Procedure to design RC walls when transferred slabs are involved

[RyanSt]

Can anyone guide me on how to design RC walls when transferred slabs are involved?
I realize that the transferred slab will have great impact for the bending moments in the walls above.
Say I have a 3 stories building with PT transferred slab on 1st floor. When I use ETABS for analysis and design and I realize the bending moment in walls above 1st floor will be huge and thus lead to extensive reinforcement required because of the existence of transferred slab below in my model. When I get rid of the transferred slab below (ie delete first floor) and the wall results look normal then (only min reo are required). I think the reason is the PT transferred slab is too 'soft' in my model (I model as shell-thin) and thus affect the bending moments in the walls.
One more thing I think of is when we use FEA package say EATBS, we have the whole structure constructed 'all at once' without 'staged construction analysis' and thus the results may be weird and should not be used in wall design?
I know it is a 'too general' question when still wonder how you normally design RC walls when transferred slab is involved?
Can anyone explain to me in details if possible? Thanks.

 

[Agent666]

Can you post some examples of what you are looking at, I'm struggling to visualise the scenario being described whereby it causes huge moments (under gravity case?) above a transfer slab?

 

[jayrod12]

How thick is the transfer slab? Generally I would expect the slab to do nothing but support the walls. It shouldn't be putting any additional stresses back into the walls above unless, as you indicated, it's too thin and therefore deflecting too much.

 

[RyanSt]

It is a general question because I have this kind of issue many times so I cannot tell how thick the slabs are. But I will upload a model with 350mm thick conventional transfer slab.
Same. That's why I detected the transfer floor and then got 'reasonable' results.
I think the 'abnormal' bending moments in the walls are caused by the interaction with the transfer slabs below. Yes, when it is PT slab it is thin but the deflection shouldn't be much as it is PT (I simply define the slab as thin shell so the PT effect was ignored though). But when the slab are not PT, I have make the transfer slab to be thick enough so the long term deflection of the slab is less than 25mm and L/500 but I still get the 'unwanted' results in walls.

 

[RyanSt]

Yes it is gravity. Sorry for that.

It is a general question but I have uploaded a simply etabs model here.
below is the plan view.

upper floor​

lower transfer slab​

The dimension of the upper floor is two 10m x 7m . slab is 300mm thick.
The dimension of the lower transfer slab is two 12m x 10m. slab is 350mm thick (not PT).

I did a quick wall design in Etabs and the vertical reinforcement ratio under gravity comb is as below. Sometimes I will get near 2% ratio!

 

[KootK]

1) Sadly for me, I don't currently have ETABS. Please patient with my attempt to fly blind on this.

2) My guess is that, on that right wall, your monster wall moments are strong axis and accompanied by monster shears.

3) As shown below, I hypothesize that you might be getting some story high Vierendeel frame action happening.

4) These high moments will crack your walls flexurally and, most likely, open up the relatively crappy moment joints between the walls and the slabs. The high shears may also induce some shear slip at the shear friction joints where the walls and slabs connect.

5) The cracking in movement in #4 can be used to advantage here. I'd look at using those things to justify softening the walls flexurally in ETABS via the wall property modifiers. That will reduce the stiffness of the Vierendeel action and, hopefully, steer things back in the direction of more palatable results. You could even pin the column tops and bottoms in your model if you're not needing those walls for your lateral load system.

6) 3D transfer slab models tend to produce a lot of weird results. And, frankly, there's a fair bit of truth to them despite our dogged attempts to make the weirdness go away. As previously mentioned, stiffness is very much your friend here. It does sound as though you've got that covered though as your 1"/L/500 seem reasonable to me.

 

[RyanSt]

First of all, thank you very much for your reply. Really appreciate that.
In regard to modifying the wall stiffness and let it crack under gravity, I am wondering how wide the crack width will be and will it be considered as service failure? Also, I am not sure if I fully understand the mechanism when it comes to cracking. Let's say I have min. reinforcement in the wall, which has lower moment capacity than the analysis one with unmodified stiffness, what will happen before the wall cracks? Does that mean the vertical reinforcement will yield before the wall cracks as illustrated on your second figure? If so, is it okay to let it yield under gravity? Also, even if the walls are not designed for lateral purpose, is it allowed to let it crack under gravity? Do we need something like boundary element to make sure the vertical reinforcement be restrained under seismic as concrete already crack under gravity?

I think the high Vierendeel frame action as you mentioned is worst by the fact that I didn't run construction sequence case. The moment should be less than what I got if I take construction sequence into consideration. I know ETABS provide construction sequence analysis but I don't think many people will perform this fancy analysis. So do you take it into consideration for wall design? BTW, may I ask what FEA software you use for building?

Thank you.

 

[r13]

When I get rid of the transferred slab below (ie delete first floor) and the wall results look normal then (only min reo are required).

When get ride of the transfer slab, where is the base of the walls above? Ca you please elaborate? Thanks.

 

[RyanSt]

I simply deleted that whole story and assigned both ends of walls as pin restraints.
The moment was caused by the transfer slab below or, like KootK mentioned, the vierendeel frame action.

 

[r13]

So, when with the transfer slab, the walls were pined to the slab too? Seems the rigidity of the connection has something to do with the different results.

Another note is, on the end walls, if Mx is excessive, it could be vierendeel action in play, otherwise, just reactions from rigid frame action.

 

[RyanSt]

No, just rigid connection for walls-slab.
I am pretty sure it is vierendeel action caused by the transfer slab. Not sure about the mechanism of modifying wall stiffness to let it crack like KootK mentioned though.

 

[r13]

Two approaches to reduce the moment in the walls above the transfer slab: 1)thicken the transfer slab, 2) specify partial moment capacity for wall-slab connection. I'll avoid the latter, as it is not a simple method.

 

[RyanSt]

I don't think thicken the transfer slab is necessary when the long term deflection is under limit. Also, sometimes transfer slab is PT, which means the deflection will be under control even when the slab is thin. However, in the full model (not designing the transfer slab itself), I (also many others I think) will simply assign the slab as normal thin shell without adding tendons etc, so this vierendeel action will still exist.

 

[blihpandgeorge]

i have seen similar problems as what KootK has said with viereendeel action giving the wall moments. a further effect of this is that i have had the transfer moments reported to be much lower than i expected. i cant recall if you ETABS allows you to place moment releases in the columns -if it can then this will help?

Whatever you i strongly recommend a hand calc of the column load and use that to check the transfer moments.

 

[KootK]

First of all, thank you very much for your reply. Really appreciate that.

You're most welcome Ryan.

In regard to modifying the wall stiffness and let it crack under gravity, I am wondering how wide the crack width will be and will it be considered as service failure?

It's very difficult to say but my suspicion would be that this would be fine given that:

1) You'll design the system to the tight deflection limits without the benefit of that Vierendeel truss action.

2) Given that the slab to wall joints at the top will suck for the transfer of both large moments and large shears, I suspect that the walls will draw less load than your modeling suggests.

3) You'll have a cold joint "crack" at the top and bottom of the walls. I'd expect much of the cracking that occurs to be concentrated disproportionately in those preexisting cracks which should a) rarely be in the view of occupants and b) look somewhat orderly, like slab on grade control joint cracking.

4) As you mentioned, consideration of construction sequencing here will reduce the Vierendeel effect if transfer level formwork and shoring will be removed before the upper level slabs will be cast.

Another avenue of attack for you might be to mesh the slab more finely in the vicinity of the walls that are drawing undesirable moments. You want to ensure that course modeling and/or goofy element shapes are not adversely affecting slab flexibility near the joint.

lso, I am not sure if I fully understand the mechanism when it comes to cracking. Let's say I have min. reinforcement in the wall, which has lower moment capacity than the analysis one with unmodified stiffness, what will happen before the wall cracks? Does that mean the vertical reinforcement will yield before the wall cracks as illustrated on your second figure?

Nobody knows what will happen with any certainty. I would expect the walls to crack before the vertical reinforcement yields. Those "cracks" may well be predominantly the cold joints as I mentioned above however. Also keep in mind that the moment connections at the tops and bottoms of the walls will be lousy. I suspect that you'd tear the bars from the slabs long before you yielded them.

If so, is it okay to let it yield under gravity?

It would be fine so long as the walls are not part of your lateral system. And I'd hope that to be the case for this setup precisely because the moment connections between the wall and slab at the top and bottom of the walls are dubious to begin with.

So do you take it into consideration for wall design?

I don't unless it's a very large or very complex building that I'm investigating. For modest structures weir I don't expect weird things to be going on, my default will be to model the transfer slab on its own with the structure above treated simply as applied load.

BTW, may I ask what FEA software you use for building?

I'm currently using RISA. It's Achilles heel is that it doesn't yet do a reasonable job of concrete deflections. I'm hoping that will change with RISA's recent acquisition of ADAPT.

 

[RyanSt]

So what happens if the reinforcement does yield before concrete cracks? This is one of my concern when modifying wall stiffness to let it crack.

 

[RyanSt]

Yes it allows me to. I can also modify the stiffness to achieve that. What really concern me is I am not 100% confident about the consequence of doing this.

 

[KootK]

So what happens if the reinforcement does yield before concrete cracks?

Does mild reinforcement ever yield in tension before the surrounding concrete cracks in conventional situations? I thought not. It would be wise to proportion your reinforcement to meet the flexural minimums so that you get distributed concrete cracking as much as possible but I would assume that you're already doing this and that the corresponding reinforcement ratios would be practical.

 

[rapt]

Ryanst,

How can the reinforcement yield before the concrete cracks? Before cracking the strain in the concrete and the reinforcement is the same. The stress in the reinforcement is n times the stress in the concrete which is about 7 * concrete stress. Concrete cracks at about 3 - 5 MPa (500 - 700psi) so the stress in the reinforcement at the time of cracking is about 30Mpa, and yield is about 400 - 500 MPa.

The problem is if the concrete cracks and there is not sufficient reinforcement to carry the tension transferred to the steel. Then you get a single large crack rather then thinner well distributed cracks.

So the area of reinforcement must always be greater than is required for the cracking moment.

Why will the moments in the walls above be so huge anyway? If everything is sized logically they will be no worse than the moments from the floor above.

Is ETABS allowing for the reduced deflection of the slab due to the prestress in determining the wall moments. I doubt it!

Most designers before the days of unrealistic super accurate and misleading FEM analysis would assume pinned bases on the walls above and ignore the walls above in the transfer slab analysis. (Note there is a lot of sarcasm in this last sentence!)