摘要
近年来,我国电力负荷峰值增长速度较快,尤其是华东地区,负荷峰值不断刷新。文章研究负荷峰值特性分布,基于向前逐步选择正则化提出两阶段法进行模型因素选择,并以实例验证选择出6个最佳影响因素组合。在两阶段法模型因素选择研究基础上,结合k-means聚类降低计算工作量,设计了基于贝叶斯网络的电力负荷峰值预测模型和分类预测算法,并以上海市浦东区为例进行验证,预测结果精度较高,验证了该方法的可行性及有效性。
In recent years,power load peak in China is growing rapidly,especially in East of China.In view of this,the load peak characteristic distribution is studied.Based on the forward stepwise selection regularization,a two-stage method is proposed to select the model factors,and the six best influencing factors are selected by example verification.Combining with k-means clustering to reduce the computational workload,a Bayesian network-based power load peak forecasting model and a classification forecasting algorithm are designed,based on the research of two-stage model factor selection.Taking Shanghai Pudong District as an example,the accuracy of the forecasting results is high,and the feasibility and effectiveness of the method are verified.
作者
王文秀
田世明
王泽忠
谢伟
卜凡鹏
田英杰
苏运
WANG Wenxiu;TIAN Shiming;WANG Zezhong;XIE Wei;BU Fanpeng;TIAN Yingjie;SU Yun(China Electric Power Research Institute,Beijing 100192,China;North China Electric Power University,Beijing 102206,China;State Grid Shanghai Electric Power Company,Shanghai 200122,China)
出处
《供用电》
2019年第7期57-64,共8页
Distribution & Utilization
基金
国家重点研发计划项目(2016YFB0901100)
国家电网公司科技项目(52094017002U)~~
