High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods

Deep learning technology is identified as a valid tool for transient stability assessment (TSA). Moreover, the superior performance of the TSA model depends on generously labeled samples. However, the power grid is dynamic, and some topologies or operation conditions change substantially. The tradit...

Full description

Saved in:
Bibliographic Details
Main Authors: Baoqin Li, Pengfei Fan, Qixin Chen, Rong Li, Kaijun Lin
Format: Article
Language:English
Published: China electric power research institute 2025-01-01
Series:CSEE Journal of Power and Energy Systems
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10838237/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849390520598528000
author Baoqin Li
Pengfei Fan
Qixin Chen
Rong Li
Kaijun Lin
author_facet Baoqin Li
Pengfei Fan
Qixin Chen
Rong Li
Kaijun Lin
author_sort Baoqin Li
collection DOAJ
description Deep learning technology is identified as a valid tool for transient stability assessment (TSA). Moreover, the superior performance of the TSA model depends on generously labeled samples. However, the power grid is dynamic, and some topologies or operation conditions change substantially. The traditional method generates a significant quantity of samples for each specific topology. Nonetheless, generating these labeled samples and establishing TSA models is very time-consuming. This paper proposes a high-quality sample generation framework based on data-driven methods to build a high-quality offline samples database for TSA model training and updating. Firstly, the representative topologies provided by the system operator are clustered into four different categories by density-based spatial clustering of applications with noise (DBSCAN). Thus the corresponding samples are collected. Then, when a new topology is encountered in the online application, scenario matching is used to match the most similar topology category. After that, instance-based transfer learning is implemented from a database of the best-matched topology category. Finally, a deep convolutional generative adversarial network (DCGAN) is constructed to mitigate the class imbalance problem. That is, unstable scenarios occur far more rarely than stable scenarios. Consequently, a high-quality and balanced TSA model training and updating database is constructed. The comprehensive test results on the Central China Power Grid illustrate that the proposed framework can generate high-quality and balanced TSA samples. Furthermore, the sample generation time is dramatically shortened. In addition, the metrics of accuracy, reliability and adaptability of the TSA model are significantly enhanced.
format Article
id doaj-art-c424e5c0bfba4d69b18ec0627caaac89
institution Kabale University
issn 2096-0042
language English
publishDate 2025-01-01
publisher China electric power research institute
record_format Article
series CSEE Journal of Power and Energy Systems
spelling doaj-art-c424e5c0bfba4d69b18ec0627caaac892025-08-20T03:41:34ZengChina electric power research instituteCSEE Journal of Power and Energy Systems2096-00422025-01-011141681169210.17775/CSEEJPES.2023.0707010838237High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven MethodsBaoqin Li0Pengfei Fan1Qixin Chen2Rong Li3Kaijun Lin4China Electric Power Planning & Engineering Institute,Beijing,China,100120China Electric Power Planning & Engineering Institute,Beijing,China,100120Tsinghua University,Department of Electrical Engineering,Beijing,China,100084China Electric Power Planning & Engineering Institute,Beijing,China,100120China Electric Power Planning & Engineering Institute,Beijing,China,100120Deep learning technology is identified as a valid tool for transient stability assessment (TSA). Moreover, the superior performance of the TSA model depends on generously labeled samples. However, the power grid is dynamic, and some topologies or operation conditions change substantially. The traditional method generates a significant quantity of samples for each specific topology. Nonetheless, generating these labeled samples and establishing TSA models is very time-consuming. This paper proposes a high-quality sample generation framework based on data-driven methods to build a high-quality offline samples database for TSA model training and updating. Firstly, the representative topologies provided by the system operator are clustered into four different categories by density-based spatial clustering of applications with noise (DBSCAN). Thus the corresponding samples are collected. Then, when a new topology is encountered in the online application, scenario matching is used to match the most similar topology category. After that, instance-based transfer learning is implemented from a database of the best-matched topology category. Finally, a deep convolutional generative adversarial network (DCGAN) is constructed to mitigate the class imbalance problem. That is, unstable scenarios occur far more rarely than stable scenarios. Consequently, a high-quality and balanced TSA model training and updating database is constructed. The comprehensive test results on the Central China Power Grid illustrate that the proposed framework can generate high-quality and balanced TSA samples. Furthermore, the sample generation time is dramatically shortened. In addition, the metrics of accuracy, reliability and adaptability of the TSA model are significantly enhanced.https://ieeexplore.ieee.org/document/10838237/Data augmentdeep learningsample generationtransfer learningtransient stability assessment
spellingShingle Baoqin Li
Pengfei Fan
Qixin Chen
Rong Li
Kaijun Lin
High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
CSEE Journal of Power and Energy Systems
Data augment
deep learning
sample generation
transfer learning
transient stability assessment
title High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
title_full High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
title_fullStr High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
title_full_unstemmed High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
title_short High-Quality Sample Generation for Power System Transient Stability Assessment Based on Data-Driven Methods
title_sort high quality sample generation for power system transient stability assessment based on data driven methods
topic Data augment
deep learning
sample generation
transfer learning
transient stability assessment
url https://ieeexplore.ieee.org/document/10838237/
work_keys_str_mv AT baoqinli highqualitysamplegenerationforpowersystemtransientstabilityassessmentbasedondatadrivenmethods
AT pengfeifan highqualitysamplegenerationforpowersystemtransientstabilityassessmentbasedondatadrivenmethods
AT qixinchen highqualitysamplegenerationforpowersystemtransientstabilityassessmentbasedondatadrivenmethods
AT rongli highqualitysamplegenerationforpowersystemtransientstabilityassessmentbasedondatadrivenmethods
AT kaijunlin highqualitysamplegenerationforpowersystemtransientstabilityassessmentbasedondatadrivenmethods