Research On The Structural Vulnerability Of Electrical Cyber-Physical System Based On Interdependent Networks Theory

With the development of smart grid and the the increasing sizes of interconnected grid, the power grid, which is refered to as primary side, relies on the reliable operations of the advanced information and communication system more than ever. The information and communication system, which is deeply combined with power grid, has already become its brain and nervous system and gradually evolves to the electrical cyber-physical system. The advanced ICT does enhance the contrability and observability of power grid. On the other hand, the failures of cyber or communication system may spread to power grid by the coupling relationships, even conducting to interactive cascading failures between power grid and its cyber network. The multiple blackouts happening in recent years also indicate that the failures of cyber or communication system deteriorate the scales of blackous. So it's of therotical value and realistic significance to conduct the research on the vulnerability of electrical communication network, electrical cyber-physical system and their protection strategies. Based on interdependent networks theory, the main works of this dissertation, which focuses on the aspect of structural vulnerability, are organized as follows:1) Firstly, this chapter reviews on interdependent networks theory and its applications in the structure vulnerability analysis of electrical cyber-physical system. The currently existed model method, interdependent networks model and the vulnerability analysis algorithm are summarized. The vulnerability factors, i.e., topological characterists, dependency link and topological inter-similarity, and the corresponding protection strategies are analyzed. This paragraph can improve the understanding of the interdependent networks.2) Secondly, this chapter investigates the structure vulnerability of electrical communication network and its protection strategy. We find that the degree distribution of electrical communication network is similar to scale-free distribution. It's the structure that lead to which the electrical communication network is robust to random failures but extremely vulnerable to targeted attacks. To improve the situation, this chapter proposes three nodes protection strategis and 4 link addition strategies to improve the structural vulnerability. The simulation results show that high betweenness node protection strategy and low degree link addition strategy can effectively improve the structural robustnss of electrical communication networks.3) Thirdly, the structural vulnerability of electrical cyber-physical system is invsstigated based on interdependent networks theory. The "partial one-to-one correspondence" electrical cyber-physical interdependent networks model is built and the structural vulnerabilities of three realistic electrical cyber-physical interdependent networks are analyzed. Contrary to the previous researches, this chapter proves that the realistic electrical cyber-physical interdependent networks will not undergo first-order transition under random failures. A novel inter-similarity index is proposed and the reasons for the phenomenon are explained from the perspective of topological inter-similarty. In the end, this chapter verifies that the vulnerability curve obtained by complex network can still reflect the vulnerability of electrical cyber-physical interdependent networks.4) Fourthly, this chapter researches the link addition strategy of interdependent networks. Four link addition strategies previously proposed are compared in interdependent networks and the simulation results show that allocating the connectivity links equally to the both networks is more effective in enhancing the robustness of interdependent networks than allocate to the isolated networks. According to the interdependency relationships, this chapter proposes the notion of inter degree-degree difference and two novel links addition strategies. By comparing the six link addition strategies in three different types of interdependent network and two realistic electrical cyber-physical ststems, the effectiveness of our methods is verified.