The nonlinear behaviour of a FOWT TLP during tendon tensioning

Abstract

Human activities, particularly fossil fuel consumption, are accelerating global warming. Allseas, in collaboration with the report’s author, is interested in the installation procedures for Floating Offshore Wind Turbines (FOWT) to mitigate this global warming. This research focuses on Tension Leg Platforms (TLPs), a structure that has more buoyancy forces than gravitational forces and thus uses tendons for stability when installed. There is potential in TLPs due to their motion stability, minimal seabed impact, and cost efficiency when installed, however, installation is a challenge because a TLP relies on its tendons for stability. The installation should be safe, reliable, and cost-effective. Current installation methods are inefficient and not used on a commercial scale. Allseas has designed a concept ship
to install FOWT TLPs cost-effectively, potentially reducing emissions. Research performed with free-body diagrams and a literature study into hydrodynamic behaviour during installation reveals several challenges and knowledge gaps. Most papers about the behaviour of TLPs are about the moment that they are already installed. The papers that are found about the installation and include time-domain simulations are on TLPs used for the oil and gas industry. These do not have a flexible tower structure on top of them, while in the wind industry the accelerations of the flexible structure might be one of the key limiters of the operation. Next steps need to be taken in understanding what is necessary to model the installation of tension moored structures; in particular, tension-moored structures for FOWT in the process of going from slack to taut. The safety of the installation of FOWT TLPs would improve when their behaviour during installation is known. To study this gap in the existing literature, the following research objective is proposed: To find the minimal viable method to investigate the feasibility of the tendon tensioning phase of a FOWT TLP in different sea states.

What makes the TLP installation studied in this paper state of the art, is that a unique installation method is applied. For installation, an external ballast deck will be used, which is connected to the installation vessel with an installation system. This research will be performed on the test case of a 20𝑀W TLP that is brought to its design draught starting the moment that the tendons are connected. For the case study, the maximum nacelle acceleration that may happen during installation is 3.0𝑚/𝑠2 (Taboada et al., 2020). The maximum tendon tension may be 25𝑀𝑁; this is a criterion for a 20𝑀𝑊 TLP and can change for other wind turbines. The following steps have been taken to successfully fulfil this research. First, the method of numerical modelling of a TLP had to be validated. To do this, an
experiment of a 5𝑀𝑊 TLP was remodelled. After this was validated, a 20𝑀𝑊 model could be made following the same procedure. When the behaviour of an already installed TLP was validated, the installation model could be built. For the installation, only the installation deck was modelled at first to build on the complexity of the model. When this was done, the ship’s presence could be taken into account. To research the implications of simplifications, the ship first responded to displacement RAOs before load RAOs were used to model the hydrodynamic interaction between both bodies. When the model was finished, different sea states could be modelled. Because every spectrum has different seeds that all correspond to the spectrum, a statistical analysis had to be performed on the results.

The results of the installation without the ship present show that with a perfect motion compensation the installation of a TLP would be possible. The influence of waves is significant during the installation. When there is a slight disturbance at the moment when the tendons go from slack to taut, peaks in the nacelle accelerations and tendon tensions arise. The tendon loads peak with the same frequency as the wave period, which means that they are directly generated by wave loads. The higher the waves, the higher the peak loads.

The influence of the installation ship on the results is that the maximum nacelle acceleration and tendon tension during installation get higher. This holds for the model with the displacement RAOs and the model with load RAOs. The model with the load RAOs is the most realistic: the ship’s movements influence the TLP motions and the TLP impacts the ship.

Adding a sea state to the simulation can increase the movements compared to when a regular
wave is used. If the sea state includes periods closer to the natural frequencies, the response is
higher. This can also be due to the second order wave drift and sum frequency loads that are included on the TLP. A randomly generated wave time series can also include lower and higher waves during the simulation. To account for this randomness, the simulations of a chosen set of sea states are run with 20 different seeds. Four of them have the same significant wave height and differ in average wave period, three of them have the same average wave period and differ in significant wave height.

The research has shown that when the response is higher, the spreading of the maximum nacelle acceleration and maximum tendon tension observed per simulation is also higher. This means that near the installation criteria, the most seeds have to be run to make sure that the operation is feasible. With the results obtained during the 20 seeds, the change of the P90 remains below 5% after 11 runs, this would thus be the recommended amount of seeds. The sensitivity study also found that the tensioning phase, where the tendons go from slack to taut till the moment that the deballasting is finished, is the normative phase for installation. The taut phase is the phase where the deballasting is finished till the end of the simulation. For the nacelle accelerations a maximum factor of 2.0 is found between the two phases. For the tendon loads a maximum amplification factor of 1.2 is found. The study also established that the tendon loads were not governing for the design of the TLP. It can be concluded that the current method for modelling can be used to find the limit of workability. With other designs for the turbine or TLP, one could still implement the method used.

The main limitation of this study was the concept nature of the installation method, and thus, not every part of the design is known yet. Additionally, the model simplifies reality. In spite of its limitations, the study gives insight into the minimum model necessary for a safe TLP installation.