Impact of cyber dependencies in critical infrastructures: The reliability of grid splitting in power systems

Abstract

The growing presence of cyber interdependencies in critical infrastructures can be exemplified by grid splitting, which is an application in power system stability and control dependent on communication services. It involves the controlled separation of a power system into islands in reaction to an imminent instability. Communication is necessary for gathering system-wide synchronized measurements for state estimation as well as for the dispatch of line switch-opening signals. However, grid splitting may become ineffective due to a degradation in communication, which is accrued by the use of non-dedicated, open communication networks. To address this issue, this article aims to quantify the reliability of grid splitting under degraded communication conditions. A simulation framework is developed that allows the integration of a transient electrical model and a stochastic communication delay model that captures the effects of congestion and traffic uncertainty. The application to the IEEE 39-Bus Test System shows that the reliability of grid splitting is reduced under the influence of degraded communication leading to increased time delays. Furthermore, the results identify a critical transition range in the grid splitting reliability for a very narrow range of external interfering traffic and network data rate. Finally, the interdependency of the electric and the communication network is quantified by the expected efficiency of the communication network subject to degraded communication following the grid splitting action. The decrease in the expected communication efficiency under degraded communication is driven by increased congestion levels and by failures of electrical and co-located communication nodes and is related to the critical transition range for reliability. The developed methodology allows identifying the communication requirements for a reliable grid splitting execution under traffic uncertainty and congestion.