Observer-based fault detection for large-scale systems with event-triggered protocols.

In this note, the problem of continuous linear large-scale systems fault detection based on event-triggered interval observers is considered. First, an interval observer-based fault detection scheme is proposed. The interval observer design process is mainly implemented based on positive system theory and Lyapunov stability theory, and the residual interval is constructed to determine whether the system behavior is normal. To improve the fault sensitivity and maximize the impact of the fault signal on the residual signal, the corresponding performance indicators are introduced in the interval observer design process. Second, an event-triggered protocol is added to the interval observer design to reduce the occupation of communication resources. The solution conditions of the observer matrix are derived by constructing an augmented system. In addition, it is demonstrated that no Zeno behavior occurs under the adopted event-triggered protocol. Finally, the correctness of conclusions is verified by a practical example of a large-scale system consisting of several F-18 aircraft and a numerical example.