Power System Fault Diagnosis Method Based on Deep Reinforcement Learning

Zirui Wang (), Ziqi Zhang, Xu Zhang, Mingxuan Du, Huiting Zhang and Bowen Liu
Additional contact information
Zirui Wang: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
Ziqi Zhang: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
Xu Zhang: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
Mingxuan Du: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China
Huiting Zhang: State Grid Shanxi Electric Power Company Skills Training Center, Shanxi Electric Power Vocational and Technical Institute, Taiyuan 030021, China
Bowen Liu: School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, China

Energies, 2022, vol. 15, issue 20, 1-15

Abstract: Intelligent power grid fault diagnosis is of great significance for speeding up fault processing and improving fault diagnosis efficiency. However, most of the current fault diagnosis methods focus on rule diagnosis, relying on expert experience and logical rules to build a diagnosis model, and lack the ability to automatically extract fault knowledge. For switch refusal events, it is difficult to determine a refusal switch without network topology. In order to realize the non-operating switch identification without network topology, this paper proposes a power grid fault diagnosis method based on deep reinforcement learning for alarm information text. Taking the single alarm information of the non-switch refusal sample as the research object, through the self-learning ability of deep reinforcement learning, it learns the topology connection relationship and action logic relationship between equipment, protection and circuit breakers contained in the alarm information, and realizes the detection of fault events. The correct prediction of the fault removal process after the occurrence, based on this, determines the refusal switch when the switch refuses to operate during the fault removal process. The calculation example results show that the proposed method can effectively diagnose the refusal switch of the switch refusal event, which is feasible and effective.

Keywords: alarm information; deep reinforcement learning; fault diagnosis; deep Q-network (search for similar items in EconPapers)
JEL-codes: Q Q0 Q4 Q40 Q41 Q42 Q43 Q47 Q48 Q49 (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
https://www.mdpi.com/1996-1073/15/20/7639/pdf (application/pdf)
https://www.mdpi.com/1996-1073/15/20/7639/ (text/html)

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:gam:jeners:v:15:y:2022:i:20:p:7639-:d:943792

Access Statistics for this article

Energies is currently edited by Ms. Agatha Cao

More articles in Energies from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().