Coupling a Mesoscale NWP model with Large-Eddy-Simulation CFD for realistic wind-plant aerodynamics simulations

Wind plant aerodynamics are influenced by a combination of microscale and mesoscale phenomena. The current state of the art is to simulate the microscales using the large-eddy simulation (LES) technique, which directly resolves the energy-containing scales of turbulence. Such microscale simulations are usually idealized, using periodic lateral boundaries and homogeneous surface characteristics over flat terrain. Current practices of computationally studying wind plant aerodynamics involves embedding actuator turbine models within such microscale LES. Although this method has led to improved understanding, the absence of realistic atmospheric forcing limits its applicability. For example, this approach does not capture the non-homogeneity of real inflow winds due to the influence of mesoscale weather, such as diurnal cycles or frontal passages.

Mesoscale numerical weather prediction (NWP) models offer a possibility for providing realistic forcing for microscale LES. However, since in NWP models the boundary layer turbulence is parameterized and not directly resolved, it is not straightforward to produce realistic turbulence in the microscale domain when coupling a mesoscale model with microscale LES. This is an area of research that needs more exploration. In our presentation we focus therefore on coupling techniques from the Weather Research and Forecasting (WRF) model to OpenFOAM, a microscale CFD toolbox. We will discuss nesting down from WRF to WRF-LES, and use the latter as input to OpenFOAM. We will quantify, using spectra, how turbulence spins up with distance from the domain boundaries. Specific research questions further include the effect of resolution in the WRF-LES domain, the decrease of wind speeds in the WRF LES domain as opposed to measurements and the mesoscale domain, and using the WRF mesoscale model output directly (without an intermediary WRF LES domain) as input to OpenFOAM. Simulations will be compared with observations from the Crop and Wind Energy eXperiment (CWEX).