Multi-angle And Multi-dimensional Vulnerability Analysis Of Infrastructure Systems

Infrastructure systems play a crucial role in the stability and development of society. However, infrastructure systems are threatened by many inevitable factors, such as components ageing, natural hazards, man-made faults etc. So many scholars apply vulnerability to study performance of infrastructure systems under different kinds of attacks and hazards, which can identify vulnerable parts of systems to better protect infrastructure systems. Recently,many scholars have applied complex network based models and metrics to study vulnerability of infrastructure systems from different angles and perspectives. But how these models and metrics perform in vulnerability analysis and how correlated are these models and metrics are interesting questions. To study above problems, this paper introduces framework of multi-angle and multi-dimensional vulnerability analysis of infrastructure systems, then the framework is applied to analyze vulnerability of power grids and Chinese railway network.Taking power grids as an example, this paper selects three typical models, including purely topological model(PTM), betweenness based model(BBM) and direct current power flow model(DCPFM), and six frequently used vulnerability metrics, including efficiency based vulnerability VE, source-demand considered efficiency based vulnerability VSDE, largest component size based vulnerability VLCS, connectivity level based vulnerability VCL, clustering coefficient based vulnerability VCCO, and power supply based vulnerability VPS. Based on above models and metrics, this paper analyzes difference between models and correlation of different metrics on IEEE 300 power grid under random failures. And results show: the three models can produce almost identical topology-based vulnerability results at a large line tolerance parameter or large failure probability, and can produce almost identical flow-based vulnerability results only when all components fail. Also, the priority of PTM and BBM to better approach the DCPFM for vulnerability analysis mainly depends on the value of failure probability. In addition, source-demand considered topology-based metrics VSDE and VCL are strong correlated, while flow-based metric VPS has mild correlation with source-demand considered topology-based metrics VSDE and VCL.Taking Chinese railway network as an example, this paper selects four typical models, including purely topological model(PTM), purely shortest path model(PSPM), weight(link length) based shortest path model(WBSPM) and real train flow model(RTFM), and two vulnerability metrics, including accessibility based vulnerability VACC and flow-based vulnerability quantified by the fraction of trains which can run during a typical weekday VFOT. Based on above models and metrics, this paper analyzes difference between models and correlation of different metrics of Chinese railway network under random failures. In addition, this paper studies vulnerability of Chinese railway network from complementary perspective with considering the complementary effects of Chinese airline system or not. And results show: the purely topological model can be used to estimate the lower bound value of real vulnerability, and under some extreme events which can cause almost 60–70% of components fail, complex network based models can provide well approximation on the real accessibility-based vulnerability, but to produce almost identical real flow-based vulnerability from complex network based models, it requires almost all components failed. Also, the flow-based vulnerability and accessibility-based vulnerability have mild correlation. In addition, the complementary strength of railway network on airline network is much larger than that for airline network on railway network. In the case of that two networks are subjected to random failures, when 15% railway nodes are damaged, the complementary strength of airline network on railway network reaches the largest value when the complementary airline network is subjected to different levels of damage; while the complementary strength of railway network on airline network is a decreasing function of the airline node failure fraction.