多区域电力系统实时分散鲁棒调度的求解方法  

A Real-time Decentralized Robust Dispatch Approach of Multi-area Power System

作　　者：王晨曦 朱建全[1] 夏云睿 郭野 莫勰旻 罗添允 WANG Chcnxi;ZHU Jianquan;XIA Yunrui;GUO Ye;MO Xicmin;LUO Tianyun(School of Electric Power,South China University of Technology,Guangzhou 510640,Guangdong Province,China)

机构地区：[1]华南理工大学电力学院

出　　处：《电网技术》2019年第12期4478-4485,共8页Power System Technology

基　　金：国家自然科学基金项目(51307064);中央高校基本科研业务费专项资金资助项目(2017MS024);广东省自然科学基金项目(2018A0303131001)~~

摘　　要：针对风电的不确定性以及各个区域的信息私密性和调度独立性要求,提出一种多区域电力系统分散鲁棒调度问题的求解方法。首先,将风电出力的不确定转化为预测区间的边界信息,并利用机组的自动发电控制应对风电的出力波动,基于仿射可调节鲁棒优化技术建立集中式的鲁棒调度模型。其次,基于动态规划理论对所提的集中式模型进行解耦,通过值函数描述某一区域的决策对其他区域的影响情况,从而得到各个区域的鲁棒调度子模型,并利用近似动态规划对其进行迭代求解。最后,利用1个2区域6节点系统和实际4区域2298节点系统进行仿真测试,所得结果验证了文中方法能够准确、快速、稳定地求解电力系统分散鲁棒优化调度问题。In order to deal with uncertainty of wind power and information privacy and dispatch independence of each area, a decentralized robust optimization method is proposed for solving the real-time decentralized robust dispatch problem of multi-area power system. Firstly, the uncertainty of wind power is converted into the boundary information of prediction interval. And an automatic generation control unit is used to cope with wind power fluctuation. Then a centralized robust dispatch model is established based on affine adjustable robust optimization technique. Secondly, the proposed centralized model is decoupled according to dynamic programming theory. The influence of one area’s decision on other areas is described with value function, and then the robust dispatch sub-model for each area is obtained. The sub-models are solved iteratively using approximate dynamic programming. Finally, tests are carried out in a 2-area 6-node system and an actual 4-area 2298-node system. Results indicate accuracy, rapidity, and stability of the proposed approach for solving the decentralized robust optimal dispatch problems.

关 键 词：多区域电力系统 风电 实时调度 分散鲁棒优化 近似动态规划 

分 类 号：TM721[电气工程—电力系统及自动化]