Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

ALE- Incompressible Navier stoks

Please login with a confirmed email address before reporting spam

Hi ,

I have created a moving boundary model. I get some results when the pressure gradient is low. As soon as I increase the pressure gradient or velocity, although it is still laminar, solver diverges. Anybody is working on the same modules? I can share more data on the problem then cause I did not to send a long message now.

Regards,

11 Replies Last Post 28 janv. 2010, 03:49 UTC−5

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 janv. 2010, 07:23 UTC−5
It would help if you describe your model in some lines.
It would help if you describe your model in some lines.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 janv. 2010, 19:00 UTC−5

It would help if you describe your model in some lines.


Hi Daniel

Thank you very much for your message. The model is attached below:

www.4shared.com/file/203837477/2bf3f9b7/test.html

As you can see, I have added some support line to prevent the mesh to get inverted. As I increase the pressure gradient the solver diverges. I am not sure about the solving method and time steps that I have chosen but I have tried different mesh models and different support lines. Non of them helped me so much. The desired pressure gradient is 900 ...

Please have a look at the file and kindly let me know your suggestions,

Regards,
Hani
[QUOTE] It would help if you describe your model in some lines. [/QUOTE] Hi Daniel Thank you very much for your message. The model is attached below: http://www.4shared.com/file/203837477/2bf3f9b7/test.html As you can see, I have added some support line to prevent the mesh to get inverted. As I increase the pressure gradient the solver diverges. I am not sure about the solving method and time steps that I have chosen but I have tried different mesh models and different support lines. Non of them helped me so much. The desired pressure gradient is 900 ... Please have a look at the file and kindly let me know your suggestions, Regards, Hani

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 janv. 2010, 13:59 UTC−5
Hi Hani,

I cannot access your file. Maybe you try another file sharing?

The file link that you requested is not valid.

Danial
Hi Hani, I cannot access your file. Maybe you try another file sharing? [quote]The file link that you requested is not valid.[/quote] Danial

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 janv. 2010, 16:04 UTC−5
Sorry,

I attached the file.

Regards,
Hani
Sorry, I attached the file. Regards, Hani


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 janv. 2010, 17:43 UTC−5

Sorry,

I attached the file.

Regards,
Hani


Try using Heaviside functions on your pressure to ramp it up to the final value. In your inlet pressure I used P*flc2hs(t-0.1,.1), and I got the following result (attached screenshot).



[QUOTE] Sorry, I attached the file. Regards, Hani [/QUOTE] Try using Heaviside functions on your pressure to ramp it up to the final value. In your inlet pressure I used P*flc2hs(t-0.1,.1), and I got the following result (attached screenshot).


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 janv. 2010, 18:21 UTC−5
Thank you very much. In fact I have tried that as well. I made a ramp from time zero to 2, that the pressure reaches 300 and from t=2 to t=3 pressure is constant p=300. Again the cycle is not completed and it crashes at time 2.67. The file is attached.

What about the model that you tested, as far as I see, it has crashed at t=0.52. Am I right?

The other possibility is to make the channel longer after the moving parts in order to reduce the effect of outlet on the fluid. But again in that case, I could not get any result....
Thank you very much. In fact I have tried that as well. I made a ramp from time zero to 2, that the pressure reaches 300 and from t=2 to t=3 pressure is constant p=300. Again the cycle is not completed and it crashes at time 2.67. The file is attached. What about the model that you tested, as far as I see, it has crashed at t=0.52. Am I right? The other possibility is to make the channel longer after the moving parts in order to reduce the effect of outlet on the fluid. But again in that case, I could not get any result....


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 janv. 2010, 05:28 UTC−5
No it didn't crash; I just stopped it, as I was running it my laptop.

I will try again today and report back
No it didn't crash; I just stopped it, as I was running it my laptop. I will try again today and report back

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 janv. 2010, 11:33 UTC−5
I will appreciate that.
I will appreciate that.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 janv. 2010, 11:43 UTC−5
Hi Hani,

I tried some more simulations, and I can say so far that mesh around the neck is a very important part. So I refined the mesh in that area and I could simulate a bit further. How about you draw the mesh quality and improve the parts that goes low.
Hi Hani, I tried some more simulations, and I can say so far that mesh around the neck is a very important part. So I refined the mesh in that area and I could simulate a bit further. How about you draw the mesh quality and improve the parts that goes low.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 janv. 2010, 22:39 UTC−5
Yes. That would be a good idea. I will do that. What about the solver. In fact I ma not a finite element man, I do not know how to decide between different solvers and which one is more suitable for this problem. Do you have any idea? UFMPACK, SPOOLES, PARDISO?...

Thanks
Yes. That would be a good idea. I will do that. What about the solver. In fact I ma not a finite element man, I do not know how to decide between different solvers and which one is more suitable for this problem. Do you have any idea? UFMPACK, SPOOLES, PARDISO?... Thanks

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28 janv. 2010, 03:49 UTC−5
I usually use Pardiso, maybe sometimes UMFPACK. I set the Gen-a solver with 0.5 for frequencies and type of scaling to manual (or none). But it really depends on your problem and needs a feeling for it.
I usually use Pardiso, maybe sometimes UMFPACK. I set the Gen-a solver with 0.5 for frequencies and type of scaling to manual (or none). But it really depends on your problem and needs a feeling for it.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.