Construction Joint in STM Designed Transfer Beam 1

[KootK]

Please refer to the PDF sketch attached.

A construction joint has been constructed that crosses a major concrete transfer beam that we've designed using strut and tie methods. The joint cannot be relocated to a more favourable position.

For a non-STM designed beam, I would address this issue using shear friction provisions. Additionally, I would count on my flexural tension reinforcement to pull double duty as flexural / shear friction steel. If any additional shear friction reinforcement was required, I'd add it to the beam tension steel per ACI recommendations.

Because the girder under consideration was designed using STM methods, I'm wondering if the procedure described in the previous paragraph is still valid. To be consistent with my STM design, I feel that:

1) I should be able to use the horizontal component of the compression strut crossing the joint as an effective clamping force for shear friction computations.

2) I should concentrate any additional reinforcement required for shear friction over the calculated height of the compression strut.

3) I should calculate the maximum permitted shear stress using the vertically projected area of the compression strut rather than the area of the entire beam.

On the one hand, I feel that my shear friction design should be consistent with my STM design, as described above. On the other hand, I feel that the "right" answer should be the same for both STM and sectional design methods.

Please advise.

KootK

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[dcarr82775]

STM?

 

[KootK]

Strut and tie method / strut and tie model. TLA's are confusing. My bad.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[CELinOttawa]

Well I have designed several pile caps that look not at all unlike what you've got there... In one case I wound up with an unplanned cold joint.

In repairing that one, I put two rows of four rebar (Reidbar to be specific, but the closest you'd be able to find would be DYWIDAG) and developped them with Hilti Hit HY-150. On the free end I had bar terminators installed.

Effectively both the concrete design manual of connections I used in New Zealand (XXX) and the Hilti International Manual recommended treating this as developed steel crossing the joint. I guess that puts me in the Hokkie66 camp of non-shear friction engineers.

Since you have the ability to put whatever you want in to allow for the cold joint, I would solve this by providing fully lapped tension and compression steel, possibly using face couplers (Lenton if you had to, but DYWIDAG would be best with screw-in formwork style types as a fall back) and then use coil steel to ensure good shear transfer across the joint.

Bear in mind that many, MANY, of these cold joints have been built successfully over the years with just surface roughening to + or - 3/8" and SSD condition at pour. Some steel across the joint is a nice security blanket, but with good workmanship it is not necessary.

 

[KootK]

@CEL: thanks for your response. While you've shared some good information about your solution, you haven't really provided any feedback on mine. Do you think that an STM designed beam warrants a different solution as I've proposed? Do you agree that my first point is probably the reason why so many unreinforced joints have been successful in the past?

Effectively both the concrete design manual of connections I used in New Zealand (XXX) and the Hilti International Manual recommended treating this as developed steel crossing the joint. I guess that puts me in the Hokkie66 camp of non-shear friction engineers.

What you've described is shear friction. I'll just have to cross my fingers that Hokie doesn't see this. Can you supply the title of your NZ connections manual. That sounds like something that I might like to procure.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[CELinOttawa]

Crap... I usually search posts for "XXX", which is why that way there... The name is: "Guidelines for Structural Precast Concrete in Buildings".

You can try to find it online, or just click here:

https://duckduckgo.com/l/?kh=-1&uddg=http://caenz.squarespace.com/s/Precast_Concrete.pdf

*evil smiles* Some really great detailing in there...

 

[CELinOttawa]

P.S. I don't think what I do is properly shear friction. I believe in the natural tendency for concrete to adhere extremely well to existing concrete, and then I put in some "paranoia steel".

As best as I understand it, in shear-friction you're calculating the strength you'll get from the amount of shear and area of roughened concrete. I don't buy that. I think you can safely rely upon the Vc and Vs if you do your detailing well, but if you try to go beyond this you're a fool.

I hope Hokkie66 DOES post; I am a huge fan of learning here and he's very much the kind of engineer we can learn alot from...

 

[slickdeals]

@KootK,
My $0.02. I would distribute the shear friction reinforcement over the entire depth of the section between top and bottom reinf. Wouldn't the compression force have a vertical component at the joint? I would also calculate the max shear stress based on the depth of the section rather than the strut.

 

[Trouser]

KootK,

Interesting issue. I have a few thoughts and some experience in this arena that I'll try to layout logically.

I am a bridge engineer with the Oregon DOT. I have used STM during repair work (bearing replacements) to asses the adequacy of diaphgrams/backwalls for use during jacking operations. I only do this when a section design model shows inadequate capacity. These reinforced concrete beams typically have a couple small bars (top and bottom) crossing the joint between the diapgragm and the girder. I am unsure of whether these were placed monolithically, but in the past I have made use of the entire section capacity for interface shear friction (calculated via AAHSTO LRFD) with success.

If I were in your shoes, with your instincts/concerns, I'd draw a free-body diagram. Split the compression strut forces into vertical and horizontal loads and make use of the normal force (clamping) in my calculations. I don't see how it is any different than a prestressing force. In my gut, I feel that the vertical component of the strut is a true shear force resisted by the entire section. I guess I'd ask myself the question, "What if the strut failed in interface shear?". The answer I see is that the entire section would jump in to help. In a true STM model, you could at least count on the remaining section below the strut to act as a support. This section wouldn't be in compression, but it would still add to the overall capacity (and likely contains your steel).

Just some thoughts. Let me know if I need to clarify anything.

 

[KootK]

@CEL: thanks for the great reference and thoughtful comments.

@Slick: thanks for your 0.02. This is a poll of sorts to that's perfect. I agree 100% that the compressive strut would have a vertical component at the joint.

@Trouser: how have we not met before? You, my friend, are practically reading my mind. I respectfully disagree with this:

In my gut, I feel that the vertical component of the strut is a true shear force resisted by the entire section. I guess I'd ask myself the question, "What if the strut failed in interface shear?". The answer I see is that the entire section would jump in to help. In a true STM model, you could at least count on the remaining section below the strut to act as a support.

Despite being a native speaker of the King's English, I can't really think of a way to articulate my concern with words. Please refer to the muddled sketch attached to this email. I feel that the strut couldn't "hang" from the sections of concrete highlighted in green without those sections of concrete failing in tension. The only way around this, that I can see, would be to provide some hanger reinforcement beside the cold joint to deliver the shear up to the top of the beam. But then that would essentially be adding an extra tie to the STM model. And I've done that on occasion for exactly this reason. That reason being either paranoia or a lack of fundamental understanding I suppose.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[dcarr82775]

I think if you provide stirrups each side of the joint (what I assume you are calling hanger steel) you prevent the spalls. After doing that I would provide shear friction steel generally where you show it, but I would be willing to spread it out a bit more. I would size it based on the vertical component of the strut. I imagine the loads are high or you might not have gone through the STM and just done a 'normal' beam design. I do agree with CEL that with good construction practices the bond between new and old is very good, maybe even enough, but then I so rarely see good construction practice that I tend to add belts, suspenders, and whatever else I can think of when dealt this type of hand.

 

[KootK]

@Dcarr: agreed on all counts. It's probably a subject for another thread but I see parallels to the more common situation of a beam dowel connected to an existing wall. I've been providing hanger steel tight to the wall for the very same reasons. This is another area in which I struggle with shear friction. Designers often space their SF reinforcement uniformly through out the connected cross section. In many instances, I think that might be inappropriate without something akin to the hanger steel that we're discussing here.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Trouser]

@KootK,

Test your GoogleFu:
EVALUATION OF BENT CAPS IN REINFORCED CONCRETE DECK GIRDER BRIDGES, PART 2
FINAL REPORT
ODOT RS 121
Dr. Christopher Higgins, Oregon State University

I believe that you can find this report online for free. The report is quite useful for visualization and application. I’d suggest taking a look at how these beams fail. The report uses common Oregon inventory for the selection of bent caps breaks down the analytics using STM. The study finds that STM is still quite conservative, even with poor column details and poor bent cap longitudinal steel (straight bars – not developed). In essence, we already have suspenders and belts in place…we don’t need to overdesign these sections. More steel could cause as many issues as it fixes (poor placement = bad interface friction). Don't be that kind of engineer. If you're concerned, talk with your inspectors!

Some other thoughts…have you considered addition another truss configuration to the analysis? Re-arranging your truss so that a vertical member is located at the joint could be an alternate way of analyzing this section? This would let you evaluate the joint as another case within the model used for other design elements.

Again, I’d really stress how the compression strut will fail in interface shear. For the interface shear failure to occur within the compression strut, strains would be sufficient to mobilize the entire section. This won’t be an exploding beam and there is still plenty of interface shear friction in your remaining section, outside the compression strut.

Something else that I've noticed, the saltier engineers seem to be comfortable with the fact that past practice shows this is likely of no concern. If it isn't codified, it hasn't caused many issues.

 

[KootK]

I found the document without any difficulty Trouser. Thanks for the recommendation. Can't beat DOT research for contemporary STM stuff. I'll check it out over the weekend. I agree that precedent leads one to believe that cold joint issues aren't a big deal. This post is actually an offshoot of another that I've got going: Link. If you check out the attachment to my original post there, you'll see my attempt at explaining why that might be the case from a mechanics perspective. As you say, however, testing trumps everything else.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[hokie66]

In order to deal with the vertical component of the strut force, I would want the strut to have a joint normal to its axis. Therefore, I would want the already cast section of concrete to be chipped out to provide this normal axis, which could be in steps. For a strut in a highly loaded transfer beam or pile cap, I wouldn't like to depend on resistance to sliding, whatever the assumptions. Columns don't have inclined joints, and neither should STM struts.

 

[KootK]

You raise an interesting point regarding columns Hokie. As you know, you've been helping me out with a related shear friction issue lately (Link). While you're no doubt sick of indulging me on the SF front, I believe that the concepts in the other thread have application here. Try his wacky supposition on for size:

I believe that columns -- all columns -- do in fact have inclined joints. Several actually: 15deg, 30 deg, 60 deg... and at every location along the columns too. Those joints are just monolithic, held together by shear friction (or the atheist equivalent), and perhaps... imaginary.

Why would I bother to say something so fru-fru and bizzare? Imagine me in seated lotus position as I type this. For me, the imaginary shear friction plane idea from the other thread has a very practical application. It allows me to be less conservative when dealing with surprise cold joints. Here's how it goes:

1) Contractor calls me up and proposes / tells me about a unplanned cold joint.
2) I freak out. You put a cold joint WHERE? It's the worst possible location.
3) I remind myself that there was always a slip plane at the proposed cold joint location. The only difference is that it was originally a monolithic shear friction joint and now it will be a roughened shear friction joint.
4) I check the numbers (mu = 1.4 vs 1.0) and, if things work out, I carry on.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[KootK]

Although it's clearly the consensus here, I'm still sceptical about using the full cross section shear friction capacity to deal with shear across my STM cold joint. Please refer to the attached sketch for my last attempt at frightening others. It's my shot at producing FBD's for the various parts and pieces. I believe that using the concrete above the strut creates a demand for tension capacity that isn't present without providing designated hanger stirrups.

@Trouser: In the interest of full disclosure, I still haven't read your OR DOT document yet.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[CELinOttawa]

Too existential for me... Choose an analysis method, choose a design methodology, and then believe in both as you apply them throughout.

We could never know the truth state of stress in any case. Wooten FTW.

 

[hokie66]

I agree the part above the strut is useless in the given model. I wouldn't like to depend on hanging reinforcement in this situation, and with the previously cast part, you don't necessarily even have those bars.

I can't see your belief that there are inclined planes in columns, except at crush load. Think failure modes of concrete test cylinders.

For struts, I still want bearing normal to the force, not on an incline.

 

[KootK]

For shame CEL! For shame. You, of all people... No e-Christmas card in your inbox this year.

To follow a recipe and advocate faith in place of understanding is to turn your back on what being an engineer is all about. Not knowing the true state of stress is only a valid cop out if you're dealing with a ductile system that can redistribute the load to some other capable path. Not so here.


The greatest trick that bond stress ever pulled was convincing the world it didn't exist.