Very-short-term prediction of wind speed based on chaos phase space reconstruction and NWP

Shuang Gao, Lei Dong, Xiaozhong Liao, Yang Gao

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Citations (Scopus)

Abstract

Wind speed forecasting has already been a vital part of wind farm. The operational planning of power grids are with the aim of reducing greenhouse gas emissions. This paper presents a very short term prediction scheme that combined chaos phase space reconstruction with numerical weather prediction (NWP) method. Historical wind speed data, which are reconstructed as phase space vectors, are taken as the first input part of hybrid prediction model; the NWP data at the prediction time are taken as the second input part. General regression neural network (GRNN) is used to map the non-linear relationship in the study and wind speed at the height of turbine hub is derived from neural network model. The data from a wind farm are used to verify the proposed method. The prediction results are presented and compared to the chaos GRNN model, NWP GRNN model and persistence model. The results show that the method presented in this paper has an improved prediction precision.

Original languageEnglish
Title of host publicationProceedings of the 32nd Chinese Control Conference, CCC 2013
PublisherIEEE Computer Society
Pages8863-8867
Number of pages5
ISBN (Print)9789881563835
Publication statusPublished - 18 Oct 2013
Event32nd Chinese Control Conference, CCC 2013 - Xi'an, China
Duration: 26 Jul 201328 Jul 2013

Publication series

NameChinese Control Conference, CCC
ISSN (Print)1934-1768
ISSN (Electronic)2161-2927

Conference

Conference32nd Chinese Control Conference, CCC 2013
Country/TerritoryChina
CityXi'an
Period26/07/1328/07/13

Keywords

  • General Regression Neural Network
  • Numerical Weather Prediction
  • Wind speed prediction
  • phase space reconstruction

Fingerprint

Dive into the research topics of 'Very-short-term prediction of wind speed based on chaos phase space reconstruction and NWP'. Together they form a unique fingerprint.

Cite this