ETABS slab meshing

[Enhineyero]

Hi, I just want to ask what is the usual size of auto-meshing do you assign for slabs (assigned as membrane) in a framed structure (those with beams & col). I found out that the more i mesh the better the accuracy but i am wondering if you have a high rise structure and you mesh the slabs by 1m x 1m. that would take alot of time. so I am just want to know the usual rule of thumb regarding the auto-meshing of slabs.

another thing i usually assign the interior wall loads as a null beam with a load, the interior wall (usually light bricks) will be placed on slabs. i also noticed that if i auto mesh the slabs into smaller sizes (say 1x1m), the moment & shear diagrams on the beams become much realistic, but if i use the default meshing, the resulting moment & shear is a bit far off from what i am expecting. then again the question is, if i have a high rise structure (say 40 sty) and i mesh all my slabs by say 1m x 1m. think it will take ETABS a long time just to analyze my model.

any advice on this would be appreciated.

 

[MrHershey]

I have stopped allowing ETABS to auto mesh my slabs in most cases. If you've got a thicker concrete slab, then you should be able to justify the slab as a rigid diaphragm, in which case really fine meshes don't really help that much. Would make sure you have your slabs meshed at supports and load points and then at intermediate areas so distance doesn't get too crazy where you don't have supports or load points(every 15-20'/4-6meter seems to be a happy median for me). Finer mesh would be appropriate near changes in plan orientation or openings. The auto-meshing for ETABS also doesn't appear to be that intelligent, so if you've got any unusual geometry such as changes in orientation from a standard orthogonal grid system, the mesh produced by Etabs' automesh is essentially useless.

If you've got a composite slab, then perhaps the 1' x 1' mesh size is a bit more appropriate. Could still look at going a little coarser. Compare the outputs and see what you get. If your run time is cut in half while member forces and errors in modal analysis only change by a couple percent, than I would say you're fine.

 

[Enhineyero]

thanks mark, appreciate your input.