Shear load transfer at new concrete beam

[StrEng007]

Hey guys, I’d like to hear about your approach to transfer shear loads through dowel connections in concrete. In this scenario, imagine a new concrete beam being poured between two existing concrete columns. The ends of the beam are being epoxy doweled into the existing concrete column.

This is my approach, which leaves some questions up in the air for me. I’ve left the equations out of the steps listed below. These are basically the (3) steps provided by ACI for calculation transfer of horizontal forces (ie shear forces). Steps “b” and “c” essentially go hand in hand since they are used to determine development/bond.

The steps are:
a. Determine minimum shear friction reinforcing requirement. This calculation from ACI assumes the slipping faces of the concrete will place the anchor dowels into tension due to the faces separating from each other (Clamping force).
b. Check tensile development length of the dowel placed inside the new beam.
c. Check the embedment & bond strength through appropriate strength design method provided by epoxy manufacturer. Based on step “a” from above, only the tension needs to be considered. I believe there should be a shear check as well.
d. Check the maximum shear transfer permitted by code based on concrete contact (friction area).
Some questions I’m asking myself:
1. Why does ACI not have a check for the shear transfer through the dowel? Would these dowels not experience some form of shear or combined tension and shear? I understand the faces will slip past each other, so shear stress would be present.
2. Same for the epoxy side. I see shear being present.
3. If you combine the two (tension and shear), what interaction do you use?
4. Does ACI address the edge distance of these load transfer dowels to avoid breakout at the top of the beam (edge with the least amount of concrete cover)?

 

[KootK]

Check out the attached document from Hilti Europe. The snippet below is an excerpt from that publication. We in North America don't really seem to have a great, established method for this.

Before I get into the meat of your questions, I'll say this: while I am a believer in shear friction, this is one of those situations where I'd be looking to chip a bearing key into the columns. Belt and suspenders if nothing else.

So... in my opinion, here's what needs to happen:

1) You probably need to accept that the new beam will be pin ended rather than being fixed to the columns. It's just too damn hard to make a go of that connection with significant moment transfer included. The beam should be designed accordingly.

2) You need to make shear friction work at the joint.

3) You need to anchor the bottom steel into the column in a manner consistent with the "development of positive reinforcement" section of your code. This is the part that often gets missed.

4) With regard to the bottom steel demands resulting from #1 & #2, you can double dip. The demand is not additive.

5) I would add some doweled in top steel for a) joint crack control and b) redundant shear friction reinforcing. The best place for shear friction reinforcement is really the tension zone up top. If it were possible, a fixed connection between beam and column would be ideal. Since this generally is not feasible, the approach to take is to dowel in as many small bars as you are practically able to and ensure that they can develop fy without suffering a bond failure or ripping a breakout cone out of the columns. The strategy, since you won't be able to dowel in enough bars to create full moment transfer, is to make sure that the bars that you do get in there don't fail in a brittle manner upon overload.

Now for your questions:

#1/#2/#3) The idea with shear friction is that the reinforcement itself does not resist load via direct shear stress on the rebar. The rebar provides the clamping; the concrete resists the shear. This is a bit muddied by the commentary for shear friction which indicates that, for smooth concrete surfaces, shear friction essentially is dowel action. This is somewhat moot for you as you should absolutely be roughening your connection surfaces. So no shear in the rebar and no need for a combined tension/shear check. Check.

#4) No, ACI does not address edge distances for shear friction where true dowel action would be at play. One could probably adapt appendix D provisions, or manufacturer's load tables to investigate such a failure but I don't think that is commonly done. I share your concern though. That's why I'll usually provide 100% shear friction transfer capacity in the bottom dowels and then also add as close to 100% shear friction transfer capacity in the top dowels as I can manage.

For what it's worth, there are methods in non-US codes for evaluating true dowel shear connections. To my knowledge, however, we don't have anything analogous in North America. At least nothing "official".

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[StrEng007]

Thanks for the information.

Unfortunately, the beam I'm casting has size restrictions. It's difficult to achieve 100% shear friction transfer in bottom dowels and an additional 100% at the top.

I considered a work around to this. While using a "roughed surface" (u=1.0), 100% of the transverse shear requirement gets transferred to tension in the dowels. I designed my dowel/bond for 100% of the load as tension and applied an additional 40% of the transverse shear applied to the dowel/bond as actual shear (Combined tension and shear). I figured, in the worst case ACI only lets you utilize 60% of that shear friction to tension conversion when the surface is smooth. Here I'm requiring that my connection carry that additional 40% that is "lost", even though it's technically not required when you intentionally roughen the concrete surface.