still floor meshing

[hannes964]

Hi folks,

i’m still really unhappy by meshing floors with areas. I handled to mesh the geometry manually, which is irregular, in six big quads. So far the model runs without warnings. I tried to generate a finer mesh by dividing the areas into sub-elements to get the proper deformation of my floor. But then I receive that some areas are too close by checking the model. Someone gave me the hint to delete these areas and rebuilt these areas manually (thanks civilungo). So far so good if the number is small. In my example the number of areas which are too close is about 100. So I have no chance to redraw these areas manually. My question is actually really simple, how can I mesh irregular floor areas in etabs? Or are these warnings harmless???
I’m really depressed…I’m wondering that nobody has any problems by meshing floors in etabs (I use version 9.0.7)

Thanks a lot

hannes

 

[zeemas]

Hi
In my opinion if you have very regular building then go for Auto meshing, because as you know that manual meshing makes the life so much difficult specially Architects are involved too much. Any way if you want to go for manual mesh, I have one more method which i usuall do. Make a very basic model and export to ROBOT or AUTOCAD, generate the mesh if you are using ROBOT by number of meshing method or just create manual mesh in AUTOCAD and then import back to ETABS, Assign properties and loads and annalyse, this is my experience that chances of warnings are very low. Also for warnings i tell you one thing, if error is less than Exp-11, then I think you can neglegt the warnings, in this case check the reactions and other things carefully.

 

[opa123]

you can also build your floor model in safe and export it as dxf abd then take it to etabs.

 

[zeemas]

can do with SAFE, but many places mesh aspect ratio becomes very bad, which can cause these warnings again.

 

[GarthJ]

I generally use automesh and control the meshing at special locations with grids and/or null lines.
I would look at your force balance of your model and then I would look at how many digits of accuracy are being lost. They don't seem to worry too much if there are six digits or more lost.
You must gain confidence in your results. Remember you can use the process of elimination to get rid of your problem, study the forces and deflections of this new model and then compare with the model that has the trouble. This may be a way to gain some confidence in results.

Hope this helps