Fault Detection Of Bearing Base On Statistical Methods

Rolling bearings are the most widely used in machinery and equipment, but also more crucial parts and components, whether it is normal for job or not is directly related to the quality and safety of production. To diagnosis and monitoring the bearing fault are the focus of the study.Firstly, the introduction of recent research on the rolling bearing, have been given in this thesis, as well as commonly used methods of research, analysis of the vibration signal of rolling bearing fault diagnosis principle. Then statistics have been put forward in the idea of mathematical statistic. The statistics at a fixed periods or around period must have analysis in both cases, respectively, several treatment methods.Because of the exist of period's bias, we must make more contents take into account in this thesis. So we focused on three ways under the conditions: general methods, average moving and kernel smoothing . These three methods introduce the distribution function of the statistics, at the same time the results of stochastic simulation confirms the correctness of the result derived above. By hypothesis testing and the distribution function, critical value can be derived , then it can be used to determine whether fault or not.