Study On Predicting The Remaining Service Life Of Rolling Bearing Based On Signal Processing Method By Vibration Signal

Rolling bearing has a universal application in actual production. It directly affects the health of the entire system as an important part in chemical rotating machine. Therefore, prediction of rolling bearing has a great importance. The research object is the vibration signal which encompasses a wealth of characteristic information. In this paper, rolling excitation mechanism is explained and illustrated. The regular of vibration characteristics changed with the evolution of degradation are revealed. The prediction of remaining service life is achieved by processing the vibration signal, extracting degradation index and using prediction model based on data-driven method. Life prediction has a great significance in the design and manufacture of rolling bearings. It has a great help to establish assessment criteria life, improve equipment and machinery performance.1. The degradation characteristic is revealed from the view of vibration amplitude, frequency and the energy. The feature parameters can reflect the degradation status in the whole life cycle of rolling bearing, so the performance of feature parameters is compared in the whole life. The results show that kurtosis, root mean square, frequency of the center of gravity, variance of frequency and energy of IMF have a relatively better performance than others in the degeneration of rolling bearing, but relying solely on a single characteristic parameter in one domain cannot be effectively characterized as performance indicator because single characteristic is often unable to take into account the sensitivity and stability, and it is active only for a defect or fault. In actual operation, because of the complexity and variability of rolling bearing, it is not effectively that one parameter in a single domain is selected as performance indicator.2.The SPSS software is used to the correlation analysis of kurtosis, root mean square, frequency of the center of gravity, variance of frequency and energy of IMF1,IMF2,IMF3,IMF4. The conclusion that they are relevant is deduced. The correlation analysis shows that the characteristics of these parameters can reflect the information of degradation in different degrees. Principal component analysis is used to fuse these parameters and established the PCA model. The validation data that represent different degrees of fault data are take in the model to prove the effectiveness of the model, and then the performance indicator of degradation is built based on the first principal component.3. Reliability Analysis and Proportional Harzards Model based on Weibull distribution are used to analyze the life situation of rolling bearing. The covariates in the Proportional Harzards Model are the performance indicator of degradation and its process variation factor. In this way, the relationship between running state and the residual life of rolling bearing is established, and then a fitting curve is obtained to predict the remaining service life of rolling bearing. The actual experimental data demonstrate the effectiveness of the method.