Design and implementation of the fault diagnostic system using interpreted and colored petri net

Abstract

This work shows the design and implementation of a fault diagnosis system for a hydraulic system that presents symmetry in its design. The diagnostic system uses the controller commands and signals from the sensors to diagnose eleven individual faults or some combinations. The paper presents the methodology for designing the diagnoser, which consists of eight steps that seek the extraction of information from the model base of the hydraulic system. As a base model we use a Petri net interpreted to discretize by events the system. On the other hand, the model of the diagnose is derived in a new type of colored Petri net that allows it to be independent of the controller of the hydraulic system, which implies that it can be implemented in an external computer and perform diagnoses in real time. The implementation of the diagnoser was done with NI LabView language. A characteristic of the diagnose  is that its model allows to be scaled without great effort to large systems that present symmetry in its structure, since each new subsystem that is incorporated to the original system does not alter the design of the diagnostician system. It also does not alter the matrix or vectors used because the new subsystem is incorporated as “colors” or copies of the selected base subsystem.