智能算法在风电场布置中的应用研究

发布时间：2018-02-23 20:37

本文关键词： 粒子群算法邻域结构小世界网络模型风电场布置　出处：《石家庄经济学院》2015年硕士论文　论文类型：学位论文

【摘要】：目前这些可发现的绿色能源中,发展最快的能源之一为风能源。基于我国是一个含有巨大风能资源的国家,如果此资源能够得到充分利用,将会带来巨大的经济和社会效益。现如今很多研究人员通过设计合理的风力发电场布置将此资源得以充分利用,但是目前出现的问题有,风电场优化布置设计过度依赖商业软件、风电场的发电成本高和单位风力发电机出力低等。现在通过大量实验证明,通过智能算法对风电场布置进行优化后,上述问题得到了有效改善,如提高了单位风力发电机的出力同时也降低了发电成本等。本文对风电场布置问题进行优化主要在以下两方面:风力发电机优化布置和风力发电机机组选型优化。首先针对风力发电机布置问题分别进行三种智能算法求解、比较得出更适合针对此问题进行优化的粒子群算法。虽然比较得出更适合优化此问题的算法,但是与其他文献比较得出,此算法还未达到优化目标,需要对算法继续进行改进。考虑到算法搜索模式及策略对最终找到最优解的影响,因此可以试图对算法中粒子的搜索模式及策略进行改进来达到进一步优化其问题的目标。粒子群算法中粒子的搜索模式所具有的网络结构可以认作为数学中节点的拓扑网络结构,完全规则拓扑网络拥有高度的聚集性,完全随机拓扑网络具有小的平均路径长度这些显著优点,但是这两类拓扑网络并不能显示真实的网络拓扑结构。本文中所介绍的小世界网络模型则符合真实的网络拓扑结构,并且此模型具有平均路径短,聚集系数高等特点。基于小世界网络模型的特点,将粒子群算法中粒子的搜索模式设定为局部模式,且每个粒子的邻域结构(搜索空间)构造为小世界网络模型的结构。此种方法改进后的粒子群算法中粒子的搜索深度有所增加,也避免出现过早收敛现象,从而有效地提高了解的质量,促使算法最终能够找到更好的最优解。最后通过实验证明,改进后的粒子群算法进行优化风电场布置问题能够得到很好的效果。
[Abstract]:One of the fastest-growing sources of green energy is wind energy. Based on the fact that China is a country with huge wind energy resources, if this resource can be fully utilized, There will be tremendous economic and social benefits. Nowadays, many researchers use this resource by designing a reasonable wind farm layout, but the problems that arise now are:. The optimal layout design of wind farm is too dependent on commercial software, the cost of wind farm is high and the output of unit wind turbine is low. Now, through a large number of experiments, it is proved that after optimizing the layout of wind farm by intelligent algorithm, These problems have been effectively improved, In this paper, the layout of wind farms is optimized in the following two aspects: the optimal layout of wind turbines and the selection of wind turbine units. Firstly, three kinds of intelligent algorithms are used to solve the problem of wind turbine layout. The particle swarm optimization algorithm, which is more suitable for the optimization of the problem, is compared. Although the algorithm is more suitable for the optimization of the problem, compared with other literatures, the algorithm has not reached the optimization goal. It is necessary to continue to improve the algorithm. Considering the influence of algorithm search mode and strategy on finding the optimal solution, Therefore, we can try to improve the particle search mode and strategy in the algorithm to achieve the goal of further optimizing its problem. The network structure of particle search pattern in particle swarm optimization algorithm can be regarded as the node in mathematics. The topology of the network, Complete regular topological networks have a high degree of aggregation, and complete random topological networks have the remarkable advantages of small average path length. However, these two kinds of networks can not show the real network topology. The small-world network model described in this paper accords with the real network topology, and the model has a short average path. Based on the characteristics of small-world network model, the particle search mode in particle swarm optimization algorithm is set as local mode. The neighborhood structure (search space) of each particle is constructed as the structure of the small-world network model. The improved particle swarm optimization algorithm increases the searching depth of particles and avoids the phenomenon of premature convergence. Thus the quality of the solution can be improved effectively, and the algorithm can finally find a better optimal solution. Finally, it is proved by experiments that the improved particle swarm optimization algorithm can achieve good results in wind farm layout optimization.
【学位授予单位】：石家庄经济学院
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP18

【相似文献】