基于分布式电源功率预测的虚拟电厂优化调度模型研究

发布时间：2018-03-04 11:13

本文选题：分布式电源　切入点：虚拟电厂　出处：《华北电力大学(北京)》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,分布式电源技术作为能源互联网以及智能电网技术中的重要分支正在迅猛发展。与此同时,分布式电源设备大规模接入电网后所引起的系统不稳定性正给传统电网带来巨大的挑战。虚拟电厂(Virtual Power Plant,VPP)作为一种将分布式电源有效整合的方式应运而生,其在运行控制层面更趋近于并网型的微电网[1],而就整体性而言则呈现出了较为明显的外部性效果,尤其在经济效益方面。本文对分布式电源接入条件下虚拟电厂的联合优化调度问题进行了研究。首先,就虚拟电厂中分布式电源的运行特性进行了分析与研究,得到风机设备输出功率与风速之间关系的数学模型;其次分别应用时间序列预测算法中的自回归滑动平均(Auto-Regressive and Moving Average,ARMA)模型与Bayes网络对风机系统的输出功率进行了预测,得到在不同自然条件下风速—功率拟合误差概率分布并通过风速与风机出力的关系得到了风机的输出功率预测值,通过两种算法的对比分析,得到基于风速—功率拟合的风机出力预测结果;再次,建立虚拟电厂多类型分布式能源综合优化调度模型,选取优化调度中经济效益最优作为目标函数,约束条件的选择考虑了虚拟电厂中可调控的负荷量,并将其作为联合调度优化的可控变量得到最优输出值;最后,在优化算法方面本文在人群搜索者算法(Seeker Optimization Algorithm,SOA)的基础上引进了禁忌搜索技术(Tabu Search,TS),得到改进的TS-SOA算法对某15节点虚拟电厂系统进行了算例仿真与分析,得到了优化调度的最优输出功率以及虚拟电厂优化调度的最优经济效益,同时对标准SOA算法与TS-SOA两种优化算法进行算法的收敛性与优化特性进行了对比分析,验证了TS-SOA算法的正确性与适用性。
[Abstract]:In recent years, distributed power generation technology as an important branch of the energy Internet and smart grid technology is developing rapidly. The system instability caused by the large-scale access of distributed generation equipment to the power network is bringing great challenges to the traditional power network. Virtual Power Plan VPP (Virtual Power Plant) has emerged as a way to effectively integrate the distributed power supply (DG). It is closer to grid-connected microgrid [1] at the operation control level, but it has obvious externality in terms of integrity. Especially in the aspect of economic benefit, this paper studies the joint optimal dispatching problem of virtual power plant under the condition of distributed power access. Firstly, the operation characteristics of distributed power generation in virtual power plant are analyzed and studied. The mathematical model of the relationship between wind speed and output power of fan equipment is obtained. Secondly, the autoregressive and Moving average ARMA-model and Bayes network are used to predict the output power of fan system, respectively. The probability distribution of the wind speed power fitting error is obtained under different natural conditions, and the output power prediction value of the fan is obtained by the relationship between the wind speed and the output force of the fan. The comparison and analysis of the two algorithms are carried out. The prediction results of fan output force based on wind speed and power fitting are obtained. Thirdly, the multi-type distributed energy comprehensive optimal scheduling model of virtual power plant is established, and the optimal economic benefit is selected as the objective function. The selection of constraints takes into account the controllable load in the virtual power plant and takes it as the controllable variable of the joint scheduling optimization to obtain the optimal output value. In the aspect of optimization algorithm, the Tabu search algorithm is introduced on the basis of the crowd searcher Optimization algorithm. The improved TS-SOA algorithm is used to simulate and analyze a 15-bus virtual power plant system. The optimal output power of optimal dispatching and the optimal economic benefit of optimal dispatching of virtual power plant are obtained. The convergence and optimization characteristics of standard SOA algorithm and TS-SOA algorithm are compared and analyzed. The correctness and applicability of TS-SOA algorithm are verified.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【参考文献】