Abstract
We investigate the use of spatial interpolation methods for reconstructing the horizontal near-surface wind field given a sparse set of measurements. In particular, random Fourier features is compared with a set of benchmark methods including kriging and inverse distance weighting. Random Fourier features is a linear model approximating the velocity field, with randomly sampled frequencies and amplitudes trained to minimize a loss function. We include a physically motivated divergence penalty , as well as a penalty on the Sobolev norm of . We derive a bound on the generalization error and a sampling density that minimizes the bound. We then devise an adaptive Metropolis–Hastings algorithm for sampling the frequencies of the optimal distribution. In our experiments, our random Fourier features model outperforms the benchmark models.
Highlights
An integral part in wind farm planning and weather prediction is access to high-quality wind data [1,2]
|K| needs to be sufficiently large for the error to be reasonably small
We explored the potential for wind field reconstruction with sparse data using interpolation models
Summary
An integral part in wind farm planning and weather prediction is access to high-quality wind data [1,2]. Meteorological stations are often heterogeneously distributed with many gaps, and interpolation techniques have to be employed in order to increase spatial resolution [3]. A crucial step in building wind farms is site prospecting, in which national or state-level measurements are used to estimate the. Expected aggregate yearly or monthly energy output [1]. A common approach is to approximate the probabilistic distribution of the wind speed over time with some parametric model, apply royalsocietypublishing.org/journal/rspa Proc.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
