Nested simulations for wind energy applications

Operational wind power forecasting, wind turbine micrositing, and turbine design require high-resolution simulations of atmospheric flow over complex terrain. We are using the Weather Research and Forecasting (WRF) model to develop large-eddy simulations (LES) for wind energy applications. Grids are refined from mesoscale to finer scales using grid nesting to adequately resolve turbulence and terrain in the atmospheric boundary layer. On the finest grids, LES turbulence closures are used. Thus, changing weather conditions are incorporated even in the smallest domains by time dependent lateral boundary conditions provided by grid nesting, which seamlessly transfers information between large and small domains. This framework combines synoptic-scale and local information to improve estimates of wind speed and direction, wind shear, and atmospheric stability at wind turbine hub heights.

Here, we investigate the effect of different grid nesting configurations and the choice of turbulence model on predictions of wind speed at a west coast wind farm. Preliminary results show that nesting down to finer resolutions does not significantly improve model results compared to field observations for days with strong synoptic forcing such as frontal passages. There is only slight improvement in the results when comparing coarser to finer vertical resolution. A comparison between two-way nesting, where the finer grid feeds information back to the parent, and one-way nesting does not yield significant differences either. On the other hand, the model results appear sensitive to the choice of turbulence model. We will examine “mixed models”, which combine an eddy-viscosity component with a scale-similarity component and have shown improved representation of shear-generated turbulence in other work. These results for synoptically-forced conditions will be compared to days dominated by local forcing to determine when higher resolution is beneficial at this particular wind farm.