Cosmos 2004: stress on bonded surfaces

[MoorparkRay]

I have an assembly comprised of two parts bonded together. These two parts have different material properties like modulus and thermal expansion coefficient.

When I applied a uniform temperature on the assembly, both parts expanded and generated thermal stresses. I noticed that the stresses on the bonded interfaces are the same for both parts.

I think it doesn't make sense. Since two parts are bonded together, their interfaces should have the same displacements and strain, but the stress should be very different because of the different moduli.

Can anybody help to resolve this mystery? Thanks a lot.

 

[cbrn]

Hi,

I think you are in error here: for the equilibrium condition at the interface, you have necessarily sigma(radial, disk1) = sigma(radial,disk2) = p(interface). All this is written in absolute value. The strains will consequently be different, i.e. the EQUILIBRIUM (which must be enforced) will happen in a CONGRUENT displacement field where the strains for disk1 and disk2 are different.

Waiting for other comments to confirm or invalidate... ;-)

Regards

 

[MoorparkRay]

Cbrn,

Thank you for your comment.

Do you mean the stresses at the interfaces are the same for two parts but the strains are different?

I do have a question: on the interface, every node on one surface is bonded to its corresponding node on the other surface. For the corresponding nodes, they should experience the same displacement. And because the initial dimension of the two bonded surfaces are the same, the strain should be the same for these correponding nodes. So the two bonded surfaces should have same strain distributions, right?

 

[cbrn]

Hi,
radial stresses at the interface should be the same for the equilibrium.

xsi=r*epsilon_t
where xsi is the radial deformation and epsilon_t is the tangential strain.
epsilon_t=(sigma_t-nu*sigma_r)/E
where nu is the Poisson ratio and E is the elastic module (different for each material you have)
Working a bit with these should lead you to the solution...