华南理工大学学报(自然科学版) ›› 2024, Vol. 52 ›› Issue (9): 1-11.doi: 10.12141/j.issn.1000-565X.240117

• 能源、动力与电气工程 •    下一篇

面向光伏集群扩展规划的两阶段分布鲁棒优化

曾君1(), 王天伦1, 黄智鹏1,3, 张轩1,2   

  1. 1.华南理工大学 电力学院, 广东 广州 510640
    2.中国南方电网有限责任公司 北京分公司, 北京 100020
    3.广东电网有限责任公司 电力科学研究院, 广东 广州 510080
  • 收稿日期:2024-03-13 出版日期:2024-09-25 发布日期:2024-04-12
  • 作者简介:曾君(1979—),女,博士,教授,主要从事微电网能量管理及优化、可再生能源发电系统中的电力电子及控制技术研究。E-mail: junzeng@ scut.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(62173148);广东省自然科学基金资助项目(2022A1515010150);广东省基础与应用基础研究基金资助项目(2022A1515240026)

Two-Stage Distributional Robust Optimization for the Expansion Planning of Photovoltaic Cluster

ZENG Jun1(), WANG Tianlun1, HUANG Zhipeng1,3, ZHANG Xuan1,2   

  1. 1.School of Electric Power Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
    2.China Southern Power Gird Beijing Company, Beijing 100020, China
    3.Electric Power Research Institute of Guangdong Power Grid Co. , Ltd. , Guangzhou 510080, Guangdong, China
  • Received:2024-03-13 Online:2024-09-25 Published:2024-04-12
  • About author:曾君(1979—),女,博士,教授,主要从事微电网能量管理及优化、可再生能源发电系统中的电力电子及控制技术研究。E-mail: junzeng@ scut.edu.cn
  • Supported by:
    the National Natural Science Foundation of China(62173148);the Natural Science Foundation of Guangdong Province(2022A1515010150);the Basic and Applied Basic Research Foundation of Guangdong Province(2022A1515240026)

摘要:

随着双碳目标的深入推进,可再生能源的渗透率逐年攀升,其消纳问题备受关注。分布式可再生能源集群是一种消纳可再生能源的新模式,在规划时需要着重考虑源荷不确定性的影响。以分布式光伏集群新增光伏并网规划为背景,计及源荷的不确定性,提出了一种基于两阶段分布鲁棒优化的光伏集群扩展规划方法。考虑规划阶段和运行阶段侧重的不同,建立了以年等值成本最小为目标、考虑机组出力约束和电网承载能力的两阶段分布鲁棒优化模型。为了提高计算效率,结合K-means聚类和极限场景法对区域分布式光伏和随机负荷的历史数据进行削减和修正,并基于修正后的场景集构造了基于Wasserstein距离的概率分布模糊集。接着采用列和约束生成算法将建立的两阶段分布鲁棒优化模型分解为主问题和子问题,通过主问题和子问题的迭代进行求解,进一步提高了求解效率。其中,为实现子问题的求解,引入了拉格朗日对偶法将子问题转化为确定性优化问题。最后,以某分布式光伏集群为例开展算例分析,结果表明:所提的基于两阶段分布鲁棒优化的光伏集群扩展规划方法可以有效协调规划运行方案的经济性和鲁棒性,根据历史场景集的大小和可靠程度可以灵活调整模型控制参数以满足多种工程应用场景下对可靠性和经济性的不同需求。

关键词: 光伏集群, 极限场景, Wasserstein距离, 分布鲁棒优划, 列与约束生成算法

Abstract:

With the deepening of the two-carbon target, the penetration rate of renewable energy is increasing year by year, and its consumption problem has attracted much attention. Distributed renewable energy cluster is a new mode of accommodating renewable energy. It is necessary to consider the influence of source-load uncertainty in planning and operation. In this paper, based on the new photovoltaic grid-connected planning of distributed photovoltaic cluster, considering the uncertainty of source and load, a distributed photovoltaic cluster expansion planning method based on two-stage robust optimization was proposed. Considering the difference between the planning stage and the operation stage, it established a two-stage distributed robust optimization model, which takes the minimum annual equivalent cost as the objective and considers the unit output constraint and the power grid carrying capacity. In order to improve the computational efficiency, the historical data of regional distributed renewable energy and random load were reduced and modified by combining K-means clustering with extreme scenario method. Based on the modified scenario set, a probability distribution fuzzy set based on Wasserstein distance was constructed. The column and constraint generation algorithm was used to decompose the two-stage distributed robust optimization model into the main problem and the sub-problem. The main problem and the sub-problem were solved by iteration, which further improves the efficiency of the solution. In order to solve the sub-problem, Lagrange duality was introduced to transform the sub-problem into a deterministic optimization problem. Finally, a distributed photovoltaic cluster was taken as an example to carry out an example analysis. The results show that the proposed two-stage distributional robust optimization method for distributed photovoltaic clusters can coordinate the economy and robustness of the planning operation scheme. Model control parameters can be flexibly adjusted according to the size and reliability of historical scene sets to meet the different requirements of reliability and economy in various engineering application scenarios.

Key words: photovoltaic cluster, extreme scenario, Wasserstein distance, distributional robust optimization, column and constraints generation algorithm

中图分类号: