Realization and Characterization of a Research Platform for the Use of Wind Energy in Complex Terrain in Southern Germany (WINSENT)

Principal Investigator:
Stefan Emeis

Institute for Meteorology and Climate, Atmospheric Environmental Research, Karlsruhe Institute of Technology

Local Project ID:

HPC Platform used:
SuperMUC and SuperMUC-NG of LRZ

Date published:

Hilltops are of particular interest for the wind energy industry, since the flow is sped up as it passes over an obstacle. However, locations in complex terrain pose many problems with regards to predictability of the wind and the impact of turbulence on the generation of electric power and material fatigue. The WINSENT test-site is now under construction to provide a platform at which research can be done to answer these questions. It is located at the Swabian Alb near a 200 m high terrain edge. It consists of two 750 kW wind turbines with a hub height of 75 m, surrounded by four 100-m high towers at which meteorological measurements are taken.

Regional weather models typically run at mesh sizes of 1 km. At that resolution, these models are unable to resolve the terrain adequately for which reason the simulated mean flow is erroneous. Also, these models do not resolve turbulence, a critical variable for wind energy. A much finer mesh (~100 m) is required for a better simulation of the flow field around the test site. Simulations with such fine meshes are computationally costly (approx. 104 core hours per single day), especially if one is interested in long-term simulations. However, even though such simulations are much better than mesoscale models, they still cannot fully resolve the turbulence nor is it possible to resolve the impact of the wind turbine on the flow. It is for this reason desirable to create a model chain by coupling this model with CFD models with an even finer grid.

Key goals of this project are:

  • To perform and validate model simulations to create the first link of the model chain.
  • To assess the predictability of the flow at the test-site.
  • To deepen the knowledge of the generation and transport of turbulence and its impact on wind turbines in complex terrain.

To assess the flow conditions at the test-field, we performed a simulation over a period of two months using the Weather Research and Forecasting model (WRF). This model uses multiple domains with increasingly fine meshes to arrive at a mesh size of 150 m in the horizontal and 15 m in the vertical for the innermost domain. A statistical analysis of this simulation is shown in fig. 1. It demonstrates that the wind direction is simulated well, whereas the higher wind speed (10-20 m/s) is overrepresented in the model compared to observations.

To investigate the source of this discrepancy we simulated specific days of that period in large eddy simulation model with a mesh size of 50 m. We were able to show that the main source of the bias is due to the drag caused by the nearby forest. This drag is poorly accounted for in WRF and a parameterization of this effect after Shaw and Schumann (1992) had to be implemented. Figure 2 shows streamlines of the horizontal flow over the escarpment at 60 m above ground (black) and at 190 m above ground (blue) for a simulation without (left) and with forest parameterization (right).

The increased drag not only leads to a reduction of the wind speed at 60 m, but in some locations causes a change in wind direction of up to 90°. The forest parameterization successfully reduces the positive bias of the wind speed in 100 m above ground, but leads to a negative bias (i.e. an overcorrection) in 45 m a.g. and below. A revision of the parameterization would be desirable.

Improvements of meteorological models over complex terrain help of identify suitable locations for wind turbines when coupled with RANS or LES models. Including forest drag is vital for accurate simulations both in flat terrain and over hills.

A publication is in preparation at the moment.


Shaw, Roger H., and Ulrich Schumann. 1992. “Large-Eddy Simulation of Turbulent Flow above and within a Forest.” Boundary-Layer Meteorology 61 (1): 47–64. https://doi.org/10.1007/BF02033994.


PI: Prof. Dr. Stefan Emeis (Professor of Meteorology at the University of Cologne), Institute for Meteorology and Climate, Atmospheric Environmental Research, Karlsruhe Institute of Technology (KIT)

Daniel Leukauf, PhD, Institute for Meteorology and Climate, Atmospheric Environmental Research, Karlsruhe Institute of Technology (KIT)

Scientific Contact

Daniel Leukauf, PhD
Karlsruhe Institute of Technology
Institute for Meteorology and Climate, Atmospheric Environmental Research
Kreuzeckbahnstrasse 19, D-82467 Garmisch-Partenkirchen (Germany)
e-mail: daniel.leukauf [@] kit.edu

Local project ID: pr27po

July 2020

Tags: LRZ KIT Environmental Science Meteorology