Integrity Reinforcement in Transfer Slabs 9

[KutEng]

Just want to bring back an old post by KootK that got no real answer:

https://www.eng-tips.com/viewthread.cfm?qid=388712

Integrity reo has just been introduced into our Australian concrete standards so I don't think many of us will have this figured out yet.

This is a tough one because transfer slabs really benefit from integrity reo since a failure in a transfer slab can be catastrophic, however trying to get adequate integrity reo over your columns seems like a huge ask.

It's almost as if this clause is pushing us away from designing flat plate transfer slabs by making it so unfeasible that no one is willing to use them anymore. Generally, on some of our transfer decks we transfer columns at ground floor that continue up to 15 stories (around 7000kn of load). It would be almost impossible to fit enough reo over your column in these cases.

Would love to hear from some people who have managed to satisfy this clause in a transfer slab, or if its generally left out of transfers.

 

[Tomfh]

Yes, If adding robustness is the underlying rationale for adding bottom integrity steel then it makes sense to consider if PT is an adequate substitute.

 

[rapt]

StructEng23

Answering your questions in order

- Commentary some time next year, hopefully. Along with at least one more amendment.

- Would have been wonderful to have 9 years. But someone else would have had to do the 2009 commentary to free us up to do it. As I said, I am not able to comment on the cause of errors, 2009, Commentary or 2018, but they cost us a lot of time and frustration as well, finding and fixing them.

- I explained the logic behind the clause above in about June/July and the difference between it and CSA/ACI. It is not to provide overall Robustness. Simply a more ductile punching shear failure by providing Bottom Reinforcement at the column. Overall Structural Robustness still needs to be checked and provided in accordance with 2.1.3 and the Building Code of Australia. AS3600 does not give specific provisions on how to provide Robustness, just the motherhood statement in 2.1.3. Designers have to resolve the Robustness issue themselves and PT tendons can be used for that. Unfortunately, the building code also needs a rethink on its Deemed To Comply Robustness provisions.

- The clause is very specific on what is required. It is not CSA23 or ACI218, so you cannot read those clauses to interpret it.

- No, it was not added for disaster situations as I have implied several times in earlier comments. It is simply to provide a more ductile punching shear failure mode rather than the brittle failure you get with no bottom reinforcement through the column core. Unfortunately the default detailing rules in AS3600 before 2018 basically allowed a designer to provide no continuous reinforcement through the bottom of a slab/beam at an internal column and for PT slabs at an edge column.

The clause specifically says N* is the "column reaction from the floor at the Ultimate Limit State"! 1.2G + 1.5Q is the Ultimate Limit State for gravity load and always has to be checked for Flexure, one way shear and Punching Shear. Why is interpretation required? If it was required only for Earthquake or Fire, it would only be in those code sections or would have been limited to those situations.

Agreed Earthquake and Fire are other ultimate limit states but their loading tends to be less, though detailing is as or more important. Earthquake design would require bottom reinforcement there anyway, in any RC or PT frame structure. How many PT designers are adding bottom reinforcement at columns for earthquake (and top reinforcement at mid span)? We specifically made Earthquake more prominent in 2018 moving it from an Appendix to a main section to try to get designers to actually apply it.

 

[rapt]

Re Ingenuity's Northbridge picture, the prestress in that case was unbonded. Do not start thinking that bonded PT will act in the same way.

There are special provisions for prestress in special moment frames in ACI318. The main one for us is that bonded tendons must be unbonded over plastic hinge regions and the strain limited to .01 under the design displacement as the fully bonded tendons would fracture due to their relatively low ductility.

Fully Bonded PT will act very differently to unbonded PT in these situations.

 

[aaronPTeng]

RAPT I know you disagree with my last statements. However I have one for you.

If we are after a ductile failure shouldn't N* be higher than 1.2 DL + 1.5LL? As our punching shear around our column should already be designed for this load combination. In testing I have seen (And an old building that failed many years ago that was tested) it proved to be much stronger than this. So if we are getting a failure at a Higher load then N*, then how is reinforcement in the bottom at N* going to provide ductility?. If it has a punching shear capacity of 2xN* or 3xN* (columns may be for a tower and nice and big) then reo in the bottom at N* aint going to help much?

On the logic stated above, shouldn't the design load for this reo match the load at which punching failure will occur based on the column size and slab thickness. Not the actual 1.2 DL + 1.5 LL. And as we all know this would then be near impossible to feed through the columns.

Your above comments implies you and the committee have spent so much time on this topic to think about it, which is good. Would be great for you to point out to us the research/ testing that verifies this and formed part of the conclusion.


"Structural Engineering is the Art of moulding materials we do not wholly understand into shapes we cannot precisely analyse, so as to withstand forces we cannot really assess, in such a way that the community at large has no reason to suspect the extent of our ignorance." Dr. Dykes, 1976

 

[rapt]

AaronPTeng,

Punching shear is generally a non-ductile failure. It reaches capacity and collapses without warning. Providing this amount of bottom reinforcement is supposed to provide sufficient ductility to make the failure more ductile, providing sufficient warning against collapse. It does not increase the punching capacity significantly, simply reduces the likelihood of sudden collapse after punchiung failure.

AS3600 has to assume that a designer can calculate N* correctly for his column. That is the load that it is designed for so is the load it needs to be ductile for. If the designer is incapable of calculating N* correctly, then that person should read my comment about finding another profession. A design code cannot completely protect us against incompetence or all of our buildings would be mass concrete. Read AS3600 clause 1.1.1 Note 2.

It is not overload we are attempting to save, though this will help, it is sudden collapse. ACI352 refers to papers suggesting we should design it for 2 * G. We decided 1.2G + 1.5Q would be adequate and may be lower than 2 * G in domestic/parking structures.

In the Darwin problem buildings, my understanding is that some of them have punching capacity (unfactored) Nu of half or less of the applied N*. So either the designer did not check punching shear or could not calculate N*. You cannot blame AS3600 for either possibility. None actually collapsed though at least one had obviously failed. They were RC slabs so presumably had some bottom reinforcement through the column cores, even if they were lapped bars rather than continuous. Without being involved directly we cannot know how much.

I was fist told about the German Studrail tests that showed this in the early 1990's but do not have copies of the results and no longer have contact with the researchers who showed the effects (If Dr Nadia C???? is lurking she can comment). My understanding was that they showed punching capacity little changed either way , but where the rails were continuous through the column core, collapse load was about 2.5 times the collapse load where the rails for the studs stopped at or just inside the column face.

ACI352.1R-11 section 6.3 covers this (and has a diagram amazingly similar to Kootk's first on this topic!!), and specifically says that only bottom reinforcement can provide this and precludes the use of draped tendons to replace the bottom reinforcement. ACI318 and CSA 23 obviously did not listen to them. We did not completely listen either as we did not require this reinforcement to be continuous in the bottom or you would be adding 500t to your building, not 100t. As I said earlier, I would ignore bonded tendons at the face of the column in the plastic hinge zone anyway due to their relatively low ductility under the strains we are talking about.

 

[Settingsun]

Rapt, are you saying that the reinforcement area based on N* provides the desired ductility in cases where punching occurs at significantly greater column load than N*?

I echo the request for the research that this is based on to be named here in advance of the commentary. We're supposed to self-educate but it sounds as though AS is different from other codes. Either AS is from other research than we're likely to find without a pointer, or the AS committee is using the same experiments and drawing different conclusions. Either way, self-education doesn't seem feasible from the available information.

From the outside, I don't understand why the commentary is significant additional work. The science is presumably settled when the code is published and, I would expect, written up by the people responsible for each change.

 

[rapt]

steveh49,

Not sure how much more I can say. It provides a much more ductile collapse mechanism. It does not increase Vu significantly. It is not going to save you if you have grossly under-designed punching shear which would be the case if V* >> phi Vu.. I think I said or implied all of that in my last post!

You obviously did not read my last paragraph!

After this post, I will never again be commenting on anything to do with AS3600 development. I have gone outside the rules in trying to help. I heave learned my lesson. Never again!

 

[Settingsun]

We're asking about the case of V*<<phi.Vu.

 

[Toby43]

Personally I don't believe ductility is the best terminology to explain the integrity reo requirements.
i.e. 1.2G+1.5Q=950kN
Predicted Punching Design Capacity = 1000kN
Predicted Post Punching Design Capacity = 1000kN
Actual Punching Failure = 1200kN (Over-strength)
Actual Post punching Capacity = 1100kN
Result = Pancaking with no ductility
If ductility is the intent, then "Capacity Based Design" principles would need to be applied to ensure Post punching had a Design Capacity greater than the Punching Over-strength.
As stated in my previous post, I feel that integrity reinforcement is a cheap safety net to account for the inherent uncertainties in what we do. As for its practicality in major transfer slabs, I can't comment as I have no 1st hand experience in high-rise construction.

Regards
Toby

 

[Trenno]

After this post, I will never again be commenting on anything to do with AS3600 development. I have gone outside the rules in trying to help. I heave learned my lesson. Never again!

Out of curiosity, what are the proper channels to contact the AS3600 committee?

As engineers we are often required to defend our designs and decisions - I don't think it's unreasonable for engineers to question and interrogate the logic behind code clauses and/or the research it's based upon. But the question remains, without a coincident code commentary, who/what will shed the light?

Engineering judgement comes into play here, but it's obviously to the benefit of everyone if these decisions are well informed.

 

[KootK]

After this post, I will never again be commenting on anything to do with AS3600 development. I have gone outside the rules in trying to help. I heave learned my lesson. Never again!

I just want to go on record as saying that I do not relish this outcome. In my opinion, the value of your contributions to content in this area vastly outweigh the emotional cost of any abrasiveness endured in the production of that content. That said, if you feel that you've been breaking the rules in this space, I'll accept your decision on that basis alone.

 

[rapt]

Toby,

Ductile Punching shear failure is definitely the correct term for the slab bottom reinforcement requirements, as the opposite to the normal brittle punching shear failure mechanism, no matter what your personal feelings. We are not looking at overall structure robustness for a lost support, just trying to avoid sudden brittle collapse at punching shear failure, essentially stopping us from losing a support, or at least delaying it.

So your column in question has phi Vu = 1000. But assuming material over capacities, its actual capacity = 1200KN.

Applied Load = 950KN.

Why does it fail in punching if the applied shear force is only 80% of the capacity, phi Vu = 1.25 V*?

If this does fail,
- Why does it fail? Is there another cause increasing the shear load or reducing the capacity that I do not know about?
- What is the slab load transferring to the column at failure, 950KN, or something much higher, > 1200KN presumably to cause failure? If much higher, why is the loading 25% or more greater than the maximum design load?

If the actual slab loading transferring to the column is still 950KN or less. Then why is it going to "pancake" if the collapse capacity = 1100KN?

The bottom reinforcement does not have to substitute for the column head capacity, it has to transfer the slab load to the column. The maximum value of this should be the ULS factored gravity load. Under earthquake it should be less than that. Under fire loading, it should be less than 70% of the ULS gravity load, and the reinforcement will be reduced to about 70% of the normal ultimate strength due to temperature increase as long as the cover is adequate, so still ok.

The only thing that needs clarification in the rule is distribution to faces. ACI's rule does this the way I suggested it should be done, based on the force coming onto each face. I will get that clarified, either in the code or the commentary. Otherwise there is nothing wrong with the rule as it is!

I assume all of this also applies to steveh49 and aaronPTeng's questions also.

 

[Tomfh]

RAPT, please don’t go.

 

[Toby43]

All i was trying to say was that in this case, the ductile mechanism would need to be reliably stronger than the brittle mechanism (owing to the fact that it is a secondary mechanism that only comes into play post brittle punching)

 

[Settingsun]

I think the question is simply whether this reinforcement is intended as a foolproof way of achieving a collapse load greater than N* in the event that the initial punching capacity was too low for some reason (incorrect design, construction error etc).

Then the question moves to whether the reinforcement (sized based on N* rather than Vu) is effective if slabs are correctly designed and built with punching capacity 50% or more in excess of N*.

 

[jayrod12]

Maybe this is obvious, and you guys are going on just for the sake of going on, but it feels like most of you think the intent of the integrity reinforcement is to improve punching shear capacity. But rapt is saying, and I agree with this, it has negligible effect on the punching shear capacity, but rather the failure mechanism. The integrity reinforcement is intended to improve the ductility of the punching shear failure. Yes the slab still fails in punching shear, yes it still requires major repairs and is likely started via a brittle initial failure. But the integrity reinforcement is intended to provide SOME ductility to avoid a total brittle collapse.

 

[aaronPTeng]

All

I think we all agree there is something that needs to be done for ductility. Also agree with others engineering discussion is based on numbers and not emotion. If there is no trail to the roots of a solution then its always going to be hard to ask people to blindly follow. Also this is not just a debate on the current AS 3600. Its a code that is constantly under development. There are always additional considerations which are fair and reasonable to discuss. Otherwise we would be sitting around a camp fire hitting each other with sticks in a cave.

As per my previous question and Steveh49, Toby 43 comments

You wont get ductility if N* is much less than Vuc. Other parts of the code relate ductility to section parameters and not load.

I look forward to seeing some research and testing come out around this topic and hopefully a more practical solution is developed otherwise industry wont incorporate it.

Enjoy all!

regards,

"Structural Engineering is the Art of moulding materials we do not wholly understand into shapes we cannot precisely analyse, so as to withstand forces we cannot really assess, in such a way that the community at large has no reason to suspect the extent of our ignorance." Dr. Dykes, 1976

 

[rapt]

ArronPTeng , steveh49 and Toby43

I assume the above post is from all 3 of you, or at least you all agree on it.

I cannot understand your conclusions, which I assumed would be based on logic considering your desire for an engineering discussion. Contrary to you comment regarding a "trail to the roots of a solution", both Kootk and I have given you the name of a reference document covering this. ACI318 has had a clause at least since 2011 covering it and possibly earlier, CSA 23 has had it since 1994 that I know of and possibly earlier. The Canadians possibly "got into it first" as someone suggested several months ago, because they were heavily involved in the development of Studrail and would have seen the test German results I saw for it in the early 1990's and possibly had their own corroborating tests.

It is not like compression ductility in over reinforced sections requiring capacity to control it. As suggested above in several posts by several people it simply provides a mechanism to provide support for the applied loads on the floor to transfer to the column in the event of a punching shear failure. It allows the connection to behave in a more ductile manner in the case of a punching failure instead of the normal very brittle failure mode. The methodology requires reinforcement to be supplied to support that load, V* as defined in AS3600. if you think it requires sufficient reinforcement to support the shear head capacity, Vuc, that is ok. You can provide the extra reinforcement. AS3600 requires you to provide a minimum based on V*. You can always add extra based on Vuc if you want to.

It does not supply a Robustness solution in the case of the loss of a support, the designer still needs to consider this as an extra design requirement.

The research and testing has been done since the 1980's and possibly earlier. That is why it has been in some codes for so long. AS3600 has now caught up and it is now a mandatory requirement in AS3600-2018. It has been discussed by sub-committees and the main committee developing AS3600 and went out in draft form for Public Comment. It passed through all of that and is now included in the code. It is not optional. If you do not provide it then you are designing outside AS3600 and outside the Building Code of Australia. And presumably outside the requirements of your PI insurance and any contracts you have signed with clients.

 

[Toby43]

Rapt,
I wholly agree in the need for integrity reinforcement, I'm just not convinced on the "ductile" part of it. Most of the tests I've read about about (Mitchell and Cook, Melo and Regan) firstly break the slab by getting it to punch, and then proceed to re-apply load (through some form of jack) to assess the post-punching capacity. Their is a load drop in this scenario (see image), whereas, I suspect if they used kentledge to load the slabs until punching failure, then the post punching capacity would quickly be exhausted aswell. A real structure would not experience a load drop at a column unless their was an alternate load path, thus I feel ductility cannot be assured unless capacity design principles are adopted.
So as I have stated earlier in this discussion, I most definitely see the need for it, but more so to minimise the risk of disproportionate collapse, in the event that the inherent uncertainties related to punching shear, cause failure.

Regards
Toby