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.

Heat Transfer in Solids (motion of objects inside a domain)

Please login with a confirmed email address before reporting spam

Hi,

I am trying to model a simple 2D transient heat transfer problem. For this I move a sphere with a constant boundary temperature trough a plane making use of the Deformed Mesh and the Heat Transfer in Solids Interfaces. But what I can not understand is...

1. Is it possible to move the sphere through the plane, if the both are solid (... does FEM allowed this motion)?

2. How does Comsol calculate the temperature in the space left of the overlap area between the time steps? Diffusion? (see attachment!)

best regards,

Antoni Artinov


9 Replies Last Post 2016年9月1日 GMT-4 02:46
Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月26日 GMT-4 11:57
Hi Antoni,

I wouldn’t use the Deformed Mesh interface. Instead use Heat Transfer and Solid Mechanics and prescribe a displacement (or velocity) to the sphere in the solid mechanics interface. That way the temperature is attached to the material points on the solid and it “moves” as the solid moves.

Nagi Elabbasi
Veryst Engineering
Hi Antoni, I wouldn’t use the Deformed Mesh interface. Instead use Heat Transfer and Solid Mechanics and prescribe a displacement (or velocity) to the sphere in the solid mechanics interface. That way the temperature is attached to the material points on the solid and it “moves” as the solid moves. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月26日 GMT-4 12:20
Dear Nagi,

thank you for your reply! I am not that familiar with the Solid Mechanics Interface, but I will give it a try. With the Deformed Mesh the temperature is also attached to the material points on the solid (as a boundary condition). I also use identity pairs and continuity bc exactly as in this blog:

www.comsol.com/blogs/deformed-mesh-interfaces-rotations-and-linear-translations/

The thing is, I do not understand, why is it possible to model a motion of a solid object inside a solid object (in the meaning, that you can not move a solid object through another one without breaking it)? Why can I do this with FEM?

...so assuming it is possible to do it with Comsol, the other question will be, how does Comsol calculate the temperature in the remaining area left of the overlap area (this is the part of the sphere for t1 for in which we didn't callculate the temperature, but for t2 we need full temperature field). When the object moves from point 1 to point 2 (t1->t2 see attachment) there is a gap in the calculated temperature field for t1 (the area where we assumed/set the temperature of the sphere), so how Comsol calculates the temperature field for t2 without this information. Do these cells get just an initial value or are calculated as a conduction in the time interval from t1 to t2?

I really hope somebody could explain how it works internally :).

Best regards
Antoni
Dear Nagi, thank you for your reply! I am not that familiar with the Solid Mechanics Interface, but I will give it a try. With the Deformed Mesh the temperature is also attached to the material points on the solid (as a boundary condition). I also use identity pairs and continuity bc exactly as in this blog: https://www.comsol.com/blogs/deformed-mesh-interfaces-rotations-and-linear-translations/ The thing is, I do not understand, why is it possible to model a motion of a solid object inside a solid object (in the meaning, that you can not move a solid object through another one without breaking it)? Why can I do this with FEM? ...so assuming it is possible to do it with Comsol, the other question will be, how does Comsol calculate the temperature in the remaining area left of the overlap area (this is the part of the sphere for t1 for in which we didn't callculate the temperature, but for t2 we need full temperature field). When the object moves from point 1 to point 2 (t1->t2 see attachment) there is a gap in the calculated temperature field for t1 (the area where we assumed/set the temperature of the sphere), so how Comsol calculates the temperature field for t2 without this information. Do these cells get just an initial value or are calculated as a conduction in the time interval from t1 to t2? I really hope somebody could explain how it works internally :). Best regards Antoni

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月26日 GMT-4 18:11
Hi Antoni,

The Deformed Mesh interface should work as the nice Blog you pointed to shows. If I understand your problem correctly there is no overlaps or gaps. The region ahead of the sphere will be compressed and the region behind it will be expanded. Or put it another way, the mesh in of front and behind the sphere will deform. Therefore a point right in front of the sphere will move and remain in front of the sphere at it travels, and the same for a point behind the sphere, and there will be no confusion at any point in time if a certain coordinate is “sphere” or “plate”. I hope that helps.

Nagi Elabbasi
Veryst Engineering
Hi Antoni, The Deformed Mesh interface should work as the nice Blog you pointed to shows. If I understand your problem correctly there is no overlaps or gaps. The region ahead of the sphere will be compressed and the region behind it will be expanded. Or put it another way, the mesh in of front and behind the sphere will deform. Therefore a point right in front of the sphere will move and remain in front of the sphere at it travels, and the same for a point behind the sphere, and there will be no confusion at any point in time if a certain coordinate is “sphere” or “plate”. I hope that helps. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月29日 GMT-4 03:15
Hi Nagi,

thank you for the explanation, it really helped me, because I was quite confused about the way it works. If you don't mind, I would like to ask you one more question :). In the doc file "Deformed Geometry vs. Moving Mesh" is written:

"In the Deformed Geometry interface, the material does not follow the change in shape. Deformation of the geometry boundaries therefore corresponds to addition or removal of material."

But in my case I don't really change the geometry boundaries, because the sphere is moving inside the volume, so the question is, do I get the right solution (how does Comsol calculate the undependent variables in the deformed cells?), because after streching/distorting the elements we need to do something to get the information for the new part of the streched cell, because it is bigger now. Let's say for instance we solve the heat transfer. If Comsol just interpolate between the already known nodes temperature and do not take the diffussion into account, we will get an error for the temperature which I wouldn't neglect.

Thanks in advance! Wish you a good start in the week :).

Best regards
Antoni
Hi Nagi, thank you for the explanation, it really helped me, because I was quite confused about the way it works. If you don't mind, I would like to ask you one more question :). In the doc file "Deformed Geometry vs. Moving Mesh" is written: "In the Deformed Geometry interface, the material does not follow the change in shape. Deformation of the geometry boundaries therefore corresponds to addition or removal of material." But in my case I don't really change the geometry boundaries, because the sphere is moving inside the volume, so the question is, do I get the right solution (how does Comsol calculate the undependent variables in the deformed cells?), because after streching/distorting the elements we need to do something to get the information for the new part of the streched cell, because it is bigger now. Let's say for instance we solve the heat transfer. If Comsol just interpolate between the already known nodes temperature and do not take the diffussion into account, we will get an error for the temperature which I wouldn't neglect. Thanks in advance! Wish you a good start in the week :). Best regards Antoni

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月31日 GMT-4 11:39
Hi Antoni,

I wouldn’t use Deformed Geometry for your problem, I would use Moving Mesh, or Solid Mechanics. That way the mesh movement means material movement. The temperature would move also since it is associated with material points. If your deformed mesh volume is bigger like you described then you are adding thermal energy to the model (assuming there are no heat transfer/sources/sinks, so the temperature remains constant in the sphere). If you are just moving the sphere not enlarging it, even if it’s a very big displacement, then there should be no problem. I hope that helps.

Nagi Elabbasi
Veryst Engineering
Hi Antoni, I wouldn’t use Deformed Geometry for your problem, I would use Moving Mesh, or Solid Mechanics. That way the mesh movement means material movement. The temperature would move also since it is associated with material points. If your deformed mesh volume is bigger like you described then you are adding thermal energy to the model (assuming there are no heat transfer/sources/sinks, so the temperature remains constant in the sphere). If you are just moving the sphere not enlarging it, even if it’s a very big displacement, then there should be no problem. I hope that helps. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月31日 GMT-4 13:50
Hi Nagi,

I have a similar problem. I am trying to model a bolt on a surface. Basically, I want to simulate the compression of the bolt on the surface, and also see the stress on the bolt if I pull the material under the surface. It's like pulling an RJ45 cable, after it has been crimped. How do I make Comsol deform the surface due to the force on the bolt?
Hi Nagi, I have a similar problem. I am trying to model a bolt on a surface. Basically, I want to simulate the compression of the bolt on the surface, and also see the stress on the bolt if I pull the material under the surface. It's like pulling an RJ45 cable, after it has been crimped. How do I make Comsol deform the surface due to the force on the bolt?

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月31日 GMT-4 18:54
Hi Abdulrasaq,

That problem is quite different from the one discussed above! For that you certainly need the Structural Mechanics physics. COMSOL has some good features for bolt loading. Check out the Prestressed Bolts in a Tube Connection (Application 185) in the COMSOL Application Gallery for a good example.

Regards,
Nagi
Hi Abdulrasaq, That problem is quite different from the one discussed above! For that you certainly need the Structural Mechanics physics. COMSOL has some good features for bolt loading. Check out the Prestressed Bolts in a Tube Connection (Application 185) in the COMSOL Application Gallery for a good example. Regards, Nagi

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年8月31日 GMT-4 21:13
Dear Nagi,

Thanks for pointing that out. However, the prestressed bolt setting in Comsol only works if the bolt is selected from the parts library. My geometry contains a bolt already, how should I put a stress on it because the prestress option isn't working.

Kind regards,

Abdulrasaq
Dear Nagi, Thanks for pointing that out. However, the prestressed bolt setting in Comsol only works if the bolt is selected from the parts library. My geometry contains a bolt already, how should I put a stress on it because the prestress option isn't working. Kind regards, Abdulrasaq

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 2016年9月1日 GMT-4 02:46
Dear Nagi,

thanks again. I have already modelled it in 2D and it works really fine. I wasn't sure about Deformed Geometry as well, but if you do not chnage the initial volume trough deformation, then you get a reasonable solution. Thanks for sharing your knowledge with us!

Wish you all the best :).

Best regards
Antoni
Dear Nagi, thanks again. I have already modelled it in 2D and it works really fine. I wasn't sure about Deformed Geometry as well, but if you do not chnage the initial volume trough deformation, then you get a reasonable solution. Thanks for sharing your knowledge with us! Wish you all the best :). Best regards Antoni

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.