Detection and Localization of Manipulated Smart Meters Using Super State Hidden Markov Models

Abstract

Manipulated energy metering devices lead not only to severe economic damages, they also can directly lead to unfortunate decisions concerning stable grid operation, especially for Smart Grids of the future with less and less rotating inertia and more and more volatile infeed of renewables. For this reason, we propose in the present paper a new manipulation detection method using Super State Hidden Markov Models (SSHMMs) in combination with an additional central metering device. The proposed new method does not only allow the detection of the manipulation of the Smart Meters, but additionally infers which meters are most likely manipulated. Thus, the new method is able to detect attacks on the grid and to pinpoint the manipulated devices, enabling the network provider to initiate further counteractions. The accuracy and effectiveness of the new method are shown by its application to a distribution grid example. It is ready-to-use for network providers of classical distribution grids.