基于SOM神经网络的负荷特性分类  

The classification of load characteristics based on SOM neural network

作　　者：严金花 

机构地区：[1]华东交通大学电气与电子工程学院,江西南昌330013

出　　处：《大众科技》2013年第12期31-33,共3页Popular Science & Technology

摘　　要：负荷模型对电力系统仿真结果有重要影响,由于负荷特性的辨识是负荷建模的主要方面之一,故提高负荷模型的准确度就需要对负荷特性分类进行研究。文章在详细分析SOM自组织映射神经网络结构的基础上,采用了基于SOM神经网络的负荷分类方法,以负荷模型参数作为负荷动态特性分类特征向量,应用SOM神经网络对负荷特性进行分类,并对分类结果进行测试,结果表明该方法可有效地对负荷样本进行分类。Load models for power system simulation results have important implications, since the load characteristic identification is one of the main aspects of load modeling, thus improve the accuracy of load model requires study of load characteristic classification.Based on the detailed analysis of the SOM self-organizing map neural network structure,on the basis of the classification method based on SOM neural network load, load model parameters as the load dynamic characteristics classification characteristic vector and the application of SOM neural network to classify load characteristic,and classification results were tested and the results show that this method can effectively load the sample classification.

关 键 词：负荷特性分类 参数辨识 SOM神经网络算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]