In this thesis a model is proposed for establishing the coupled analysis of a spar floating wind turbine considering both ice and aerodynamic loads. This topic is important within the field of renewable energies research given that wind energy has known one of the fastest growths among renewable energies. And, in the case of cold climate regions such as the Baltic Sea, ice loads become an important point to consider in the design of offshore wind turbines. The central issue to be addressed within this work is the action of ice and aerodynamic loads on a spar floating wind turbine which is of relevance in determining the design requirements for structural checking of such structure. The aim is the diagnosis of the factors relevant to the spar floating wind turbines design and the investigation of their potential for inducing significant dynamic structural responses.A numerical model for ice loads calculations has been implemented in the aero-hydro-servo-elastic simulation tool HAWC2 using a Fortran module. The work has been derived from Xian Tan s thesis and papers (Tan, et al., 2013) and Wei Shi s work (Shi, et al., November 24-26, 2014). The ice loads are determined by defining the structure and ice sheet geometry at the mean sea level and then by integrating the contact loads over the waterline.First, an eigenfrequency analysis and a convergence study have been conducted to gain knowledge on the system and the simulation settings. Then, the developed model was applied in order to determine the coupled action of wind and ice loads along with the effect of ice drifting speed and thickness variations. The simulations were defined for ice conditions corresponding to the ones encountered in the Baltic Sea. These results are compared to the results obtained with a decoupled analysis realized in a previous work to state on the goodness of the model applied.The application of the developed model to the coupled analysis of a spar floating wind turbine considering both ice and aerodynamic loads has shown that ice thickness is of critical importance in the determination of the dynamic response while ice drifting speed does not seem to have a significant influence. It is explained by the direct link between the ice loads value and the contact area between the ice sheet and the structure. Indeed, thicker ice will leads to a larger contact area for the same ice drifting speed and as a result to higher loads. Thus, these results are in agreement with the accepted knowledge within ice loads studies. Coupled and decoupled models present similar output shapes but they differ in magnitude. This difference increases for increasing ice drifting speed and ice thickness. However, a trend in the divergence is hard to identify. Then, the simulations performed including both ice and wind loads have shown that the wind has a predominant influence on the loads. But, ice loads participate to the dynamic component of the response by causing amplified oscillations around the mean value. Thus, this could have a significant influence in the lifetime of the wind turbine by accelerating fatigue damages. However, the power production does not seems to be significantly impacted, at the rated speed at least. The results achieved are not providing an extensive enough basis to state on the relative importance of ice loads in regards to aerodynamic loads. However, it is a good first insight of the subject and knowledge was gained in the simulation settings that will be a good asset in the future. Due to convergence problems in the module and the time needed to run a full simulation, only a restricted number of cases where tested and this work should be continued to obtain more extensive data and thus draw more accurate conclusions. During this investigation, the possibility offered by the coupled model to run analysis of a spar floating wind turbine considering both ice and aerodynamic loads where demonstrated. Moreover, this work has given a first validation on the settings to apply through a convergence study on both simulation time and time step influences. The cases including both wind and ice should be investigated further to allow longer simulations. It would be necessary to also complete the Fortran code to include randomly varying ice conditions. This way, the simulations would be run in more realistic conditions varying ice properties along the ice sheet and turbulent wind. A possible continuation of this work could be to include a fatigue module and look more closely on the influence of the ice loads in the energy production. Besides, now that the model gives long enough time simulations and thus stable results, it would be necessary to assess the real quality of this model by comparing the numerical results to model tests or full scale data. Nonetheless, this work demonstrate that future modelling design improvements for floating wind turbines are possible.
展开▼
