期刊论文

Based on the ART2wNF Neural Network and K-means Algorithm of Photovoltaic Power Generation Power Curve Clustering

中文

电力学报

1005-6548

2014

29

2

119-123+155

10.13357/j.cnki.jep.002324

本校为第一机构

为弥补传统单一聚类方法存在的各种缺陷,提出一种将K-means算法和ART2wNF网络相结合的聚类方法来对数据进行处理,以提高数据聚类的精确性和可靠性.结合Ashland光伏电站(5 kW) 2001全年的实测数据,详细说明了数据在该方法下的聚类流程,并绘制出聚类效果图,完成了对各类数据的对比分析,验证了该聚类方法的精度和可靠性.

To make up for the defects existing in the tradi- tional single clustering methods, this paper creatively combines classical ART2wNF neural network clustering algorithm with K-means algorithm and form a new ＂double standard, double clustering＂ comprehensive clustering method. It incorporates Ashland photovoltaic power station （5 kW） measured data of 2001 to get detailed instructions on how the data according to the procedures followed in the K-means algorithm and the ART2wNF neural network to complete clustering, draw out the clustering res...