基于神经网络的电力系统负荷特性辨识

负荷建模在电力系统分析中起着十分重要的作用。参数辨识是负荷建模的关键,好的辨识方法能够在最短的时间内找出最优的辨识结果,提高建模的效率。首先介绍了静态负荷模型和动态负荷模型,其次介绍了神经网络,最后基于神经网络对电力系统负荷特性辨识。应用线性BP（LBP）网络的参数辨识方法,分别对静态负荷模型（幂函数模型、多项式模型）和动态负荷模型（差分方程模型）的参数进行辨识。通过现场的实测数据辨识了模型的参数,并验证了模型的有效性。

Identification of Load Characteristic of the Electric Power System Based on a Neural Network

The load modeling plays a very important role in the power system analysis.Parameter identification is the key to load modeling,a good way for identifying parameter can obtain a best result in the shortest time,which can greatly enhance the efficiency of the modeling.This paper introduces the static load model and dynamic load model,then introduces neural network,at last based on neural network we identified power system load characteristic.The parameters of static（exponent function,polynomial） and dynamic（differential equations） models can be identified by the method which is used in the Linear Back-Propagation network.The model parameters can be identified by the data of the field,and verify the effectiveness of the model.