The non-Gaussianity of wind turbine loads

Description

Recent studies have shown that wind fluctuations tend not to follow Gaussian distributions, al­ though, when designing wind turbines, it has always been assumed so to simplify the complex calculations and to reduce the time spent solving the computational models. It is a common belief that this assumption can result in an overestimation of the fatigue loads that the structure suffers and that consequently, in some scenarios, this could end up in a reduction of the tur­ bine lifetime. Even though studies have been focusing on whether the wind follows or not the Gaussian distributions, the real consequences that this hypothesis would imply have not been explored yet.
The aim of this project is to demonstrate that wind fluctuations are indeed non­Gaussian and how this non Gaussianity varies with the different wind parameters. To fulfill this purpose, the analysis is carried out with a 10­minute sample since it can be assumed that during this time period the wind maintains a constant average speed. The turbulence increment is the parameter used to represent the probability density function and to study the type of distribution.
Afterwards, the fatigue loads on the blades and the tower structure are analyzed, and several comparisons are made. In particular, loads are studied according to different wind parameters. Once their evolution is known, the samples are then classified between the ones following a Gaussian distribution and the ones that are not, and a comparison is made in between the damage equivalent load that each sample presents. The result of this analysis is that both groups are characterized by the same regression curve, although the Gaussian samples tend to converge more into this curve, while the non­Gaussian samples have more spread values. Thus, it is concluded that no evidence of an overestimation of the fatigue loads has been found.

Period	Jan 2021 → Jul 2021
Examinee	Ana Luna González
Degree of Recognition	International