摘要
同步发电机动态状态估计对电力系统的分析与控制具有重要意义。然而,基于卡尔曼滤波的动态状态估计算法计算量大,易受不良数据干扰。建立考虑励磁和调速的同步发电系统模型,进行初值优化。使用隐式梯形积分方法对机电暂态过程离散化,针对CKF算法特点,提出雅可比矩阵重用策略,提高算法效率,满足实时性要求。通过假设检验,辨识输入-输出不良数据,并进行修正,增强鲁棒性。IEEE9节点系统和华东电网某台实际运行机组算例证明了算法的有效性。
Dynamic state estimation for synchronous machines is of great significance for analysis and control of power system. However, the Kalman filter-based dynamic estimation algorithm is computationally intensive and susceptible to be disturbed by bad measurement data. In this paper, a synchronous power generation system model considering excitation and speed governor is established, and an initial value optimization method is proposed. The implicit trapezoidal integral method is used to discretize the electromechanical transient process. According to the characteristics of Cubature Kalman Filter (CKF) algorithm, a Jacobian matrix reuse strategy is put forward to improve efficiency and meet real-time requirements. Through hypothesis test, the input-output bad data is identified, and a correction method is put forward to improve robustness of the algorithm. Finally, case studies in IEEE 9-bus system and an actual unit in East China Power Grid are performed, verifying effectiveness of the proposed algorithm.
作者
刘朋成
项中明
江全元
耿光超
孙维真
熊鸿韬
LIU Pengcheng;XIANG Zhongming;JIANG Quanyuan;GENG Guangchao;SUN Weizhen;XIONG Hongtao(College of Electrical Engineering, Zhejiang University, Hangzhou 310027, Zhejiang Province, China;State Grid Zhejiang Province Electric Power Company Limited, Hangzhou 330106, Zhejiang Province, China;Electric Power Research Institute of State Grid Zhejiang Province Electric Power Company,Hangzhou 310014, Zhejiang Province, China)
出处
《电网技术》
EI
CSCD
北大核心
2019年第8期2860-2867,共8页
Power System Technology
基金
国家重点研发计划项目(2017YFB0902000)~~
关键词
实时动态状态估计
隐式梯形积分
鲁棒性
不良数据校正
real-time dynamic state estimation
implicit trapezoidal integral
robustness
bad data identification and correction