Spectral Fatigue and LCOE Assessment of Floating Offshore Wind Turbines

Abstract

Floating offshore wind turbines (FOWT) are an emerging technology that will contribute towards a more sustainable energy system in the near future. Some of the main challenges in an early design stage of floating offshore wind turbines are the estimation of costs, the site selection and analyzing their fatigue damage. These challenges are investigated in this work for the WindStar TLP system. In the first part a LCOE-model from the literature for other FOWT concepts is adapted for the WindStar TLP system. The model calculates the costs as a function of the water depth and the distance to the closest shore and puts them in relation to the annual energy production. By using MetOcean-data as input, the LCOE for possible site locations on a grid of the North Sea area are estimated. The results show that the calculated LCOE of the WindStar TLP system is comparable to the LCOE of other concepts from literature. Furthermore, the results show that, based on the LCOE, good locations for floating wind farms are north of Scotland and west of Denmark. In the second part, four approaches to perform a spectral fatigue assessment using FAST simulations are studied. With the so-called White-Noise approach the fatigue damage for every possible sea state can be calculated with just a very limited number of simulations. The short-term fatigue damage for every approach is calculated using several spectral approximation methods and correction factors for non-Gaussian loads. Comparing the results with results obtained with the rainflowcounting method in the time domain leads to the following main conclusions: It is necessary to use a non-Gaussian correction factor and, if second order wave effects are considered, a spectral fatigue approximation method for widebanded spectra has to be used to get results close to the reference results. It should be noted that wind effects are not considered in the White-Noise approach for this thesis since the focus is on proposing approaches for a quick spectral fatigue assessment with FAST, but not on the detailed estimation of loads.