Distributed fault detection of nonlinear large-scale dynamic systems

This paper deals with the problem of designing a distributed fault detection algorithm for nonlinear large-scale systems. In the proposed algorithm, instead of a central detection node, several interconnected local detectors (LD) are employed. Each LD has a limited observation of the system's state and communicates with its neighbors to exchange processed information. The outlet of the detection nodes is the collective probability of failure associated with the system's fault mode. Simulation results illustrate the efficiency of the proposed approach and prove that the stronger communication amongst the LDs will lead to more reliable and faster results.