TY - JOUR
T1 - Multi-Stage Statistical Approach to Wind Power Forecast Errors Evaluation
T2 - A Southern Sulawesi Case Study
AU - Barus, Dhany Harmeidy
AU - Dalimi, Rinaldy
N1 - Funding Information:
ACKNOWLEDGMENT The authors are obliged to Suroso Isnandar, Nurdin Pabi, Titi Yustiana, Goklas Batapanmuda, and Tommy Kaendo from PLN UIKL Sulawesi for providing the operational data. The authors are also grateful to Universitas Indonesia (UI) to fund this research through PUTI Q2 grant 2020 launched by DRPM UI.
Publisher Copyright:
© 2021. All Rights Reserved.
PY - 2021
Y1 - 2021
N2 - Wind Power Plant (WPP) is part of renewable energy sources, with rapid expansion worldwide. It has the advantages of clean and green energy, but its uncertainty leads to an additional grid integration cost. The uncertainty of wind power output is much dependent on the accuracy of the wind power forecast (WPF) result. Since there is no perfect wind power forecast, understanding the current system's forecast accuracy characteristics is essential in expecting typical errors faced in the future. This paper proposed a new algorithm of the statistical approach method to evaluate characteristics of wind power forecast errors (WPFE) from an observed power system with high-penetration WPP. This method combined the approach of scatter diagram, statistical distribution, standard error performance, and score weighting in a multi-stage algorithm. It consists of serial and parallel processes to check the consistency of the results. In this study, a comprehensive analysis was made of various scenarios based on location and timescale. This proposed algorithm has been successfully tested on statistical data of Sidrap WPP and Jeneponto WPP in the Southern Sulawesi power system. The result showed that the scenario with the aggregation of both WPPs in hour-ahead timescale has the most accurate and consistent performance among all scenarios. It demonstrated specific characteristics of WPFE in the observed power system that can be used as an essential starting point in conducting future wind integration expansion studies.
AB - Wind Power Plant (WPP) is part of renewable energy sources, with rapid expansion worldwide. It has the advantages of clean and green energy, but its uncertainty leads to an additional grid integration cost. The uncertainty of wind power output is much dependent on the accuracy of the wind power forecast (WPF) result. Since there is no perfect wind power forecast, understanding the current system's forecast accuracy characteristics is essential in expecting typical errors faced in the future. This paper proposed a new algorithm of the statistical approach method to evaluate characteristics of wind power forecast errors (WPFE) from an observed power system with high-penetration WPP. This method combined the approach of scatter diagram, statistical distribution, standard error performance, and score weighting in a multi-stage algorithm. It consists of serial and parallel processes to check the consistency of the results. In this study, a comprehensive analysis was made of various scenarios based on location and timescale. This proposed algorithm has been successfully tested on statistical data of Sidrap WPP and Jeneponto WPP in the Southern Sulawesi power system. The result showed that the scenario with the aggregation of both WPPs in hour-ahead timescale has the most accurate and consistent performance among all scenarios. It demonstrated specific characteristics of WPFE in the observed power system that can be used as an essential starting point in conducting future wind integration expansion studies.
KW - multi-stage
KW - Statistical approach
KW - wind power forecast errors
UR - http://www.scopus.com/inward/record.url?scp=85106665746&partnerID=8YFLogxK
U2 - 10.18517/ijaseit.11.2.12385
DO - 10.18517/ijaseit.11.2.12385
M3 - Article
AN - SCOPUS:85106665746
SN - 2088-5334
VL - 11
SP - 633
EP - 641
JO - International Journal on Advanced Science, Engineering and Information Technology
JF - International Journal on Advanced Science, Engineering and Information Technology
IS - 2
ER -