Knowledge-Informed Deep Learning Method for Multiple Oscillation Sources Localization

Zhenjie Cui; Weihao Hu; Guozhou Zhang; Qi Huang; Zhe Chen; Frede Blaabjerg

doi:10.1109/TPWRS.2025.3533968

Knowledge-Informed Deep Learning Method for Multiple Oscillation Sources Localization

Zhenjie Cui, Weihao Hu, Guozhou Zhang^*, Qi Huang, Zhe Chen, Frede Blaabjerg

^*Corresponding author for this work

Research output: Contribution to journal › Journal article › Research › peer-review

Abstract

This letter presents a novel knowledge-informed deep learning method for the fine-grained localization of forced oscillation sources (FOSs). This method can effectively identify multiple FOSs under anomalous measurements. First, a knowledge-guided block based on dissipated energy flow (DEF) is proposed. In this block, phasor measurement unit (PMU) signals are disassembled and reconstructed in the time-frequency domain to extract DEF knowledge. Subsequently, a spatial-temporal graph attention (ST-GAT) network is employed. Topology information is embedded into this network to capture the spreading patterns of FOSs.

Original language	English
Journal	IEEE Transactions on Power Systems
ISSN	0885-8950
DOIs	https://doi.org/10.1109/TPWRS.2025.3533968
Publication status	Accepted/In press - 2025

Bibliographical note

Publisher Copyright:
© 1969-2012 IEEE.

Keywords

dissipated energy flow method
forced oscillation
knowledge-informed deep learning
oscillation localization

Access to Document

10.1109/TPWRS.2025.3533968

Cite this

@article{7bd0385a1d1a46a28dbaa55f3e621672,

title = "Knowledge-Informed Deep Learning Method for Multiple Oscillation Sources Localization",

abstract = "This letter presents a novel knowledge-informed deep learning method for the fine-grained localization of forced oscillation sources (FOSs). This method can effectively identify multiple FOSs under anomalous measurements. First, a knowledge-guided block based on dissipated energy flow (DEF) is proposed. In this block, phasor measurement unit (PMU) signals are disassembled and reconstructed in the time-frequency domain to extract DEF knowledge. Subsequently, a spatial-temporal graph attention (ST-GAT) network is employed. Topology information is embedded into this network to capture the spreading patterns of FOSs.",

keywords = "dissipated energy flow method, forced oscillation, knowledge-informed deep learning, oscillation localization",

author = "Zhenjie Cui and Weihao Hu and Guozhou Zhang and Qi Huang and Zhe Chen and Frede Blaabjerg",

note = "Publisher Copyright: {\textcopyright} 1969-2012 IEEE.",

year = "2025",

doi = "10.1109/TPWRS.2025.3533968",

language = "English",

journal = "IEEE Transactions on Power Systems",

issn = "0885-8950",

publisher = "IEEE (Institute of Electrical and Electronics Engineers)",

}

TY - JOUR

T1 - Knowledge-Informed Deep Learning Method for Multiple Oscillation Sources Localization

AU - Cui, Zhenjie

AU - Hu, Weihao

AU - Zhang, Guozhou

AU - Huang, Qi

AU - Chen, Zhe

AU - Blaabjerg, Frede

PY - 2025

Y1 - 2025

N2 - This letter presents a novel knowledge-informed deep learning method for the fine-grained localization of forced oscillation sources (FOSs). This method can effectively identify multiple FOSs under anomalous measurements. First, a knowledge-guided block based on dissipated energy flow (DEF) is proposed. In this block, phasor measurement unit (PMU) signals are disassembled and reconstructed in the time-frequency domain to extract DEF knowledge. Subsequently, a spatial-temporal graph attention (ST-GAT) network is employed. Topology information is embedded into this network to capture the spreading patterns of FOSs.

AB - This letter presents a novel knowledge-informed deep learning method for the fine-grained localization of forced oscillation sources (FOSs). This method can effectively identify multiple FOSs under anomalous measurements. First, a knowledge-guided block based on dissipated energy flow (DEF) is proposed. In this block, phasor measurement unit (PMU) signals are disassembled and reconstructed in the time-frequency domain to extract DEF knowledge. Subsequently, a spatial-temporal graph attention (ST-GAT) network is employed. Topology information is embedded into this network to capture the spreading patterns of FOSs.

KW - dissipated energy flow method

KW - forced oscillation

KW - knowledge-informed deep learning

KW - oscillation localization

UR - http://www.scopus.com/inward/record.url?scp=85216613645&partnerID=8YFLogxK

U2 - 10.1109/TPWRS.2025.3533968

DO - 10.1109/TPWRS.2025.3533968

M3 - Journal article

AN - SCOPUS:85216613645

SN - 0885-8950

JO - IEEE Transactions on Power Systems

JF - IEEE Transactions on Power Systems

ER -

Knowledge-Informed Deep Learning Method for Multiple Oscillation Sources Localization

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this