Rolling bearing performance variation research train of thought

At present, _ some scholars studied the method to solve the problem of existing, 34 research methods can be divided into the statistical method and statistical method two types, mainly include (information office 7 grey method *) Method, fuzzy network classical statistical methods, bayesian method 141, self-help method 4, etc. Generally refers to the traditional statistical method on the basis of classical statistics and probability theory of research methods, the important theoretical basis for the theorem of large Numbers and central limit theorem, for the experimental data, the number of random variables and the overall effect is small, have mutual independence; Information entropy method requires system known probability distribution or frequency, also asked that the limited number of data, the probability distribution is unknown, the data, the extrapolation error is very big; The grey system theory prediction confidence could not be determined in advance, thus predicted results is uncertain, at the same time have too strict requirements on the original data sequence, the gray forecast area, in the dynamic forecast will be more difficult to solve problems; The main problem with fuzzy set theory is the probability distribution of the unknown system establishment and the selection of membership degree of membership function is difficult; Bayesian approach requires or frequency values known probability distribution function, without these information is unable to obtain the accurate experience value, the inference result error can be large; Self-help method is entirely dependent on the initial samples for distributed general representative, simulation of the sampling results not more than the initial deserves a more credible information, especially under the condition of small sample of the lack of information, the reliability of the inference is very low. The study shows that a single method is difficult to effectively solve the problem of lack of information, each method has its limitations, as well as merits if effective fusion methods, give full play to the strengths of the single method, abandon the shortcomings of the single method, reasonable use of fusion method, can get different aspects and Angle information. Amoi tao, etc. 147. 30 in this respect did a lot of work on the analysis of the bearing performance, as well as rolling bearing performance evaluation provides a good approach to resolve the problem. Variation is a common phenomenon in nature, because the living environment have a huge change, to adapt to the environment variation, lead to changes in appearance of species. Materialism believed that genetic variation is the internal cause of species appearance change, appearance change is the concentrated reflection of internal cause. Therefore, the problem in turn, if the appearance change suggested that internal cause of change, a species by evaluating its internal cause variation characteristics, appearance change can predict the evolution of species. Based on this, can use a variety of methods, from the perspective of multiple profiles, rolling bearing genetic factor and variation factor characteristic information, in order to solve the problem of performance evaluation of rolling bearing. To study the method for the data analysis, find out the current advantages and disadvantages of various methods, and the fusion, from the perspectives of different sides, dig out more information, in order to get more reliable performance information bearing, for bearing performance variation of judgment and prediction to provide powerful evidence. Based on this, the book puts forward rolling bearing robust performance evaluation method research. In the process of data processing, data robustness is the precondition of data analysis, however, in the process of bearing operation, its performance will gradually variations, may be in a good direction transition may also be toward the direction of performance degradation. The reason is that the performance data is not stable, variation data. Variable data, the more the more serious the bearing performance of alienation; In robust statistics, robust processing of data can get reliable robust data, robust, in turn, consider the data processing, data can be isolated unsound, namely can separate the variation factor, which can identify genetic factors and variation factors of bearing performance. Rolling bearing research mainly relates to the performance of static and dynamic performance, performance parameter estimation and prediction, performance degradation and other issues. Research methods mainly depends on the modern statistics theory and chaos theory, the goal is to uncover some unknown characteristics of rolling bearing performance variation. The train of thought to solve the problem for using the method of modern statistics theory, to steady, rolling bearing performance data to identify data and unsound; Then can reflect the characteristics of the data of stable data for static and dynamic analysis; The analysis of data, use the parameters with the fusion method analysis rolling bearing static performance, dynamic performance with chaos theory analysis of the rolling bearing, using bayesian methods to analyze reliability, performance parameters of the rolling bearings with solid fusion method analysis of rolling bearing performance degradation problems, in order to realize the rolling bearing performance variation process fusion of modern statistical analysis and evaluation.