Mesh size and material dependency

[niranjanbt]

Dear Friends,

I have two similar structures. The first is modeled with viscoelastic material and subjected to quasi-static analysis,lets call this 'model A'. The other is modeled with isotropic elastic material and subjected to static analysis, 'model B'. The two abaqus models are exactly the same except the material and solution method.

A mesh convergence study is performed for Model A and lets say that a mesh size of 'x' units gave close-to-exact solution with least computational effort. Is it right to assume that the same mesh size will provide me the accurate results with model B?

In other words, is the optimal mesh size for a given structure dependent on the material and solution method?

Any thoughts on this would be helpful. A reference publication that might support your argument will be greatly appreciated.

Thanks,
Niranjan

 

[Elabbasi]

Yes the optimal mesh size is (unfortunately) dependent on the material and solution.

Take for example a loaded cantilever beam modeled with 3D solid elements. If an elastic material is used the stress along a cross-section will vary linearly with distance from the neutral axis, and a coarse mesh using an appropriate element type should give you good results. Make it now an elastic-plastic material, and that mesh will be poor because the stress is no longer linearly varying.


Nagi Elabbasi

http://www.veryst.com

 

[ESPcomposites]

Niranjan,

Is this an actual engineering problem or a "book" problem? Wouldn't the "viscoelastic material subjected to a quasi-static" load effectively reduce this to an static elastic solution? If that is the case, then I am not sure there is a difference between A and B.

Though as Nagi has stated, the general case is dependent on the mesh density. Have a look at the "technical documentation" link on this page :

http://www.espcomposites.com/software/eBolt.html

You will see that the degree of orthotropy affected the discretization error.

Brian

http://www.espcomposites.com