Application of Bayesian networks in reliability assessment: A systematic literature review

Bayesian networks are robust and powerful probabilistic knowledge representation and inference models that are widely used in engineering structures for reliability assessment. This study presents a literature review of Bayesian networks methods proposed for reliability assessment within the past tw...

Full description

Saved in:
Bibliographic Details
Published in:Structures (Oxford) Vol. 71; p. 108098
Main Authors: Wang, Qi-Ang, Chen, Jin, Ni, YiQing, Xiao, YuFeng, Liu, NingBo, Liu, ShuKui, Feng, WangSheng
Format: Journal Article
Language:English
Published: Elsevier Ltd 01.01.2025
Subjects:
Bayesian networks
Engineering structures
Machine learning
Probability assessment
Reliability evaluation
Bayesian networks
Reliability evaluation
Engineering structures
Probability assessment
Machine learning
ISSN:2352-0124, 2352-0124
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Bayesian networks are robust and powerful probabilistic knowledge representation and inference models that are widely used in engineering structures for reliability assessment. This study presents a literature review of Bayesian networks methods proposed for reliability assessment within the past two decades, with a focus on combining traditional methods with modern techniques. The development process of Bayesian networks was summarized, and their applications in several typical engineering structural scenarios were discussed. This research categorizes reliability assessment subjects into bridge structures, underground structures, building structures, and offshore structures. Detailed reviews of the research progress and outcomes for each category are provided. Furthermore, the current problems of Bayesian networks in engineering reliability assessment, e.g., algorithm efficiency and applicability, are discussed and several research directions worthy of researchers’ attention are provided, e.g., broadening the applicability of BNs, improving Bayesian networks algorithm, and establishment of objective reliability assessment method. Research indicates that Bayesian networks are suited for reliability assessment across various structural types, compared to traditional reliability assessment methods. Bayesian networks combine probabilistic models with detection and monitoring data, enabling dynamic updates of the structure's status and risk assessment, and providing timely reflection of the actual health condition of the engineering structure, thus offering more effective analysis for structural reliability evaluation. This study can provide some directional references for researchers in the field of reliability assessment.
ISSN:2352-0124
2352-0124
DOI:10.1016/j.istruc.2024.108098