Suggestions for modeling a flat plate

Famous_Mortimer

I have a flat plate with some holes that varies in thickness from 1" thick to 4" thick. I am trying to create a .liml file generator in excel similar to the mesh generator that is in the sample files. Could the analysis be done using the shell options for materials using quad elements or 4" too thick? Suggestions please.

prop_design

well a perfect cube is the ideal fea element shape. if you meshed the surface with 4" quads then the shell model would be 4" thick. i'm not sure of your geometry. but seems like something you could do in mecway directly. the other option would be to use laminate elements. you can set that up to have however many elements you want through the thickness. from what i have seen with those, it doesn't seem to matter if they are very thin. so the aspect ratio can be rather odd but the results look ok. i didn't do a detailed look at it though. however, they seem to tolerate odd aspect ratios for whatever reason.

Victor

I'm not sure if you're using the internal solver or CCX. prop_design's answer is more relevant to CCX where thicker shells can perform better because they're expanded to solids.

With the internal solver, thinner shells are usually better and "thin" depends on the other dimensions of the part. If the plate is much bigger (say 10x 4"), then it's OK. However if the holes or other details are smaller, say 1" diameter, you may be losing accuracy in the stress in those regions compared to solid elements. The internal solver's shells use the plane stress assumption and that won't be valid for a small detail on a relatively thick plate.

Not clear if one plate has multiple thicknesses but if so, consider the transition between thicknesses. Shell elements won't capture the stress concentration there properly, which is OK if you're not interested in that or if it doesn't even exist on the real plate.