Process fault isolation for linear systems with unknown inputs

A systematic and straightforward fault estimation approach for process fault detection, isolation and accommodation is presented. The approach includes the design of a reduced-order observer and an algebraic fault estimator. The observer is designed for an unknown input and fault-free system, which is obtained by co-ordinate transformations of original systems with unknown inputs and faults. The observer information is devoted to the fault estimation for fault detection and isolation. The fault estimates can be used to form an additional control input to accommodate the fault. The suggested scheme is verified through simulation studies performed on the control of a vertical takeoff and landing aircraft in the vertical plane.