Home // ENERGY 2020, The Tenth International Conference on Smart Grids, Green Communications and IT Energy-aware Technologies // View article

Intrusion Detection in Smart Grid Distribution Domain Using Deep Ruptures Detection

Authors:
Sarah Chahine
Chafic Mokbel

Keywords: Smart Grid; Intrusion Detection; Ruptures Detection; Deep Filters; LSTM

Abstract:
Smart grids brought huge added value to classical power grids in terms of advanced monitoring, metering control, trustworthiness and efficiency. This comes with some challenges, a major one being assuring the security of the grid against cyber-attacks. Obviously, such concerns are serious because of the impact and risks on electrical energy provisioning. To prevent and react to possible attacks, intrusion detection appears as a critical component. Previous literature work shows that an intrusion onto the grid translates into a small glitch that a phasor may help in identifying. In this work, we suggest to detect the glitches directly from electrical signals (current, voltage, frequency and power). We suggest using the detection of changes in the signals properties as an indicator of intrusion. To this end, classical approaches in ruptures detection have been experimented. A new approach based on deep Long Short-Term Memory (LSTM) filtering is proposed. The main focus of our work is on intrusions occurring in the distribution domain. In order to conduct experiments for the validation of the techniques, simulated data have been produced. The built simulator is also described in the paper. Benchmark results permit to confirm that our newly proposed deep nonlinear LSTM-based method is a viable solution to consider for intrusion detection for the distribution domain in a smart grid.

Pages: 37 to 43

Copyright: Copyright (c) IARIA, 2020

Publication date: September 27, 2020

Published in: conference

ISSN: 2308-412X

ISBN: 978-1-61208-788-7

Location: Lisbon, Portugal

Dates: from September 27, 2020 to October 1, 2020