Decentralized and Distributed Active Fault Diagnosis for Stochastic Systems with Indirect Observations

The active fault diagnosis (AFD) framework for stochastic large scale systems is treated. The systems are assumed to be decomposed into weakly coupled input-decentralized subsystems. Each subsystem is represented by a set of models describing fault-free and faulty behavior of the subsystems. The paper focuses on the problem formulation where the state of each subsystem is observed only indirectly using noisy measurements. The decentralized and distributed architectures of the AFD are considered, the corresponding estimation and optimization problems are set up and approximate solutions to these problems are proposed. The performance of the active fault detector in both architectures is illustrated using a numerical example.