Simulations of an Offshore Wind Farm Using Large-Eddy Simulation and a Torque-Controlled Actuator Disc Model

Abstract

We present here a computational fluid dynamics (CFD) simulation of Lillgrund offshore wind farm, which is located in the Øresund Strait between Sweden and Denmark. The simulation combines a dynamic representation of wind turbines embedded within a Large-Eddy Simulation CFD solver, and uses hr-adaptive meshing to increase or decrease mesh resolution where required. This allows the resolution of both large scale flow structures around the wind farm, and the local flow conditions at individual turbines; consequently, the response of each turbine to local conditions can be modelled, as well as the resulting evolution of the turbine wakes. This paper provides a detailed description of the turbine model which simulates the interaction between the wind, the turbine rotors, and the turbine generators by calculating the forces on the rotor, the body forces on the air, and instantaneous power output. This model was used to investigate a selection of key wind speeds and directions, investigating cases where a row of turbines would be fully aligned with the wind or at specific angles to the wind. Results shown here include presentations of the spin-up of turbines, the observation of eddies moving through the turbine array, meandering turbine wakes, and an extensive wind farm wake several kilometres in length. The key measurement available for cross-validation with operational wind farm data is the power output from the individual turbines, where the effect of unsteady turbine wakes on the performance of downstream turbines was a main point of interest. The results from the simulations were compared to performance measurements from the real wind farm to provide a firm quantitative validation of this methodology. Having achieved good agreement between the model results and actual wind farm measurements, the potential of the methodology to provide a tool for further investigations of engineering and atmospheric science problems is outlined.

Effects of Support Structures in an LES Actuator Line Model of a Tidal Turbine with Contra-Rotating Rotors

“…Building upon previous work [4], both the mean eddy length-scale and Reynolds stress profiles were specified as a function of height above the seabed for SEM, so that realistic turbulent inflow was generated. Eddy length-scales were taken from Milne et al [41], whose measurements from the Sound of Islay agreed with Nezu and Nakagawa [40].…”

“…Therefore, a power law regression fit was also applied to the mean vertical velocity profile. As the roughness on the channel bottom is not explicitly resolved, the roughness height z R was instead derived from the skin friction coefficient C F [4] for a more appropriate fit. This gave the power law:…”

“…The turbine model developed here builds upon previous work, where dynamic torque-controlled actuator discs with active-pitch correction were used to model wind turbines [15] and wind farms [4,16]. In the present model, actuator line techniques [5] have been used to represent the rotor, whereby the blades themselves are not resolved, but the forces exerted by them on the fluid are still present.…”

“…Models range from early empirical linear wake superposition approaches such as the Park model [1], through to Reynolds-averaged Navier-Stokes (RANS) CFD actuator disc models, large eddy simulation (LES) actuator disc [2][3][4] and actuator line models [5][6][7][8]. Porté-Agel et al [9] simulated a wind farm using actuator disc and actuator line methods, the results comparing favourably with experimental data.…”

“…One striking feature of these models is that, barring a few exceptions [13,14], the turbine support structure is not modelled explicitly, and so, only the rotors affect the downwind flow. It is quite likely that, in the mid-to-far wake region, wake effects due to the structure are not important in wind farms; indeed, previous, validated studies of single wind turbines [15] and wind farms [4,16] have indicated that the tower and nacelle have negligible impact on the wake and consequently the performance of downwind turbines. The pertinent question here is, then, can the same be said for tidal turbines sited in swiftly flowing water, whose density is over 800 times that of air?…”