Study On Forecasting Methods Of Rolling Bearing Friction Torque

The bearing friction torque is an important index of evaluating the operational sensitivity and life of rolling bearing. Presently, there are few researches on the prediction of rolling bearing friction torque. Because the friction torque values show random and unstable periodic change, using the traditional statistical theory to predict and control often can not achieve the desired results. So chaotic theory is introduced in the prediction field of rolling bearing friction torque. In the paper, chaos characteristics and local-region prediction problem of rolling bearing friction torque time series are discussed, and based on the poor information system theory, the problem of chaotic local interval prediction is discussed.Firstly, rolling bearing friction torque chaotic characteristics and identification methods are discussed. A novel method for solving phase space reconstruction parameters - the grey relation method, which synchronously calculate the optimal delay time and embedding dimension is proposed. The small data sets method is adopted to calculate the maximum of Lyapunov exponents. The results of application study on the four types of bearing have shown that the studied time series have chaotic characteristics, and there are low-dimensional strange attractors.Secondly, based on chaos identification of bearing friction torque, the grey chaos evaluation model (GCEM) is proposed. On the basis of poor information system theory and add-weighted one-rank local-region model , first, the optimum embedding dimension and optimum delay time for the phase space reconstruction are synchronously calculated with the grey relation method, and then, at different reconstruction parameters, based on the grey bootstrap fusion method, a novel model called the grey chaos evaluation model is presented . It can effectively predict the value and its fluctuation interval. The test data of HKTB bearing friction torque as an example is analyzed, the calculated results validate the availability.Rolling friction torque studies a variety of chaotic time series forecasting models. Considering the limitations of a single chaotic time series prediction, selecting the five kinds of chaotic forecasting methods, a chaotic dynamical fusion prediction models (CGBM) is established, selected the five kinds of chaotic forecasting methods. The model has high prediction accuracy and reliability. Case analysis shows that the model is effective.Elementary theories and methods studied in this paper are of the versatility and feasibility. Studies show that the two models established in this paper using rolling bearing friction torque analysis are correct and reliable, the prediction accuracy can reach the requirements of the actual engineering; On the different confidence levels, its reliability is also high.