Reliability Evaluation Of Turning And Milling Maching Center Of Gantry-moving Type

Reliability assessment is one of the key component in the reliability engineeringtechnology of CNC machine tools. At present, it has been very mature that the study oflarge sample of NC machine tool reliability assessment, which is based on the theory ofprobability and statistics. However, there is no perfect idea of reliability assessment of thesmall sample of NC machine tools. At the present stage, reliability assessment of the CXK×××turning and milling machining center is difficult to perform due to theinsufficiency of the experimental data and the difficult to reliability evaluation. Makingfull use of the enough data which is related reference model the characteristics of themachine tool reliability, the Bayes method that fusion of multi-source prior information isproposed in this paper.According to the structural and function of the CXK×××milling machiningcenter, the reliability test and the field test specification of machining center is established.The field fault data of machining center is collected.According to the characters of structure, the design, manufacture and materialcomposition of the machining center, choosing CK5250and CK5263as the similarmachine tools, at the same time, we collect failure data of the two models for2years.Using the failure data of the two models of similar machine tools as the priorinformation for machining center, the resample processing to the failure data of twomodels is done. Then, the mathematical model of the prior distribution parameters isobtained respectively. By this way, the credibility of the two models is calculated with thecredibility weighted method. And the weighting factor of two models is determinedaccording to the credibility. After that, the weighting factor is distributed to thecorresponding parameters of mathematical model. Thus we can obtain the fusion priordistribution and realize the effective fusion of machining center prior information.After determining the machining center prior information, we can get the Bayeslikelihood function by the situ test data that we have collected. Then, the machining centerprior information and the likelihood function are fused by Bayes formula. By this way, the machining center MTBF distribution density function is determined. Lastly, the machiningcenter prior situ test data and the prior information are effective fused.To solve the problems that the posterior distribution is difficult to calculate，by usingthe Markov Chain Monte Carlo method and the Gibbs sampling, iterated simulation ofMTBF distribution function is achieved，which is based on the OpenBUGS software. Bythis method, we get the parameter estimation of the MTBF distribution mathematicalmodel. And then, the estimated value of MTBF is determined. Finally, the reliabilityassessment of the machining center is completed.Through the comparative analysis of this method and small sample evaluationmethods of Bootstrap, the correctness of the Bayes method that fusion of multi-sourceprior information is verified.Paper provides reference to the reliability of the similar products.