Designing a Bayesian network for preventive maintenance from expert opinions in a rapid and reliable way

Abstract

In this study, a Bayesian Network (BN) is considered to represent a nuclear plant mechanical system degradation. It describes a causal representation of the phenomena involved in the degradation process. Inference from such a BN needs to specify a great number of marginal and conditional probabilities. As, in the present context, information is based essentially on expert knowledge, this task becomes very complex and rapidly impossible. We present a solution which consists of considering the BN as a log-linear model on which simplification constraints are assumed. This approach results in a considerable decrease in the number of probabilities to be given by experts. In addition, we give some simple rules to choose the most reliable probabilities. We show that making use of those rules allows to check the consistency of the derived probabilities. Moreover, we propose a feedback procedure to eliminate inconsistent probabilities. Finally, the derived probabilities that we propose to solve the equations involved in a realistic Bayesian network are expected to be reliable. The resulting methodology to design a significant and powerful BN is applied to a reactor coolant sub-component in EDF Nuclear plants in an illustrative purpose.

Optimizing the preventive maintenance scheduling by genetic algorithm based on cost and reliability in National Iranian Drilling Company

“…• To model the system lifetime and to quantify the degradation of functionality or failure probability, • To detect important variables involved in the functionality degradation process and to design maintenance events to eliminate ageing effect of equipment, • To determine the effect of maintenance activities on the system behavior, • To propose diagnosis and help in decision making, • To propose data extraction and sensibility analysis (Celeux et al 2006).…”

A risk based model for quantifying the impact of information quality

“…Expert judgment is a widely used as data input in risk management, e.g. (Cooke and Goossens 2004;Boholm 2010;Celeux et al 2006;Cooke and Goossens 2004;Evans et al 1994;Otway and Winterfeldt 1992). Furthermore, numerical ranges were used if it was difficult for the experts to provide precise numbers, as suggested by (Hubbard 2010).…”

VirMaLab — atelier virtuel de maintenance : un outil d’aide à la décision pour l’optimisation des politiques de maintenance

“…Des exemples d'études de fiabilité sur des systèmes composés de nombreuses variables interdépendantes à partir de MGP statiques ont été présentés dans la littérature [3]. L'article de Weber décrit l'utilisation des MGPM afin de modé-liser des CdM dépendant de variables exogènes pour étudier la fiabilité d'un système dynamique à temps discret en tenant compte de son contexte [2].…”

“…Nous avons opté pour un angle de vue à la fois probabiliste et graphique, et les modèles graphiques probabilistes (MGP) ou réseaux bayésiens (RB) nous sont apparus comme une solution pertinente pour développer des modèles de maintenance de systèmes dynamiques complexes [1][2][3].…”