Vulnerability analysis of power grid with the network science approach based on actual grid characteristics: A case study in Iran

https://doi.org/10.1016/j.physa.2018.08.059Get rights and content

Highlights

  • We do not just use the centrality measure. We combined use of grid network topology and centrality measures with real and physical characteristics of power grid.

  • The advantage of this approach is that it eliminates complex and time-consuming differential calculations.

  • We used real dataset and the results were compared with dispatching center and verified.

Abstract

The vulnerability analysis of power grids has long been of paramount interest to researchers and authorities. Network vulnerability analysis refers to beforehand evaluation of the impact of local failures on the network as a whole so that proper measures could be adopted before occurrence of any major crisis. Previous studies on power grid vulnerabilities that are based on network science concept have been mostly concentrated on topological measures and unweighted networks. In this paper, grid vulnerability identification is carried out by combined use of grid network topology and centrality measures along with real and physical characteristics of power grid. Namely line load and failure rate, with the help of Weighted PageRank algorithm. The main advantage of this approach is that it eliminates complex and time-consuming differential calculations, which results in reduced computation time, and allows real-time updating of the result based on changes in the actual grid characteristics. The proposed model was validated by implementation over a section of Iranian 400 kV and 230 kV power grids. The striking accuracy of the achieved result was confirmed by comparison against results that obtained through calculations by Iran’s national dispatching body.

Introduction

Power grid is one of the most vital infrastructures of any country, and due to the dependency of a whole nation on this public utility, utmost reliability and constant availability of power supply is a basic necessity. Power grid topology refers to the network of high-voltage transmission lines that transmit electricity over long distances within and between countries. The vertices of aforementioned network are the generators and substations and its edges are high-voltage lines. While all transmission substations are important components of power grid, position and operating condition of some substations make them more essential for network stability, which in turn make their reliability, safety, and security all the more imperative. In this context, the term vulnerability analysis refers to the process of identifying key substations and evaluating their role in the reliability of electricity grid [1].

Two important aspects of power grid vulnerability analysis are: network topology and network operating condition. Network topology is concerned with how its components are connected while network operating condition deals with how the load is distributed over the grid. Network topology is essentially static while operating condition is dynamic and in constant change depending on grid characteristics such as peak load.

The majority of published works on power grid vulnerability analysis that employ the network science approach have focused on network topology and paid little attention to weights and directions of its components [2]. In the few cases where direction is taken into account, lack of attention to core characteristics of a power grid, namely line load and line failure rate, have made the results inconsistent with network reality.

In this paper, we first introduce a novel power grid vulnerability prediction method and then use this method both to analyze the condition, structure and vulnerabilities of Iran’s 400 kV and 230 kV power grids and to predict their vulnerabilities at different time horizons.

The following section, firstly, the literature on power grid analysis that is based on the network science is reviewed. Secondly, weighted power grid is considered based on the proposed method and identify power grid vulnerabilities by taking both network topology and grid characteristics. In the next section, the proposed method is applied for a fragment of Iranian power grids at 400 kV and 230 kV, and the results are evaluated against those obtained by engineers at the dispatching centers.

Section snippets

Power grid analysis based on the network science

In recent years, power grid vulnerabilities have received increasing attention from researchers in the field of complex networks and have been a topic of immense research. For example, vulnerabilities of US power grid were studied by Albert [1] (2004), Chassin (2005) [2], Wang (2009) [3], Hines (2010) [4], and Xu (2014) [5], and those of European power grid were explored by Rosas-Casals [6], [7], [8], [9] (2007–2010). This same concern in Italy was investigated by Crucittia (2004) [10], Bompard

The proposed method

The proposed method aims to identify power grid vulnerabilities by taking both network topology and grid characteristics (line loads and failure rates) into account simultaneously.

Network topology is determined by taking a snapshot of power grid at peak hours, which reflects the network behavior and thus vulnerabilities under worst case scenario. Network topology calculations are carried out using a slightly modified version of Weighted PageRank (WPR) algorithm. Shams and Khansari [30] have

Implementation of the proposed method on a section of Iranian power grid

The proposed method was tested on a section of Iranian power grid with 68 substations (nodes) and 98 transmission lines (edges). Data pertaining to this grid was gathered from the official site of Iran Grid Management Company (www.igmc.ir). After importing the data into the software Pajek [33], the diagram of this directed network under peak load was obtained as shown in Fig. 3.

Utilizing the proposed model required reliability calculations for each transmission line. The reliability

Conclusion

Transmission substations are all key components of power grids, however, due to particular location and operating conditions of some substations within the network, which affects power grid’s stability, these substations are attended with relatively higher sensitivity and their reliability is of more concern than their counterparts. A survey of previous studies on power grid vulnerability analysis that is based on the network science deal with power grids as unweighted and undirected graph,

References (35)

  • Rosas-CasalsM. et al.

    Assessing European power grid reliability by means of topological measures

    WIT Trans. Ecol. Environ.

    (2009)
  • Rosas-CasalsM.

    Power grids as complex networks: topology and fragility

    IEEE Comput. Eng

    (2010)
  • CrucittiaP. et al.

    Error and attack tolerance of complex networks

    Physica A

    (2004)
  • BompardE. et al.

    The concept of betweenness in the analysis of power grid vulnerability

    IEEE Comput. Eng.

    (2010)
  • FenuaG. et al.

    A model of assessment of collateral damage on power grids based on complex network theory

    Procedia Comput. Sci.

    (2014)
  • MeiS. et al.

    Power Grid Complexity

    (2011)
  • DingJ. et al.

    The improvement of the small-world network model and its application research in bulk power system

  • Cited by (0)

    View full text