The assessment of dynamic wake effects on loading

Abstract

The wake effect from upstream wind turbines is a hot topic recently and many researches on wake modeling have been carried out in the past few years. The wake models have evolved from the simple semi-empirical approach to the dynamic meandering model using the Computational Fluid Dynamics (CFD) codes. However, some of current wake models cannot simulate the dynamic wake meandering process while the other CFD models are too complicated for engineering applications. So the purpose of this thesis is to develop a simple numerical wake model including the detailed wake velocity deficits and dynamic wake meandering process in the atmospheric boundary layer. The Simplified Eddy Viscosity (SEV) model is developed based on the simplified Navier-Stokes equation in Ainslie's work. It describes the wake mixing process in the atmosphere and gives an estimation of quasi-steady two dimensional velocity deficits. Furthermore, the turbulence generator is implemented in the SEV model and provides a three-dimensional turbulence flow field in the wake. An experimental validation study of stationary wake deficits is carried out by using measurements from ECN‘s wind turbine test station. Moreover, the SEV model is modified according to ECN‘s measurements and it shows a fair agreement with GH Bladed eddy viscosity model. Besides the stationary wake deficits, a simple dynamic wake meandering mechanism is introduced in this thesis. According to this mechanism, the Dynamic Wake Meandering (DWM) model is developed, which is an integration of the wake meandering process, the wake deficits from the SEV model, and the aeroelastic model from GH Bladed. The DWM model is validated with Risoe‘s model and GH Bladed and there is good agreement between these models. The DWM model is used for load analysis and then compared with the traditional estimation method MET prescribed by the IEC 61400-1 standard. The flow fields in the DWM model are closer to the physical processes of wake transportation in the atmospheric boundary layer. The results show a significant difference between the methods chosen for the fatigue damage estimation.