A Fault Detection and Identification System for Testing Liquid Rocket Engines

Levi M. Araújo¹; João Luiz F. Azevedo²

¹ Instituto Tecnológico de Aeronáutica; ² Instituto de Aeronáutica e Espaço

doi:10.20906/CPS/COB-2015-2124

Resumo

This work deals with the fault detection and identification (FDI) problem for a liquid rocket engine in a test bench. The approach adopted uses a Kalman filter residual-based hypothesis testing. A simple dynamic model, with 17 continuous states and 11 measurements of the engine, is developed and typical failure modes of the engine and sensors are inserted into the model in order to simulate the development of a fault. The strategy adopted regarding the faults is to identify the presence of a fault and assess whether it is a fault in the engine/test bench, i.e., a system fault, or a fault in a sensor. In the case of a system fault, the test must be stopped at once, but, in the case of a sensor fault, it is more interesting to simply ignore the failed sensor readings and continue with the test. The proposed FDI system also estimates unknown parameters in the engine dynamics, without any effect in the filtering scheme. In order to validate the designed FDI system, several Monte Carlo simulations are made and the results are effective for all the tested fault modes.

Palavras-chave: Fault detection and identification system; Liquid rocket engine; Engine testing; Kalman filtering; Hypothesis testing