基于径向基函数神经网络的配电网参数估计  

Parameter estimation of distribution system based on radial basis neural network

作　　者：黄睿[1] 郭谋发[1] 陈永往 Huang Rui;Guo Moufa;Chen Yongwang(College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350116;State Grid Fujian Jinjiang County Electric Power Supply Co., Ltd, Jinjiang, Fujian 362200)

机构地区：[1]福州大学电气工程与自动化学院,福州350116 [2]国网福建晋江市供电有限公司,福建晋江362200

出　　处：《电气技术》2019年第4期42-46,共5页Electrical Engineering

基　　金：晋江市科技计划项目(2017C006)

摘　　要：配电网的线路参数会受温度、周围环境、集肤效应等因素的影响而发生变化,导致在进行配电网常规分析应用时降低计算结果的精度。本文提出一种基于径向基函数神经网络的三相不平衡配电网线路参数估计方法,通过对配电网三相不平衡线路等效模型的数学推导建立参数估计模型,利用径向基函数神经网络去拟合线路两端支路功率、节点电压和线路参数之间的非线性关系。对于训练完成的径向基函数神经网络,只需要知道线路两端测量值便可以获得准确的线路参数,可以有效地解决线路参数,估计数学模型中的病态矩阵问题。The effect of temperature, surrounding environment, skin effect and many other factors can contribute to the difference of distribution lines. A big error happened in distribution system analysis when distribution line parameter for calculation is much different from actual parameter data. To address parameter issue, this paper introduces a novel parameter estimation method based on radial basis function neural network (RBFNN) for three-phase unbalanced distribution line to map the nonlinear relation between measured values of branch power and node voltage and line parameter which is deduced from accurate equivalent distribution line three-phase unbalanced model. For a well trained RBFNN, accurate distribution line parameter can be obtained directly from measurement at two end of distribution line which can decrease the error caused by ill-conditioned matrix in parameter estimation.

关 键 词：径向基函数神经网络 三相不平衡配电网 参数估计 非线性关系 

分 类 号：TM744[电气工程—电力系统及自动化] TP183[自动化与计算机技术—控制理论与控制工程]