Unbraced length of reinforcing plate 1

[StrEng007]

Let's say you have an existing concrete beam, and you attached (1) flitch plate on each side (vertical face) of the beam in order to take the full gravity load on said concrete beam. What do you take as the unbraced length of that plate? The anchors are located at the top 3rd of the plate and are spaced at 12" O.C. along the concrete beam length.

Is the Lb of those plate 12"? I guess it depends on the stiffness of the member on which it is attached, correct?

Replace the concrete beam with a wood beam. Does this now depend on the unbraced length of the wood beam?

 

[KootK]

Is this proposed reinforcement for the shear cracked grade beams of you other thread? I doubt this will be viable but we'll see where it goes...

 

[StrEng007]

No, removing block below an existing tie-beam.

Fairly conservative approach... typically install flitch plate(s) with bolts along it's span, terminate each end with bolts, fully grout the cells at each end with vertical bars (faux columns), then remove the load bearing CMU below the flitch. It creates a redundant beam and the idea is to let the steel plate(s) take the load and weight of beam.

We do this in older existing buildings. The tie-beams usually have some flexural reinforcing, but this alleviates the guess work and is extremely fast. Ergo, what is the unbraced length of the plate when it is only attached to the member that it is reinforcing?

My thought is, as long as the O.C. field bolts are located along the top edge of the plate (compression edge), it doesn't really buckle since the concrete beam itself will have a decent amount of torsional resistance. This is why unbraced length of concrete beams isn't really a "thing" in the same manner as a steel W-shape. For concrete beams, ACI recommends something like 50 x h for beam bracing intervals.

 

[KootK]

1) Gotcha, I've done plenty of this too.

2) When I do this, I try pretty hard to do it with channels if space and envelope detailing permit. In addition to improved Ix/Sx, the channel flanges brace the web against precisely the kind of buckling that you're concerned about.

3) Plate buckling between bolts is, of course, a legitimate concern. I'd be inclined to look at the plate as little K=1 columns between the bolts with effective widths of maybe the lesser of [s/4, 8t, 2x edge distance]. This is surely an underestimate of capacity but, in my opinion, the situation doesn't justify a more refined approach. The stiffness of the thing being flitched surely does matter but not enough that I'd bother to evaluate explicitly. I've seen some people check plate buckling but not many.

4) At a more macro level, you're creating a pair of non-composite steel beams that probably depend on the remaining CMU to brace them against lateral torsional buckling. I've never checked that and have not seen anyone else check it either. It's prudent to retain some awareness of this though, I think, such that you might check or eliminate that concern in situations that warrant that.

5) An often overlooked, governing failure mode with these things is the need for the bolts at the supports to transfer the entire beams shear back into the wall locally. This can be significant and sometimes leads me to put channel columns under the channel beams to get this done.

 

[StrEng007]

OK, your thinking is in line with mine.

I do the same exact thing with channels when loads get high. The anchor cluster iteration always becomes a issue at each end, and I result to the same exact channel column you're talking about.