Open set compound fault recognition method for nuclear power plant based on label mask weighted prototype learning

Shiqi Zhou, Meng Lin, Shilong Huang and Kai Xiao

Applied Energy, 2024, vol. 369, issue C, No S0306261924009863

Abstract: Most of the existing data-driven methods for diagnosing faults in nuclear power plants (NPP) concentrate on addressing single fault problems under the closed set hypothesis. However, the practical application of these methods is severely limited by unknown faults and compound faults caused by the complex operation characteristics of NPP. In order to extend the NPP fault diagnosis model to the Open Set Compound Fault Recognition (OSCFR), a method for OSCFR based on residual capsule network and label mask weighted prototype learning (Mask-CCPN) is proposed. In this method, the residual capsule network is constructed to achieve the decoupled classification of compound faults. Additionally, the label mask weighted method is used to enhance the existing prototype learning, which using the prototype information from single faults to identify the compound faults. Numerical experiments on complex NPP simulation data demonstrate that the proposed method can effectively solve the OSCFR of NPP without requiring additional storage space. The test results under a single fault data set are superior to the state-of-the-art method. In addition, this study confirms that the uniform distribution hypothesis is more suitable for the decision-making process of prototype learning, while the Gaussian distribution hypothesis and Weibull distribution hypothesis have poor results. Overall, the proposed method is promising to expand the application range of NPP fault diagnosis to various types of faults.

Keywords: Open set recognition; Nuclear power plant; Compound fault diagnosis; Label mask weighted (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261924009863
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:369:y:2024:i:c:s0306261924009863

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2024.123603

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().