Font Size: a A A

Reliability Evaluation Of CXK5463 Based On Multiple Similar Samples

Posted on:2016-08-03Degree:MasterType:Thesis
Country:ChinaCandidate:Q WangFull Text:PDF
GTID:2191330479950822Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
Reliability evaluation is an important part of CNC machine tools reliability engineering and technology, which can verify whether the reliability of machine tools has achieved the desired goal, test whether the machine tool design is reasonable, find the weakness of the machine tool, and provide scientific guidelines for improving machine tool performance. However, the reliability evaluation of high-end CNC machine tools in small samples has become an important and difficult problem in machine reliability area, since effective evaluation can not be achieved in small sample reliability issues using conventional methods. In this paper, a mixed prior distribution Bayesian reliability evaluation method is proposed to improve traditional Bayesian approach which has the problem that the determined prior distribution is not reasonable using similar machine tools data. And the reliability of a small sample of CXK5463 lathe-mill machining center is studied using our improved method.Firstly, a regenerated large sample was made through a re-sample process of the machining center fault data. And then the compatibility test between the regenerated data and two similar machine tool fault data was performed, which confirmed that the fault data of the two similar machine tools and the CXK5463 machine tool field failure data are compatible. Mathematical model of history prior distribution parameters in the mixed prior distribution was then calculated based on the fault data of similar machine tool. And mathematical model of update prior distribution parameters in the mixed prior distribution was calculated based on Reference prior method.Secondly, by means of the fault data re-sampling processing, probability density distributions of the two models of similar machine tool fault data and machining center fault data were obtained, respectively. And thus the objective method based inheritance factor was calculated basing on the differences between probability density distribution of similar machine tools and that of the machining center. By analyzing the similarity scores of two similar machine tools and machining center in each index that experts rated, weight of each similarity index was calculated using weighting method, and then subjective methods-based inheritance factor was obtained according to the weight and value of each similarity index. Therefore, the objective-subjective synthetic method based inheritance factor was synthetically calculated using the subjective methods-based inheritance factor value and objective methods-based inheritance factor value. The mixed prior distribution was obtained by the combination of history prior distribution, update prior distribution, and inheritance factor according to information fusion principles.Thirdly, posterior distribution function of the time between failures of machining center was obtained by combining the mixed prior distribution and machining center field experiment sample according to the Bayesian formula.Finally, to solve the problem that posterior distribution of the machining center is multiple integrals and hard to be calculated, the posterior distribution of machining center was simulated using the Open BUGS software with Markov chain Monte Carlo method as the theoretical basis, and the estimation value of MTBF of the machining center was then obtained. The correctness of method in this paper was verified by comparing it with the traditional Bayesian method and Bayesian method without information, which also shows the importance of reasonable determination of prior distribution.
Keywords/Search Tags:machining center, reliability assessment, Bayesian method, inheritance factor, Weibull distribution
PDF Full Text Request
Related items