摘要
针对电网线损的理论计算值与实际值之间的差别问题,提出利用基于直接神经动态规划的电网状态估计结果来计算线损。首先采用模糊聚类算法进行网络拓扑辨识,根据聚类向量弥补量测向量的不足,避免了不良数据的影响,得到正确的系统网络结构;然后对自适应动态规划算法进行扩维改进,建立了基于直接神经动态规划的电网状态估计模型,利用该模型的状态估计结果进行理论线损计算,得到逼近电网真实情况的线损数据。仿真结果证明了本文算法的可靠性和实用性。
To solve the difference between the theoretical value of power network line loss and its practical value, it is proposed to calculate line loss by use of power network state estimation based on direct neural dynamic programming. Firstly, network topology identification is performed by fuzzy clustering, according to clustering vector the insufficient measured vectors are made up to avoid the affect of bad data and correct system network architecture is achieved; then the dimension extension and improvement on adaptive dynamic programming algorithm are carried out to build power network state estimation model based on direct neural dynamic programming and by use of the results of state estimation the theoretical line loss is calculated which is close to the true value of power network line loss. The reliability and practicality of the proposed method are verified by simulation results.
出处
《电网技术》
EI
CSCD
北大核心
2008年第23期50-55,共6页
Power System Technology
关键词
状态估计
直接神经动态规划
网络拓扑辨识
模糊聚类
线损
state estimation： direct neural dynamic programming： network topology identification
fttzzy clustering
line loss
