【作者】 张宇栋；

【导师】 曹一家；

【作者基本信息】 浙江大学 ， 电气工程， 2013， 博士

【摘要】 近年来国内外发生了多起由连锁故障导致的大停电,给社会和经济带来了巨大的损失。因此迫切需要对连锁故障大停电的内在机理进行分析,提出预防大停电的措施。传统的电力系统安全分析方法在分析连锁反应事故和大停电机理上存在局限性,基于复杂系统理论的连锁故障大停电研究已经成为电力系统可靠性研究的热点问题。本论文结合复杂系统理论,从复杂网络拓扑演化、自组织临界性以及电网和通信网络的交互影响这3个方面对电力系统连锁故障大停电进行了分析,并提出预防大停电的相关措施。本文的具体工作如下：改进了OPA模型慢动态过程的电网拓扑演化方式。通过调节该模型的演化参数可以控制电网拓扑特征参数演变过程,从而研究各个网络特征参数的变化趋势对大停电分布的长期影响。仿真结果表明,单独地调节任意一种网络特征参数都可以减小大规模停电发生概率。但是在多数情况下,各个网络特征参数之间相互影响,往往同时发生变化,并且每个网络特征参数的变化对大停电分布的影响可能不同,甚至出现相反作用。在实际电网规划中,随着新输电线路的建设,电网的拓扑特性也在逐渐变化。在仿真中我们发现,将新建变电站节点或者发电厂节点连接到度数较低的节点时更有利于降低大规模停电的发生概率。引入复杂系统理论中同配性的概念,将线路的负载率和其脆弱性进行匹配,提出了一种新型的量化线路潮流分布特性的指标一一线路同配性指标,并应用该指标对系统的自组织临界性进行了识别。仿真结果表明,结合系统的平均负载率和潮流熵,该指标可以在多个系统的不同运行状态下有效地辨识系统的自组织临界状态,并且在系统平均负载率较低时仍具有较好的辨识能力。在预防大停电方面,线路同配性指标是现有指标的重要补充。以信息网络中输电线路的开断状态信息传输为例,结合隐性故障模型研究了电力信息网络节点故障对于电力系统连锁故障大停电的影响,并提出了对连锁故障大停电产生重要影响的关键信息节点的辨识指标。该指标可以在不同的状态下有效地辨识出对电力系统大停电产生重要影响的信息节点。仿真结果表明,信息系统的故障对于电力系统连锁故障起到了推波助澜的作用,且信息网络的高度数和高介数节点并不是对连锁故障传播产生重要影响的关键信息节点。信息节点对应的电力线路的脆弱性与信息节点在自身网络中的拓扑特征相比,前者更能表征信息节点在电力系统连锁故障大停电中的重要性。建立了较为符合电力系统运行实际的电力网络和通信网络交互作用模型,并通过调节相关参数设置了不同的路由策略来研究电力通信网络的传输特性对于电力系统连锁故障大停电的影响。最后,结合复杂网络的理论讨论了2个网络之间的内在相似策略对大停电的影响,并在此基础上提出了内在相似度指标。仿真结果表明,通过调整通信网络参数来设置恰当的路由策略,可以降低大停电发生的概率。在不同电力系统下仿真发现,网络间最优相似策略是相同的,采用该相似策略可以有效地降低大停电发生的概率。并且,内在相似度指标能很好地度量电网和通信网络的关联关系对大停电的影响,提高网络间的相似度可以有效抑制大停电的发生。更多还原

【Abstract】 In recent years, a number of large blackouts caused by cascading failures have taken place all over the world. Therefore, it is important to analyze the mechanism of cascading blackouts and propose preventive strategies for large blackouts. However, conventional security analysis strategies for electric power system confront the limitation in studying cascading failures and the mechanisms of large blackouts. And the research on cascading blackouts based on complex system theory has become a hot topic in the power system reliability research.This dissertation focuses on employing complex system theories to analyze cascading blackouts of power system from the aspects of complex network evolution, self-organized criticality and interactions between power grid and communication network, and propose some preventive strategies for large blackouts. This dissertation is organized as follows:The way of power grid topology evolution in slow dynamics of OPA model is improved to investigate the long-term effect of topology characteristics of a power grid on the probability distribution of blackout by adjusting the parameters of the novel model. Simulation results show that the probability of large blackouts in a power grid can be reduced by separately changing any one of these topological parameters. But in most cases, the variation of one topological parameter may cause simultaneous deviations on the others, which might have opposite effects on the distribution of blackout probability. Also in realistic power systems, the topology varies with the construction of transmission lines. From the simulation we found that linking a new substation or power plant to a low-degree bus is a better choice to mitigate large-scale blackouts than linking to a well-connected one.Inspired by the concept of assortativity in complex system theory, a novel index, called line assortativity, is proposed to quantify the distribution characteristic of lines power flows by matching load rate and vulnerability of lines. Line assortativity is used to identify the self-organized criticality of the power system. The simulation results indicate that the index is effective in identifying the self-organized criticality in various power systems operating in some states, by combining it with other indices, such as the average load rate and the power flow entropy. Moreover, the index proposed in this paper is still effective for the power system with a lower load rate, which is an important complement to the existing indices for mitigating large blackouts.Taking the transmission of line-outage-state information as an example, the information network is integrated with the hidden failure model to investigate the impact of node failures in the information network on the cascading failures in power systems. A new index is proposed to identify the key-buses which exert vital influence on blackouts, and it is very effective in power systems operating in various states. Simulation results show that the failures of information network might make the situation even worse and lead to large blackouts. Moreover, the nodes with high-degree or high-betweenness in the information network are not the key information nodes that have important influence in the spread of cascading failures. Rather, compared with the topological characteristics of information nodes in their own network, the vulnerability of the information nodes corresponding transmission lines are better able to characterize the importance of information nodes in the cascading failures of the power system.The model of interaction between a power grid and its communication network in accord with practical power system is proposed. And various route strategies could be generated by adjusting corresponding communication network parameters, by which the impact of transmission characteristics of electrical power communication on cascading blackouts is investigated. Moreover, the inter-similarity between coupled networks is discussed based on complex network theory. And an index is presented to assess the level of inter-similarity between interdependent networks. Simulative results indicate that, with proper routing strategy of communication network, the probability of large blackouts of power system can be reduced and the optimal inter-similarity strategy between coupled networks of different power systems is same, which can be adopted to effectively decrease the probability of large-scale blackout. The inter-similarity index is useful to assess the effect of interdependency between coupled networks on blackouts. And improving the inter-similarity between coupled networks is effective to mitigate large blackouts.更多还原