Breakout of anchors vs punching shear 3

[JIMEY]

I have always used the formulas in Appendix D to calculate the strength of anchors in breakout, pullout, blowout, etc. Just yesterday I was designing anchors for a slab, and the problem that I'm getting is that even if I put the anchors all the way down to the bottom of the slab, I'm still getting a breakout failure. This didn't make sense to me because I thought that if the load is that high, then shouldn't the slab also have a punching shear failure? After all, isn't breakout just a form of punching shear? So I checked the punching shear of the slab using the standard 2 way shear formula in ACI 318, and sure enough it passes. Can someone please explain to me why the formula for breakout results in way less capacity than the formula for punching shear. In my mind, an anchor with a washer on the end is no different than a bearing plate sitting on a slab. So why is it that something bearing on the outside of the slab is given way more capacity? Thanks for any assistance.

 

[JedClampett]

I think that all Appendix D anchor design is more conservative than the "traditional" concrete capacities (shear, moment, bearing, etc.). Maybe it's because Appendix D all testing based.
And I can't think of any punching shear failures I've seen, but I have seen anchor failures (thankfully not many).

 

[KootK]

After all, isn't breakout just a form of punching shear?

I'd agree with that in general.

Can someone please explain to me why the formula for breakout results in way less capacity than the formula for punching shear.

Can you share the relative capacities of the two failure evaluations? I'm curious to know just for my own edification.

an anchor with a washer on the end is no different than a bearing plate sitting on a slab.

Obviously, both capacities are testing based. I can think of two likely explanations:

1) You loose some effective depth in going from punching (column below) to anchor pullout simply because the washers are not positioned at the bottom of the slab.

2) Punching shear provisions assume that there's a monstrous amount of flexural stress in play at the connection (typical column/slab connection). The compression field that forms as a result of that flexure benefits the connection greatly and facilitates the use of shear stress considerably higher than our one way values. Such a flexural compression field would not have been in play during Appendix D testing.

In my mind, an anchor with a washer on the end is no different than a bearing plate sitting on a slab. So why is it that something bearing on the outside of the slab is given way more capacity?

I believe that most of us see it that way. I've seen many engineers go to a through bolt & plate just so that they could call it punching shear. Numerically it pans out but, rationally, I'm dubious that it's a significantly stronger connection. And anchors are certainly better with regard to cost and fire proofing. I worry about this same thing when you have transfer columns that are so near the offset columns below that they really don't produce any serious slab flexure. I suspect shear stress limits closer to one way make more sense but that's not what's commonly done.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JoshPlumSE]

I've seen many engineers go to a through bolt & plate just so that they could call it punching shear. Numerically it pans out but, rationally, I'm dubious that it's a significantly stronger connection.

I agree with both of KootK's comments here. I've seen engineers using through bolts so that they can call it punching shear. But, there is something about this that makes me inherently uncomfortable.... I would really like to a) see these methods converge towards the same numbers, or b) understand the physical differences in the models or assumptions that lead to the two significantly different numbers.

 

[JIMEY]

Thanks for your responses. I agree 100% with KootK's first comment. In my punching shear calculation, I actually only used the embedment of the anchor as my effective depth to make up for this. I completely ignored the concrete beyond the washer.

I've posted my calculations here. Feel free to have a look. As you can see, I'm getting over double the capacity using the punching shear formula. It is difficult for me to accept this. To JoshPlum's point, you could just make the anchors a little bit longer so they protrude right out the bottom of the slab, and call it punching shear. But what's the difference whether they go all the way through, or not quite all the way through? As long as you account for less embed, I would think the principles of design should be identical.

By the way, my calculations are done using the Canadian version of ACI 318, so I apologize if some of the symbols are different. But I'm pretty sure the procedures are identical.

 

[KootK]

I've posted my calculations here. Feel free to have a look. As you can see, I'm getting over double the capacity using the punching shear formula.

Thanks for that. Quite the discrepancy it is. I practice in Canada too so that reads just fine to me.

I agree 100% with KootK's first comment.

Do you disagree with my second comment? I believe that to account for the lion's share of the difference that you're seeing.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JoshPlumSE]

Some additional thoughts on the subject:
a)Punching shear equations do assume flexural reinforcement, and are based on the depth to that flexural reinforcement. So, what KootK was saying in his 2nd point makes some sense. I don't know enough to wholly agree with it. My tendency is to thing of this punching failure almost as a strut and tie action between the concrete shear and the flexural reinforcement.

b) The Anchor rods formulas, I believe, are mostly based on "free field" failures. Meaning that they neglect the presence of other forces or reinforcement that may serve to provide some confinement to the concrete or increase its capacity.

That's just my thought process as I rationalize my way through the discrepancy. I don't do a whole lot of anchor rod design these days (not since the 1997 UBC days). My feeling is that there has to be justification for using the existing flexural reinforcement to increase the pullout strength.

 

[Deker]

I recommend reading this paper for a discussion on the differences between Appendix D and punching shear capacities, at least as they relate to anchor groups that share an embedded anchor plate: Link. The test results from the study show that Appendix D tends to underestimate the actual connection capacity (no surprise), while the the punching shear provisions tend to overestimate the capacity. The authors offer the following explanation for the inability of the punching shear provisions to accurately predict anchorage capacities:

However, a closer inspection of the sources (ACI-ASCE, 1962) used to formulate this approach reveal two interesting factors. First, it does not explicitly incorporate fracture mechanics or the “size effect” in concrete (Bažant, 1984), which implies that the unit strength (or failure stress) of geometrically self-similar concrete components varies inversely with their size because failure is controlled by localized fracture rather than by large-scale yielding. The presence of reinforcement mitigates this effect by distributing deformations. Second, this equation is based on 198 tests on concrete slabs that were reinforced and included a large set of data on slabs thinner than 10 to 12 in. When considered together, these two points present an obvious challenge in extrapolating the equation to the anchorages tested, which have significantly deeper embedment and are also dissimilar physically as compared to a concrete column on a slab.

 

[KootK]

So, what KootK was saying in his 2nd point makes some sense. I don't know enough to wholly agree with it.

My opinion was honed based on a recent read of the book below which provided the following two insights:

1) The high flexural compression stresses in slabs over column supports is what make it possible for us to use shear stress values up to 2X our one way values. In the past, that had always bothered me as it seemed as though shear ought to be just shear, whether one way or two.

2) You get a punching shear bump when your slab is prestressed appropriately. I'd always figured that was because the prestress in the slab provided beneficial compression and confinement to the punching shear joint. For the most part, that's not the case. What really happens is that prestressed slabs tend to be slender, that slenderness exacerbates flexural compression stresses and then, once again, it is those flexural compression stresses that improve the punching shear joint. Here, the flexural compression dwarfs the P/A compression.

my tendency is to thing of this punching failure almost as a strut and tie action between the concrete shear and the flexural reinforcement.

Me too. In the absence of shear reinforcing though, this leads to your tension webs needing to be represented by diagonal tension in the concrete. And both that diagonal tension and the diagonal concrete struts would benefit substantially from being in state of bi/triaxial compression. So that checks out intuitively.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[JIMEY]

Thanks for all the responses guys. Some really good discussion here. So it sounds to me like the discrepancy between the two methods is a result of different testing conditions. Punching shear was checked with at least a minimum amount of reinforcing in the slab, and a certain amount of flexural stress as well. Breakout was tested without these factors. In reality, I think we can assume that the actual strength will be somewhere between the two methods. So I suppose it might just be a judgement call.

One interesting thing to note is that I checked the punching shear formula in the "Plain Concrete" section of the standard (ie. unreinforced concrete) and it simply applies a 2/3 factor to the regular punching shear formulas. So I guess this gives us an idea of the increase in strength that flexural steel provides.

 

[Toby43]

KootK's point regarding flexural stress would, to me present the most significant difference between the behavior of the two mechanisms. If you look at the work by Kotsovos and Pavlovic, which was coined as the "Compressive Force-Path Method" (CFPM) you will see the heavy reliance on the tri-axial stress state of the concrete in compression. Indeed the CFPM points to shear been pre-dominantly carried by the compressed (uncracked) zone of concrete at a cross-section due to its higher shear stiffness in relation to the cracked zone.

Toby