Cantilevered Retaining Wall

[RyUIUC]

I am designing a Cantilevered Retaining Wall using the CRSI 2008 Manual as a reference. I have used the tables at the back of Chapter 14 as a starting point and do all the checks by hand. One comment I got from another engineer checking my work is that the vertical stem bars do not properly develop the ldh into the base slab. For example, if a #7 bar is specified for the stem flexural bars and these bars extend into the toe of the base, the base slab should be thick enough to develop ldh of the #7 before it turns into the toe. The CRSI manual does not seem to check this and their tables often specify bars that could not develop this length for the corresponding base thickness'.

I tried to rationalize with the checker that this bar is not really a 'hook' and that it is actually checked to fully develop beyond the face of the stem into the toe thus is actually being spliced into the base reinforcing.

Does anyone have an idea if the ldh really should be developed by these 'O' bars before the 90 degree turn into the toe or a good reason why its not necessary?

 

[ishvaaag]

This one informs that the steel force is almost dissipated into the concrete at 15 diameters from section of application of the force to be developed. That said, hooks and 90º bends become favourable to shorten the required thickness for effective anchorage of some tensile force.

Since this is just a study on some tests, it is not meant to justify any reduction on any development length as required by the code (it is well known that codes ask for development lengths 2 to 3 times the rupture values).

It, however, explains why in some of the many cases where development lengths are taken mistakenly or purportedly short from code values (as could be for rebar nascent from relatively thin slabs) are not generally showing evident distress.

http://www.vncold.vn/modules/cms/upload/10/New folder/vca/Proceeding/Session B/B06.pdf

 

[TXStructural]

From the OP:
"Does anyone have an idea if the ldh really should be developed by these 'O' bars before the 90 degree turn into the toe or a good reason why its not necessary?"

The horizontal portion of the "O" bar is developed before it reaches the face of the wall, the location of maximum moment in the toe. In the current Design Handbook (2009), page 14-11, under "Toe": "If the 'O' bars ... do not have enough available length for the tension development of the bar, then the bars must be hooked or headed."

The issue should not whether the bar has been developed for axial strength of the bar. The bar turns vertical and remains developed, and the vertical portion develops fully into the wall above the critical section at the base.

This condition is similar to a fully-developed corner bar through a wall to wall joint, or a bent bar which extends up from a wall and into a slab. Detailing Corner in the Sept 2009 Concrete International, and Detailing Corner RFI 09-3 in the November 2009 issue, discuss reinforcing wall intersections which need to transfer significant moment. The same discussion is relevant for the base-to-wall intersection of retaining walls.

There can be an issue with the strength of the embedment at the depth of the bar detailed. I agree there should be no question where this depth exceeds ldh, but several of the values in the CRSI Design Handbook are less than ldh. In the case where the depth from the base of the wall to the bottom of the hook is less than ldh, but greater than the value shown on page 5-23, Table 5-7 (2 inches above the tail of a code-compliant 180 degree hook), engineering judgement should be used. The CRSI wall table values should provide at least the required Table 5-7 values. Also note that in the last paragraph of the commentary to ACI 318-05 12.5, they say that the "minimum value of ldh is specified to prevent failure by direct pullout where a hook may be located very near the critical section." There is nothing else with reference to the depth of concrete required to restrain a bend in a fully developed bar. Lacking more detailed analysis, a compression strut can be postulated to exist between the inside of the bend and the compression zone at the face of the wall (and the compression face of the toe), with ties provided by vertical "O" and horizontal "P" bars.

The concrete around and above the "O" bars is confined and restrained by the wall above, and by "P", "D", and longitudinal bars. These bars act similar to supplemental anchor reinforcement discussed in ACI 318 Appendix D for anchors, crossing the notional shear planes/cone.

I will see if a strut-and-tie and/or other analysis can be provided in the Design Handbook revision to cover this kind of question.

 

[JAE]

TXStructural - everything you state may be true - I'm not sure I can use references to refute what you are implying - but my little picture's worth a thousand words in describing how I think about it....as for me and my designs....I'll require Ldh.

 

[Teguci]

Yup, bottom slab needs to 15" thick. Has anyone contacted CRSI about this?

 

[hetgen]

Respect to all clever people posted above, but the joint between the stem wall and the front base slab is a typical closing corner discontinues joint. The vertical bar could be developed into the front base slab as long as the bearing inside the bent bars is not excessive, it may be desirable to increase the radius of this bend.

If the stem wall is thicker than the base slab the tension force in the base slab steel will be larger than in the stem steel, to avoid what JAE noted above stirrups might be required on the base slab. See attached strut and tie model that account for change in bar force in such a region. (Source RC book- by Mac Gregor)

 

[hokie66]

Teguci,
TXStructural is our own CRSI rep.

 

[TXStructural]

First, thank you to all who contributed to this thread. It has resulted in identifying some information that may improve our retaining wall chapter in the Design Handbook and other publications. The material below is my reading of an article or two on some research, and has not been vetted through the Institute. It will be placed into our formal process for editing and review prior to a CRSI Technical Note or article being developed.

Nothing I have found conflicts with the information presented in the CRSI Design Handbook chapter on cantilever retaining walls.
--
Research was conducted into wall corner joint reinforcement, including specific work on retaining wall reinforcement details, at Chalmers University of Technology in Sweden, from 1965 through 1973. A summary of this was published in the ASCE Journal of the Structural Division in June 1976.

Specific reinforcement detailing recommendations for retaining walls included:
- Area of main reinforcement is “designed on the basis of the moments and normal forces” in the adjacent members, ignoring inclined bars.
- Bars are placed such that they are as close to surfaces as cover requirements permit, and extend to the ends of the heel and toe, preserving the same cover.
- Area of steel of inclined (“D”) bars should be approximately ½ the larger area of main reinforcement.
- Toe length should be the same or longer than the wall stem thickness, and should be long enough to provide anchorage of reinforcement. Bottom reinforcement (“O” bars) should extend continuously through the length of the toe.
- Observe minimum bending rules and spacing limitations for bars.
- Inclined bars (“D” bars) are used to limit corner cracking through the joint at the back of the wall. These are not strictly required, but will improve performance and increase the capacity about 25%.
- Reinforcement percentage should not exceed 1.8% (fy=57 ksi) or 1.2% (fy=80 ksi). (Based on 4300 psi concrete cube strength; the article says that other values may be interpolated and extrapolated.)
- Bars are to be shaped and spaced so that the design is readily constructable and “concreting is possible.”
--
Based on a read and re-read of this summary article, it appears that it is unnecessary to provide embedment to the depth required to develop a standard hook, the issue was not mentioned, but there is a requirement to develop the bars in the toe before reaching the critical tension region (red in the attached sketch.) I am attempting to get a copy of the full research report and see if it is addressed.

I have attached a sketch that shows a standard reinforcement layout and joint (taken from the CRSI tables), and on that I have superimposed the truss idealization and an approximation of the location of the concrete tensile stresses, as taken from the report mentioned above. The research summary discussed a theoretical model where walls and footings were the same thickness (said to be more easily computed.) The research experiments were conducted on samples with 8 inch (20 cm) thick walls and 10 inch (25 cm) thick footings. (By observation, the "d" value for the wall stem and footing are approximately the same due to increased cover in the footing.) Similarly, the CRSI tables typically list a footing thicker than the wall stem.

(Reinforced Concrete Corners and Joints Subjected to Bending Moment, Nilsson and Losberg, Journal of the Structural Division, Vol. 102, No. 6, June 1976, pp. 1229-1254)

 

[hokie66]

TX,
On a bit of a different subject, I hope you also emphasize what Nilsson says about opening corner joints.

 

[TXStructural]

Hokie, since it is a serviceability and durability issue, I'm sure we will cover that.

Hetgen, ironically, the article I discuss above says that a primary thing to avoid is ties/stirrups in the footing for constructability. I figured that was one reason they looked strictly at footings which were thicker that the wall stem. Also, the research found that as long at the toe was wider than the thickness of the wall stem, and the "O" and "P" bars are properly developed, the behavior was different from a regular "L"-shaped corner.

 

[hokie66]

TX,
It is also a strength issue for opening corners. Nilsson says that the typical arrangement (with bars in the inside faces going across and bending into the outside faces) is only 10 to 30% efficient. With the European practice of using hairpins, 80% efficiency can be achieved. For more, it is necessary to provide stirrups, including on the diagonal.

This is mainly applicable to rigid concrete frames, and to wall corners in concrete tanks.

 

[Teguci]

Just want to have a go from the legal standpoint:
Per ACI 318 12.1.1, we are allowed to consider development as a combination of hook, length and mechanical device, but, per 12.12.1, negative moment reinforcement (appropriate for cantilevers) must be developed "into or through" by only one of these methods. In our case, the bars are fully developed by length "in" the toe (I am concerned with ACI's use of "into or through" since it implies a limitiation of orientation). Also revealing, 12.10.1 allows tension reinforcement to be fully developed by making it continuous with reinforcement on the opposite face of the member. This is what we are doing but doesn't necessarily jive with 12.12.1.

From an engineering standpoint, I am still concerned. If we are OK with developing the bar into the toe could we get away with #11s instead of #7s? What becomes our basis for footing depth? How do we verify the confinement of the hook in the footing?

 

[TXStructural]

Hokie, I agree with your observations, including that special detailing and "D" bars apply most directly to corners rather than "T" intersections. I did not discuss the case of a minimal toe or "L" configuration, since these are not typical in cantilever walls, and must follow different design schemes. The article I mentioned covers these in some depth. Do you design any "L"-configured walls?

Teguci, ACI language is "shall be anchored in or through..." which says that a bar, once developed remains developed. We are developing the "O" bars in the toe, then making a bend to vertical. In the original JAE sketch, there was a "block failure" as the whole bar and compression strut rotated and lifted. The "P" and "D" bars should prevent this action.

Interestingly, in ACI 318 Figure R12.21 (a), you would likely get better frame action by providing full development in the vertical before turning the bar into the beam. This would more correctly engage the beam reinforcement with the column reinforcement. The fact the we do not see problems in the short hook configuration demonstrates how conservative our design requirements are.

To your point about bar proportioning, I think you are correct that a poorly-proportioned reinforcing layout could be a problem, and it might be useful to have some logical limits, (i.e. As= 1.56 sq. in. = #8 @ 9" v. #11 @ 18", in the same wall/footing combination.)

I will certainly discuss this with a few authorities on development and bond at the upcoming ACI convention.

 

[JAE]

TXStructural,
The "D" bars do not cross the block shear plane so I doubt that they will help in the failure mechanism of my sketch. They do add to the As tension reinforcement though.

The "P" bars would have to work in shear friction and be fully developed on either side of the block plane. Right now, the CRSI manual only requires the "P" bars to extend ld from the back face of the wall while the shear block plane is located perhaps mid-way across the stem thickness. A subtle difference but one that should be considered.

Also, the "P" bars, if looked at as a type of shear friction reinforcement, are located only 2" clear from the loaded face side which to me is questionable in their efficacy to do the job.

In your further research and discussions you might consider the above. I'm curious to know how the individual ACI requirements (hook development, shear friction behavior, etc.) can all be technically violated yet still work in a retaining wall as you suggest.

Keep us updated - thanks.