Fusing Distributed Extended Multi-Model Adaptive Estimation Method for Fault Diagnosis of Coupling Faults in a Liquid Rocket Engine

In this paper, a fusing distributed extended multi-model adaptive estimation (FDEMMAE) fault detection and isolation (FDI) scheme is presented for FDI of coupling faults in a liquid rocket engine (LRE). Local extended multi-model adaptive estimation (EMMAE) models consisting of extended Kalman filters (EKF) are designed to describe the sensor failures in LRE subsystems, and the fusing weighted sum of squared residuals (FWSSR) of the LRE system is calculated based on filtering information of the local extended multi-model adaptive estimation (EMMAE) models, then a threshold test is employed to distinguish components failures with sensor failures of the LRE. Experimental results based on data from simulation are used to decorate the efficiency of the method. This study may help to improve LRE health management and reduce launch risk by solving the problem of coupling FDI of LRE with low computational consumption.