In this thesis, the stability and performance of closed-loop systems
following the loss of sensors or feedback signals (sensor faults) are
studied. The objective is to guarantee stability in the face of sensor
faults while optimising performance under nominal (no sensor fault)
condition. One of the main contributions of this work is to deal effectively
with the combinatorial binary nature of the problem when
the number of sensors is large. Several fault-tolerant controller and
observer architectures that are suitable for different applications are
proposed and their effectiveness demonstrated. The problems are formulated
in terms of the existence of feasible solutions to linear matrix
inequalities. The formulations presented in this work are described
in a general form and can be applied to a large class of systems. In
particular, the use of fault-tolerant architectures for damping inter-area
oscillations in power systems using wide-area signals has been
demonstrated. As an extension of the proposed formulations, regional
pole placement to enhance the damping of inter-area modes has been
incorporated. The objective is to achieve specified damping ratios
for the inter-area modes and maximise the closed-loop performance
under nominal condition while guaranteeing stability for all possible
combinations of sensors faults. The performances of the proposed
fault-tolerant architectures are validated through extensive nonlinear
simulations using a simplified equivalent model of the Nordic power
system.