Post Installed anchors to an existing concrete column

[stevenspm]

Hello all,
I am hoping some of you may shed some light on the concrete breakout strength equations of ACI 318-14. The background is we are trying to connect a new steel beam to either an existing 24" wide x 11" deep concrete column with a min of (4) #11 vertical bars and #4 ties spaced at 11" o.c. or to a 24" x 24" concrete column with a min of (8) #11 vertical bars and #4 ties at 22" o.c. The building was built in the mid '70s. The load on the connection is 121.2 kips factored. I am looking at (2) rows of 1" dia anchors spaced @ 8" apart with the anchors spaced 13" apart within each row, please see attached diagram. The load in parallel to the edges of the column. I have used both Simpson's Anchor design program and Hilti Profis program as well as going through the calcs by hand. For the most part I match both programs, expect for the loads used when comparing the concrete breakout strength in shear. Hilti is using the full load on one row where as Simpson is using only 1/2 the shear. The ACI, for concrete breakout, states to look at 2 conditions, one using the shear split evenly between the rows of bolts and the second to look at just the back row of bolts. The issue is that ACI does not make reference to when you have 2 equal edge distances parallel to the direction of shear. So when looking at trying to design this connection how you go about it?
Thanks for your help in advance,
Phil S

 

[canwesteng]

You need to determine for yourself what is correct. I believe ACI is unconservative for cast in anchors, as the slop in the base plate holes can be taken by any pair of anchors, with no guarantee the anchors nearest to the edge do not fail first. Once the concrete is broken you have compromised tension and shear capacity of the group. For post installed anchors, the holes can be much closer in size to the anchors and I'm comfortable dividing the load over 4 anchors. If you sketch out the breakout condition in this case it should be no different than the breakout condition for two anchors furthest away from the edge.

 

[hokie66]

That is a lot of load to try to take on bolts into a concrete column, and also a lot of load to add to the column itself on its face. I think you need another column. And what happens below, with the footing or whatever?

 

[stevenspm]

Thanks for the responses so far. It looks like I may need to add a bit more information. The beams are being added to reinforce and existing pre-stressed concrete waffle slab system that is showing signs of failing. The only weight we are adding to the system is the weight of the steel beams. We are not able to add more columns as that was done during a previous repair and is disruptive to the use of the building, which is a trade school. I was more interested in methodology of checking post installed anchors when the load is parallel to 2 edges each with the same edge distance to the nearest row of anchors. I have been discussing this with my colleague and we are leaning toward using 1/2 the shear on a row of bolts as the edge distance is the same.
Again thanks for you thoughts.
Phil

 

[Leftwow]

I have never encountered a situation of this magnitude for a retrofit for a concrete to steel connection. If you are slapping an endplate on the end of the beam of that connection you should be designing this as a fixed connection and it will transfer any moment present in the structure to that column. That connection will transfer all of the forces, it is not pinned by any means. If you want it to be a shear connection I suggest welding a shear plate onto the end plate, then bolting the shear plate to the beam. Your connection is very stiff, also check the maximum bolt spacing in a steel connection. AISC states if you space out your bolts too far, then you can have corrosion issues. For breakout you are using the correct method if you are using appendix D for ACI 318. I would consider a pretty in-depth analysis of this considering the amount of force applied to your column. I would definitely use hilti bolts for this, we have software (which it looks that you are using this) and a design guide. but I would call their technical representative too. For double checks. Also a steel check from AISC design guides would be advised when designing any connection that isn't typical.

 

[hokie66]

I know you just want help with design of the connection, but I think you need to be fair to us by providing more information about this prestressed waffle slab which is "showind signs of failing". What kind of signs? Are these beams just a safety net, or will you be jacking between the beams and the slab? At the columns where you are proposing to bolt the end plate to the concrete, how does the top of the end plate relate to the waffle slab? I can't visualize how there is a place for the plate to fit.

I suppose this floor is unbonded PT. If the US were to ban it, like a lot of other countries do, it would save a lot of headaches.

 

[ProgrammingPE]

There's technically 4 cases that should be considered for shear concrete breakout for your condition.
1. Load in vertical direction divided evenly amongst all anchors.
2. Load in vertical direction only on top two anchors.
3. Load in horizontal direction divided evenly amongst all anchors (Vcbg is multiplied by a factor of 2).
4. Load in horizontal direction only on furthest row of anchors (Vcbg is multiplied by a factor of 2).

You could make an argument that cases 1 and 2 could never cause shear breakout since there is not an edge distance in that direction because the column is continuous to the foundation (ca1 is infinite). Also, shear breakout fails due to tension in the concrete and load in that direction does not cause this, plus there is developed reinforcing on both sides of the breakout plane (which would be the entire area of the column) which D.6.2.9 permits you to use in lieu of concrete breakout strength (must still check the capacity of the reinforcing).

Cases 3 and 4 are required per D.6.2.1c (this is where you get a factor of 2 on Vcbg). Fig RD.6.2.1(b) shows the two assumptions that are required for multiple rows of bolts so you must check half the load in both rows AND all the load on the furthest row. You are checking a column so you're right that there's two edges but only one is relevant in each horizontal load direction so that should not affect your results. For Case 3 you would use half the load with the breakout area from the row of bolts closest to the edge. For Case 4 you would use the entire load with the breakout area from the row of bolts furthest to the edge. Case 4 will likely control the design since it is twice the load with only a slightly larger breakout area.

As Leftwow mentioned, this connection will likely be transferring moment so remember to account for that. If adding a plate and bolting to it in order to reduce the end moment, I would still include the eccentricity from the column to the bolts since a small amount of moment will have a large effect on your anchor capacity.

http://www.structuralcentral.com

 

[ProgrammingPE]

I threw your model into PROFIS since I was curious how it was addressing the edge distance issue. It looks like it is checking for load in the horizontal direction with the row nearest the edge taking all the load. You can verify that this is the check they are using since c[sub]a1[/sub] is 8" and A[sub]vc[/sub] is based on the nearest row (11" * (5*13" + 2*1.5*8") = 979in[sup]2[/sup]). For this check, the load should be cut in half since the further row would be taking the other half of the load. They do appropriately add a factor of 2 for the capacity per D.6.2.1c by creating what they call a Psi[sub]parallel,V[/sub] factor of 2.

Their check appears to be slightly conservative since I think they should be checking the furthest row for all the load which has a c[sub]a1[/sub] of 16" (which only increases the capacity by 27% since your breakout area is already through the depth of the column).

So I guess my conclusion is just that the result from PROFIS appears to be only slightly conservative. I would definitely NOT be doubling the values from the program like it seems you are suggesting since ACI 318 pretty clearly states that the furthest row must be capable of taking the entire load itself and that only increases the capacity from the PROFIS results by 27%.

Also, if you are attaching another beam to the opposite side of the column, it would cut your breakout capacity in half since the beam on the opposite side is also using up the column's breakout capacity.

http://www.structuralcentral.com

 

[stevenspm]

Thank you for your replies and ideas. This project is still in the very early stages / preliminary. We are not connecting the beam to the plate as shown in the detail I was just to lazy to take it out before posting it.
ProgrammingPE - I was not not thinking about doubling the values, I was more curious as to why Hilti and Simpson use different applied shear values when checking the concrete breakout. I agree that Hilti seems to be conservative and that it should be applied to the back rows. What my colleague and I were discussing was since the back row of bolts is also 8" from the back edge shouldn't the applied shear be split in half, as Simpson is doing, when checking against the breakout strength. As for the second beam, yes there will be, but I was only checking for a little less than half the column depth.
Thanks Again,
Phil

 

[WARose]

One thing I would like to note (as someone who has retrofitted concrete structures myself) is: you are going to need to budget for some type of testing to locate the existing re-bar in the RC column. (GPR is pretty popular now. There are some EM methods too.)

Without that......you are risking the people installing your Hiltis cutting the column reinforcement. Especially with the stirrups, its never where you think it is.

You've probably already thought of this.....but I wanted to mention it just in case because it's one of those pop up costs that people get mad at you for not anticipating.