pressure drop, laminar, use of prism?

galap · November 11, 2009, 12:09

Hi community,

I'm confused by some results calculated with Ansys CFX. I meshed a 3D structure (air flows through) with tetras and calculate the pressure drop at several inlet normal speeds. So far, results fits quite well to practical measurement data. Now I repeat these calculations with an additional prism mesh (4 layers, tetra settings remaining equal) for a better velocity resolution on surfaces. Though I get pressure drop values, increasing for instance from 600mbar/m to 900mbar/m. Quite a difference and none relation to practical measurements!

I don't know to handle that. The flow is laminar, help manual says prism mesh captures boundary layer physics. Maybe prism should only be used when I have turbulent flows..? Im confused, better solution on surfaces should give me more realistic results, isn't it..

ghorrocks · November 11, 2009, 21:16

Laminar flows have boundary layers too. If you have thin laminar boundary layers then a prism mesh is still a good idea.

If when you change the mesh your results change a lot then it sounds like you are a long way from a mesh converged solution. You probably need a much finer grid.

galap · November 12, 2009, 03:50

Thanks, thats a good point. My residuals were set to 10-4 and prism quality aren't good.. I refine my mesh and check results.

November 11, 2009, 12:09	pressure drop, laminar, use of prism?	#1
galap Member Fabian E. Join Date: Nov 2009 Posts: 38 Rep Power: 16	Hi community, I'm confused by some results calculated with Ansys CFX. I meshed a 3D structure (air flows through) with tetras and calculate the pressure drop at several inlet normal speeds. So far, results fits quite well to practical measurement data. Now I repeat these calculations with an additional prism mesh (4 layers, tetra settings remaining equal) for a better velocity resolution on surfaces. Though I get pressure drop values, increasing for instance from 600mbar/m to 900mbar/m. Quite a difference and none relation to practical measurements! I don't know to handle that. The flow is laminar, help manual says prism mesh captures boundary layer physics. Maybe prism should only be used when I have turbulent flows..? Im confused, better solution on surfaces should give me more realistic results, isn't it..

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Total and static pressure drop	Janakiram	Siemens	1	February 8, 2009 01:32
Pressure Drop at entrance of a rotor-stator.	Resnick	Main CFD Forum	0	November 20, 2007 14:50
Does star cd takes reference pressure?	monica	Siemens	1	April 19, 2007 11:26
pressure drop	siddiqui	FLUENT	2	March 15, 2006 06:13
specified pressure drop	subhra	Main CFD Forum	0	May 29, 2005 12:43

November 11, 2009, 21:16		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,700 Rep Power: 143	Laminar flows have boundary layers too. If you have thin laminar boundary layers then a prism mesh is still a good idea. If when you change the mesh your results change a lot then it sounds like you are a long way from a mesh converged solution. You probably need a much finer grid.

November 12, 2009, 03:50		#3
galap Member Fabian E. Join Date: Nov 2009 Posts: 38 Rep Power: 16	Thanks, thats a good point. My residuals were set to 10-4 and prism quality aren't good.. I refine my mesh and check results.