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 The information provided may be out of date.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.
Does COMSOL treat the nodes at the boundary also as unknowns?
Posted May 20, 2012, 3:53 a.m. EDT Results & Visualization Version 4.2a 3 Replies
Please login with a confirmed email address before reporting spam
Hi,
Does COMSOL treat the nodes at the boundary also as unknowns?
I am trying to solve colloid transport inside a pipe with constant concentration boundary condition at the inlet. The unknown is concentration. There is sharp gradient in the concentration between the boundary node and the node next to that in the flow direction. I use very fine mesh.
If I plot the concentration at the inlet boundary using domain plot parameters, it is showing whatever the boundary condition I have given (Dirichlet) . But when I use cross-section plot parameters dialogue box to specify the line plot and plot the concentration at the inlet boundary, it is showing wrong values of boundary condition. Also the plotted values at the boundary changes with time where as the original B.C. is constant. In the data exported also the values at the points in the inlet boundary are different from my boundary condition.
So does comsol interpolate from the results to find the value of variable at the boundary if we give the co-ordinates?
Does anybody know why is this problem happening?
Thanks in advance.
Regards,
Seetha
Does COMSOL treat the nodes at the boundary also as unknowns?
I am trying to solve colloid transport inside a pipe with constant concentration boundary condition at the inlet. The unknown is concentration. There is sharp gradient in the concentration between the boundary node and the node next to that in the flow direction. I use very fine mesh.
If I plot the concentration at the inlet boundary using domain plot parameters, it is showing whatever the boundary condition I have given (Dirichlet) . But when I use cross-section plot parameters dialogue box to specify the line plot and plot the concentration at the inlet boundary, it is showing wrong values of boundary condition. Also the plotted values at the boundary changes with time where as the original B.C. is constant. In the data exported also the values at the points in the inlet boundary are different from my boundary condition.
So does comsol interpolate from the results to find the value of variable at the boundary if we give the co-ordinates?
Does anybody know why is this problem happening?
Thanks in advance.
Regards,
Seetha
3 Replies Last Post May 21, 2012, 5:06 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not sure I catch fully your question, normally the BC values you define remains and do not change, just as boundary values for dependent variables on boundaries that are free to change uses the solution.
If you plot along a selected boundary you should get the values you defined on that boundary, or the one COMSOl has calculated if not constraints are defined
It's only when you define Data Set cut lines / planes etc that COMSOL interpolate and you might then get some edge effects
It's true if you define a fixed i.e. Temperature T on a domain, then the boundaries inherite this value, except if any other domain common to these boundaries have different values, then the boundary gets an average T value (valid for any dependent variable. You can check this by solving for your initial conditions and look how COMSOL prepares the model for the solving (sharp edges gets smoothed).
When you have steep gradeints you should, as you say you do, use a fine mesh in this region, I often use boundary meshes when I know the gradient is steep perpendicular to the boundary, for most types of diffusion models
--
Good luck
Ivar
I'm not sure I catch fully your question, normally the BC values you define remains and do not change, just as boundary values for dependent variables on boundaries that are free to change uses the solution.
If you plot along a selected boundary you should get the values you defined on that boundary, or the one COMSOl has calculated if not constraints are defined
It's only when you define Data Set cut lines / planes etc that COMSOL interpolate and you might then get some edge effects
It's true if you define a fixed i.e. Temperature T on a domain, then the boundaries inherite this value, except if any other domain common to these boundaries have different values, then the boundary gets an average T value (valid for any dependent variable. You can check this by solving for your initial conditions and look how COMSOL prepares the model for the solving (sharp edges gets smoothed).
When you have steep gradeints you should, as you say you do, use a fine mesh in this region, I often use boundary meshes when I know the gradient is steep perpendicular to the boundary, for most types of diffusion models
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks Ivar for your reply.
But while exporting the post-processing data also (format- nodes, elements data), it is showing different values at the boundary (actual B.C. is dirichlet). Instead of the B.C. which is equal to one , it is showing a value of 62 at the boundary. The value of the variable inside the domain is 180. So is it because of diffusion or interpolation?
If it is an error only because of interpolation from inside the domain to boundary, can I assume that COMSOL is taking the B.C. which I had specified?
I tried using artificial diffusion, in that case the concentration inside the domain also gets diffused so much.
Also the boundary is the inlet boundary and it is not a common boundary between two domains.
Using boundary layer mesh also gives the same results.
What can I do now?
Can anybody please help in this regard, as I am struggling a lot with this?
Regards,
Seetha
But while exporting the post-processing data also (format- nodes, elements data), it is showing different values at the boundary (actual B.C. is dirichlet). Instead of the B.C. which is equal to one , it is showing a value of 62 at the boundary. The value of the variable inside the domain is 180. So is it because of diffusion or interpolation?
If it is an error only because of interpolation from inside the domain to boundary, can I assume that COMSOL is taking the B.C. which I had specified?
I tried using artificial diffusion, in that case the concentration inside the domain also gets diffused so much.
Also the boundary is the inlet boundary and it is not a common boundary between two domains.
Using boundary layer mesh also gives the same results.
What can I do now?
Can anybody please help in this regard, as I am struggling a lot with this?
Regards,
Seetha
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
Can you upload a simple example, as its difficult to fully be sure we have understood you correctly.
On the other side, these are the types of issues I typicaly send to "support" as its often linked to COMSOL internal behaviour
--
Good luck
Ivar
Can you upload a simple example, as its difficult to fully be sure we have understood you correctly.
On the other side, these are the types of issues I typicaly send to "support" as its often linked to COMSOL internal behaviour
--
Good luck
Ivar