Abstract: Aiming at the problem of the solid rocket motor tail shell burnt-through fault which caused system inactivation， the Bayesian network model was obtained for getting weakness of influenced system reliability. The fault mechanics was analyzed based on specimen performance and debris data. Finite element method was used for ensuring system structure optimization scheme. Flow field numerical simulation was employed for analyzing the inner flow field situation of stationary barrier in the damage state. The test results of optimized structure show that the temperature increment ratio of the rocket motor shell reduced by 62.17%， the system reliability increased by 0.62%， and the reliability growth aim achieved.

Key words: Solid rocket motor；Bayesian network；Finite element analysis；Structure optimization；Reliability growth

摘要： 针对火箭发动机尾部壳体烧穿故障导致系统级失效的突出问题，运用贝叶斯网络模型推得影响系统可靠性的薄弱环节，结合试件性能与残骸分解数据，分析系统故障机理。采用有限元数值计算手段，由流场计算分析固定支架结构破坏后发动机内流场情况，进而明确系统结构优化方案。经结构优化考核试验，火箭发动机外壁温升降低62.17%，系统可靠性提高0.62%，实现可靠性增长的目的。

关键词: 固体火箭发动机；贝叶斯网络；有限元分析；结构优化；可靠性增长