Training Strategies for Autoencoder-based Detection of False Data Injection Attacks

Chenguang Wang; Kaikai Pan; Simon Tindemans; Peter Palensky

doi:10.1109/ISGT-Europe47291.2020.9248894

Training Strategies for Autoencoder-based Detection of False Data Injection Attacks

Chenguang Wang, Kaikai Pan, Simon Tindemans, Peter Palensky

Intelligent Electrical Power Grids

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

2 Citations (Scopus)

40 Downloads (Pure)

Abstract

The security of energy supply in a power grid critically depends on the ability to accurately estimate the state of the system. However, manipulated power flow measurements can potentially hide overloads and bypass the bad data detection scheme to interfere the validity of estimated states. In this paper, we use an autoencoder neural network to detect anomalous system states and investigate the impact of hyperparameters on the detection performance for false data injection attacks that target power flows. Experimental results on the IEEE 118 bus system indicate that the proposed mechanism has the ability to achieve satisfactory learning efficiency and detection accuracy.

Original language	English
Title of host publication	2020 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe)
Subtitle of host publication	Proceedings
Publisher	IEEE
Pages	1-5
Number of pages	5
ISBN (Electronic)	978-1-7281-7100-5
ISBN (Print)	978-1-7281-7101-2
DOIs	https://doi.org/10.1109/ISGT-Europe47291.2020.9248894
Publication status	Published - 2020
Event	10th IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2020 - Virtual/online event due to COVID-19, Delft, Netherlands Duration: 26 Oct 2020 → 28 Oct 2020 Conference number: 10 https://ieee-isgt-europe.org/

Conference

Conference	10th IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2020
Abbreviated title	ISGT-Europe 2020
Country/Territory	Netherlands
City	Delft
Period	26/10/20 → 28/10/20
Other	Virtual/online event due to COVID-19
Internet address	https://ieee-isgt-europe.org/

Bibliographical note

"Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public."

Keywords

Anomaly detection
autoencoder
false data injection attack
hyperparameter tuning

Access to Document

10.1109/ISGT-Europe47291.2020.9248894

09248894Final published version, 557 KB

Cite this

@inproceedings{aacbe5dde7794271ae674b90af323692,

title = "Training Strategies for Autoencoder-based Detection of False Data Injection Attacks",

abstract = "The security of energy supply in a power grid critically depends on the ability to accurately estimate the state of the system. However, manipulated power flow measurements can potentially hide overloads and bypass the bad data detection scheme to interfere the validity of estimated states. In this paper, we use an autoencoder neural network to detect anomalous system states and investigate the impact of hyperparameters on the detection performance for false data injection attacks that target power flows. Experimental results on the IEEE 118 bus system indicate that the proposed mechanism has the ability to achieve satisfactory learning efficiency and detection accuracy.",

keywords = "Anomaly detection, autoencoder, false data injection attack, hyperparameter tuning",

author = "Chenguang Wang and Kaikai Pan and Simon Tindemans and Peter Palensky",

note = "{"}Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.{"} ; 10th IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2020, ISGT-Europe 2020 ; Conference date: 26-10-2020 Through 28-10-2020",

year = "2020",

doi = "10.1109/ISGT-Europe47291.2020.9248894",

language = "English",

isbn = "978-1-7281-7101-2",

pages = "1--5",

booktitle = "2020 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe)",

publisher = "IEEE",

address = "United States",

url = "https://ieee-isgt-europe.org/",

}

Wang, C, Pan, K, Tindemans, S & Palensky, P 2020, Training Strategies for Autoencoder-based Detection of False Data Injection Attacks. in 2020 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe): Proceedings., 9248894, IEEE, pp. 1-5, 10th IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2020, Delft, Netherlands, 26/10/20. https://doi.org/10.1109/ISGT-Europe47291.2020.9248894

Training Strategies for Autoencoder-based Detection of False Data Injection Attacks. / Wang, Chenguang; Pan, Kaikai; Tindemans, Simon et al.
2020 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe): Proceedings. IEEE, 2020. p. 1-5 9248894.

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Training Strategies for Autoencoder-based Detection of False Data Injection Attacks

AU - Wang, Chenguang

AU - Pan, Kaikai

AU - Tindemans, Simon

AU - Palensky, Peter

N1 - Conference code: 10

PY - 2020

Y1 - 2020

N2 - The security of energy supply in a power grid critically depends on the ability to accurately estimate the state of the system. However, manipulated power flow measurements can potentially hide overloads and bypass the bad data detection scheme to interfere the validity of estimated states. In this paper, we use an autoencoder neural network to detect anomalous system states and investigate the impact of hyperparameters on the detection performance for false data injection attacks that target power flows. Experimental results on the IEEE 118 bus system indicate that the proposed mechanism has the ability to achieve satisfactory learning efficiency and detection accuracy.

AB - The security of energy supply in a power grid critically depends on the ability to accurately estimate the state of the system. However, manipulated power flow measurements can potentially hide overloads and bypass the bad data detection scheme to interfere the validity of estimated states. In this paper, we use an autoencoder neural network to detect anomalous system states and investigate the impact of hyperparameters on the detection performance for false data injection attacks that target power flows. Experimental results on the IEEE 118 bus system indicate that the proposed mechanism has the ability to achieve satisfactory learning efficiency and detection accuracy.

KW - Anomaly detection

KW - autoencoder

KW - false data injection attack

KW - hyperparameter tuning

UR - http://www.scopus.com/inward/record.url?scp=85094331654&partnerID=8YFLogxK

U2 - 10.1109/ISGT-Europe47291.2020.9248894

DO - 10.1109/ISGT-Europe47291.2020.9248894

M3 - Conference contribution

SN - 978-1-7281-7101-2

SP - 1

EP - 5

BT - 2020 IEEE PES Innovative Smart Grid Technologies Europe (ISGT-Europe)

PB - IEEE

T2 - 10th IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2020

Y2 - 26 October 2020 through 28 October 2020

ER -

Training Strategies for Autoencoder-based Detection of False Data Injection Attacks

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this