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 The information provided may be out of date.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.

undesired rolling during a symmetric deformation

Please login with a confirmed email address before reporting spam

Hi,

I am simulating the effect of pressure change inside a flexible tube (cylinder) that causes it to radially stretch. This is not an axisymmetric problem, but is a 2D problem (look at it from the circular section of that and parallel to the axis of the tube). Due to symmetry, I only have simulated the upper half of the cylinder. On the inner boundary I have a pressure load and the two walls cut sections are considered symmetry. Although the deformation seems to be symmetric, the deformed geometry translates (rolls) to one side; for example, if the center of the cylinder before pressurizing it is at (x,y)=(0,0) after the symmetric pressure load it moves to (x,y)=(0.5,0) rather than staying at (0,0). In addition, when I repeated the solution, I noticed that the amount of rolling differs from the previous run (x,y)=(0.25,0). This means that there is not a unique solution for the problem or probably the problem is not very well-defined. I do not think that if I wanted to simulate a full cylinder I would get this translation, but there is a justified reason that I opted for a half cylinder.

Also, the mesh is pretty fine, although it is not symmetric because it is of an unstructured type. So I do not think them mesh has any effect on it.

I wonder why there is no unique solution, and how can I restrict the undesired translation without fixing any point on the domain.

Your input is appreciated.

Thanks,
Jessica

5 Replies Last Post 15 mars 2012, 16:53 UTC−4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 mars 2012, 04:30 UTC−5
Hi

ifyou do nt constrain enough DoF your final object might move around, often one can get rid of such effects by adding a symmetry boundary, or splitting the object along the symmetry bounary, mesh one half and mirror the mesh onto the other symmetric part (at least this should reduce the motion).

But in real life if you blow up a baloon, it will move left and right as you bow it up, due to small inhomogeniousities in the air flow, turbulence, material non homogeinities, so in some sens its physical, but can be unwanted for the modelling, I agree

--
Good luck
Ivar
Hi ifyou do nt constrain enough DoF your final object might move around, often one can get rid of such effects by adding a symmetry boundary, or splitting the object along the symmetry bounary, mesh one half and mirror the mesh onto the other symmetric part (at least this should reduce the motion). But in real life if you blow up a baloon, it will move left and right as you bow it up, due to small inhomogeniousities in the air flow, turbulence, material non homogeinities, so in some sens its physical, but can be unwanted for the modelling, I agree -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 mars 2012, 09:47 UTC−5
I agree, for this problem I would use a 2D asymmetric simulation with the centre of the tube as the edge of symmetry.
Simulation wise the centre would be impossible to move.
I agree, for this problem I would use a 2D asymmetric simulation with the centre of the tube as the edge of symmetry. Simulation wise the centre would be impossible to move.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 mars 2012, 15:12 UTC−5

I agree, for this problem I would use a 2D asymmetric simulation with the centre of the tube as the edge of symmetry.
Simulation wise the centre would be impossible to move.


Hi Jaap,

Thanks for the your comments. In fact, my problem is not axisymmetric but it is a 2D problem. Imagine you are looking at the circular section of the tube.

Jessica.
[QUOTE] I agree, for this problem I would use a 2D asymmetric simulation with the centre of the tube as the edge of symmetry. Simulation wise the centre would be impossible to move. [/QUOTE] Hi Jaap, Thanks for the your comments. In fact, my problem is not axisymmetric but it is a 2D problem. Imagine you are looking at the circular section of the tube. Jessica.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 mars 2012, 15:39 UTC−5

Hi

ifyou do nt constrain enough DoF your final object might move around, often one can get rid of such effects by adding a symmetry boundary, or splitting the object along the symmetry bounary, mesh one half and mirror the mesh onto the other symmetric part (at least this should reduce the motion).

But in real life if you blow up a baloon, it will move left and right as you bow it up, due to small inhomogeniousities in the air flow, turbulence, material non homogeinities, so in some sens its physical, but can be unwanted for the modelling, I agree

--
Good luck
Ivar


Hi Ivar,

Thanks for your comments.
I completely agree with you about the actual physical conditions that may result in non-symmetric deformation and displacement. I believe that I have sufficiently constrained my model by assigning symmetry condition. I also tried a structured (mapped) mesh to eliminate any chance of asymetric results due to asymetric unstructured mesh. This method didn't work either. I am basically simulating an axially cut blood vessel which has been already opened from 360 degrees to 140 degrees. Now I want to close it again and see how much stress will be developed when I close it. For reducing the simulation time and simplifying formulation I am considering just half of the cut vessel (from 70 degrees to 140 degrees). Then I want to pressurize it and see what happens.
So I may want to try only to simulate one quarter of the tube and see what happens.

Thanks again.
Jessica
[QUOTE] Hi ifyou do nt constrain enough DoF your final object might move around, often one can get rid of such effects by adding a symmetry boundary, or splitting the object along the symmetry bounary, mesh one half and mirror the mesh onto the other symmetric part (at least this should reduce the motion). But in real life if you blow up a baloon, it will move left and right as you bow it up, due to small inhomogeniousities in the air flow, turbulence, material non homogeinities, so in some sens its physical, but can be unwanted for the modelling, I agree -- Good luck Ivar [/QUOTE] Hi Ivar, Thanks for your comments. I completely agree with you about the actual physical conditions that may result in non-symmetric deformation and displacement. I believe that I have sufficiently constrained my model by assigning symmetry condition. I also tried a structured (mapped) mesh to eliminate any chance of asymetric results due to asymetric unstructured mesh. This method didn't work either. I am basically simulating an axially cut blood vessel which has been already opened from 360 degrees to 140 degrees. Now I want to close it again and see how much stress will be developed when I close it. For reducing the simulation time and simplifying formulation I am considering just half of the cut vessel (from 70 degrees to 140 degrees). Then I want to pressurize it and see what happens. So I may want to try only to simulate one quarter of the tube and see what happens. Thanks again. Jessica

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 mars 2012, 16:53 UTC−4
I have an update on this:

In my 2D half cylinder, I added a line from the inner to the outer boundary, exactly half way from the two ends. COMSOL automatically considers it an interior boundary, where the continuity is also met, and I did not need to add any boundary condition for it. Then I constrained that interior boundary by setting the Prescribed Displacement in x direction equal to zero, and it worked very well as the attached picture shows.

Jessica
I have an update on this: In my 2D half cylinder, I added a line from the inner to the outer boundary, exactly half way from the two ends. COMSOL automatically considers it an interior boundary, where the continuity is also met, and I did not need to add any boundary condition for it. Then I constrained that interior boundary by setting the Prescribed Displacement in x direction equal to zero, and it worked very well as the attached picture shows. Jessica

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.