[OneDotThree]

I need some help...

I've to represent in ICEM a volume (the section is in the figure below)

Cattura.PNG

Is it possible to represent only 1/8 (in red) of the volume to do simulation? if it is, which boundary condition I have to set? symmetry? periodicity or both?

If it isn't possible I think I have to represent 1/4 of the volume (in green) and I have to set periodicity conditions, is it correct?

Thanks...

[Far]

Is it possible to represent only 1/8 (in red) of the volume to do simulation? if it is, which boundary condition I have to set? symmetry? periodicity or both?

No, 1/8 is not enough. You need symmetry condition.

If it isn't possible I think I have to represent 1/4 of the volume (in green) and I have to set periodicity conditions, is it correct?

1/4 volume and symmetry condition.

[OneDotThree]

Thanks a lot Far

[Far]

Although I believe that the periodic condition is not suitable for this model, but you should use both conditions and check accuracy of both conditions.

[evcelica]

What type of symmetry you can use depends on what your actually modeling, if its non-buoyant flow through this cross section you can use 1/8 symmetry. If its buoyant flow through this cross section you can only use half symmetry. But you haven't told us what you are modeling, you've only shown the shape.

[Far]

well 1/8 is not possible as evident from the attached pic at post # 1

[OneDotThree]

What type of symmetry you can use depends on what your actually modeling, if its non-buoyant flow through this cross section you can use 1/8 symmetry. If its buoyant flow through this cross section you can only use half symmetry. But you haven't told us what you are modeling, you've only shown the shape.

The fluid is gas, in particular a CH4/H2 mixture. I can't understand why it depends on the type of fluid.

Could you give me some more explanations? Thanks a lot

[OneDotThree]

In any case, I think I will semplify the section to a circular one and represent only 1/8 of the whole because I can't afford to increase the cell number...

[Far]

In any case, I think I will semplify the section to a circular one and represent only 1/8 of the whole because I can't afford to increase the cell number...

Then you must assess the error incurred due to this assumption!!!! Also compare the accuracy vs. computational cost.
In case of 1/8 model you must use periodic condition

[energy382]

Originally Posted by Far

Then you must assess the error incurred due to this assumption!!!! Also compare the accuracy vs. computational cost.
In case of 1/8 model you must use periodic condition

i completely agree with this conclusion! it's highly questionable to put the 1/8 model.

[OneDotThree]

Quote:
Originally Posted by Far
Then you must assess the error incurred due to this assumption!!!! Also compare the accuracy vs. computational cost.
In case of 1/8 model you must use periodic condition

i completely agree with this conclusion! it's highly questionable to put the 1/8 model.

The section I've posted above is very large in comparison to the other complex volumes in it. So I'm not so interested in what happens at the external boundaries. It's for this reason that I'm thinking to semplify the geometry.