[Rockda]

I need help concerning the right way to set up a conjugate heat transfer condition at the interface between a liquid zone and a solid.

In my case, I have water at 300K flowing (2m/s) along the surface of a solid rectangular piece (channel case type). I have two well delineate zones, a fluid zone and a solid zone. I created two named selections at the boundary of the two zones, one at the boundary surface of the fluid zone and the other at the boundary of the solid zone.

Then using the mesh interface panel, I created a “coupled wall" between the two zones which should allow the conjugate heat transfer.

Once I activated the themal resolution, I patched the liquid filling the liquid zone at 300K and patched the solid zone at 500K. I’m using the multiphasic Eulerian model. In a first time I didn’t activate the boiling model because I wanted to make sure that the heat transfer between the solid and the liquid was working correctly.

My problem comes from the boundary condition definition. When I created a mesh interface, using the two zones boundaries, 4 new surfaces have been created: “a boundary zone” and an “interface wall zone” for each liquid and solid zone, respectively wall 13, wall 14, wall 6 and wall 6 shadow.

I well understand that the two “interface wall zone” created, and put in “coupled” automatically in the boundary conditions panel allow the heat exchange between the two zones but the two "boundary zones" need to be complete in either with dirichlet condition (wall temperature given) , or by a neumann condition (Flux at the wall given). his is not what I want, because I’m expecting to see my solid zone getting colder along the simulation and consequently the heat flux exchanged at the boundary between liquid and solid changing along the simulation (indeed if my solid part temperature is different the heat flux exchanged should be different). But here, the heat flux will stay the same and my solid will stay at 500K.

Is anyone has already succeeded to set up that kind of “quenching” configuration? I’m thinking that it is maybe compulsory to use UDF for that kind of problem ?

Thank you very much

[LuckyTran]

This problem is doable without a complicated udf. You simply need to define the coupled interface the proper way. I have done it before, but it was awhile ago so I do not remember the exact steps now.

Sorry this post may teach you exactly how to fix your problem but what you are trying to do is doable! You're on the right track, just need to find the right buttons!

[syavash]

Originally Posted by Rockda

I need help concerning the right way to set up a conjugate heat transfer condition at the interface between a liquid zone and a solid.

In my case, I have water at 300K flowing (2m/s) along the surface of a solid rectangular piece (channel case type). I have two well delineate zones, a fluid zone and a solid zone. I created two named selections at the boundary of the two zones, one at the boundary surface of the fluid zone and the other at the boundary of the solid zone.

Then using the mesh interface panel, I created a “coupled wall" between the two zones which should allow the conjugate heat transfer.

Once I activated the themal resolution, I patched the liquid filling the liquid zone at 300K and patched the solid zone at 500K. I’m using the multiphasic Eulerian model. In a first time I didn’t activate the boiling model because I wanted to make sure that the heat transfer between the solid and the liquid was working correctly.

My problem comes from the boundary condition definition. When I created a mesh interface, using the two zones boundaries, 4 new surfaces have been created: “a boundary zone” and an “interface wall zone” for each liquid and solid zone, respectively wall 13, wall 14, wall 6 and wall 6 shadow.

I well understand that the two “interface wall zone” created, and put in “coupled” automatically in the boundary conditions panel allow the heat exchange between the two zones but the two "boundary zones" need to be complete in either with dirichlet condition (wall temperature given) , or by a neumann condition (Flux at the wall given). his is not what I want, because I’m expecting to see my solid zone getting colder along the simulation and consequently the heat flux exchanged at the boundary between liquid and solid changing along the simulation (indeed if my solid part temperature is different the heat flux exchanged should be different). But here, the heat flux will stay the same and my solid will stay at 500K.

Is anyone has already succeeded to set up that kind of “quenching” configuration? I’m thinking that it is maybe compulsory to use UDF for that kind of problem ?

Thank you very much

Dear Rockda,

Could you manage to solve your problem?
I have faced a similar one and I need to know how to resolve it.
tnx.

[misagh]

conjugate heat transfer should be simulated in 3D.In 2D in fluent in b.c you can not set up shell conduction between solid and the liquid but when you use 3D a new b.c will be appear in fluent in b.c named SHELL CONDUCTION you can just enable this choise and give it the thermal conductivity of solid and also its thickness.(check it out please)
bests

[syavash]

Originally Posted by misagh

conjugate heat transfer should be simulated in 3D.In 2D in fluent in b.c you can not set up shell conduction between solid and the liquid but when you use 3D a new b.c will be appear in fluent in b.c named SHELL CONDUCTION you can just enable this choise and give it the thermal conductivity of solid and also its thickness.(check it out please)
bests

Dear Misagh,

I appreciate your comment. Please confirm(or deny) what I suppose to be right.
I have a thick wall in adjacent with fluid flow and case is 2-D. I have a b.c. with the name "wall" and another with the name "wall shadow" in Fluent. I want to model solid(wall) cooling process (air blows and solid gets cooled). I can specify air inlet temperature but I can not do the same for wall b.c.(at least I couldn't find any field to specify wall initial temperature) while I have checked "coupled" in wall b.c. panel.
According to what you just mentioned, I would be able to specify an initial value as wall temperature if the case was 3-D?

[misagh]

dear syavash
actually your problem is not setting initial temp. since it is available in both 2.d and 3.d .in this way:define>boundary condition>zone(sth like wall)>type:wall>set>thermal<thermal condition>tempreture
3d model should be considered in order to make fluent understand the conduction heat transfer process(the shell conduction choice available in just 3d not 2.d)
bests

[Markat]

Originally Posted by misagh

dear syavash
actually your problem is not setting initial temp. since it is available in both 2.d and 3.d .in this way:define>boundary condition>zone(sth like wall)>type:wall>set>thermal<thermal condition>tempreture
3d model should be considered in order to make fluent understand the conduction heat transfer process(the shell conduction choice available in just 3d not 2.d)
bests

I am new in CFD and I have the same problem as syavash.
But I can not see why shell conduction must be used. If you mesh the wall thickness then you probably don't have to use shell conduction. I Think that in a 2D model you can have conduction but only in thickness direction.

I also think that the temperature you mention is a constant temperature. I probably think that this problem is not a steady state problem because when the solid is being cooled the boundary conditions are accually being changed.

All I have mentioned are only estimation due to my lack of experience with CFD.

[syavash]

I think the same as Markat. I suppose in 2-D the conduction is normal to wall and if you have a solid zone there should be no problem. I somehow managed to solve specifying the initial temperature of wall by using patch option. But there is no change in wall temperature as the problem is solved(Transient of course). I don't know where the problem lies!

[misagh]

dear markat
what you mentioned is true.i just said the way some one can set the wall b.c in fluent.yes the tempreture of wall is not cte so the convection b.c or mixed b.c should be considered in fluent DEFINE>.... or other b.c conditions depending to the problem.
and i have read in fluents tutorial that shell conduction can be solved in 3.d modeling.actually it is one of the problems of the fluent not to be capable of solving it in 2.d.
best regards

[Markat]

Misagh thank you for you reply.

Could you please check also my thread for any suggestions?
What step should be done in order to define appropriate BC, that model conjugate heat transfer? Is it impossible to have a CHT model in 2D at all?
I have noticed (in my model) that both the wall and the wall shadow are from the same material (both solid). Is it right? I actually think that one should be fluid and the other solid but I can not define it in this way.

Thank you in advance

[misagh]

dear Markat,in order to not mislead you please check these links.i found them useful.
https://www.sharcnet.ca/Software/Flu...c_bc_wall.html (((7.3.13.3)))
http://hpce.iitm.ac.in/website/Manua...ug/node573.htm
i hope they are related to your case...

[dsking]

I am not sure if it is the same for 2d (i have only done conjugate heat transfer in 3d), but make sure when you create the geometry in Ansys that the fluid zone and solid zone are under the same "Multi-Body" part. Then when you mesh them, you should have conformity at the solid/fluid interface.

[Markat]

I have already done this so that the meshing in the common line-surface between the solid and the fluid is the same for both fluid ans solid.
Would you suggest to mesh to solid part or not? What's the exact difference?

[syavash]

Originally Posted by dsking

I am not sure if it is the same for 2d (i have only done conjugate heat transfer in 3d), but make sure when you create the geometry in Ansys that the fluid zone and solid zone are under the same "Multi-Body" part. Then when you mesh them, you should have conformity at the solid/fluid interface.

Actually I am using Gambit to generate geometry and mesh. I have checked 3d tutorial of Fluent and there is no difference in B.C. type.
There is something that might be interesting. I have assigned a heat source to my solid zone instead of specifying an initial temperature and it seems working! but continuity residual does not converge.

[misagh]

Originally Posted by syavash

Actually I am using Gambit to generate geometry and mesh. I have checked 3d tutorial of Fluent and there is no difference in B.C. type.
There is something that might be interesting. I have assigned a heat source to my solid zone instead of specifying an initial temperature and it seems working! but continuity residual does not converge.

i am not sure assigning heat source be a good way.you had better check the results with a theoretical or sth else to validate your work otherwise it cant be reliable.

[syavash]

Originally Posted by misagh

i am not sure assigning heat source be a good way.you had better check the results with a theoretical or sth else to validate your work otherwise it cant be reliable.

You are tight, I was just trying to understand why fluent does not solve fluid-solid heat transfer. I wonder why it does not act just like 3-D case.

[misagh]

Originally Posted by Markat

I have already done this so that the meshing in the common line-surface between the solid and the fluid is the same for both fluid ans solid.
Would you suggest to mesh to solid part or not? What's the exact difference?

as mentioned in the above links,shell conduction method(which can be utilized only in 3d model) is better to be implemented for THIN WALLS.actually fluent itself uses mesh to solve the conductivity equations.
but for other cases you can mesh the solid zone(in 2d or 3d no difference) and also the fluid zone.

[Markat]

Originally Posted by misagh

as mentioned in the above links,shell conduction method(which can be utilized only in 3d model) is better to be implemented for THIN WALLS.actually fluent itself uses mesh to solve the conductivity equations.
but for other cases you can mesh the solid zone(in 2d or 3d no difference) and also the fluid zone.

This is exactly what I would like to now. Is conduction the only equation that fluent solves when solis is meshed? Is there any way to model conduction in a 2D model (1axial conduction)?

[misagh]

Originally Posted by Markat

This is exactly what I would like to now. Is conduction the only equation that fluent solves when solis is meshed? Is there any way to model conduction in a 2D model (1axial conduction)?

conduction seems to be the only equ. that fluent solves for solid zone but for the fluid it solves other equ. like energy,continuity...
yes meshing the solid zone is the only way in 2.d for conjugate heat transfer to get rid of 3.d model(SHELL CONDUCTION b.c that appears only in 3.d)

[syavash]

Originally Posted by misagh

as mentioned in the above links,shell conduction method(which can be utilized only in 3d model) is better to be implemented for THIN WALLS.actually fluent itself uses mesh to solve the conductivity equations.
but for other cases you can mesh the solid zone(in 2d or 3d no difference) and also the fluid zone.

Concerning this issue, there might be some problems which should be modeled in 2-D. For example a hot airfoil in front of cold air flow needs to be thermally analyzed and conjugate heat transfer matters.