Retaining Wall - Flexural Reinforcement from Stem Into Footing 21

[CWEngineer]

I am trying to get some clarification regarding the flexural reinforcement of the stem of a retaining wall into the footing.

Does the flexural reinforcement in the stem of a wall, need to be developed into the toe, such as show in Figure 1 of the attached document. Or is providing a standard hook (12db), sufficient, such as that show on Figure 2 of the attached document? If providing a standard hook is sufficient, can the hook be turned towards the heel?

Thanks in advance

 

[KootK]

KootK, I apologize for calling you a liar, I should have just stuck to the facts.

I'm sure you can appreciate how frustrating it is to hear you insist that the horizontal bars needs a full lap onto the vertical bars, and then completely dismiss the fact that a 180 hook example gets close to full capacity without any such lap, and then go off on some tangent about trombones...

So your sorry but I'm still a flip flopping chaser of irrelevant tangents? Gosh... thanks Tom.

I'm sure that you can appreciate how frustrating is for me to hear you repeatedly misinterpreting my statements and sketches. Nowhere have I or Nilsson been discussing 180 hooks, disparagingly or otherwise. Surely you must see that by now?

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.

 

[Tomfh]

misinterpreting my statements and sketches. Nowhere have I or Nilsson been discussing 180 hooks

Your sketch shows regular 180 bend without flat section at bottom, not a trombone.

Nilsson shows regular 180 bends without flat section, not trombones.

 

[KootK]

Did you totally miss my 4 Oct 16 23:02 post? Nilsson and I both explicitly described our bends as hairpins rather than 180 hooks.

Personally, I don't want to be done with this thread. I think that it's interesting and important and I look forward to discussing it further with any other folks who may be late to the party. And I'll continue to make myself available for that.

That said, it seems that the discussion between you and I has ceased be fruitful. How about this Tom:

1) Make one last post on this thread if you like. Say whatever you wish. And then say no more, at least not to me, about this.

2) I'll not respond to your next post. You shall have have the last word forever more, whatever you choose those last words to be.

How about it?





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.

 

[Tomfh]

1) Make one last post on this thread if you like. Say whatever you wish. And then say no more, at least not to me, about this.

I will sum up my understanding:

KootK believes that a 90 degree standard cog turning into the toe is “abominable” (28 Sep 16 01:35):

He has presented STM/pulley models which he argues show this to be the case. After much toing and froing, and my misundertstanding his concern as being vertical bar anchorage, we arrived at his “precise” concern, which is that a standard 90 degree cog provides insufficient lap with the footing bottom bars to transfer the bottom bar tensile load around the corner up into the joint. (4 Oct 16 01:12)

I had not fully considered his concern, but nonetheless felt the pulley analogy is misleading. So I asked for Nilsson's results, which KootK kindly offered up. Nilsson’s T13 (4 Oct 16 02:01) showed even less lap between horizontal and vertical bars than a 90 degree bend provides, and yet T13 achieves 79% efficiency. KootK presumably didn’t like these numbers, referring to them as “unexpected” (4 Oct 16 03:23).

SCALE DRAWING
To help explain away this apparently anomalous results, and preserve the pulley idea, KootK produced a “scale drawing” (4 Oct 16 03:23) of a 180 degree hook superimposed on a very incorrectly drawn 90 degree hook, to show that – despite appearances - a 180 hook in fact provides *superior* lapping to the bottom bars than does a standard 90 degree hook. Hallelulia! That’s why Nilsson’s examples work! A 180 hook touching a horizontal bar is actual a very effective lap, the pulley model is saved.

TROMBONES
Also, to bolster his argument, KootK said he wasn’t even talking about 180 hooks (despite drawing one), he in fact meant trombone hairpin (not that they satisfy Ld + ??? anyway).
And regarding Nilsson’s hairpins, KootK claims that the word “hairpin” in Nilsson’s text proves it is trombones, regardless of what Nilsson has drawn, and regardless that the actual crack pattern traces out an ordinary 180 bend.


SUMMARY
T13 has negligible lap between vertical and horizontal bars, and T12a has complete bar continuity, and yet – to KootKs surprise - they provide almost identical capacity (79% vs 82%). This contradicts KootKs primary claim that the horizontal bars need a full lap onto the vertical steel in order for the joint to function, and thus undermines the basis of his claim that a standard hook is an abominable detail.

 

[RFreund]

So I get the pulley concept and why that improves the joint. Quick question regarding the hairpin/180 degree hook whatever -

@kootk - your sketch (see below) is the STM explanation for why the hairpin/180 works but not as well due to the reduced "reduced flexural depth". Is that right?

EIT

http://www.HowToEngineer.com

 

[TehMightyEngineer]

KootK; while I do agree with you (both on your rigorous debate mentality and this topic at hand); I feel that Tomfh is correct about the technicality of the "hairpin". Your drawing does appear somewhat continuous in the curve (if you intended otherwise I would look at such intent, but I'm just stating what I'm observing). Further, Nilsson does seem to be showing a continuous curve in his sketches more in line with a 180° hook. While a large enough radius curve seems fundamentally identical to a "hairpin" or "trombone" as you put it, I feel it's worth pointing out that quibbling over the minutia of what is shown is missing the key points of this debate. Finally, I've seen multiple various definitions of "hairpin" that are defined as either "trombones" or "180° hooks" or even V-shaped bars (like the ones shown for supplemental reinforcement in ACI 318 appendix D for shear) so just because Nilsson did use the term "hairpin" doesn't mean to me explicitly that they were or were not 180 degree hooks as the term has too much ambiguity to it.

All that said though, I agree with your point. So, if anyone has no objections, can we try to focus on what the differences between the reinforcements gives us and why we have some strength (79%, etc.) with one detail and more strength with others. Trying to decide if Nilsson's sketches show X or Y seems important but not the key topic we should be focusing on as it is entirely subject to opinion (in my opinion). Let's not forget the ultimate goal of this thread: "what is the proper joint detail at a retaining wall". At the most basic level; wouldn't the 79% strength level be "acceptable" as long as we knew what the reduction in strength was?

Seems like identifying the cause of this reduction in effective strength is the key; not how many engineers it takes to define a hairpin.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[TehMightyEngineer]

I'm going to attempt to summarize what I'm seeing as the crux of each persons argument. This is GROSSLY simplified; no yelling please if I misrepresent something you said. My opinion/interpretation below:

KootK: Providing anything but my detail potentially gives you a strength of X% less than the ideal full flexural capacity of the joint. This reduction in strength isn't addressed by most retaining wall designs. Therefore I declare this joint detail verboten as it's potentially dangerous unless we can understand how much strength is lost. It's also not the most efficient as my strut-and-tie model shows so you're probably going to get best results with my detail (further recommended by CRSI, ACI joint details, etc.).

Tomfh: (To paraphrase Tomfh's own summary, which was well written). T16 has bar continuity giving almost identical capacity (79% vs 82%) to the abominable detail. Thus a standard hook is not an abominable detail, it's simply a reduction in capacity at worst and nearly as efficient. If we design for that reduction then there's nothing wrong (and this is why walls aren't falling down all over the place) and saying absolutes like "we must have full lap splices" and "we must have maximum effective joint strength" isn't the only solution.

[Ducks under desk after having painted a huge bullseye and possibly caused both parties to be forced to repeat themselves to clarify my simplifications.]

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[KootK]

@kootk - your sketch (see below) is the STM explanation for why the hairpin/180 works but not as well due to the reduced "reduced flexural depth". Is that right?

Uhh... In my heart of hearts, I want to just say yes. However, this thread has become hyper sensitive to semantics. So I'll say this instead in the interest of hyper precision:

The STM is an explanation for why a stem bar that is developed vertically within the footing depth, but is not explicitly designed to transfer tension around the corner, will work but with a reduced efficiency which, fundamentally, I believe is a result of that "reduced flexural depth" that I mentioned. In the context of the previous statement, what I mean by developed vertically is one of the following:

1) A developed vertical with no hook

2) A developed vertical with a 90 degree hook.

3) A developed vertical with a 180 degree hook.

4) Perhaps one leg of a trombone style hairpin to the extent that it would function as vertical development for the stem bars.

So, yeah. Yes.

I'll point out that "trombones" is not a KootK term. It was just included in that diagram that I pilfered off of the inter-web. The silliness of the term has been used to poke fun of me and my arguments. That said, it's kinda growing on me and folks seem to know what that means so I'm just gonna roll with it. Trombone = KootK Style Hairpin.

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.

 

[TehMightyEngineer]

Oh, I thought "hairpin" was hilarious and taught it to all the guys in the precast plant. I can't wait to see what they do when I call out a "trombone" on my next project.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[Tomfh]

TME,

what specific point of kootks is it that you are agreeing with?

As for the cause of the strength reduction, as Nillson captions - hairpin anchorage failure. With continuous anchor bar or long laps you can milk a bit more capacity out of the joint.

 

[KootK]

Your drawing does appear somewhat continuous in the curve

My green rebar 90 has a definite curve in the corner. My green rebar 90 has a definite straight, horizontal extension. These things are congruous with reality. Where I went wrong is that I measured the end of the hook as 12db from the vertical bar centerline rather from the point of tangency on the curve. So yeah, the straight part is a little short (~60 mm). Hari Kari for me.

Further, Nilsson does seem to be showing a continuous curve in his sketches more in line with a 180° hook. While a large enough radius curve seems fundamentally identical to a "hairpin" or "trombone" as you put it, I feel it's worth pointing out that quibbling over the minutia of what is shown is missing the key points of this debate.

It's a small sketch. Maybe it's not drawn in incredibly intricate detail. Maybe it's incorrect of me to interpret the width of the hairpin extending to cover as meaning trombone style. Maybe when Nilsson says hairpin, he actually means 180. Maybe it is a big radius rather than a true "trombone".

- Nilsson said it was a hairpin.
- It looks like a hairpin to me.
- There is nothing that explicitly said that it is a 180.

While my interpretation may indeed not be correct, when you add all of that up, it leads me to believe that my interpretation is at least as valid as anyone else's. Additionally, when I posted my apparently ambiguous sketch, I included text clarify my intent. And then, subsequently, I re-clarified it over and over again. I don't see how, at this point, anyone could be unclear with regard to what my intention was.

I feel it's worth pointing out that quibbling over the minutia of what is shown is missing the key points of this debate.

I agree but I can only respond to the points that people choose to query me on.

At the most basic level; wouldn't the 79% strength level be "acceptable" as long as we knew what the reduction in strength was?

Yup. And I have previously mentioned that sometimes increase the rebar by 1/%Efficiency rather than include the diagonal bars for small retaining walls where I'm unconcerned with crack width. One point of concern for me, however, is how reliable the inefficiencies scale up for larger bars. All of the testing was done on 10M. And my MathCAD work above suggests that things may get worse, quickly, for larger bars.


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.

 

[KootK]

KootK: Providing anything but my detail potentially gives you a strength of X% less than the ideal full flexural capacity of the joint. This reduction in strength isn't addressed by most retaining wall designs. Therefore I declare this joint detail verboten as it's potentially dangerous unless we can understand how much strength is lost. It's also not the most efficient as my strut-and-tie model shows so you're probably going to get best results with my detail (further recommended by CRSI, ACI joint details, etc.).

Firstly, what the hell are you doing man?!? I just about had this settled down.

Secondly, yeah, you pretty much nailed it. Minor clarifications:

1) I might have toned down some of the absolutes for mass consumption.

2) I believe the 90 degree hook termination to be a fundamentally poor detailing choice, regardless of the efficiency numbers etc. Anecdotally, every time that I've encountered a proponent of that detail, that proponent has proven themselves to harbor misunderstandings about how concrete joints function.

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.

 

[KootK]

Your "scale drawing" went wrong in at least three ways- The bend diameter is all wrong. The straight extension is all wrong (you've drawn 4db not 12db). The overall length is all wrong (a standard 10M 90 degree bend within a 200 wall with 25 cover will extend past the wall face).

I thought that we had a truce in place Tom? Can we not abide by that? Self discipline, it's what separates us from the animals.

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.

 

[KootK]

Yes it's fairly clear by now that your beliefs are impervious to the results.

I'm just going to respond to others while doing my best to stick to the truce Tom. Please don't interpret that as any kind of slight as that's certainly not my intent.

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.

 

[Tomfh]

Therefore I declare this joint detail verboten as it's potentially dangerous unless we can understand how much strength is lost.

Well then you're going to have to ban just about all the joints in this table, including the case of an unbroken L-bar, as the only joints which reliably give you 100%+ are the ones with diagonals and ties bandaging the knee up.:

The 79% is in the ball park for all these sort of joints.

 

[TehMightyEngineer]

Quote (KootK)
Therefore I declare this joint detail verboten as it's potentially dangerous unless we can understand how much strength is lost.

I just want to point out that this was my interpretation of KootK's position. While he agreed with my assessment, I want to make sure people don't think this is a direct quote of KootK.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[TehMightyEngineer]

TME, what specific point of kootks is it that you are agreeing with?

First, let me point out that this is half my opinion and half my understanding/conclusions based on weighing all the comments presented here. I will fully admit I am no expert in concrete joint detail though I believe I have enough of an understanding to draw my own conclusions. Also, while I share KootK's abrasive debate style I do not have any "favorite" person here and considered all positions equally.

This is my own interpretations of the arguments presented, please correct my if I misunderstood someone's position. That said, these are the points I believe KootK and I agree on.

[ul]
[li]Failure of a straight or hooked bar which is fully developed does not preclude an "appendix D" style tension pull-out cone. While it may not control the strength of the joint, not checking the capacity of the frustum pull-out failure is neglecting a failure mode. Providing force transfer around a corner is one way to ensure that this failure mode will not control. It's worth pointing out that I believe even a 90° hook or straight dowel will still transfer some force around a corner, as both KootK and pretty much everyone else has indicated.[/li]
[li]Curved bar nodes provide for force transfer around corner joints. Force transfer around the corner is required for the cantilevered retaining wall to function as idealized.[/li]
[li]Fully developing a large radius curved bar around the joint as KootK shows appears the be the most efficient detail when comparing tested results, modeling of forces, recommendations from CRSI/ACI/etc., constructibility, crack control, and ease of analysis. This is probably the biggest reason I'd side with KootK's argument over others.[/li]
[/ul]

Things I may not agree on fully with KootK's argument.

[ul]
[li]There are other ways to get strength out of the joint. A simple 90° hooked bar or a hairpin will transmit some force around the corner and likely will work sufficiently. However, considering the full flexural capacity of the joint without taking in account the potential deleterious effects of this setup is not appropriate. Use of this detail, properly accounting for all the strengths, is perfectly acceptable. I do agree with KootK that confirming that you have properly scaled Nilsson's test results is difficult without understanding the reason for the loss in strength.[/li]
[li]I'm not convinced that CELinOttowa is correct in his force diagram for the curved node bar forces where he included the development forces. But, neither am I convinced he's wrong. I believe there's some definite merit to his comments regarding the forces involved.[/li]
[/ul]

By all means discuss some of the above with me; but I don't want to derail this thread by forcing people to stop and repeat what they said to correct my interpretation.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[KootK]

There are other ways to get strength out of the joint. A simple 90° hooked bar or a hairpin will transmit some force around the corner and likely will work sufficiently. However, considering the full flexural capacity of the joint without taking in account the potential deleterious effects of this setup is not appropriate.

So I woke up this morning, popped out of bed, and realized that I still have a bit more to give on this. And not just petty bickering either. Good stuff.

Everybody seems to want to know how, and how well, the joints work when the stem bars are terminated in a developed hook and no more. Above, I've pitched two explanations for the mechanism/efficiency in such joints: the STM model and the joint clamping mechanism analogous to a perimeter beam column joint.

The major drawback with STM models, of course, is that nobody has time for that stuff in a design office. Additionally, there are some features of the joint behavior that are actually less intuitive when presented in STM than they might otherwise be if presented in terms of traditional sectional design methods.

So... I've decided to generate a proposed sectional method for the design of these joints as illustrated in the two sketches at the bottom of this post as follows:

1) Detail A shows the excellent clamping mechanism that could be relied on if only the strut would not slip off of the stem bars due to lack of development. I've assumed that the tension forces in the footing rebar can be conceived of as being pulled across the joint and manifested as additional compression on the other side. We do similar things in strut and tie so I'm comfortable with the validity of the assumption. Really, it's a big part of what makes the clamping mechanism so effective.

2) Detail B shows a proposed method of adjusting the design for the amount of stem bar development actually available. Again, I've assumed that the tension forces in the footing rebar can be conceived of as being pulled across the joint and manifested as additional compression on the other side. Basically, it amounts to this:

2a) Figure out the dimension La to make things pan out. Some iteration might be required.

2b) Increase the heel reinforcing to account for the reduced flexural depth (jd_heel_reduced).

2c) Increase the stem reinforcing to account for the partial rebar development (La vs Ldh).

2c) Check one way shear on heel based on reduced flexural depth.

2d) Ignore the compression strut as we would with other sectional methods (I guess).

2e) Have faith that the system will "find" this load path rather than initiate an anchorage failure along the way.


An interesting feature of this is that it provides a couple more ways to explain why we don't see failures with this detailing:

1) We know that current, code specified development lengths are quite conservative. The reduced flexural depth in the heel is probably better than it seems. And the ratio La/Ldh is probably greater than it seems.

2) We know that the probable yield strength of rebar closer 1.25 x fy. This will help to offset the additional rebar required by the inefficiencies inherent in the joint.

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.

 

[Tomfh]

Fully developing a large radius curved bar around the joint as KootK shows appears the be the most efficient detail when comparing tested results, modeling of forces, recommendations from CRSI/ACI/etc., constructibility, crack control, and ease of analysis. This is probably the biggest reason I'd side with KootK's argument over others.

I agree an unbroken bar is more more efficient. It is slightly more efficient. The hairpin (T13) gives 79%. When all else is equal and the bar continues into a decent length lap (T12b) the capacity goes to 82%. Different reinforcement ratios (T12, T16) can see the lapped/continuous bar capacity increase up to 100% capacity (perhaps the hairpin case would increase too, or maybe it maxxes out around 80%)

What I disagree with is that this proves the STM/pulley model, that it proves a bar with less than full lap is "abominable" and that it proves simply wrapping a continuous bar "guarantees" idealized performance. It is a silly to say 79% reflects a dangerous joint and that 82% is perfectly good reliable joint that we needn't look twice at. If one is questionable, the other is too.

There are other ways to get strength out of the joint. A simple 90° hooked bar or a hairpin will transmit some force around the corner and likely will work sufficiently.

Yes I agree a standard hook will provide some lap strength and would help keep the joint tied up. Such a case lies midway between the 180 hairpin vertical bars with no lap and a continuous 90 degree bar, so you expect intermediate performance.

However, considering the full flexural capacity of the joint without taking in account the potential deleterious effects of this setup is not appropriate. Use of this detail, properly accounting for all the strengths, is perfectly acceptable. I do agree with KootK that confirming that you have properly scaled Nilsson's test results is difficult without understanding the reason for the loss in strength.

I'm all for testing and understanding joints. It just strikes me as a real double standard to condemn less than full lap as dangerous whilst maintaing that a continuous bar is safe and reliable. How can 79% be dangerous and 82% safe?

As for the reason for the loss in strength in the hairpin - it is anchorage failure, as noted on nilssons images. The hairpin can pull out easier than a continuous/90degrees bar beacause there is nothing crossing the pullout interface - see image below. This is in line with KootKs ideas, but is not nearly as bad as he imagines.

Since the continuous bar cases start failing at only 3% more load, at 82%, these joint has other problems anyway. You can't simply make a continuous bar, dust your hands, and give yourself a pat on the back.

If you want the guaranteed 100% performance you are demanding, with no questions as to reduced capacity, you need diagonals and bandage bars, etc

 

[Tomfh]

Fixed.