Interesting Results in Rigid Building Analysis?

[Backcheckrage]

So check this out:

It's a 6-storey concrete shearwall and concrete diaphragm building modeled in ETABS. Diaphragm is somewhat irregular and is modeled as completely RIGID. Concrete shearwalls are same configuration from floor to floor. The interesting bit is that shears in the wall increase as you go down the building and then get "sucked out" at some levels... these shears appear to be a redistribution of the shear in the walls above. What is going on here? If I were to do this building "by hand" I would simply be performing rigid diaphragm analysis at each level to understand how shears in the diaphragm are distributed to the shearwalls -but using only diaphragm tributary forces! I wouldnt be redistributing diaphragm mass above the diaphragm!... would I?

 

[sponton]

what do you mean by sucked out, as in non-existent? I've had many problems with ETABS lately, I used to run concrete buildings and would have Tension on the basement walls even when I had no uplift forces, I think most of these problems come from over-detailing the 'model' rather than simplifying it. For some reason, sometimes node-merging isn't working quite as expected so matrices go bananas and you get funky results. Have you run the check model to see if there are connectivity issues in your model?

 

[Backcheckrage]

Sponton:

By "sucked out" I mean that the shear in this particular shearwall would DECREASE SIGNIFICANTLY from one level to the next. This is against the idea that the shears ought to INCREASE as you go from top to bottom of a shearwall. I think you're right regarding over-detailing in my model, the only difference in the building shearwall configuration is that the ground floor is ALL shearwall all around it -so its very stiff at ground. I had deleted the ground floor walls and simply extended the shearwalls to the ground... this gave a more reasonable answer -unless someone explain the phenomenon of shears in the wall "leaving" the wall and entering the diaphragm.

Have you had an instance where the shear in the wall would decrease significantly from one level to the next? (presuming shearwall configuration is symmetrical from floor to floor, L-shaped rigid diaphragm, 6 storey building).

I'd appreciate the discussion.

 

[sponton]

Yes and it seems that as models get more complex this is a trend. Sometimes it's due to aliasing issues in the diaphragm-wall boundary when drawing, etabs has the bad habit of getting extremely annoying when you have a set of points lying too close, so it wants to condense them and it normally isn't right. Right now i am fighting etabs, I am analyzing a mezzanine and I did it by hand but etabs keeps giving me lighter loads so the results obviously don't make any sense.

In your case, I'd try to keep diaphragms with regular shapes and finely meshed. try to run the check model for connectivity issues, most problems lie there. Then it's a pain in the but fixing it, I'd rather just start from scratch.

 

[ahmedhegazi]

Do you have basement walls at lower level?

can you post elevation for shear force in wall

 

[Backcheckrage]

Hi ahmed.
no, no basement. However, the groundfloor walls are solid all around (mostly). Meaning to say, it would indeed look like a basement wall i elevation... thoughts?

 

[slickdeals]

Are there more walls at one or 2 levels below where this change is happening. Usually it is due to stiff floor below. Try running it with semi rigid diaphragm at the levels where this happens

 

[ahmedhegazi]

ok , Please do this exercise which will answer your question

Model 1: remove the ground floor walls
Model 2: assign semi rigid diaphragm with ground floor walls
Model 3: your current model


compare RC walls, shear values between the three models

 

[Backcheckrage]

Slickdeals: yes there are all solid walls at the level below where this is happening.

A semirigid diaphragm yields results more inline with what i was predicting. Any theories why?

 

[slickdeals]

A rigid diaphragm has infinite stiffness in-plane, which is a reason for unrealistic shear reversals.

 

[Backcheckrage]

slickdeals:

I figured that out now, even after some research as well. Irregular concrete diaphragms (for example, L-shaped), should not be idealised as infinitely rigid in it's plane. We have standardised rigid diaphragm analysis spreadsheets which should only be used for box-like diaphragms. So basically, an e-tabs model would be necessary for more complex diaphragm analysis.