling bearings.Because of the complexity of the operating conditions of most machines and equipment,the fault signals of rolling bearings are nonlinear and unstable.If the running condition of the rolling bearing can be effectively monitored,the vibration signals produced by the rolling bearing analyzed and the effective characteristic information extracted,the obtained effective characteristic information diagnosed,the faults found in time and maintenance measures taken,it is of great significance for bearing system to run healthily and stably and prevent bearing accidents.In this paper,the feature extraction of Vibration Signal of rolling bearing fault is analyzed and studied.The main work is as follows:(1)The Working Environment of rolling bearing is bad.The vibration signal noise is big.The signal-to-noise ratio is low,thereafter,the characteristic parameter is easy to be buried when extracting the vibration signal,which causes the difficulty of fault feature extraction.In order to solve this problem,the method of Maximum Correlated Kurtosis Deconvolution(MCKD)is used to denoise the original signal.The method can enhance the periodic impact characteristics of the signal.In order to improve the performance of the MCKD filter and determine the relevant parameters,the adaptive MCKD algorithm is used to filter the vibration signal,and the Kurtosis is taken as the index to optimize the length of the filter.Then,the simulation test based on the rolling bearing fault model verifies the filtering performance of the adaptive MCKD.The test results show that the adaptive MCKD has a good filtering and denoising effect.(2)The vibration signals of rolling bearing fault are complex,unstable and non-linear,which makes it difficult to extract fault features.To solve this problem,a feature extraction method based on variable Mode Decomposition(VMD)and Dispersion Entropy(DE)is proposed.And The fault model of rolling bearing is simulated and analyzed to observe the effect of VMD decomposition.In order to quantify the characteristics of the decomposed components,the distribution entropy of the decomposed components is used as the characteristic matrix of the signal.According to the data of bearing data center of Western Reserve University,the fault signal is denoised by using adaptive MCKD,and then the denoised signal is decomposed by VMD.Finally,the eigenvalues of the decomposed components are constructed based on the entropy of dispersion,and the discrimination degree of the EIGENVALUES is verified.In order to automatically identify rolling bearing fault patterns,this paper classifies the feature set of rolling bearing based on the Support Vector Machine Machine(SVM).Based on the above algorithm,features are extracted from the simulation data of Inner Ring Fault,Outer Ring Fault,Rolling Element Fault and normal condition.The training set and test set are constructed to verify the effectiveness of the feature extraction algorithm proposed in this paper.