Shear Punching for Slab to Column Joint Vintage 1980 1

Wrong forum. I suggest you use the red flag to "report" yourself to the admin and ask them to move this to the Structural Engineering General Discussion. You'll get much more focused responses from structural engineers there. Not too many of pay attention to this one.

Why are you posting these questions here? This is a forum for discussing significant 'engineering failures', usually ones involving which loss of life or at least a large amount of physical damage. If you're looking for information dealing with a problem that YOU have with a structure, you may be better served by posting these questions in one of the dedicated structural forums.

John R. Baker, P.E. (ret)
Irvine, CA
Siemens PLM:

The secret of life is not finding someone to live with
It's finding someone you can't live without

New to the site, sorry for the confusion.

Looks like you now have the thread in the appropriate forum.

Yes, there have been changes to punching shear provisions, and there are many buildings with issues in this regard.

Yes, there have been changes to the rules but in most cases they should not be that significant.

No-one can answer you questions. You would have to do the checks yourself using older versions of ACI code.

If the wall is acting as a bearing wall, and has sufficient capacity, why wouldn't you count it?

(4) Could the designer have used the presence of the masonry walls to justify this connection as a one-way slab?

Click to expand...

Seems insane that the original designer would call a wall out as non-loadbearing on plan but rely on it for slab design. I wouldn't consider the wall load bearing unless you feel like checking that everything both above and below it are also working.

----------------------------------------------------------------------

Why yes, I do in fact have no idea what I'm talking about

(1) I would say this is a definite possibility. This construction date is right around the time of the Harbour Cay collapse (during construction) and the eventual collapse of Champlain Towers South. Both had concern with punching shear, as do "all concrete buildings." I don't mean that in a derogatory sense, I mean it's a major design consideration and it has to drive construction if the slabs are thin (i.e. slabs are too thin and deflection checks are thus required, that seems to be a common element). Your slabs seem to be a bit thicker.

While these are cast-in-place versus post-tensioned, I'd encourage you to peruse at least the NIST report on Harbour Cay for some context.


You also need a code book from that timeframe as well, in addition to the relevant building code (Broward has numerous old SFBC codes if you poke around, that's where I got my copies). The file names "look like this" I can't get to a main folder where they are all are listed now. Shrug.

(2) YES.

(3) I can't give you a history on post-tensioned slabs, or slab-column joints, or edge-slab column joints, but when I first heard of PT slabs, let's say 1990, they were more popular in Minnesota and Texas, so I'd look into your original designer and see what the story is there, on the FBPE site and poke around a few other states. I haven't done any PT work, and if you aren't familiar with it, well, it sounds like you're trying to recertify something you may not be familiar with. Further, the corner columns are probably something that was done without a lot of design/testing guidance and then it was noticed to be a problem and research was done in later codes, you'll be able to determine that via researching the codes and following their evolution.

If you haven't looked at Champlain Towers South design drawings, I would encourage you to look at those, again, for context. It is unlikely that a non-bearing masonry wall can do anything but transfer the problem to a lower floor, provided it has zero clearance between the bottom of the slab and the top of the masonry wall, which seems unlikely and something you'd probably be able to see by going to the site. What I mean about the CTS drawings is there are shear walls but there aren't many and they are rather awkwardly placed.

Regards,
Brian

Brian
Thanks for the very insightful comments. The Cay Harbor report is a good relative to the code and construction capabilities for analysis at that time. I'm familiar with ICES-Strudl code. Even though the ACI 318-77 code they used just generally points to looking at shear near a concentrated load (column), the CH report diagrams seem to show some variation in the shear stress along the planes of the critical section due to applying the bending moments. The latest version of ACI gives very detailed pictures and guidance on the shear punching calculations. Probably depending on the skills of the engineers at the time, the moments may have been considered in some manner.

The shear punching repair business does seem to be going on out there based on a number of interesting repair techniques i've seen on the web. Unfortunately, they aren't code worthy as yet.
Tom

Tom -

This isn't exactly what I was hoping to share with you, but it's thematically "there". I've seen some reinforcement schemes involving supplemental concrete and I always wonder how development length is being achieved. This article involves a steel collar, which at first glance at least doesn't have that issue. I suppose the other potential refuge is that the concrete may be stronger than specified, but rebar crowding was such a trend at the time, and over-reinforcing columns.


There is a lot of research on punching shear, and a lot of code changes over the years, although the side of it I've researched is the cast in place, not the post-tensioned.

Another approach -

I think I found a ACI 318-1963 code online, perhaps archive.org. Not that it's that useful for a 1980 project. Not sure it even mentions PT.

Regards,
Brian

Hey Brian,
Some good ideas at those links, thank-you. I will take a look at them. The worse case is an 8x36 inch edge column with unbalanced moments of 78 ft-kips about both the weak and strong axes and axial load of 81.7 kips. It will be interesting to see if the steel plate would be sufficient for the loads without being so heavy that its a problem to install. But it would sure beat the look of a slab of concrete on the ceiling.
Tom

Yeah the "add a drop panel" always looked iffy to me, not that there are good options, given the relative lack of research and working to fix an existing deficiency. Usually you're dealing with a slab that's too thin to omit the deflection checks, and that's where problems start - the slab is too thin, so depth of rebar is critical, and then punching shear wasn't checked correctly in the initial design. (or it's never checked, like Harbour Cay). How do you fix that?

To me, I really wonder about shoring because you need a fair area to add the concrete for the drop panel and you're chopping away at the column concrete to thread bars through. If you aren't shoring, you're asking the vertical bars to take 100% of the load in place, and the bars are unsupported for what, 8"? I suppose buckling or yielding isn't that likely, but development length of the column vertical reinforcement? Splices?

Plus development length (of the added bars) and all that epoxy being done properly has always seemed pretty optimistic to me. And can you get a U shaped bar through the column anyway? And I'm not clear the approach is trying to use shear friction to bypass the deficient design, or what the exact design approach is. Is it supposed to be some kind of load-sharing between the original and the added? Or is the drop panel supposed to take it all? Also, if they overdid the reinforcing in the column that would complicate getting the bars through in the first place. And then there's the question about the floor concrete strength versus the column concrete strength.... I've noticed that a few times as well. I think the column concrete is supposed to be no more than 40% stronger than the floor concrete but it seems that little fine print information gets missed by some. (ACI 318-63 talks about it in section 917. CTS had that issue on floors 1-4)

I've not designed or sealed a retrofit drop panel (partly because I've never gotten into a spot where I needed one, and multistory concrete repair isn't really a main component of my work/experience either), but I did have a project where some ding-dong used EnerCalc for footing design and (because they didn't read the manual) got the bar location wrong so it had punching shear issues "from below". You really can't add a drop panel above a footing because it conflicts with the parking in the basement and naturally they'd poured the slab by that time as well. Why anyone would bother with EnerCalc for a footing when they can just pull them from CRSI, well, that's another discussion, but not using CRSI created the problem when they used EnerCalc wrong.

As a reference detail for the "add a drop panel", I've attached something I found (for Clarity, it's from Champlain Towers South, work that was never performed, or, perhaps, never finalized and sealed). It seems likely something similar was done (and successfully, I suppose) on Dolphin Towers. The original design of CTS in at least a few areas had too much column reinforcing (several beyond 4% others beyond 8%). (Just like Harbour Cay). I would check for that here as well.


(Note there's no "existing hatch" on the column)

Oh, also - "REEXAMINATION OF PUNCHING SHEAR STRENGTH AND DEFORMATION CAPACITY OF CORNER SLAB-COLUMN CONNECTION" 2017

Regards,
Brian

It has been pretty common practice to reduce perimeter column stiffnesses such that less unbalanced moment is drawn to these joints. Is that something that you are availing yourself of? Ostensibly, this is a half assed way to account for the flexural cracking at columns with large moments and low axial loads relative to interior conditions.

Another historically common practice -- albeit a very dubious one -- has been to only consider the joint moments uniaxially, along with whatever design strip is being examined

Kootk,

if you are going to do the stiffness reduction logic, the designer has to justify the reduction on the actual effects for each column.

We see a lot of designers making blanket reductions in stiffness (as far as 0). This is not logical.

The absolute minimum condition that has to be accounted for is the loading and resulting moment that will crack a column. The column slab connection will experience that combined M and N, no matter how much the column stiffness reduces after that. And the decision on cracking loading should be based on the upper estimate of the tensile strength of the concrete, not the average or lower estimate.

rapt said:

if you are going to do the stiffness reduction logic, the designer has to justify the reduction on the actual effects for each column.

Click to expand...

You're preaching to the choir on this rapt. It wasn't my intention tell tomzeb what I think should be done. I was telling him what I know has been done. That is often necessary if one wishes to understand the sources of apparent deficiencies with historical designs.

In North America to this day, there are marquee firms cranking out millions of square feet of floor slab each year based on a blanket assumption of 50% reduction in stiffness at edge columns. And 100% as you say. I've seen 100% reduction, in person, and from one of the "big boys".

I feel that column stiffness evaluation would be a nice feature for FEM vendors to provide. Understandably, however, FEM vendors tend to be slow to add features that would make their customers' designs more costly. I was seeing inflection point bracing options in beam design packages long after it was a settled debate.

I often hear the "Field of Dreams" mantra of rebar and bending moment. "Reinforce for the moments and they will come!". And, Conversely, "don't reinforce for the moments and they'll stay away!". There is some measure of truth to it but it lacks rigor and nuance in ways similar to the issues that you've raised.

Interesting comments Gentlemen - Thank you. I am trying to understand how the building design fails the shear punching limits based on recent analyses. I think the link on "corner slab/column reexamination" that Brian provided presents a good history of the shear punching requirements and changes that have occurred since 1910. It notes that the term "shear punching" was used for the first time in 1913, but it was unclear at that time how to compute the demand. I think the report answers a number of my question as explained below:
Based on the corner reexamination report, in the 1920's the code defined vertical shear and diagonal tension to be looked for shear modes of failure with limit of 0.1f'c. After iterations on the shape of the failure zone and more limit changes, in 1956 the code came up with shear demand = V/(bo j d) and used a critical section at distance d from the column. In 1963 revision, the Ultimate Strength Design (USD) method was introduced, and was used along with the Working Stress Design (WSD) and provided the methods to address the diagonal tension failure mode. They changed to a distance d/2 from the column for the critical section and added v = V0/A +- M (c+d/2) / Jc. This was called the Eccentric Shear Stress (ESS) model which was revolutionary at the time (probably). It looks pretty close to what we have today. The equation was adjusted slightly in 1971 with the term gammaV and addressed un-balanced moment. In addition, in 1971 the WSD for shear punching was eliminated. In 1974 the slab width was redefined at 1.5 x slab thickness for the sides of the critical section (must have changed back to d/2 in late 80's?) and an adjustment to v_c was made based on larger rectangular column shapes.
From all this, it does seem that the ESS model was in code in 1963 and slightly improved in 1971. So shear and moments in at least the analysis direction should have been looked at for construction in the late 1970's, early 80's. Possibly, the moment transverse to the analysis direction could have been left out of an analysis in the 70's and if large enough could be a root cause to fail limits in a contemporary analysis. The corner reexamination report noted that the question of using biaxial moment contributions simultaneously came up in 2005; and in 2008,2010 the code explicitly suggested to consider them simultaneously. But, since the peak stress was at a point, they allowed for a 15% reduction.

Kootk,

Just wanted to get the message out there that the amount of reduction has to be justified. I see too many 0, 5%, 10% suggestions based on "but everyone reduces the column stiffness" or "Joe Blogs (or KootK) said I can reduce it" without considering that it is physically impossible to get a column stiffness anywhere near that low from cracking even without any axial load, unless you leave out the reinforcement completely.

Once upon a time there was a bulk stiffness reduction that was allowed to be used in modelling for lateral forces as I recall, they were bulk percentages applied to the gross moment of inertia. Without saying it's appropriate or not (I believe the intent was for it to be applied to the lateral force analysis as an acceptable simplification, but not intended for detailed member checks), but that's possibly where the column stiffness reduction you've seen in calculations, Rapt.

Good analysis Tomzeb, I haven't been through that document in detail, since it's a side research project, I've found a lot of the references but haven't gone through them, so I appreciate it, it should really be an article in ACI journal....

I would mention if you're trying to prove it works, somehow, you'd probably be applying the loads from the time of design against the code at the time of design, (Florida dinged a guy for doing an analysis per current codes some time ago in a defect report,
I've also seen some big gaps in code adoption in Florida, so if they don't adopt a current code for say, 20 years (cough Miami-Dade cough 1966-1988 - the "enforced" code might refer to an older ACI 318 than you'd think, and then you're left wondering what code the engineer applied or CLAIMED to have applied, similar to that what, 10 year period where California was still under the 1997 UBC which referenced the 1994 NDS? Circa 2005? And from the looks of it, New York City has something similar going on.... ( that's from 2014, so I guess as of 2001 New York City still referenced the 1991 NDS...) One would like to think they'd use the current code, but it's really murky if they did.

Regards,
Brian