Monitoring and assistance diagnosis of a centrifuge pump by using the fuzzy Petri nets

In an industrial context, aiming to increasing the productivity and to reduce the maintenance costs of the systems of production; it seems to us interesting to develop the techniques of the diagnosis. Within this framework, this work consists in working out of the diagnosis assistance system of the revolving machines. A model based on the fuzzy Petri nets is presented and validated by simulation. Thus the detection of the defects by using the spectral analysis of the system was made starting from the thresholds of pre alarms. Indeed, the detection of a defect, at an early stage, and followed by its evolution are of great interest; Nevertheless, the owner of maintenance wishes more and more to know the nature of this anomaly and to estimate his gravity to make the suitable decision. To answer this requirement, fuzzy logic was introduced; and consequently, this tool was reliable in the treatment of the inaccuracies and the decision-making on the state of degradation of the system.