GPS spoofing detection for the power grid network via a multi-receiver hierarchical architecture

In the process of modernizing the North American electric power grid with the creation of the Smart Grid, thousands of devices called phasor measurement units (PMUs) have been deployed across the U.S. continent to continuously monitor the power grid state in real-time. Each PMU measures voltage and current phasors at its local substation, then synchronizes these measurements across the continental network using the Global Positioning System (GPS) as a common timing reference. GPS serves as an excellent timing source due to its global coverage as well as its precise, sub-microsecond level timing accuracy. However, because civilian GPS signals are unencrypted with a publicly available signal structure, any individual with the appropriate equipment can mimic these signals in order to establish a false timing solution at the PMU sites. This type of attack, commonly known as GPS spoofing, presents a major concern to our future power grid infrastructure. Indeed, even minor timing manipulations can cause inaccurate power flow representations and corresponding corrective measures, potentially inducing large-scale power disruptions, instability within the power grid, and/or damage to generators and other power equipment. In this thesis, we present a multi-receiver spoofing detection algorithm for PMU devices, utilizing a hierarchical architecture framework. For the received GPS signal at each PMU station, we create conditioned signal fragments containing the military P(Y) GPS signal, which bears a binary spreading code sequence that is unavailable to civilian users and thus cannot be forged by an attacker. As a result, the military P(Y) signal establishes an encrypted signature in the background of all authentic GPS signals. The presence of the authentic signature can be verified, without knowledge of the precise bit sequence, by correlating amongst conditioned signal fragments obtained from other PMU sites in a sub-network of cross-check receivers, thereby leveraging the secure communication network available within the power grid infrastructure. We further defend against coordinated spoofing attacks conducted against the sub-network of PMU devices by comparing condensed, representative signals generated for each sub-network within the power grid. Using real-world data recorded during a government-sponsored, live-sky spoofing event, we demonstrate that our algorithm successfully evaluates the authenticity of each receiver in a widely dispersed network