基于复杂网络理论的连锁故障建模与预防研究

发布时间：2018-07-24 09:32

【摘要】：复杂网络连锁故障是一种时常发生在基础设施网络上的动力学现象,尤其是近年来发生在复杂电网中的大规模停电事故,给社会经济和人们生活带来了巨大的损失和影响。因此,复杂电网连锁故障分析与预防已经引起了电力工作者的高度重视,并且已经成为一个亟待解决的问题。为了克服过于注重个体动态特性的传统安全分析方法在揭示复杂电网整体动态行为上的缺陷,从系统论和整体论角度出发的系统科学理论与方法为揭示连锁故障的全局动态行为提供了新的研究思路。复杂网络理论是系统科学中的重要分支之一,从网络拓扑的视角研究网络结构与网络功能之间存在的密切关系。根据复杂电网的拓扑连接关系,建立相应的电力网络拓扑模型,并依此构建适合复杂电网连锁性停电事故分析的连锁故障模型,以探讨发生在电力系统中的连锁故障动力学特性,继而提出复杂电网连锁故障预防策略是抑制连锁故障传播以提高电力系统安全性的一种有力途径。为此,本论文从复杂网络的视角出发,考虑电力网络所具有的特殊电气特性以及节点重要性对复杂电网连锁故障动力学行为影响上存在的差异,重点研究了复杂电网重要节点的识别算法以及连锁故障的建模过程,并进一步提出了连锁故障预防模型。本论文的具体研究工作如下：1、复杂电网拓扑特性分析。基于复杂网络理论分析网络的动力学性质主要是从构成网络骨架的节点和连边入手,研究网络拓扑对网络功能的影响。该方法有别于传统电力系统分析,因此本文着重探讨了这两种分析方法之间存在的区别与联系。进一步地,介绍了复杂电网的基本建模方法并依此构建了复杂网络拓扑模型,分析了复杂电网拓扑统计特性以及其与连锁故障动力学行为之间存在的复杂的关系。2、复杂网络节点重要性识别算法研究。根据网络拓扑结构的统计参量,首先提出了一种考虑节点对之间存在相互依赖关系的关联度中心性指标,并在其基础上提出了一种基于信息熵的多属性节点重要识别算法；进一步地,建立了考虑复杂电网自身电气特性的节点重要度评价指标,并依此提出了一种基于基尼系数的多属性节点重要度识别算法。最后通过在IEEE-39节点系统网络中的仿真分析,对比分析了本论文提出的两种复杂电网节点重要度识别算法在评价效果上的可行性和准确性。3、复杂电网连锁故障建模研究。考虑到电力系统中的大规模潮流转移和隐性故障是触发复杂电网连锁故障的两大主要因素,故本文从复杂网络的研究视角分别建立了基于拓扑的电力系统隐性连锁故障概率模型、电气负载下的连锁故障模型以及计及隐性故障的电气负载下的连锁故障模型。通过在IEEE-300节点系统网络中的仿真研究,分析了隐性故障与复杂电网的小世界性对连锁故障规模的影响以及不同负载定义形式与连锁故障鲁棒性之间的关系。4、复杂电网连锁故障预防建模研究。考虑到复杂网络节点重要度对连锁故障动力学行为之间有着密切的联系,建立了相应的节点重要度基础上的连锁故障预防模型,分析了其在降低网络建设成本与提高复杂网络鲁棒性之间的关系,并进一步地探讨了根据不同节点重要度评价结果建立的复杂电网节点容量优化配置模型对建立低成本且高鲁棒性网络的影响。
[Abstract]:Complex network cascading failure is a dynamic phenomenon that often occurs on the infrastructure network, especially in recent years, which has occurred in a complex power network, which has brought huge loss and influence to the social economy and people's life. Therefore, the analysis and prevention of complex grid linkage barrier has caused the power workers. In order to overcome the shortcomings of the traditional security analysis method which overly pays attention to the dynamic characteristics of the individual, the system theory and methods of system science and holism are provided to reveal the global dynamic behavior of the cascading failures. The complex network theory is one of the important branches of system science. It studies the close relationship between network structure and network function from the perspective of network topology. Based on the topology connection of complex grid, a corresponding topology model of power network is set up. The linkage fault model is analyzed to discuss the dynamic characteristics of the cascading failure in the power system, and then it is proposed that the chain fault prevention strategy of complex power grid is a powerful way to suppress the transmission of cascading failures in order to improve the security of the power system. The difference between the electrical characteristics and the importance of the node importance to the dynamic behavior of the complex network cascading failure is studied. The recognition algorithm of the important nodes of the complex grid and the modeling process of the cascading failures are studied, and the chain fault prevention model is further proposed. The research work of this thesis is as follows: 1, the topology characteristics of the complex power grid Analysis of the dynamic properties of the network based on the complex network theory is mainly from the nodes and sides that constitute the network skeleton and the influence of network topology on the network function. This method is different from the traditional power system analysis. Therefore, this paper focuses on the differences and relations between the two methods. The basic modeling method of complex grid is introduced and a complex network topology model is constructed. The topological statistical characteristics of complex power network and the complex relationship between the complex network and the dynamic behavior of the cascading failure.2, the research of the importance recognition algorithm for complex network nodes are studied. On the basis of it, a kind of important recognition algorithm for multi attribute nodes based on information entropy is proposed, and the evaluation index of node importance considering the electrical characteristics of complex power grid is established, and a kind of Gini coefficient is proposed. Multi attribute node importance recognition algorithm. Finally, through the simulation analysis in the IEEE-39 node system network, the feasibility and accuracy of the evaluation results of two kinds of complex grid node importance recognition algorithms proposed in this paper are compared and analyzed.3, the research of complex network cascading failure modeling. Transfer and recessive faults are the two main factors that trigger the cascading failures of the complex power grid. Therefore, this paper sets up the hidden chain fault probability model of the power system based on the topology of the complex network, the chain fault model under the electric load and the cascading failure model under the electric load of the recessive fault. The simulation research on EE-300 node system network analyzes the influence of the small world on the scale of the cascading failures and the relationship between the different load definition forms and the robustness of the cascading failures, and the research on the modeling of the complex network cascading failure prevention. Considering the dynamics of the complex network nodes, the linkage failure dynamics is considered. There is a close relationship between behavior, and a chain fault prevention model based on the corresponding node importance is established. The relationship between the cost reduction of network construction and the improvement of the complex network robustness is analyzed, and the optimal configuration of the complex network node capacity based on the evaluation results of different node importance is further discussed. The impact of the model on low cost and high robustness network is established.
【学位授予单位】：西南交通大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TM73

【参考文献】