Modeling And Prevention Of Cascading Failures Based On The Complex Network Theory

The cascading failure is a common dynamic phenomenon in infrastructure networks, especially the large-scale blackouts happened in power grids in recent years, which affect people lives and inflict huge economic loss to the society. Therefore, many scholars have paid attention to the analysis and prevention of the cascading failures in complex power grids, which need to be solved urgently. As we know, the traditional safety analysis method focuses too much on the individually dynamic characteristics in revealing the wholly dynamic characteristics of the complex power grids. To solve the problem, we have to recognize and analyze the power grids from the perspectives of the system theory and holism, and combining the system science principles. Complex network theory is one of the main branches of system science, and can investigate the relationship of the network structure and function from the network topology perspective. According to the topology connection relationship of the complex power grids, we establish the network topology model and build the cascading failure model which fits to analyze the cascading failure blackout of the complex power grids. And then we give the prevention strategies for cascading failures of the complex power grids based on node importance evaluations. The prevention strategies are useful to prevent the spread of cascading failures and improve the safety of the power systems. So in this thesis, from the angle of the complex networks, we put emphasis on the node importance identification and modeling cascading failures in terms of considering the electrical properties of the complex power grids itself and the influence disparity of the node importance on dynamic behaviors of cascading failures of the complex power grids. Furthermore, we propose the cascading failure prevention model. The details of the thesis are as follow:1. The topological characteristics analysis of the complex power grids.The dynamics analysis is to investigate the effcts of the network topology on the network function from the perspective of the nodes and edges, which constitute the network together. Since there are distinct differences between the complex network theory method and the traditional power system analysis, the relationship of the two methods mentioned above is explored in this thesis. By building the complex network topological model, we analyzed the relationship of the statistical characteristics of the complex power grids and the dynamic behavior of the cascading failures.2. The node importance identification of the complex networks.According to the topological statistics of the network, we firstly propose a correlation centrality index, based on which we propose a multi-attribute node importance decision making method based on entropy. Besides, we present some node importance evaluation indices which consider the electrical characteristics of the power grids, and then we put forward a ginicoefficient-based multi-attribute node importance evaluation method. At last, we do simulations on the IEEE-39 node system, and make comparison of the feasibility and accuracy between the two node importance evaluation methods mentioned above.3. Modeling cascading failures in complex power grids.Considering the large-scale power transferring and the hidden failures are two main contributing factors of triggering cascading failures in complex power grids, we build three models from the angle of the complex networks, and they are hidden cascading failure probability model, cascading failure model under electrical loads, and cascading failure model under electrical loads taking hidden cascading failures into account. We analyze the impacts of hidden failures and small world properties of the complex networks on the scales of the cascading failures, and also the relationship of different load definitions and the network robustness under cascading failures by simulations on the IEEE-300 network.4. Modeling cascading failure prevention strategies in complex power grids.Due to the fact that different node importance affects the dynamic behaviors of the cascading failures differently, we build the cascading failure prevention model based on the node importance evaluation. Furthermore, we investigate the relationship of lowering the construction cost and improving the network robustness under this model. Besides, we discuss the effects of node capacity allocation on constructing a low cost and high robustness network under different node importance results.