Coding Approaches for Fault Detection and Identification in Discrete Event Systems

Abstract

This chapter applies coding techniques in the context of detecting and identifying faults in complex discrete event systems (DES’s) that can be modeled as Petri nets [Hadjicostis, 1999; Hadjicostis and Verghese, 1999]. The approach is based on replacing the Petri net model of a given DES with a redundant Petri net model in a way that preserves the state, evolution and properties of the original system in some encoded form. This redundant Petri net model enables straightforward fault detection and identification based on simple parity checks that are used to verify the validity of artificially-imposed invariant conditions. Criteria and methods for designing redundant Petri net models that achieve the desired objective while minimizing the cost associated with them (e.g., by minimizing the number of sensors or communication links) are not pursued here, but several examples illustrate how such problems can be approached.