Power grid security improves cost with decision-focused learning
Decision-Focused Learning for Cyber-Physical Risk-Aware Power Grid Operations
Cryptography and Security
Summary
Protecting power grids from cyberattacks is tricky because mistakes in detecting attacks can cost a lot operationally. The authors propose a new method that helps machine learning models not just detect attacks better but also reduce the actual costs caused by errors. Their approach links detection mistakes directly to how much money these errors cost in running the grid safely. Tests on two datasets showed this method cut power grid operation costs compared to older approaches that treat detection and operation separately.
What this means in practice
- •For power grid operators: Optimize power grid operations by integrating cyberattack detection with cost-aware decision making to reduce overall risk and operational costs.
- •For industrial control system managers: Enhance security response strategies by linking intrusion detection errors with their real impacts on industrial control processes.
Authors
Ayesha Siddiqua Dina, Ebisa D. Wollega, Afzal Hossain Sohag
Abstract
Power grid cyber-physical security requires intrusion detection methods that account for the operational consequences of detection errors. However, the existing intrusion detection approaches primarily focus on the predictive performance of learning algorithms, while downstream operational costs are often optimized independently, which can lead to inferior operational performance. To help alleviate this problem, we propose a novel optimization-learning cyber-physical-risk-informed framework that communicates iteratively to minimize power grid operational costs. This framework translates machine learning prediction errors into power grid operational cyber-physical-risk costs. Experimental results on two power grid operation datasets, Industrial Control System (ICS) Cyber Attack and Sherlock, show that the proposed approach improves power grid operation costs compared with the traditional approach.