An Intelligent Hierarchical Approach to Actuator Fault Diagnosis and Accommodation

This proposal presents a novel intelligent hierarchical approach to detection, isolation, and accommodation of primary aerodynamic actuator failures. The proposed architecture has three main components. First, a nonlinear fault diagnosis scheme is used to detect any fault occurrence and to determine the particular fault type. The proposed method can directly deal with nonlinear systems and nonlinear faults, unstructured modeling uncertainty, and new and unanticipated faults. Second, a controller module consists of a primary nominal controller and a secondary adaptive fault tolerant controller. While the nominal controller can be any existing conventional flight control system, the secondary neural network (NN) based adaptive controller is designed to accommodate primary control surface failures by utilizing control redundancy. A pseudo-control hedging method is used to prevent the NN from adapting to various actuation anomalies. Third, a reconfiguration supervisor makes decision regarding controller reconfiguration and control reallocation by using on-line diagnostic information. The proposed architecture is attractive in particular as a retrofit to previously certified flight control systems for improved flight safety. Our primary Phase 1 research objective is feasibility demonstration through extensive simulation studies. In Phase 2, we will refine the algorithms and develop the real-time control software.