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 The information provided may be out of date.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.
Laminar flow combined with plane strain and ALE, need help please!
Posted 12 apr 2011, 12:14 GMT-4 Version 3.5a, Version 4.1 4 Replies
Please login with a confirmed email address before reporting spam
Hi Mr ivar,
According to your suggestions, I try to fix it and I still have something really confused!
This is the precious suggestions that you give to me!
there are many things here that I do not really catch, but I believe also it's because I do not fully understand your model, a few items I can suggest to check:
1) you could probably use the FSI module, as there all three physics are already combined
2) the inlet 1 you set up the velocity with an "s" but you have 3 segments, and s is redefined on each going from 0, to 1 along the edge, have you checked that the initial value velocity is already more or less as expected ?
3) I do not understand "inlet 2" this is a wall not an inlet ? I'm missing something ;)
4) boundary loads 1 and 3 of "0" do not bring anything
5) in the solid Mechanics do not leave an entry empty, add a "0" to the xref (even if you do not use it)
6) in the moving mesh, by default everything is fixed, I believe you should set a free deformation to the fluid part
7) your prescribed mesh displacement is not set you need to define something also for the boundaries
8) remove the not used solver cases and keep only one
9) try to solve only one physics at the time, to check that they are ok, you can also just get the initial conditions loaded to test those also
Here is my confusions:
About 2) the three segments in inlet 1 ,3,5 has the same conditions. Does I make some mistakes here?
3) inlet 2 , actually it is the inner boundary of the fluids, it has the same velocity as the inner boundary of the membrane.
4)boundary loads 1 and 3 , I want to do is they are like fixed, should I use the prescribed displacement? I check in the old version 3.5, it should use “standard notation" , Rx 0m Ry 0 m
7) prescribed mesh displacement is the membrane's mesh displacement, so I think it should be u2 and v2, is it right?
My model is about the single valve, the domain 4 8 9 13 15 is the solid membrane, when the flow enters from left to right, I need to find out how the membrane acts and how the fluids act.
Thank you so much!
According to your suggestions, I try to fix it and I still have something really confused!
This is the precious suggestions that you give to me!
there are many things here that I do not really catch, but I believe also it's because I do not fully understand your model, a few items I can suggest to check:
1) you could probably use the FSI module, as there all three physics are already combined
2) the inlet 1 you set up the velocity with an "s" but you have 3 segments, and s is redefined on each going from 0, to 1 along the edge, have you checked that the initial value velocity is already more or less as expected ?
3) I do not understand "inlet 2" this is a wall not an inlet ? I'm missing something ;)
4) boundary loads 1 and 3 of "0" do not bring anything
5) in the solid Mechanics do not leave an entry empty, add a "0" to the xref (even if you do not use it)
6) in the moving mesh, by default everything is fixed, I believe you should set a free deformation to the fluid part
7) your prescribed mesh displacement is not set you need to define something also for the boundaries
8) remove the not used solver cases and keep only one
9) try to solve only one physics at the time, to check that they are ok, you can also just get the initial conditions loaded to test those also
Here is my confusions:
About 2) the three segments in inlet 1 ,3,5 has the same conditions. Does I make some mistakes here?
3) inlet 2 , actually it is the inner boundary of the fluids, it has the same velocity as the inner boundary of the membrane.
4)boundary loads 1 and 3 , I want to do is they are like fixed, should I use the prescribed displacement? I check in the old version 3.5, it should use “standard notation" , Rx 0m Ry 0 m
7) prescribed mesh displacement is the membrane's mesh displacement, so I think it should be u2 and v2, is it right?
My model is about the single valve, the domain 4 8 9 13 15 is the solid membrane, when the flow enters from left to right, I need to find out how the membrane acts and how the fluids act.
Thank you so much!
Attachments:
4 Replies Last Post 13 apr 2011, 04:06 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I attached some of my model pictures and the related behaviors of the fluids which is done in 3.5a
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I added the fixed points in my geometry, these points are fixed, like fix the membrane to the tube. But still, error occurs!
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I search the discuss forum and find that my interior boundaries like 3, 7, 11 .....need to be the continuity, and I do the ' form assembly' in the geometry node, so I should create 3, 7, 11, 12, 8 and all these interior boundaries as "identity pairs" and set them to be continuity, because other interior boundaries like 23, 38...these are the membrane interface holding u2, v2 boundaries!
but I don't know how to do it? When I try to use the pair----flow continuity----I can choose any boundary? Can you give me some suggestions?
About the inlet 2, I search the user's guide, I think the laminar flow physics is not active in the solid domain, so the interior boundary 11,20,23,27...serves as the exterior boundary of the fluid domain, so I can set it as inlet, did my understanding right?
but I don't know how to do it? When I try to use the pair----flow continuity----I can choose any boundary? Can you give me some suggestions?
About the inlet 2, I search the user's guide, I think the laminar flow physics is not active in the solid domain, so the interior boundary 11,20,23,27...serves as the exterior boundary of the fluid domain, so I can set it as inlet, did my understanding right?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I have several problems with your model, it might also be partly due to my newer release/patch.
I gave you one wrong hint, use rather for your velocity profile AND for the initial values velocity "x" component something like
u_max*(0.004-y[1/m])^2
Because of the symmetry boundary, you want maximum velocity there
By the way you are not respecting the units so this is also making life more difficult to check if it's coherent, it's worth to correct that
To get your example to run I need to :
- disable all the "weak constraints", disable your 4 "weak constraints", also the "inlet 2" and the symmetry nodes.
- Add a wall node with the same selection as the "symmetry", and select "SLIP"
then ONLY I can solve the SPF (alone) in a reasonable way
If I then add the structural and there:
- disable Prescribed displacement 2,3, as well as boundary loads 3, 5 and Point Load 1. And then replace the Boundary Load 4 values by the pressure "p" (from spf) and turn off the weak constraints in the prescribed displacements
- Normally I would replace your Prescribed Displacement 3 by "Fixed Boundary" but it's the same for COMSOL
The in ALE I'll leave as is
Then I can run and it gives some results for a stationary case (by adding a Parameter t = 0.1[sec] to force some flow, remove it when using the time stepping, even if COMSOL will automatically overwrite it when you add a time solver
If I replace the stationary solver by a transient one, then do not forget to change the "Time Dependent Solver 1 node "Time Stepping" from "Free" to "Intermediate" if you want to see your repeating pulses as otherwise it will jump over some and have convergence issues. And define enough steps to resolve the pulses (not too many 3-6 per pulse should do)
But I cannot get your model to run in transient yet.
I believe its easier to reconstruct it, by exporting the geometry and reimporting it into a fresh v4 model, I suspect translation issues, or just that the model has been modified too many times ;)
I agree my comments are somewhat messy. It would be easier to restart and add one physics at the time, solve for it and check that its running, using SPF, ALE and then Solid.
--
Good luck
Ivar
I have several problems with your model, it might also be partly due to my newer release/patch.
I gave you one wrong hint, use rather for your velocity profile AND for the initial values velocity "x" component something like
u_max*(0.004-y[1/m])^2
Because of the symmetry boundary, you want maximum velocity there
By the way you are not respecting the units so this is also making life more difficult to check if it's coherent, it's worth to correct that
To get your example to run I need to :
- disable all the "weak constraints", disable your 4 "weak constraints", also the "inlet 2" and the symmetry nodes.
- Add a wall node with the same selection as the "symmetry", and select "SLIP"
then ONLY I can solve the SPF (alone) in a reasonable way
If I then add the structural and there:
- disable Prescribed displacement 2,3, as well as boundary loads 3, 5 and Point Load 1. And then replace the Boundary Load 4 values by the pressure "p" (from spf) and turn off the weak constraints in the prescribed displacements
- Normally I would replace your Prescribed Displacement 3 by "Fixed Boundary" but it's the same for COMSOL
The in ALE I'll leave as is
Then I can run and it gives some results for a stationary case (by adding a Parameter t = 0.1[sec] to force some flow, remove it when using the time stepping, even if COMSOL will automatically overwrite it when you add a time solver
If I replace the stationary solver by a transient one, then do not forget to change the "Time Dependent Solver 1 node "Time Stepping" from "Free" to "Intermediate" if you want to see your repeating pulses as otherwise it will jump over some and have convergence issues. And define enough steps to resolve the pulses (not too many 3-6 per pulse should do)
But I cannot get your model to run in transient yet.
I believe its easier to reconstruct it, by exporting the geometry and reimporting it into a fresh v4 model, I suspect translation issues, or just that the model has been modified too many times ;)
I agree my comments are somewhat messy. It would be easier to restart and add one physics at the time, solve for it and check that its running, using SPF, ALE and then Solid.
--
Good luck
Ivar