Sensor fault detection and isolation for electro-mechanical actuators in a reusable launch vehicle TVC system

Abstract

This paper introduces a model-based Fault Detection and Isolation (FDI) approach for a Reusable Launch Vehicle (RLV) Thrust Vector Control (TVC) system operated by Electro-Mechanical Actuators (EMAs). The focus is on the sensors required for the EMA embedded control system to track the on-board computer control commands. The nullspace FDI method is considered and applied to detect and isolate additive faults affecting the mentioned sensors. A detailed formulation of the problem and the EMA-based TVC system modelling for FDI synthesis is provided, including the mechanical load exerted by the rocket nozzle. The FDI synthesis framework is introduced and the application of the nullspace-based strategy is described, including considerations about isolability of the faults. Vehicle-induced loads can potentially disrupt the fault detection process, therefore they are included in the problem formulation to achieve decoupling from the residual generator output and not incur into false alarms. The generator performance is then assessed in fault-free and faulty scenarios using a high-fidelity TVC physical model, and successively benchmarked at the example of an RLV mission scenario.