Bayesian Framework for In-Flight SRM Data Management and Decision Support

We report progress in the development of a novel Bayesian framework for an in-flight failure decision and prognostic (FD&P) system for solid rocket boosters (SRBs) based on a combination of low-dimensional performance models and a Bayesian framework for diagnostics and prognostics of the parameters of nonlinear flow of combustion products in the combustion chamber. To simulate faults we introduce high-fidelity models of these faults based on stochastic partial differential equations (SPDE). To infer parameters of the model, the SPDE is reduced to a low dimensional performance model (LDPM). It is shown by example of the nozzle blocking fault that using a novel Bayesian framework, it becomes possible both to infer the variations of SRB parameters stimulated by the fault and to predict values of the pressure and time of the overpressure fault even in the case of highly nonlinear fault dynamics. The extension of the method to the diagnostic and prognostic of the case burning fault is discussed.