arrow
返回

Wind speed estimation using multilayer perceptron

delete2014-05-01
delete79
PRE
AI
R
R. Velo *
P
Paz López
F
F. Maseda
DOI:10.1016/j.enconman.2014.02.017delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Wind speed knowledge is prerequisite in the siting of wind turbines. In consequence the wind energy use requires meticulous and specified knowledge of the wind characteristics at a location. This paper presents a method for determining the annual average wind speed at a complex terrain site by using neural networks, when only short term data are available for that site. This information is useful for preliminary calculations of the wind resource at a remote area having only a short time period of wind measurements measurement in a site. Artificial neural networks are useful for implementing non-linear process variables over time, and therefore are a useful tool for estimating the wind speed. The neural network used is multilayer perceptron with three layers and the supervised learning algorithm used is backpropagation. The inputs used in the neural network were wind speed and direction data from a single station, and the training patterns used correspond to sixty days data. The results obtained by simulating. the annual average wind speed at the selected site based on data from nearby stations with correlation coefficients above 0.5 were satisfactory, compared with actual values. Reliable estimations were obtained, with errors below 6%. (C) 2014 Elsevier Ltd. All rights reserved.
Keyword:
Artificial neural network
Forecasting
Multilayer perceptron
Wind speed

期刊

Energy Conversion and Management 封面图
Energy Conversion and Management
IF:
10.9
论文数:
2.0W
被引数:
11.3W

机构

U
Universidade de Santiago de Compostela
学者数:
1.5W
论文数: 1.3W
被引数: 1.4W
引用论文

引用论文

err分享
err收藏
Long-term wind speed and power forecasting using local recurrent neural network models
err2006-03-01
err448
PREAI
errBarbounis, TG; Theocharis, JB; Alexiadis, MC; Dokopoulos, PS
err分享
err收藏
Forecasting wind with neural networks
err2003-01-01
err218
PREAI
errMore, A; Deo, MC
err分享
err收藏
学者 查看更多内容