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.
Modelling a simple heat exchanger
Posted Nov 24, 2010, 2:55 a.m. EST 6 Replies
Please login with a confirmed email address before reporting spam
Hi,
I am trying to model a simple heat exchanger with a hot fluid flowing in the inner pipe and cooling water flows in a ticker pipe around the inner pipe. I have attached a drawing from Inventor.
I successfully applied all boundary conditions and i can solve the model. The problem, however, is that the solution is highly dependent on the mesh. I use a quadratic extruded mesh in the inner pipe and tetrahedral elements in the rest of my model. Having an element size of "finer" i get a temperature drop in the hot fluid from ~600-300K; if i refine the mesh to "very fine" the temperature drop changes to 600-450K and finally "extremely fine" mesh results in a drop of 600-599,5K.
I have tried the following so far:
Study 1 -> Solver Configurations -> Solver 1 -> Stationary Solver and set "relative tolerance" to 1e-6. This had the same effect as refining the mesh, i.e. a lower temperature drop.
I wanted to increase the order of the shape function polynomial, but i have been unable to find the setting to do so.
Now i have no more ideas, so i turn to you hoping that someone is able to help me.
I will attach a model with the coarsest mesh as it doesn't take up so much space as the other models.
Thank you for reading and i hope someone is able to help me
Kind regards,
Anders J M
I am trying to model a simple heat exchanger with a hot fluid flowing in the inner pipe and cooling water flows in a ticker pipe around the inner pipe. I have attached a drawing from Inventor.
I successfully applied all boundary conditions and i can solve the model. The problem, however, is that the solution is highly dependent on the mesh. I use a quadratic extruded mesh in the inner pipe and tetrahedral elements in the rest of my model. Having an element size of "finer" i get a temperature drop in the hot fluid from ~600-300K; if i refine the mesh to "very fine" the temperature drop changes to 600-450K and finally "extremely fine" mesh results in a drop of 600-599,5K.
I have tried the following so far:
Study 1 -> Solver Configurations -> Solver 1 -> Stationary Solver and set "relative tolerance" to 1e-6. This had the same effect as refining the mesh, i.e. a lower temperature drop.
I wanted to increase the order of the shape function polynomial, but i have been unable to find the setting to do so.
Now i have no more ideas, so i turn to you hoping that someone is able to help me.
I will attach a model with the coarsest mesh as it doesn't take up so much space as the other models.
Thank you for reading and i hope someone is able to help me
Kind regards,
Anders J M
6 Replies Last Post Nov 25, 2010, 2:06 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
hello,
I have been looking quickly at your model and I dont have definitive comment to offer, not being an expert in heat transfer but i here are a few comment
- you use water as fluid with temperature u p to 600K in your calculation. The water model is good up to 553.75K in Comsol. Not sure what it does about that.[ and we have steam here.. not sure that the calculation goes smoothly]
-I am surprised that you model the inlet and outlet in 3d but not the outside wall..
any reason for that..? look either something that could be simplified or need to be completed... it will remove a few meshes from your simulation.
not sure if any of that will resolve your issues... but these are the stuff that came to my mind while going through your model..
jf
I have been looking quickly at your model and I dont have definitive comment to offer, not being an expert in heat transfer but i here are a few comment
- you use water as fluid with temperature u p to 600K in your calculation. The water model is good up to 553.75K in Comsol. Not sure what it does about that.[ and we have steam here.. not sure that the calculation goes smoothly]
-I am surprised that you model the inlet and outlet in 3d but not the outside wall..
any reason for that..? look either something that could be simplified or need to be completed... it will remove a few meshes from your simulation.
not sure if any of that will resolve your issues... but these are the stuff that came to my mind while going through your model..
jf
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello and thanks for your reply.
I should probably be more specific here - the fluid is in a subcritical state and the pressure is 260[bar], so it is still a liquid at this temperature. I've had some problems in inputting this correctly to the model - i tried setting the liquid to 260[bar] and initial value for the liquid also to 260[bar]. However when plotting the pressure along the axis of the pipe it is around 1[Pa] so obviously i'm not doing this correctly.
Regarding the temperature of 600K contra 554.75K, i seem to recall that comsol will just extrapolate a value at 600K or just use the values from 554.75K. I will lower the temperature just to check.
I'm not sure i understand what you mean here. Do you mean the outside wall of the outer pipe?
If so, my reasoning was that i'm assuming the heat loss through this wall is negligible so i just modelled it as a thermally insulated wall.
hello,
I have been looking quickly at your model and I dont have definitive comment to offer, not being an expert in heat transfer but i here are a few comment
- you use water as fluid with temperature u p to 600K in your calculation. The water model is good up to 553.75K in Comsol. Not sure what it does about that.[ and we have steam here.. not sure that the calculation goes smoothly]
I should probably be more specific here - the fluid is in a subcritical state and the pressure is 260[bar], so it is still a liquid at this temperature. I've had some problems in inputting this correctly to the model - i tried setting the liquid to 260[bar] and initial value for the liquid also to 260[bar]. However when plotting the pressure along the axis of the pipe it is around 1[Pa] so obviously i'm not doing this correctly.
Regarding the temperature of 600K contra 554.75K, i seem to recall that comsol will just extrapolate a value at 600K or just use the values from 554.75K. I will lower the temperature just to check.
-I am surprised that you model the inlet and outlet in 3d but not the outside wall..
any reason for that..? look either something that could be simplified or need to be completed... it will remove a few meshes from your simulation.
I'm not sure i understand what you mean here. Do you mean the outside wall of the outer pipe?
If so, my reasoning was that i'm assuming the heat loss through this wall is negligible so i just modelled it as a thermally insulated wall.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Regarding my second comment.. if you are Ok with BC for the outer wall then you dont need to model heat transfer through the external outlet inlet.
I remember [vaguely] than constitutive model/thermodynamic can have LARGE impact on CFD simulation ...I will suggest that you run a few case at more mundane temperature/pression [ say ambient and peak temperature below boiling temperature and check if you have the same numerical problems... If not it will point in the direction that the water model is not adequate.
if yes it will force you to look elsewhere. to resolve your issue
jf
I remember [vaguely] than constitutive model/thermodynamic can have LARGE impact on CFD simulation ...I will suggest that you run a few case at more mundane temperature/pression [ say ambient and peak temperature below boiling temperature and check if you have the same numerical problems... If not it will point in the direction that the water model is not adequate.
if yes it will force you to look elsewhere. to resolve your issue
jf
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Good idea, i will try using more reasonable temperatures.
I will report back with my findings
I will report back with my findings
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Using 99[degC] water and 12[degC] cooling water i am experiencing the same problems. Atleast i now know that it isnt the temperature range that is causing the problems-
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not sure it's linked, but you have for both fluids and main physics a "strain reference temperature" of T=0[K] but I'm not fully sure if its used in your case.
Then there are all the (in)coherences around the pressure, initial values and settings absolute / gauge values in the different nodes + outlet settings, I often get that wrong and must run a few check cases regularly to get it back into my mind the "right" COMSOL way. SO like that I cannot tell immediatly if its OK in your model
Then if I remember right, COMSOL doe not extrapolate material properties, but keep them constant at the min-max values, check the material graphs
Finally you have a set of warning signs on 3 material parameters that COMSOL is claiming that it's missing, such issues leads often to strange results, but can also pass undetected !
(I'm running 4.1.0.88 just now (still 2 days left ;)
--
Good luck
Ivar
I'm not sure it's linked, but you have for both fluids and main physics a "strain reference temperature" of T=0[K] but I'm not fully sure if its used in your case.
Then there are all the (in)coherences around the pressure, initial values and settings absolute / gauge values in the different nodes + outlet settings, I often get that wrong and must run a few check cases regularly to get it back into my mind the "right" COMSOL way. SO like that I cannot tell immediatly if its OK in your model
Then if I remember right, COMSOL doe not extrapolate material properties, but keep them constant at the min-max values, check the material graphs
Finally you have a set of warning signs on 3 material parameters that COMSOL is claiming that it's missing, such issues leads often to strange results, but can also pass undetected !
(I'm running 4.1.0.88 just now (still 2 days left ;)
--
Good luck
Ivar