Stochastic Dynamic Response Analysis of a Tension Leg Spar-Type Offshore Wind Turbine

Floating wind turbines can be the most practical and economical way to extract the vast offshore wind energy resources at deep and intermediate water depths. The Norwegian Ministry of Petroleum and Energy is strongly committed to developing offshore wind technology that utilises available renewable energy sources. As the wind is steadier and stronger over the sea than over land, the wind industry recently moved to offshore areas. Analysis of the structural dynamic response of offshore wind turbines subjected to stochastic wave and wind loads is an important aspect of the assessment of their potential for power production and of their structural integrity. Of the concepts that have been proposed for floating wind turbines, spar-types such as the catenary moored spar (CMS) and tension leg spar (TLS) wind turbines seem to be well-suited to the harsh environmental conditions that exist in the North Sea. Hywind and Sway are two examples of such Norwegian concepts; they are based on the CMS and TLS, respectively. Floating wind turbines are sophisticated structures that are subjected to simultaneous wind and wave actions. The coupled nonlinear structural dynamics and motion response equations of these turbines introduce geometrical nonlinearities through the relative motions and velocities. Moreover, the hydrodynamic and aerodynamic loading of this type of structure is nonlinear. A floating wind turbine is a multibody aero-hydro-servo-elastic structural system; for such structures, the coupled nonlinear equations of motion considering nonlinear excitation and damping forces, including all wave- and wind-induced features, should be solved in the time domain. In this thesis, the motion and structural responses for operational and extreme environmental conditions were considered to investigate the performance and the structural integrity of spar-type floating wind turbines. The power production and the effects of aerodynamic and hydrodynamic damping, including wind-induced hydrodynamic and wave-induced aerodynamic damping, were investigated. Negative damping adversely affects the power performance and structural integrity. In this thesis, the controller gains were tuned to remove servo-induced instabilities. The rotor configuration effect on the responses and power production was investigated by comparing the upwind and downwind turbines. To develop robust design tools for offshore wind power, the competencies of the offshore technology and wind technology must be combined. Both the offshore and wind energy industries have begun to extend their existing numerical codes to account for the combined aerodynamic and hydrodynamic effects on the structure. As a result verifications of extended codes by doing experiments and code-to-code comparisons are needed. One of the aspects of the present research was to fill this gap by performing hydrodynamic and hydro-elastic comparison between commercial codes. For both CMS and TLS concepts, the comparisons were carried out prior to using the tools to study the behaviour of the CMS and TLS under wave- and wind-induced loads. Offshore structures encounter a variety of operational and harsh environmental conditions. Limit states such as ultimate, fatigue, accidental collapse and serviceability limit states (ULS, FLS, ALS and SLS) are defined as the design criteria for offshore structures. In performing realistic ultimate limit state analysis, the extreme responses of a floating wind turbine over its life should be estimated. This estimation requires detailed analysis of the extreme response. In the present thesis, extreme value analysis for spar-type wind turbines subjected to simultaneous wave and wind actions was preformed. The structural responses and the effect of modelled forces such as turbulence on these responses were investigated. The joint distribution of the environmental characteristics of the wave and wind was applied through the contour surface method. Stochastic wave and wind analysis showed that, while rigid body modelling was sufficient for obtaining accurate motions, consideration of the elastic behaviour of the tower/support structure was necessary to predict structural responses. The blades structural responses were found to be significantly affected by the turbulent wind. However, the mean and standard deviation of global motion and structural responses were not affected by the turbulence. Thus, to reduce the simulation time in fatigue analysis, a constant wind speed model can be applied. The CMS and TLS wind turbines are inertia-dominated structures, and the hydrodynamic viscous drag did not affect their wave-induced responses, while an increase in viscous drag could effectively reduce the resonant responses of such turbines. Under operational conditions, aerodynamic damping was found to be active in reducing both wave frequency and resonant responses. The results showed that, for a floating wind turbine, extreme response could occur in survival conditions, while for a fixed wind turbine, the extreme response occurs in operational cases related to the rated wind speed. To estimate the extreme value responses, extrapolation methods were used to reduce the sample size in Monte Carlo simulations. The accuracy of methods to estimate the extreme responses as a function of sample size and methods applied was investigated. The normalized responses for both CMS and TLS offshore wind turbines were presented to draw more generalized conclusions.