摘要
将数据挖掘的思想应用到对配变历史负荷数据的分析中,通过聚类分析将预处理过的负荷数据进行相关性分类。在分析负荷特性的基础上,提出了基于粒子滤波的负荷预测方法。应用该方法并结合负荷相关性分类对每类负荷分别进行预测,加权处理后得到配变的负荷预测数据。经算例验证该方法有较高的负荷预测准确率。
In this paper,the idea of data mining is applied to the analysis of historical load data of distribution transformers,and the pre-processed load data is classified by cluster analysis according to the correlation.Based on the analysis of load characteristics,a load forecasting method based on particle filter is proposed.By combing the proposed method with the load correlation classification method,the prediction of each type of load is carried out,and the load forecasting data of the distribution transformers is obtained after weighting process.The results show that this method has higher load forecasting accuracy.
作者
王倩
王麒翔
张建宾
张韦维
牛雨
王磊
WANG Qian;WANG Qixiang;ZHANG Jianbin;ZHANG Weiwei;NIU Yu;WANG Lei(State Grid HAEPC Electric Power Research Institute,Zhengzhou 450052,China;School of Electrical and Electronic Engineering,North China Electric Power University,Beijing 102206,China)
出处
《电力科学与工程》
2018年第3期35-39,共5页
Electric Power Science and Engineering
关键词
数据挖掘
聚类分析
粒子滤波
负荷预测
配电变压器
data mining
cluster analysis
particle filter
load forecasting
distribution transformers