Cast-in plates, concrete shear breakout dilemmas... 2

[Agent666]

Hi All

This is a tricky little scenario that we've been doing some circles around for a while trying to figure out a best practice practical approach.

It's related to concrete breakout capacity of anchor groups in shear, particularly for shear acting towards an edge (but equally applies to shear parallel to an edge by extension), and typically for a cast in plate which might support a beam with a shear tab type connection with multiple vertical rows of anchors (3, 4, 5 rows of welded shear studs typically).

Now ACI318-19 states the following, which all makes sense for the examples given in ACI318 which include only 2 rows of anchors to keep the concepts simple (referring to the three CASES noted in the second image).

The question is how do people practically handle this?

Or what is the correct approach/interpretation when you have multiple rows of anchors (>2) in the direction of the load?

Basically if welded studs to a common cast in plate, then ACI318 is saying CASE 2 applies and you take the breakout surface on the back anchors furthest from the edge. That's all good for 2 rows like shown and makes perfect sense for that type of arrangement. The dilemma we are having is for larger plates with correspondingly larger loads if you were to perhaps take all of the shear on the back anchors, then the anchors themselves might not have sufficient strength in the steel anchor checks. This seems fundamentally an incorrect way to think of it as the other anchors in their own right might have sufficient breakout strength closer to the free edge, but the load still needs to go through the same potential breakout surface as the back studs irrespective.

Some alternative approaches we've kicked around:-

1. Check breakout assuming all anchors share a proportion of the load (neglecting any close to the edge that might fail), so the closer studs are checked progressively with their own individual breakout surface but with the accumulated load from earlier studs going through the same concrete breakout surface, but the studs them selves just see the proportion of stud shear load (shear/number effective studs) at each effective breakout surface.

2. Just adopt CASE 1 type check, checking each row individually, neglecting any rows closest to the edge that do not work for the proportion of the load (shear/number effective studs). This method seems to neglect by the time you get to rear rows that the load from the front studs is in the concrete already. But this seems to be a direct application as ACI318 spells it out, albeit we often do not comply with the spacing 's' being greater than or equal to 'c_a1' edge distance for practical connections as noted for CASE 1 (or 2 for that matter).

3. Adopt a hybrid CASE 1/2 type check, but only check the first stud closest to the edge that works for it's proportion of load, rear studs by virtue of being further from the edge imply a larger capacity and not inferred a check is necessary with CASE 1 for example (not always the case depending on how the side edges influence the breakout surface). Keep neglecting those close to the edge until you arrive at an effective number of rows that work for their proportion of the load. However, 's' is usually not going to satisfy the greater than or equal to edge distance limit.

4. Use CASE2 as stated with all the load on two rear anchors (providing some monster sized anchors at the top of the plate). This seems a bit silly, but it is a direct application of CASE 2 as stated.

I note as well in most cases your stud spacing 's' will typically be less than 'c_a1' the minimum edge distance, which only satisfies CASE 3. Which for CASE 3 states it specifically does not apply to the case of common plates! Even if it did this check is telling you to put all of the shear on the front anchors which clearly will likely result in failure if you have any decent loads and seems extremely conservative and not reflectiuve of the expectation that the shear will be distributed between the anchors furthest from the edge in some manner!


Lets ignore the fact here that you may in fact mobilise some supplementary or anchor reinforcement that is developed/provided the further from the edge you might get, and any tension that might be present. Focus the discussion on shear concrete breakout mechanism checks in isolation as that is the root of the headscatching we've got going on here in coming up with a method to generally apply to the problems like this.

Thoughts and guidance appreciated!

I will say as well that I have looked through a few international standards (ACI/NZ) and guidance such as the PCI Design handbook and SCI P416 and it never quite touches on this aspect. There was some discussion I found in some fib bulletins (#216) which basically had the same logic as ACI but again was only talking about 2 rows of anchors parallel to an edge.

(Let me know if the above makes no sense at all and I'll sketch something up)

 

[Tomfh]

If the studs are welded to the plate it’s valid to assume the load is shared. It’s when it’s bolted through holes in the plate that the outside anchor can end up copping the brunt of the load.

 

[Agent666]

If the studs are welded to the plate it’s valid to assume the load is shared. It’s when it’s bolted through holes in the plate that the outside anchor can end up copping the brunt of the load.

Yeah thats my assumption that it is shared somehow. Apart from the implication in ACI that (as per the clause I quoted above and CASE2) that all of the shear shall be assumed to be resisted by the critical anchor row alone, which is the anchor studs furthest from the edge. Now maybe they are only talking about the concrete interface as described resists all of the shear, and not the steel anchors alone at this location as I'm potentially misinterpreting? I guess that's fundamentally the question I'm asking?

 

[Tomfh]

Your first example is correct in my opinion. If you're using cone 4 then V/2 per anchor. You can't count anchors inside your assumed breakout cone.

If you want to use V/8 per anchor then you need to use cone 1 as the limiting concrete condition. This is analogous to Case 1 in the official guide - you are spreading the load over all anchors and then checking the worst case concrete condition.

You could also use intermediate cases. E.g V/4 with cone 3 critical.

 

[Agent666]

You could also use intermediate cases. E.g V/4 with cone 3

I think only if the first two rows form the edge were not good for respectively 2*V*/8, and 2*V*/6 though.

Obviously once you start discounting rows at the edge due to failure, then this has implications for tension and combined actions (cannot intentionally fail concrete near edge if it's being relied on for tension breakout). Possibly the only way to get combined actions to play nice is to have the lower bound strength based on row closest to the edge limiting the capacity (V*/8 on all anchors, V*/4 on the bottom 2 anchors).

 

[Tomfh]

Case 2 suggests you needn't worry about the anchors closer to the edge, provided the deeper anchor(s) can carry V. So presumably it doesn't matter if the outer cone fails provided deeper cones and anchors are sufficient?

 

[Settingsun]

I didn't read everything so ignore if not relevant, but is the commentary halfway down pg. 262 relevant? R17.7.2.1. It gives a condition when you can share the load between anchors while still using the rear cone.

 

[Agent666]

Yeah that might possibly be the answer, if you're referring to this paragraph.

Will try track down the paper referred to and see if it sheds more light on those statements. Edit paper

 

[Agent666]

As a followup for anyone that follows, this is what we decided to adopt until more definitive advice comes along:-

1. check shear breakout perpendicular to the edge for all stud rows (i.e. with two columns and 5 rows, you check 5 breakout surfaces for a shear of V*/5 per breakout surface). Basically following logic of CASE 1 in ACI but extending to further rows from edge.

2. If it does not work (usually your bottom studs will be critical), and you are governed by the lowest stud row, provided that the anchor closest to the edge is not required to resist tension resulting from moment from a shear tab type connection in our case and you can accept some localised cracking, then neglect this set of studs from further calculations. Or you can just remove this stud row outright from the weld plate as not required for shear or tension.

3. Essentially this is a combination of a few of ACI's CASES, if anchor closest to edge doesn't work, neglect it, then reevaluate remaining 4 rows for breakout, with a shear of V*/4.

4. Rinse and repeat until you run out of stud rows if required.

5. Follow similar procedure for parallel to edge cases, on the assumption of two columns of studs check following 4 conditions:-

check left hand column of studs breaking out to left hand edge,​

check right hand column of studs breaking out to left hand edge,​

check left hand column of studs breaking out to right hand edge,​

check right hand column of studs breaking out to right hand edge,​

6. if you neglected a row of studs as part of the perpendicular to edge cases, then it is also neglected as part of the parallel case checks, because the assumption is in getting to that point this row closest to the edge has some concrete cracking that affects the parallel to edge breakout surface as well.

7. work out critical ratio for either perpendicular and parallel cases and use this to proceed with.​

I must say even after reviewing that paper I linked to I'm still at a bit of a loss as to how they got the s/c_a1,1<0.6 ratio recommendation as there is no such thing specifically noted in the paper unless it's inferred by some of the combinations of various factors. The paper just repeats the PCI Design Handbook criteria which follows a slightly different approach to the ACI approach.

But luckily, if you are really close to the bottom edge for a vertical load and you neglect a row, the actual edge distance and normal stud spacing typically mean you will actually satisfy this ratio. Some judgement defiantly required in assessing these things, and obviously for more serious loads you should consider some form of anchor anchor reinforcement so concrete breakout capacities can be effectively replaced with anchor reinforcement capacities as per ACI recommendations.

 

[Agent666]

Problem #2 - somewhat related, shear pryout checks

For concrete shear pryout, most codes treat it as being of the form ΦV_cp = Φ * k * N_cb.
In ACI318-19 the requirements are in CL17.7.3.

where k is a multiplier on the concrete tension breakout strength (typically 1 or 2 depending on the stud anchor effective depth).
N_cb is the concrete tension breakout strength.
Φ = strength reduction factor.​

So say you have a shear tab carrying vertical load like shown below, this creates a moment which is resisted by the top rows of studs (say my previous 2 columns of studs with multiple rows of studs), where top three rows resolve the tension. So we have a breakout strength in tension for these top 3 rows acting as a group, but now the shear is applied to all the stud rows.

Basically the question is how are people making the comparison between the applied shear and the shear pryout capacity in this type of scenario?

1. Are people comparing 3/8's of the shear to the N_cb value for the top 3 rows?

2. Or are you required to calculate a different N_cb for the assumption of all anchors being in tension?

3. Or are you comparing it to a different individual N_cb value for a critical single anchor in breakout, like the one with the lowest breakout tributary area?

4. Or other?.... cross your fingers/bury your head in the sand and hope for the best?​

The more I thought about it, because it is a shear mechanism, the tension breakout is not the same as for the tension loads, but what to consider?

Anyone have any guidance they follow, have found on this?

While we have all these good code equations to cover everything, once you start looking at it in detail you realise there are a lot of gaps in what to do when you start looking at what "if type" arrangements when generating a robust methodology for dealing with specific/certain cases!

 

[Agent666]

Well to answer my own question for anyone that follows regarding the pryout question:-

It seems Eurocode 2 and common adopted European practices has some specific advice around pryout, where Eurocode states that you simply evaluate based on an individual anchors tributary tension breakout surface for evaluating pryout. (So my option 3 from above)
3. Or are you comparing it to a different individual N_cb value for a critical single anchor in breakout, like the one with the lowest breakout tributary area?

Hopefully it's of some help to someone in the future.