Packet-data anomaly detection in PMU-based state estimator using convolutional neural network

Sagnik Basumallik, Rui Ma, Sara Eftekharnejad

Research output: Contribution to journalArticle

Abstract

With more phasor measurement units (PMU) being deployed by utilities for reliable monitoring of power systems, there grows an increasing risk of vulnerability to various cyber-attacks. This paper is concerned with a class of false data injection attacks (FDIA) which aim to modify PMU measurements resulting in incorrect state estimation solutions. We extract multi-variate time-series signals from PMU data packets aggregated in phasor data concentrators (PDC) corresponding to different events such as line faults and trips, generation and load fluctuations, shunt disconnections and FDIA prior to every cycle of state estimation (SE). A Convolutional Neural Network (CNN) data filter with Nesterov Adam gradient descent and categorical cross entropy loss is proposed to validate the PMU data. This filter extracts inter time-series relationships to classify different power system events by comparing the temporal structure of PMU packet data. The performance of the filter is then compared with (a) deep learning algorithms such as Recurrent Neural Networks (RNN) and Long Short Term Memory (LSTM) and (b) traditional classifiers such as SVM and ensemble methods. It is seen that the proposed CNN-based filter results in higher classification accuracies among all classifiers. This makes the CNN classifier suitable to serve as an independent data filter to identify falsified data streams targeted to alter SE. In order to verify the accuracy of the proposed filter, a hybrid state estimator (HSE) has been used in this study which obtains measurements from both PMU and traditional meters. All simulations are carried out on IEEE-30 bus and IEEE-118 bus systems.

Original languageEnglish (US)
Pages (from-to)690-702
Number of pages13
JournalInternational Journal of Electrical Power and Energy Systems
Volume107
DOIs
StatePublished - May 1 2019

Fingerprint

Phasor measurement units
Neural networks
State estimation
Classifiers
Time series
Recurrent neural networks
Learning algorithms
Entropy
Monitoring

Keywords

  • Convolutional neural network
  • False data injection
  • Hybrid state estimation
  • Multi-variate time series
  • Nesterov Adam

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering

Cite this

Packet-data anomaly detection in PMU-based state estimator using convolutional neural network. / Basumallik, Sagnik; Ma, Rui; Eftekharnejad, Sara.

In: International Journal of Electrical Power and Energy Systems, Vol. 107, 01.05.2019, p. 690-702.

Research output: Contribution to journalArticle

@article{bb8e9edc325f429fbdf7288c82fae55b,
title = "Packet-data anomaly detection in PMU-based state estimator using convolutional neural network",
abstract = "With more phasor measurement units (PMU) being deployed by utilities for reliable monitoring of power systems, there grows an increasing risk of vulnerability to various cyber-attacks. This paper is concerned with a class of false data injection attacks (FDIA) which aim to modify PMU measurements resulting in incorrect state estimation solutions. We extract multi-variate time-series signals from PMU data packets aggregated in phasor data concentrators (PDC) corresponding to different events such as line faults and trips, generation and load fluctuations, shunt disconnections and FDIA prior to every cycle of state estimation (SE). A Convolutional Neural Network (CNN) data filter with Nesterov Adam gradient descent and categorical cross entropy loss is proposed to validate the PMU data. This filter extracts inter time-series relationships to classify different power system events by comparing the temporal structure of PMU packet data. The performance of the filter is then compared with (a) deep learning algorithms such as Recurrent Neural Networks (RNN) and Long Short Term Memory (LSTM) and (b) traditional classifiers such as SVM and ensemble methods. It is seen that the proposed CNN-based filter results in higher classification accuracies among all classifiers. This makes the CNN classifier suitable to serve as an independent data filter to identify falsified data streams targeted to alter SE. In order to verify the accuracy of the proposed filter, a hybrid state estimator (HSE) has been used in this study which obtains measurements from both PMU and traditional meters. All simulations are carried out on IEEE-30 bus and IEEE-118 bus systems.",
keywords = "Convolutional neural network, False data injection, Hybrid state estimation, Multi-variate time series, Nesterov Adam",
author = "Sagnik Basumallik and Rui Ma and Sara Eftekharnejad",
year = "2019",
month = "5",
day = "1",
doi = "10.1016/j.ijepes.2018.11.013",
language = "English (US)",
volume = "107",
pages = "690--702",
journal = "International Journal of Electrical Power and Energy Systems",
issn = "0142-0615",
publisher = "Elsevier",

}

TY - JOUR

T1 - Packet-data anomaly detection in PMU-based state estimator using convolutional neural network

AU - Basumallik, Sagnik

AU - Ma, Rui

AU - Eftekharnejad, Sara

PY - 2019/5/1

Y1 - 2019/5/1

N2 - With more phasor measurement units (PMU) being deployed by utilities for reliable monitoring of power systems, there grows an increasing risk of vulnerability to various cyber-attacks. This paper is concerned with a class of false data injection attacks (FDIA) which aim to modify PMU measurements resulting in incorrect state estimation solutions. We extract multi-variate time-series signals from PMU data packets aggregated in phasor data concentrators (PDC) corresponding to different events such as line faults and trips, generation and load fluctuations, shunt disconnections and FDIA prior to every cycle of state estimation (SE). A Convolutional Neural Network (CNN) data filter with Nesterov Adam gradient descent and categorical cross entropy loss is proposed to validate the PMU data. This filter extracts inter time-series relationships to classify different power system events by comparing the temporal structure of PMU packet data. The performance of the filter is then compared with (a) deep learning algorithms such as Recurrent Neural Networks (RNN) and Long Short Term Memory (LSTM) and (b) traditional classifiers such as SVM and ensemble methods. It is seen that the proposed CNN-based filter results in higher classification accuracies among all classifiers. This makes the CNN classifier suitable to serve as an independent data filter to identify falsified data streams targeted to alter SE. In order to verify the accuracy of the proposed filter, a hybrid state estimator (HSE) has been used in this study which obtains measurements from both PMU and traditional meters. All simulations are carried out on IEEE-30 bus and IEEE-118 bus systems.

AB - With more phasor measurement units (PMU) being deployed by utilities for reliable monitoring of power systems, there grows an increasing risk of vulnerability to various cyber-attacks. This paper is concerned with a class of false data injection attacks (FDIA) which aim to modify PMU measurements resulting in incorrect state estimation solutions. We extract multi-variate time-series signals from PMU data packets aggregated in phasor data concentrators (PDC) corresponding to different events such as line faults and trips, generation and load fluctuations, shunt disconnections and FDIA prior to every cycle of state estimation (SE). A Convolutional Neural Network (CNN) data filter with Nesterov Adam gradient descent and categorical cross entropy loss is proposed to validate the PMU data. This filter extracts inter time-series relationships to classify different power system events by comparing the temporal structure of PMU packet data. The performance of the filter is then compared with (a) deep learning algorithms such as Recurrent Neural Networks (RNN) and Long Short Term Memory (LSTM) and (b) traditional classifiers such as SVM and ensemble methods. It is seen that the proposed CNN-based filter results in higher classification accuracies among all classifiers. This makes the CNN classifier suitable to serve as an independent data filter to identify falsified data streams targeted to alter SE. In order to verify the accuracy of the proposed filter, a hybrid state estimator (HSE) has been used in this study which obtains measurements from both PMU and traditional meters. All simulations are carried out on IEEE-30 bus and IEEE-118 bus systems.

KW - Convolutional neural network

KW - False data injection

KW - Hybrid state estimation

KW - Multi-variate time series

KW - Nesterov Adam

UR - http://www.scopus.com/inward/record.url?scp=85059134877&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059134877&partnerID=8YFLogxK

U2 - 10.1016/j.ijepes.2018.11.013

DO - 10.1016/j.ijepes.2018.11.013

M3 - Article

VL - 107

SP - 690

EP - 702

JO - International Journal of Electrical Power and Energy Systems

JF - International Journal of Electrical Power and Energy Systems

SN - 0142-0615

ER -