This paper presents a model-based method for fault detection and isolation of actuator damage and failure on an aircraft. The proposed method is based on multiple linearized models that approximate the nonlinear dynamics of the aircraft and allow for uncertainty. In this paper, linearized actuator fault models are introduced and a Luenberger observer is modified with a piecewise constant update law for robust fault estimation. For the purposes of identifying actuator faults, a fault identification algorithm is suggested using the estimated fault information from the observer. Simulation results demonstrate the performance of the proposed method in models of a transport class aircraft.