Share this page on:

Conference programme 

Back to the programme printer.gif Print

Poster session

Lead Session Chair:
Stephan Barth, Managing Director, ForWind - Center for Wind Energy Research, Germany
Johannes Lindvall Kjeller Vindteknikk AS, Sweden
Co-authors:
Johannes Lindvall (1) F P Øyvind Byrkjedal (1) Ola Eriksson (2) Stefan Ivanell (2)
(1) Kjeller Vindteknikk AS, Stockholm, Sweden (2) Wind Energy Campus Gotland, Uppsala University, Visby, Sweden

Printer friendly version: printer.gif Print

Poster
Download poster(1.36 MB)

Presenter's biography

Biographies are supplied directly by presenters at OFFSHORE 2015 and are published here unedited

Johannes Lindvall has been working in the wind power industry for 1.5 years and is currently an adviser at Kjeller Vindteknikk AS. He holds an PhD in Atmospheric Science from Stockholm University. After finishing his thesis and before joining the wind industry he spent 4 years doing research on boundary layer cloudiness and the Arctic surface energy balance at the NASA Jet Propulsion laboratory, Pasadena and at the Department of Meteorology Stockholm University.

Abstract

Simulating wind farms in the weather research and forecasting model, resolution sensitivities

Introduction

The numerical weather prediction model WRF has recently included a parameterization that accounts for wind farms (Fitch et al., 2012) by including a turbine drag coefficient. The wind farm is felt by the model as a sink of the resolved atmospheric momentum. The energy extracted by the wind farm is divided into electric energy and turbulent kinetic energy (TKE), following a specified power coefficient. While the fraction of the resolved atmospheric momentum that is extracted is given by specified thrust coefficients. In this study we investigate the resolution dependency of the wind farm parameterization.

Approach

Lillgrund wind farm, located between Copenhagen and Malmö, are simulated using WRF for a 10-day period of prevailing southwesterly winds. Simulations are conducted with and without WRF’s built-in wind farm parameterization for a Control simulation and for two sensitivity simulations (HiVert and HiHor). In HiVert the number of vertical layers below 250m is increased from 5 to 10, and in HiHor the horizontal resolution is increased from 1km to 333m. The normalized production at individual turbines is compared with operational data and simulated changes in velocity deficit and TKE are analyzed. The velocity deficit is also compared to LES data.

Main body of abstract

The mean free wind speed at the hub height during the simulation is approximately 9.4 m/s. In the Control and HiHor simulation the velocity deficit is about 2.5 m/s 4 km downstream the wind farm and approximately 0.5 m/s 18 km downstream. In the HiVert-simulation the wake recovery is considerably faster and the velocity deficit is approximately 0.8 m/s and smaller than 0.3 m/s at a downstream distance of 4 and 18 km, respectively. The faster wake recovery in the HiVert is attributed to sharper vertical gradients that generate more efficient mixing of momentum from above.
The largest difference between the Control and HiHor simulations is seen in the normalized production of the individual turbines. In the event that several turbines are contained in the same WRF grid point, the drag from each of the turbines, regardless of their internal orientation, will be the same. With internal turbine spacing as small as 6.6 rotor radii (approx 300 m) in some directions, the Lillgrund wind farm is considered very dense. In our Control simulation there are grid points containing 10 turbines, in the HiHor simulation the maximum number of turbines in any grid point is 2. In terms of normalized production the HiHor simulation results in more reasonable and realistic wake effects internally in the wind farm.


Conclusion

Our analysis of the wind farm sub-grid parameterization in the WRF-model highlights that internally wake effects as well as the characteristics of the downstream wake depend on the horizontal and vertical resolution of the model setup.


Learning objectives
The WRF built-in wind farm parameterization can be a useful tool to study meso-scale effects of a wind farm, e.g. wake interactions between parks. Since the scheme runs online with the model, two-way interactions between the wake and the planetary boundary layer will be captured. However, our study highlights the resolution sensitivity of the wind farm parameterization and that care is needed in setting up the model experiments.