Efficient Nonlinear Actuator Fault Reconstruction System

Abstract

This paper presents an actuator fault reconstruction system for civil aircraft. The fault reconstruction uses the 'local' actuator model and the sensor measurements. Five different actuator benchmark faults are considered: liquid and solid Oscillatory Failure Case (OFC), liquid and solid jamming, and runaway. The OFC reconstruction is regarded as an input fault reconstruction and is achieved by designing an Iterated Optimal Two-Stage Extended Kalman Filter (IOTSEKF) whereas jamming and runaway faults are considered as output faults and are reconstructed by designing a Robust Three-Step Extended Kalman Filter (RTS-EKF). In all the scenarios, both the state and faults are estimated in an unbiased sense. The accurate estimated state is fed back to the control system of the actuator which is designed using the Nonlinear Dynamic Inversion (NDI) approach. The complete actuator fault reconstruction system is presented. Simulation results demonstrate the effectiveness of the proposed system, which shows that the presented system can be applied in practice.