Purpose of tie beam to pile cap 2

[dccd]

What's the purpose of tie beam to pile cap ?

I was told that When the pile is overloaded, the tie beam from to the next pile cap can help to overcome the problem, When the pile is overloaded, it simply means that the axial compression on the pile is too high How can tie beam overcome the problem? The rebar of the tie beam is to provide some sort of moment resistance I guess?

But, again when the pile is overloaded, the main purpose is to reduce the axial load on pile. How can providing tie beam reducing the axial load on pile?

 

[1503-44]

I assume the context was overloading caused by a lateral load. That's typical weakness of piles. Very strong for axial load, but limited lateral resistance.

 

[dccd]

No, I was told that the purpose of tie beam is to overcome the overloading of vertical force of pile. why ? I just dont understand...

 

[1503-44]

Pretty hard to understand that logic. Its for lateral load.

A tie beam will tend to dump more vertical load into the pile cap as well.

 

[phamENG]

Tie beams are typically used in seismic regions. During a seismic event, they keep the foundation tied together so the building doesn't do a split.

They are also used to stabilize pile caps. Let's assume low to negligible seismic region. If I have a row of columns that only require 1 pile to support the load, I can't put a single pile under it and call it okay - it's unstable when you consider driving tolerances, etc. If I don't want tie beams, I have to put at least three piles under it (for stability in all directions). Or, I can put 2 piles under each one and run a tie beam along the column line. If there's another column line, I can put 1 pile under each and run a grid of tie beams to each to provide stability. It's just a matter of weight the economics of the foundation system. If the piles are 15' timber, then the P-3 probably isn't too bad. But if I need 120' precast concrete piles, then it might be better to use some tie beams to reduce the amount of pile driving. In this case, the tie beams are designed to resist the moment caused by eccentricities from the construction tolerances.

 

[phamENG]

And yes, if you have several closely spaced piles with tie beams between them and they are properly reinforced, they can redistribute some of the load to adjacent piles if overloaded. But I wouldn't count on it unless I was designing for progressive collapse resistance or something and that was part of my alternate load path.

 

[milkshakelake]

Yes, it is for the moment. Tie beams for piles is to avoid/"normalize" the eccentric loading at edges of the property line. And it's also as phamENG said. You can't have a condition where there is 1-2 piles without tie beam because it's unstable. 3 piles in a triangle are the minimum for stability in x-y axis. But it's generally cheaper to drive 1-2 piles and use a tie beam instead of driving 3 piles.

For lateral capacity, I batter the piles (put them at an angle).

 

[KootK]

I was told that When the pile is overloaded, the tie beam from to the next pile cap can help to overcome the problem

That's close. I suspect that you're really just speaking of generic grade beams rather than true tie beams where the dominant load resisting function is axial. The old school view of grade beams connecting piles, which the world has largely forgotten somehow, was not so much for improving capacity but, rather, for ironing out differential settlement. See one recommendation, below, from ACI 314-11.

 

[dccd]

It seems that the purpose of tie beam is just to minimize the settlement rather than overcome the pile overloading problem. How does the tie beam help to overcome the pile overloading and reducing the loading on the pile so that the vertical compresssion force doesnt exceed the pile vertical compression capacity ?

 

[KootK]

How does the tie beam help to overcome the pile overloading and reducing the loading on the pile so that the vertical compresssion force doesnt exceed the pile vertical compression capacity ?

It doesn't, at least not meaningfully in conventional situations. You can tell by the equations that I posted above that the load that might be transferred from one pile to its neighbors will be a small percentage of its total load.

 

[1503-44]

Exactly. Its like trying to redistribute a load directly over a support to other supports in a continuous beam. Doesn't happen.

 

[dccd]

I don’t think the distribution of load to other support is effective , especially when the ECC is small. When the ECC is large , maybe another additional support can take more load. However, when the ECC is small, more load will still concentrate on the old support. So, how does adding a tie beam effective when the ECC is small?

 

[1503-44]

That's what a number of us have said already. Tie beams are really only good for distributing lateral loads from one pile to others nearby. That's the only thing I've ever used tie beams between piles for.

Using a tie beam to carry vertical loads would require a (column) load somewhere in mid span, just as would on a typical beam somewhere in a frame. If you load the supports, basically you do not load the beam.

 

[dccd]

Even when there's tie beam, the load will ultimately goes into the support as well. The amouht of load goes into the support depends on the load location, if the ECC is small, the load is closer to the old piles (which overloaded), adding a new tie beam to another pile wouidnt allievate the overloading of old pile , right ? Unless the ECC is very large, adding a new tie beam would help to distribute the noticable amount of load to the new pile,right ?

 

[1503-44]

Theoretically yes, but if that's going to be less than significant, why bother with the obvious, unless you absolutely have to.

 

[bones206]

dccd - If I had to guess, the person who told you, "when the pile is overloaded, the tie beam from to the next pile cap can help to overcome the problem", was referring to structural redundancy. In an area with poor and/or variable soils, the likelihood of any individual pile failing or performing poorly increases. So having a tie beam/grade beam can help redistribute the load to adjacent piles, minimizing the consequences of the overloaded pile.

 

[dccd]

So having a tie beam/grade beam can help redistribute the load to adjacent piles, minimizing the consequences of the overloaded pile.

I also having the same thought, but the adding a tie beam is only effective when the ECC of pile was very large , right ? Meaning that if the pile is overloaded, the column load must be close enough to the 'new pile' so the loading on the overloaded first pile can be mitigated effectively.

If the ECC is small, means the column loading is still closer to the overloaded first pile, adding a tie beam to a 'new pile' wouldn't help much, am I right ?

 

[bones206]

Here's an example using RISA:

- 3 equivalent piles with 50 kip load (red arrows)
- 24"x18" continuous grade beam connecting the 3 piles (density set to zero for this example)

Results show, as expected, each pile has a 50 kip reaction (green arrows):

and settlement is uniform:

Now say the middle pile has an unanticipated problem that reduces it's effective stiffness. Reducing the stiffness by 50% in the model, we can see the adjacent piles are taking up the load that was shed by the "softer" middle pile:

there is some differential settlement, but not terrible:

The grade beam is necessary to distribute the load out to the adjacent piles. Without the grade beam, the settlement of the middle softer pile would be: (50 kip) / (5 k/in) = 10 in, while the settlement of the outer piles would remain at 5 inches.

 

[1503-44]

There's something that still bothers me about designing tie beams for distributing loads across weak piles. Is this something that is done on a routine basis? Whenever you have piles, you just assume piles will settle too much and you design the beams to "span over" a potential weak pile, like bones206 has done there? Thanks for the example bones.

I would also guess that nobody knows which pile might turn up weak, so you have to do the same for all tie beams, even the ones on building corners, not only those in the middle. Tie beams connecting corners would be a kind of semi-cantilever, load at the end partially supported?

It also occurs to me that making the tie beams larger, add extra load to the pile, especially one in the center of three that would take 5/8ths of the extra load. Guess the extra load wouldn't be a lot, but any additional load to a suspected weak pile isn't ideal.

I'm asking because I used to design (long long time ago) a lot of piled foundations for refineries and chem plants, all built on piles in poor soils (800 psf surface loading limits) along the Mississippi River outside New Orleans and we never designed any tie beams to redistribute weak pile loads in this manner. Many foundations for settlement sensitive equipment and structures. I'm wondering why we didn't do that. It was a major EPC company specializing in refineries and chem plants.

Bones206, One other question if I may. What span lengths are you using between piles? I didn't see the dimensions. And I assume the tie beams are 24" high, 18" wide?
If I include the weight of the tie beam, the center pile load is 43% higher at 48K.
Does that sound right? That's the part that seems counterproductive, but that is from the entire tie beam load, not just the extra bit you need to add to make it stiff enough to span.

I'm trying to work the same problem in Excel.
I still have a bit of a difference from your Risa results, but its getting very close.
20ft and 30ft spans
Pile A Pile B Pile C
59.9K 33.5K 56.6K
5.99in 6.70in 5.66in

 

[geotechguy1]

I thought it was just about making the foundation system more rigid to reduce differential settlements. Load redistribution would also occur - if the foundation elements are tied together with beams then the stiffer parts of the foundation will always end up picking up more load.