Reliability Modelling Method of Computerized Numerical Control Machine Tools Based on the Improved Posterior Distribution

Abstract

To overcome the instability and low accuracy of the reliability model established by Bayes estimation method, especially when there are small size sample and incomplete failure data, a new reliability modelling method is proposed based on the improved posterior distribution. Firstly, a deep analysis and discussion of posterior distribution of Bayes estimation method are given, and then the disadvantage of Bayes estimation method is pointed out. Secondly, to improve the reliability model, the likelihood function is divided into three levels, and then the improved posterior distribution function is established. Thirdly, the proposed method is compared with traditional modelling methods. It follows that the proposed method is more stable and accurate.