Reliability Estimation of Mechanical Aerospace Systems with Few or No Failure–A Case Study of Thrust Regulator of Liquid Rocket Engine

Reliability estimate of the individual system is an essential input to carry out Probabilistic Risk Assessment (PRA) of a launch vehicle. This paper presents three methods to estimate risk and reliability of mechanical aerospace systems when failure data is scarce. The thrust regulator of a liquid rocket engine is used as a case study, although these methods are applicable to any similar system. The engines under consideration are pump-fed, gas-generator-cycle rocket engines that use hypergolic propellants as fuel and oxidizer. The thrust regulator is a single-point failure for the engine as it lacks redundancy. The first method involves fault tree (FT) analysis, where fault trees are constructed for each possible failure mode of the regulator, ultimately leading to the top event labeled “Regulator Unable to Regulate.” The second method utilizes flight data to estimate failure probability using a special case of the Weibull distribution known as the Nelson method. The third method applies Bayesian analysis, where the prior information is derived as the mean estimate from the first method, and the likelihood function is modeled using an exponential distribution based on the available flight data. These methods provide a systematic approach to estimate risk and reliability, offering insights for enhancing the safety and performance of mechanical aerospace systems with few or no failures.