Flat slab designing for shear- peak shear or average shear? 1

[sharp910sh]

If you have a flat slab, and you are checking punching shear. you have made a FEA model of the slab, and you see there is a peak shear force. 8 times larger than the average along the perimeter. do you design for the average or the peak? will some sort of redistribution occur?

thanks sharp910sh

 

[DaveAtkins]

You should design for the average shear over the appropriate shear perimeter.

FEA programs will show a very high peak shear, which tapers off very quickly, around a support. But in order for a punching shear failure to occur, the entire shear perimeter must fail.

DaveAtkins

 

[sharp910sh]

Thank you I was thinking so. Do you get some kind or redistribution of the peak shear stress?

 

[sharp910sh]

what if the peak stress is near constant over one portion and then drops off. would you still design for the average stress across the whole perimeter.

I have included the sketch of what im getting.

 

[hokie66]

FEA as a method of analyzing concrete structures requires a lot of interpretative ability, as Dave implied. The shear given by your FEA analysis does not correspond exactly to the way shear is calculated in the code, but it does show how eccentric loading at an edge or corner condition amplifies the shear stress. Based on the output of your FEA model, I would judge your shear stress at that corner to be much too high.

 

[ishvaaag]

sharp910sh presents a doubt that is interesting from a professional practice viewpoint. What if what you are getting seems to be against your understanding of what would be proper? I would say that in the instance what you need is to clarify in whatever the way what happens to your acceptance of that it meets both your understanding and the ruling code.

For example, for cases clearly defined in the codes (and should there be scarce ones that are not, but still there are, and many times quite intentionally) you can proceed as the code says and if in the end you see the thing correct, end of the issue. DaveAtkins correctly points that following the code normally may relieve you of the problem. In this case he comments as well on why, so at the same time is providing supporting engineering judgement for the code.

You may discover some conceptual error that leads to the problem, as hokie66 says, since at 50 kgf/cm2 shear stress ordinary concrete may be (at least conceptually) utterly cracked and just be rubble within cages (there have been failures where such kind of concentration of stresses have initiated locally a failure that from then the remaining reinforcement has not been able to restrain). Then, for the case, the scarce thickness leads to high (and quite likely unacceptable from a code standpoint) shear stresses, and the deformations are also high contributing to the concentration of stresses and your problem. Augmenting the thickness may reduce the problem to more acceptable averaging.

 

[hokie66]

Not sure I am reading your post correctly, ishvaaag, but I wish you wouldn't use those units. What do you have against the SI units of N/mm^2, more commonly known as MPa?

 

[hetgen]

Flat slab punching is scary stuff.Forget the FEA, do it by hand.

 

[ishvaaag]

Nothing at all, hokie66, only custom, since I have a nicer mental understanding of what a kgf is and a cm2 is. It was called the "technical" system once, maybe for such reason. Anyway as of now I am cozy enough with SI units (yet not as much as this old metric).

 

[dcarr82775]

If you have the slab support modeled as a frame element, just take the axial load in the column as your punching shear load. Also, take the moment and adjust be the gamma factor to get the moment transfer via shear.

Similar if you have the support modeled as a point instead of a frame element

 

[EricCA]

hetgen is correct. FEA is a complex animal and each piece of software handles it differently. At a minimum you should do some rough hand calcs to verify the FEA results.

I don't totally agree with DaveAtkins about averaging over the the entire perimeter. I agree that you do not need to design for the peak, but averaging over the entire perimeter may be non conservative. ACI 421.1 says "When the maximum vu occurs at a single point on the critical section, rather than on a side, the peak value of vu does not govern the strength due to stress redistribution (Brown and Dilger 1994). In this case, vu may be investigated at a point located at a distance 0.4d from the peak point. This will give a reduced vu value compared with the peak value; the reduction should not be allowed to exceed 15%."

 

[PicoStruc]

1) First, How can you check the punching shear with FEA without defining polar inertia of the critical section ? Do it by hand or use a FEA software done for that purpose, example : SAFE.

2) Not sure if I understand correctly but it is normal that you have a peak in one of the four corners of the critical perimeter because of the effect of the moment in both direction added to the averaged shear stress!.

 

[SethGuthrie]

I'm obviously a big fan of RAM Concept for the evaluation of punching in a flat slab. The approach we take there is to determine the critical demand from the net column unbalanced reaction (or just the column reaction if there is no column above or point loads to deduct). Then we apply the traditional ACI or other code punching shear equations to determine the failure plane and the capacity (with or without reinforcing). There are tools available (generally from studded shear reinforcing suppliers) if you don’t have the luxury of an integrated application like Concept.

I would never directly correlate the Finite Element analysis through-shear stresses from the mesh of shells to code allowable shear stresses.