基于分布鲁棒模型预测控制的微电网多时间尺度优化调度  

Multi-time scale distributed robust optimal scheduling of microgrid based on model predictive control

作　　者：李嘉伟 巨云涛 张璐[1] 刘文武 王杰 LI Jiawei;JU Yuntao;ZHANG Lu;LIU Wenwu;WANG Jie(College of Information and Electrical Engineering,China Agricultural University,Beijing 100083,China;School of Electrical and Control Engineering,North China University of Technology,Beijing 100144,China;State Grid Hangzhou Power Supply Company of Zhejiang Electric Power Co.,Ltd.,Hangzhou 311400,China)

机构地区：[1]中国农业大学信息与电气工程学院,北京100083 [2]北方工业大学电气与控制工程学院,北京100144 [3]国网浙江省电力有限公司杭州供电公司,浙江杭州311400

出　　处：《电力工程技术》2024年第4期45-55,共11页Electric Power Engineering Technology

基　　金：国家自然科学基金资助项目(52177125)。

摘　　要：源荷多元不确定性给“源荷储”一体化微电网优化调度带来了诸多挑战,但传统方案存在优化模型过于片面极端和时间尺度单一导致调度不合理的问题,无法兼顾可靠性和经济性,并且难以协调不确定性分析方法与不同时间尺度之间的配合关系。文中基于数据驱动的多离散场景分布鲁棒方法,提出一种微电网两阶段分布鲁棒日前优化调度模型,使用列和约束生成算法进行求解。结合改进分布鲁棒优化的不确定方法、多时间尺度调度策略和模型预测控制理论,通过逐级细化调度时间尺度和减小预测周期的长度来提高精度,以最小化发电成本以及调节成本等为目标建立日前-日内多时间尺度滚动优化调度模型,较大程度上抵抗系统不确定性因素的影响。结合算例仿真分析,进一步说明了所提模型能够有效消纳新能源、降低运行成本同时兼顾安全性和经济性。The multi-uncertainty of source and load poses significant challenges to the optimal scheduling of′source-load-storage′integrated microgrid.However,a limitation of the traditional optimization model is its one-sidedness and use of a single time scale,which can result in suboptimal scheduling outcomes.Striking a balance between reliability and economy presents a considerable obstacle,as does coordinating the relationship between uncertainty analysis methods and varying time scales.Based on the data-driven multi-discrete scene distribution robust method,a two-stage distributed robust day-ahead optimal scheduling model of microgrid is proposed,which is solved by column and constraint generation algorithm.By combining the improved distributed robust optimization uncertainty method with a multi-time scale scheduling strategy and model predictive control theory,the accuracy of the scheduling can be enhanced.This is achieved through the gradual refinement of the scheduling time scale and reduction of the prediction period length.The day-ahead-day multi-time scale rolling optimization scheduling model is established to minimize the generation cost and adjustment cost,while also exhibiting a high degree of resilience to system uncertainties.Combined with the simulation analysis of the example,the proposed model has demonstrated advantages in incorporating new energy sources,reducing operating costs,and balancing considerations of safety and economy.

关 键 词：微电网 模型预测控制 多时间尺度优化调度 分布鲁棒优化 多场景技术 数据驱动 

分 类 号：TM73[电气工程—电力系统及自动化]