Fast simulations of wind farms over complex terrains

Background

Offshore and onshore wind farms are nowadays designed by means of simplified methods where the wind is simulated over a terrain and the turbines are introduced by means of analytical methods (wake-model methods). Each turbine is modelled as a wake with prescribed properties, empirically determined from other similar cases. This step is crucial because the wake of each turbine determines the inflow for the downstream turbines and, therefore, the power they can extract. It is known that wakes interaction provides a significant loss in the energy yield of a farm (on the order of 10-30% of the installed power) but, despite this relevant magnitude, the error of these methods is subjected to uncertainties of the order of 25%.

Computational Fluid Dynamics (CFD) is a valuable alternative to wake models. In CFD the flow equations are solved with the turbines and a more accurate result is expected. However, the cost of the CFD simulation is large for a given wind direction and magnitude (on the order of hours to days) while wake models run in less than a second on a regular computer. When considering the number of tests needed to design a wind farm, it is clear that industry still prefers using the wake-model method, although better tools are required. Furthermore, wake models do not affect the wind field more than what due to the wake presence, prohibiting to characterize several cases where significant two-way interactions can take place, such as in presence of special atmospheric conditions

Aims

This project aims to develop a new solver, based on approximated flow equations, that is faster than traditional CFD simulations and more accurate than wake models. The code is called ORFEUS and it is two orders of magnitude faster than CFD. It can perform several types of simulations, including the effect of atmospheric stratification and gravity waves, providing a useful tool to perform fast simulations for fundamental and applied research on wind power.

Approaches

The code is written in Python as a spectral code. The turbines are simulated as body forces distributed over the rotor area following the actuator-disk method. Various terrain properties can be accounted in the simulations, including topography, roughness and even forest, an aspect of crucial importance in the Nordic countries. ORFEUS has been tested already in many complex wind farms with good results when comparing the farm power performance with simulations from ORFEUS and traditional CFD.

Keywords

Wind turbine simulation, Fast linearized methods, Gravity waves

Links

Link to research article ”Linearized simulation of flowover wind farms and complexterrains

Link to article ”Wind-farm simulation over moderately complexterrain

Link to the article ”Wind power in forests II: Forest wind”

Link to the article ”Data analysis and simulation of the Lillgrund wind farm”

Project leaders

Antonio Segalini, KTH Royal Institute of Technology, FLOW and Department of Earth Sciences, Uppsala University