Machine Learning Methods for Attack Detection in the Smart Grid
This work addresses security vulnerabilities in smart grids, which is critical for infrastructure protection, but it is incremental as it applies existing machine learning techniques to this domain.
The paper tackles attack detection in smart grids by framing it as a statistical learning problem, using machine learning algorithms to classify measurements as secure or attacked, and shows that these methods outperform state vector estimation-based algorithms in experiments on IEEE test systems.
Attack detection problems in the smart grid are posed as statistical learning problems for different attack scenarios in which the measurements are observed in batch or online settings. In this approach, machine learning algorithms are used to classify measurements as being either secure or attacked. An attack detection framework is provided to exploit any available prior knowledge about the system and surmount constraints arising from the sparse structure of the problem in the proposed approach. Well-known batch and online learning algorithms (supervised and semi-supervised) are employed with decision and feature level fusion to model the attack detection problem. The relationships between statistical and geometric properties of attack vectors employed in the attack scenarios and learning algorithms are analyzed to detect unobservable attacks using statistical learning methods. The proposed algorithms are examined on various IEEE test systems. Experimental analyses show that machine learning algorithms can detect attacks with performances higher than the attack detection algorithms which employ state vector estimation methods in the proposed attack detection framework.