基于RBF循环神经网络的电力系统负荷建模  

Power System Load Modeling Based on Recurrent RBF Neural Network

作　　者：陈幸琼[1] 邓长虹[1] 潘章达[1] 董超[1] 

机构地区：[1]武汉大学电气工程学院,湖北省武汉市430072

出　　处：《电网技术》2007年第17期56-59,共4页Power System Technology

基　　金：湖北省自然科学基金资助项目(2006ABA214)~~

摘　　要：针对负荷模型难以精确建立的问题以及负荷非线性动态仿真的复杂性,提出了一种基于径向基循环神经网络的负荷建模方法。将循环神经网络和径向基网络相结合,利用循环神经网络对时间序列的学习能力和径向基网络具有结构自适应确定、快速收敛的优点,建立新的电力系统综合负荷模型。典型新英格兰测试系统的仿真证明了该模型对电力系统负荷模型辨识的有效性和准确性。Due to the existing problems that accurate load model is difficult to built and the complexity of dynamic simulation of nonlinear loads, a load modeling method based on recurrent radial basic function （RBF） neural network is proposed. In order to utilize the tinie series earning ability of recurrent neural network （RNN） and the self-structuring and fast convergence of RBF, the proposed method integrates RNN with RBF network to build new composite load model of power system. Simulation results of New England 10-machine 39-bus test system prove that the proposed load modeling method is valid and accurate in power system load model identification.

关 键 词：负荷建模 循环神经网络 RBF网络 模型辨识 

分 类 号：TM715[电气工程—电力系统及自动化]