Since the introduction of the first offshore wind farm in 1991, the demand for offshore renewable wind energy has experienced exponential growth all around the world. To supply this demand, the power rating and corresponding dimensions of offshore wind turbines have grown significantly. Due to the ever-shrinking availability of easily accessible shallow water sites and the abundance of high quality wind resources in deeper water, the industry is stimulated to come up with innovative, yet cost effective, solutions to tap into these deep water sites. Historically, the use of monopile foundations has been an important facilitator of cost reduction due to its relative ease of manufacturability, transportability and installability. Monopile foundations have, however, thus far only been used in relatively shallow water depths. With jacket-type and floating support structures remaining relatively costly, the question arises if the monopile could yet be scaled up further to be used in water depths beyond the current 40-60mfor future, 10+ MW wind turbines. The goal of this research is to investigate the technical feasibility of monopile foundations in the water depth ’gap’ of 60 to 120 meters, which is currently claimed by jackets, for large wind turbines and determine critical design parameters for up-scaling monopiles to these depths. With an eye on future developments a 15 MW reference turbine is adopted and to make the research widely applicable the Hywind Scotland site in the Northern North Sea is selected, a well-documented reference site with a very severe wind and wave climate. To define monopile designs, a parametric (static) monopile geometry optimization tool is developed in Excel, which transfers the environmental data to forcing components. The monopile geometry is optimized for first natural frequency and ULS resistance (yield and global buckling) by varying the outer diameter and wall thickness along the structure. It was found that the ULS check is governed by the inertial wave forcing during survival case (50 year return period storm conditions). As the acting wave frequency is way lower than the system natural frequency, the monopile can be adequately assessed using a static approach. To assess the effect of critical variables, the tool is used to define monopile geometries for water depths ranging between 60 and 120 m, target first natural frequencies of 0.15, 0.17 and 0.20 Hz and a range of soil types, while complying with known manufacturing limits. The results show that all designs are within manufacturing limits and resistance against ULS decreases for lower target frequencies, making 0.15 Hz monopiles unfeasible. In contrast to the ultimate limit failures, the fatigue damage is largely incurred from normal rather than ultimate wave states, which embrace the first natural frequency of the system, thus warranting a full dynamic analysis. In order to assess the fatigue resistance of 0.17 and 0.20 Hz monopiles, which pass the ULS check, an analytical full dynamic model is developed in Maple based on the fundamental equations of motion, including (added) mass, aerodynamic damping and soil/structural stiffness. The model is verified and validated against numerical modelling using ANSYS Finite Element Analysis software and found to be in very good agreement. The Maple model is then used to test the fatigue performance. It is found that the fatigue damage accumulated over the 25-year design lifetime is governed by the bending stress cycles induced by inertial wave forces. The model is used to generate a transfer function between this wave forcing and the resulting bending stress over the entire range of present wave frequencies. The transfer functions are used to transfer the wave scatter diagram to accumulated fatigue damage against a B2, C1 and D graded S-N curve. The analysis results show increasing fatigue damage for decreasing first natural frequency due to the increasing slenderness and more prominent interference with present wave frequencies, which causes the 0.15 and 0.17 Hz monopiles to fail on fatigue damage. Given the very demanding site conditions assumed in this research, more slender monopiles may be feasible in deep waters with a milder wave climate. Over all, the results show no fundamental technical limitations for a monopile supported 15 MW wind turbine in water depths of up to 120 meters, provided that the stiffness of the structure is sufficiently high. Since steel usage increases for increasing stiffness, technically feasible monopiles will need to be relatively heavy, thus costly. Therefore, a range of strategies to reduce steel weight have been quantitatively assessed. It is found that the amount of steel can be reduced up to 35% by adopting higher-grade steel types and improved weld quality. Based on the fundamental limiting factors for monopiles found, also a novel hybrid floating-fixed bottom concept is proposed aiming at steel weight reduction. Although the concept can be considered promising, it does add complexity to the system and unfortunately does not (yet) result in steel reduction.

The world faces an increasing energy problem, forcing people to search for sustainable energy sources. Offshore wind energy has shown great potential to financially compete with traditional energy sources. Recent developments like the slip-joint connection increase this potential. However, for further optimization of the design of a slip-joint, the location of the contact areas between the two cones must be known. Previous attempts to detect these contact areas based on techniques such as heat transfer or ultrasonic measurements have proven insufficient. A possible new way of detecting contact areas, is through the behaviour of Energy Flux. Energy Flux methods have shown great potential as a damping identification tool in other applications. Therefore, in this study the relation between Energy Flux behaviour and the presence of a contact point in the time-, frequency- and time-frequency-domain is studied. To this end, a numerical analysis of a vibrating simply supported Euler Bernoulli beam is conducted, simulating a contact area with a point load. The analysis in the frequency-domain showed the most promising results. Presence of a contact point (i) introduces peaks at twice the first and twice the second eigenfrequencies, and (ii) increases peak height at the location of the contact point. The pressure of the simulated contact point increased these effects. In the time-domain the presence of a contact point increased the amplitude of the cumulative energy flux. This change was most significant at the antinode of the first eigenmode. The location of the contact point was of little influence on this effect. These results show that the presence of a contact area influences the behaviour of the energy flux. The results are encouraging for a later implementation of the energy flux method for the detection of contact areas in a slip-joint. As a validation of these results, an experiment has been proposed. After execution of this experiment, further research is needed in (i) the behaviour of Energy Flux in a conical shape, and (ii) Energy Flux measurements of higher frequencies.