Titel: Mapping of Icing for Wind Turbine Applications – A Feasibility Study, Elforsk rapport 08:40
Författare: Göran Ronsten
Utgivare: Elforsk, Vindforsk
Årtal: 2008
Ämnesord: Vindkraft i kallt klimat
Sökord/keywords: COST 727, de-icing
Rapport
Sammanfattning/Abstract: Onshore wind energy is currently significantly more cost-effective compared to offshore ditto. Swedish wind farms will therefore, as suitable sites in open terrain are scarcer, have to be developed in turbulent forests and cold climates. The wind energy developers are, however, lacking both verified and detailed mapping of icing as well as commercially available de- or anti-icing systems 3. Icing of wind turbine rotor blades and wind sensors may, in affected areas, lead to low or no energy production during extended periods of time. The safety risk, caused by ice being thrown from the blades is not only important to consider, but may also lead to prohibition of operation with icedup blades. Icing of rotor blades can also, due to the effects of unbalanced aerodynamic drag/lift and mass, cause a decrease in the useful life of the wind turbine.
The relative humidity, too often believed to be a key parameter for estimating icing, can be used to indicate a risk for icing but cannot be used to calculate the rate of icing on a particular object. Instead, the main parameters for icing, apart from the obvious temperature and wind speed, are the droplet size distribution and the liquid water content of air. Today, automatically and in the field, these two last items can only be measured and estimated indirectly, via visibility and vertical velocity, respectively.
An updraft may increase the risk of icing. Therefore, a sufficiently high terrain model resolution will be a key to more correctly estimate the vertical velocity. For site estimations, topographically induced vertical velocities can’t be seen if the model’s resolution, i.e. the orography4, is too low. Here, an analogy with wind speed estimations can be useful: The rich wind resource on the isolated mountain of Uljabuouda in the Municipality of Arjeplog (SE) can’t even be seen in the 1 km grid used in the most current national wind resource atlas. As a comparison, the icing forecast currently in use by the National Weather Service (SMHI) uses a 5 km grid.
The importance of cloud schemes used by the models and, most lately, the turbulence models in use, are currently being investigated. Without corresponding measurements, it’s hardly possible to develop weather models, which can be believed to correctly estimate droplet size distribution and liquid water content of air. And without this knowledge, it is impossible to estimate the rate of icing on an object. Such measurements are thus critical in order to move research forward.
Icing measurements in tall towers and at nearby wind farm sites can be used to verify and develop weather forecasts and models. Such modelling, being computer intensive at high resolutions and over climatologically relevant time periods, may preferably start on a local scale, i.e. at a < 50 km innermost grid, and from there, after thorough verification, expand to regions and nations. The areas to cover should be prone to icing at heights up to 200 m above ground and of interest for the deployment of future wind farms.
³Currently available de-icing systems (from Enercon) are “only” able to cope
efficiently with light icing conditions.
4Definition of orography: http://en.wikipedia.org/wiki/Orography (2008-11-17)