Solid Rocket Motor Reliability Evaluation Based On BAYES Method

The assessment of Solid Rocket Motor(SRM) system is a complex systems engineering, involving various information. Bayesian method can fuse various prior information that gained from design, manufacture and examination effectively, and assess the performance of SRM system combining a little field information.First, this paper discussed the normal model of SRM. Analysed the different failure modes by investigating the related data, some amelioration methods were proposed too. Meanwhile, this paper analysed the representative failure modes of SRM by use of FTA(Fault Tree Analysis) method.Second, this paper discussed the general theory about prior information, including the acquirement and consistent test of prior information, fusion of prior information of multiple sources, noninformation priors and prior distribution obtain under small sample. Proposed the Bayes Bootstrap Maximum Entropy method to gain the prior distribution, which combined the virtue of Bayes Bootstrap method and Maximum Entropy method, reduced the subjectivity of prior distribution hypothesis. Analyzed the assessment of component reliability by adopting Bayesian approach. A comparative research of different reliability evaluation methods was given.Third, this paper analyzed the data convert methods between different data types, and proposed the Two Points Optimize method on the basis of existing data convert methods. Its high precision compared with normal convert methods, such as Two Moments method, Point-Low Confidence Limit method and Two Points method, shows a priority in reliability evaluation. Using data convert methods, discussed the Bayesian's application in system reliability assessment with maximum entropy method, Chebyshev expasion method and Monte-Carlo method etc.Finally, analysed the Bayesian reliability assessment of the SRM's mechanical system under different stress-strength interference models. A mechanical system reliability assesmenting process was given.