Grade Beam with elevation changes

[WillyTheG]

We have an 18" x 24" grade beam that's supposed to be continuous with 3 - 5" T&B on top of 18" piles spaced 10' apart. Because it's a hillside with an existing house above us we have an elevation difference of 12-16" at pile locations where the grade beams meet. How can we make that continuous?

Maybe this sounds dumb and I'm hoping it makes sense, but I'm thinking we lap the bottom of one beam to the top of the other, and a Z bar lapping the top of the higher beam to the bottom of the lower.

 

[r13]

No, you can't make it continuous, the grade beams end on 10' pile span. You need to hook bars in each beam at ends, and extend the vertical bars in the piles thru the top of the upper grade beam to make the joint an integral piece to resist beam end moment.

 

[jayrod12]

There's a few ways to skin a cat, but with the information provided it's hard to tailor a response. You could make the gradebeam deeper as required to suit the slope, you could have the gradebeam not be continuous at the step, you could cantilever part of the gradebeam to the junction.

Are you the contractor or the designer?

 

[r13]

Beware of drastic dimension changes in continuous structures, the state of stresses is not very clear. If I adopt Jayrod12's approach, I'll treat the upper grade beam as added weight rather than a part of load carrying member.

 

[RPMG]

Step it at span / 4

 

[KootK]

1) I encounter this situation often on fancy houses on the sides of mountains in Colorado and the sides of cliffs by the see in California. So that's the context for my solution.

2) I support the L/4 solution in many instances but, where my piles are tightly spaced (usually about 10') I find that difficult to make work with the proportions. The joint winds up taking up 1/3 of the span length. So, for this application, I've found it slick to locate the joints at the piles. This also make is easy to specify where I want the steps.

3) In my opinion, grade beams do a good deal more than just transfer loads over to the piles. They also perform the important function of helping to smooth out differential deflection between piles which is usually important given that crap soil is usually the reason for deep foundations to begin with. For this reason, I like to maintain flexural continuity across grade beam steps, at least nominally.

4) I don't treat these moment connections with the same seriousness that I would, say, a special moment frame connection. I take a few liberties to suit the seriousness of the situation and the skill levels that I anticipate with residential contractors. Simplicity all the way & judgement in all things.

 

[r13]

For a smaller elevation difference, and maybe longer beam span, KootK's approach is the prefect/excellent solution. But note the OP has a 18"x24" grade beam, and the elevation difference is 12"-16" (h in KootK's sketch), with 18" dia. pile at 10' spacing. Let's assume development length for #5 bar is 14", therefore the length of the thickened section to the right side of the pile is 14+2*12=38", or 14+2*16=48", it is going to extend near, or beyond the mid-span. Then why not just extend the thickened section all way through?

 

[KootK]

Ln = 8.5' but, either way, I don't see why a 4' long connection should present a problem.

 

[r13]

KootK,

Sorry, I edited my comment (ln = 7') while you were reading.

I personally couldn't accept this detail for this particular application.

 

[KootK]

Can you elaborate on the nature of your objection to the detail?

 

[r13]

Also, if we consider non-contact splice as typically in a continuous beam over sport, the length will further increases.

 

[KootK]

Also, if we consider non-contact splice as typically in a continuous beam over sport, the length will further increases.

The Ld + 2h is the non-contact splice to which you've referred retired13. Nothing additional is required in that regard. You haven't stated what it is that you object to about my detail but I'm guessing that it is the relative proportions of the joint versus the beam span. If so, one could make a defensible argument for changing my Ld + 2h to something as little as Ld_hook + 1h. At that point, it's hard to imagine a shorter joint that would still make sense as a plausible moment connection.

With regard to proportions, our depth to shear span ratio here for the beam alone is already down to 2.125 ((10'-18")/2/24") without consideration of the joint. As such, the beam alone is already technically in strut & tie / disturbed region territory and, therefore, any expectation of Bernoulli flexural response is probably wishful thinking.

 

[KootK]

Then why not just extend the thickened section all way through?

I agree, and have done that in some instances. The one drawback is that, if you need a step for each pile in sequence, there's no way to do this in adjacent spans. You'd basically just be back to the same problem with 24"+16" = 40" beam.

 

[r13]

KootK,

1) For h is small = min[h < Ld/6, or h < 6"], as said before, I won't have problem with your detail.
2) Yes, except the first span, all remaining spans should be thickened to cover the elevation drops, with a step at each pile location.
3) I remain convinced that each grade beam shall terminate at piles. For which, I'll design for ±M = wL²/10, check shear, and call the day. I don't see why to make this problem more complicate.

 

[r13]

One more thing to consider when increase beam depth.

Let, H = Beam depth = 24", Ln = 8.5' = 102", d = 24-3 = 21", Ln = 102/24 = 4.25 > 4 ---> Check shear as normal beam. Otherwise deep beam behavior need to be considered.