.png){kind=icon}
.png){kind=icon}
Return
Improving probabilistic wind speed forecasting using M-Rice distribution and spatial data integration
.png){kind=icon}
OA
AI
DOI：10.1016/j.apenergy.2024.122840.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We consider the problem of short-term forecasting of surface wind speed probability distribution. Our approach simply consists in predicting the parameters of a given probability density function by training a neural network model whose loss function is chosen as the log -likelihood provided by this distribution. We compare different possibilities among a set of distributions that have been previously considered in the context of modeling wind fluctuations. Our results rely on two different hourly wind speed datasets: the first one has been recorded by Meteo-France in Corsica (South France), a very mountainous Mediterranean island while the other one relies on KNMI database that provides records of various stations over The Netherlands, a very flat country in Northwestern Europe. A first part of our work globally unveils the superiority of the so-called Multifractal Rice (M -Rice) distribution over alternative parametric models, showcasing its potential as a reliable tool for wind speed forecasting. This family of distributions has been proposed in the context of modeling wind speed fluctuations as a random cascade model along the same picture as fully developed turbulence. For all stations in both regions, it consistently provides better results regardless of the considered probabilistic scoring rule or forecasting horizon. Our second findings demonstrate significant enhancements in forecasting accuracy when one incorporates wind speed data from proximate weather stations, in full agreement with the results obtained formerly for point -wise wind speed prediction. Moreover, we reveal that the incorporation of ERA5 reanalysis of 10 m wind data from neighboring grid points contributes to a substantial improvement at time horizon h = 6 hours while at a shorter time horizon (h = 1 h) it does not lead to any noticeable improvement. It turns out that accounting for pertinent features and explanatory factors, notably those related the spatial distribution and wind speed and direction, emerges as a more critical factor in enhancing accuracy than the choice of the optimal parametric distribution. We also find out that accounting for more explanatory factors mainly increases the resolution performances while it does not change the reliability contribution to the prediction performance metric considered (CRPS).
Keywords:
Surface wind speed
Wind power
Deep learning models
Probabilistic forecasting
Short-term forecasting
Spatio-temporal correlations
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
11
Papers:
2.6W
Citations:
17.8W