A Dependency Links Removal Strategy to Improve Robustness of Cyber-Physical Power Systems

Cascading failures in smart grids can lead to widespread blackouts, which result in huge economic losses and catastrophic consequences. We find that the propagation of cascading failures can be expended by the topological imbalance between nodes at both ends of the dependency links cyberphysical power systems (CPPS), further weakening robustness of the system. Therefore, we propose a strategy to improve network robustness via imbalanced dependency links partial removal. First, we define an imbalance index of dependency links to assess the imbalance of nodes at both ends of an interdependent link by nodes’ importance, which integrates topology and function because the difference assessment between dependency links only from topology is incomplete. Then, we perform dependency links partial removal according to the ranking of dependency links imbalance index (DLI) in descending order. Finally, we employ IEEE39-bus case and scale-free network to model a cyber-physical power system. The numerical simulation illustrates that the proposed dependency links strategy is practical and feasible, robustness can be improved 50%-100%, the effect is even more obvious for high importance nodes. Moreover, the proposed strategy has obvious advantages over using an index in terms of a single topological metric (e.g., node betweenness) for interdependent links removal.