6 Replies Latest reply on Jun 5, 2012 4:24 PM by Boyko Tchavdarov

# Analyzing Wind Turbine with Flow

Hello all,

I am interesting in how to use flow to simulate flow through a wind turbine/wind mill/similar type devices. I know that it can do the opposite, where you can specify a rotating impeller and then determine flow that it created. But I want to see  the flow through the turbine so that I can use that data to calculate how fast the turbine will rotate as a result. Anyone know how or done this before?

Thanks!

-Mike

• ###### Re: Analyzing Wind Turbine with Flow

We wanted to do something similar but were told by our VAR that such an analysis was not possible.

Our best option so far has been to do a study with everything static, get the force component on the blades in the direction of motion then calculate angular velocity from that (taking in to account rotating mass, bearing friction etc.).

• ###### Re: Analyzing Wind Turbine with Flow

Hi Mike,

Perhaps you can get some input from http://grabcad.com/questions/tutorial-how-to-do-a-flow-simulation-to-a-rotating-propeller ? Not sure if that's what you are after, but the tutorial seems ok.

//Bengt

• ###### Re: Analyzing Wind Turbine with Flow

Dave Laban is on the right track in my opinion. You can solve for torque on the fluid surfaces, as well as on a shaft. However, you still need to factor in any sort of gearbox as well as the generator/load being used in the windmill.

With something of this nature, you could simplify things by running just one blade and then multiplying the forces/number of blades.

If running the full assembly, I would mark the 'gravity' checkbox as well.

• ###### Re: Analyzing Wind Turbine with Flow

Thanks everyone. I had originally been thinking along the lines of computing the force on the blades with a given flow speed to determine the torque and subsequent rotational speed taking into account bearinf friction and rotational inertia.

The grabcad approch seems like a good method as well as far as how to set up the simulation. I'd be able to obtain the pressure/force results from that I'd think.

• ###### Re: Analyzing Wind Turbine with Flow

There are two ways to approach rotational problems, global rotation or using a rotating region.

I just finished working on a water turbine simulation that is very similar to this one.

The tutorial posted is good.  The only thing that is missing from what I see are the goals (which are critical).  To measure the torque you need to create surface goals for the Torque in the X,Y,Z directions using the prop surfaces.  Then make an equation goal for the resultant torque.  I would use this equation goal to determine convergence (rather than other goals such as average or max fluid velocity).

I found when running a similar analysis that the simulation converged more easily with the torque goal rather than some fluid type of goal.  But it all depends on your rotational speed and fluid velocity.  With these sort of problems (in some instances) you will notice that things do not converge but rather oscillate (in a sine wave manner).

Anyway I thought I would add my tip about using a resultant torque equation for convergence.

Also for these sort of simulation make sure to check that your results are not mesh sensitive.  The meshing capabilities in 2012 are greatly improved.  Check to see that the blade surfaces are close to what you want.  I found a good way to check this is by creating a velocity surface plot of all the surfaces, then check off the "use CAD geometry" box.  This gives you a picture of what the meshed prop really looks like (rather than overlaying the colors onto the CAD geometry).  You  might be surpised at what you see if your mesh is too coarse.

• ###### Re: Analyzing Wind Turbine with Flow

You can use the fan model to determine the speed of the turbine quite accurately. Set up for a Parametric Study to minimize the axial torque by varying the rotational speed at given wind speed & direction while the Environment Pressure B.C. is set-up  on the rest of the computational boundaries. Use a global rotating region with a cylindrical shape of the computational domain with axis aligned to the wind direction (not to the turbine axis, if the two directions are different). Allow for at least 8xD distance between the turbine location and the far-field exit boundary.

Let me know how it works.