A Novel Active Fault-Tolerant Control Strategy Based on Output Feedback Model Predictive Control for Aero-Engines

Abstract The efficacy of the aero-engine control system, especially in the presence of faults, plays a significant role in maintaining the safe and stable operation of aero-engines. Aimed at the fault-tolerant system, this paper proposes a novel active fault-tolerant control (AFTC) strategy for aero-engines using output feedback model predictive control (OF-MPC). The aero-engine linear models under two typical types of faults, namely, sensor fault and gas path fault, are first established. To achieve the analytical redundancy under sensor fault, the other measurable output sensors are utilized to estimate the state variables online. Furthermore, the robust positive invariant (RPI) set of the estimation errors is designed and effectively integrated with the tube-based model predictive control (MPC). In this way, the recursive feasibility and robust stability of the newly proposed OF-MPC are rigorously proven. In particular, the decision variable (i.e., constriction factor λ) is introduced in the constraint such that the feasible region for the first prediction step is expanded, thus resulting in improved dynamic control performance. In allusion to the gas path fault, the health parameters are further estimated online through measurable outputs, meanwhile taking into account the estimation errors in the MPC optimization problem. Finally, two simulation examples of the dual-shaft turbofan engine are provided under sensor fault and gas path fault, respectively. The results have demonstrated the effectiveness of the proposed AFTC strategy and performed superior control performance.