Adaptive Methods for Flight Control Diagnostics

Model-based diagnostic and prognostic techniques depend upon mathematical models of the plant. In the presence of uncertainties, modeling errors can decrease system sensitivity to faults and reduce the accuracy of failure prognoses. Parameters of many physical systems may change over time and become uncertain as the system ages, and such nominal parameter variations are often not indicative of impending system failures. When failures or faults happen, the system can experience structural changes and have additional large parameter changes from failure induced signals acting on the system. Such structural and parametric uncertainties can significantly deteriorate control system performance in an essentially different way from the effects of system nominal parameter uncertainties. Clearly, a desirable method to address this problem is to employ an approach that can adapt to those system uncertainties. Adaptive techniques have been proven to be able to effectively handle system uncertainties in many applications. In this paper, a new adaptive model-based diagnostic design is developed for flight control health monitoring systems. The designs leverage framework from adaptive observer techniques and are applied for diagnosis of uncertain actuator failures in systems having unmeasured states and uncertain dynamics. The proposed adaptive designs, which can handle both nominal system parameter uncertainties and failure uncertainties, are computationally efficient and suitable for real-time implementation. The stability and robustness properties of the adaptive failure detection systems are assessed analytically and in simulation.