Health Monitoring and Prognosis of Hybrid Systems
Résumé
Maintenance and repair of complex systems are an increas- ing part of the total cost of final product. Efficient diagno- sis and prognosis techniques have to be adopted to detect, isolate and anticipate faults. Moreover the recent industrial systems are naturally hybrid: their dynamic behavior is both continuous and discrete. This paper presents an architecture of health monitoring and prognosis for hybrid systems. By using model and experience-based approach we propose an implementation of an integrated diagnosis/prognosis process based on Weibull probabilistic model. This article focuses particularly on the prognosis algorithm description. The pro- cess has been implemented and tested on Matlab. Simulation results on a water tank system show how prognosis and diag- nosis interact into the architecture.
Domaines
Autre
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...