NonLinear Fault Tolerant Flight Control for Generic Actuators Fault Models

This paper presents a Fault Detection, Isolation and Reconfiguration approach applied to the aircraft nonlinear longitudinal control in presence of even simultaneous faults on both elevator and thrust actuators. The overall control system is based on a Fault Detection and Diagnosis module, consisting of adaptive filters based on Radial Basis Function Neural Networks, which provide the isolation of faults and their estimates. These estimates are exploited by a further feedback loop to reconfigure the controller, thus resulting in a Fault Tolerant Flight Control System. It is worth observing that the integration of the NonLinear Geometric Approach with the Radial Basis Function Neural Network structure provides efficient fault estimation capabilities. In particular, the NonLinear Geometric Approach exploits a coordinate change that leads to two one-dimensional subsystems affected by only one fault and decoupled from the other. On the basis of these subsystems, a bank of two Radial Basis Function Neural Networks is designed, thus resulting in fault isolation and adaptive estimation of generic fault functions. To the best of the authors’ knowledge, this work represents the first application of the proposed Fault Detection, Isolation and Reconfiguration methodology to an aircraft control system. The simulation results obtained with a high-fidelity flight simulator demonstrates the efficacy of the proposed method, which maintain the aircraft in a safe flight envelope even in case of actuator faults.