Meshing Composites in ANSYS

[Chantzypoo]

I am trying to model a composite plate in ANSYS using SHELL91 or SHELL 99 elements and it seems that the results are highly dependent on the mesh no matter how fine or 'clean' the mesh is. I have tried many variations of mapped and free meshes with different area divisions, and I seem to get different results every time, at least as far as contour plots and deformation are concerned.

Does anybody have experience with these elements, and/or any tips on achieving mesh independence with them?

My mesh has already gotten very fine relative to an equivalent isotropic analysis and I can't imagine that going 'ultra fine' is the answer since it would make modelling of complex geometry extremely innefficient.

 

[thenga]

I have had similar experience before when using shell elements for orthotropic material modelling. of particular interest is the stress values that ansys generates. be careful about what we are looking for interms of stresses and strains.

I am modelling a curved surface (a cylinder for instance) and try to mesh using shell99. since i have created area for the geometry, it is not allowing me to mesh along the surface.

i am looking for global twist dof output and so i cannot use solid elements as they have only ux, uy, uz dof inbuilt.

any help would be highly appreciated.
thanks
thenga

 

[Chantzypoo]

Although I haven't tried it yet, I was informed that results are much nicer if you use the SHELL99 elements with the material ABD matrix instead of using the layer feature.

I am currently working out my various ABD matrices now and should be able to try it out later this week. I am also modelling a 3D surface so I will let you know how it goes.

I was also told that althogh the distributions of stresses and strains over the area and through the thickness may be innaccurate, the maximum and minimum values for the laminate may be ok, so that may be something to consider.