ETABS Beam torsion problem

[SemiPE]

I have recently done a sample model in ETABS, to verify the results of the program, a one storey 6mx6m square grid concrete structure. I have modeled the 150mm slab as membrane. Beams size was 200x600mm and column size was 300x300mm.

No imposed DL was applied (only self wt.). Live load = 5kPa. I then added an additional load case named Wall (for brick wall loads). After that I added a null line object (1m away from a beam) and applied a 5kN/m load on it, to simulate a brick wall load.

I assigned an auto mesh for the slab to limit the size of the mesh to 1m x 1m, since I want the brick load to be transferred properly.

After performing the analysis. I checked the results for bending moments on each of the 3 load cases and the figures seems in line with my hand calculation, except that for the Dead load case and live load case there is a Torsional moments on the beams! Which is supposed to be Zero! I even tried reducing the torsional constant to 0.01 but the Torsional moments is still there.

How could this be? Does anyone know how to correct this?

 

[TLycan]

need to see the model.
anyway, mostly you will have torsional moments on the edge beams due to the rigid connection between the slab element and the beam element.

Its been so long since I worked in RC structure but as I remember we should design the beam on the torsional moment resulting from W*l*l/24

 

[ingdaniel]

agree with TLycan you need to post the model.

But if the slab is a membrane dont apply the mesh... only apply for shell.

regards

 

[SemiPE]

Ingdaniel - if we dont mesh the slab membrane how would the line load; from the brick wall, be distributed properly?

I attached my etabs model with the link below. I modeled two identical structure, one with a slab mesh and the other without. You may notice that for the Wall load, the slab without mesh doesn't distribute the load properly as evidenced by the Shear and moment diagram. While The meshed slab distributes the wall load properly, it shows that there is a significant torsional moments for the dead and live load case.

http://files.engineering.com/getfil...c-46de-4d4f-b43a-262fe9ee82c1&file=SAMPLE.EDB

http://files.engineering.com/getfil...a-1710-44a8-b41c-97c898d28341&file=SAMPLE.$ET

 

[ingdaniel]

Sorry but I cant check out the model, I have Etabs 9.7.2 and when I import the .$ET file the software with me "Error, Divion by Zero"

 

[TLycan]

For the problem of the torsional moment:

assign all the frame end offsets to Zero to know exactly the bending values at the ends of each member you will find the following for the case of dead load:

M1 on beam end = 7Kn.m
M2 on column upper end = 20 kn.m
Torsion on the perpendicular beam = 13kn.m

that is the difference between M1 and M2 is transferred as torsion on the perpendicular beam which is completely logic.

that is of course because you did not release the beam ends; and you cannot release it otherwise you system will be unstable