Contribution onto Prognosability Based on Discrete Events Systems

摘要

Developing a method of prognosis, allows to avoid any accidental stop of the system, by predicting the future failure occurrence. Which makes it possible to preserve performance, availability and safety of both operators and equipment. To strengthen this technique, various research works have developed the prognosability, as a necessary and sufficient condition for prediction of a failure without missed detection and false alarm. This paper is a continuation of the works that has been realized on the prognosis of failure on extended Petri Net (PN) based on discrete events systems (DES); a prognostic method based on time, to predict the future failure occurrence at the earliest and at the latest. The aim is to prove that prognosability is possible on extended PN, while respecting these necessary and sufficient conditions.