Investigating the impact of wakes on single turbine availability

Abstract

The current trend in wind energy leads to closely packed wind farms. These short distances result in a higher turbulence and consequently higher loads. In this paper an approach is proposed to evaluate single turbine availability losses due to the wake effect by modeling material fatigue behavior of selected turbine components. First, a characteristic wake scenario is defined: While one turbine is exposed to a wind field conforming to wind class IIB, another turbine experiences higher turbulence intensities according to Frandsen’s wake model. Using a dynamic multibody simulation, turbine loads are calculated which then are analyzed using the rainflow counting algorithm (giving load range and cycle counts) before being extrapolated to design lifetime of 20 years. Subsequently, artificial material curves are estimated representing the minimum load cycles to withstand during design lifetime. The slope of these curves is determined using the slope of a representative S-N-curve for each component. Position is then determined applying Miner’s rule with unity damage at the end of design lifetime. Lastly, failure behavior is approached by using a lognormal probability distribution. As a result increased failure probabilities for the rotor hub unit (main bearing and rotor hub) are found in the wake case. First results for an annual time in wake fraction of 20 % show an increase in total failure probability from 2.30 % to up to 41.85 % at a distance between the two turbines of 3D. Applying a mean down time per failure from literature (based on modern EU wind farms) finally leads to an overall technical availability decrease due to hub unit failure of 0. 035 % in the wake case compared to the free stream case.