Test Design of Small Sample Launch Vehicle Based on Composite Equivalency Bayesian Fusion

Abstract

In view of the high development cost of the Long March series of launch vehicle and the difficulty of implementing reliability tests, a small sample reliability sampling method based on composite equivalency Bayesian fusion is proposed. First, in order to avoid large amounts of prior data submerging the field data with small sample, a comprehensive use of physical equivalency credibility and data compatibility test are used to fully integrate multi-source test data. Then, according to Bayesian theory, the reliability of the fusion data of the launch vehicle that obeys the normal distribution is statistically verified under the complex assumptions. Finally, considering the experimental cost and the constraints of the two types of risks, a nonlinear constraint programming model for solving the minimum sample size is established.