Method for estimating wind turbine production losses due to icing

Atmospheric icing causes production losses to wind turbines, poses a risk of ice throw and increases dynamic loading of wind turbine which might reduce the lifetime of turbine components. These effects are qualitatively widely known but not quantitatively. In order to estimate effects of icing to power production of a typical 3 MW wind turbine, a simulation based study was made. Three rime ice cases were selected with the meteorological conditions typical for Finnish climate. Conditions were the same for each case. The lengths of the icing events were varied to represent different phases of an icing event; beginning of icing, short icing event and long lasted icing event. Thus, three different ice masses accreted on a wind turbine blade were simulated. Accretion simulations were performed with a VTT inhouse code TURBICE. The aerodynamical properties of the iced profiles were modelled using computational fluid dynamics (CFD) with ANSYS FLUENT flow solver. As a result, lift and drag coefficients were drawn as a function of an angle of attack. Small-scale surface roughness effect on drag coefficient was determined analytically. Finally, the power curves were generated with FAST software for the clean wind turbine and for the same turbine with the different ice accretions. The results showed relatively large impact of small-scale surface roughness on power production. In the beginning of an icing event, where ice causes basically only increased surface roughness, power production was discovered to reduce by approximately 17 % below rated wind speeds compared to the no-ice case. As the ice mass was increased, production reduction was 18 % for short icing event and 24 % for long lasted icing event. However, the relative reduction was smaller than in the beginning of icing, mainly due to the small-scale surface roughness that remained at the same level. The results indicate that the surface roughness is crucial to take into account when defining the aerodynamic penalty caused by icing of wind turbine blades. The generated power curves for iced up wind turbine in this study estimate higher power production than can be expected from the observations in real life. This is caused by the less complex ice shapes resulting from simulating only rime ice conditions. The results of this study were used in the Finnish Icing Atlas (2012), where time dependent numerical weather simulations were carried out to calculate icing conditions and energy production losses. The results were also used in ICEWIND project for production loss estimation process.