The Assessment of the Fatigue Behaviour of Offshore Wind Farm Support Structures

Abstract

Renewable energy generated by offshore wind turbines (OWT) are nowadays more frequently used in the offshore industry. The design of OWT foundations share similarities with a typical offshore oil and gas construction, however a OWT farm usually consist of multiple foundations whose installation are more cost-sensitive compared to one oil and gas platform. Furthermore the profit target of the wind farm is generated at a later stage of the project. Thereby it is preferable to design a light OWT foundation that is cheaply fabricated and easily installed.
Due to these reasons the OWTs are designed as slender constructions. When designing a structure the aim is to generate a design of which the natural frequencies do not overlap with the applied forcing frequencies. This is required to avoid resonance which results in low estimated fatigue life of the structure. A challenge regarding the slender OWT structures is that the natural frequencies are close to the forcing frequencies.
This thesis is based on a project having OWT structures with the foundation consisting of a jacket design having a framework of four bays of cross braces. During a later stage of the project it was found by the substructure designer that the structure would have a low estimated fatigue life. The cause being that the applied external forces on the structure are in the same frequency range as the natural frequency of particular cross braces in the jackets, in the range of 2 – 2.5Hz. Consequently, due to resonance large brace excitation was found in the jacket model, created by the designer, which resulted in a low expected fatigue life. To mitigate this, double sided welds and external toe grinding were implemented. However no clear explanation regarding this issue was provided and not all parties involved obtained the same results and conclusions. Therefore further investigation regarding this theoretic fatigue issue was requested.
The goal of this research is to identify and provide a clear explanation of the vibration amplification movement of the braces and thus independently assess if there is a fatigue problem with the OWT jackets. Furthermore this thesis focusses on alternative solutions to this problem, had the vibration issue been found during an earlier stage of the project.
The first part of this thesis focusses on the identification of the problem. The approach was structured in three stages;
First the identification of the forcing vibration in the critical frequency zone was performed with the use of a Fast Fourier analysis of time domain data of the applied forces on the OWT. The obtained frequency response demonstrated excitations around the critical frequency zone, which are due to the eigen-frequencies of the tower and the blades of the turbine.
This was followed by the analysis of the eigen-frequencies of the cross braces of the jacket in bay 3 and 4 for the local and global response. This was performed with the use of ANSYS modal analysis in combination with a build-up of simplified brace models. Hand calculations were performed to verify the obtained results.
Finally, a comparison was made regarding the overlapping of frequencies of the forcing frequencies and eigen-frequencies of the braces. This was done to determine if large brace amplification due to resonance would occur, resulting in a low estimated fatigue life.
The second part of this research evaluates alternative solutions to this problem. Possible concepts were elaborated with their ad- and disadvantages, followed by a multi criteria analysis. Three concepts were selected for further investigation and examined in terms of whether a successful outcome could be achieved by their implementation.
This thesis also provides a procedure with regards to identifying to the potential for resonance at an early stage of the design process. Keeping track of changes during all design phases will lead to avoiding the unexpected discovery of a short fatigue life, due to resonance, at a late design stage.