Improvement of recurrent fuzzy detection/diagnosis functions modeling. A temporal modeling application for fault states in a manufacturing system

This article proposes a new approach in modeling the detection/diagnostic functions using PN. It assumes knowledge of the evolution of successive degradation states of the monitored system. Using the logical model, corresponding to the change in system degradation state, hierarchical levels are identified, corresponding to the successive external events with which this model is synchronized. A dedicated hierarchical structure is constructed, capable to repeatedly carry out a fixed defect detection scheme. Each hierarchical level executes this action in temporal windows which are progressively increased. The proposed structure allows the detection function to be rerun on a hierarchical level by preserving information from the previous level. This is the recurrent aspect of the paper. The present proposal is a refined modeling of the detection function because it leads to the identification of temporal windows in which a given degradation state might appear. For modeling the recurrent aspect, between hierarchical levels there is an information subsystem based and LIFO and OOPN (Object Oriented Petri Nets). The present paper enhances similar approaches of the authors, carried out through the recurrent modeling of the monitoring function.