Abstract

Due to the increasing complexity of engineered products, it is of great importance to develop a tool to assess reliability dependencies among components and systems under the uncertainty of system reliability structure. In this paper, a Bayesian network approach is proposed for evaluating the conditional probability of failure within a complex system, using a multilevel system configuration. Coupling with Bayesian inference, the posterior distributions of these conditional probabilities can be estimated by combining failure information and expert opinions at both system and component levels. Three data scenarios are considered in this study, and they demonstrate that, with the quantification of the stochastic relationship of reliability within a system, the dependency structure in system reliability can be gradually revealed by the data collected at different system levels.

Original languageEnglish (US)
Pages (from-to)104-114
Number of pages11
JournalReliability Engineering and System Safety
Volume152
DOIs
StatePublished - Aug 1 2016

Keywords

  • Bayesian inference
  • Bayesian network
  • Dependency
  • Incomplete information
  • Multi-level data

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering
  • Safety, Risk, Reliability and Quality
  • Applied Mathematics

Fingerprint Dive into the research topics of 'A computational Bayesian approach to dependency assessment in system reliability'. Together they form a unique fingerprint.

  • Cite this