Wind of change

Wind power divides opinion perhaps more than any other renewable energy. Consequently, whenever possible, wind farm developers attempt to site turbines in uninhabited areas.

The problem with this is that uninhabited areas are frequently hilly, where wind conditions are difficult to predict. So a project run by Spain’s national renewable energy centre (CENER) recently looked into applying wind calculation software to assess the best location for wind farms in hilly areas.

While some proprietary software does exist within the wind industry it is generally not particularly well suited to hilly terrain.

The Spanish team turned to CFD (Computational Fluid Dynamics) software which is more commonly associated with the aerodynamic design of car bodies or boat hulls.

Javier Villanueva, project manager of CENER’s wind turbine department explained that the team chose CFD because of its ability to simulate wind conditions for any type of terrain.

Historically, wind farm developers looking to evaluate the suitability of specific sites have turned to software developed specifically for the wind industry. However, these proprietary tools can be unreliable in hilly areas, claimed Villanueva.

‘The experience some years later has shown that there’s an important difference between the experimental annual energy produced by a wind farm and what the designer expected.

This difference increased with the complexity of the topography,’ he claimed.

Villanueva’s team discovered that unlike WASP, a CFD model can take into account the topographical effects as well as the roughness of different kinds of terrain.

Such parameters modify the wind speed profile and consequently the kw/h produced by turbines. It can also predict the effect that the wake from one turbine can have on its neighbours.

The team evaluated the benefits of CFD during a year-long project based in Navarra in northern Spain. Wind modelling was carried out using Fluent 6.1, and wind data was obtained from seven masts placed over an area of 24sq km at a height of 1050m to collect wind speed and direction data.

Correlations from two measurement stations were used to validate the CFD predictions for average wind speed and power density. Villanueva said that for comparison purposes, the same simulations were carried out with WASP software.

The results showed that the total error using CFD was approximately seven per cent. This, said Villanueva, was a vast improvement over traditional methods which are typically out by between 15 and 20 per cent.

Velocity contours on a plane slicing through the highest peak in the landscape show a variety of wind speeds near ground level.

Velocity contours are more uniform in regions where the topographical variation is less.

‘CFD uses dynamics equations which improve the accuracy of the results. In the real complex terrain studied, improvement in wind assessment was achieved,’ he said.

Villanueva believes that CFD can now be used with confidence to evaluate wind conditions when a wind farm is being planned.

‘The developed methodology could be useful at each step of the building process,’ he said. ‘By combining the CFD model with other calculation software, a better estimation may be made at the design stage so the financial risk can be reduced.’