Dynamic behaviour of floating wind farms with shared mooring in waves - Part II: Modal description and mooring line model fidelity

• Modal interpretation of cyber-physical test and numerical simulation results for shared mooring. • 5-9 modes capture the line-tension standard deviation of selected lines within 15% in the studied farms. • Quasi-static mooring models provides good approximation of floater horizontal motions in shared mooring but shared line tensions can be underestimated by up to 80% • Reconstructing line tensions using measured platform displacements and a dynamic mooring line model for better accuracy. • Modal approach explains comb-like frequency response in the power spectral density of line tensions and modal responses. Cyber-physical test and numerical simulation results for two shared mooring farms, consisting of 2 and 9 floating wind turbines, are analysed using a modal basis. Decay tests show a single dominant modal response. In irregular wave tests, comparisons with numerical simulations are limited to 5-9 modes that contribute significantly to line tension. These modes could reproduce the selected line tension standard deviation within 15%. The power spectral density of shared line tension and modal responses in the wave-frequency range exhibits a distinct comb-like frequency response associated with modal excitation from first-order wave loads. To meet the real-time computational requirement of cyber-physical tests, the mooring line model used during the test was a quasi-static model. Numerical comparisons are made between simulations with quasi-static and dynamic mooring line models for the tested shared-mooring farms to assess the impact of this simplification. In irregular wave tests, the quasi-static mooring line model overestimated the horizontal translational motion standard deviation by up to 20% (0.29 m full scale). Although direct application of the quasi-static mooring model underestimated the shared line tension standard deviation by up to approximately 80%, line tensions can be accurately reconstructed during post-processing.