Wind power forecasts using Gaussian processes and numerical weather prediction

Niya Chen, Zheng Qian, Ian T. Nabney, Xiaofeng Meng

Research output: Contribution to journalArticle

Abstract

Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate wind power forecasts are necessary for the safe and economic use of wind energy. In this paper, we investigated a combination of numeric and probabilistic models: a Gaussian process (GP) combined with a numerical weather prediction (NWP) model was applied to wind-power forecasting up to one day ahead. First, the wind-speed data from NWP was corrected by a GP, then, as there is always a defined limit on power generated in a wind turbine due to the turbine controlling strategy, wind power forecasts were realized by modeling the relationship between the corrected wind speed and power output using a censored GP. To validate the proposed approach, three real-world datasets were used for model training and testing. The empirical results were compared with several classical wind forecast models, and based on the mean absolute error (MAE), the proposed model provides around 9% to 14% improvement in forecasting accuracy compared to an artificial neural network (ANN) model, and nearly 17% improvement on a third dataset which is from a newly-built wind farm for which there is a limited amount of training data.

Original languageEnglish
Pages (from-to)656-665
Number of pages10
JournalIEEE Transactions on Power Systems
Volume29
Issue number2
Early online date2 Oct 2013
DOIs
Publication statusPublished - 1 Mar 2014

Fingerprint

Wind power
Wind turbines
Farms
Turbines
Earth (planet)
Neural networks
Economics
Testing

Bibliographical note

© 2013 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Funding: National Fund for Creative Groups of China (Grant No. 61121003) and in part by a BUAA Scholarship

Keywords

  • censored data
  • Gaussian process
  • numerical weather prediction
  • wind power forecasting

Cite this

Chen, Niya ; Qian, Zheng ; Nabney, Ian T. ; Meng, Xiaofeng. / Wind power forecasts using Gaussian processes and numerical weather prediction. In: IEEE Transactions on Power Systems. 2014 ; Vol. 29, No. 2. pp. 656-665.
@article{c00301cd4b0c4e25a5a3567a00e466a1,
title = "Wind power forecasts using Gaussian processes and numerical weather prediction",
abstract = "Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate wind power forecasts are necessary for the safe and economic use of wind energy. In this paper, we investigated a combination of numeric and probabilistic models: a Gaussian process (GP) combined with a numerical weather prediction (NWP) model was applied to wind-power forecasting up to one day ahead. First, the wind-speed data from NWP was corrected by a GP, then, as there is always a defined limit on power generated in a wind turbine due to the turbine controlling strategy, wind power forecasts were realized by modeling the relationship between the corrected wind speed and power output using a censored GP. To validate the proposed approach, three real-world datasets were used for model training and testing. The empirical results were compared with several classical wind forecast models, and based on the mean absolute error (MAE), the proposed model provides around 9{\%} to 14{\%} improvement in forecasting accuracy compared to an artificial neural network (ANN) model, and nearly 17{\%} improvement on a third dataset which is from a newly-built wind farm for which there is a limited amount of training data.",
keywords = "censored data, Gaussian process, numerical weather prediction, wind power forecasting",
author = "Niya Chen and Zheng Qian and Nabney, {Ian T.} and Xiaofeng Meng",
note = "{\circledC} 2013 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Funding: National Fund for Creative Groups of China (Grant No. 61121003) and in part by a BUAA Scholarship",
year = "2014",
month = "3",
day = "1",
doi = "10.1109/TPWRS.2013.2282366",
language = "English",
volume = "29",
pages = "656--665",
journal = "IEEE Transactions on Power Systems",
issn = "0885-8950",
publisher = "IEEE",
number = "2",

}

Wind power forecasts using Gaussian processes and numerical weather prediction. / Chen, Niya; Qian, Zheng; Nabney, Ian T.; Meng, Xiaofeng.

In: IEEE Transactions on Power Systems, Vol. 29, No. 2, 01.03.2014, p. 656-665.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Wind power forecasts using Gaussian processes and numerical weather prediction

AU - Chen, Niya

AU - Qian, Zheng

AU - Nabney, Ian T.

AU - Meng, Xiaofeng

N1 - © 2013 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Funding: National Fund for Creative Groups of China (Grant No. 61121003) and in part by a BUAA Scholarship

PY - 2014/3/1

Y1 - 2014/3/1

N2 - Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate wind power forecasts are necessary for the safe and economic use of wind energy. In this paper, we investigated a combination of numeric and probabilistic models: a Gaussian process (GP) combined with a numerical weather prediction (NWP) model was applied to wind-power forecasting up to one day ahead. First, the wind-speed data from NWP was corrected by a GP, then, as there is always a defined limit on power generated in a wind turbine due to the turbine controlling strategy, wind power forecasts were realized by modeling the relationship between the corrected wind speed and power output using a censored GP. To validate the proposed approach, three real-world datasets were used for model training and testing. The empirical results were compared with several classical wind forecast models, and based on the mean absolute error (MAE), the proposed model provides around 9% to 14% improvement in forecasting accuracy compared to an artificial neural network (ANN) model, and nearly 17% improvement on a third dataset which is from a newly-built wind farm for which there is a limited amount of training data.

AB - Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate wind power forecasts are necessary for the safe and economic use of wind energy. In this paper, we investigated a combination of numeric and probabilistic models: a Gaussian process (GP) combined with a numerical weather prediction (NWP) model was applied to wind-power forecasting up to one day ahead. First, the wind-speed data from NWP was corrected by a GP, then, as there is always a defined limit on power generated in a wind turbine due to the turbine controlling strategy, wind power forecasts were realized by modeling the relationship between the corrected wind speed and power output using a censored GP. To validate the proposed approach, three real-world datasets were used for model training and testing. The empirical results were compared with several classical wind forecast models, and based on the mean absolute error (MAE), the proposed model provides around 9% to 14% improvement in forecasting accuracy compared to an artificial neural network (ANN) model, and nearly 17% improvement on a third dataset which is from a newly-built wind farm for which there is a limited amount of training data.

KW - censored data

KW - Gaussian process

KW - numerical weather prediction

KW - wind power forecasting

UR - http://www.scopus.com/inward/record.url?scp=84897656346&partnerID=8YFLogxK

U2 - 10.1109/TPWRS.2013.2282366

DO - 10.1109/TPWRS.2013.2282366

M3 - Article

VL - 29

SP - 656

EP - 665

JO - IEEE Transactions on Power Systems

JF - IEEE Transactions on Power Systems

SN - 0885-8950

IS - 2

ER -