Fault tolerant control system against actuator failures based on re-configuring reference input
Résumé
This paper deals with Active Fault Tolerant Control System when performance degradation occurs on system due to actuator faults. Fault Tolerant Control methods are generally focused their attention to design a reconfigurable controller in order to reach the nominal closed-loop performances as close as possible. In the presence of such faults, the steady-state performance can be also degraded due to the physical actuator limits. However, only few contributions concern the reference (also called command) input adjustments to prevent the actuators from saturation. Inspired by [6], the main contribution is to consider the reconfigured system as an open loop controlled by a classical Model Predictive Control strategy in order to design “one-line” new reference input trajectories. The added value of this method is to reduce the energy spent to achieve desired closed-loop performance and consequently to maintain a reliable system in dynamical way. The effectiveness of the proposed approach is illustrated using a classical benchmark corrupted by abrupt actuator faults: the three-tank system.