Highly-loaded anchor bolts in shear

[JSA2]

I have some very highly loaded anchor bolts ("through bolts") in shear. The service load shear force is 32kips to the worst case bolt of an 8-bolt group. The bolts are "through bolts" through an 8" slab with a thick steel plate at both the top and bottom of the slab. The concrete crushing of the bolts in shear will be restrained by the top plate. (The bolts are shear-loaded from the top of the slab.)

I imagine that ACI 318 Appendix D will be too restrictive and will not provide guidance on this condition. Does anyone know of any resources which would provide contemporary limit states for highly-loaded anchor bolts in shear when the concrete crushing limit state is restrained by the top steel plate?

 

[abusementpark]

Are you trying to develop the shear load into the slab or through the slab?

 

[a2mfk]

It would help to provide a sketch and show the loading and edge distances at a minimum. Is the structure already built, and you are trying to add this? Or is this for something new?

 

[ishvaaag]

See an approximate FEM model attachment for a case close to yours and included considerations.

If in the end you would be dimensioning your device by shear friction, just make the capacity of the set of bolts be higher than the factored shear, and the thickness of the plates enough to sustain quite rigidly at factored level the magnitude of shear force uniformly distributed as compression to be hold by the bolt hardpoints (implied friction factor = 1, or modify for yours).

 

[JSA2]

Thank you all for your contributions and attempts to help with this.

My condition, as is shown in the attached pdf, is a plate with large shear loads. Because of the confining effect of the plate below the slab, the limit state for shear in the bolts is unclear to me.

I met with the local Hilti rep this afternoon, who believes that South Korean testing has shown that in this condition, where the top plate "confines" the concrete shear failure at the bolts, the bolts themselves will fail in shear before there will be a concrete consequence, at least for bolts up to 2" in diameter. This includes high strength bolts.

I am now looking for support for these statements. I thought I was looking for some kind of limit state assessment of the concrete below the confining top plate, but now I'm thinking that only testing can show this and I am now looking for this testing.

Interestingly, running these numbers through the Hilti Profis program seemed to indicate that 1 3/4" diameter 105 ksi yield bolts will work for these loads, assuming only failure of the steel.

Now my question is, does anyone know where we could find this reputed South Korean testing?

 

[dik]

The AISC specifically recommends against doing this... should consider shear lugs...

Dik

 

[a2mfk]

"and edge distances at a minimum. Is the structure already built, and you are trying to add this? Or is this for something new?"

We still don't have all of the basic parameters of your problem..

 

[JSA2]

dik, I will check AISC to see what they say; however, it's seems strange to me that a steel organization would opine on what seems to be a concrete performance issue under the steel plate. Do you have the doc and section of the AISC reference?

a2mfk, one plate is far away from edges. The other is say 18" plate edge to concrete edge. The highest direction shear load is toward the free edge, of course The 8" slab is not yet placed. I intend the bolts to be threaded rod cast into to the concrete.

 

[dik]

I will try to dig it up tonight...

Dik

 

[a2mfk]

With that info (new slab) I am with DIK, go to a shear lug and reinforce the slab as necessary. This gets you out of the dreaded App D...

 

[JSA2]

a2mfk,

In my case, I am a peer reviewer for a design that is already done. One of my tasks as such is to determine a bolt size and capacity of this connection, as designed. If I were convinced that the connection is inadequate, then of course I would have to recommend revision. At this point, I do not yet believe the connection is invalid. I do believe there is testing out there which addresses this.

You know, theoretically-speaking, if the concrete were to crush under shear, you would have a translation of the top of the bolt, followed by a shear friction mechanism developing between the plate and concrete, assuming adequate stiffness of the plate. (This may be why AISC does not recommend this??) I expect that this development of the shear friction would serve to arrest the increase of shear in the bolt and concurrently increase the tension until you eventually get a shear/tension failure in the bolt, given that the bottom plate will preclude a bolt pullout failure.

Thank you for your contributions to this thread. I will post what I learn and my conclusions.

 

[ishvaaag]

I have been trying the case in the sketch in FEM (I still am) and in the case that surfaces contact, no bond and no friction a solution is found but I think not surprisingly the crushing stresses in concrete are too much; then I add friction and I don't manage to get a solution without warning error (yet) (1.25 inch diameter bolts) what makes me think that the stresses imparted by the solicitations to just 8 inserts may be found too much. Whilst the computer crunches numbers I am thinking that a simple evaluation of the force at each bolt can show this, on a linear rotation of the plate, and doing it such way I get that the worst inset rod would have to pass 13.18 metric tonnes, what would make 40 MPa crushing stress on an optimistic half depth of the concrete inset (projection) (actual bearing should be less). The model that closed did at around 70 MPa. The other is still crunching.

So I start to think that 8 1 1/4" bolts are not going to work for the forces in the sketch.

 

[ishvaaag]

If this is the case after so becoming convinced, I would introduce a modification that would ensure positive introduction of the forces. The bolts inserts would disappear and something far more substantial would take their function within the concrete.

 

[JSA2]

Ishvaaag,

I ran the problem through Hilti's Profis program. Based on steel strength alone, it indicated that 1 3/4" bolts at 105 ksi yield would work. We ignored concrete effects, just to see how strong the bolts would have to be.

Thanks for presenting these numbers. What program are you using to do this analysis?

 

[ishvaaag]

I discover than in my worksheet estimate I have made an error in a final vectorial combination, that I made as if the two vectors were at 90 deg when are not; angle yet to be determined; anyway the forces to be passed stand above 12.84 tonnes (and might near 16 tonnes) and so the comments stand.

 

[ishvaaag]

The bolts also passed in the contact without friction evaluation that closed. Was under 40 MPa (that I remember, for I have been overwriting the model). It is the concrete that it seems to me is failing.

 

[TXStructural]

Grouting or using a rigid adhesive/filler around the bolts would reduce problems associated with local crushing where the bolt diameter is smaller than the hole or where the holes are not perfectly aligned. I do not see anything wrong with the proposed solution, and it will likely be stronger than anything involving embedded, headed studs.

As large as the plate is, I agree that something to stiffen it from out-of-plane bending would be useful. If pre-installed plates or cutting slots into the slab was an option, attaching stiffeners to the back of the plate would also provide lugs to transfer the load into the concrete.

 

[ishvaaag]

The program is the current reencarnation of Algor, that its now buyer-owner-developer Autodesk calls now Autodesk Simulation Multiphysics 2012. The solution with friction has failed again but maybe on its no convergence to a solution. I am not precisely the most reliant nor reliable user of FEM yet, but I like to use it to learn, someday I may be more. Also, for this particular case, I find the "hand" check (really a worksheet on instantaneous center of rotation for bolts on shear and 1 eccentricity -so I have yet to properly add the transversal shear) made with Mathcad quite consistent and promising in that some day I can use these things with more confidence.

 

[ishvaaag]

I could concurr with TXStructural but I frankly wouldn't feel cozy with evaluated crushing stresses at 40 MPa or over level on 35 MPa concrete, even with allowance for 2D confinement; the tests may pass marginally because would be facing actual limit strength, but I don't see these are the levels to be imparted to concrete as per a calculation, I mean, you need to leave some allowance for crushing stress even at factored level.

 

[ishvaaag]

If you want, I would admit the level of stress permitted for struts in strut and tie schemes multiplied by a factor derived from some evaluation of bidimensional confinement effect, I think to remember ACI 318 has one. Noting that there maybe close to that point ugly tensile stresses trying to spall the concrete that seems enough.

There may be more to exact from concrete, and anchor vendors may be making a life from it; but then, they test and warrant. And if some particular is not in their catalog, no warrant. So the designer/reviewer decides.