中国电力 ›› 2013, Vol. 46 ›› Issue (7): 91-94.DOI: 10.11930/j.issn.1004-9649.2013.7.91.3

• 电网 • 上一篇    下一篇

基于日特征量相似日的PSO-SVM短期负荷预测

陈超1, 黄国勇1, 邵宗凯1, 2, 王晓东1, 2, 范玉刚1, 2   

  1. 1. 昆明理工大学 信息工程与自动化学院,云南 昆明 650500; 2. 云南省矿物管道输送工程技术研究中心,云南 昆明 650500
  • 收稿日期:2013-01-22 出版日期:2013-07-23 发布日期:2015-12-10
  • 作者简介:陈超(1989—),男,云南楚雄人,硕士研究生,从事电力系统负荷预测、智能电网方面的研究。
  • 基金资助:
    国家自然科学基金资助项目(51169007); 云南省科技计划项目(2010DH004,2011DA005,2011FZ036)

Short-Term Load Forecasting for Similar Days Based on PSO-SVM and Daily Feature Vector

CHEN Chao1, HUANG Guo-yong1, SHAO Zong-kai1, 2, WANG Xiao-dong1, 2, FAN Yu-gang1, 2   

  1. 1. School of Information Engineering & Automation, Kunming University of Science and Technology, Kunming 650500, China; 2. Yunnan Provincial Mineral Pipeline Technology Research Center, Kunming 650500, China
  • Received:2013-01-22 Online:2013-07-23 Published:2015-12-10

摘要: 通过引入人体舒适度指数,综合分析了气象因素对电力负荷的影响,并加入星期类型、日天气类型、日期差3个主要影响因素,构成了日特征向量,采用求取相似度的方法来选取相似日,利用相似日的日特征向量和负荷数据来建立PSO-SVM预测模型。经2001年EUNITE负荷预测竞赛的数据预测分析表明,该方法适应性较强,能够选取较合适的相似日,有较高的预测精度和推广能力。

关键词: 人体舒适度指数, 日特征向量, 相似日, 支持向量机, 短期负荷预测

Abstract: The human body amenity indicator was introduced to make a comprehensive analysis of the influence of the meteorological factors on power load, and three main influence factors, including week type, daily weather type and date difference, were added to constitute the daily feature vector. By using the method for calculating the similarity degree to select similar days, the PSO-SVM forecasting model was built up with the daily feature vector and load data of the similar days. An forecasting analysis of the EUNITE load prediction competition data in 2001 shows that this method has a good adaptability, and can easily select the suitable similar days, and has a high prediction accuracy and good potential for promotion.

Key words: human body amenity indicator, daily feature vector, similar days, SVM, short-term load forecast

中图分类号: