Multiple fault diagnosis of discrete event systems using Petri Nets

Discrete event systems is a class of dynamic systems, which evolutions are characterized by asynchronous occurrences of discrete events. The complexity of these systems is important. This complexity can make fault appearance very often. For this reason, the diagnosis task is needed as it is one of the key technologies guaranteeing the system safety and reliability. In this context, this paper deals with discrete event systems diagnosis. We propose a general approach based on the use of Petri Nets in order to on-line diagnose multiple event and state faults. Petri Nets are particularly useful as they are intuitive graphical modelling language and advanced formal analysis method. Nowadays, they are considered as one of the main formalisms for modelling, analysis and control of discrete event systems (DES). Petri Nets are used, in our work, for modelling and diagnosis purposes. The proposed approach consists in defining a diagnosis Coverability tree and using invariants to check the system place and state faults. To test the effectiveness of the proposed approach a diagnosis benchmark is used: the three-tank hydraulic system. A comparison between the results given by using the proposed Petri Nets approach and those given by automata tool is finally made.