CFD Modeling of a Neutral Atmospheric Boundary Layer over Complex Terrain

Abstract

As wind energy production continues to grow the most optimal areas for wind energy generation have quickly been utilised. As wind farms are more frequently placed in less optimal areas, the need for methods to accurately site these areas becomes paramount. One such method is to use Computational Fluid Dynamics (CFD) models of the atmospheric boundary layer to predict the flow structures above sites which are located in complex terrain. This study investigates the use of Detached Eddy Simulations (DES) as well as Large Eddy Simulations (LES) to quantify the effects of flow over complex terrain. The simulated wind fields are used in combination with an aero-elastic solver to assess the dynamic response of a wind turbine. It is found that the protective properties of boundary layers in DES is such that it makes it inconvenient for generating a fluctuating wind field since it dampens out fluctuations in altitudes occupied by wind turbines. LES is found to replicate well the predicted turbulence as well as time-averaged velocity profiles even though no rough wall treatment is used. However, discrepancies are found when investigating the Reynolds stresses that result from a combination of a too short sampling time and symmetric boundary conditions. Albeit these discrepancies the wind fields generated through LES are used with an aero-elastic solver to simulate the dynamic response of a "NREL 5-MW Baseline Wind Turbine". The results show that the wind turbine located in complex terrain is subject to more fatigue than if it is placed on flat terrain. The identified load cycles are however relatively few and so the simulation time should be increased to get a better statistical representation. Due to the discrepancies resulting from symmetric boundaries and short sampling time the effect of complex terrain is not quantified with any certainty. However, LES in the absence of complex wall functions is proven useful as a tool to assess wind fields over complex terrain. Further investigation is necessary where it is recommended that boundaries be assigned differently and longer sampling times be used.