Fault diagnoser design for untimed continuous Petri nets

Abstract

Continuous Petri nets can be viewed as an approximation of the classical discrete models introduced to cope with the state explosion problem typical of discrete event systems. In this paper we consider free-labeled Petri net systems, and assume that certain transitions, including all those modeling faulty behaviors, are unobservable, i.e., they are labeled with the empty word. Based on the notion and characterization of the set of justifications of a given observation (and the corresponding set of j-vectors), we provide a diagnosis approach using the solution of a certain number of linear programming problems.