Fault tolerant control using virtual actuators and invariant-set based fault detection and identification

In this paper we combine an invariant-set approach to fault detection and identification (FDI), recently proposed by the authors, with the virtual actuator approach to controller reconfiguration of Steffen, Lunze and co-authors. The FDI approach is based on the separation of invariant sets that characterise healthy system operation from invariant sets that characterise faulty operation. The derivation of these sets takes into account the closed-loop system reconfigured by means of the virtual actuator under all considered fault situations. We provide analytic conditions, in terms of closed-loop system parameters and bounds on external signals, which guarantee the aforementioned set separation and, thus, closed-loop stability of the scheme under all considered fault situations.