Nonlinear Wave Loads, and Resulting Global Response Statistics of a Semi-Submersible Wind Turbine Platform with Heave Plates

Abstract

Floating wind turbines (FWTs) are considered to be a promising way to harness the energy from winds over deep water and farther offshore. However, there are some challenges to bring this technology to full maturity. In deep water, FWTs may be exposed to harsh environments and steep waves which induce highly nonlinear wave loads. Given that the natural frequencies of FWTs are designed to be outside the wave frequency range, these nonlinear wave loads can excite eigenfrequencies of FWTs, leading to larger dynamic responses that strain the mooring system or to structural vibrations. However, engineering tools limit hydrodynamic modeling to linear and weakly nonlinear models, and underpredict the dynamic responses of FWTs, especially at the low- and high-frequency regions. Therefore, well-validated modelling tools are needed to capture these nonlinear wave loads and resulting global responses more accurately while keeping the computational efficiency at a reasonable level. The focus is on semi-submersible FWTs due to their wide applicability across a range of water depths. In this thesis, a computational fluid dynamics (CFD) model (OpenFOAM) and an engineering model based on potential-flow theory with Morison-type drag (SIMA) are developed to investigate nonlinear wave diffraction and radiation loads on the DeepCwind semi-submersible FWT. Then, the estimated second-order difference-frequency wave load quadratic transfer functions (QTFs) and frequency-dependent added mass and linearized damping from the CFD simulations with turbulence model are used to improve the engineering model. The nonlinear wave loads and resulting global responses estimated from the CFD model, the original and modified engineering models are validated against experimental measurements. Compared to the experimental measurements in regular waves, the CFD model gives better estimations for the higher order wave diffraction loads, especially for the CFD with turbulence model. The SIMA model has large discrepancies in predcition of amplitude of higher order wave diffraction loads. For the difference-frequency wave diffraction loads, CFD and SIMA agree well at the lower frequencies, while CFD predicts larger wave loads at higher wave frequencies. Additionally, large discrepancies in the phases are found for both high order and difference-frequency wave diffraction loads. The modified engineering model reduces the underprediction of low-frequency wave diffraction loads compared to the original engineering model and CFD with a laminar flow model. The low-frequency added mass derived from the CFD simulation is generally around 12\% larger than that estimated by the potential flow theory. This additional added mass in the CFD simulation is due to viscous effects. The linearized damping shows a small dependence on the oscillation period and a larger dependence on the oscillation amplitude near resonant frequencies of the DeepCwind semi-submersible FWT. At these frequencies, radiation damping is completely negligible compared to the viscous damping, and the accuracy of Morison’s drag forces in capturing the viscous damping is sensitive to the drag coefficient. In the free decay tests, the modified engineering model predicts natural periods close to the experimental results, and the underprediction of the damping is reduced compared to the original engineering model. The low-frequency motions, mooring line tensions and tower-base loads response to an irregular wave are underestimated using the original engineering model. The additional linear damping estimated by matching the decay motions from the CFD simulations increases this underestimation, while the modified QTFs based on CFD simulations result in larger low-frequency responses. The overestimation is reduced by modifying the frequency-dependent damping at the same time and the best agreement with the experimental measurements is achieved. Meanwhile, the combined modifications give improved agreement with experimental data in terms of damage equivalent loads for the mooring lines and tower base.