Composite Behavior of Adjacent Concrete Beams

[msrrr77]

Here is a hypothetical problem. I would appreciate any input to support or reject the concept.

There is an existing concrete beam (say 5'W x 5.5'D) supporting concentrated loads from both sides of the beam. Now we need to relocate/ divide these concentrated loads beyond either edge of the beam. In order to do this, we are adding 2.5'W x 5.5'D sections on either side of the existing beam. So in plan view this modified configuration will have 2.5'W new section, 5'W existing section & 2.5' wide new section. These new concrete sections will be connected to the existing section with dowels and transverse post-tensioning (straight). Please note that all the concentrated loads (previously on existing section) will entire fall on these new concrete sections. Also, another important criteria to note would be that we are not altering the support conditions. Thus, the new beams are connected to the existing and only the existing beam is directly supported by the columns.

So, the question is, can we consider this arrangement (3 sections stitched together with dowels & PT) as a composite section?

Please consider the fact that the existing section do not have adequate shear capacity to support the loading condition alone. The arrangement will work if we assume composite property and count the shear capacity of the new section near supports.

I will sincerely look forward to your responses.
Thanks.

 

[TXStructural]

If you are not changing the support condition, and not connecting the new beams to the column or another support, how will you increase shear capacity at face of support? Just because the critical section is nominally away from the face of support, for design purposes, does not mean that the existing beam will not fail in shear at a weak plane along the face of support.
Or, am I reading your question wrong?

 

[msrrr77]

TXStructural..Thanks for your reply..

I totally agree with you.. I too think that it will not be prudent to assume composite section just by connecting the new to the existing with dowels/ PT. Because ultimately the load path to the supports is through the existing section only.

However, it has been argued otherwise. I will definitely wait for any reply that can agree with this concept and would like to know the plausible explanation if any.

 

[hokie66]

With enough transverse PT, you can certainly make the three sections work together. But for the shear to be shared into the support, the two new sections must also extend into/through the support.

 

[dcarr82775]

To play devil's advocate, if you assume the new beams are 'adequately' connected to the old beam, and the controlling failure mode is beam shear, then why wouldn't it work? Sure the load has to get from the new, through the old, and on into the column, but the old beam can't simply shear off without also shearing off the new beams. One heck of a lot of thought needs to go into the detailing between the beams, and maybe you can't get enough of a connection to transfer that kind of load, but in theory I think it could work.

If somehow a punching shear failure mode is what controls then certainly the added concrete does not good.

Given the size of the members in question I imagine the loads are substantial. I probably wouldn't do it in my own design, but I think it could work because you can't break one without the other. Of course I could be wrong and if so I don't doubt someone will put me in my place

 

[JAE]

I agree with hokie66.

The idea of "composite" action doesn't really apply here - the three separate beams all can act to share the load if they are adequately tied together to transfer the outer load application into the center beam.

There is not "composite" action here in the traditional sense in that the beams are not adding to the flexural capacity through horizontal shear behavior...only transferring a portion of the load to the center beam in proportion to the individual beam stiffnesses.

The shear at the end sections might be mutually shared up to a point near the end, but as you approach the support, with only the center beam being supported, the outer sections must shed the shear into the center beam to deliver it to the support.

 

[dcarr82775]

Assume the side pieces are tied into the original beam sufficiently, which is no easy task. If the assumption is the new side beams do not contribute to the shear strength of the overall assembly at the support, how does the shear failure occur? You can't simply shear off the old beam without also shearing off the new ones.

 

[bkal]

I apologise if I am missing a point here, but have you consider if the new beams are going to see only additional loads applied after they have been installed (unless you preload them somehow). In other words, the new beams are not going to relieve the loads from the existing beam (which it is carrying at the moment) unless you specifically design it.

 

[JAE]

dcarr2775,
Attached sketch - if the two side beams are not attached to the support, they simply slip down with the shear-failed center beam.

 

[msrrr77]

Thanks to everyone for your opinions.

"dcarr82775" - You have correctly assumed that these concentrated loads are substantially high similar to bridge beam bearing points on pier cap. I agree with you that the new concrete will reinforce the shear capacity of the existing beam. However considering that the existing beam has inadequate shear capacity, if the new pieces are not directly connected to the column it will encounter a global failure as sketched out by "JAE".

The junction of the existing beam and column is monolithic and is crisscrossed by several rebars in all directions. Nevertheless, we need to connect the new pieces at this beam-column junction to provide a load path directly to the column as an alternate to the existing beam only. We will have to use PT and/or steel dowel. Detailing this connection will be essential and critical as this will be the primary load path for this new arrangement. I believe we will have to use the concept of shear-friction to design the dowel and PT. I am also thinking of using stainless steel to avoid future corrosion issues.

Any thoughts on the type of analysis/detail of this connection will be greatly appreciated.

 

[dcarr82775]

JAE,

My assumption has always been with the caveat that some careful detailing goes into the connection between the old and new beams. If the connection is properly detailed than you can not get that crack in the old beam without getting a similar crack in the new beams. I am thinking something akin to a 'C' bar that in doweled into the bottom and top of the existing beam (crossing that inclined failure plane similar to a stirrup) and sticking into the new beam each side. These bar prevent the failure from occurring per your sketch. I wouldn't count on any shear strength from the new concrete, but I think you get to count on the new Vs or at least some respectable portion thereof with proper detailing. I think the detail is more complex than just epoxying in some #4 bars, but I think it could be done.

We have installed vertical threaded rods through concrete beams in the past to serve as additional shear steel. It would be similar to that.

 

[JAE]

dcarr,
Not sure I agree at all with you. My sketch shows that the three beams are completely tied together and if the center beam shears off the column the two side beams simply drop with the center beam because the OP indicated that there was NO connectivity between the side beams and the support.

Now you can argue that this is more like a shear friction failure down a vertical plane at the end of the center beam (vs. a diagonal arching shear failure), but it still can fail that way and the two side beams aren't taking any shear because they have no place to deliver it.

If you tie the three beams with an infinite glue you could still get the failure in my sketch.

 

[hokie66]

msrrr77 has come to the same conclusion as JAE. Now it is just a matter of detailing it. If you can get at the sides of the column as per JAE's sketch, I think stressbars would be my first choice, along with roughening the column face. I wouldn't use stainless steel. Maybe galvanized stressbars, if they are available.

 

[dcarr82775]

JAE,

We will have to agree to disagree. You simply can not get that crack without also failing the added steel in the new beam. The steel in the new beam delivers the shear load across the inclined crack in the old beam. The new concrete gets you nothing, but the steel can get you something if done right.

Again, I wouldn't do it this way. There are certainly more efficient better ways to do it with cleaner load paths.

 

[msrrr77]

dcarr82775;

I like your concept and I kind of agree that this detail will engage the new bars in its shear capacity. However, I am not sure how to quantify (100%, 75%, etc.) the contribution. Any thoughts? Also, I would modify the detail a little bit. I would drill a threaded bar @ top and bottom completely through the existing (and grout obviously) and will use mechanical anchors (to avoid lapping of bars) in the new sections at each ends to provide continuity between the three elements. So, it will be like a big stirrup enclosing the 3 pieces.

But still I feel that it will be difficult to overcome the deficiency of the existing (as per the calcs it will be more than 100% overstressed in shear). So, we definitely have to rely on the direct connection to the column junction, most likely with similar concept along with PT.

hokie66;
You mentioned galvanized stress bars instead of stainless steel. Usually galvanized bars have lesser capacity (about 5% less) than regular PT threaded bars. So, I am thinking if we grout the full length of PT and also the PT anchorage ends (which will be in a pocket), do we still need galvanized PT bars?

Thanks to all for the ideas.

 

[msrrr77]

By mechanical anchors I meant couplers. sorry for any confusion.

 

[JAE]

dcarr - I don't think you understand the point we were discussing above.
hokie66 correctly identified the concern:

But for the shear to be shared into the support, the two new sections must also extend into/through the support.

My sketch was responding to that issue - [blue]whether the shear delivered to the support/column is an issue in the OLD beam since it is the ONLY thing attached to the column.
The assumption in my sketch is that he supporting column extends up between the two new side beams. The two new side beams DO NOT attach to that column.

The rebar you show in your sketch does NOTHING to keep the two side beams from sliding down the sides of the column. Or....what if the two side beams don't even reach the column/support face but stop 1 inch away from it?

The center beam (the ONLY thing attached to the column) can crack through its width and then the center beam with the two attached side beams WILL be able to slide down the column.

The OP specifically said that the side beams are NOT attached to the support.

Thus, the new beams are connected to the existing and only the existing beam is directly supported by the columns.

My sketch is a 3D view so perhaps difficult to visualize but the crack can occur through the center beam and the thing fail despite your [red]Super-glue-infinitely-fastened-completely-attached[/red] side beams.
The side beams just go down with the center beam.

 

[dcarr82775]

JAE,
I also followed the OP's restriction that the new side beams are in no way connected to the column. I may have gone beyond the OP's quote in proposing a manner in which you could gain some additional shear capacity in the existing beam by utilizing the new beams. I assumed that was the ultimate end he was getting at. I took the leap of faith and assumed his question was really, "Is there a way to add to the shear strength to the system if the new beams are only attached to the existing beam and not the column?" So I looked at it from that end, and feel I was very clear on that point.

I don't propose any super-infinity-fastened-smartass-magic-glue as you claim. What I do propose is a plain old simple piece of reinforcing steel crossing an inclined crack functioning similar to every other beam stirrup in the known world. The new concrete gives you nothing, we all agree, but the new steel can give you something if it is anchored into the existing beam such that it crosses the crack. It would function in a similar way to carbon fiber wrapped around a beam and anchored on the top side of the beam to add shear strength (see Fyfe for examples of just this method). IMO, the added steel means the crack you drew can't happen at as low a load as it would if the added steel wasn't there.

Cheers

MSRRR77
My thought process was just kicking around an engineering idea that could work, but like I said I probably wouldn't do it that way. So my sketch was simplistic to just show a conceptual idea. I would probably core vertically through the beam and add vertical rods to act as additional shear steel if I needed a reasonable amount of additional strength. But since you are looking for 100% more than you should just try tying into the column since you may be bumping up against the shear limits.

 

[hokie66]

dcarr,

You are talking about increasing the beam capacity in diagonal tension, while JAE and I were concerned about failure in more or less direct shear at the column face, which is another potential mode of failure. Extending the side beams past the column and clamping them to it (shear friction, some call it) may be a way of addressing that mode of failure.

msrrr77,

I didn't remember that galvanized PT bars had less capacity, but doubtless they would, as you have to allow for the galvanizing in the threads. But if the PT is fully grouted, I think that gives the corrosion protection you need.

 

[JAE]

hokie66 - you are correct in describing the difference in failure modes - dcarr is talking about general bonding of side beams to increase the Vn value of the overall beam.

dcarr - I agree with you completely that with the proper detailing you can get some improvement in the overall shear capacity of the beam along the span where the three sections are attached to each other. No complaints with that.