大规模电网连续时段概率潮流计算仿真与分析

论文总字数：36756字

摘 要

16012327 孙启锐

指导教师 喻洁

近阶段，随着能源利用结构的调整，我国开始大力推广新能源发电技术，风力发电就是其中一种新能源形式。但是由于风力发电过于依靠环境，不能稳定控制发电量，大力发展风力发电给电网的稳定运行带来不确定性和冲击性。传统的确定性潮流计算的局限性逐步展现，考虑到系统中存在的随机因素，概率潮流计算逐渐得到推广和发展。

考虑大规模电网机组出力中存在的不确定性因素，本文采用概率潮流计算方法，即是结合概率思想，应用概率论中的方法对大量潮流计算结果进行概率性分析。通过概率潮流计算，可以得到线路承载有功功率的概率，进一步分析得到其越限概率值，将系统运行的不确定性具体量化到线路越限概率，便于调度部门进行合理的调度安排。

对于复杂的电力数据网络，首先建立电力系统模型，对冗杂的数据进行整理，通过对线路中所包含的变压器进行处理，弱化变压器对于线路的影响，并且进行标幺值转化，构成一个整体不包含变比的电力网络；根据建立起的电力系统模型，找出其中的风电场，基于拉丁超立方采样法的原理，为了考虑随机变量概率分布的尾部特性因素，本文提出改进拉丁超立方重要采样法；对于包含多个风电场的电力系统，我们需要考虑不同风电场之间的相关性，在此引入Copula函数，将风电场之间的相关性模型化和数据化；根据改进拉丁超立方重要采样法和Copula函数对风电场的出力数据进行抽样，将此数据应用在已经建立的电力系统模型的概率潮流计算中，并在MATLAB中以核密度估计方法对概率潮流计算的离散结果进行拟合；对标准IEEE-118节点系统进行测试，并以蒙特卡罗模拟法计算结果为标准，结果验证了基于改进拉丁超立方重要抽样法的概率潮流计算是有效的；在通过标准测试系统的验证后，将改进拉丁超立方重要采样法应用于宁夏实际电力系统中，并且对于风电场出力分布服从规则分布的情况，通过观察概率潮流计算的结果，分析风电场出力对于不同线路在空间位置上的影响。

关键词：大规模电网；概率潮流计算；拉丁超立方抽样；核密度估计

Abstract

Recently, with the adjustment of energy using structure, our country began to vigorously promote new energy power generation technology, and wind power is one of the new forms of energy. However, wind power is so related to environment that the farms cannot steadily control output. The rapid development of wind power brings uncertainty and impact to the stable operation of the power grid. The limitations of traditional deterministic flow calculation show step by step. So probabilistic power flow calculation is promoted and get development, considering the random factors that exist in the system.

Considering uncertainty factors, which exist in large-scale power grid unit output, the probabilistic power flow calculation method is used in this paper, which is combined with probability theory. Apply the methods of probability theory for large-scale power flow calculation results to get probabilistic analysis. By probabilistic power flow calculation, this paper can get the probability of the line bearing active power, further analysis to get the more limited probability values. The uncertainty of the system is to quantify to the line limited probability, facilitating the scheduling department to do reasonable scheduling.

For complex electric power data network, power system model is established firstly. Sort a large number of data, and weaken the influence of the transformer for the line through dealing with the transformer in the line. Using the MAO value transformation, constitute a whole power network that does not contain transformation ratio. According to the model of power system, find out the wind farm. Based on the principle of Latin hypercube sampling method, in order to consider the end of the random variable probability distribution characteristic factor, this paper points improvement Latin hypercube important sampling method. For including more than one wind power system, correlation between different wind farms needs to be considered. So, Copula function is introduced to make the correlation between wind farms model and digital. According to the improvement Latin hypercube important sampling method and Copula function, sample the wind farm output data, and this data can be applied to the established probabilistic power flow calculation of power system model, and the kernel density estimation method can fit the discrete probabilistic power flow calculation results. Test standard IEEE - 118 node system, and on the basis of Monte Carlo simulation results, the results prove that the probabilistic power flow calculation based on the improved Latin hypercube important sampling method is effective. Through the verification of standard test system, the improvement Latin hypercube important sampling method can be used in practical power system. And for the wind farms that obey the rules of distribution, by observing the results of the probabilistic power flow calculation, wind farm output has different influence on different line in space position.

Key words: large-scale power grid; the probabilistic power flow calculation; Latin hypercube sampling method; kernel density estimation.

目录

摘要 II

Abstract III

第一章 绪论 1

1.1 研究背景 1

1.2 概率潮流计算的研究现状 1

1.3 本文的主要研究内容 2

第二章 概率潮流计算原理 4

2.1 潮流计算 4

2.2 基于改进拉丁超立方重要抽样法的概率潮流计算 6

2.2.1 拉丁超立方抽样法 6

2.2.2 改进拉丁超立方重要抽样法原理 7

2.3 基于因子表共用的概率潮流计算加速方法 8

2.4 基于核密度估计的数据拟合方法 9

第三章 考虑新能源相关性的概率潮流计算 13

3.1 相关性描述 13

3.2 Copula函数基本概念 13

3.2.1几种常用Copula函数 13

3.2.2 Copula函数的选择 14

3.2.3 基于正态Copula函数考虑风电相关性的概率潮流计算方法 15

第四章 概率潮流计算结果及分析 18

4.1 标准IEEE-118算例分析 18

4.1.1 IEEE-118节点概率潮流计算 19

4.1.2 概率潮流计算精度分析 21

4.2 宁夏电力系统算例分析 28

4.2.1 参数处理 28

4.2.2 风电场有功出力数据抽样 30

4.2.3 潮流计算结果及分析 34

第五章 总结 39

致 谢 40

参考文献： 41

剩余内容已隐藏，请支付后下载全文，论文总字数：36756字