Cantilevered Retaining Wall ACI 318-19 Hooked Bar Development

[jdgengineer]

I wanted to see how people are addressing the ACI 318-19 provision which no longer allows reductions in hooked development length based on excess steel. Two examples we commonly run across are on residential cantilevered retaining walls where forces can be somewhat significant, and with temperature and shrinkage steel at wall corners.

Where steel reinforcing demand is not directly quantified (say T&S steel) are you putting in reduced hooks and ignoring the provision as the demand is very small?

Where steel reinforcing demand is quantified (say at the interface between the stem and footing of a retaining wall), what are you doing? Really thick footings if steel reinforcing exceeds #5 or #6 bars? I know an alternate (and perhaps a more justifiable approach) is to use the strut and tie method utilizing curved-bar nodes to determine the radius of the bend that is required for the compression strut to develop the tension steel. However, this often leads to a radius that exceeds a standard bar bend. We could specify a higher radius, but for a residential project I'm not sure how likely it is that it will actually be installed that way.

How would you get this joint to calc for a column supporting a thin flat slab with minimal demands on the hooked dowels?

 

[KootK]

Where steel reinforcing demand is not directly quantified (say T&S steel) are you putting in reduced hooks and ignoring the provision as the demand is very small?

I don't sweat the anchorage of T&S reinforcement. For a wall corner, I'll run the inside face bars over to the outside face and hook them there and call that good. Just standard detailing.

Where steel reinforcing demand is quantified (say at the interface between the stem and footing of a retaining wall), what are you doing?

I usually just mimic what the CRSI manual shows. In the absence of an STM to prove that it worked, I would have always been uncomfortable with a footing too thin to develop the yield strength of the incoming verticals via hooks. As you probably know, I view this as a problem of transferring bar tension around a corner rather than one of development.

How would you get this joint to calc for a column supporting a thin flat slab with minimal demands on the hooked dowels?

Those joints invariably suck. And usually have the hooks turned the wrong way anyhow. I hook outwards like the contractors want and call those joints pins in my modelling efforts with respect to slab bending and deflection. I fix the joints to check punching shear and the columns.

 

[jhnblgr]

A thick footing is often required to meet the overturning moment and sliding. Also shear of the heel of the footing is often quite large. I had an engineering firm provide a design for a wall to substitute a precast one and it was clear that they never checked their work against any standards. There were numerous items that were not accounted for and I sent it back for them to check their work. They then used a computer program with out any idea of what they were doing and gave me something that was more footing than wall. I gave up at that point since what they provided would work just super overkill. ACI has a design guide and there are a few new books based on the 2019 code.

 

[jdgengineer]

Thanks you guys. And yes KootK I knew your position and was hoping you'd comment

In our market, basement slabs are often relatively thin "mat slabs" that also serve as the footing for the retaining walls. Commonly in the neighborhood of 12-18" thick. Contractors typically want to keep the matslab uniform thickness to simplify excavation, rebar and waterproofing so this joint tends to drive the thickness of the slab. We've often used 14" slabs with small reductions in hooked bar lengths for #5/#6 bars but I don't think that will be feasible anymore. The walls are typically designed as cantilevered so they can be backfilled before the floor is in place (and would also behave as one at lightwell conditions)but they end up functioning somewhere between a propped cantilever and a free cantilever once construction is complete.

 

[jhnblgr]

Ok so you are more talking about foundation walls. Like you say they are braced up top by the floor system. Also the basement floor slab provides stability to the footings. I typically like to specify clean stone backfill to reduce soil pressure and have better drainage. In my experience typically water pressure or soils on slopes will cause failure. Any long wall should have pilasters to help keep it straight. One wall failure I saw was a frost wall that pushed out like 2 ft from lack of steel into footing. The rebar seemed to have failed in shear first rather than pullout. I think too often people try to exceed the spacing that is logical and that is the cause of failure.

 

[Celt83]

jdgengineer:
You may end up finding this to be a bit self-resolving once you get into the new shear provisions for the mat design, we are finding on average ACI 318-19 results in about 50% of the shear capacity that you would get using ACI 318-14. Running a real quick one-way shear check against CRSI's spread foundation tables was finding anywhere between 4" up to 14" uptick in thickness to satisfy just one-way shear under ACI 318-19.

 

[KootK]

So the "right" answer is, of course, to figure out what your peers are doing and replicate it. Hence this thread.

I thought of an idea in the shower this morning that might provide a more technically satisfying solution however.

I know an alternate (and perhaps a more justifiable approach) is to use the strut and tie method utilizing curved-bar nodes to determine the radius of the bend that is required for the compression strut to develop the tension steel. However, this often leads to a radius that exceeds a standard bar bend.

Tru dat. However... that large radius is associated with a tension demand assumed to yield the bar, right? I propose that one could provide excess reinforcement such that tension in each bar would satisfy the STM using conventional bends. Basically, for any bar size bent conventionally, there should be some tension force at which the conventional bend radius would not crush the concrete inside the bend.

I would argue that doing the "excess reinforcement" dance in this context would still be street legal precisely because this, fundamentally, is not an anchorage or development problem but, rather, a problem a dragging bar tension around a corner.

How about that? For once, maybe all of my technical fussing can produce something of practical value. Contrary to common perception, I do strive to be genuinely helpful from time to time.

 

[AaronMcD]

I always provide full hook development. #6 can be developed in 14", (10" with 4" cover, hooks tied to bottom reinforcing on 3" chairs). Or just use #5 at closer spacing if the footing is 12". I need the thickness anyway for the mat slab/footing bar hooks anyway if not using U bars/ties. To develop the slab bars, I need a bit of a heel beyond the wall. If they have to form the wall they like the heel anyway.

http://www.hellodwell.design

 

[jdgengineer]

Thanks AaronMcD, turns out my hooked development length assumed a confinement reinforcement factor of 1.6 which likely doesn't apply to these conditions as s>6db. Oops.