Large Valve Flow Analysis - Inflation Problem

sercan85 · March 4, 2010, 08:40

THE FLOW GEOMETRY IS

MY AIM:

Creating inflation on pipe and valve walls (~D=0,3m, first layer thickness=0,0001m, about 5 layers)
Also do this by using sweep method for 2 pipe sections.

NOTE: I didn't select very complex small surfaces on valve geometry as can be seen from picture, only major surfaces are selected.

PROBLEM I FACED:

I tried to use sweep method for two pipe sections, automatic method for valve section.

*Inflation couldn't be created on pipe walls by sweeping inflation. (no problem for some surfaces of
valve section walls)

I tried lots of parameters (increased number of sections, growth rate) for inflation but it didn't work.

Then I removed sweep method.

Mesher created tetrahedron meshes for pipe geometry.

Inflation created on walls (as seen in the picture below)

but 30-40 of 1 millon mesh elements have very bad aspect ratios, skewness exceeds 0,98 for these 30-40 elements.
This causes epsilon to diverge in k-e turbulent analysis at start.

SUMMARY:
IF SWEEP IS REMOVED, I CAN GET INFLATION, BUT SOME (30-40) CELLS THEN HAVE VERY BAD SKEWNESS.

What can I do? what parameters can I change?

I try to merge some surfaces on valve geometry but it deformates geometry a bit.

My aim is to get inflation without removing sweep method and get good skewness for all cells.

izumi11 · June 1, 2010, 05:08

Hi have you been able to solve the problem. I've had problems with this for the last 6 months. Have you tried changing the inflation method to post/pre?
Anyways if someone could please explain how to get rid of the highly skewed elements..

sercan85 · June 1, 2010, 15:02

Yes I have solved it by changing mesh sizing parameters.

From main sizing parameters, three parameters (something like max min tet face size, etc) must be reduced in order to fit the mesh to the geometry.

But the ratio between the three size must remain the same. If the order of magnitude of the one at the top is 1x, then the others should follow like 100x and 200x.

March 4, 2010, 08:40	Large Valve Flow Analysis - Inflation Problem	#1
sercan85 New Member Sercan Acarer Join Date: Mar 2010 Posts: 10 Rep Power: 16	THE FLOW GEOMETRY IS MY AIM: Creating inflation on pipe and valve walls (~D=0,3m, first layer thickness=0,0001m, about 5 layers) Also do this by using sweep method for 2 pipe sections. NOTE: I didn't select very complex small surfaces on valve geometry as can be seen from picture, only major surfaces are selected. PROBLEM I FACED: I tried to use sweep method for two pipe sections, automatic method for valve section. Inflation couldn't be created on pipe walls by sweeping inflation. (no problem for some surfaces of valve section walls) I tried lots of parameters (increased number of sections, growth rate) for inflation but it didn't work. Then I removed sweep method. Mesher created tetrahedron meshes for pipe geometry. Inflation created on walls (as seen in the picture below) but 30-40 of 1 millon mesh elements have very bad aspect ratios, skewness exceeds 0,98 for these 30-40 elements. This causes epsilon to diverge in k-e turbulent analysis at start. SUMMARY:* IF SWEEP IS REMOVED, I CAN GET INFLATION, BUT SOME (30-40) CELLS THEN HAVE VERY BAD SKEWNESS. What can I do? what parameters can I change? I try to merge some surfaces on valve geometry but it deformates geometry a bit. My aim is to get inflation without removing sweep method and get good skewness for all cells. Last edited by sercan85; March 5, 2010 at 04:54. Reason: improved visualisation

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Flow meter Design	CD adapco Group Marketing	Siemens	3	June 21, 2011 09:33
Problem with Convergence at high flow rates	Syed	Siemens	1	April 10, 2007 16:18
supersonic flow in compressor valve?	Perceptron	CFX	2	January 12, 2007 01:16
oscillating valve flow	Matthias Finke	Main CFD Forum	2	September 10, 2003 14:59
CFD Application in hydraulics valve flow	Roger Yang	Main CFD Forum	11	February 11, 1999 17:53

June 1, 2010, 05:08		#2
izumi11 New Member Join Date: Feb 2010 Posts: 11 Rep Power: 16	Hi have you been able to solve the problem. I've had problems with this for the last 6 months. Have you tried changing the inflation method to post/pre? Anyways if someone could please explain how to get rid of the highly skewed elements..

June 1, 2010, 15:02		#3
sercan85 New Member Sercan Acarer Join Date: Mar 2010 Posts: 10 Rep Power: 16	Yes I have solved it by changing mesh sizing parameters. From main sizing parameters, three parameters (something like max min tet face size, etc) must be reduced in order to fit the mesh to the geometry. But the ratio between the three size must remain the same. If the order of magnitude of the one at the top is 1x, then the others should follow like 100x and 200x.