[ICEM] 2 way FSI, CFX + Mechanical + ICEM, Remesh, Deformation due to fluid flow

binoyjjohn · July 30, 2013, 10:26

Hi,

This is my first thread in CFD online.
I am new to ICEM. Have been working with Ansys WB 14.5 and CFX for a while. Now i am attempting a 2D, 2 way FSI, with CFX and Mechanical, to simulate a tilt pad thrust bearing with 10 micron min film thickness. While i run the analysis, it exits giving mesh error. I came to know that i need to use remeshing.

But, my problem here is that i cannot use remesh option in CFX, as it is a 2 way FSI and the structural deformation is due to fluid flow.
I have ICEM, but do not know how to integrate it into WB. I wish to do almost all activities in WB. Also, from these forums i came to know that i need to do scripting in ICEM using TCL, again that i do not know.

Can anyone post a remeshing script of ICEM. I will rework back to suit my requirement. Or else please advice.

Thanks,

Rexyrox

binoyjjohn · August 10, 2013, 06:50

No updates to my question?? How sad!

mvoss · August 13, 2013, 04:23

hey

some things good to know:
What´s the mesh error?
How do you model the mesh movement?
Can you post a picture of your geometry?
Did you run a solo case to check only for the meshmovement?
I suppose you use Icem-Hexa!?
I´ve done quite a lot of 2W-FSI simulations, also with ICEM, but never needed to do a remeshing - i solved all issues with the mesh folding etc. by defining a clever meshstiffness function via CEL.

Matthias

binoyjjohn · August 13, 2013, 07:12

Hi Matt,

Thanks for the response. The mesh error is 'negative volume'.
I have certain limitation in sharing the model. So i am sharing some representative figure. There are 2 pictures, a 3D model and a 2D model
The 2D model is a simplified one when a cut is taken at the section shown in the 3D model.

The oil gap is of the order of 15 microns.
Yes, i ran solo case. The mesh distortion too much and due to perturbation along pad pivot and due to structural deformation. The oil film has a many layers and slight change in surrounding geometry distorts it very much.
Yes i use ICEM hexa. Varying mesh stiffness will not work, i think you might have understood why.

http://imageshack.us/photo/my-images/824/e56w.jpg/
http://imageshack.us/photo/my-images/703/4ozm.jpg/
Thanks,

Binoy

mvoss · August 13, 2013, 07:51

Can you prescribe the mesh movement with some arrows in the pictures? Actually i didn´t get it. Sorry.

binoyjjohn · August 13, 2013, 10:12

The collar has 2 DOF, it can rotate (in 2d, its linear velocity) and move up or down. The tilt pad can tilt along the pivot.
The mesh movement at the fluid region happens due to motion of these regions. Sorry..the mesh is only representative. I know its bad quality without inflation.
http://imageshack.us/photo/my-images/35/e6pg.jpg/
Thanks,
Binoy

binoyjjohn · June 2, 2014, 07:46

Hi,

Again back to the same problem. I see that when i import fluid hex mesh from ICEM to CFX and do the run, the CFX solver comes out with an error, for mesh displacement. When i look at the mesh in post, i see that some of the nodes in fluid mesh in CFX near the solid boundary has collapsed for unknown reason. Have anyone seen this before, i am not getting much help from ansys. CAn anyone help??

mvoss · June 2, 2014, 08:03

Try to increase the calculation of the meshmovement down to like e.g 1E-07 with enough iterations to succeed (>15). If the bc of the domains get close to each other, you will face serious problems, since the angels and volumes are going to get worse and worse. Does the mesh fold at the corners or somewhere arbitrary along the bc (picture...)?

Sujit · June 25, 2014, 10:00

I am also facing same problem of meshing object of 15 micron thickness.
What if use upscalled part for meshing and then transform the mesh?

Link https://onedrive.live.com/redir?resi...CE93EAD3%21628

I have scalled this geometry by 100x in y axis..

July 30, 2013, 10:26	2 way FSI, CFX + Mechanical + ICEM, Remesh, Deformation due to fluid flow	#1
binoyjjohn New Member Binoy Join Date: Jul 2013 Location: Canada Posts: 7 Rep Power: 13	Hi, This is my first thread in CFD online. I am new to ICEM. Have been working with Ansys WB 14.5 and CFX for a while. Now i am attempting a 2D, 2 way FSI, with CFX and Mechanical, to simulate a tilt pad thrust bearing with 10 micron min film thickness. While i run the analysis, it exits giving mesh error. I came to know that i need to use remeshing. But, my problem here is that i cannot use remesh option in CFX, as it is a 2 way FSI and the structural deformation is due to fluid flow. I have ICEM, but do not know how to integrate it into WB. I wish to do almost all activities in WB. Also, from these forums i came to know that i need to do scripting in ICEM using TCL, again that i do not know. Can anyone post a remeshing script of ICEM. I will rework back to suit my requirement. Or else please advice. Thanks, Rexyrox

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Want Impeller Driven Fluid Flow: What Inlet and Outlet BC to use for Centrifugal Pump	Zev Xavier	FLUENT	3	May 9, 2016 07:42
Is 2-way FSI valve modeling with seperate fluid domains possible?	Maurosso	FLUENT	4	March 7, 2013 09:56
ATTENTION! Reliability problems in CFX 5.7	Joseph	CFX	14	April 20, 2010 16:45
ICEM 10 mesh scale foe CFX	Korsh Mik	CFX	2	October 7, 2005 10:07
CFX 4 for two continuous fluid flow solution	sinjae	CFX	1	May 19, 2001 04:07

August 10, 2013, 06:50		#2
binoyjjohn New Member Binoy Join Date: Jul 2013 Location: Canada Posts: 7 Rep Power: 13	No updates to my question?? How sad!

August 13, 2013, 04:23		#3
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	hey some things good to know: What´s the mesh error? How do you model the mesh movement? Can you post a picture of your geometry? Did you run a solo case to check only for the meshmovement? I suppose you use Icem-Hexa!? I´ve done quite a lot of 2W-FSI simulations, also with ICEM, but never needed to do a remeshing - i solved all issues with the mesh folding etc. by defining a clever meshstiffness function via CEL. Matthias

August 13, 2013, 07:12		#4
binoyjjohn New Member Binoy Join Date: Jul 2013 Location: Canada Posts: 7 Rep Power: 13	Hi Matt, Thanks for the response. The mesh error is 'negative volume'. I have certain limitation in sharing the model. So i am sharing some representative figure. There are 2 pictures, a 3D model and a 2D model The 2D model is a simplified one when a cut is taken at the section shown in the 3D model. The oil gap is of the order of 15 microns. Yes, i ran solo case. The mesh distortion too much and due to perturbation along pad pivot and due to structural deformation. The oil film has a many layers and slight change in surrounding geometry distorts it very much. Yes i use ICEM hexa. Varying mesh stiffness will not work, i think you might have understood why. http://imageshack.us/photo/my-images/824/e56w.jpg/ http://imageshack.us/photo/my-images/703/4ozm.jpg/ Thanks, Binoy

August 13, 2013, 07:51		#5
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	Can you prescribe the mesh movement with some arrows in the pictures? Actually i didn´t get it. Sorry.

August 13, 2013, 10:12		#6
binoyjjohn New Member Binoy Join Date: Jul 2013 Location: Canada Posts: 7 Rep Power: 13	The collar has 2 DOF, it can rotate (in 2d, its linear velocity) and move up or down. The tilt pad can tilt along the pivot. The mesh movement at the fluid region happens due to motion of these regions. Sorry..the mesh is only representative. I know its bad quality without inflation. http://imageshack.us/photo/my-images/35/e6pg.jpg/ Thanks, Binoy

June 2, 2014, 07:46		#7
binoyjjohn New Member Binoy Join Date: Jul 2013 Location: Canada Posts: 7 Rep Power: 13	Hi, Again back to the same problem. I see that when i import fluid hex mesh from ICEM to CFX and do the run, the CFX solver comes out with an error, for mesh displacement. When i look at the mesh in post, i see that some of the nodes in fluid mesh in CFX near the solid boundary has collapsed for unknown reason. Have anyone seen this before, i am not getting much help from ansys. CAn anyone help??

June 2, 2014, 08:03		#8
mvoss Senior Member Matthias Voß Join Date: Mar 2009 Location: Berlin, Germany Posts: 449 Rep Power: 20	Try to increase the calculation of the meshmovement down to like e.g 1E-07 with enough iterations to succeed (>15). If the bc of the domains get close to each other, you will face serious problems, since the angels and volumes are going to get worse and worse. Does the mesh fold at the corners or somewhere arbitrary along the bc (picture...)?

June 25, 2014, 10:00		#9
Sujit New Member sujit Join Date: Apr 2014 Posts: 4 Rep Power: 12	I am also facing same problem of meshing object of 15 micron thickness. What if use upscalled part for meshing and then transform the mesh? Link https://onedrive.live.com/redir?resi...CE93EAD3%21628 I have scalled this geometry by 100x in y axis..