Understanding ACI 318-19 - Rebar as Anchorage 2

[ChickenBake]

I have a couple of questions on one of the provisions that has been added in ACI 318-19. The clause on 17.1.6 states:

Reinforcement used as part of an embedment shall have development length established in accordance with other parts of this Code. If reinforcement is used as anchorage, concrete breakout failure shall be considered. Alternatively, anchor reinforcement in accordance with 17.5.2.1 shall be provided.

The commentary to the same section states:

As an alternative to explicit determination of the concrete breakout strength of a group, anchor reinforcement provided in accordance with 17.5.2.1 may be used, or the reinforcement should be extended.

It is common in my world to specify embed plates with DBAs to facilitate connections between steel beams and the ends of walls or columns. The design philosophy is that the DBAs are extended a full development length into the wall and the connection is designed using the ACI Chapter 22 shear friction provisions for concrete against steel. Chapter 17 limit state checks are never done. If concrete breakout was checked, this would almost never pencil. See attached for an example detail.

Questions are as follows:

1. What is the difference between reinforcement as "part of an embedment" and reinforcement "used as anchorage?" If they're the same, why does the clause use two different terms?
2. If I am using the shear friction provisions, do I still need to check concrete breakout?
3. Why does the code suggest extending the reinforcement as a solution instead of calculating the breakout capacity of the group? Extending the bars beyond development length would have no effect on the calculated Chapter 17 shear breakout capacity, and if you were talking tension, you wouldn't be able to determine what that extension should be without actually doing the calculation.

I am curious if other firms have encountered this issue yet. The internet appears to be mostly silent.

 

[ANE91]

1. Up to 20d embedment as an anchor (Chapter 17); up to 60d embedment as “part of an embedment” (Chapter 26). Basically, you can treat a post-installed rebar the same as a cast-in rebar, under certain circumstances. https://www.hilti.com/engineering/a...-post-installed-reinforcing-bar-design/xbcwum
2. If over 20d, then no. But, testing has shown that groups of rebar close together can and do still fail the concrete in breakout when the ratio of steel is high.
3. The whole point is for when designers are strapped for embedment. Consider embedded top-of-slab facade attachments. So ACI says either confine the anchorage or extend the bars. You’re SOL if space is tight. Rarely, you can justify shallower embedments with Chapter 17 and still get full development.

 

[ANE91]

Here’s another article re: shear-friction https://www.hilti.com/engineering/a...method-for-post-installed-rebar-design/63jbgp. TL;DR: Palieraki method for shear-friction without full bar development.

 

[canwesteng]

I read it as full development is needed for shear friction - additional length of rebar provided = length until the bar is fully developed and shear friction can be used. Until the bar is fully developed, it needs to be checked as an anchor.

 

[TRAK.Structural]

This has always confused me as well.

Consider a reinforced concrete flexural element cast in separate pours. The flexural capacity of the section is a function of the bars reaching the code prescribed level of stress. There is no check for some sort of breakout of the bars in tension. Hopefully others can point out the complexities and nuance that I am missing, but I've never really understood why development length embedments would require a breakout type check.

 

[ANE91]

Consider a reinforced concrete flexural element cast in separate pours.

Here, concrete would split before you get classical breakout, because the long. bars hug the tension face (i.e., too close to an edge). Stirrups preclude this by wrapping around the bottom bars, effectively acting as “anchor reinforcement.” It’s an indelicate but practical analogy.

Chapter 17 is relatively over-conservative compared to the rest of the code. The reasons and consequences of this conservatism are:
1. Failures in the case history tend to occur at connections / joints.
2. These failures are often brittle.
3. Connections / joints typically have no load-path redundancy.
4. 5% fractile is used for nominal concrete breakout strength.
6. The behavior of an anchorage is still not 100% understood. For example, the code (2019) does not acknowledge the beneficial, confining effect of flexural compression (e.g., under one edge of a baseplate) on the breakout cone.
7. Strengths often calculated assuming cracked concrete, even when uncracked concrete may be justified.

The concrete itself doesn’t distinguish between a rebar and an anchor except for the embedment and the head conditions (e.g., 180 hook or nut/washer). The line drawn between the two, by the code, is a somewhat arbitrary 20d. Indeed, development can occur sub-20d (refer to Tanaka and Oba [2001]), and breakout can occur beyond 20d. So whatever Chapter you’re using, be mindful of the physical behavior, first and foremost. The competing mechanisms are ultimately bond strength, breakout strength, and steel yield.

Read Silva’s work on the matter. He’s an authority on Chapter 17. One of many papers linked below.

https://www.acifoundation.org/Portals/12/xBlog/uploads/2023/10/18/SJ_September_2022_V_119_No_5.pdf