Electric power system security: the case for an integrated cyber-physical risk management framework

TL;DR

The paper addresses the security of the electric power transmission grid against cyber-physical attackers by modeling a tri-level decision problem that co-optimizes preventive physical reserve procurement and cyber-firewall updates under uncertainty about attacker effectiveness. It treats cyber-security as non-absolute, introducing attacker classes with varying capabilities and probabilities, and uses a DC power-flow framework solved via a modified Column & Constraint Generation (CC&G) algorithm to obtain near-optimal planning and operational strategies. The main finding is that physical- and cyber-security measures are non-exchangeable complements; their synergy allows the grid to withstand sophisticated attacks and adapt to varying threat landscapes, as demonstrated on the IEEE 24-bus and 118-bus benchmarks. The work highlights practical implications for integrated risk management, acknowledges solvability and computational cost limitations, and suggests future directions such as parallelization and reinforcement learning to scale to more detailed models.

Abstract

This paper concerns the security of the electric power transmission grid facing the threat of malicious cyber-physical attackers. We posit that there is no such thing as perfectly effective cyber-security. Rather, any cyber-security measure comes with the possibility that a highly skilled attacker could (eventually find a way to) bypass it. On these grounds, we formulate a tri-level decision making problem seeking to co-optimize preventive physical and cyber-security measures under uncertainty on the ability of an exogenous cyber-physical attacker to overcome the latter. Preventive physical security measures refer to the \emph{ex-ante} procurement of reserve capacity, which translates into ramping restrictions in real-time. Cyber-security measures refer to updating the firewall rules so as to impede an intruder from taking over the cyber infrastructure of the grid and disconnecting power generators and transmission branches. We adopt standard assumptions to formalize the inner optimization problems corresponding to the cyber-physical attacker and power grid operator and focus on uncertainty management at the uppermost level of the problem. Our findings establish that physical- and cyber-security measures are non-exchangeable complements in keeping the power grid operation secure.

Figures (4)

• Figure 1: Cyber-physical power grid security actors
• Figure 2: Column & constraint generation flowchart
• Figure 3: Single-line diagram of the IEEE 24-bus benchmark 24-diagram
• Figure 4: Single-line diagram of the IEEE 118-bus benchmark Al-Roomi2015