A Real-time Quantitative Framework for Survivability Evaluation of Smart Grids

Responding to high impact events, such as cyberattacks, during the first phase of resilience, i.e., survive, plays an important role in a system resilience enhancement. To this end, a real-time quantitative framework is highly needed. Motivated by this, this paper proposes a novel real-time cyber-physical resilience-based survivability metric for smart grids’ survivability assessment. The proposed method takes into account the survivability of both the electrical, i.e., physical, and cyber systems. A survivability margin is established for Power-side Survivability (PsS) to monitor the power system’s capacity to maintain the functionality of its critical components. Available Generation Capacity (AGC) and Network Accordant Connectivity (NAC) are factors that are considered while calculating PsS. Based on the available pathways between the Human-Machine Interface (HMI) and Intelligent Electronic Devices (IEDs) in a substation, the Cyber-side Survivability (CsS) is calculated for a given cyber configuration. The computational complexity of the metric calculation limits the metric to real-time measurement. To address this issue, the Histogram-based Gradient Boosting Regression Tree (HGBRT) model is used to forecast the metric depending on system circumstances. To demonstrate its efficacy, the proposed metric is tested on the PJM 5-bus system.