Application of node centrality in transmission expansion planning under uncertainty

Abstract

Methods for transmission expansion planning have long been associated with detailed scenarios for the location and types of future generation and expected load growth. Such expansion futures might seek transmission additions to meet short/long term needs, while addressing anticipated reliability issues, reduction in CO2 emissions, integration of renewable sources of energy, among other objectives and constraints. Today, there is tremendous uncertainty and a myriad of possibilities in future generation mix and placement, load growth, and storage technology. Also, stringent requirements in CO2 emission policy have emerged in recent years. These uncertainties are putting great strain on present practice in transmission expansion planning. In lieu of scenario driven methods, we explore a different philosophy in approaching the transmission expansion planning problem. Based on properties that are inherent to the structure of the network, the techniques presented here make use of results from spectral partitioning and concepts from graph centrality to suggest most suitable bus choices for placement of transmission lines. Derived from the circuit concepts of effective resistance and resistance distance, a measure closely associated with the bus impedance matrix (Zbus) widely used in power system fault studies, the centrality measure adopted in this work will correspond to the information centrality (IC) metric employed in other graph problems. While there exist a number of other well-known centrality indices, e.g., shortest path betweenness, closeness, degree, eigenvector centrality, this work will argue that IC proves best suited to the transmission expansion planning problem. These results will be demonstrated via an illustrative case based on IEEE 300-bus system.