CFD wind turbines wake assessment in complex topography Hafida Daaou Nedjari a,b,⇑ , Ouahiba Guerri a , Mohamed Saighi b a Centre de Développement des Energies Renouvelables, CDER B.P. 62, Route de l’Observatoire, Bouzaréah, 16340 Alger, Algeria b Thermodynamics and Energetical Systems Laboratory, Faculty of Physics, USTHB, B.P 32 El Alia, 16111 Bab Ezzouar-Algiers, Algeria article info Article history: Received 8 November 2016 Received in revised form 10 January 2017 Accepted 30 January 2017 Keywords: CFD Hybrid method RANS Complex topography Wake-ground interaction k-epsilon model abstract The purpose of this work is to numerically study the wind turbine wake evolution in farm over both com- plex and flat terrain. The nonlinear, aerodynamic interaction between the rotor wake and the wind farm terrain is modeled using the Hybrid method combining CFD (Computational Fluid dynamics) with the actuator disk model. The rotor defined as a function of the wind speed and the thrust coefficient is applied through a source term added in Navier-Stokes equations within (RANS) decomposition. The framework is structured and resolved via an Open Source Computational fluid dynamics program based on the finite volume method. The interactions between atmospheric upwind boundary layer, downwind wake and ground effects are evaluated considering different wind farm configurations: First, simulations are con- ducted for flat terrain by varying the wind turbine hub height. The soil effects on the wake evolution are estimated by means of the size length of the eddy areas of low speed recorded behind the rotor. In the second configuration concerning the complex terrain, the proposed hybrid method is adapted to the local wind field significantly disturbed by the topography singularities. The flow field obtained at the hub level is then analyzed and used to define the corresponding actuator disk model. This approach is applied to a small region located in north of Algeria. However, the accuracy and performance of the proposed model to predict the near wake and the far wake are demonstrated by a comparison with wake measurement over flat terrain and are in good agreement with experimental data. Results obtained in all cases gives interesting information involved in wind farm layout. Ó 2017 Elsevier Ltd. All rights reserved. 1. Introduction Due to the energy crisis, the development of wind energy is now expanding worldwide. Algeria has recently included in the new renewable energy programs a part of 5Gwatt of wind energy installed by 2030. To achieve this goal, the wind farms optimiza- tion and the windy site studies are undertaken by different govern- mental research centers such as CDER, CREDEG and SKTM. However, wind farms are particularly difficult to model because of the wake impact with their multiple wind turbines which con- siderably reduce the wind farm efficiency. On the other side, the most Algerian windy sites are characterized by complex topogra- phy. Therefore, the wind farm implementation remains a great challenge. To improve the wind farm yielding, numerous numerical and experimental works have focused the wake evolution. Baker [1] and Taylor [2,3] were among the first to experiment horizontal axis wind turbine in full scale, considering homogeneous terrain and storing an interesting databases of wake at different distances downstream. Few experimental works have concerned the wake in complex terrain because of the high complexity degree [4,5]. In parallel, numerical studies have focused on the development of wake model that can nearly approximate the experimental data in a wind farm. The main problem is how to accurately represent the rotor while considering the turbulent atmospheric structures and the rotational aspect of the fluid flow. However, in order to simplify the wake model, the terrain features adding an extra parameter to model are generally not treated at the same time [6]. Several rotor models have been developed and listed in litera- ture [6–9]. The most commonly used in the wind energy industry are based on the engineering BEM model (Blade Element Momen- tum theory) [10,11] consisting on the momentum and the blade element theory. This approach can easily provide the loads on the blade and general information on the deficit abaft the rotor but not on wake evolution downstream involved in the wake- wake and wake-ground interactions [6,7]. Other simplified formu- lations proposed by O. Jensen were frequently used in wind poten- tial assessment software as the WAsP [12]. Based on the statistical treatment of the measured wind data, a linear equation assuming a http://dx.doi.org/10.1016/j.enconman.2017.01.070 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved. ⇑ Corresponding author at: Centre de Développement des Energies Renouve- lables, CDER B.P. 62, Route de l’Observatoire, Bouzaréah, 16340 Alger, Algeria. E-mail address: h.daaou@cder.dz (H. Daaou Nedjari). Energy Conversion and Management 138 (2017) 224–236 Contents lists available at ScienceDirect Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman