Discussion Groups

Simulation CFD

Reply
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Some clarifications

279 Views, 11 Replies
10-09-2012 08:45 AM

Hello,

 

I have started using Autodesk Simulation CFD at work and I have some questions:

 

1) I am modelling conical perforated filter as resistance and using a shell approach. Meaning, I will model the internals of filter as a fluid domain and the surface as resistance with a finite thickness. Admittedly, this approach is easier. But I am afraid if am losing some physics here. All the flow emanating from the resistance is perpendicular to conical surface, regardless of the direction of its entry; a fact that can be attributed to the inherent modelling of "surface resistances" in Simulation CFD.  When I try to model the filter as a cylindrical part with thickness, SimCFD throws an error that says material devices (resistance) should have at least two surfaces in contact with fluid. I know there are two surfaces in fluid contact. Yet why do I get this error?

 

2) Does SimCFD have any bounded scheme for discretisation, which will smear out the wiggles at the discontinuities or large gradient areas?

 

3) The jump of mesh size is way too high when mesh switches from  prism layers adjacent to wall to inner cells. Would this jump cause loss of physics?

 

4) Does SimCFD support the meshes created in other dedicated meshes, like say ICEM CFD or TGrid?

 

Sorry for the long post, but I switched from ANSYS solvers and there are many things I am curious about.

 

Cheers

OJ

 

Please use plain text.
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Re: Some clarifications

10-12-2012 07:02 AM in reply to: OmkarJ

No replies!?

Please use plain text.
Product Support
wildej
Posts: 643
Registered: ‎08-25-2011

Re: Some clarifications

10-15-2012 02:12 AM in reply to: OmkarJ

Hi, I can help with some of this:

 

1) Try using the other directions available within the material setup, the 2 normal directions. This ought to allow you greater scope regarding the flow direction

 

2) Do you mean something like the mesh adaption or meshing regions or purely a graphical solution?

 

3) No, you should be fine, would help to see an example though. If you wish to refine the mesh though, you of course can do. There are many mesh controls available within the advanced and enhancement settings.

 

4) No, best to use our mesher. It does work very well, although the approach can seem different.

 



Jon Wilde
Please use plain text.
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Re: Some clarifications

10-16-2012 05:18 AM in reply to: wildej

#1 : Please refer to the attached image. The conical filter is perfectly surrounded by fluid domain. I have tried flow direction as both "Radial" and "perpendicular to conical surface". Additionally, I define the axial (to cone) direction as another normal direction. But no success. To reiterate, I am modelling the cylinder as actual cylindrical part with thickness, and not using surface shell resistance.

 

#2: I was taking about discretization schemes for advection/turbulence. If the second order schemes tend to oscillate near discontinuities and sharp gradients. While first order schemes, though smooth near discontinuities, are inaccurate due to diffusion. I wanted to understand if SimCFD has a bounded second order scheme, preferably monotonous.

 

Thanks for clarifications.

 

OJ

Please use plain text.
Product Support
wildej
Posts: 643
Registered: ‎08-25-2011

Re: Some clarifications

10-16-2012 05:41 AM in reply to: OmkarJ

How are the directional settings within the material setup though? Could you attach a cfz file?



Jon Wilde
Please use plain text.
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Re: Some clarifications

10-16-2012 05:53 AM in reply to: wildej

Am afraid I can not share the files, given the confidential nature. Can you help me without them?

Please use plain text.
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Re: Some clarifications

10-16-2012 05:57 AM in reply to: OmkarJ

Additionally, for directional settings, please refer point #1 of my earlier post.

Please use plain text.
Product Support
wildej
Posts: 643
Registered: ‎08-25-2011

Re: Some clarifications

10-16-2012 05:58 AM in reply to: OmkarJ

I can try!

 

Have you tried changing the directional settings within the material? There are 3.



Jon Wilde
Please use plain text.
*Expert Elite*
Posts: 377
Registered: ‎10-02-2012

Re: Some clarifications

10-17-2012 03:07 AM in reply to: wildej

Am not sure if I understand what you mean by directional settings. The only one that seems appropriate is flow direction towards radial and normal direction 1 along axis (going up). Admittedly, I haven't explicitly defined normal dimension 2. Is that what you are hinting? I chose this approach as it is in alignment with the HVAC tutorial.

 

Can you suggest specifically, how I should try to set the flow direction, normal direction 1 and normal direction 2, in all possible combination that you consider appropriate for this modelling?

 

Thanks

OJ

Please use plain text.
Product Support
wildej
Posts: 643
Registered: ‎08-25-2011

Re: Some clarifications

10-17-2012 03:40 AM in reply to: OmkarJ

If each direction were set to 0.5 (Free Area Ratio) then the flow would be able to travel at any direction through the media. This is what you were asking initially if I am correct? How to allow the flow to travel in more than just the radial direction?



Jon Wilde
Please use plain text.