Fatigue Considerations based on Measurements from an Offshore Wind Turbine
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2615029Utgivelsesdato
2017Metadata
Vis full innførselSamlinger
- Institutt for marin teknikk [3437]
Sammendrag
This thesis analyses full scale measurements from a 5 meter diameter, monopile supported, 3.6 MW Siemens offshore wind turbine, standing at about 20 meters water depth at Sheringham Shoal offshore wind farm. Strain measurements at about +7.6 and -20 meters wrt. sea level, environmental parameters and a numerical model of the turbine, are provided by Statoil ASA.
An overview of the site specific environmental conditions has been established through marginal and joint probability distributions. Further, full scale strain series have been compared with strain series from time domain simulations using the SIMA Workbench. Some differences in the strain series can be observed, and are thought to be due to the different wind turbine controller for the full scale turbine and the numerical model, but also due to the uncertainties in the wave measurements. The reason for the observed differences from the full scale turbine to the numerical model should be investigated further before strong conclusions are drawn.
Total damping has been estimated from both the full scale measurements and the numerical model, which for full scale has been found from analysis of the decaying strain responses triggered by shut-down events. The total full scale damping has been found to be twice as large as damping found from the numerical model, which is thought to be due to the large displacement soil damping connected with the large displacements from full scale shut-down events.
By transforming the stress signals in the fore-aft direction to the frequency domain, the difference between the numerical model and the full scale turbine, for the first and second tower natural frequency, has been found to be +10 \% and +20 \%, respectively. This is thought to be due to too low soil stiffness in the numerical model.
The fatigue damage contribution from shut-down events has been found by comparing the 1-minute fatigue damage over a shut-down event with the 1-minute fatigue damage just before the shut-down events. By analysing 70 shut-down events, the average fatigue damage from a shutdown event has been found to be equivalent to 1.1 days of normal operation, which could be important if the shut-down frequency of occurence is high. High scatter in the results suggests that further investigations are neccessary before strong conclusions are drawn. However, it is fair to say that unnecessary shut-down events should be avoided.