摘要
为了更准确地预测风速,首先利用可见图将风速时间序列映射到有向加权网络中,通过计算邻居时间节点的相似度并结合最短路径的方法确定网络中节点的相似度矩阵。然后在有向加权网络节点相似度分析的基础上结合相邻预测法及线性逼近法进行预测。在预测实验中,通过和其他模型的误差比较证明其适用性和可预测性,说明该方法能更准确地预测风速,可为电力系统的运行提供借鉴作用。
Wind speed prediction is widely concerned in wind power industry because of its wide application in wind power generation.In order to predict wind speed more accurately,a new method is proposed in this paper.First,the visibility graph is used to map the wind speed series to a directed weighted network,the similarity matrix of nodes in the network is determined by calculating the similarity of neighbor time nodes and combining the shortest path method,then on the basis of the node similarity analysis of directed weighted network,the adjacent prediction method and linear approximation method are combined to predict.In the prediction experiment,the applicability and predictability of this method are illustrated by comparing the errors with other models.This method can predict wind speed more accurately and provide reference for power system operation.
作者
张董极
肖琴
Zhang Dongji;Xiao Qin(College of Sciences,Shanghai Institute of Technology,Shanghai 201418,China)
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2023年第3期90-96,共7页
Acta Energiae Solaris Sinica
基金
上海市自然科学基金(16ZR1447200)
国家青年科学基金(11701379)。
关键词
风速
复杂网络
时间序列分析
可见图
最短路径
节点相似性
wind speed
complex network
time series analysis
visibility graph
shortest path
node similarity
