Numerical Prototyping of Floating Offshore Wind Turbines: Virtual Operation in Real Environmental Conditions

Abstract

Abstract The numerical fatigue analysis of floating offshore wind turbines (FOWTs) must account for the environmental loading over a typical design life of 25 years, and the stochastic nature of wind and waves is represented by design load cases (DLCs). In this statistical approach, combinations of wind speeds and directions are associated with different sea states, commonly defined via simplified wave spectra (Pierson-Moskowitz, JONSWAP), and their probability of occurrence is identified based on past observations. However, little is known about the difference between discretizing the wind/wave direction bins into (e.g.) 10deg bins rather than 30deg bins, and the impact it has on FOWT analyses. In addition, there is an interest in identifying the parameters that best represent real sea states (significant wave height, peak period) and wind fields (profile, turbulence) in lumped load cases. In this context, the aim of this work is to better understand the uncertainties associated to wind/wave direction bin size and to the use of metocean parameters as opposed to real wind and sea state conditions. A computational model was developed in order to couple offshore wind turbine models with realistic numerical metocean models, referred to as numerical prototype due to the highly realistic wind/wave conditions in which it operates. This method allows the virtual installation of FOWTs anywhere within a considered spatial domain (e.g. the Mediterranean Sea or the North Sea) and their behaviour to be evaluated in measured wind and modelled wave conditions. The work presented in this paper compares the long-term dynamic behaviour of a tension-leg platform (TLP) FOWT design subject to the numerical prototype and to lumped load cases with different direction bin sizes. Different approaches to representing the wind filed are also investigated, and the modelling choices that have the greatest impact on the fidelity of lumped load cases are identified. The fatigue analysis suggests that 30deg direction bins are sufficient to reliably represent long-term wind/wave conditions, while the use of a constant surface roughness length (as suggested by the IEC standards) seems to significantly overestimate the cumulated damage on the tower of the FOWT.