Hardware and software signal extraction for power grid resilience

Power grid is the heart of recent concerns due to climate, economic and geostrategic risks. If the safety and performance of these systems are essential subjects to study, the resilience of energy systems to cyber-attacks is an issue that cannot be ignored. Indeed, energy systems are now the target of computer attacks carried out by state, mafia or political groups with objectives ranging from ransom to sabotage for war purposes. Intrusion detection within network devices is now recognized as an essential countermeasure to identify an attack and remediate it before its effects are critical. This thesis proposes to study two scientific issues: (i) the correlation between signals and device behavior in order to finely identify what impact attacks have on signals, and (ii) the distributed aspect of the network in order to aggregate data measured on several devices to obtain a global view of the energy system. From a use case around energy networks (simulation of the energy network and emulation of devices), real attacks that have impacted energy systems (especially on the Ukrainian energy network in 2015 and 2016) could serve as test vectors. Based on different extracted data, data fusion methods combined with artificial intelligence will be used to detect these cyber-attacks within an energy system.

Facebook

Twitter