Global Sensitivity Analysis for Cyber-Physical Power Distribution Network

Abstract

Cyber-physical power systems (CPPS) operate under uncertainties from both the power and communication layers, yet there are limited systematic methods to evaluate uncertain cyber-physical impacts. To address this gap, we build a CPPS model by coupling a power distribution network represented by differential–algebraic equations (DAEs) and a discrete-event communication network with node-specific communication delays. On this basis, global sensitivity analysis (GSA) is introduced to quantify the impact of cyber-physical uncertainties on the power distribution network, thereby identifying the most critical factors that affect CPPS operation. Probabilistic models are used to simulate load, photovoltaic (PV) generation, and communication delays, and a Monte Carlo framework is built to estimate Sobol’ sensitivity indices. Finally, a CPPS digital co-simulation platform is developed which integrates OPAL-RT power system simulator and OMNeT++ communication network simulator via shared memory mechanism to validate the CPPS framework. Case studies demonstrate the quantitative relationship between the sources of uncertainty and their impacts within the CPPS, indicating that PV generation dominates the variance of bus voltages, whereas communication node delays mainly affect local voltage stability in the test system.