Wall Shear Stress Dramatically Increases Everytime I Refine My Mesh

kartjo13 · April 22, 2019, 07:49

Hi there,

I'm currently modelling the fluid flow from a catheter (100% saline) into a vein (100% blood) on Star CCM as part of my thesis project.

I've chosen the SST (Menter) K-Omega to model the turbulent regions as the saline interacts with the blood in the vein. I've selected the all y+ wall treatment model and kept my y+~1. I'm currently attempting to do a mesh independence study (GCI to be exact), however every time I increase my mesh by a refinement ratio of 2 (doubling the number of cells), my Max Wall Shear Stress (WSS), Average WSS and Std WSS drastically increases and results in my GCI values to very large for WSS. Also, I believe that my mesh is refined enough as i've gone up to 8,000,000 cells and still my WSS values are still increasing.

I've attached a picture of what I'm trying to model and also the WSS plots with each mesh refinement.

I'm a complete newbie when it comes to CFD and Star CCM+, so any pointers or words of advice would very much be appreciated (and the fact i don't know who else to ask).

Thank you in advance

FMDenaro · April 22, 2019, 08:31

From the second diagram I suppose you restarted from the solutions of the previous grids, right? It seems you are getting towards a mesh independent solution, try to refine it further. In turbulence the wall stress requires to describe very well the velocity profile in the viscous sub-layer.

However, you should be aware of the fact that your problem would require a non-newtonian model.

kartjo13 · April 22, 2019, 13:16

Hi FMDenaro, that is correct i've restarted the solution from previous grids. How much further refinement (how many cells?) would you say is appropriate? as i have already gone up to 8,000,000 cells already.

Thanks.

steve lee · April 22, 2019, 23:02

You should check the Re number in your problem.
Check velocity first..
I'm not sure you chose right physics model

Katt · April 23, 2019, 02:56

Hi,

Is this a 2 phase problem? Then I believe this is a valid behaviour at the triple contact point (region where the meniscus touches the wall). The more you refine, the shear stresses on the wall increase.

This phenomena is predominant at large capillary numbers. I see you are using a turbulence modelling, so I suppose this should be even worse in your case.

If this is the case, try using a partial slip on the wall incase you are using a no-slip wall boundary. Choose the slip length equal to or slighty larger than the boundary cell size.

Look into the work of S. Afkhami et al. (2009), however, they dealt with capillary flows which is not similar to your flow conditions.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
sliding mesh problem in CFX	Saima	CFX	46	September 11, 2021 08:38
Natural convection in a closed domain STILL NEEDING help!	Yr0gErG	FLUENT	4	December 2, 2019 01:04
Export Wall Shear Stress Data	trooken	FLUENT	6	February 5, 2019 16:09
Wall Shear Stress prediction varies with fix y+	Jeeloong	FLUENT	0	November 24, 2016 18:17
Re: About wall shear stress	Mike	FLUENT	9	November 17, 2003 15:41

April 22, 2019, 08:31		#2
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,897 Rep Power: 73	From the second diagram I suppose you restarted from the solutions of the previous grids, right? It seems you are getting towards a mesh independent solution, try to refine it further. In turbulence the wall stress requires to describe very well the velocity profile in the viscous sub-layer. However, you should be aware of the fact that your problem would require a non-newtonian model.

April 22, 2019, 13:16		#3
kartjo13 New Member Jonathan Kartawihardja Join Date: Apr 2019 Posts: 4 Rep Power: 7	Hi FMDenaro, that is correct i've restarted the solution from previous grids. How much further refinement (how many cells?) would you say is appropriate? as i have already gone up to 8,000,000 cells already. Thanks.

April 22, 2019, 23:02		#4
steve lee Member Steve Join Date: Mar 2018 Location: South Korea Posts: 69 Rep Power: 8	You should check the Re number in your problem. Check velocity first.. I'm not sure you chose right physics model

April 23, 2019, 02:56		#5
Katt Member Katt Join Date: May 2017 Posts: 30 Rep Power: 9	Hi, Is this a 2 phase problem? Then I believe this is a valid behaviour at the triple contact point (region where the meniscus touches the wall). The more you refine, the shear stresses on the wall increase. This phenomena is predominant at large capillary numbers. I see you are using a turbulence modelling, so I suppose this should be even worse in your case. If this is the case, try using a partial slip on the wall incase you are using a no-slip wall boundary. Choose the slip length equal to or slighty larger than the boundary cell size. Look into the work of S. Afkhami et al. (2009), however, they dealt with capillary flows which is not similar to your flow conditions.