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 The information provided may be out of date.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.

Cylindrical anisotropic stiffness

Please login with a confirmed email address before reporting spam

Hi,

Problem:

Imagine a cylindrical shell, with a certain wall thickness. The material properties of the shell have an isotropic density (rho) and an anisotropic stiffness described by a constant radial component (E_r) and a constant tangential component (E_perp). How to define that using the anisotropic solid model for a linear elastic material?



Thanks !
C

7 Replies Last Post 07.04.2013, 21:10 GMT-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 01.04.2013, 14:43 GMT-4
Hi

you can define the coordinate system to be used when setting up the material property call, and you can use Definitions Coordinate systems to add your own, or use the "boundary coordinate system COMSOL adds up by default (not theat the tangeant directions might be tricky to identify due to boundary direction vectors

--
Good luck
Ivar
Hi you can define the coordinate system to be used when setting up the material property call, and you can use Definitions Coordinate systems to add your own, or use the "boundary coordinate system COMSOL adds up by default (not theat the tangeant directions might be tricky to identify due to boundary direction vectors -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 01.04.2013, 15:15 GMT-4

Hi

you can define the coordinate system to be used when setting up the material property call, and you can use Definitions Coordinate systems to add your own, or use the "boundary coordinate system COMSOL adds up by default (not theat the tangeant directions might be tricky to identify due to boundary direction vectors

--
Good luck
Ivar


Thanks for replying. After playing around a bit, I might have figured my issue out. I believe my anisotropy equates to an orthotropic symmetry, and I use a 2D axisymmetric geometry with default coordinates R, PHI, Z, with Z terms =0. Then just input the E_R, E_PHI, v_RPHI and G_RPHI.

The next issue is how to do this with a fluid model, to define radial and tangential bulk modulus of a fluid domain?

~Chris
[QUOTE] Hi you can define the coordinate system to be used when setting up the material property call, and you can use Definitions Coordinate systems to add your own, or use the "boundary coordinate system COMSOL adds up by default (not theat the tangeant directions might be tricky to identify due to boundary direction vectors -- Good luck Ivar [/QUOTE] Thanks for replying. After playing around a bit, I might have figured my issue out. I believe my anisotropy equates to an orthotropic symmetry, and I use a 2D axisymmetric geometry with default coordinates R, PHI, Z, with Z terms =0. Then just input the E_R, E_PHI, v_RPHI and G_RPHI. The next issue is how to do this with a fluid model, to define radial and tangential bulk modulus of a fluid domain? ~Chris

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 01.04.2013, 16:03 GMT-4
Hi

but a fluid is incompressible and isotropic, no ?
hence should only have diagonal elements in their stiffness matrix
note you cannot use nu=0.5, try 0.49 or 0.499 and then you might gain in solver stability by adding the pressure variable (tick "nearly incompressible material" ...) in your physics node, for the appropriate material

--
Good luck
Ivar
Hi but a fluid is incompressible and isotropic, no ? hence should only have diagonal elements in their stiffness matrix note you cannot use nu=0.5, try 0.49 or 0.499 and then you might gain in solver stability by adding the pressure variable (tick "nearly incompressible material" ...) in your physics node, for the appropriate material -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 01.04.2013, 16:14 GMT-4

Hi

but a fluid is incompressible and isotropic, no ?
hence should only have diagonal elements in their stiffness matrix
note you cannot use nu=0.5, try 0.49 or 0.499 and then you might gain in solver stability by adding the pressure variable (tick "nearly incompressible material" ...) in your physics node, for the appropriate material

--
Good luck
Ivar


In wave homogenization theory, you very often describe anisotropic stiff materials with a fluid-like equation with anisotropic bulk modulus or anisotropic mass density. So yes you're right a physical fluid doesn't have an anisotropic modulus, but the effective fluid can, this is what I want to model.
[QUOTE] Hi but a fluid is incompressible and isotropic, no ? hence should only have diagonal elements in their stiffness matrix note you cannot use nu=0.5, try 0.49 or 0.499 and then you might gain in solver stability by adding the pressure variable (tick "nearly incompressible material" ...) in your physics node, for the appropriate material -- Good luck Ivar [/QUOTE] In wave homogenization theory, you very often describe anisotropic stiff materials with a fluid-like equation with anisotropic bulk modulus or anisotropic mass density. So yes you're right a physical fluid doesn't have an anisotropic modulus, but the effective fluid can, this is what I want to model.

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 01.04.2013, 17:38 GMT-4
Hi

that is interesting ...,
as I have often the issue how to model simply, but reasonably correctly, a fluid (static) interacting with a solid when I do not want a full FSI Navier stockes solving (basically a fluid in a solid closed volume) where the solid undergoes small or medium deformation and where the fluid is transferring presure and volume invariant to the solid.

So far I have used a "solid with nu=0.495, + nearly incrompressible checked, and Young modulus around the bulk modulus calculated from the square of the sound velocity times the density, hence ignoring the viscosity, and I do not see what else really should give off diagonal terms, and for some of my examples I can really neglect "flow" assuming infinetly slow motion.

But is this more or less correct ?
it gives me convergence issues when I deform the surrounding solid by a contact identation for example, but this I get also for a full FSI model, appart that it solves even slower ;)

Another case is approximations of a vibrating structure (canteliever) eigenmodes when dipping into a fluid (infinite in extent) here it's easy to add the fluid mass effect on the solid canteliever, but the fluid viscosity should come in too, and again a full FSI with Navier Stokes is heavy (and not implemented as is in COMSOL for eigenfrequencies)

Do you have any good and simple examples, or references, to use to validate such approaches, that would interest me highly, and a few other I have sen strugling with the same issue on the Forum

--
Having fun COMSOLING
Ivar
Hi that is interesting ..., as I have often the issue how to model simply, but reasonably correctly, a fluid (static) interacting with a solid when I do not want a full FSI Navier stockes solving (basically a fluid in a solid closed volume) where the solid undergoes small or medium deformation and where the fluid is transferring presure and volume invariant to the solid. So far I have used a "solid with nu=0.495, + nearly incrompressible checked, and Young modulus around the bulk modulus calculated from the square of the sound velocity times the density, hence ignoring the viscosity, and I do not see what else really should give off diagonal terms, and for some of my examples I can really neglect "flow" assuming infinetly slow motion. But is this more or less correct ? it gives me convergence issues when I deform the surrounding solid by a contact identation for example, but this I get also for a full FSI model, appart that it solves even slower ;) Another case is approximations of a vibrating structure (canteliever) eigenmodes when dipping into a fluid (infinite in extent) here it's easy to add the fluid mass effect on the solid canteliever, but the fluid viscosity should come in too, and again a full FSI with Navier Stokes is heavy (and not implemented as is in COMSOL for eigenfrequencies) Do you have any good and simple examples, or references, to use to validate such approaches, that would interest me highly, and a few other I have sen strugling with the same issue on the Forum -- Having fun COMSOLING Ivar

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 01.04.2013, 17:44 GMT-4
Hi Chris,

What are the equations in the case of “anisotropic bulk modulus”? If you provide the equations someone from the Forum may be able to help. Anisotropic density is more intuitive to me and also easy to implement in COMSOL.

Nagi Elabbasi
Veryst Engineering
Hi Chris, What are the equations in the case of “anisotropic bulk modulus”? If you provide the equations someone from the Forum may be able to help. Anisotropic density is more intuitive to me and also easy to implement in COMSOL. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 07.04.2013, 21:10 GMT-4
Here is another quandary.

Imagine a simple acoustic scattering problem off an elastic solid cylinder (2D space dimension). I assume this is operating under plane strain conditions (thickness 1[m]), though in the acoustic-structure interaction mode you cannot directly select versus plane strain or plane stress, as you can in the structural mechanics.

Anyway, I am comparing the two following cases.

(a) First case the cylinder is modeled as an isotropic solid with E0, nu0 and rho.

(b) Just to see what happens, the second case modeled is the same cylinder with orthotropic anisotropic selected but using the values Ex = E0 , Ey = E0 , Ez = 0 , nuxy = nu0, nuyz =0 and nuxz =0.

I was expecting this to match the isotropic case. But it does not. Why not?

The only way to match to the isotropic case is to also set Ez = E0, nuyz = nuxz = nu0. But if this is plane strain the out-of-plain strains are zero and hence the out of plane Poisson's ratios are zero.

It makes me think the isotropic case is not plane strain at all, though the Acoust_Struct_Interaction documentation states that the 1m 'thickness' is "suitable for plane strain models".



thanks in advance !


~Chris
Here is another quandary. Imagine a simple acoustic scattering problem off an elastic solid cylinder (2D space dimension). I assume this is operating under plane strain conditions (thickness 1[m]), though in the acoustic-structure interaction mode you cannot directly select versus plane strain or plane stress, as you can in the structural mechanics. Anyway, I am comparing the two following cases. (a) First case the cylinder is modeled as an isotropic solid with E0, nu0 and rho. (b) Just to see what happens, the second case modeled is the same cylinder with orthotropic anisotropic selected but using the values Ex = E0 , Ey = E0 , Ez = 0 , nuxy = nu0, nuyz =0 and nuxz =0. I was expecting this to match the isotropic case. But it does not. Why not? The only way to match to the isotropic case is to also set Ez = E0, nuyz = nuxz = nu0. But if this is plane strain the out-of-plain strains are zero and hence the out of plane Poisson's ratios are zero. It makes me think the isotropic case is not plane strain at all, though the Acoust_Struct_Interaction documentation states that the 1m 'thickness' is "suitable for plane strain models". thanks in advance ! ~Chris

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.