strut and tie modelling in precast deep beam/ wall 1

[4dmodeller]

hi all,

is it possible to use strut and tie modelling in precast wall (deep beam) - 2 stories tall.
I understand that you can use wet joint to connect the tie together but what would you do at the horizontal joint between precast on one floor and the other?

 

[rapt]

You would have to provide horizontal and vertical shear transfer across the joint. Not a good idea in my opinion!

make the precast 2 floors high!

 

[4dmodeller]

Rapt,

I understand that it is not the best design but ive been forced to due to circumstances

How would you design shear transfer across the joint? do i assume it works like a normal flexural beam and use VQ/I to work out amount of shear dowel across the horizontal joint? would that be acceptable?

is there a way to keep the design consistent with strut and tie i.e. use horizontal component of diagonal strut forces that goes through the joint as the longitudinal shear demand needed through the joint?

 

[KootK]

do i assume it works like a normal flexural beam and use VQ/I to work out amount of shear dowel across the horizontal joint? would that be acceptable?

This would be acceptable were you using conventional, sectional flexural design. It may be inconsistent with S&T however depending on your particular arrangement.

is there a way to keep the design consistent with strut and tie i.e. use horizontal component of diagonal strut forces that goes through the joint as the longitudinal shear demand needed through the joint?

I believe so. And I think that you're pretty close to the mark with respect to what needs to be done. Better still, other than providing for lateral stability at the joint, which should be easy, I see no reason to positively connect the two panels. Below, I've pitched two alternatives:

1) Straight up S&T as if the joint wasn't there. Shear friction across the horizontal joint needs to work without reinforcing. And it may well work as you can utilize the joint clamping created by the vertical component of strut compression at the joint. You could enhance joint horizontal shear capacity with vertical dowel connections but, once you do that, it becomes rather difficult to know just where those dowels should be located to be effective.

2) A two tiered S&T system that should be about as efficient as #1. This is absolutely the way that I would go. The extra rebar tie is cheap insurance to produce a system that should be pretty robust. I'd oversize the upper tie a fair bit to keep strut spreading at the joint to a minimum.

Obviously, I'm guessing at the loading and proportions here. If you post a sketch of the real thing, that may change my recommendation some.

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]

Technically, option #1 could be considered in violation of vertical skin reinforcement requirements at the joint. I'd be inclined not to worry about that, however, as I have a hard time imagining just what horrors would result.

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.

 

[4dmodeller]

hey KootK
Thanks for your two sketches which exactly confirm what i have in mind!

I have chosen to go to option 2 as well as reo are must stronger in tension than in shear. Like you said, it is also hard to justify effective zone of dowel bars that would help prevent longitudinal shear failure at the joint.
For option 1, I believe that for a simply supported deep beam, all the vertical dowels in one half of the panel must resist the horizontal component of that side strutting force. I believe this is in line with composite beam design where there must be enough shear studs from mid span to end of span to resist that longitudinal shear (and mirror on the other side). What are your thoughts on that?

for option 2, can you clarify "oversize the upper tie a fair bit to keep strut spreading at the joint to a minimum. " . esp. strut spreading at the joint bit.
in terms of tie requirement, i believe it should be the same as what is required on the bottom level (assuming the strutting angle is constant all the way to bottom support and ignoring how the strut strut vertically locally at joint before it redirect itself to the support again). Would that be fair assumption?

Thanks.

 

[4dmodeller]

KootK,

What about if there is a vertical joint? for example in the 2 scenarios attached. yes they are real scenarios. I believe they work but the detailing of it and how to justify using a uniform strut and tie approach is still a work in progress i.e. apply stitch plates to strut and tie through joint and compression strut through vertical joint
do you agree with the approach i proposed?

 

[KootK]

For option 1, I believe that for a simply supported deep beam, all the vertical dowels in one half of the panel must resist the horizontal component of that side strutting force. I believe this is in line with composite beam design where there must be enough shear studs from mid span to end of span to resist that longitudinal shear (and mirror on the other side). What are your thoughts on that?

I don't agree with this approach. Like I mentioned above, I feel that VQ/It stuff is only appropriate for conventional, sectional methods of flexural design. For STM, I believe that you need localized, discrete horizontal shear transfer commensurate with the localized, discrete load path that STM implies.

for option 2, can you clarify "oversize the upper tie a fair bit to keep strut spreading at the joint to a minimum. " . esp. strut spreading at the joint bit.

If the upper struts spread apart, due to tie elongation, that would imply cracking the snot out of the lower panel. Let's avoid that.

in terms of tie requirement, i believe it should be the same as what is required on the bottom level (assuming the strutting angle is constant all the way to bottom support and ignoring how the strut strut vertically locally at joint before it redirect itself to the support again). Would that be fair assumption?

Sounds about right to me.

What about if there is a vertical joint? for example in the 2 scenarios attached. yes they are real scenarios. I believe they work but the detailing of it and how to justify using a uniform strut and tie approach is still a work in progress i.e. apply stitch plates to strut and tie through joint and compression strut through vertical joint
do you agree with the approach i proposed?

Issues with #1:

1) It would be unstable with asymmetrical loading. You'll need some shear capacity along the vertical joint to deal with that.
2) I'd shift your piles in a bit such that they are completely beneath the walls. This isn't the place to be compressing your node geometry.
3) You'd need to provide lateral stability to the vertical joint where the horizontal struts come together. And it's not simple matter to know where that will occur.

Issues with #2:

1) I'd try to get all of your tension resolved by connection to the lower panel. That way you wouldn't need to sweat a tension tie connection from upper panel to lower panel.
2) You'll need to effectively lap your panel tension tie with the vertical steel in the pile directly below. That might require more space than you're showing and is another great reason to shift you piles inwards.



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.

 

[rapt]

If there is a vertical joint, I would use bonded PT tendons installed after the 2 panels are positioned to provide the tension tie at the bottom.

Your vertical and horizontal shear requirements could also be satisfied with bonded stressed bar to tie everything together.

 

[4dmodeller]

KootK,

If the upper struts spread apart, due to tie elongation, that would imply cracking the snot out of the lower panel

by providing antibursting reo along the strut and the tension tie at the bottom of top panel, that should prevent it no? I do feel it's a bit dodgy at the joint though..as this does usually mean going through weaker concrete slab = spreading the load. i think axial transmission should be checked.
on a side note, with bearing check that has an equation of A1/A2 to increase bearing confinement. What is your opinion of the definition of "supporting structure" in A2, let say the panel bear against the pile cap and is strutting to the piles, is A2 the pile? or the pile cap? I feel like it's pile cap because its about confinement i.e. if the pile cap is really deep it doesnt make sense to limit the confinement to the tiny pile..you have so much of the pile cap depth to confine it. what do you think?

Panel No .1
#1.Agree, I have provided a few heavy duty stitch plates at the vertical joint to allow for asymmetrical load. I few that i should consider taking out the unstable load by bringing it back to the lateral stability element as well i.e. load strut straight to the pile and getting strapped out by slab back to the core.
#2 agree, will see if it's too late to change it
#3 2 systems provided in #1 should cover it.

Panel No. 2
#1 do you mean putting all the stich plate shear connection in the lower panel? can you explain graphically? I feel that if i put all the stitch plates in the bottom panel, the strut and tie model wouldn't have changed. I would still need a tie at the LHS of panel. to bring the load down to the shear connection in the lower panel.
#2 agree. thanks for pointing it out.

 

[4dmodeller]

Rapt, how can you install a PT tendon into already installed precast concrete?

 

[KootK]

Rapt, how can you install a PT tendon into already installed precast concrete?

Until now, it was not clear to me that:

1) The precast was already installed.
2) The upper and lower wall panels were seperated by a floor slab.

by providing antibursting reo along the strut and the tension tie at the bottom of top panel, that should prevent it no?

I don't think so. The spreading of the struts will be restrained by the reinforcing that forms the primary tension tie between struts.

is A2 the pile? or the pile cap?

Both. Or neither. You have two separate bearing conditions that may require consideration:

1) Wall strut to pile cap.
2) Pile cap to pile.

If you are using S&T, however, your nodal stress checks should take precedence over, and obviate the need for, conventional bearing checks.

#1 do you mean putting all the stich plate shear connection in the lower panel?

Yes, if the connections can be made to work.

can you explain graphically?

I will if you require it but do I really need to? This seems simple enough to explain in prose that I wouldn't need to expend 15 minutes of my life sketching and scanning.

I feel that if i put all the stitch plates in the bottom panel, the strut and tie model wouldn't have changed. I would still need a tie at the LHS of panel. to bring the load down to the shear connection in the lower panel.

I don't see the problem. Your sketch already shows a full depth tension tie at the left hand side of the cantilevered panel. If you keep the connection to the upper panel then you'll need a tension tie connecting the upper and lower panels which seems as though it would be an necessary hassle.

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.

 

[4dmodeller]

KootK,

The panel has not been installed. It is in design stage. Rapt mentioned adding PT when the panels are installed. So I queried what he meant.

Both. Or neither. You have two separate bearing conditions that may require consideration:

1) Wall strut to pile cap.
2) Pile cap to pile.

If you are using S&T, however, your nodal stress checks should take precedence over, and obviate the need for, conventional bearing checks.

I feel that you should check for both bearing and nodal check. On the matter of A2, thanks for resolving the definition for me. So it is the pile cap..it's just that i have to do another check between pile cap to pile

Regarding the tension tie only in the lower panel, that makes sense. Thanks for clarifying. I dont think I can fit 2 stitch plates that big in top part of lower panel though. Have to resort to tension dowel i guess .

 

[KootK]

The panel has not been installed. It is in design stage. Rapt mentioned adding PT when the panels are installed. So I queried what he meant.

Got it. Perhaps one could cast a tendon and duct into each panel and provide a coupling device at the joint between panels. Post tensioning the tie would actually be a great way to deal with the strut spreading issue that we've been discussing.

I dont think I can fit 2 stitch plates that big in top part of lower panel though. Have to resort to tension dowel i guess .

Somewhere in ACI, there is a provision for tension tie members that requires designers to use only mechanical splices in the tie rebar rather than relying on lap splices. While that wouldn't strictly apply here, I would be inclined to apply it anyhow as this is a critical connection that will be permanently in tension. Another way to tackle this would be to use vertical post tensioning for the tie.

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.

 

[rapt]

They install post-tensioned tendons in precast bridges all the time. And PT is the best way to get the anchorage force to the ends and fully develop it past the support centre at the ends as some design codes require (I know the Canadian code does not require it Kootk, but that does not mean they are right. Maybe Collins and Mitchell got one wrong!).

Cast the anchorages and duct into the precast (properly aligned hopefully). After everything is assembled, thread in the tendon, stress it and grout it. Normally the precast joint is coated with epoxy to seal the joint but it is not relied on for tension. The full force is still taken by the tendons as it is for any strut-tie arrangement. Sometimes the precast elements are match cast against each other to ensure a perfect joint.

Same could be done vertically using high tensile stressed bars as they do often in water retaining structures and sometimes where tie downs are required in lift shafts etc.

 

[hokie66]

Maybe 4dmodeller is not familiar with post-tensioning? That's the way you do it, after the concrete is in place, whether it is precast of cast in place.

 

[KootK]

fully develop it past the support centre at the ends as some design codes require (I know the Canadian code does not require it Kootk, but that does not mean they are right. Maybe Collins and Mitchell got one wrong!).

We have a version of this but it's not quite so stringent. Full development past the centre of support seems excessive to me. Even analyzed with STM, you'd only need to develop for required force past the interior face of the support.

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.

 

[4dmodeller]

I was just a bit surprised to see PT tendons in walls. PT tendon in precast wall is very uncommon practice and would frown upon by builders (lower mid tier builder)where I work. also, it will be too congested with dowel bars and reo each face already. The panel is 180 thick.
As KootK mentioned, using normal ductile reo bar should do the job. regarding vertical joint, we will simply have to splice the bar with wet concrete joint. I know there has been discussion about only using mechanical splice in STM. However, i don't see it being a problem esp in the case that I am dealing with.

 

[rapt]

Yes, PT in precast or any wall is uncommon. But so is a beep beam made up of partial length and partial depth precast elements. It is the best solution for what you are asking, or for any deep beam for that matter.

Kootk, Depends on who does the STM. I would argue that development is required past the intersection of the compressive strut and the tension tie, which is normally the centre of the support! That is what is required by most design codes also.

 

[KootK]

I would argue that development is required past the intersection of the compressive strut and the tension tie, which is normally the centre of the support! That is what is required by most design codes also.

Really? From what I've seen in the codes and literature in Canada, Europe, and the US, the length available for tie development always seems to be as shown below. It's the intersection of the interior edge of the compression strut and the centroid of the tie group and is well to the inside of the bearing centerline.

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.