Abstract
Floating offshore wind turbines (FOWTs) enable wind energy deployment in deep waters with abundant wind resources. However, many potential deployment sites in cold regions are seasonally or permanently covered by sea ice, which introduces complex ice–structure interaction processes and additional risks to the structural integrity, station-keeping performance, and operational reliability of FOWTs. Compared with bottom-fixed offshore wind turbines, FOWTs are compliant systems that can experience large-amplitude motions in six degrees of freedom, making their response to ice loads significantly different and less understood. Therefore, it is necessary to evaluate the dynamic response of FOWTs deployed in ice-covered seas.
Various ice load prediction models have been proposed for marine structures. However, their high computational cost and limited capability to represent the full range of ice failure modes make them impractical for time-domain global response analyses of FOWTs under long-duration stochastic wind and ice loads. In this thesis, the dynamic response of FOWTs subjected to ice loads is investigated through the development and application of a coupled numerical framework. The framework integrates the wind turbine global response simulation tool SIMA with a semi-empirical ice load prediction model that considers both ice bending and crushing failure modes. The coupling between SIMA and the ice model is implemented through a dynamic-link library (DLL), enabling time-domain simulations that account for aerodynamic, hydrodynamic, structural, and ice-induced loads.
A feasibility study of a 5 MW spar-type FOWT shows that ice thickness and wind–ice velocity misalignment angle have significant effects on the structural response of the system. Ice loads can induce pronounced low-frequency platform motions; however, motion mitigation measures such as ice-breaking cones and optimized mooring layouts can effectively reduce the dynamic response. The results indicate that the OC3 spar platform operates within acceptable motion limits under the range of ice conditions suggested in the ISO 19906 standard for the Baltic Sea.
As a key component of FOWTs, the mooring system governs the station-keeping performance and strongly influences the global dynamic response of the system. The behaviour of the mooring system of the 5 MW spar-type FOWT under extreme wind and ice conditions in the Gotland Basin is further studied through ultimate limit state (ULS) analyses. The results show that a modified mooring configuration can survive extreme environmental conditions. In addition, optimized mooring line lengths can satisfy ULS requirements while reducing system cost.
A comparative analysis between a 15 MW semi-submersible FOWT and a 20 MW spar-type FOWT is also conducted under identical environmental conditions in the Gotland Basin. The results indicate that the 20 MW spar-type FOWT is more suitable for deployment in ice-covered seas, such as the Gotland Basin, in terms of platform motion performance.
Finally, a two-stage long short-term memory (LSTM) framework is developed to predict the dynamic response of a 20 MW spar-type FOWT subjected to ice loads. The proposed machine-learning model successfully captures the nonlinear relationship between environmental loads, platform motions, and mooring line tensions. The LSTM-based
framework demonstrates strong potential for efficient response prediction and digital-twin applications.
Overall, this thesis improves the understanding of ice–FOWT interaction and provides a computational framework for evaluating the structural performance and design of floating wind turbines operating in cold regions.