A self-organizing forecast of day-ahead wind speed: Selective ensemble strategy based on numerical weather predictions

Jing Zhao, Zhenhai Guo, Yanling Guo, Wantao Lin and Wenjin Zhu

Energy, 2021, vol. 218, issue C

Abstract: Accurate wind predictions are required in modern wind power generations, for which numerical weather predictions are generally used. This study focuses on short-term wind speed forecast based on ensemble numerical simulations. As for typical ensemble methods, all the single members are intuitively included, while an interesting idea is that the ensemble performance could be improved if the low-performing members are recognized and removed. With this conception, a selective ensemble system is developed in this study, which is a new method of ensemble numerical simulations especially for day-ahead wind speed predictions at local wind farms. The developed method is constructed by ensemble atmospheric simulations of weather research and forecasting models, a sample clustering process using Gaussian mixture model, an automatic modeling mechanism using group method of data handling network, and a multi-objective Cuckoo search optimization. The results obtained led to the following original conclusions: (i) the proposed method automatically excludes some low-performing members, signifying that ensembles with selected members could generate better forecasts than those with a combination of all members; (ii) by partitioning wind series into wave-like segments and applying sample clustering, model’s performance can be further improved.

Keywords: Wind speed forecasts; Numerical weather prediction; Gaussian mixture model; Group method of data handling; Ensemble method; Evolutionary algorithm (search for similar items in EconPapers)
Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (19)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544220326165
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:218:y:2021:i:c:s0360544220326165

DOI: 10.1016/j.energy.2020.119509

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().