Detection of Malicious Command Injection Attacks Against Static Var Compensators in Smart Grids

Abstract

In a smart grid, voltage control and reactive power control are crucial to its safe and reliable operation. Static Var compensators (SVCs) are widely used to achieve these controls in a transmission system, either through centralized control by Energy Management Systems or through local closed loop control systems. In smart grids, the command channels to control SVCs are vulnerable to cyber attacks. In this paper, attack scenarios involving injection of malicious commands to SVCs are studied in detail. Attack models are established for both methods of control. Based on these attack models, two detection algorithms are proposed. The principle behind these algorithms involves the notion that even though an adversary can manipulate the commands and measurements related to SVCs, it is nearly impossible to hide the effect on other state variables and measurements in the system. The algorithms are developed mathematically using electrical quantities, making these schemes independent of the underlying Information and Communication Technologies (ICT) used. The rationale behind the choices made during the development of the algorithms are proven formally. Finally, two algorithms are formally proposed, which are easy to implement and computationally less intensive, when compared to iterative and multi-stage algorithms. These algorithms are then tested on various test-cases on the IEEE 118-bus system and found to be effective.