基于人工智能算法的配网重构和拓扑分析的研究

发布时间：2018-08-17 15:38

【摘要】：配电网网络重构是配电系统极为重要的优化手段。随着配电网网络结构的日趋复杂和人们对电能质量要求的提高,配电网自动化成为必然的趋势,其中,网络重构是实现配电网自动化的关键环节。配电网重构是一种不需要增加新设备的优化方法,仅通过联络开关或者分段开关的开合状态的调整,从而改善电网品质。重构可有效提高电网运行的安全性,与用户用电息息相关,因此,网络重构作为一个重要优化手段,对配电系统的重要性日益显现。配电网络拓扑分析是电力系统分析的基础板块,也是配电系统其它分析顺利进行的保障。在阐述了配电网网络设备的描述、配电网的简化、以及面向对象技术在配电网中的应用等基础上,本文探讨了配电网络拓扑的生成和配电网拓扑结构的遍历方法,包括广度优先搜索方法和深度优先搜索方法。并在配电网拓扑结构的遍历方法的基础上,研究了静态拓扑的生成和动态拓扑的生成。包括对静态拓扑的等值节点分析和电气岛分析进行介绍,以及动态拓扑跟踪中断路器“开”、“合”两种状态下的等值节点与电气岛的分析。最后,以一个简易的网络模型作为算例,进行配电网拓扑分析研究。通过配电网拓扑分析,文中利用图论理论简化配电网拓扑结构,在原始数据的基础上建立拓扑结构表,并形成层次关系,简化潮流计算的复杂性。系统的阐述了当前的配电网潮流计算的方法,利用拓扑分析形成的层次关系进行前推回带算法进行潮流计算,并结合算例验证。随着时代的变迁,配网重构的方法也日趋多样化,配电网重构方法主要有数学优化方法、启发式方法以及人工智能算法等。其中,人工智能算法在配网重构中的运用得到了相当的重视。本文将混沌技术引入二进制粒子群算法中,并对其进行改进,以克服二进制粒子群算法容易陷入局部最优的缺点,得到更精确地优化结果。文章最后,以网络损耗最小为目标函数进行网络重构,将算法进行matlab编码,并通过算例进行验算。实验结果证明,该方法能有效降低网络损耗。同时,提高了最低节点电压。通过算例验证,该方法具有可行性和有效性。
[Abstract]:Distribution network reconfiguration is an important optimization method of distribution system. With the increasing complexity of distribution network structure and the improvement of power quality requirements, distribution network automation has become an inevitable trend. Among them, network reconfiguration is the key link to realize distribution network automation. Distribution network reconfiguration is an optimization method that does not need to add new equipment. It can only improve the quality of power network by adjusting the opening and closing state of the connection switch or segmented switch. Reconfiguration can effectively improve the security of power grid operation and is closely related to consumer power consumption. Therefore, as an important means of optimization, network reconfiguration is becoming increasingly important to distribution system. Distribution network topology analysis is the basic block of power system analysis, and it is also the guarantee of other analysis of distribution system. Based on the description of distribution network equipment, the simplification of distribution network and the application of object-oriented technology in distribution network, this paper discusses the generation of distribution network topology and the traversal method of distribution network topology structure. It includes breadth first search method and depth first search method. Based on the traversal method of distribution network topology, the generation of static topology and dynamic topology is studied. This paper introduces the equivalent node analysis and the electric island analysis of the static topology, as well as the analysis of the equivalent node and the electric island in the two states of the circuit breaker "on" and "close" in the dynamic topology tracking. Finally, a simple network model is taken as an example to analyze the topology of distribution network. Based on the topology analysis of distribution network, the topology structure of distribution network is simplified by using graph theory, the topological structure table is established on the basis of original data, and the hierarchical relationship is formed, which simplifies the complexity of power flow calculation. The current power flow calculation method of distribution network is systematically expounded. The forward push back band algorithm is used to calculate the power flow by using the hierarchical relation formed by topological analysis, and an example is given to verify it. With the change of times, the methods of distribution network reconfiguration are becoming more and more diverse. The methods of distribution network reconfiguration include mathematical optimization, heuristic and artificial intelligence. Among them, the application of artificial intelligence algorithm in distribution network reconstruction has received considerable attention. This paper introduces chaos technique into binary particle swarm optimization algorithm and improves it to overcome the shortcoming of binary particle swarm optimization algorithm which is easy to fall into local optimum and obtain more accurate optimization results. Finally, taking the minimum network loss as the objective function to reconstruct the network, the algorithm is coded by matlab and verified by an example. Experimental results show that the proposed method can effectively reduce the network loss. At the same time, the minimum node voltage is increased. An example shows that the method is feasible and effective.
【学位授予单位】：福州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM711;TP18

【相似文献】