| 引用本文: | 巨云涛,张紫枫,刘文武,等.考虑限幅约束的多区域互联电网全分布式潮流算法[J].电力系统保护与控制,2022,50(7):58-68.[点击复制] |
| JU Yuntao,ZHANG Zifeng,LIU Wenwu,et al.A fully distributed power flow algorithm for multi-region interconnected power grids considering limiting constraints[J].Power System Protection and Control,2022,50(7):58-68[点击复制] |
|
| 摘要: |
| 多区域互联电网中有大量限幅非光滑约束，增加了潮流模型的非凸性。针对传统基于启发式规则处理限幅的分布式潮流算法在计算过程中容易出现收敛性问题，提出了一种能鲁棒性处理非光滑约束的互联电网全分布式潮流计算方法。首先根据分层分区调度模式对电网进行分区，并将模型中非光滑约束进行光滑处理，然后基于具有二阶收敛速度的双层交替方向非精确牛顿法，将潮流问题转化为求解最优步长增量的问题。基于零空间法(Null-Space) 对状态变量的系数矩阵进行降维处理，采用共轭梯度(Conjugate Gradient, CG)算法更新每个分区的对偶乘子，乘子更新过程中利用二阶信息提高了算法的收敛性。多区域间进行分布式计算时无需协调层参与，仅需通信少量边界信息，因此通信负担轻。最后，以30节点和182节点系统为测试算例，验证了所提方法在设置恶劣初值和处理非光滑约束时具有较高的精确性和较好的收敛性。 |
| 关键词: 互联电网分区潮流 分布式计算 非光滑限幅约束 隐私保护 |
| DOI：DOI: 10.19783/j.cnki.pspc.211128 |
| 投稿时间：2021-08-18修订日期：2021-12-16 |
| 基金项目:国家自然科学基金项目资助(52177125，51707196) |
|
| A fully distributed power flow algorithm for multi-region interconnected power grids considering limiting constraints |
| JU Yuntao,ZHANG Zifeng,LIU Wenwu,CHEN Xi |
| (College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China) |
| Abstract: |
| There are a large number of limiting non-smooth constraints in multi-region interconnected power grids. These constraints increase the non-convexity of the power flow model. Given that the traditional distributed power flow algorithm, based on a heuristic rule processing limit, is prone to convergence problems in the calculation process, a fully distributed power flow calculation method for an interconnected power grid that can robustly handle non-smooth constraints is proposed. First, the grid is partitioned according to the hierarchical and partitioned dispatch mode, and the non-smooth constraints in the model are smoothly processed. Then, based on the bi-level alternate direction inexact Newton method with a second-order convergence rate, the power flow problem is transformed into a problem of solving the optimal step increment. Based on the null space method, the coefficient matrix of the state variable is reduced in dimension, and the conjugate gradient algorithm is used to update the dual multiplier of each partition, and the second-order information is used in the multiplier update process to improve the convergence of the algorithm. The coordination layer does not need to be involved in distributed computing between multiple regions, and only a small amount of boundary information needs to be communicated, so the communication burden is light. Finally, the 30-bus system and 182-bus system are used as test examples to verify that the proposed method has higher accuracy and better convergence when setting bad initial values and dealing with non-smooth constraints.
This work is supported by the National Natural Science Foundation of China (No. 52177125 and No. 51707196). |
| Key words: interconnected power grids partition power flow distributed computing non-smooth limiting constraints privacy protection |
