The offshore wind energy industry has been developing rapidly due to the increasing demand for renewable energy. As a result, offshore wind turbines are increasing in mass and height and are installed at larger water depths and further away from shore. A floating vessel is not limited by water depth and soil conditions, but is sensitive to wave-induced motions. Heerema Marine Contractors developed a new installation method using a floating vessel, where the rotor nacelle assembly (RNA) of a wind turbine is assembled on an assembly tower on deck and is installed in one lift. The research objective is to gain a better understanding of the dynamic behaviour of a rotor nacelle assembly, particularly due to the wind, focusing on the horizontal motions of the RNA, when it is hanging in the crane above the tower. To examine the dynamic behaviour of the RNA, separate models are made to assess the effect of the wind and waves individually in OrcaFlex and Liftdyn. This is done to identify which motions are due to the wind, and which motions are due to the waves. Finally, the wind and waves are coupled to gain insight in the coupled response. Furthermore, a case study is performed to prove its applicability for implementing and assessing the effect of using tugger lines on mitigating horizontal motions. In the models the blade pitch angle, crane position and responses for different wind and wave conditions are examined. From the different models is observed that the response due to wind cannot be neglected. Wave-induced motions are governing, due to large crane tip motions. The governing motion is a pendulum motion of the RNA, which is larger for longer wave peak periods. For both short and long wave peak periods, the horizontal motions have to be mitigated. In the case study, tugger lines are implemented to examine the effect on the RNA motions and to assess if the model can be used for developing concepts to reduce the motions. It is shown that providing stiffness in the horizontal plane and in roll help to reduce the sideward pendulum. For long peak periods, the RNA motion has to follow the opposing motion of the crane tip. More research is required to develop a concept that can follow the crane tip for short peak periods and oppose the motion of the crane tip for long wave peak periods.

A new development aiming to improve the safety and/or operability of offshore operations is deterministic motion forecasting. Various parties target at developing a radar which is able to detect wave trains. This deterministic wave forecast can be used to forecast vessel motions up to several minutes ahead. With such a forecast a quiescent period can be searched for, where ship motions are at a minimum. Possible advantages considered for a heavy lift operation are: Safer operations, less waiting on weather and extended lifetime of equipment. With the use of the Heerema Simulation Center (HSC), the possibilities to extend a deterministic wave forecast into a motion forecast is investigated. Thewave components are known in the HSC, so a perfect wave radar is available and the focus is entirely on (1) predicting motions and (2) presenting it in a useful way. The first part of the objective is approached by investigating the possibilities and limitations of making a motion forecast in the frequency domain. This method is convenient since the calculation is easy to understand and very fast, a quality which is most wanted when making a future prediction. To quantify the quality of a wave- andmotion forecast, the correlation coefficient (½) and RMS ratio (¾) are used (forecastedvs logged motion). The model showed good results for heave roll and pitch (½ 0.85 - 1.0, ¾ 0.86 - 1.15) for mild, severe and complex sea-states (all wave directions, respectively). For extreme swell a forecast could not be made. This can be explained by the fact that when motions become larger, a frequency domain approach is no longer valid. However, it should be noted that it is very unlikely a lift operation will take place with Hs > 2m. The sensitivity of themodel to variations in the center of gravity and radii of gyration of the vessel (model input), is found to be within acceptable bounds. The impact is negligible on the phase and small (<10%) on the amplitude of the forecast. With data offshore available with an accuracy of +- 2m, this is acceptable. Two test cases are investigated. A single crane lift (1000 mT module) and a dual crane (7000 mT module). In both cases the model proved to be well capable of forecasting themotion of interest (Heave) of the module (relative to a fixed and floating platform). The second part of the objective has a more qualitative character, where the use of a forecast is tested in the HSC by using a tool that is created for this purpose: the VirtualWave Radar Tool (VWRT). This tool is able to make and display a forecast in the HSC. The tests executed in the HSC showed positive results for the use of a forecast in the crane cabin during a lift operation. A reduction of secondary impact loads in a challenging environment was realized due to the use of the VWRT. Furthermore, a forecasted time span of two minutes (current performance of wave radars on the marked) proved to be long enough to benefit from during a lift operation. The tool demonstrated that simplicity of the dashboard is key and should be understandable at a glance. With many possible sources causing a forecast to deviate from the actual motion, a self learning, optimization model could reduce the need for exact model input offshore (not always available), even without knowing the source of deviation. v