Robustness analysis of electricity networks against failure or attack: The case of the Australian National Electricity Market (NEM)

Abstract

This study explores network science algorithms for the robustness analysis of electricity networks. We first investigate the characteristics of key network models including random graphs, small-world, and scale-free networks. Then, various measures are explored for the robustness of such networks against failure or attack, utilizing topological features and percolation theory. Both weighted and unweighted scenarios are studied, with network voltage considered as the edge weight. For a case study, we investigate the network characteristics as well as the robustness of the Australian National Electricity Market (NEM) network on the basis of these models and theories.

The NEM is the world's longest interconnected power system, with an end-to-end distance of over 5000 km between the state of Queensland in the north and the state of South Australia. Our data contains 2375 transmission lines and 1538 nodes as generators or large demand customers. Our study shows that the NEM as an unweighted network is a small-world network (with exponential degree distribution). However, as a weighted network (considering the voltage capacity of nodes), it has a scale-free topology (following a power-law degree distribution). Robustness analysis revealed that the NEM presents relatively stronger robustness when facing random errors than when facing intentional attacks to nodes with a high degree centrality. It also revealed the sensitivity of the scale-free network to deliberate attacks directed toward important “hubs” (interconnected nodes).