Statistical and physical modelling of large wind farm clusters
Abstract
As the first large wind farms are constructed the issue of the effect of large wind farms on local climates is being raised. The main concern currently is that, in some countries, areas in which large offshore wind farms can be constructed over the next 10 to 20 years are fairly limited due to technical and economic constraints. This means that wind farms will be built in clusters of up to 100 wind turbines but within 20 km of the nearest cluster. Theoretical considerations suggest that the effects of a wind farm on a downwind wind farm maybe more noticeable offshore than onshore where higher turbulence assists wind speed recovery. Added to this many offshore areas are dominated by stable and neutral atmospheres where wakes propagate over longer distances than on land where unstable conditions also occur for a significant fraction of the time. On the other hand the large turbulence generated by the wind farm itself may be sufficient to assist wind recovery but possibly provide a higher than expected turbulence at the neighbouring wind farm or cluster. While some progress has been made with single wake modelling offshore, these models have not been evaluated for more than 5 wakes. Hence it is difficult to evaluate the impact of large wind farms and to optimise the spacing of clusters. A new project STORPARK is underway which is using statistical and physical modelling methods to make preliminary estimates of large wind farm impacts. The work described in this paper is a combination of statistical methods using observations from offshore wind monitoring sites at Vindeby/Omø Stålgrunde and Rødsand/Gedser in Denmark to evaluate in the first instance how far the effects of land can be detected on wind speed and turbulence intensity. These results will be compared with model simulations from WAsP and the Coastal Discontinuity Model (CDM) where large wind farms are currently represented by large roughness elements in accord with models developed by Crespo, Frandsen and others. An alternative approach is to represent reduced velocity by a momentum deficit at hub-height which is also being implemented in the CDM.
- Publication:
-
EGS - AGU - EUG Joint Assembly
- Pub Date:
- April 2003
- Bibcode:
- 2003EAEJA.....2475B