ICEM Prism Layer transition between surface with prism layers and one without

TWaung · September 25, 2009, 16:20

Hi,

I'm hoping someone could suggest a prism layer strategy for the following problem. I am generating an unstructured tetrahedral mesh in ANSYS ICEM 11.0.

I am modeling a 3D air intake geometry. My domain extends downstream of the true outflow plane of the intake geometry because I expect limited regions of flow separation at the outflow plane.

Upstream of the outflow plane the domain surfaces are defined as no-slip walls, and downstream of the outflow plane I have defined the surfaces to be slip walls.

I want the prism layer along the intake walls to extend all the way to the outflow plane. In order to do this, I have had to let the prism layer extend onto the slip wall surfaces.

I am finding that my residuals (P-Mass, u,v,w momentum) are very high in the regions where there are prism layers along slip walls.

I would like to not use prism layers along slip walls but still have prism layers along the no-slip walls right up to the outflow plane.

Is this possible? Does anyone have suggestions on how this could be done?

Thanks!

Tim

PSYMN · October 12, 2009, 15:54

If you just stop the prism at the internal outflow boundary, the pyramids and tetras attached would likely be very poor quality. I recomend putting a much taller initial height in the slip area so the prism can transition to a more reasonable aspect ratio before you delete them and then fill the region with bottom up tetra... Actually, if you increase the height on these prisms, you may not have the problem with your solver...

Another option would be to break the slip section into two regions. Use the upstream region, adjacent to the viscous flow domain to increase the height of the prisms. Then set no prisms on the lower section. You will see the prisms start to swell and then pyramid caps will form to stair step the prisms away.

Simon

PSYMN · October 12, 2009, 15:56

I just noticed I didn't explain the problem... If your boundary layer has very thin prisms along the wall, these will not cap well because the pyramids and tetras would have to be very high aspect ratio, which would translate to poor internal angles, high skewness, etc.

Thickening the layer in the transition regions reduces these problems.

September 25, 2009, 16:20	ICEM Prism Layer transition between surface with prism layers and one without	#1
TWaung New Member Join Date: Mar 2009 Posts: 14 Rep Power: 17	Hi, I'm hoping someone could suggest a prism layer strategy for the following problem. I am generating an unstructured tetrahedral mesh in ANSYS ICEM 11.0. I am modeling a 3D air intake geometry. My domain extends downstream of the true outflow plane of the intake geometry because I expect limited regions of flow separation at the outflow plane. Upstream of the outflow plane the domain surfaces are defined as no-slip walls, and downstream of the outflow plane I have defined the surfaces to be slip walls. I want the prism layer along the intake walls to extend all the way to the outflow plane. In order to do this, I have had to let the prism layer extend onto the slip wall surfaces. I am finding that my residuals (P-Mass, u,v,w momentum) are very high in the regions where there are prism layers along slip walls. I would like to not use prism layers along slip walls but still have prism layers along the no-slip walls right up to the outflow plane. Is this possible? Does anyone have suggestions on how this could be done? Thanks! Tim

October 12, 2009, 15:54	...Not very timely, but better late than never?	#2
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	If you just stop the prism at the internal outflow boundary, the pyramids and tetras attached would likely be very poor quality. I recomend putting a much taller initial height in the slip area so the prism can transition to a more reasonable aspect ratio before you delete them and then fill the region with bottom up tetra... Actually, if you increase the height on these prisms, you may not have the problem with your solver... Another option would be to break the slip section into two regions. Use the upstream region, adjacent to the viscous flow domain to increase the height of the prisms. Then set no prisms on the lower section. You will see the prisms start to swell and then pyramid caps will form to stair step the prisms away. Simon

October 12, 2009, 15:56	In case I didn't explain enough...	#3
PSYMN Senior Member Simon Pereira Join Date: Mar 2009 Location: Ann Arbor, MI Posts: 2,663 Blog Entries: 1 Rep Power: 47	I just noticed I didn't explain the problem... If your boundary layer has very thin prisms along the wall, these will not cap well because the pyramids and tetras would have to be very high aspect ratio, which would translate to poor internal angles, high skewness, etc. Thickening the layer in the transition regions reduces these problems.