Increasing spatial resolution of wind resource prediction using NWP and RANS simulation

The detailed prediction of the upcoming wind on wind farms can support optimization of wind energy production and operation and maintenance. Numerical Weather Prediction (NWP) tools allow to simulate the wind over long-term forecasting horizons (up to several days) with a spatial resolution ranging between the continental level down to a few hundred meters. We present a methodology, based upon Computational Fluid Dynamics (CFD) and Reynolds Averaged Navier Stokes (RANS) modelling, that allows to downscale the spatial resolution of the wind prediction supplied by a NWP model down to the typical length-scale of wind energy applications. The proposed approach combines a number of standard tools, including: Geographical Information Systems (GIS), Advanced Research - Weather Research and Forecasting (WRF-ARW) and OpenFOAM, and proposes methods to interface these tools and set-up the local-scale simulation. Models and problem sizes are selected to keep the computational cost of the system sustainable in view of its implementation in operational forecasting. Finally, we present the application of the method on a given onshore site, and for three different meteorological conditions, showing the potential of the approach, but also giving an account of the limitations that it may encounter when dealing with complex planetary boundary layers.

Automated summary

The article presents a methodology based on Computational Fluid Dynamics (CFD) and Reynolds Averaged Navier Stokes (RANS) modeling to downscale the spatial resolution of wind prediction for wind energy production optimization. By combining standard tools such as GIS, WRF-ARW, and OpenFOAM, the approach allows for local-scale simulation with sustainable computational cost.