Fault diagnosis in a class of differential-algebraic systems

An iterative learning observer (ILO) is proposed for fault diagnosis in a class of differential-algebraic nonlinear systems described by so-called semi-explicit form with index 1. The main feature of this ILO-based fault diagnosis strategy is that the ILO can estimate both system states and algebraic variable. This is important since both fault detection and estimation can be achieved. As a result the ILO can still track the post-fault system. Moreover, the ILO input can be used to isolate faults. The simulation study presented shows the effectiveness of this ILO-based fault detection and estimation strategy for differential-algebraic systems.