VAWT simulation with pimpleDyMFoam
 

Dear all,
Hello. I have been simulating VAWT in 3-D computational domain for the last couple of months. I am using pimpleDyMFoam as my solver. My current time step is very small like 2e-05 in 3-D computational domain, when the maximum courant number is 10. I am wondering, is it normal? Because I have seen the propeller tutorial. There the time step is around 1e-04. That's why I am a little bit worried.

I have also simulated and validated my results with experimental paper in 2-D. In case of 2-D I was able to manage a time step around 1e-04. In 3-D it would be tough, as there is separation around blade tip. But still I am confused. Again I was not able to validate my results with the experimental results.

Another problem is the pressure. My pressure is fluctuating over the entire computational domain. I don't know why.
Thank you all.

Have you compared your cell sizes and velocities to the ones of the propeller tutorial? These usually determine the Courantnumber, as you probably know from you CFD textbooks.

With regards to the pressure fluctuations: What are your settings in fvSolution and how large is your domain? Especially the space between your turbine and the farfield boundaries. Do you use AMI/GGI or rotate the entire mesh uniformly?

Jens

Your courant number will increase with rotor radius, rotational speed and inlet velocity, the two latters being related by TSR. Real size rotor with real inlet velocity will require very small timesteps. You might want to decrease one of these parameters, while keeping an appropriate reynolds number (you can go from 10⁷ to 10⁶ but not from 10⁶ to 10⁵ for instance). it'll give you a feasible timestep.

Fluctuating pressure is a bad sign. Check your boundary conditions, your very first timesteps, your GGI interface etc... I've had issues of fluctuating pressure with inlet mesh problems.

Sorry for my late reply. Yes I have compared my case with the propeller tutorial. It is quite similar. My turbine is small in size and the inlet velocity is high.

@Benn I am uploading my fvSchemes and fvSolutions. Would you please take a look at it?

Thanks in advance.:)


[ATTACH]Attachment 26821[/ATTACH]

Your fvsolutions/schemes won't tell me much for that issue.

I'd need screen captures of your mesh, and your turbine characteristics.

If your inlet velocity is high then yes you'll need a very small timestep. However, you can reduce this velocity, and "extrapolate" your results to higher velocities by using coefficient of performance, and normal/tangential force coeffs. etc....

I am also attaching some snapshots of fluctuating pressure.

Attachment 26823
Figure 1: Fluctuating pressure

Attachment 26824
Figure 2: Meshing of a slice

Attachment 26822
Figure 3: Overall meshing

As you may see I have used structured mesh in farfield (inlet and outlet). Do you think it might be a reason?

My boundary conditions are same as propeller tutorial. I have just changed the fvSolutions and fvSchemes a little bit.

What did you generate the mesh with ? is it hex or prism ? (can't really tell from your mesh) If it's prisms and cells are shared between inlet/outlet and side patches it will ruin your computation.

The pressure fluctuation looks indeed pretty nasty.

I have used pointwise to generate the mesh. It is a combination of hex and prism mesh. Sorry I did not understand the "cells are shared between inlet/outlet and side patches" part. As you may assume the prism mesh is around the blades. I am attaching some snapshots.
Attachment 26831

As you may see the prism mesh is around the blade. The other meshing elements are tetrahedra and pyramids. I have used unstructured meshing around the tip of the blade surface and structured mesh in other parts in order to save computational time. I am also attaching a snapshot.
Attachment 26833

I have extruded the meshing on the blade surface into prism mesh. After that I used unstructured meshing only.

Do you have any advice regarding meshing? If you want I could send you my meshing file. I am afraid if I am asking for too much.

Post your mesh in parafoam readable format, I'll check it out.

Dear Bennn,
I have uploaded the VTK file in the following link.
https://drive.google.com/file/d/0Byl...it?usp=sharing
Please take a look if you have time. :)

dude 1.3 gig ?

Sorry for the inconvenience and my late reply. I have uploaded another VTK format file. https://drive.google.com/file/d/0Byl...it?usp=sharing

please have a look. It is only 450MB. :)

Hi,

I am also working on a VAWT problem being new to OpenFoam i am trying to go through the propeller tutorial in OF. Since i also intend to perform a 3D analysis can you suggest or share some reading material about defining moving body in OF.How the rotating patches are defined? :confused:also How at start a 2D case can be setup by importing geometry and meshing:confused: it.I generally use ICEM CFD for meshing.

Thank you.

Hi! I use pointwise to generate HAWT mesh too, but not with the glf file.I am now confused about how to set the rotating zones,could you give me some advices? How do you set the rotation?with GGI or MRF or some other methods?I am not sure how to set the interface when use MRF,because when I generate mesh,the interface are disappered,I coulf only set the outer mesh boundary.So I use the topoSet/fvoption file to difine the ratating zones,but not sure is it right,plz help me.Thank you very much!