Process fault diagnosis via the integrated use of graphical lasso and Markov random fields learning & inference

• Published in: Computers & chemical engineering Vol. 125; pp. 460 - 475
• Main Authors: Kim, Changsoo, Lee, Hodong, Lee, Won Bo
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 09.06.2019
• Subjects: Fault diagnosis; Graphical lasso; Kernel belief propagation; Markov random fields; Process monitoring; Tennessee Eastman process; Process monitoring; Fault diagnosis; Kernel belief propagation; Graphical lasso; Tennessee Eastman process; Markov random fields
• ISSN: 0098-1354; 1873-4375
• DOI: 10.1016/j.compchemeng.2019.03.018

•A novel process fault diagnosis method is proposed.•Variables are modelled as Markov random fields for efficient diagnosis.•The graphical lasso algorithm is used for structure learning of MRFs.•Kernel belief propagation is used for learning and inferencing MRFs.•Newly defined conditional contribution is used for fault diagnosis. In this study, a novel methodology for process fault diagnosis is proposed, where the monitored variables are modelled into pairwise Markov random fields (MRFs), and the conditional contribution values are calculated for each node, with respect to the occurring fault. First the monitored variables are modelled into a MRF framework, and the parameters of the MRF are learned using normal process data. Then when a fault occurs, the conditional marginal probability of each of the variables are obtained by using the kernel belief propagation (KBP) method, which is converted into the conditional contribution value for fault diagnosis. Compared to state-of-the art fault diagnosis methods, the proposed methodology successfully detected the root cause nodes for all of the fault types, as well as allowing detailed analysis of the characteristic of the fault. Also, the propagation paths of faults were detectable according to the conditional contribution plots.