Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
thermally resistive layer
Posted 7 mars 2013, 22:19 UTC−5 Version 4.2, Version 4.2a, Version 4.3, Version 4.3a 6 Replies
Please login with a confirmed email address before reporting spam
Hi everyone,
I am modeling a pipe with fluid circulating within by coupling heat transfer and fluid flow and the pipe diameter is defined in the geometry section.
i need to define the pipe wall thickness for my pipe as well which i do it with adding a thermally resistive layer in heat transfer physics and define the conductivity as well as the wall thickness.
my question is that considering the pipe wall thickness, is the pipe diameter defined in building the geometry inner diameter or the outer one?
Regards
ASAL
I am modeling a pipe with fluid circulating within by coupling heat transfer and fluid flow and the pipe diameter is defined in the geometry section.
i need to define the pipe wall thickness for my pipe as well which i do it with adding a thermally resistive layer in heat transfer physics and define the conductivity as well as the wall thickness.
my question is that considering the pipe wall thickness, is the pipe diameter defined in building the geometry inner diameter or the outer one?
Regards
ASAL
6 Replies Last Post 25 mars 2013, 02:25 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not sure I catch you fully, you can simulate a solid domain with a tube inside of given wall thickness (domain2) and therein a domain of fluid. Then you have 3 domains, but you can also simulate the tube by a thin layer between the fluid and the externa solid domain (2 domains and one "thin" boundary "domain".
But you cannot put a thin resisiive domain on an external boundary i.e. with convective cooling to the outside
--
Good luck
Ivar
I'm not sure I catch you fully, you can simulate a solid domain with a tube inside of given wall thickness (domain2) and therein a domain of fluid. Then you have 3 domains, but you can also simulate the tube by a thin layer between the fluid and the externa solid domain (2 domains and one "thin" boundary "domain".
But you cannot put a thin resisiive domain on an external boundary i.e. with convective cooling to the outside
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
actually I define a a U-pipe which is in contact with surrounding soil, and a thermally resistive layer between the soil and the pipe wall. i think this will be the second case you mentioned, the fluid, the soil and a resistive layer in between. attached is the geometry of my model and would appreciate if you have a look at that.
regards
ASAL
actually I define a a U-pipe which is in contact with surrounding soil, and a thermally resistive layer between the soil and the pipe wall. i think this will be the second case you mentioned, the fluid, the soil and a resistive layer in between. attached is the geometry of my model and would appreciate if you have a look at that.
regards
ASAL
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
actually I define a a U-pipe which is in contact with surrounding soil, and a thermally resistive layer between the soil and the pipe wall. i think this will be the second case you mentioned, the fluid, the soil and a resistive layer in between. attached is the geometry of my model and would appreciate if you have a look at that.
regards
ASAL
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
looking at your model, my impression is really that you are putting a large effort in modelling the details, but the main issues can be estimated well from a lumped approach. Your fluid flow is so thin compared to the length and the diameter that your are in a steady state flow condition with a given surface exchange that can be estimated as an integration of a more or less linear varying value.
For me 1) you should ideally use the "pipe flow" module , its really made up specifically for such kind of cases (long piping's embedded in large volumes compared to diameter and flow) But it's still feasible like you have set up your model, provided some good initial condition settings and time to wayt for the convergence
Have you tried to 1) only the piping to get the flow velocity (constant Temp to start with), then take a 1m long pipe and get the heat exchange coefficients and add those to the boundary of the fluid, and solve the rest in even 2D axi (I know its not 100 correct as you have a flow up and down, strictly breaking the symmetry.
--
Good luck
Ivar
--
Good luck
Ivar
looking at your model, my impression is really that you are putting a large effort in modelling the details, but the main issues can be estimated well from a lumped approach. Your fluid flow is so thin compared to the length and the diameter that your are in a steady state flow condition with a given surface exchange that can be estimated as an integration of a more or less linear varying value.
For me 1) you should ideally use the "pipe flow" module , its really made up specifically for such kind of cases (long piping's embedded in large volumes compared to diameter and flow) But it's still feasible like you have set up your model, provided some good initial condition settings and time to wayt for the convergence
Have you tried to 1) only the piping to get the flow velocity (constant Temp to start with), then take a 1m long pipe and get the heat exchange coefficients and add those to the boundary of the fluid, and solve the rest in even 2D axi (I know its not 100 correct as you have a flow up and down, strictly breaking the symmetry.
--
Good luck
Ivar
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
yes you are right but this is the way I have proceeded with comsol so far.
now I have another question. I am modelling very turbulent flow in a narrow pipe as well and use the normal k_epsilon model but the velocity profile does not look ok at the pipe wall. if I change it to low re number K-Epsilon model results are significantly different with better velocity profile and the solving time is about 10 times more and sometime it takes for ever and does not converge finally. I am not sure how reliable this low Re number model is. I was wondering if there is any adjustments to do to make the model converge with low Re number k-e models.
Regards
ASAL
yes you are right but this is the way I have proceeded with comsol so far.
now I have another question. I am modelling very turbulent flow in a narrow pipe as well and use the normal k_epsilon model but the velocity profile does not look ok at the pipe wall. if I change it to low re number K-Epsilon model results are significantly different with better velocity profile and the solving time is about 10 times more and sometime it takes for ever and does not converge finally. I am not sure how reliable this low Re number model is. I was wondering if there is any adjustments to do to make the model converge with low Re number k-e models.
Regards
ASAL
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm no CFD specialist, so I have no trained "physics feeling" for how a CFD code, in turbulent should look at along the boundaries. But as FEM uses often phenomenological approaches to simulate the surface layers, these might be correct for the solving but still look odd for us.
But there are a few CFD specialist on the Forum, they should be in a better position to tell, and then I can learn something new too ;)
--
Good luck
Ivar
I'm no CFD specialist, so I have no trained "physics feeling" for how a CFD code, in turbulent should look at along the boundaries. But as FEM uses often phenomenological approaches to simulate the surface layers, these might be correct for the solving but still look odd for us.
But there are a few CFD specialist on the Forum, they should be in a better position to tell, and then I can learn something new too ;)
--
Good luck
Ivar