Industrial fault detection and diagnosis using Bayesian belief network

Gharahbagheri, Hassan (2016) Industrial fault detection and diagnosis using Bayesian belief network. Masters thesis, Memorial University of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (3MB)


Rapid development in industry have contributed to more complex systems that are prone to failure. In applications where the presence of faults may lead to premature failure, fault detection and diagnostics tools are often implemented. The goal of this research is to improve the diagnostic ability of existing FDD methods. Kernel Principal Component Analysis has good fault detection capability, however it can only detect the fault and identify few variables that have contribution on occurrence of fault and thus not precise in diagnosing. Hence, KPCA was used to detect abnormal events and the most contributed variables were taken out for more analysis in diagnosis phase. The diagnosis phase was done in both qualitative and quantitative manner. In qualitative mode, a networked-base causality analysis method was developed to show the causal effect between the most contributing variables in occurrence of the fault. In order to have more quantitative diagnosis, a Bayesian network was constructed to analyze the problem in probabilistic perspective.

Item Type: Thesis (Masters)
Item ID: 12145
Additional Information: Includes bibliographical references (pages 104-108).
Keywords: Fault detection and diagnosis, Kernel principal component analysis, Causality analysis, Bayesian network
Department(s): Engineering and Applied Science, Faculty of
Date: May 2016
Date Type: Submission
Library of Congress Subject Heading: Fault location (Engineering); Multivariate analysis; System failures (Engineering); Automatic control

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics