I am not sure of the meaning of your question. Forcing is likely to follow a power law due to turbulence. However, tip motion will likely to be driven by the blade resonant frequencies.

Thank you Yvan. Yes, I am finding from a review of response frequencies that the first natural frequency is dominant. I want to find a way to summarise the displacements and range of displacements in some statistical form that I can use later in an optimisation routine.

im afraid that i can't follow the point of your question, but i think you would like to know how to identify this system using an empirical method.

i suggest that you use an identification tools from MATLAB: you can use either ARX or ARMAX tools. you can just take a few input(either wind force/speed) and output (the tip's displacement) data tha you have recorded. thus way, you will have transfer fuction estimation from this system.

Thanks Zaid. My goal is to describe the tip motion of a given simulation by using a few numbers - possibly a mean and variance so that these can be the targets for a neural network. Other results such as peak displacement are also available at this stage. Input and output could be used to train a neural network and hopefully generalize results of a large number of forced simulations. Therefore I am wondering if using a normal distribution is too crude to describe tip mean displacement values from each cycle and similarly if using an exponential distribution to describe the range of displacements is also too crude. Does this make sense?