A CFD framework for offshore and onshore wind farm simulation

Abstract

Abstract. We present a wind simulation framework for offshore and onshore wind farms. The simulation framework involves an automatic hybrid high-quality mesh generation process, a preprocessing to impose initial and boundary conditions, and a solver for the Reynolds Averaged Navier-Stokes (RANS) equations with two different turbulence models, a modified standard k-ε model and a realizable k-ε model in which we included the Coriolis effects. Wind turbines are modeled as actuator discs. The wind farm simulation framework has been implemented in Alya, an in-house High Performance Computing (HPC) multi-physics finite element parallel solver. An application example is shown for an onshore wind farm composed of 165 turbines. IntroductionSimulation of wind farms with Computational Fluid Dynamics (CFD) models involves the resolution of the turbulent Atmospheric Boundary Layer (ABL) and the effects induced by wind turbines, including wind speed deficit, increase of turbulent kinetic energy, and interaction among wakes. Commonly, wind turbines are modeled as uniformly loaded actuator discs [1], an option which allows a compromise between a reasonable model accuracy and a low computational cost.However, it is well known that the Reynolds-Averaged Navier-Stokes (RANS) standard k-ε model significantly overestimates Reynolds stresses [2] behind actuator discs, resulting on a significant underestimation of velocities and on an excessive wake damping. Several modifications of the standard k-ε model have been proposed for delaying the simulated wake flow recovery [3], but most of them require the adjustment of model parameters that depend on the characteristics of the wind turbine at stake. Shi et al. [4] proposed a realizable k-ε model that enhances the wake predictions of actuator discs [5] and needs no adjustment of parameters. On the other hand, it is also well-established that the k-ε model needs an additional mixing length limitation model [6] to accurately predict wind intensity profiles and wind veering with height in the ABL when considering Coriolis forces. However, no attempts have been made to modify the realizable model for the simulation of Coriolis forces in the ABL.Several aspects of wind farm modeling pose requirements on the computational mesh. Firstly, the existence of a boundary layer imposes stretching requirements in order to capture nearsurface sharp gradients. Secondly and for onshore farms, the underlying terrain must be discretized and properly approximated by the mesh. Finally, the actuator discs have to be embedded in the mesh with a subsequent mesh refinement around and downstream the turbines in order to capture near wake flow effects properly. This is a constrain for conformal structured

Micro-scale model comparison (benchmark) at the moderately complex forested site Ryningsnäs

“…Zero diffusion through the wall is imposed for the TKE, and the dissipation ε is imposed in terms of the TKE value at a distance δ w from the ground. For a detailed description see Avila et al (2017).…”

Diurnal cycle RANS simulations applied to wind resource assessment

“…Alya‐CFDWind() is a CFD code for the ABL implemented in Alya, a High Performance Computing multiphysics parallel solver based on the finite element method. () In order to consider atmospheric stability and Coriolis forces, we have implemented the thermal model of Sogachev et al The resulting Alya‐CFDWind model equations are the incompressible unsteady Reynolds averaged Navier‐Stokes (URANS) Equations , together with a k ‐ ε turbulence model to and the energy conservation Equation .…”

A Hybrid Meshing Framework Adapted to the Topography to Simulate Atmospheric Boundary Layer Flows

“…Herein, we present a short overview of our CFD model for ABL flows, linking it with the required mesh specifications for our specific purpose mesher. Our complete CFD model, featuring the coupling with the actuator disc model for wind farm simulation, and the implementation, can be found in [23,24].…”

“…We highlight that the model is implemented in the finite-element multi-physics parallel solver Alya [28,29]. However, the proposed meshing algorithm has been implemented in the external model-independent pre-process code WindMesh [23,9,10,30]. As a result, meshes generated with this utility can be used to simulate both with RANS or Large-Eddy Simulation (LES) turbulence models and, in addition, are also valid for solvers based on other numerical methods such as Finite Volumes.…”

“…One of the targets of the hybrid mesher is to reduce the degrees of freedom required to discretize a complex topography and the Atmospheric Boundary Layer in simulations featuring Coriolis effects. To perform this comparison, we use our in-house semi-structured code WindMesh [9,10], that generates a semi-structured cross-type mesh, see Figure 9(a), which constitutes the standard in the field of ABL flow simulation [6,7,8,9,23,11,10]. As previously highlighted, these semi-structured meshers are devoted to generate an hexahedral mesh aligned with the inflow direction.…”