ECN publication
Some ideas on the modelling of atmospheric turbulence [ECN-RX--94-122]
Published by: Publication date:
ECN 1995
ECN report number: Document type:
ECN-RX--94-122 Other
Number of pages: Full text:
11 Download PDF  

In the introductory note to the IEA Symposium on Wind Conditions for WindTurbine Design 1994, the question is raised whether the actually occurring extreme wind conditions, which recently lead to severely damaged wind turbines, make a reconsideration and evaluation of the commonly used design basis necessary. In addition several related questions were formulated of which one is addressed here: 'Is the format of currently used design guidelines concerning wind, i.e. extreme conditions and turbulence, sufficient, is it too simple, does it cover all the relevant situations?' The main line of thought is that we should seek for characteristic features in the wind that are important for the (extreme) response of the wind turbine structure. In the first section it is shown that next to strong gusts the spatial distribution of wind speed variations should be included as a characteristic feature that is important to system response. Gust models should be replaced by (extreme) wind models based on stochastic wind field simulation. The great number of time scales is another characteristic feature of atmospheric turbulence. Only the shortest time scales, with a maximum period of 10 minutes are included in load calculations. In section 3 it is argued that the variation of hourly averages should be included, because they can lead to a significant number of large load ranges, which in turn can lead to low cycle fatigue. The last characteristic feature that is dealt with is the randomness of the wind. Different realizations of a stochastic wind field, which have the same statistics will generally lead to different load spectra. Several realizations are needed to get a reliable estimate of the highest load ranges. A few examples are shown. 8 figs., 6 refs.

Back to List