App. D

[Lion06]

Here is a very generic question on ACI App. D. I have a very good grasp of this material since I wrote a very detailed spreadsheet for it, but I still question the validity of it. Is the only reason that this is required is because of the high stress concentrations associated with headed anchors? If the anchors were deformed and embedments met development length requirements would App. D even exist?
I'm just thinking about a #6 rebar with a development length of 15" (with no reductions).
Let's say for the attached sketch you have a #6 bar hooked with 15" embedment - you have nothing else to worry about, no breakout checks, no side face blowout checks, no pull-out checks, etc. You also get a capacity of 47.5k of tension (for the two anchors in tension).

Now replace those #6 rebar with 3/4" diameter headed anchors with 15" embedment and design per App. D. Now for the same 2 anchors in tension, you get a capacity of 15.3k and it's controlled by concrete breakout (assuming cracked concrete, but using supplementary steel). That's a HUGE difference!

Is the only reason because of the stress concentrations at the head of the anchor?

 

[Lion06]

Here is the attachment.

If the pier shrinks to 18" and the edge distance of the anchors drops to 2.5" then the headed anchor capacity drops down to 6K, from the 47.5k of the rebar. That's crazy.

 

[frv]

I think it has to do with confinement. The difference is reinforced vs. unreinforced concrete. If you had ties around your studs, I think you can treat it like an ordinary development length (and lap splice) problem, not per Appendix D. I'd be interested to see if others share this view.

 

[Lion06]

I believe you can lap headed anchors with reinforcement regardless of confinement steel.
I don't believe that confinement is the issue since the basic hook development length doesn't require ties for confinement - you get a reduction if they are present, but they're not required.

 

[JedClampett]

StructuralEIT, you're right. There's many issues that we ignore for standard reinforcing that are critical for anchoring to concrete, like edge distances, spacing, etc. I can't explain it.
It might have something to do with the fact that reinforcing is developed with much longer lengths than practical with anchor bolts, but logically that can't be the whole reason.
Any time I try to get my arms around Appendix D, I get a massive headache. I follow it, the best I can, but I've never understood the necessity/complexity for it.

 

[StructuralEngGuy]

You bring up a very good point...one that I've often pondered as well. I also have no explanation for it. Either App D is very conservative for headed anchor design or we are all very unconservative when designing standard reinforcement. But a spreadsheet or anchor design program is a definitely a must for anchorage calcs.

 

[vandede427]

I hate Appendix D.

 

[JStephen]

I think perhaps multiple factors. One is the difference in the bar versus bolt- with all that load going at one point or being distributed along the bolt. One thing is a lack of analysis. Just because there's not a requirement for breakout check doesn't mean there shouldn't be.

You might also notice that in your original example, if you use a hooked bolt with the same geometery as the hooked rebar, then capacity of the bolt will be less than the headed bolt.

 

[haynewp]

I had considered that maybe the concrete failure with rebar is a bond failure (or crushing directly above the hook then bond failure) versus a cone failure with headed studs, so the cone interactions with edges and with each other do not apply in the typical case with rebar. But that doesn't seem to add up. There must be some type of cone like interaction as you go along the length of each rebar/deformed bar.

 

[Lion06]

I'm not concerned at all with a hooked bolt, I would NEVER spec one. I'm concerned with the difference in behavior and the drastically different capacities that can be generated.

 

[JStephen]

Maybe I'm missing something here, but it looks like you've got a block of concrete with two bars sticking out of it, you figure the pullout capacity of those bars, but then in the plane just behind them, you have an unreinforced concrete cross-section that is loaded in tension? Is that correct?

 

[Lion06]

Sort of. I'm not concerned with the member itself, only getting the tension load into the member. Either way, what does that have to do with your first post?

 

[msquared48]

Kinda why I generally use a bearing plate at the end of the stud - 3 to 4" square - to increase the pullout without having to rely on any bond development. I even do this for rebar, welding the plate to the end of the bar.

Mike McCann
MMC Engineering

 

[kslee1000]

Ok, guys, let me try (without directly refer to code though).

1. If you place a #6 hooked bar as a 3/4" headed stud, the investigation method is identical, no differential treatment - the breakout, if it controls stud design, it controls the #6 bar as well. The difference of hooked bar in pedestal and in beam-column joint is due to the confining effect of the rigid joint (small rotation), and nature of the load delivery mechanism - progressively passing from point to point (not concentrated at a single load point as anchorahe does, which causes high local stress concentration as many have pointed out).
Also note, bent bars, or anchor bolt, is poor anchorge device, which prone to pull-out failure.

2. Again, without refer to code. For headed stud, I would utilize the reinforcing dowels to resist uplift. However, the dowels shall have adequate development length above the concrete failure cone. The cone radials 45 degrees from the top of the head of the stud, thus, usually resulted in longer stud, which is conservative. The upper ties shall be kept in a closer spacing (1/3 to 1/2 of the normal spacing) to prevent the dowels from poping out. Also, make sure there is adequate space in between the anchorage and the dowel to allow for bond to develop for both (comply with ACI minimum spacing requirement - would result in bigger pedestal size).

3. I believe the original design concept (anchorages in concrete subject to shear & tension) and analytical method was developed by NRC (Nuclear Regulatory Commitee) for nuclear power facilities, then adopted industrial wide with very conservative safety factor (4 ?), that meant to use in conjunction with ASD. The ACI had been tardy for many years, not until ???, finally, it started to play catch-up games. Well, you know how good it is. Good luck, fellows.

 

[sandman21]

I don't see any real difference, I will give you the fact that load is not going develop along the bolt the same as a rebar. However the ribs on a #6 bar are not that deep. Just to throw a wrench in the operation wait until you see the requirements for when rebar can be used to resist uplift forces in 318-08, something like ½ or ¼ don’t have it in front of me. It’s not based on whether the load can be developed into the tie, 12” or 6” can’t get too many ties in there. Once the load is in the concrete it is in the concrete. App. D is CCD approach, so test are run on many different anchor bolts, and they came up with these equation. Which is why App. D applies to anchors under 2” dia and under 25” embed.

Also companies do make rebar that is thread at the end, which kicks you out of App. D. Bet they like App. D

 

[JStephen]

The original question was why the strength of the rebar calculates very high when the strength of anchor bolts in an equivalent installation calculates very low. But if you have an unreinforced cross section beyond the rebar, then the strength of that section also needs to be considered. If I understand the workings of App. D, that failure is included in the breakout check. Neglecting that failure mode, you're asking "Why does rebar calculate very strong when I neglect certain failure modes while anchors calculate very weak when those modes are considered?" And the issue then is a lack of analysis, not necessarily a fault in the codes. Or if you put additional reinforcing in to eliminate that issue, you no longer have equivalent installations.

Consider the attached figure. You have an upper block of concrete with anchor bolts in the top of it, attached to the lower slab with rebar. If those anchor bolts extended all the way through the upper block into the slab, then App. D gives you equations to check breakout of the bolts from the slab. However, if the bolts terminate where shown, and rebar is extended into the slab as shown instead, there is no requirement in the standard to check breakout. Does that mean that the rebar installation is stronger? Or that you've just missed a failure mode?

Please understand that I haven't made a detailed comparison of App. D vs the rebar anchorage requirements; I'm just trying to explain some possibilities of why things are as different as they are in the results.

 

[CTW]

I think haynewp is on the right track.

The development length of rebar is determined by the bond strength which is affected by adhesion, friction and bearing of the deformations against the concrete. For the rebar to pullout, the concrete in front of each deformation must crush or the keys of concrete between the bar deformations must shear off.

With a headed anchor, the proximity of the failure cone to the edges and the depth of embedment affect the capacity. There's no consideration for adhesion and friction along the length of embedment of an anchor. Plus the only bearing is provided by the head on the anchor whereas on rebar, numerous deformations provide the bearing along the length of embedment.

sandman21 brings up an interesting subject about threaded rebar. Why go through the hassle of App D, struggle with obtaining capacity, and utilizing rebar lapped to the anchor rod to develop the required resistance when you can use a specified length of rebar threaded on one end to serve as the anchor and not bother with App D?

 

[Lion06]

jstephen-

App. D does not address the actual member itself, only the anchorage. The issue is not a lack of analysis. With rebar development, there is no "concrete breakout, pullout, etc", it is a bond failure or splitting failure (depending on cover, spacing, etc.). The point I was trying to make is that if you have two identical piers (as I detailed in a previous post) and one has #6 bars, the other has 3/4" diameter headed anchors. They both have identical spacings, edge distances, embedments, etc. the #6 give dramatically higher capacity. I am simply asking if this is a result of the deformations on the bar developing the rebar along its length as opposed to the headed anchor being concentrated at a single point. I don't care about, and have no interest in, what goes on beyond developing the load into the member. In your sketch, the only way that you don't have to worry about concrete breakout is if the rebar have enough cross-sectional area to develop the load AND the rebar extends the required development length above the potential failur plane.


kslee-

The investigation method is not the same. The headed stud requires App. D, while the rebar requires only the necessary development length. There is no "breakout" capacity of a hooked #6 bar, there is only the development length (which is limited by bond or splitting).

 

[kslee1000]

StrEIT:

Did code explicitly calls for different method for stud and rebar? If it is the case, I don't feel it's correct. And that's the reason, though I didn't point out, why it stays in "APP D", because it is subjected to potential revisions.
I think something in the appendix is not rigid provision, but a preference/guide. Once the refine works done, it will be moved into the main code body.

 

[Lion06]

App. D applies to headed anchors only, not rebar. It's not a mistake in the standard. App. D is only recently introduced. I don't believe it will be refined to incorporate deformed bar.