Residence Time Distribution

brajh11 · November 18, 2011, 17:56

Anybody know how is possible to implement in CFX Residence Time Distribution of flue gases in a combustion chamber?

ghorrocks · November 19, 2011, 06:14

You want to impose this on a simulation or extract this from a completed run as post-processing?

brajh11 · November 19, 2011, 07:54

I want extract this from a completed run as post-processing

ghorrocks · November 20, 2011, 06:31

If the simulation has already been run then put some streamlines on it (assuming it is steady state) and colour the streamlines by time. This will be your residence time distribution.

If the simulation has not run then add a user scalar with convection equation only, but a source term of 1 [s^-1]. If you do a plot of this variable it will be the residence time of that region of the flow. This will work in both steady and transient flows.

ghorrocks · November 20, 2011, 17:57

Here's some links to other posts on this issue. It is always worthwhile doing a search on the forum as many questions have been asked before.

http://www.cfd-online.com/Forums/cfx...ution-cfx.html

http://www.cfd-online.com/Forums/cfx...ime-trace.html

Note both of these links use a source term of 1 [s/s] which then means your AV has the units of seconds which makes more sense than mine which is unitless.

Danial Q · November 20, 2011, 22:43

HI ALL,
Anyone please suggest that if it is possible to model a molten droplet of metal at high temp say 2700 C hitting a slolid metal surface at some temp say 500C. this molten droplet solidifes and there lies an interface between the two. I have to model "this interface" which acts different with its variable thickness. while this thickness depends on many solid and molten droplet's properties. i have to model the heat transfer of this interface with its corresponding thickness along the surface.
Please make comment if it would be right to define the domain interface between the solid and fluid (molten droplet) domains as my required interface? i am confused about it because it would negate the idea of variable thickness of that interface modeling. someone told me that it is beyond the range of CFX modeling. he suggested me to solve it with explicit dynamics rather CFX. is dt true?
Your help would be highy appreciated

.

Thanks

ghorrocks · November 21, 2011, 05:54

Talk to CFX support about solidification modelling. And please do not hijack threads, start a new thread.

Danial Q · November 21, 2011, 17:42

sorry, it ws first time ..i dint realize dt..

emily.imdieke · February 3, 2014, 13:29

Am I correct in assuming that if the flow is transient you can't just simply color the streamlines and get the residence time distribution or does it work the same way?

ghorrocks · February 3, 2014, 17:17

You can colour streamlines by any local variable. So if you have a residence time distribution you can colour it by that.

emily.imdieke · February 3, 2014, 17:23

So basically if my timestep is 0.001, but if my streamline time distribution goes to 1.7 seconds according to CFX Post, it is just taking the streamlines that started at that individual timestep and calculating them out to completion at 1.7 seconds which is when they would leave the geometry?

ghorrocks · February 3, 2014, 17:34

Of course in a transient simulation streamlines are just what the instantaneous flow field looks like - a real particle will not necessarily follow that path as the flow field is changing in time. So be careful trying to read too much into streamlines in transient simulations.

Having said that, I think your description of how it works is correct. You should do some checks to be sure.

emily.imdieke · February 3, 2014, 18:06

Thank you for the response. Upon further looking at the other threads that you posted it seems as though someone else agrees as well that transient streamline times are difficult to interpret.

This being said. I did try the additional variable method and had mine set as a volumetric scalar with a source of 1. I was originally trying to look at this to find residence time but did not figure out how each of the different variables worked nor did I realize that graphing them would be a good option. My questions then are what would be the location that this graph would be plotted on, and what are the correct variables to be plotted? Such as what is the different between Particle Time and Particle Traveling Time?

ashtonJ · November 9, 2014, 19:17

Dear Glenn,
For residence time calculation, the unit of source term for AV is kg/m^3. How is it possible to define a source term for AV with unit [s/s]??

November 18, 2011, 17:56	Residence Time Distribution	#1
brajh11 New Member Bostjan Join Date: Apr 2009 Posts: 24 Rep Power: 17	Anybody know how is possible to implement in CFX Residence Time Distribution of flue gases in a combustion chamber?

November 19, 2011, 07:54		#3
brajh11 New Member Bostjan Join Date: Apr 2009 Posts: 24 Rep Power: 17	I want extract this from a completed run as post-processing Last edited by brajh11; November 21, 2011 at 12:28.

November 20, 2011, 22:43	Modelling suggestions??	#6
Danial Q Senior Member Danial Join Date: Nov 2011 Posts: 179 Rep Power: 15	HI ALL, Anyone please suggest that if it is possible to model a molten droplet of metal at high temp say 2700 C hitting a slolid metal surface at some temp say 500C. this molten droplet solidifes and there lies an interface between the two. I have to model "this interface" which acts different with its variable thickness. while this thickness depends on many solid and molten droplet's properties. i have to model the heat transfer of this interface with its corresponding thickness along the surface. Please make comment if it would be right to define the domain interface between the solid and fluid (molten droplet) domains as my required interface? i am confused about it because it would negate the idea of variable thickness of that interface modeling. someone told me that it is beyond the range of CFX modeling. he suggested me to solve it with explicit dynamics rather CFX. is dt true? Your help would be highy appreciated . Thanks

November 21, 2011, 17:42	HI glen	#8
Danial Q Senior Member Danial Join Date: Nov 2011 Posts: 179 Rep Power: 15	sorry, it ws first time ..i dint realize dt..

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
TimeVaryingMappedFixedValue	irishdave	OpenFOAM Running, Solving & CFD	32	June 16, 2021 07:55
Pathline colored by time = residence time	pkleb	FLUENT	1	February 26, 2021 09:00
Superlinear speedup in OpenFOAM 13	msrinath80	OpenFOAM Running, Solving & CFD	18	March 3, 2015 06:36
How to get residence time distribution?	bigfans	FLUENT	0	April 27, 2009 10:39
Residence Time Distribution	rajeev	FLUENT	11	July 26, 2001 14:49

November 19, 2011, 06:14		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	You want to impose this on a simulation or extract this from a completed run as post-processing?

November 20, 2011, 06:31		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	If the simulation has already been run then put some streamlines on it (assuming it is steady state) and colour the streamlines by time. This will be your residence time distribution. If the simulation has not run then add a user scalar with convection equation only, but a source term of 1 [s^-1]. If you do a plot of this variable it will be the residence time of that region of the flow. This will work in both steady and transient flows.

November 20, 2011, 17:57		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Here's some links to other posts on this issue. It is always worthwhile doing a search on the forum as many questions have been asked before. http://www.cfd-online.com/Forums/cfx...ution-cfx.html http://www.cfd-online.com/Forums/cfx...ime-trace.html Note both of these links use a source term of 1 [s/s] which then means your AV has the units of seconds which makes more sense than mine which is unitless.

November 21, 2011, 05:54		#7
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Talk to CFX support about solidification modelling. And please do not hijack threads, start a new thread.

February 3, 2014, 13:29		#9
emily.imdieke New Member Emily Imdieke Join Date: Apr 2013 Posts: 20 Rep Power: 13	Am I correct in assuming that if the flow is transient you can't just simply color the streamlines and get the residence time distribution or does it work the same way?

February 3, 2014, 17:17		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	You can colour streamlines by any local variable. So if you have a residence time distribution you can colour it by that.

February 3, 2014, 17:23		#11
emily.imdieke New Member Emily Imdieke Join Date: Apr 2013 Posts: 20 Rep Power: 13	So basically if my timestep is 0.001, but if my streamline time distribution goes to 1.7 seconds according to CFX Post, it is just taking the streamlines that started at that individual timestep and calculating them out to completion at 1.7 seconds which is when they would leave the geometry?

February 3, 2014, 17:34		#12
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Of course in a transient simulation streamlines are just what the instantaneous flow field looks like - a real particle will not necessarily follow that path as the flow field is changing in time. So be careful trying to read too much into streamlines in transient simulations. Having said that, I think your description of how it works is correct. You should do some checks to be sure.

February 3, 2014, 18:06		#13
emily.imdieke New Member Emily Imdieke Join Date: Apr 2013 Posts: 20 Rep Power: 13	Thank you for the response. Upon further looking at the other threads that you posted it seems as though someone else agrees as well that transient streamline times are difficult to interpret. This being said. I did try the additional variable method and had mine set as a volumetric scalar with a source of 1. I was originally trying to look at this to find residence time but did not figure out how each of the different variables worked nor did I realize that graphing them would be a good option. My questions then are what would be the location that this graph would be plotted on, and what are the correct variables to be plotted? Such as what is the different between Particle Time and Particle Traveling Time?

November 9, 2014, 19:17		#14
ashtonJ Senior Member Ashkan Javadzadegan Join Date: Sep 2010 Posts: 255 Rep Power: 17	Dear Glenn, For residence time calculation, the unit of source term for AV is kg/m^3. How is it possible to define a source term for AV with unit [s/s]??