This secure area is password protected and session limited, if you leave you may be asked for your password again.

WORKSHOP ON WAKE MODELLING AND BENCHMARKING OF MODELS

6 and 7 September 2006
Hotel Legoland, Billund, Denmark
Høvsøre Test Site

Summary and conclusion

In large wind farms with standard layout wake losses are expected to be 10-20% of total power output and wake generated turbulence significantly exceeds ambient turbulence which is critical for design loads. There is therefore considerable interest in accurately modelling wakes prior to wind farm construction to optimize wind farm layouts and for examining wakes during wind farm operation in order to evaluate optimal control strategies. Wake models developed during the 1980s were mainly related to onshore single or double wake situations.

Evaluations of state of the art for wake models have indicated large differences between the models and also between models and experiments. Also, there exist theoretical and experimental evidence of significant differences from onshore to offshore conditions. These divergences were evident from the contributions and discussions at the wake workshop where 27 participants representing the scientific world, manufacturers, developers and power companies met to review the knowledge of today and to look for possibilities for improvement of models by comparing the models in a future benchmarking process.

A variety of modelling efforts were presented at the workshop, including analyses of data not least from the two large demonstration wind farms in Denmark, Horns Rev and Nysted. The modelling efforts range from CFD studies of the wake flow structure and Ainslie-type wake modelling for wind speed deficit to simpler but extensive modelling of complete wind farms and their interaction with the atmosphere. The aim of such research spans from pure research to the development of engineering tools for determination of array efficiency and structural loads.

These and many other interesting results on structural loading and wind farm layout as separate subjects were presented before the final discussion on benchmarking of wake models. The theoretical and experimental results indicated differences in results from modelling and in the interpretation of measuring data from offshore wind farms, in particular data from the Horns Rev and Middelgrunden wind farms.

The presentations and the discussions indicated a great need for further collaboration and exchange of data in order to develop and verify computational models and to understand the physics of wakes and meteorological backgrounds.

The EU R&D project UPWIND includes similar activities to the ones discussed at the workshop and coordination is suggested to take place between the members of IEA Task 23 and partners of UPWIND.

In conclusion it was agreed to prepare a format for collecting and benchmarking data related to offshore wind farms and onshore farms in cases when it is considered relevant. The format will be suggested to the participants in the workshop and member countries of Task 23. The collaboration will be focussed on data which are important for power calculations as well as design loads. Within the next year the benchmarking experience and the results obtained from the continued collaboration also with UPWIND will be analysed and discussed at a second workshop.