Ensemble regression model-based anomaly detection for cyber-physical intrusion detection in smart grids

Anna Magdalena Kosek; Oliver Gehrke

doi:10.1109/EPEC.2016.7771704

Ensemble regression model-based anomaly detection for cyber-physical intrusion detection in smart grids

Anna Magdalena Kosek
, Oliver Gehrke

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

The shift from centralised large production to distributed energy production has several consequences for current power system operation. The replacement of large power plants by growing numbers of distributed energy resources (DERs) increases the dependency of the power system on small scale, distributed production. Many of these DERs can be accessed and controlled remotely, posing a cybersecurity risk. This paper investigates an intrusion detection system which evaluates the DER operation in order to discover unauthorized control actions. The proposed anomaly detection method is based on an ensemble of non-linear artificial neural network DER models which detect and evaluate anomalies in DER operation. The proposed method is validated against measurement data which yields a precision of 0.947 and an accuracy of 0.976. This improves the precision and accuracy of a classic model-based anomaly detection by 75.7% and 9.2%, respectively.

Original language	English
Title of host publication	Proceedings of 2016 IEEE Electrical Power and Energy Conference
Number of pages	7
Publisher	IEEE
Publication date	2016
Pages	1-7
ISBN (Print)	978-1-5090-1919-9
DOIs	https://doi.org/10.1109/EPEC.2016.7771704
Publication status	Published - 2016
Event	16th annual IEEE Electrical Power and Energy Conference 2016 - Ottawa, Canada Duration: 12 Oct 2016 → 14 Oct 2016 Conference number: 16 https://www.epec2016.ieee.ca/

Conference

Conference	16th annual IEEE Electrical Power and Energy Conference 2016
Number	16
Country/Territory	Canada
City	Ottawa
Period	12/10/2016 → 14/10/2016
Internet address	https://www.epec2016.ieee.ca/

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

Data models
Density estimation robust algorithm
Training
Correlation
Power systems
Analytical models
Intrusion detection
power system
Data-driven modelling
machine learning
cyber-physical security
model-based anomaly detection
ensemble regression

Access to Document

10.1109/EPEC.2016.7771704

OpenUrl availability

Full text

Cite this

@inproceedings{863417b1bc404effac416cb5020c7989,

title = "Ensemble regression model-based anomaly detection for cyber-physical intrusion detection in smart grids",

abstract = "The shift from centralised large production to distributed energy production has several consequences for current power system operation. The replacement of large power plants by growing numbers of distributed energy resources (DERs) increases the dependency of the power system on small scale, distributed production. Many of these DERs can be accessed and controlled remotely, posing a cybersecurity risk. This paper investigates an intrusion detection system which evaluates the DER operation in order to discover unauthorized control actions. The proposed anomaly detection method is based on an ensemble of non-linear artificial neural network DER models which detect and evaluate anomalies in DER operation. The proposed method is validated against measurement data which yields a precision of 0.947 and an accuracy of 0.976. This improves the precision and accuracy of a classic model-based anomaly detection by 75.7\% and 9.2\%, respectively.",

keywords = "Data models, Density estimation robust algorithm, Training, Correlation, Power systems, Analytical models, Intrusion detection, power system, Data-driven modelling, machine learning, cyber-physical security, model-based anomaly detection, ensemble regression",

author = "Kosek, \{Anna Magdalena\} and Oliver Gehrke",

year = "2016",

doi = "10.1109/EPEC.2016.7771704",

language = "English",

isbn = "978-1-5090-1919-9",

pages = "1--7",

booktitle = "Proceedings of 2016 IEEE Electrical Power and Energy Conference",

publisher = "IEEE",

address = "United States",

note = "16th annual IEEE Electrical Power and Energy Conference 2016, EPEC 2016 ; Conference date: 12-10-2016 Through 14-10-2016",

url = "https://www.epec2016.ieee.ca/",

}

TY - GEN

T1 - Ensemble regression model-based anomaly detection for cyber-physical intrusion detection in smart grids

AU - Kosek, Anna Magdalena

AU - Gehrke, Oliver

N1 - Conference code: 16

PY - 2016

Y1 - 2016

N2 - The shift from centralised large production to distributed energy production has several consequences for current power system operation. The replacement of large power plants by growing numbers of distributed energy resources (DERs) increases the dependency of the power system on small scale, distributed production. Many of these DERs can be accessed and controlled remotely, posing a cybersecurity risk. This paper investigates an intrusion detection system which evaluates the DER operation in order to discover unauthorized control actions. The proposed anomaly detection method is based on an ensemble of non-linear artificial neural network DER models which detect and evaluate anomalies in DER operation. The proposed method is validated against measurement data which yields a precision of 0.947 and an accuracy of 0.976. This improves the precision and accuracy of a classic model-based anomaly detection by 75.7% and 9.2%, respectively.

AB - The shift from centralised large production to distributed energy production has several consequences for current power system operation. The replacement of large power plants by growing numbers of distributed energy resources (DERs) increases the dependency of the power system on small scale, distributed production. Many of these DERs can be accessed and controlled remotely, posing a cybersecurity risk. This paper investigates an intrusion detection system which evaluates the DER operation in order to discover unauthorized control actions. The proposed anomaly detection method is based on an ensemble of non-linear artificial neural network DER models which detect and evaluate anomalies in DER operation. The proposed method is validated against measurement data which yields a precision of 0.947 and an accuracy of 0.976. This improves the precision and accuracy of a classic model-based anomaly detection by 75.7% and 9.2%, respectively.

KW - Data models

KW - Density estimation robust algorithm

KW - Training

KW - Correlation

KW - Power systems

KW - Analytical models

KW - Intrusion detection

KW - power system

KW - Data-driven modelling

KW - machine learning

KW - cyber-physical security

KW - model-based anomaly detection

KW - ensemble regression

U2 - 10.1109/EPEC.2016.7771704

DO - 10.1109/EPEC.2016.7771704

M3 - Article in proceedings

SN - 978-1-5090-1919-9

SP - 1

EP - 7

BT - Proceedings of 2016 IEEE Electrical Power and Energy Conference

PB - IEEE

T2 - 16th annual IEEE Electrical Power and Energy Conference 2016

Y2 - 12 October 2016 through 14 October 2016

ER -

Ensemble regression model-based anomaly detection for cyber-physical intrusion detection in smart grids

Abstract

Conference

UN SDGs

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this