The Poor Information Forecasting On Performance Of Rolling Bearing

As important performance indexes, the friction torque and vibration have aninfluence on the operation flexibility and life of rolling bearings. At present, the studyabout performance prediction of rolling bearing is still shallow, influence factors ofbearing performance are multiple and complex, the friction torque shows an obvioustrend of random uncertainty, changeable, unknown probability density, so thetraditional prediction methods are often difficult to realize it. In this paper, using thefriction torque and vibration of rolling bearing as the research object, both the greysystem theory and chaos theory are applied to predict the performance of rollingbearing, focus on the grey prediction and chaotic local prediction about friction torqueand vibration time series of rolling bearing, based on the poor information theory,fusion the predictive value, discuss the interval prediction problem.Study on GM(1,1) model of grey system theory, in the condition of modelingsection s=5, establish the model for predicting time series of the friction torque andvibration, and residual test and posteriori error test are used for the prediction results,the test results show that using the grey model for predicting the rolling bearingperformance is feasible, the model’s prediction accuracy grade is qualified, predictiondata is reliable.Study on the chaos characteristics of the rolling bearing performance and chaosidentification method, it mainly includes the embedding dimension, delay time, thelyapunov index etc. the characteristic quantity and qualitative, quantitative chaosidentification method. And use a small data amount calculated the largest lyapunovexponent of five sets of experimented bearings under different rotation speed, thecalculation results show that bearing friction torque time series have chaoscharacteristics, it can be predicted with chaos theory.On the basis of rolling bearing performance chaos identification, using zero-rank local-region method, first-rank local-region method, weighted one-rank local-regionmethod and modified the weighted one-rank local-region method predict theexperimental data of friction torque and vibration, and fusion the predictive value forfinding prediction interval, finally, to analyse of prediction results. Analysis shows thatthe model which is applied to predicted rolling bearing performance is feasible, andhas high prediction accuracy, in addition to individual points beyond the predictioninterval, the rest are in the prediction interval range, the reliability of fusion value ishigh.The theory and method in this paper has the versatility and feasibility, the resultsshow that the prediction model used in the paper is reliable for the calculation andanalysis on rolling bearing friction torque and vibration, research results will lay thefoundation for dynamic evaluation of the rolling bearing performance.