Multibody modeling for concept-level floating offshore wind turbine design

Abstract

AbstractExisting Floating Offshore Wind Turbine (FOWT) platforms are usually designed using static or rigid-body models for the concept stage and, subsequently, sophisticated integrated aero-hydro-servo-elastic models, applicable for design certification. For the new technology of FOWTs, a comprehensive understanding of the system dynamics at the concept phase is crucial to save costs in later design phases. This requires low- and medium-fidelity models. The proposed modeling approach aims at representing no more than the relevant physical effects for the system dynamics. It consists, in its core, of a flexible multibody system. The applied Newton–Euler algorithm is independent of the multibody layout and avoids constraint equations. From the nonlinear model a linearized counterpart is derived. First, to be used for controller design and second, for an efficient calculation of the response to stochastic load spectra in the frequency-domain. From these spectra the fatigue damage is calculated with Dirlik’s method and short-term extremes by assuming a normal distribution of the response. The set of degrees of freedom is reduced, with a response calculated only in the two-dimensional plane, in which the aligned wind and wave forces act. The aerodynamic model is a quasistatic actuator disk model. The hydrodynamic model includes a simplified radiation model, based on potential flow-derived added mass coefficients and nodal viscous drag coefficients with an approximate representation of the second-order slow-drift forces. The verification through a comparison of the nonlinear and the linearized model against a higher-fidelity model and experiments shows that even with the simplifications, the system response magnitude at the system eigenfrequencies and the forced response magnitude to wind and wave forces can be well predicted. One-hour simulations complete in about 25 seconds and even less in the case of the frequency-domain model. Hence, large sensitivity studies and even multidisciplinary optimizations for systems engineering approaches are possible.