Concrete two-way slab, localized shear stresses

[TehMightyEngineer]

Working on a small two-way slab with an opening and slightly non-parallel edges. Created a quick finite element model to get a reasonable approximation of the moments in the slab for flexural design. There's a small re-entrant corner on one side of the slab which causes a extremely high localized shear stress. In the past, I've generally provided additional reinforcement in these areas but am really wondering how "real" these stresses are? Obviously they exist, but once a crack forms the stress will be relieved and distributed out along the width of the slab, correct? Essentially, you can't shear a "portion" of the slab, you have to shear it from one edge to another, correct?

Please see the attached picture for the slab and stresses.

Also, I am aware that a one-way shear failure can occur between the re-entrant corner and the manhole opening. This will obviously be designed for.

Maine EIT, Civil/Structural.

 

[KootK]

once a crack forms the stress will be relieved and distributed out along the width of the slab, correct?

I hold similar beliefs. Moreover, I feel that there needs to be some compatibility between the moment at a given location and the shear there. Once a flexural crack opens up, your shear will start to redistribute in plan too.

That being said, it's tough to know how to take advantage of that redistribution in a conservative way. So I don't. I would treat this as a punching shear situation and add stud rails if required.

The issue arises partly result of modelling inaccuracies. Because your walls fix the slab vertically, and do so a foot or two away from the corner, the FEM simulates almost perfect fixity at the corner and therefore draws a ton of shear. It's a bit reminiscent of the back stay effect for tall shear walls with basements.

If some of your wall supports beyond the corner could be omitted or turned into springs, it might yield more accurate results and lower shears. But again, who knows how to do that accurately and, more importantly, who has the time.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[KootK]

Remembered one more thing. I doubt that the slab would ever come down due to issues at the corner. If anything, I suspect that it would be a cracking/serviceability failure. On occasion, I have let the shear overstress ride and provided bottom steel coming off of each wall proportioned like integrity steel at interior columns. That way, if the corner cracks apart, at least I know that the slab isn't going far.

I'd love to see your final rebar layout. She's small but tricky.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

Mmm, I too suspect modeling shenanigans are a big part. If I were to refine the supports to very stiff, compression only spring and refine the model's mesh at the corner I bet much of those stresses will vanish.

Actually, in the interest of comparison, and because I have a light workload today, I've done just that. See the attached image for the difference.

...Yeah, this reduced it to what is clearly a modeling stress concentration and I'm no longer concerned that this was a "real" stress. I'm still curious how other people treat this sort of modeled stress, though.

As for the rebar, it's a 8 ft span for the long direction, H-20 load rated cast-in-place concrete cover. Very small job so I'm mostly going to be conservative on the rebar when I present this to my boss for approval. I'll probably do much of the reinforcement based on conservative combination of the FEA moments and shears combined with the strip design method. I'll post a quick sketch of my rebar arrangement if I have time later today.

Maine EIT, Civil/Structural.

 

[TehMightyEngineer]

Also attached is the primary flexural stress for that load case, the moment in the perpendicular direction is about half these values. The wheel load for this load case was located at the high flexure location.

Maine EIT, Civil/Structural.

 

[TehMightyEngineer]

Because you asked Kootk, rebar design attached.

Maine EIT, Civil/Structural.

 

[KootK]

Thanks for indulging me TME. Based on your detailing, your slab is actually free to lift off from the walls. I hadn't realized it was that kind of lid. I should have been thinking pre-cast.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[TehMightyEngineer]

I designed it to work both ways, this was originally going to be precast but was changed to cast-in-place with wall rebar hooking up into the slab.

Maine EIT, Civil/Structural.