A note on prognosis of system based expert knowledge

Author

Ouladsine, Radouane ; Outbib, R.

Author_Institution

LSIS Lab., Aix-Marseilles Univ., Marseille, France

fYear

2012

fDate

15-16 Nov. 2012

Firstpage

343

Lastpage

348

Abstract

This paper is a contribution to the prognostic of complex systems based on the expert knowledge. The aim is to estimate the Remaining Useful Life (RUL) of a resource (i.e. a component of the system) based on partial knowledge. In this paper the models describing the behavior of the resource are, and in order to be more realistic, stochastic. Hence, a probabilistic method based on Maximum Relative Entropy (MRE) approach is used. Based on partial knowledge, provided by expert, the MRE allows to update the probability distribution of the damage parameters. Finally, using a Markov Chain Monte Carlo (MCMC) simulation, damage trajectory is constructed.

Keywords

Markov processes; Monte Carlo methods; expert systems; large-scale systems; statistical distributions; MCMC simulation; MRE approach; Markov Chain Monte Carlo simulation; RUL; complex systems; damage parameters; damage trajectory; expert knowledge; maximum relative entropy; probabilistic method; probability distribution; prognosis; remaining useful life; Estimation; Mathematical model; Probability distribution; Prognostics and health management; Roads; Suspensions; Trajectory;

fLanguage

English

Publisher

ieee

Conference_Titel

Control Conference (AUCC), 2012 2nd Australian

Conference_Location

Sydney, NSW

Print_ISBN

978-1-922107-63-3

Type

conf

Filename

6613220