The Weibull might work because it it is flexible, approximating many other probability distribution functions :)

Andrew, do you have any suitable reference to support your answer?

First of all, obtain normalized histograms for your data. Of course, with enough data. I suppose wind will have very unstable behavior, so take into account geography, terrain topology, month of the year, day hour, for each histogram.

You can check the attached file related to your problem.

Suitable functions can be Weibull, Johnson´s Sb, Rayleigh, Burr, and segmented variants.

Best regards

Obtain a random sample of the output power of a wind generator. Then use a Mathematica function called

FindDistribution

to find the suitable distribution that fits the data. You can use either

FindDistribution[data]

to find a simple functional form to fit the distribution of data, or use

FindDistribution[data,n]

to find up to n best distributions.

Here is my paper, which has details about the Weibull :)