Un Cadre De Multi-state Simulation Pour Modélisation De La Dégradation Dans Les Composants De Systemes Energetiques a Multi-state Simulation Framework for Modeling Degradation in Components of Energy Systems

We illustrate a framework for simulating degradation processes in components of energy systems. The logical framework revolves around a multi-state representation of the processes. The physics governing the processes is captured in the transition rates governing the random walk across the degradation states. The computational framework relies on a representation of the process dynamics by stochastic Petri nets, augmented by Bayesian networks to capture the influence of interdependent uncertain external factors. Acronyms BN Bayesian network CDF Cumulative distribution function CPT Conditional probability table CTMC Continuous time Markov chain DTMC Discrete time Markov chain MC Monte Carlo MSM Multi-state model MSPM Multi-state physics model NHCTMM Non-homogenous continuous time Markov model PDF Probability density function PN Petri-net SPN Stochastic Petri net Notations The vector of component degradation states Time The discrete function representing the stochastic degradation process, which takes values from The total number of component degradation states The vector of component state probability at time t The transition rate from state i to state j at time t, given the realizations of the uncertain external influencing factors The vector of uncertain external influencing factors The total number of transitions among the degradation states I The total number of influencing factors Probability density function of Place i of the Petri nets Transition j of the Petri nets The holding time at state i Cumulative distribution function of departure time from state i, given and The transition probability from state i to state j at time t given ha l-0 07 44 00 2, v er si on 1 22 O ct 2 01 2 Author manuscript, published in "LAMBDA MU 18 MAÎTRISE DES RISQUES ET SÛRETÉ DE FONCTIONNEMENT, France (2012)"