Distributed fault diagnosis and tolerant control for a large-scale power generator network

This paper addresses a distributed fault diagnosis and fault-tolerant control problem for a multi-agent system modeling a large-scale power generator network. The goal is to enable all the agents to achieve the control objective of asymptotic stability without losing the system tracking performance. Before solving this DFTC problem, a distributed fault detection (DFD) method is provided for fault detection. Next, the designs focus on a DFTC scheme without estimating the upper bound of the coupled, nonlinear, state-dependent unknown input. A model-based distributed state estimator (DSE) together with a proportional-integral-like nonlinear distributed identifier (DI) is then developed to identify the unknown input. By exploiting the redundancies from the estimated states and unknown input information obtained from the DSE and DI, a novel continuous DFTC is designed to enable the agents to achieve asymptotic consensus tracking without losing the system tracking performance while achieving distributed unknown input identification. A power system example and numerical simulations are provided to illustrate the effectiveness of the proposed DFTC method.