An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia

Yong Kim Hwang, Mohd Zamri Ibrahim, Ali Najah Ahmed, Aliashim Albani

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The major issues on the wind measurement campaign are the data measured in a short period and the occurrence of missing data due to the failure of the measurement instrument. Meanwhile, Measure-Correlate-Predict (MCP) method had widely been used to predict the long-term condition and missing data at the measurement site based on nearest Malaysian Meteorological Department (MMD), Meteorological Aerodrome Report (METAR) and extended Climate Forecast System Reanalysis (ECFSR) data. In this research, the long-term wind data at selected potential sites in Malaysia were predicted by optimized Artificial Neural Networks (ANNs). The Genetic Algorithm (GA) was applied to optimize the ANN. Five different ANN MCP models had been designed based on different types of reference data and different temporal scales to predict wind data at three target sites. Weibull frequency distributions and RMSE examined predicted wind data. The prediction of ANN had been improved in between 20.562% to 113.573% by GA optimization. The best R-value obtained from optimization were affected the Weibull shape and scale of predicted data. At last, the result revealed that the optimized ANN model could predict the long-term data for the target site with better accuracy.

Original languageEnglish
Title of host publicationProceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018
PublisherIEEE Computer Society
ISBN (Electronic)9789748257990
DOIs
Publication statusPublished - 05 Feb 2019
Event2018 International Conference on Green Energy for Sustainable Development, ICUE 2018 - Karon, Phuket, Thailand
Duration: 24 Oct 201826 Oct 2018

Publication series

NameProceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE
Volume2018-October
ISSN (Print)2166-0581
ISSN (Electronic)2166-059X

Conference

Conference2018 International Conference on Green Energy for Sustainable Development, ICUE 2018
CountryThailand
CityKaron, Phuket
Period24/10/1826/10/18

Fingerprint

Neural networks
Genetic algorithms

All Science Journal Classification (ASJC) codes

  • Energy Engineering and Power Technology
  • Fuel Technology
  • Renewable Energy, Sustainability and the Environment

Cite this

Hwang, Y. K., Ibrahim, M. Z., Ahmed, A. N., & Albani, A. (2019). An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia. In Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018 [8635790] (Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE; Vol. 2018-October). IEEE Computer Society. https://doi.org/10.23919/ICUE-GESD.2018.8635790
Hwang, Yong Kim ; Ibrahim, Mohd Zamri ; Ahmed, Ali Najah ; Albani, Aliashim. / An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia. Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018. IEEE Computer Society, 2019. (Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE).
@inproceedings{1d71c10d0198434687e5f6eecf2f7089,
title = "An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia",
abstract = "The major issues on the wind measurement campaign are the data measured in a short period and the occurrence of missing data due to the failure of the measurement instrument. Meanwhile, Measure-Correlate-Predict (MCP) method had widely been used to predict the long-term condition and missing data at the measurement site based on nearest Malaysian Meteorological Department (MMD), Meteorological Aerodrome Report (METAR) and extended Climate Forecast System Reanalysis (ECFSR) data. In this research, the long-term wind data at selected potential sites in Malaysia were predicted by optimized Artificial Neural Networks (ANNs). The Genetic Algorithm (GA) was applied to optimize the ANN. Five different ANN MCP models had been designed based on different types of reference data and different temporal scales to predict wind data at three target sites. Weibull frequency distributions and RMSE examined predicted wind data. The prediction of ANN had been improved in between 20.562{\%} to 113.573{\%} by GA optimization. The best R-value obtained from optimization were affected the Weibull shape and scale of predicted data. At last, the result revealed that the optimized ANN model could predict the long-term data for the target site with better accuracy.",
author = "Hwang, {Yong Kim} and Ibrahim, {Mohd Zamri} and Ahmed, {Ali Najah} and Aliashim Albani",
year = "2019",
month = "2",
day = "5",
doi = "10.23919/ICUE-GESD.2018.8635790",
language = "English",
series = "Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE",
publisher = "IEEE Computer Society",
booktitle = "Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018",
address = "United States",

}

Hwang, YK, Ibrahim, MZ, Ahmed, AN & Albani, A 2019, An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia. in Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018., 8635790, Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE, vol. 2018-October, IEEE Computer Society, 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018, Karon, Phuket, Thailand, 24/10/18. https://doi.org/10.23919/ICUE-GESD.2018.8635790

An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia. / Hwang, Yong Kim; Ibrahim, Mohd Zamri; Ahmed, Ali Najah; Albani, Aliashim.

Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018. IEEE Computer Society, 2019. 8635790 (Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE; Vol. 2018-October).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia

AU - Hwang, Yong Kim

AU - Ibrahim, Mohd Zamri

AU - Ahmed, Ali Najah

AU - Albani, Aliashim

PY - 2019/2/5

Y1 - 2019/2/5

N2 - The major issues on the wind measurement campaign are the data measured in a short period and the occurrence of missing data due to the failure of the measurement instrument. Meanwhile, Measure-Correlate-Predict (MCP) method had widely been used to predict the long-term condition and missing data at the measurement site based on nearest Malaysian Meteorological Department (MMD), Meteorological Aerodrome Report (METAR) and extended Climate Forecast System Reanalysis (ECFSR) data. In this research, the long-term wind data at selected potential sites in Malaysia were predicted by optimized Artificial Neural Networks (ANNs). The Genetic Algorithm (GA) was applied to optimize the ANN. Five different ANN MCP models had been designed based on different types of reference data and different temporal scales to predict wind data at three target sites. Weibull frequency distributions and RMSE examined predicted wind data. The prediction of ANN had been improved in between 20.562% to 113.573% by GA optimization. The best R-value obtained from optimization were affected the Weibull shape and scale of predicted data. At last, the result revealed that the optimized ANN model could predict the long-term data for the target site with better accuracy.

AB - The major issues on the wind measurement campaign are the data measured in a short period and the occurrence of missing data due to the failure of the measurement instrument. Meanwhile, Measure-Correlate-Predict (MCP) method had widely been used to predict the long-term condition and missing data at the measurement site based on nearest Malaysian Meteorological Department (MMD), Meteorological Aerodrome Report (METAR) and extended Climate Forecast System Reanalysis (ECFSR) data. In this research, the long-term wind data at selected potential sites in Malaysia were predicted by optimized Artificial Neural Networks (ANNs). The Genetic Algorithm (GA) was applied to optimize the ANN. Five different ANN MCP models had been designed based on different types of reference data and different temporal scales to predict wind data at three target sites. Weibull frequency distributions and RMSE examined predicted wind data. The prediction of ANN had been improved in between 20.562% to 113.573% by GA optimization. The best R-value obtained from optimization were affected the Weibull shape and scale of predicted data. At last, the result revealed that the optimized ANN model could predict the long-term data for the target site with better accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85062879361&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062879361&partnerID=8YFLogxK

U2 - 10.23919/ICUE-GESD.2018.8635790

DO - 10.23919/ICUE-GESD.2018.8635790

M3 - Conference contribution

AN - SCOPUS:85062879361

T3 - Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE

BT - Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018

PB - IEEE Computer Society

ER -

Hwang YK, Ibrahim MZ, Ahmed AN, Albani A. An Optimized ANN Measure-Correlate-Predict Method for Long-term Wind Prediction in Malaysia. In Proceedings of the 2018 International Conference on Green Energy for Sustainable Development, ICUE 2018. IEEE Computer Society. 2019. 8635790. (Proceedings of the Conference on the Industrial and Commercial Use of Energy, ICUE). https://doi.org/10.23919/ICUE-GESD.2018.8635790