WRF-LES Model for Wind Energy Applications: A New Milestone in Microscale Modeling

Mesoscale-microscale coupling has constituted one of the major challenges of the wind industry modeling subsector over the last few years. Wind resource datasets based on mesoscale numerical weather prediction models involve a mature technology that is drawing to a close because the set of approximations used for parameterizing the Planetary Boundary Layer (PBL) do not remain valid at grid-resolutions of tens and hundreds of meters. Furthermore, the degree of development of the industry calls for new specifications such as turbulence, inflow angle, ramps, wind gusts, wind turbines with blades larger than the nocturnal boundary layer, complexity in power curves, which cannot be resolved explicitly by means of a mesoscale model. Over the last years, Vortex has focused its research on overcoming the main limitations of the cutting-edge atmospheric model created by the National Center of Atmospheric Research (NCAR), the Weather Research and Forecasting - Large Eddy Simulation (WRF-LES): i) software efficiency , ii) lateral boundary conditions, iii) Terra Incognita and iv) performance of the surface layer parameterization. As a result of this successful investigation, the Vortex-LES model is capable of providing one year of 10-min time series at any site over the world.

Vortex-LES achieves a resolution of 100 m and it produces one year of 10-min time series at any location. This resolution will be increased in the coming years. The output of the model is saved every 0.25 s and then post-processed to 10-mean averages, fluxes and standard deviations. This process is performed for all fields in a region between the surface and 200 m covering an area of 2 km x 2 km. Indeed, Vortex-LES functions as a "poor" sonic anemometer which allows new variables, such as wind gust, to be saved without additional parameterizations.

Two years ahead, we presented a depth validation in 8 wind industry sites over Europe, South America and South Africa with promising results of the WRF-LES under real cases (Montornès et al., 2016). In this new contribution, we summarize our learning and knowledge after more than 300 sites worldwide distributed. The presentation is divided in two parts.

The former ones shows an extended validation for wind speed and turbulence intensity in 85 sites and 50 sites, respectively. These sites are characterized by different met-mast heights, topographic complexity, local weather regimes and climate features. The validation is performed with one year of 10-min time series of WRF-LES initialized with the ERA-5 reanalysis dataset.

The study of wind speed includes a discussion of several time scales (10-min, hourly, daily and monthly), daily cycle, spatial distribution, seasonal behavior and more sophisticated analyses such as energy spectrum and wind direction. Results for 10-min time series show a mean correlation of 0.683, which increases to 0.881 for monthly values. WRF-LES performs a mean absolute error of 5.1% with respect to mean wind speed.

The analysis of turbulence focuses upon 10-min values for wind standard deviation and turbulence intensity. Wind standard deviation shows a mean absolute error of 1.3%. Turbulence intensity is studied as a function of wind speed.

The latter part focuses the discussion upon two particular cases: one in on-shore flat terrain and the other in on-shore complex terrain.

This is the first step towards achieving a seamless modeling chain across the scales, from large motions in the atmosphere to small eddies at the microscale. In the coming years, new improvements are to be expected, such as an increase in grid resolution to tens of meters, and Terra Incognita and surface layer parameterizations, among others.

Montornès, A., Casso, P., Kosovic, B., Lizcano, G.: Is WRF-LES a Suitable Tool for Realistic Turbulence Analyses in Wind Resource Assessment Applications? 22nd Symp. Boundary Layer and Turbulence. 2016.