Methods for Evaluating Critical Lines and Nodes in Cyber–Physical Power Systems From Three Network Perspectives

Abstract

A typical modern power system is a multilayer network consisting of a power network coupled with a communication network, namely, a cyber–physical power system (CPPS). Some previous research methods that were applicable to single-layer networks (such as power networks) are not applicable to current CPPSs. In this article, we first establish an importance evaluation model for the lines in the power network and the nodes in the communication network of a CPPS. Then, the key nodes and lines in the CPPS are evaluated from three perspectives: network information, network properties, and network structure. Two algorithms are proposed, from network information and network property perspectives. Ten indicators are also formulated from a network structure perspective. Finally, the results of analyzing the correlations between indicators under these three perspectives using Kendall's tau-b show that the important lines and nodes in the different networks are different and that the most effective method of importance evaluation is based on information fusion from the network information perspective. The results also show that the research significance of investigating CPPS network node importance based on a Barabási–Albert communication network is greater than that of analyses based on Watts–Strogatz and Erdös–Rényi networks. The results can provide a good understanding of the operating status of a power system to help reduce the probability of system faults, thereby providing support for future intelligent and autonomous power systems.