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.

Displacement limit

Please login with a confirmed email address before reporting spam

Hi,
I'm a student in electronic engineering, and now i need to simulate an elastic material.
I'm interest at structural mechanic module with 2D axial symmetry. I need to understand the stress and strain of an ideal hookian material of unknow E (Young's modulus).
The model is a simply fill disc of Radius=0.25m and thick=0.01m with E=6kPa, nu=0.3 and rho=900 kg/m3. I set the boundary setting with Rz constrain = 0 for the lower edge and, Rr constrain = 0 for the axial symmetry edge, but I've got some problem with the upper boundary. The upper boundary is under a load of -p Pa with p parameter of parametric simulation from 0 to 100000 with step of 1000. The problem is the displacement. When the load p is too big the upper boundary don't stop run on the bottom layer but go down.
My question is: Can I limit the displacement of top boundary as the disk thick?
I try to set a displacement dependent constrain of top layer in Z-direction, but it doesn't work.

Thanks

7 Replies Last Post 11.02.2011, 10:49 GMT-5
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 11.02.2011, 01:32 GMT-5
Hi

fisrt to be sure: are you in 2D-axi mode (using a true 2D-axisymmetric analysis with a rectangular section, or are you in 3D with a axi-symmetric cylinder as domain ?

because in 2D-axi you are restraining your model such that the full stiffness tensor has specific symmetric behaviour. THis is no problem per say just to better understand.

Because if you use E, nu, rho you are also defining a specific symmetry of your stiffness tensor.

Anyhow what strikes me are your values:
the compressive stiffness of a cylindrical slab (fixed on a lower section is

k[N/m] = E[Pa] * Area[m^2] / Length[m] or about 120 kN/m,

loading that with some 1E5[Pa] or about 1[Atm] should give almost a 1m displacement/compression, which is not really compatible with the 1 [cm] thickness, or have I missed something ? These are not "small" displacements

Then if you are fixing the rim of your cylinder and its a membrane, you should get out something like the shape of a well filled balloon.

Anyhow for such large displacement you must use more advanced material properties as the linear theory

Could it be that E= 6 [GPa] and not 6 [kPa] ? that would change a lot, and still be more compatible with something like a plastic material ?

--
Good luck
Ivar
Hi fisrt to be sure: are you in 2D-axi mode (using a true 2D-axisymmetric analysis with a rectangular section, or are you in 3D with a axi-symmetric cylinder as domain ? because in 2D-axi you are restraining your model such that the full stiffness tensor has specific symmetric behaviour. THis is no problem per say just to better understand. Because if you use E, nu, rho you are also defining a specific symmetry of your stiffness tensor. Anyhow what strikes me are your values: the compressive stiffness of a cylindrical slab (fixed on a lower section is k[N/m] = E[Pa] * Area[m^2] / Length[m] or about 120 kN/m, loading that with some 1E5[Pa] or about 1[Atm] should give almost a 1m displacement/compression, which is not really compatible with the 1 [cm] thickness, or have I missed something ? These are not "small" displacements Then if you are fixing the rim of your cylinder and its a membrane, you should get out something like the shape of a well filled balloon. Anyhow for such large displacement you must use more advanced material properties as the linear theory Could it be that E= 6 [GPa] and not 6 [kPa] ? that would change a lot, and still be more compatible with something like a plastic material ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2011, 05:58 GMT-5
Hi Ivar,
thank you for very fast response.
I'm in 2D-axi mode.
I know that the Young's Modulus is very small, but the material is a polyurethane and it's very soft.
In the Analysis I expect that when the pressure is too high the compression stop, but it isn't.
I know that for this E the load on material must be low, in particular under the 8[kPa] ,but in real application the load can reach values of more than 8 [kPa]. In the Realty if the pressure grown over a limit the compression show a saturation and the displacement of upper boundary stop to about 90% of the thickness of 1 [cm]. I'd like this behavior.
I'd like see the compression of material until it become a thin material of 10% of initial 1[cm] thick. Than the displacement must stop his growth.
Can't I limit the displacement?
There is the possibility of constrain the upper boundary to stay over a particular Z-coordinate, for example 0.001 [m]?
Hi Ivar, thank you for very fast response. I'm in 2D-axi mode. I know that the Young's Modulus is very small, but the material is a polyurethane and it's very soft. In the Analysis I expect that when the pressure is too high the compression stop, but it isn't. I know that for this E the load on material must be low, in particular under the 8[kPa] ,but in real application the load can reach values of more than 8 [kPa]. In the Realty if the pressure grown over a limit the compression show a saturation and the displacement of upper boundary stop to about 90% of the thickness of 1 [cm]. I'd like this behavior. I'd like see the compression of material until it become a thin material of 10% of initial 1[cm] thick. Than the displacement must stop his growth. Can't I limit the displacement? There is the possibility of constrain the upper boundary to stay over a particular Z-coordinate, for example 0.001 [m]?

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 11.02.2011, 07:10 GMT-5
Hi

well that means that you have a non-linear material, and you need to add that to the material physics/parameters. It's probybly more a hyperelastic material than truely a "linear" one. So check with the other material definition possibilities and enter rather the correct physics for your material, than searching for "tricks" or workarounds.

For me that seems to make more sens

--
Good luck
Ivar
Hi well that means that you have a non-linear material, and you need to add that to the material physics/parameters. It's probybly more a hyperelastic material than truely a "linear" one. So check with the other material definition possibilities and enter rather the correct physics for your material, than searching for "tricks" or workarounds. For me that seems to make more sens -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2011, 09:32 GMT-5
Hi,
I don't interested about the non linearity of material, I'd like make analysis for linear material, but I'd like take a constrain on the maximum displacement for upper boundary. Isn't it possible?
Thank You
Hi, I don't interested about the non linearity of material, I'd like make analysis for linear material, but I'd like take a constrain on the maximum displacement for upper boundary. Isn't it possible? Thank You

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 11.02.2011, 09:43 GMT-5
Hi

you can adapt at will so long you respect physics that is add in some equations that governs correctly your displacemen. If you apply with a large force on a soft material it will compress, but linear theory is based on "small displacements up to a few permille or a percent, you are at 10'000%, the hypithesis of "small displacement is not really being respected.

You can say you add asmall metallic spacerthat touches (i.e. contact problem) or you add a non-linear spring that sets the system in equilibrium, before you have sweezed tzhe shim to "0" thickness. But you must respect the physics nd the underlaying hypothesis, or have I missed something in you explanations ?

--
Good luck
Ivar
Hi you can adapt at will so long you respect physics that is add in some equations that governs correctly your displacemen. If you apply with a large force on a soft material it will compress, but linear theory is based on "small displacements up to a few permille or a percent, you are at 10'000%, the hypithesis of "small displacement is not really being respected. You can say you add asmall metallic spacerthat touches (i.e. contact problem) or you add a non-linear spring that sets the system in equilibrium, before you have sweezed tzhe shim to "0" thickness. But you must respect the physics nd the underlaying hypothesis, or have I missed something in you explanations ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.02.2011, 09:58 GMT-5
Great, these can be the solutions I search. In particular I've tried to put a small metallic spacer under the elastic material, but with the same error: the elastic material penetrates in metallic material.
I don't try the idea of spring, but I'm not a mechanics so I search the easiest.
I think we are on the right way, you understand my problem.
Thank You
Great, these can be the solutions I search. In particular I've tried to put a small metallic spacer under the elastic material, but with the same error: the elastic material penetrates in metallic material. I don't try the idea of spring, but I'm not a mechanics so I search the easiest. I think we are on the right way, you understand my problem. Thank You

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 11.02.2011, 10:49 GMT-5
Hi

just be aware that contact problems are also rather tricky to get solved, and are based on non linear springs

--
Good luck
Ivar
Hi just be aware that contact problems are also rather tricky to get solved, and are based on non linear springs -- Good luck Ivar

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.