[Dorit]

Hello everyone,

I am doing research on wind turbines (HAWT's) aerodynamics using Fluent. Conceptually I want to analyse the staring-up performance of a HAWT by generating cp curves and looking at corresponding flow features. There are two ways of how to go about this:

1) Set up a steady state simulation (if possible!) of a HAWT, once converged extract the torque to plot the cp curve and then use the solution as an initial guess for a simulation at a higher rotational rate. The advantage of this approach is two fold: only 1/3 of the full domain is required and the cp curve can be generated reasonably quickly. However the other approach is:

2) Run a simulation at 0 rpm, extract the moment once converged and then use an interia model to calculate the new rotational speed at a certain time step. Repeat the process until a constant rpm is achieved. While this approach is computationally very expensive, I think that it may be able to capture unsteady processes that the first method is not able to predict.

Are there any people out there who have any experience/ideas/interesting thoughts on that subject? I would like to start an open conversation about the topic with anyone who's interested. For instance I'm especially interested in suggestions of how many time steps would be required at what time step size for method 2, if anyone has done something like that.

Dorit