Model-based fault isolability and isolation of persistent faults: Centralized and distributed implementations

Research on system monitoring, fault detection, and fault isolation is extremely important for guaranteeing reliability, safety, integrity, and efficiency of plants, technical processes, and especially safety-critical systems. In this paper we first discuss, analyze, and compare two known centralized observer-based fault isolation schemes. Secondly we devise and discuss their distributed implementations. We finally evaluate and compare the potentialities of the different fault isolation schemes and their distributed implementations considering a benchmark case study.