Font Size: a A A

Remaining Useful Life Prediction Of Flexible Bearing Based On Vibration Feature

Posted on:2021-04-05Degree:MasterType:Thesis
Country:ChinaCandidate:W B HuangFull Text:PDF
GTID:2392330611966196Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
Harmonic reducer is recognized as the most excellent transmission gear for robot,and flexible bearing is the key part of harmonic reducer,which has an important impact on the rotation accuracy and service life of harmonic reducer.The vibration signal of bearing contains abundant information including the performance state information.To analyze the vibration signal,extract the performance degradation sensitive features and predict the remaining useful life of flexible bearing is necessary for equipment maintenance.The main work of this paper are as follows:1.Based on the working principle of harmonic reducer and the operating condition of flexible bearing,a test rig is built,and the vibration data of flexible bearing life cycle is obtained through accelerated life tests.2.The trend of 14 time-domain features,4 frequency-domain features and sample entropy features based on wavelet packet transform of vibration signal are compared through correlation analysis.The result shows that the wavelet packet sample entropy of vibration signal has the largest correlation coefficient with the test time,so it is selected as the performance degradation sensitive feature.3.Aiming at uncertain failure values of the performance degradation sensitive feature,a three parameter Weibull distribution model is established by the genetic algorithm based on moment estimation.The results shows that failure values meet the three parameter Weibull distribution,in which the genetic algorithm based on moment estimation can accurately estimate the model parameters.4.To solve the uncertainty relationship between performance degradation sensitive feature and remaining useful life of flexible bearings,a Gaussian process regression model is established by combining genetic algorithm.The result shows that the Gaussian process regression model can accurately give the point estimation value and confidence interval of the remaining useful life of flexible bearings.5.Based on the flexible bearing life test rig,a flexible bearing remaining useful life prediction system is designed to realize the online collection of flexible bearing vibration data and the prediction of flexible bearing remaining useful life.
Keywords/Search Tags:Flexible bearing, Remaining useful life prediction, Sample entropy, Genetic algorithm, Gaussian process regression
PDF Full Text Request
Related items