Modelling of wind turbine wake using large eddy simulation

• Published in: Renewable energy 2018-01, Vol.115, p.1166-1176
• Main Authors: Sedaghatizadeh, Nima, Arjomandi, Maziar, Kelso, Richard, Cazzolato, Benjamin, Ghayesh, Mergen H.
• Format: Article
• Language: English
• Subjects: Engineering wake models; Velocity deficit in the wake; Wake expansion; Wind farm; Wind turbine wake
• ISSN: 0960-1481; 1879-0682
• DOI: 10.1016/j.renene.2017.09.017

In an array of wind turbines, the interaction of the downstream machines with the wakes from the upstream ones results in a reduction in the overall wind farm performance. Turbine wakes are a major source of turbulence which exerts fluctuating loads on the blades of the downstream turbines, resulting in the generation of noise and fatigue of the turbine blades. There are many semi-empirical wind turbine wake models in the literature. This paper, develops a fully numerical model of wind turbine wakes using CFD by means of a Large Eddy Simulation (LES). The new LES model is tested against experimental data, showing very good agreement. The advantages of the LES model compared to the available semi-empirical models in the literature are discussed and it is shown that the LES model is very accurate compared to the conventional semi-empirical wake models usually used in industry. Moreover, the LES model is used as a benchmark to compare the accuracy of these semi-empirical models; it is shown that the model proposed by Jensen can predict the velocity deficit most accurately among the semi-empirical models, while the highest degree of accuracy in the wake expansion is achieved by using the Larsen model. •Four well-known semi-empirical wake models are assessed and compared with the well-resolved validated Large Eddy Simulation model of a wind turbine.•It is found that the models proposed by Jensen and Frandsen can provide more accurate results in near-wake region in comparison with Larsen models.•The velocity deficit and wake diameter in far-wake is better predicted by Larsen models.•In order to increase the accuracy while maintaining the computation cost to minimum, it is suggested to use a combination of the models based on their strength in predicting desirable factors in different regions.