Mass Transfer: save results for the expressions from every solver step?

AliLemprex · August 11, 2014, 10:11

Hello everyone,

I am currently encountering a problem in CFX while modelling mass transfer in a fiber bundle.
Flow has been calculated beforehand and is given as initial values file so the solver will only solve mass transfer equations.
Boundary conditions are the inlet concentration and initial concentration in the fluid. Mass transfer shall occur from fibers to the fluid.
My diffusion coefficient for mass transfer is being calculated from current concentration values.

After a few solver steps, concentration drops below zero though and the solver will abort. Has anyone else ever encountered a similar phenomenon and could give any impulses on how to resolve this issue?

I have tried to get an insight on why my values drop all of a sudden and would like to save the solver results for several expressions from every solver step. Is there a possibility to do so?

Thank you very much!

ghorrocks · August 11, 2014, 19:52

This question is essentially this FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

AliLemprex · August 15, 2014, 09:44

Thanks Glenn, but my initial problem would be a bit different from what I find in the FAQ.

For my mass diffusion coefficient, I need to calculate some other values, one of which is my concentration. At some point my concentration is changing to negative values. This causes the solver to quit.

I have achieved solutions with the same set of equations and boundary conditions, but on other geometry.

CFX post fails to point out problematic regions, so I would like to get an output of several expression values for each solver step to help identify where my problem may lie. Are there any strategies on how to tackle such a problem?

Thanks again!

ghorrocks · August 17, 2014, 07:29

Your problem is your simulation is numerically unstable and liable to diverge. The FAQ gives tips on how to deal with numerical instability.

If you have had similar models run successfully before then either the new geometry has introduced a new flow feature which is harder to solve (maybe a choked flow point, or a higher Reynolds number or some other non-linearity) - or more likely your mesh quality on the new geometry is poorer quality.

So I recommend you improve your mesh quality - this always helps.

If you want CFX to point out problem regions then add residuals to your output file and save a results file just before it crashes. You should see high residuals in the problem region.

Opaque · August 18, 2014, 09:55

Based on your description, I assume you are solving the "mass transfer" problem using additional variables, and not the multicomponent material/fluid approach.

If the material composition creates density gradient different from the ones used to solve the previous fluid flow conditions, the results of this combination is unreliable at best. If the density gradients are different between simulations, and you are no longer solving the fluid flow (momentum + mass conservation), the mass conservation is not being satisfied in your new simulation.

August 11, 2014, 10:11	Mass Transfer: save results for the expressions from every solver step?	#1
AliLemprex New Member Join Date: Mar 2014 Posts: 7 Rep Power: 12	Hello everyone, I am currently encountering a problem in CFX while modelling mass transfer in a fiber bundle. Flow has been calculated beforehand and is given as initial values file so the solver will only solve mass transfer equations. Boundary conditions are the inlet concentration and initial concentration in the fluid. Mass transfer shall occur from fibers to the fluid. My diffusion coefficient for mass transfer is being calculated from current concentration values. After a few solver steps, concentration drops below zero though and the solver will abort. Has anyone else ever encountered a similar phenomenon and could give any impulses on how to resolve this issue? I have tried to get an insight on why my values drop all of a sudden and would like to save the solver results for several expressions from every solver step. Is there a possibility to do so? Thank you very much!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Error message: Insufficient Catalogue Size	Paresh Jain	CFX	33	August 16, 2024 06:09
Help for the small implementation in turbulence model	shipman	OpenFOAM Programming & Development	25	March 19, 2014 11:08
Question about heat transfer coefficient setting for CFX	Anna Tian	CFX	1	June 16, 2013 07:28
Source terms due to mass transfer in VOF model	ssamton	FLUENT	0	March 5, 2012 01:03
How to calculate Volumetric Mass transfer coefficient using CFX?	tuks_123	CFX	2	July 22, 2010 02:15

August 11, 2014, 19:52		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	This question is essentially this FAQ: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

August 15, 2014, 09:44		#3
AliLemprex New Member Join Date: Mar 2014 Posts: 7 Rep Power: 12	Thanks Glenn, but my initial problem would be a bit different from what I find in the FAQ. For my mass diffusion coefficient, I need to calculate some other values, one of which is my concentration. At some point my concentration is changing to negative values. This causes the solver to quit. I have achieved solutions with the same set of equations and boundary conditions, but on other geometry. CFX post fails to point out problematic regions, so I would like to get an output of several expression values for each solver step to help identify where my problem may lie. Are there any strategies on how to tackle such a problem? Thanks again!

August 17, 2014, 07:29		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Your problem is your simulation is numerically unstable and liable to diverge. The FAQ gives tips on how to deal with numerical instability. If you have had similar models run successfully before then either the new geometry has introduced a new flow feature which is harder to solve (maybe a choked flow point, or a higher Reynolds number or some other non-linearity) - or more likely your mesh quality on the new geometry is poorer quality. So I recommend you improve your mesh quality - this always helps. If you want CFX to point out problem regions then add residuals to your output file and save a results file just before it crashes. You should see high residuals in the problem region.

August 18, 2014, 09:55		#5
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	Based on your description, I assume you are solving the "mass transfer" problem using additional variables, and not the multicomponent material/fluid approach. If the material composition creates density gradient different from the ones used to solve the previous fluid flow conditions, the results of this combination is unreliable at best. If the density gradients are different between simulations, and you are no longer solving the fluid flow (momentum + mass conservation), the mass conservation is not being satisfied in your new simulation.