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.
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.
Product Support Specialist
#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.
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?
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?
Product Support Specialist