skip using local solution in mass source

lostking18 · May 8, 2018, 20:07

Hi, there:

I have known from CFX-Solver Modeling Guide Ch1.3.2.3. Mass (Continuity) Sources
https://www.sharcnet.ca/Software/Ans....html#i1299353

that the mass source will induce corresponding secondary source terms in all other equations; and if the continuity source is positive, the secondary source terms will use the specified values; and if the continuity source is negative, the specified values will be ignored by the solver, and local solution variables will be used.

Most importantly in multi-component flow we may force the solver to use the specified temperature even if the mass source is negative in 'MCF/Energy Sink Option'. May I know is there any way(such as expert parameters) that we can use to force the solver to skip the usage of the 'local solution' also in the single component flow? I just want to use the specified values in my single component simulation.

Or alternatively, is there any method we can use to skip the implementation of the 'secondary source term', just like the one did in Fluent, and purely introduce the mass source?

Thank you very much!

ghorrocks · May 9, 2018, 07:32

You have to define properties to the other variables when you create mass. It is not physical to not define it.

But rather than giving it a fixed value I suspect you might be thinking of something like the added mass has the other variables define as having the same value as currently exists in that location. If this is what you want to do I suspect if you simply define the temperature as "T" and U velocity as "u" etc then it should use the local value for the added mass. Is this what you want?

Opaque · May 9, 2018, 09:54

Would you mind explaining what you are trying to achieve?

When the flow is leaving the domain, i.e. a sink mass source, I do not understand how can we "physically valid" force the outgoing value.

Lets say the requested output is 0.5 mass fraction, but nearby mass fractions is 0.000001. How will the mathematical model come up with such amount?

Imagine trying to extract more mass that is available. Usually these approaches are a sign of ill-conditioned mathematical model

lostking18 · May 11, 2018, 13:25

Quote:

Originally Posted by ghorrocks

You have to define properties to the other variables when you create mass. It is not physical to not define it.

But rather than giving it a fixed value I suspect you might be thinking of something like the added mass has the other variables define as having the same value as currently exists in that location. If this is what you want to do I suspect if you simply define the temperature as "T" and U velocity as "u" etc then it should use the local value for the added mass. Is this what you want?

Hi, Ghorrocks:

Thanks for your reply! I actually want to do the opposite: I hope the solver to use my specified value even when mass source is negative. Do you know how to do that?
I know with the MCF/energy sink option we can use the specified value for the temperature. But I have no idea on how to do that for the velocity...

ghorrocks · May 11, 2018, 23:53

Defining the value for mass which is disappearing is usually not physical. So can you explain why are you trying to do it?

And why would you want to do it - the mass then disappears out the mass source, and the variables you defined along with it.

lostking18 · May 12, 2018, 19:44

Quote:

Originally Posted by Opaque

Would you mind explaining what you are trying to achieve?

When the flow is leaving the domain, i.e. a sink mass source, I do not understand how can we "physically valid" force the outgoing value.

Lets say the requested output is 0.5 mass fraction, but nearby mass fractions is 0.000001. How will the mathematical model come up with such amount?

Imagine trying to extract more mass that is available. Usually these approaches are a sign of ill-conditioned mathematical model

Hi, Opaque, Thanks for your always helpful answers! I want to do that because in such way I can easily cancelled out the effect of the secondary source term caused by the mass source. What I would prefer is actually the mass source without bringing any secondary source.

And the reason for me to get the mass source without bringing any secondary source is because now I have a residual in continuity equation caused purely by numerical errors, I want to introduce an additional source to compensate this residual. In this case the mass source seems to have no physical meanings and therefore I suppose it should not bring any further source in the other equations. Do you think the mass source will bring an additional source even when the usage is purely for numerical reasons? Thanks!

lostking18 · May 12, 2018, 19:47

Quote:

Originally Posted by ghorrocks

Defining the value for mass which is disappearing is usually not physical. So can you explain why are you trying to do it?

And why would you want to do it - the mass then disappears out the mass source, and the variables you defined along with it.

Hi, Ghorrocks, in my case I have a residual in continuity equation caused purely by numerical errors, I want to introduce an additional source to compensate this residual. In this case the mass source seems to have no physical meanings and therefore I suppose it should not bring any further source in the other equations. Do you think the mass source will bring an additional source even when the usage is purely for numerical reasons? Thanks!

ghorrocks · May 13, 2018, 07:50

You cannot compensate a residual error by introducing a source term. They are totally different things. So what you are proposing will not work.

Your problem is poor convergence, not anything to do with the secondary sources linked to a mass source term.

This looks like a classic XY problem: http://xyproblem.info/

Can you explain the convergence problem you are having and what you are trying to model? I am sure you will find the issue has nothing to do with the secondary source terms for mass sources at all.

Opaque · May 14, 2018, 09:33

Tinkering with the continuity equation is dangerous since the "residual's residual" you are observing is either lack of convergence (as Glenn point out) or an inconsistency in the implementation.

You are better off looking at the location of the maximum continuity residual in CFD-Post, and see what it may cause the lack of convergence: recirculation zone, poor mesh quality, etc. Fix the possible issue, and re-run.

Tinkering with the continuity equation may produce a different pressure profile, and it cascades down from there. I would not do that.

May 8, 2018, 20:07	skip using local solution in mass source	#1
lostking18 Member phd Join Date: Oct 2013 Posts: 76 Rep Power: 13	Hi, there: I have known from CFX-Solver Modeling Guide Ch1.3.2.3. Mass (Continuity) Sources https://www.sharcnet.ca/Software/Ans....html#i1299353 that the mass source will induce corresponding secondary source terms in all other equations; and if the continuity source is positive, the secondary source terms will use the specified values; and if the continuity source is negative, the specified values will be ignored by the solver, and local solution variables will be used. Most importantly in multi-component flow we may force the solver to use the specified temperature even if the mass source is negative in 'MCF/Energy Sink Option'. May I know is there any way(such as expert parameters) that we can use to force the solver to skip the usage of the 'local solution' also in the single component flow? I just want to use the specified values in my single component simulation. Or alternatively, is there any method we can use to skip the implementation of the 'secondary source term', just like the one did in Fluent, and purely introduce the mass source? Thank you very much!

May 9, 2018, 07:32		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	You have to define properties to the other variables when you create mass. It is not physical to not define it. But rather than giving it a fixed value I suspect you might be thinking of something like the added mass has the other variables define as having the same value as currently exists in that location. If this is what you want to do I suspect if you simply define the temperature as "T" and U velocity as "u" etc then it should use the local value for the added mass. Is this what you want? lostking18 likes this. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 11, 2018, 23:53		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Defining the value for mass which is disappearing is usually not physical. So can you explain why are you trying to do it? And why would you want to do it - the mass then disappears out the mass source, and the variables you defined along with it. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

May 13, 2018, 07:50		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	You cannot compensate a residual error by introducing a source term. They are totally different things. So what you are proposing will not work. Your problem is poor convergence, not anything to do with the secondary sources linked to a mass source term. This looks like a classic XY problem: http://xyproblem.info/ Can you explain the convergence problem you are having and what you are trying to model? I am sure you will find the issue has nothing to do with the secondary source terms for mass sources at all. __________________ Note: I do not answer CFD questions by PM. CFD questions should be posted on the forum.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
pimpleDyMFoam computation randomly stops	babapeti	OpenFOAM Running, Solving & CFD	5	January 24, 2018 06:28
[foam-extend.org] problem when installing foam-extend-1.6	Thomas pan	OpenFOAM Installation	7	September 9, 2015 22:53
[swak4Foam] swak4foam building problem	GGerber	OpenFOAM Community Contributions	54	April 24, 2015 17:02
IcoFoam parallel woes	msrinath80	OpenFOAM Running, Solving & CFD	9	July 22, 2007 03:58
Could anybody help me see this error and give help	liugx212	OpenFOAM Running, Solving & CFD	3	January 4, 2006 19:07

May 9, 2018, 09:54		#3
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	Would you mind explaining what you are trying to achieve? When the flow is leaving the domain, i.e. a sink mass source, I do not understand how can we "physically valid" force the outgoing value. Lets say the requested output is 0.5 mass fraction, but nearby mass fractions is 0.000001. How will the mathematical model come up with such amount? Imagine trying to extract more mass that is available. Usually these approaches are a sign of ill-conditioned mathematical model

May 14, 2018, 09:33		#9
Opaque Senior Member Join Date: Jun 2009 Posts: 1,880 Rep Power: 33	Tinkering with the continuity equation is dangerous since the "residual's residual" you are observing is either lack of convergence (as Glenn point out) or an inconsistency in the implementation. You are better off looking at the location of the maximum continuity residual in CFD-Post, and see what it may cause the lack of convergence: recirculation zone, poor mesh quality, etc. Fix the possible issue, and re-run. Tinkering with the continuity equation may produce a different pressure profile, and it cascades down from there. I would not do that.