| 引用本文: | 周任军,任青青,闵雄帮,郑权国,吴献祥,王仰之.基于改进矩不确定分布鲁棒优化算法的多机调频比例决策[J].电力自动化设备,2019,39(6): |
| ZHOU Renjun,REN Qingqing,MIN Xiongbang,ZHENG Quanguo,WU Xianxiang,WANG Yangzhi.Frequency regulation ratio decision-making of multi-generator based on modified distributional robust optimization under moment uncertainty[J].Electric Power Automation Equipment,2019,39(6): |
|
| 摘要: |
| 针对单一平衡节点承担全网不平衡功率容易导致超出其调节范围的问题，考虑多台自动发电控制(AGC)机组参与调频，并将调频比例作为决策变量，建立调频比例决策模型。考虑到风电短期出力分布的矩具有不确定性的特点，将该模型变换为矩不确定分布鲁棒优化模型，并采用对偶原理将其转化为一个确定半定规划模型。对偶转换后模型中的半定约束矩阵的元素含有调频比例的二次方，常规半定规划算法无法对其进行求解。因此引入Benders方法对原问题进行分解，引入矩阵最小特征值方程对原半定约束进行松弛。仿真结果表明，基于改进广义Benders分解方法的半定规划模型具有良好的收敛性；发电调频总成本随着不确定集半径参数的增大而增大；与给定参与比例相比，将调频比例作为决策变量可获得更好的经济性和安全性。 |
| 关键词: 调频比例 自动发电控制 分布鲁棒优化 半定规划 广义Benders分解 矩阵最小特征值方程 |
| DOI：10.16081/j.issn.1006-6047.2019.06.019 |
| 分类号:TM73 |
| 基金项目:国家自然科学基金资助项目(51277016,71331001) |
|
| Frequency regulation ratio decision-making of multi-generator based on modified distributional robust optimization under moment uncertainty |
|
ZHOU Renjun1, REN Qingqing1, MIN Xiongbang2, ZHENG Quanguo3, WU Xianxiang1, WANG Yangzhi1
|
|
1.Hunan Province Collaborative Innovation Center of Clean Energy and Smart Grid, Changsha University of Science and Technology, Changsha 410004, China;2.Hydropower branch of Hunan Electric Power Co.,Ltd.,Changsha 410000, China;3.State Grid Changde Power Supply Company, Changde 415000, China
|
| Abstract: |
| When a single slack bus undertakes the unbalanced power of the whole network, it is easy to exceed its regulating range. Aiming at this problem, multiple AGC(Automatic Generation Control) units are considered to participate in the frequency regulation, and the decision-making model of frequency regulation ratio for each AGC unit is established, and the frequency regulation ratio is taken as a decision variable. Considering the uncertainty of the moment of short-term wind power output distribution, the proposed model is transformed into a distributional robust optimization model under moment uncertainty, which is further transformed into a determined semi-definite programming model based on the duality principle. After dual transformation, the element of semi-definite constraint matrix in the model contains the quadratic of frequency modulation ratio, which cannot be solved by the conventional semi-definite programming algorithm. Thus, Benders method is introduced to decompose the original problem by introducing the matrix minimum eigenvalue equation to slack the original semi-definite constraints. Simulative results show that the semi-definite programming model based on modified generalized Benders decomposition method has good convergence, the general frequency regulation costs increases with the increase of the radius parameter of uncertainty set, and compared with the case of fixed participation ratio, better economy and safety can be obtained by using frequency modulation ratio as the decision-making variable. |
| Key words: frequency regulation ratio AGC distributional robust optimization semi-definite programming generalized Benders decomposition matrix minimum eigenvalue equation |
