State of the Art and Trends in Wind Resource Assessment
Received: 1 April 2010 / Revised: 4 May 2010 / Accepted: 21 May 2010 / Published: 3 June 2010
Cited by 15 | PDF Full-text (1064 KB) | HTML Full-text | XML Full-text
Given the significant rise of the utilization of wind energy the accurate assessment of the wind potential is becoming increasingly important. Direct applications of wind assessment techniques include the creation of wind maps on a local scale (typically 5 20 km) and
[...] Read more.
Given the significant rise of the utilization of wind energy the accurate assessment of the wind potential is becoming increasingly important. Direct applications of wind assessment techniques include the creation of wind maps on a local scale (typically 5 20 km) and the micrositing of wind turbines, the estimation of vertical wind speed variations, prospecting on a regional scale (>100 km), estimation of the long-term wind resource at a given site, and forecasting. The measurement of wind speed and direction still widely relies on cup anemometers, though sonic anemometers are becoming increasingly popular. Moreover, remote sensing by Doppler techniques using the backscattering of either sonic beams (SODAR) or light (LIDAR) allowing for vertical profiling well beyond hub height are quickly moving into the mainstream. Local wind maps are based on the predicted modification of the regional wind flow pattern by the local atmospheric boundary layer which in turn depends on both topographic and roughness features and the measured wind rose obtained from one or several measurement towers within the boundaries of the planned development site. Initial models were based on linearized versions of the Navier-Stokes equations, whereas more recently full CFD models have been applied to wind farm micrositing. Linear models tend to perform well for terrain slopes lower than about 25% and have the advantage of short execution times. Long-term performance is frequently estimated from correlations with nearby reference stations with concurrent information and continuous time series over a period of at least 10 years. Simple methods consider only point-to-point linear correlations; more advanced methods like multiple regression techniques and methods based on the theory of distributions will be discussed. Both for early prospecting in regions where only scarce or unreliable reference information is available, wind flow modeling on a larger scale (mesoscale) is becoming increasingly popular.
(This article belongs to the Special Issue Wind Energy