Measuring the loads on offshore wind turbines with conventional mechanical strain gauges adds maintenance effort and a risk of reduced availability. Considered over the entire lifetime and current technology this direct approach is uneconomic for large scale application. However, without the site and turbine specific fatigue an optimized maintenance schedule, wind farm load control and potential life-time extension is not possible. Based on two years of field data the accuracy of fatigue load estimation is evaluated with respect to measurement campaign related design parameters such as the choice of instrumentation, signal resolution and duration of mechanical strain measurements. A correlation study was conducted to narrow down the number of standard data statistics required for the fatigue load estimation. A brute force neural network configuration test was performed to analyse the scope of possibilities when choosing standard data statistics as neural network inputs. It was shown that the user has the exibility to choose between different neural network input conifgurations i.e. standard data statistics while achieving similar estimation accuracy. Finally, it was demonstrated that the estimation accuracy of the neural network does not further increase by providing more and more training data samples. The criteria for a minimal number of training data required is formulated by the sample data spread in mean wind speed and turbulence intensity. This results in a minimal duration requirement for the measurement campaign. There is no need to maintain the load measurement campaign longer once sufficient data has been collected. Thus the cost of measurement measurement campaigns for fatigue load monitoring can be reduced by choosing the type and quality of signals to be measured and by minimizing the required duration of strain gauge base measurements. Although the results presented are only valid for the discussed load quantities and load cases, the methods can be applied to further case studies. Future work will focus on improving the fatigue life estimation accuracy.
On the design of measurement campaigns for fatigue life monitoring of offshore wind turbines
2013
6 Seiten, 6 Bilder, 1 Tabelle, 10 Quellen
Aufsatz (Konferenz)
Datenträger
Englisch
Improvements in the fatigue design of support structures for offshore wind turbines
TIBKAT | 2021
|Performance-Based Design of Offshore Wind Turbines
British Library Conference Proceedings | 2010
|Advanced Topics in Offshore Wind Turbines Design
British Library Conference Proceedings | 2010
|Special fatigue aspects in support structures of offshore wind turbines
Tema Archiv | 2011
|Dependability of Offshore Wind Turbines
British Library Conference Proceedings | 2010
|