In the current wind industry, wind turbines and foundations are usually designed by different companies and there is a general problem of sharing wind turbine structure and controller details. As a result, the uncoupled analysis method, sometimes with iterations, has been developed. Nevertheless, whether such a method can be applied to the analysis of floating offshore wind turbine (FOWT) and whether its results are conservative or nonconservative should be studied. In this work, the OO-Star floater with a 12MW wind turbine is selected as an example to make a comparison of the fully-coupled method and the uncoupled analysis. The results show that, if the loads of the wind turbine (which is considered as the external loads) are calculated based on the first-order motion of the floater, the results would give obvious deviation in response spectrum of the motion compared to the fully-coupled analysis, although the second-order wave loads are used in the un-coupled method.
In recent years, the exploration of offshore wind has moved from shallow waters to deep waters, and the floating offshore wind turbine (FOWT) is considered a competitive candidate to harvest wind energy in waters with a depth of more than 50m (Henderson and Witcher, 2010). FOWT is a kind of complex engineering equipment consisting of a multi-subsystem, such as wind turbine, tower, floating foundation and mooring system. It is subjected to a variety of environmental loads during service including wind, wave and current loads. Besides, the floating foundation has rigid body motions. These loads and motions are coupled and nonlinear so that a fully-coupled dynamic analysis is preferable.
For the simulation of the dynamic responses of a FOWT, a number of codes have been successfully developed, such as FAST (Jonkman and Buhl Jr, 2005), HAWC2 (Larsen and Hansen, 2007). Both of them can carry out the fully aero-hydro-servo-elastic coupled analysis of a FOWT, and they are widely used and validated. Haid et al. (Haid et al., 2013) examined the appropriate length of a FOWT simulation with FAST. The results indicated that the loads did not increase for longer simulations and the fatigue analysis was more sensitive to the procedure used for counting half cycles than the simulation length itself.
