基于伪量测型协调变量的分布式状态估计算法

发布时间：2018-05-10 06:29

本文选题：状态估计 + Newton-GMRES算法　；参考：《中国电机工程学报》2014年19期

【摘要】：电力系统状态估计在实际工程中通常由各控制中心采用区域状态估计模式独立完成。考虑到在电网分层分区的控制管理模式下,多样化的电网信息本身就是按照分层分区的分布式模式采集的,分布式状态估计无疑是更加适应该体系的状态估计运行模式。利用集中式优化问题的一阶KKT条件分解技术,实现状态估计任务的分解。为保证在边界上相邻分区状态估计结果的一致性,引入描述边界协调项的等式约束。另外,将上述边界等式约束处理成零功率注入伪量测,以简化分区之间的数据交换,便于重用已有状态估计器。为了尽最大可能维持各个分区计算的独立性,该文给出了协调侧Lagrangian函数梯度的两步式计算方法,并结合Newton-GMRES算法完成求解。最后,通过IEEE标准系统的测试,验证了所提方法的有效性。
[Abstract]:The state estimation of power system is usually carried out independently by each control center using the regional state estimation model in practical engineering. Considering that under the control and management mode of the power system, the diversified power network information is collected according to the distributed mode of the hierarchical and partitioned power system, and the distributed state estimation is undoubtedly more suitable for the operation mode of the system. The state estimation task is decomposed by using the first order KKT conditional decomposition technique for centralized optimization problems. In order to ensure the consistency of the state estimation results of adjacent partition on the boundary, an equality constraint is introduced to describe the boundary coordination term. In addition, the above boundary equality constraints are treated as zero power injection pseudo measurements to simplify the data exchange between partitions and facilitate the reuse of existing state estimators. In order to maintain the independence of each partition as far as possible, this paper presents a two-step method for calculating the gradient of the Lagrangian function on the coordination side, and combines the Newton-GMRES algorithm to solve the problem. Finally, the effectiveness of the proposed method is verified by testing the IEEE standard system.
【作者单位】：电力系统国家重点实验室(清华大学电机系);
【基金】：国家863高技术基金项目(2012AA050201)~~
【分类号】：TM744

【参考文献】