| dc.description.abstract | Offshore wind energy has shown great potential and can be a great contributor when it comes to solving some of the climate challenges that the world faces today. However, both design and operational cost has been a limiting factor for the industry. The complexity of the structures is the main reason, as well difficult conditions for mainte-nance. In design of offshore wind turbines (OWT), one of the most important design criteria to evaluate is fatigue. The object for thesis is to provide a fatigue assessment for an OWT jacket structure in frequency domain.
Fatigue is damage caused by prolonged cyclic stress. For offshore structures, these loads are constantly present as they are affected from environmental conditions like wind and waves. Wind turbines are in addition affected by rotation of the blades which also influence to total dynamic properties of the system. Instead of using time-domain simulation, it is of interest to calculate the fatigue loads in frequency domain. This will shorten the simulation time significantly, which again will simplify structural optimization. To use frequency domain approach, transfer func-tions needs to be calculated. In this thesis, transfer functions have been calculated by applying white noise atop of the turbine in the point where the rotor nacelle assembly (RNA) is attached to the tower. White noise is used to be sure that all frequencies are included in the response. To get the excitation atop of the rotor, the RNA was decou-pled from rest of the system, and fixed in all directions. When a wind field then is applied to the rotor, the reaction forces at this point will correspond to the forces that would have occurred for full time domain simulation. The spectra of this excitation are then multiplied by transfer functions, and the result is a spectrum that shows the struc-tural response for the jacket. Based on these spectra is it possible to calculate damage by using Dirlik Method. This damage is compared to damage calculated by time domain simulations where Rainfow counting and Palmgren Miner rule were used. To compensate for aerodynamic damping, a uniaxial linear damper was applied to the sys-tem. The damping coefficient of this damper was found by considering the system as one degree of freedom (DOF) where horizontal displacement in wind direction in the only DOF.
Results from analyses performed in this thesis show that fatigue calculations in frequency domain have great potential for offshore structures. However, it has been difficult to recreate the effects of aerodynamic damping in an exact manner. The results shows that it is possible to recreate the response for the outer legs quite accurate, and that the damper works for the modes that it designed to damp out. The response for the braces was on the other hand hard to simulate in frequency domain. The main reason for this is the fact that they have different mode shapes than the outer legs, which the damper is not designed to damp out. In general is the response from frequency domain calculations higher than for time domain simulations since it is only one mode that is damped out, while aerodynamic damping in reality will influence several modes. When comparing damage calculated from the two different methods, there is a higher discrepancy than what is found in previous articles. For dynamic system where several frequencies is dominant in the response spectrum, it is expected that the deviation between the two methods is as high as 100%. | en |
