【作者】 王淑良；

【导师】 陈学广；

【作者基本信息】 华中科技大学 ， 控制科学与工程， 2012， 博士

【摘要】 基础设施是一类为社会生产和居民生活提供服务的物质工程设施，它们是社会赖以生存和发展的一般性物质条件，也是典型的复杂系统。当今社会越来越依赖于基础设施系统提供的各种服务以维持其正常运行。然而近年来以北美大停电、08年中国南方雪灾为代表的各类灾害频繁爆发，给基础设施系统的安全敲响了警钟。鉴于基础设施系统对社会生产和人民生活的重要性，研究基础设施系统在不同扰动下的响应情形，分析、理解、识别基础设施系统的脆弱性，进而提出相应的抗干扰措施和改进策略，提高基础设施系统的可靠性，成为了当前研究的一个热点课题。围绕基础设施系统的脆弱性研究及其应用，国内外众多研究机构和学者分别采用投入产出模型、系统动力学模型、Agent仿真模型，探讨了灾害对基础设施系统中的经济体、功能器件、业务流程等的影响并取得了若干初步成果，但是就总体而言，基础设施系统脆弱性的研究尚处于起步阶段，特别地，对于关联基础设施系统脆弱性研究目前仍然是一个较新的研究领域。复杂网络作为大量真实复杂系统的高度抽象，其理论与应用研究已渗透到数理、生命和工程学科等众多不同的领域。利用复杂网络的理论和方法能够很好的描述基础设施系统的内部结构和彼此间的关联关系，为基础设施系统脆弱性的研究提供一个全新的视角和研究方法，有助于从整体上把握系统的复杂性和系统响应的动力学特性。基于此，本文运用复杂网络的理论和方法，从单一基础设施系统入手，以关联基础设施系统为研究重点分析基础设施系统的脆弱性。主要研究内容如下：首先，利用复杂网络的理论和方法对基础设施系统进行了拓扑描述和功能刻画，介绍并总结了描述基础设施系统的聚类系数、度相关性、介数等参量，以便更深刻的理解基础设施系统结构与功能特性，给出了基础设施系统的脆弱性分析理论框架。针对以往单一基础设施系统网络脆弱性研究中对边的级联故障和社团脆弱性很少提及这一情形，本文在一般脆弱性分析的基础上，建立了节点和边的级联故障模型，分析了基础设施系统网络在节点故障和边故障下的脆弱性。同时利用模块度优化算法给出了社团结构划分方法，并结合基础设施系统网络的结构特性和功能特点分别提出了结构和功能脆弱性指标，用以寻找网络最脆弱和鲁棒的社团。对于关联基础设施系统，本文分别从关联类型、运行状态、基础设施特性、故障类型、耦合和响应行为、基础设施环境等不同方面分析了其关联性，结合基础设施系统脆弱性分析框架，对关联基础设施系统进行了脆弱性分析。在研究中对每一个具体系统进行了拓扑结构抽象，结合其运行机制建立了功能模型，考虑了不同系统间的关联关系，并给出了关联基础设施系统接口设计策略。利用网络效率、雪崩规模等性能指标分析了系统在不同接口、不同扰动策略下的脆弱性，得出了关联基础设施系统的故障容忍性和最优的接口设计标准。脆弱性分析的目的是为了更好的保护基础设施系统。在单一和关联基础设施系统脆弱性分析基础上，探讨了基础设施系统的关键组件辨识方法，以便对这些关键组件进行重点保护。通过纳入缓冲、增加冗余、故障修复等手段讨论了基础设施系统的改进策略，为控制基础设施系统的级联故障、增强系统的抗干扰能力提供决策支持。更多还原

【Abstract】 Infrastructure systems are the framework of material engineering facilities thatprovide service for the social production and the resident’s life. It is the general materialcondition of survival and development for society. It is a topical complex system.Societies today are increasingly dependent on such systems to maintain normal operation.However, events of recent years such as the North American blackout and the ChineseSouthern snowstorm have raised alarms in regards to overall safety. In view of theimportance of infrastructure systems for social well-being and sustenance, studyingresponse of infrastructure systems under different types of disturbances, analyzing,understanding and identifying vulnerability of infrastructure systems for the purpose ofimprovement, has become a focal area for the process of optimizing infrastructure designand protection.Utilizing IIM, system dynamics model and agent simulation model, manyauthoritative agencies and researchers have made an research on the tender inside theinfrastructure systems, and discussed the disaster affection to the economic system,functional device and the business processes, etc. But the researchers only get certainpreliminary achievements because of the complexity of infrastructure systems. Therefore,the study in vulnerability analysis of infrastructure systems has been at the beginning stagewithout forming a complete theoretical and methodological system.As abstraction of many real complex systems, complex network theory has beenapplied to the area of mathematics, life sciences and engineering. It thus provides a newperspective and applied research method to vulnerability analysis of infrastructure systems,allowing a level of analysis that can approach the overall complexity and dynamics ofinfrastructure systems. Using the theory and method of complex networks, this paperstudies the vulnerability of infrastructure systems, starting from a single system and thenprogressing to interdependent systems. Main contributions of this dissertation aresummarized as follows:First, this paper gives the theoretical framework for vulnerability analysis, andanalyses the complexity of infrastructure systems. Combined with the theories and methods of complex network, topology description and functional characterizations ofinfrastructure systems are studied. Meanwhile, to give a more profound understanding ofthe structural and functional features of the infrastructure systems, parameters such asclustering coefficient, degree, and betweenness are introduced and summarized.Because vulnerability analysis of single infrastructure systems often focuses on nodecascading failures and the impact of external factors (such as natural disasters,) edgecascading failures, community structure implementation and community structuralvulnerabilities are rarely mentioned. This article establishes a model of edges cascadingfailures, and analyses the vulnerability of infrastructure systems. Meanwhile, a communitydetection method and vulnerability metric are proposed to locate the weakest and mostrobust communities.Interdependent infrastructure system vulnerability analysis is a relatively new area ofresearch. Current research in this area has not yet reached maturation of theory andmethodology, especially in the field of modeling. Thus, this paper first studies theinfrastructure system from different aspects such as interdependency types, operationstates, infrastructure characteristics, failure types, coupling and response behavior, andinfrastructure environment. It then establishes a framework for vulnerability analysis.The purpose of vulnerability analysis is to better protect the infrastructure systems.On the basis of vulnerability analysis on single and interdependent infrastructure systems,by considering the performance loss of systems undergoing edge or node failures, keycomponents of an infrastructure system are proposed and identified and networkinterventions as well as improvement strategies are discussed to better protect theinfrastructure systems.更多还原