Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Trouble coupling heat transfer physics for both fluids and solids

Please login with a confirmed email address before reporting spam

Hey all,

So I'm getting stuck with my model where I try to couple three different physics. The three physics are: Electric Currents, Heat transfer in Solids, and Heat transfer in Fluids. Basically what I'm trying to model is a silicon microbridge which is doped, which makes it so that it has a temperature gradient accross it when heated. I am able to couple and run the electric currents and heat transfer in solids physics, where the power dissipated from the ec physics is inputed into the ht physics as the heat source, and then the temperature that is calculated is used to calculate a new resistivity value for the silicon, until the model converges to a value for the temperature at the given voltage.

The problem arises when I try to incorporate the ht2 physics to mimic the effects of natural convection from the sorrounding gases. I set a temperature boundary condition within the ht2 physics which tells the program that the temperature of the microbridge is given by T, which is calculated in the ht physics. However, when I run this, I get the following error:

"Constraing found for variables in different Segregated groups. Try to merge these groups.
-Feature: Stationary Solver 1 (sol1/s1)
-Error: Constraint found for variables in different Segragated groups. Try to merge these groups."

Any ideas? Attached you will find my model. Thanks.


34 Replies Last Post 17 oct. 2016, 09:33 UTC−4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 avr. 2012, 03:07 UTC−4
Hi

you should NOT use 2 HT physics, but only use one HT and define a fluid and a solid region, right click the main HT node (you might need the HT module to get the coupled one pre-cooked by COMSOL

If you use 2 separate HT you get 2 Ts etc and then you must write far more equations by hand to knot the two physics together on the common boundaries

--
Good luck
Ivar
Hi you should NOT use 2 HT physics, but only use one HT and define a fluid and a solid region, right click the main HT node (you might need the HT module to get the coupled one pre-cooked by COMSOL If you use 2 separate HT you get 2 Ts etc and then you must write far more equations by hand to knot the two physics together on the common boundaries -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 mai 2012, 06:09 UTC−4
Hello,

I also have a similar problem, in my model i expect heat transfer between heated rock and moving fluid.
I observed in my model (I have ht and ht2 physics) there is heat transfer between rock and fluid.

Dear Ivar ,

you have mentioned use only HT, but how do I specify fluid domain if I use only HT? it doesnot allow me to access fluid domain,
Could you please give any hints to make my model work.

Thanka you
Hello, I also have a similar problem, in my model i expect heat transfer between heated rock and moving fluid. I observed in my model (I have ht and ht2 physics) there is heat transfer between rock and fluid. Dear Ivar , you have mentioned use only HT, but how do I specify fluid domain if I use only HT? it doesnot allow me to access fluid domain, Could you please give any hints to make my model work. Thanka you

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 mai 2012, 06:44 UTC−4
Dear Ricardo,

How did you fix your common boundary temperature. Could please explain me, i also have a similar issue but rock domain and fluid domain. I also used two physics heat transfer in solid and heat transfer in fluid. I seems that doesnt work. I guess that my boundary temperature is not working properly.

COuld you please let me know how did you solve this common boundary temperature issue.

thank you
Dear Ricardo, How did you fix your common boundary temperature. Could please explain me, i also have a similar issue but rock domain and fluid domain. I also used two physics heat transfer in solid and heat transfer in fluid. I seems that doesnt work. I guess that my boundary temperature is not working properly. COuld you please let me know how did you solve this common boundary temperature issue. thank you

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 mai 2012, 09:23 UTC−4
Hi

HT gives you some of the principal heat exchange means, you need to define your solid and your fluid domains (right click main physics + select the appropriate DOMAIN physics (as fluid has also transport heat elements)

if you want to solve the full fluid NS + HT then you need the Conjugated Heat transfer physics

--
Good luck
Ivar
Hi HT gives you some of the principal heat exchange means, you need to define your solid and your fluid domains (right click main physics + select the appropriate DOMAIN physics (as fluid has also transport heat elements) if you want to solve the full fluid NS + HT then you need the Conjugated Heat transfer physics -- Good luck Ivar


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 02:03 UTC−4
Dear Ivar,

Thank you very much for your reply.

Well, I hope I understand you. I manged to set both fluid and solid under one 'heat transfer in solid' physics. In my common boundary i keep Temperature common (T). That means rock and fluid shares the same temperature at the common boundary.
The thickness of the rock domain is 50m. It has the temperature of 180 0C at the other and other two boundaries has isothermal boundary conditions.

Fluid has specified low velocity(ensuring laminar flow) and it is expected to collect heat from the rock when it travels along the rock.Inlet temperature of the fluid 60 oC .
However when i simulated this I find no heat again(basically no temperature difference after 1 months. this cannot be true as i have constant heat at the far field.
My doubt is about the common temperature boundary.(I have used union,therefore i believe continuity is there).
I don't have access to conjugate heat transfer module therefore i would like to solve it using heat transfer in solid or fluid modules.

Thank you very much Ivar, your help is highly appreciated.
Dear Ivar, Thank you very much for your reply. Well, I hope I understand you. I manged to set both fluid and solid under one 'heat transfer in solid' physics. In my common boundary i keep Temperature common (T). That means rock and fluid shares the same temperature at the common boundary. The thickness of the rock domain is 50m. It has the temperature of 180 0C at the other and other two boundaries has isothermal boundary conditions. Fluid has specified low velocity(ensuring laminar flow) and it is expected to collect heat from the rock when it travels along the rock.Inlet temperature of the fluid 60 oC . However when i simulated this I find no heat again(basically no temperature difference after 1 months. this cannot be true as i have constant heat at the far field. My doubt is about the common temperature boundary.(I have used union,therefore i believe continuity is there). I don't have access to conjugate heat transfer module therefore i would like to solve it using heat transfer in solid or fluid modules. Thank you very much Ivar, your help is highly appreciated.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 02:48 UTC−4
Dear Ivar,

This is my model
Dear Ivar, This is my model


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 04:48 UTC−4
Hi

I see two issues with your model:
1) you impose the temperature T on the common boundary, that is not required, COMSOL considers such an internal boundary as a boundary with continuity on any flow, including heat flux.
2) there is nowhere to excit the heat flux, if your liquid is flowing towards X then it takes with it part of the heat with the fluid mass flow, you need to define a HT outflow boundary on the right fluid domain

Then why the symmetry BC as well as the 2nd Thermal isolation, this does not bring anything else than the default frirst T isloation BC.

Now I do not really know enough to be able to say if the results are right or wrong, your flow is low, but as you have only 10 mm of water the heat conduction over such long durations should not be an issue, you could checj the energy balance in /out over the boundaries

--
Good luck
Ivar
Hi I see two issues with your model: 1) you impose the temperature T on the common boundary, that is not required, COMSOL considers such an internal boundary as a boundary with continuity on any flow, including heat flux. 2) there is nowhere to excit the heat flux, if your liquid is flowing towards X then it takes with it part of the heat with the fluid mass flow, you need to define a HT outflow boundary on the right fluid domain Then why the symmetry BC as well as the 2nd Thermal isolation, this does not bring anything else than the default frirst T isloation BC. Now I do not really know enough to be able to say if the results are right or wrong, your flow is low, but as you have only 10 mm of water the heat conduction over such long durations should not be an issue, you could checj the energy balance in /out over the boundaries -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 05:06 UTC−4
Dear Ivar,

Thank you very much for your comments,

Well boudary 7 shoul have out flow boudary condition meaning convection dominant heat flow.

This is a simple model for geothermal wells.The Fluid path way represents a fracture inside the surface about 2000m depth. That is why we can resonaby assumethe symetric bounady condition.

Normally geothermal wells run for long period giving sufficient temperature differece. This is simple model (starting point) However this model should be able to give noticible temperature differecne in the long run.

Well, I take your point regarding the boundary Temperature T. I took it away but the result it same.

Thank you very much for your comments
Dear Ivar, Thank you very much for your comments, Well boudary 7 shoul have out flow boudary condition meaning convection dominant heat flow. This is a simple model for geothermal wells.The Fluid path way represents a fracture inside the surface about 2000m depth. That is why we can resonaby assumethe symetric bounady condition. Normally geothermal wells run for long period giving sufficient temperature differece. This is simple model (starting point) However this model should be able to give noticible temperature differecne in the long run. Well, I take your point regarding the boundary Temperature T. I took it away but the result it same. Thank you very much for your comments


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 07:25 UTC−4
Hi

indeed, still I'm woundering if there is not other things here:

1) remove your "T" temperature boundary condition on the internal boundary that is definitively wrong,

You can also try changing the mesh, as you gradients are only around the water granite interface:

try : a manual mesh,
edge "5" + Distribution 400
Mapped domain "1, 2" + DIstribution 4 for boundaries "3, 7" and 50 for boundaries " 1, 6"
Boundary Layer Domain 2: Boundyry layer Properties on boundary "4" with Number of layers 10, Stretch 2, Thick 0.1

This will give you a mapped mesh that follows closer your gradients without too many elements.

Then use a time stepping of the type

2^{range(0,1,8)}*1[day]

as heat diffusion follows a power law

For the rest I do not really see what is wrong, you have heat diffusivities alpha of water of 0.14 mm^2/s and that of Granite about 10 times greater. If you compare that to the velocity of the fluid 0.01m/s or 1E4 [s] about 2.8h to travel the 100 m and the time for heat to diffuse 10 [mm] across the fluid thickness:
Delta_t = h^2/alpha = (0.01[m])^2/0.14E-6[m^2/s] = about 12 min, you see that the fluid increases it's temperature only by a few degrees through the fluid film, this explains well the constant temperature observed at the symmetry layer of your water "film"

On the other side the 50 meters of granite has a heat diffusivity time of some (50[m])^2/13E-6[m^2/s] = 64 years
So infact, you do not need the constant heat T tmeperature, the stored heat in your volume is plenty for your 1 years simulation.

All this provided I havent done any errors in my rapid hand calculations ;) Pls check carefully.
It's often useful to do some hand check before simulating too much, or at lleast to try to reproduce the results
--
Good luck
Ivar
Hi indeed, still I'm woundering if there is not other things here: 1) remove your "T" temperature boundary condition on the internal boundary that is definitively wrong, You can also try changing the mesh, as you gradients are only around the water granite interface: try : a manual mesh, edge "5" + Distribution 400 Mapped domain "1, 2" + DIstribution 4 for boundaries "3, 7" and 50 for boundaries " 1, 6" Boundary Layer Domain 2: Boundyry layer Properties on boundary "4" with Number of layers 10, Stretch 2, Thick 0.1 This will give you a mapped mesh that follows closer your gradients without too many elements. Then use a time stepping of the type 2^{range(0,1,8)}*1[day] as heat diffusion follows a power law For the rest I do not really see what is wrong, you have heat diffusivities alpha of water of 0.14 mm^2/s and that of Granite about 10 times greater. If you compare that to the velocity of the fluid 0.01m/s or 1E4 [s] about 2.8h to travel the 100 m and the time for heat to diffuse 10 [mm] across the fluid thickness: Delta_t = h^2/alpha = (0.01[m])^2/0.14E-6[m^2/s] = about 12 min, you see that the fluid increases it's temperature only by a few degrees through the fluid film, this explains well the constant temperature observed at the symmetry layer of your water "film" On the other side the 50 meters of granite has a heat diffusivity time of some (50[m])^2/13E-6[m^2/s] = 64 years So infact, you do not need the constant heat T tmeperature, the stored heat in your volume is plenty for your 1 years simulation. All this provided I havent done any errors in my rapid hand calculations ;) Pls check carefully. It's often useful to do some hand check before simulating too much, or at lleast to try to reproduce the results -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 08:49 UTC−4
Dear Ivar,

Thank you very much for your comments. Your calcualtions are correct. In fact this is my problme also. When we have a granite of 50m with 180 C. It has sufficient amont of heat energy to generate considerable temperature diffference between the inlet and out let for a long period.

I have updated the model with your mesh and time stepping, Common boundary temperature was taken away.
But i don't see a noticible difference in the out let temperature. i still fell somthing odd it taking place in the rock and fluid interface.

Thank you very much your comments are highly appreiciated !

Dear Ivar, Thank you very much for your comments. Your calcualtions are correct. In fact this is my problme also. When we have a granite of 50m with 180 C. It has sufficient amont of heat energy to generate considerable temperature diffference between the inlet and out let for a long period. I have updated the model with your mesh and time stepping, Common boundary temperature was taken away. But i don't see a noticible difference in the out let temperature. i still fell somthing odd it taking place in the rock and fluid interface. Thank you very much your comments are highly appreiciated !


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 10:30 UTC−4
Hi

one thing I got wrong in my typing, the Boundary mesh domain should be domain 1, boundary 4 it is to get some thighter mesh elements at the fluid solid interface on the granite sides

It should not change that much the results though, just get a better flux estiamtion at the boundary.

if you now plot the lines 4 and 5 for T-333[k] you will see that rapidly (few days the temperature stabilises, then the slop decreases, we do not get to the assymptote yet due to the large amount in the Granite and the finite diffusivity in the granite. The same if we plot on the "4" internal boundary the heat flux "ht.tfluxy" ot the total normal energy flux (the same in this case) we see that after a few days the transient has sattled and we enter some steady state of a few tens of W/m^2, again one would need a stationary case to see the end stable value.

I'm doing currently NITF studies between air flow and a polymer (but ina much smaller scale and the minutes durations this works nicely and I'm in laminar flow as for your case. Normally I do not believe we would see anything different with nitf or the HT because of your low flow. Still, to be checked once.

By the way you should patch your version to get the latest one (see the main Web site of COMSOL), and these days COMSOL is shipping out the new 4.3 with further enhancements, particularly in the meshing

--
Good luck
Ivar
Hi one thing I got wrong in my typing, the Boundary mesh domain should be domain 1, boundary 4 it is to get some thighter mesh elements at the fluid solid interface on the granite sides It should not change that much the results though, just get a better flux estiamtion at the boundary. if you now plot the lines 4 and 5 for T-333[k] you will see that rapidly (few days the temperature stabilises, then the slop decreases, we do not get to the assymptote yet due to the large amount in the Granite and the finite diffusivity in the granite. The same if we plot on the "4" internal boundary the heat flux "ht.tfluxy" ot the total normal energy flux (the same in this case) we see that after a few days the transient has sattled and we enter some steady state of a few tens of W/m^2, again one would need a stationary case to see the end stable value. I'm doing currently NITF studies between air flow and a polymer (but ina much smaller scale and the minutes durations this works nicely and I'm in laminar flow as for your case. Normally I do not believe we would see anything different with nitf or the HT because of your low flow. Still, to be checked once. By the way you should patch your version to get the latest one (see the main Web site of COMSOL), and these days COMSOL is shipping out the new 4.3 with further enhancements, particularly in the meshing -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 11:09 UTC−4
Dear Ivar,

Thank you very much for your reply. Well I do observe nice variation of the out let temperature, meaning significant inlet and out let temperature difference however with gradual reduction with time. During first few hours I observe the kind of practical temperature out put
Normally we don\t observe this kind of temperature drop in the field. Normally geothermal wells are run for years without significant temperature drop.
I did few more simulations with higher thermal conductivities that are higher than the granite, then I observer the long-term higher temperature.

I am wondering what the parameter that I am missing here if my interface boundary is correct.

Your comments are highly appreciated Ivar.

Thank you
Dear Ivar, Thank you very much for your reply. Well I do observe nice variation of the out let temperature, meaning significant inlet and out let temperature difference however with gradual reduction with time. During first few hours I observe the kind of practical temperature out put Normally we don\t observe this kind of temperature drop in the field. Normally geothermal wells are run for years without significant temperature drop. I did few more simulations with higher thermal conductivities that are higher than the granite, then I observer the long-term higher temperature. I am wondering what the parameter that I am missing here if my interface boundary is correct. Your comments are highly appreciated Ivar. Thank you

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 21 mai 2012, 15:51 UTC−4
Hi

Indeed your temperature drops come from the initial conditions, what you see here is what happens if you dry your well, let it stabilise a few years, and then suddenly inject water at a constant temperature, assuming it flows in very rapidly and then once filled starts to flow at low speed.

When you run a well over longer periods its the steady state response you see, or at least what gives your background for such slow variations.

Another way to run such a simulation is to set up a steady state, then use that to initiate (as initial conditions) a transient when you change the inlet temperature (your 333K) by a small amount, lets say a sinus over 1-3 months period, or even having 2 frequencies, one for a dayly response one for a seasonal response. What I get from these values is that such a well need about a century to get to a true steady state, you need some fuzzy logic controller there to take into account this slow variation on top of the rest, and to learn from the response, that would be some nice projects, I have several colleagues who would love to work on that kind of problems (optimsing the flow to stabilise the reponse in such unknown environment ;).

I'm not used to the geothermal time values, but in fact its a type of small signal harmonic response you need with very looow frequency

--
Good luck
Ivar
Hi Indeed your temperature drops come from the initial conditions, what you see here is what happens if you dry your well, let it stabilise a few years, and then suddenly inject water at a constant temperature, assuming it flows in very rapidly and then once filled starts to flow at low speed. When you run a well over longer periods its the steady state response you see, or at least what gives your background for such slow variations. Another way to run such a simulation is to set up a steady state, then use that to initiate (as initial conditions) a transient when you change the inlet temperature (your 333K) by a small amount, lets say a sinus over 1-3 months period, or even having 2 frequencies, one for a dayly response one for a seasonal response. What I get from these values is that such a well need about a century to get to a true steady state, you need some fuzzy logic controller there to take into account this slow variation on top of the rest, and to learn from the response, that would be some nice projects, I have several colleagues who would love to work on that kind of problems (optimsing the flow to stabilise the reponse in such unknown environment ;). I'm not used to the geothermal time values, but in fact its a type of small signal harmonic response you need with very looow frequency -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 02:55 UTC−4
Thank you very much Ivar for your comments.

But still i didin't solve this problem, The result given by comsol is far away from the reality.

I feel that is am missing a parameter. I understand your comments in the last mail. Well it will definelty improve the things. That is not the way how geothermal wells are run. They keep on running for years without drastic temperature drop like we observe in Comsol model.

I don' t know whether we have scale some properties of rock to improve the thermal diffusivity. Then it will be again away from the real model.

With kind regards
Thank you very much Ivar for your comments. But still i didin't solve this problem, The result given by comsol is far away from the reality. I feel that is am missing a parameter. I understand your comments in the last mail. Well it will definelty improve the things. That is not the way how geothermal wells are run. They keep on running for years without drastic temperature drop like we observe in Comsol model. I don' t know whether we have scale some properties of rock to improve the thermal diffusivity. Then it will be again away from the real model. With kind regards

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 03:00 UTC−4
Hi

indeed I understand that, but COMSOLs results are from my understanding correct.

What is missing is a more realistic initial condition, foryour model, or perhaps a Stationary solving first to settle the initial conditions then a time series, knowing that the first few months of the time series are a transient case, particularly for your long lasting inertia terms.

Initial conditions build up is part of the simulation difficulties for such physical models

--
Good luck
Ivar
Hi indeed I understand that, but COMSOLs results are from my understanding correct. What is missing is a more realistic initial condition, foryour model, or perhaps a Stationary solving first to settle the initial conditions then a time series, knowing that the first few months of the time series are a transient case, particularly for your long lasting inertia terms. Initial conditions build up is part of the simulation difficulties for such physical models -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 03:08 UTC−4
Thanks Ivar,

That is a good idea, I am going to do, I have a small problem. I can run the model for stationary state. Then i can save that model. then i want to use that saved model for my transient simulation. Is possible. Could you please let me know the procedure for that
Thanks Ivar, That is a good idea, I am going to do, I have a small problem. I can run the model for stationary state. Then i can save that model. then i want to use that saved model for my transient simulation. Is possible. Could you please let me know the procedure for that

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 03:23 UTC−4
Hi

in v4 it is straightforward:
create a new study with stationary, then right click the study node and add atime dependent study, then solve (or ask for the defult solver sequence, then update the solver settings, time range etc). This will link both, COMSOL stores the stationary case and uses it by default as initial condition for your transient case.

One important thing though: if you have time dependent BCs (using explictely or implicitely "t"), then add a Parameter "t" and set it to either "0[s]" or any other constant value such that the time dependent BC's have correct values for the stationary solver case. the time dependent solver will overwrite the parameter t value so "t" remains driven by your solver settings for the time solving, while for the stationary, a "t" must often be defined, and the best place is a Parameter (and NOT a variable)

take a closer look at the solver nodes

--
Good luck
Ivar
Hi in v4 it is straightforward: create a new study with stationary, then right click the study node and add atime dependent study, then solve (or ask for the defult solver sequence, then update the solver settings, time range etc). This will link both, COMSOL stores the stationary case and uses it by default as initial condition for your transient case. One important thing though: if you have time dependent BCs (using explictely or implicitely "t"), then add a Parameter "t" and set it to either "0[s]" or any other constant value such that the time dependent BC's have correct values for the stationary solver case. the time dependent solver will overwrite the parameter t value so "t" remains driven by your solver settings for the time solving, while for the stationary, a "t" must often be defined, and the best place is a Parameter (and NOT a variable) take a closer look at the solver nodes -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 03:39 UTC−4
Ivar,


thanks Ivar,

stationary solving bring the out put temperature to 60.1. (Almost equal to the initial value). After that there is no improvement in the out put temperature with transient calculation.
Ivar, thanks Ivar, stationary solving bring the out put temperature to 60.1. (Almost equal to the initial value). After that there is no improvement in the out put temperature with transient calculation.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 23 mai 2012, 03:44 UTC−4
Hi

yes but the 60.1 is at the output, probably you have some gradients along the line, these are no at steady state. What you will notice is that the extra energy to get to the output temperature comes from the constant T at the bottom of your rock domain.

For your transient you should vary the inlet temperature by a few degrees slowly and see how the output temperature changes (starting from the stationary case). In fact that is doing a sensitivity analysis of the output temperature w.r.t the input temperature for a given rock volume. You can get your transfer function, response lag etc from that

--
Good luck
Ivar
Hi yes but the 60.1 is at the output, probably you have some gradients along the line, these are no at steady state. What you will notice is that the extra energy to get to the output temperature comes from the constant T at the bottom of your rock domain. For your transient you should vary the inlet temperature by a few degrees slowly and see how the output temperature changes (starting from the stationary case). In fact that is doing a sensitivity analysis of the output temperature w.r.t the input temperature for a given rock volume. You can get your transfer function, response lag etc from that -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 juil. 2012, 17:43 UTC−4

Hi

I see two issues with your model:
1) you impose the temperature T on the common boundary, that is not required, COMSOL considers such an internal boundary as a boundary with continuity on any flow, including heat flux.
2) there is nowhere to excit the heat flux, if your liquid is flowing towards X then it takes with it part of the heat with the fluid mass flow, you need to define a HT outflow boundary on the right fluid domain

Then why the symmetry BC as well as the 2nd Thermal isolation, this does not bring anything else than the default frirst T isloation BC.

Now I do not really know enough to be able to say if the results are right or wrong, your flow is low, but as you have only 10 mm of water the heat conduction over such long durations should not be an issue, you could checj the energy balance in /out over the boundaries

--
Good luck
Ivar


Hi,

I have question regarding the continuity across the internal boundary that is set as highly conductive layer in HT module. I am trying to simulate a problem similar to www.comsol.com/community/forums/general/thread/13676/, which I have domain 1, domain 2 separated by boundary 1. The geometry in union to ensure continuity, and boundary 1 is assigned as a highly conductive layer. I see from the equation of highly conductive layer that it only transfer heat tangential to the boundary. It is not taking normal flux from the domain. So the question is, does the continuity exist only between domains? How do I impose continuity flux between domains and the thin layer? The paper below: www.comsol.com/community/forums/general/thread/13676/ imposed an internal boundary condition, but I do not know how to do that in HT module.
[QUOTE] Hi I see two issues with your model: 1) you impose the temperature T on the common boundary, that is not required, COMSOL considers such an internal boundary as a boundary with continuity on any flow, including heat flux. 2) there is nowhere to excit the heat flux, if your liquid is flowing towards X then it takes with it part of the heat with the fluid mass flow, you need to define a HT outflow boundary on the right fluid domain Then why the symmetry BC as well as the 2nd Thermal isolation, this does not bring anything else than the default frirst T isloation BC. Now I do not really know enough to be able to say if the results are right or wrong, your flow is low, but as you have only 10 mm of water the heat conduction over such long durations should not be an issue, you could checj the energy balance in /out over the boundaries -- Good luck Ivar [/QUOTE] Hi, I have question regarding the continuity across the internal boundary that is set as highly conductive layer in HT module. I am trying to simulate a problem similar to http://www.comsol.com/community/forums/general/thread/13676/, which I have domain 1, domain 2 separated by boundary 1. The geometry in union to ensure continuity, and boundary 1 is assigned as a highly conductive layer. I see from the equation of highly conductive layer that it only transfer heat tangential to the boundary. It is not taking normal flux from the domain. So the question is, does the continuity exist only between domains? How do I impose continuity flux between domains and the thin layer? The paper below: http://www.comsol.com/community/forums/general/thread/13676/ imposed an internal boundary condition, but I do not know how to do that in HT module.

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 27 juil. 2012, 04:03 UTC−4
Hi

I can only give a partial answer, I would need to check the formulas better but I'm not by my COMSOL WS for some days ;)
By default COMSOL imposes continuity across the boundaries of adjacent domains (in geometry UNION mode). But if you add a special "boundary physics" as the highly condutive layer, it's as if you add a thin "virtual" domain between the two adjacent domains, and the previously "continuity" conditions on the boundary is adapted, as now you have an "up and "down" boundary towards the two respective adjacent domains.

Now the thin (heat) conductive layer (see the doc) is there to mimic a thin i(.e. Cu or Ag) layer in between two domains. Such a layer is assumed to have a high tangeantial conductivity, but as it's also assumed "thin" there is no T gradient across it so the T up and T down are the same, while you might get a small gradient tangeantially
These surface lyer assumptions will therefore interact and influence the heat gradient in the vicinity of the boundary.

You have also the thin resistive (heat) layer boundary condition. This one gives a temperature step across the boundary, and strong gradients tangeantially (comapred to the previous layer physics), as this layer is assumed of low heat conductance

--
Good luck
Ivar
Hi I can only give a partial answer, I would need to check the formulas better but I'm not by my COMSOL WS for some days ;) By default COMSOL imposes continuity across the boundaries of adjacent domains (in geometry UNION mode). But if you add a special "boundary physics" as the highly condutive layer, it's as if you add a thin "virtual" domain between the two adjacent domains, and the previously "continuity" conditions on the boundary is adapted, as now you have an "up and "down" boundary towards the two respective adjacent domains. Now the thin (heat) conductive layer (see the doc) is there to mimic a thin i(.e. Cu or Ag) layer in between two domains. Such a layer is assumed to have a high tangeantial conductivity, but as it's also assumed "thin" there is no T gradient across it so the T up and T down are the same, while you might get a small gradient tangeantially These surface lyer assumptions will therefore interact and influence the heat gradient in the vicinity of the boundary. You have also the thin resistive (heat) layer boundary condition. This one gives a temperature step across the boundary, and strong gradients tangeantially (comapred to the previous layer physics), as this layer is assumed of low heat conductance -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 31 juil. 2012, 19:45 UTC−4
Could you please take a look at a sample case i created. I am trying to test the flux across the internal boundary when treating it like a thin shell and make sure it is continuous. In this model, i applied a gradient perpendicular to boundary surface to observe the flux normal to an internal boundary that is assigned as highly conductive layer but with the same properties, and the result showed a discontinuous flux at the boundary. Is there anything I can do in ht or PDE to ensure continuity?
Could you please take a look at a sample case i created. I am trying to test the flux across the internal boundary when treating it like a thin shell and make sure it is continuous. In this model, i applied a gradient perpendicular to boundary surface to observe the flux normal to an internal boundary that is assigned as highly conductive layer but with the same properties, and the result showed a discontinuous flux at the boundary. Is there anything I can do in ht or PDE to ensure continuity?


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 19 sept. 2012, 18:55 UTC−4
Hello, I'm a grad student at Univ. of Michigan with a similar issue. I tried to solve for rough vacuum heat transfer due to gas flow from a head containing a micro-nozzle to a substrate plate. At first, I used the rarified gases slip model coupled to both a heat transfer for fluids and heat xfer for solids. The calculation failed and I then simplified to one heat xfer model taking into account both phases, but again to no avail.

I finally relied upon the conjugate heat xfer model which worked, however I was relegated to using a simple laminar flow option for my fluid flow. I'm quite sure the model is not accurate as Knudsen diffusion has to be taken into account at these low pressures and tiny dimensions. If you can direct me as to how to include the rarified gases models alongside the conjugate heat transfer, I would really appreciate it.

Many thanks.
Hello, I'm a grad student at Univ. of Michigan with a similar issue. I tried to solve for rough vacuum heat transfer due to gas flow from a head containing a micro-nozzle to a substrate plate. At first, I used the rarified gases slip model coupled to both a heat transfer for fluids and heat xfer for solids. The calculation failed and I then simplified to one heat xfer model taking into account both phases, but again to no avail. I finally relied upon the conjugate heat xfer model which worked, however I was relegated to using a simple laminar flow option for my fluid flow. I'm quite sure the model is not accurate as Knudsen diffusion has to be taken into account at these low pressures and tiny dimensions. If you can direct me as to how to include the rarified gases models alongside the conjugate heat transfer, I would really appreciate it. Many thanks.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 18 déc. 2012, 08:35 UTC−5
Dear Ivar,

I am simulating a heat exchange problem with comsol. The injected water goes thorough tube(surrounded by an annulus) and comes out along the annulus which is touching the hot rock. Then during the movement it absorbs heat from the bottom and when it comes up water exchange heat between the surrounding and the injected tube. I get the expected output temperature.
However when i want observe the thermal recovery of the rock after shutting down the injection. I don't get realistic results. The closer observation of the bottom rock temperature shows that during first day it has lost from 110 oC to 88 oC which is quite a lot. (Please run the model for 16 or 20 days)
Do you have any idea about this, I have attach the model for your consideration

thank you very much
i cannot attach the file it is only 5 kB. it gives me an error message. Could you please give your email address i will the file
Dear Ivar, I am simulating a heat exchange problem with comsol. The injected water goes thorough tube(surrounded by an annulus) and comes out along the annulus which is touching the hot rock. Then during the movement it absorbs heat from the bottom and when it comes up water exchange heat between the surrounding and the injected tube. I get the expected output temperature. However when i want observe the thermal recovery of the rock after shutting down the injection. I don't get realistic results. The closer observation of the bottom rock temperature shows that during first day it has lost from 110 oC to 88 oC which is quite a lot. (Please run the model for 16 or 20 days) Do you have any idea about this, I have attach the model for your consideration thank you very much i cannot attach the file it is only 5 kB. it gives me an error message. Could you please give your email address i will the file

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 20 déc. 2012, 02:14 UTC−5
Hi

you missed the model, Indeed 24h and 40 °C is quite a lot, you must be pushing quite some "cold" water through your system.

Have you ruun the case with plotting while solving for all steps ? it makes life easier often to debug when the couplings go wrong

--
Good luck
Ivar
Hi you missed the model, Indeed 24h and 40 °C is quite a lot, you must be pushing quite some "cold" water through your system. Have you ruun the case with plotting while solving for all steps ? it makes life easier often to debug when the couplings go wrong -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 4 août 2013, 23:41 UTC−4
Hello,

This is my first time to use Comsol and I am having trouble to couple heat transfer in unsaturated soils and fluid flow(water). I have saturated sand layer at the bottom and unsaturated silt layer covering the sand. We placed a u tube heat exchanger (closed loop which circulates the hot water) in the middle which goes on top of the sand, to heat the silt. when the soil heats up water evaporates from the saturated sand layer reaching the top and condensates there (because of hydraulic barrier). And this process causes downward capillary flow. This mechanism is consistent during the test. I set my geometry and I have the material properties and boundary conditions but since I am not familiar with this software, I can not move forward! Should I use heat transfer in solids and couple it with heat transfer in fluids? or is it better to use conjugate heating interface?

Any idea to help?

Thanks!

Tugce
Hello, This is my first time to use Comsol and I am having trouble to couple heat transfer in unsaturated soils and fluid flow(water). I have saturated sand layer at the bottom and unsaturated silt layer covering the sand. We placed a u tube heat exchanger (closed loop which circulates the hot water) in the middle which goes on top of the sand, to heat the silt. when the soil heats up water evaporates from the saturated sand layer reaching the top and condensates there (because of hydraulic barrier). And this process causes downward capillary flow. This mechanism is consistent during the test. I set my geometry and I have the material properties and boundary conditions but since I am not familiar with this software, I can not move forward! Should I use heat transfer in solids and couple it with heat transfer in fluids? or is it better to use conjugate heating interface? Any idea to help? Thanks! Tugce


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 sept. 2013, 06:12 UTC−4
Hi, I have a similar problem. I am trying to couple heat transfer in fluids with laminar flow. I have modeled the heat transfer correctly where i could get a clear transition zone or mushy region in the heat treatment of the alloy during casting but i cannot obtain the fluidic behaviour assuming that the alloy solidifies before leaving the casting tube. I am getting a uniform velocity throughout the tube instead of a gradually damping velocity. I don't know how to proceed. I also did exactly the way it has been described in the comsol model 'continuous casting' but it is not working on my geometry. Please help.
Hi, I have a similar problem. I am trying to couple heat transfer in fluids with laminar flow. I have modeled the heat transfer correctly where i could get a clear transition zone or mushy region in the heat treatment of the alloy during casting but i cannot obtain the fluidic behaviour assuming that the alloy solidifies before leaving the casting tube. I am getting a uniform velocity throughout the tube instead of a gradually damping velocity. I don't know how to proceed. I also did exactly the way it has been described in the comsol model 'continuous casting' but it is not working on my geometry. Please help.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 18 sept. 2013, 17:45 UTC−4
hello to everybody! im new here.
im looking a heat transfer liquid to transfer the heat of a high power led quickly to the heatsink by a 5mm copper tube.the distance from heat starting to the heatsink is short e.g 30cm like the cooling system of personals computer have this liquid. on internet i found various liquids but everybody selling in big quantities. could someone help me out?
thank in advance!
hello to everybody! im new here. im looking a heat transfer liquid to transfer the heat of a high power led quickly to the heatsink by a 5mm copper tube.the distance from heat starting to the heatsink is short e.g 30cm like the cooling system of personals computer have this liquid. on internet i found various liquids but everybody selling in big quantities. could someone help me out? thank in advance!

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 juin 2014, 20:23 UTC−4
I suspect it has something to do with needing the conjugate heat transfer model.

To test this, i made a simple geometry: rectangular tube (2-D Cartesian coordinates). Tube is 1 mm long and 0.1 mm high. A solid is placed in contact with the fluid underneath it and is the same dimensions.

Fluid properties, for simplicity, are:
rho = 1000 kg/m3
mu = 0.001 Pa*s
k=0.1 W/m-K
Cp = 1000 J/kg-K

Solid properties are:
rho=8000 kg/m3
k = 400 W/m-K
Cp = 300 J/kg-K

Inlet is the left fluid boundary; outlet is the right; symmetry on top. Inlet velocity is set to 0.1 m/s (Re = 20, thus laminar) and a temperature boundary condition at the inlet is set to 300K. A heat flux of 1 W/cm^2 is applied to the bottom of the solid, thus, since the domain is 1 mm long, 10 W/m is applied.

If we define an average operator on the outlet to calculate mass average temperature from T_out = aveop1(u*T)/aveop1(u), we can calculate energy balance from Q = m_dot*Cp*(T_out-T_in). Since we know 10 W is applied, we can calculate temperature rise: Q = 10 W = (1000 kg/m3)*(0.0001 m) *(0.1 m/s)*(1000 J/kg-K)*dT. Solving for T_out, we find: T_out = 301 K. However, when i ran the model, i get T_out = 300.0395 K.

Interestingly enough, if we remove the solid, and apply the heat flux directly to the fluid boundary, the simulation works, and we get energy balance.

Am i doing something wrong, or can conjugate conduction not be simulated in COMSOL without conjugate heat transfer module?

Attached is also my COMSOL file.

Thanks for your help,
I suspect it has something to do with needing the conjugate heat transfer model. To test this, i made a simple geometry: rectangular tube (2-D Cartesian coordinates). Tube is 1 mm long and 0.1 mm high. A solid is placed in contact with the fluid underneath it and is the same dimensions. Fluid properties, for simplicity, are: rho = 1000 kg/m3 mu = 0.001 Pa*s k=0.1 W/m-K Cp = 1000 J/kg-K Solid properties are: rho=8000 kg/m3 k = 400 W/m-K Cp = 300 J/kg-K Inlet is the left fluid boundary; outlet is the right; symmetry on top. Inlet velocity is set to 0.1 m/s (Re = 20, thus laminar) and a temperature boundary condition at the inlet is set to 300K. A heat flux of 1 W/cm^2 is applied to the bottom of the solid, thus, since the domain is 1 mm long, 10 W/m is applied. If we define an average operator on the outlet to calculate mass average temperature from T_out = aveop1(u*T)/aveop1(u), we can calculate energy balance from Q = m_dot*Cp*(T_out-T_in). Since we know 10 W is applied, we can calculate temperature rise: Q = 10 W = (1000 kg/m3)*(0.0001 m) *(0.1 m/s)*(1000 J/kg-K)*dT. Solving for T_out, we find: T_out = 301 K. However, when i ran the model, i get T_out = 300.0395 K. Interestingly enough, if we remove the solid, and apply the heat flux directly to the fluid boundary, the simulation works, and we get energy balance. Am i doing something wrong, or can conjugate conduction not be simulated in COMSOL without conjugate heat transfer module? Attached is also my COMSOL file. Thanks for your help,


Please login with a confirmed email address before reporting spam

Posted: 8 years ago 10 août 2016, 12:49 UTC−4
Hi!! please help me
my model is solid with heat flux on the upper face and fluid moving under the lower face , I used Conjugate heat transfer
,but the result show that there are no heat transfer between solid and liquid!
Hi!! please help me my model is solid with heat flux on the upper face and fluid moving under the lower face , I used Conjugate heat transfer ,but the result show that there are no heat transfer between solid and liquid!

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 12 sept. 2016, 01:49 UTC−4
does the fluid inlet share a vertex (point) with any other boundary?

i'm noticing now that if i make my inlet farther back by even the slightest amount, such that it no longer shares a vertex with any other boundary, it gives correct answer...

very weird.
does the fluid inlet share a vertex (point) with any other boundary? i'm noticing now that if i make my inlet farther back by even the slightest amount, such that it no longer shares a vertex with any other boundary, it gives correct answer... very weird.

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 19 sept. 2016, 22:24 UTC−4

Hi!! please help me
my model is solid with heat flux on the upper face and fluid moving under the lower face , I used Conjugate heat transfer
,but the result show that there are no heat transfer between solid and liquid!


same problem here anybody?

[QUOTE] Hi!! please help me my model is solid with heat flux on the upper face and fluid moving under the lower face , I used Conjugate heat transfer ,but the result show that there are no heat transfer between solid and liquid! [/QUOTE] same problem here anybody?

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 19 sept. 2016, 23:37 UTC−4
Did you see my solution/suggestion/question?
Did you see my solution/suggestion/question?

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 17 oct. 2016, 09:33 UTC−4
Yes I have the same problem, using the conjugate heat transfer I can't get any heat transfer from the fluid flow to the surrounding ground.

Can someone help me?
Yes I have the same problem, using the conjugate heat transfer I can't get any heat transfer from the fluid flow to the surrounding ground. Can someone help me?

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.