[floooo]

Hi,

I am student. For my end of study's project, I try with 2 others students to study a new generation of vertical wind turbine with a CFD simulation .

We try to draw the curve of the Power coefficient CP according to lambda (the Tip Speed Ratio).

In order to have a reference, I are working in the same time on a Savonius wind turbine.

If we had a prototype and a wind tunnel, we would have to connect the wind turbine to a dynamic couple-meter. We would obtain the wished curve by making vary the resisting torque while keeping a constant speed of wind.

We started with fluent, but in this software it is not possible to put a resistant torque on a moving mesh.

So we made the opposite : we imposed a rotation speed of pales then we measured the couple engendered by the fluid on the rotor.

We simulated this with a stationary simulation and a rotational moving mesh boundary condition on blades. And then a in-stationary simulation with a moving mesh.

Both simulation give exactly the same results, the residuals are very good, but Cp curve is not realistic for lambda > 0.

For Lambda = 0, the result is very close to the theoretical value.

After a lot of work we concluded that the model is not good :
putting a rotating wind turbine in an air flow and measuring the couple engendered on the tree of rotation is not the same that measuring the rotation speed of a wind turbine on which we impose a resisting torque.

I have work on OpenFOAM during an internship period.
I have try the mixerVessel2D tutorial.

Is it possible to impose a resisting torque on a moving mesh area in OpenFOAM ??

[skshivas]

It is possible to feed a Resistive torque and force in fluent for moving mesh model.or otherwise u can use Rigid body solutions (6-dof) Methods in CFX.

[enry]

Hi floooo
I'm also studying a VAWT. Are you sure that resutls obtained using a fixed rotation to the blades is not reasonable? I'm using this approach, and results seems to be good.

I have a question for you: I studied the problem using FLUENT, and now I'm trying to use OpenFOAM. I use sliding interface tecnique, but OF torque is not the same of FLUENT one. After a lot of simulation I think that the problem is maybe in boundary conditions. How have you set them for your problem?
Thank in advance.
Enry.

[floooo]

Hi enry

I think that skshivas gave the solution. But it was too late for my report.
Currently I haven't the time to work on it.

There is an 6 DOF solver in Fluent. And I think it is possible to solve the problem using and UDF script.

I think the Fluent UDF file looks like this:

#include "udf.h"
DEFINE_SDOF_PROPERTIES(store, prop, dt, time, dtime)
{
prop[SDOF_MASS] = 10;
prop[SDOF_IZZ] = 200.;
if (time <= 0.3)
{ prop[SDOF_LOAD_F_X]=0; prop[SDOF_LOAD_F_Y]=0; prop[SDOF_LOAD_M_Z] = 20.0;
}
}

Florian