Beam Stiffeners 1

[mikek396]

I have a job where a piece of rotating equipment is sitting on top of steel beams. Rotation is in plane with the length of the beam, so theoretically there is no (or very little) lateral force on the top flange. The beams were supposed to be set such that the slab was poured flush with the top of the beam, however, the GC instead poured the bottom of the slab level with the bottom of the beam tension flange. This created the situation shown in the photo below. Now we are getting a very slight oscillating lateral movement at the top of the beam (~.04"), which I have been asked to come up with a fix for. My initial thought was to chip out the slab in a few locations and weld full height stiffeners at load and bearing points, but the contractor asked if it could be done without chipping out and just weld stiffeners as shown below, either with or without the addtional bearing plate between the stiffener and the slab.

I have already checked global bending capacity is fine without the full lateral bracing of the compression flange, we just need to fix the lateral deflection issue. I'm not really sure what calcs to run here as I dont believe this is a case of web yielding or crippling. I used the deflection to calculate a max lateral force on the flange but now what?? Any help here is appreciated.

TYIA

 

[jayrod12]

If you wanted to approximate the lateral deflection you're seeing, you could use roughly half of the weak axis moment of intertia and an assumed length of participating beam. To dial in the length of beam you could use your measured deflection.

All that above is strictly for a gut check if necessary.

Personally, I'd want the bearing plates for the stiffeners. Because in my mind, if there's still a portion of unstiffened flange between the top of concrete and underside of stiffener, you're still going to experience some flexibility. It will likely be an improved scenario without the bearing plate, but may exhibit movement.

 

[DaveAtkins]

I agree with using a bearing plate in combination with the stiffener plates.

I am also concerned with fatigue. Isn't it better to cut the stiffener plates a little short of the top flange, and avoid welding them to the top flange? I recall seeing somewhere that stiffener plates welded to a top flange with bridge crane traffic are subject to fatigue failure. This may not be the same situation as a bridge crane girder, but I thought I would throw that out there.

DaveAtkins

 

[XR250]

Could you weld some knee braces near the top flange and bolt them to the slab or would they get in the way?

 

[mikek396]

@jayrod12 - I approximated deflecting force with delta = PL^3/3EI, as I am assuming the deflection is coming more from rotation of the web& top flange about the "fixed" end embedded in the slab. I came up with ~2250LB, but I really cant figure out where this is truly being generated as the rotation is in the same plane as the beam length. Good point regarding the bearing plate - this was my thought as well.

@Dave - I don't see fatigue being a concern here, as the load points on the beam are stationary, but I am curious about the fatigue failure mechanism you mentioned. Do you have a link or something I can check out?

Thank you both

 

[mikek396]

@xr250 - this is an interesting idea. I realize now that I left out any dimensions - this is a W18 with top flange ~12" above the slab. If this was larger, I think knee braces would work better but I'm not sure the added complexity is economical

 

[DaveAtkins]

I did a quick Google search, and it appears it is OK to weld the stiffener plate to the flanges, as well as the web.

DaveAtkins

 

[XR250]

@xr250 - this is an interesting idea. I realize now that I left out any dimensions - this is a W18 with top flange ~12" above the slab. If this was larger, I think knee braces would work better but I'm not sure the added complexity is economical

Honestly, the knee braces are probably an easier fabrication option and be a more effective solution. They only need to be on one side of the beam. Welding the proposed stiffeners the underside of the top flange is going to be fun.

 

[jayrod12]

I assume there are a pair of these beams at least? What about just tying the tops of the beams together so they're resisting the lateral load in concert with each other.

 

[centondollar]

Discrete web stiffeners (with or without the proposed bearing plate) will not increase torsional rigidity. If you want to prevent lateral movement, your best option is to weld web plates to one or both sides of the beam, and to connect these web plates to the beam web with horizontal plates at height of the concrete slab. By welding the horizontal plates first, you have more than enough space for welding the vertical plates. The required plate thickness will be small, and can be hand-calculated (bending, shear and torsion of the box).

The result is great torsional resistance (a double-cell box section, assuming direct bearing of the horizontal plates against the concrete slab) and great lateral inertia, which will prevent any lateral movement.

 

[dold]

Like jayrod said, are there two beams, and if so, how far apart are they. I'd suggest more of a diaphragm bracing scheme. I.e., L3x3 x-brace between the beams at some interval along the beam. That also assumes there doesnt need to be an open space between the beams. Maybe even use some sort of baseplate where the bottom nodes of the x-brace / diaphragm bracing thing attaches to the beam web at the base. Probably overkill. But it sounds like a critical problem.

 

[mikek396]

@centondollar - Im not sure I agree with your assessment that stiffeners + bearing plate wont solve this issue. Its not quite a torsion problem, but cantilever bending of the web plate

There are 2 beams, 41.5" center to center.
X-brace seems like overkill, and it seems like I would need to weld a web plate anyway to make the top connection work

 

[WARose]

Agree with centondollar: web stiffeners will not change the torsional stiffness of the beam. I did a FEA study on this some years back.....and you have to add a impractical number of them to make a difference. I like the brace idea others have proposed.

 

[mikek396]

@WARose - if the issue is from cantilever bending of the web plate as opposed to torsion - adding stiffeners solves this, correct?
It seems the web plates need to be added if I am to propose the x brace, so I may just have them add the stiffeners first and if there's still any movement, then add the bracing

 

[PMR06]

If the space between beams was always intended to have concrete up to top of flanges, I would work out a fix detail that provides that.

Roughen the concrete surface between beams, add some little dowels into the existing slab, weld some rebar on to each of the exposed beams and lock it all in place with a block of concrete.

 

[dvd]

The issue may be related to "soft foot" and needs to be addressed mechanically. Some awareness of soft foot may be helpful so that you don't continue to try to solve a separate issue.

 

[dold]

PMR's idea seems like the sturdiest solution. Maybe cheapest too depending on access. As long as your floor framing can handle the extra ~150 psf of concrete.

 

[WARose]

@WARose - if the issue is from cantilever bending of the web plate as opposed to torsion - adding stiffeners solves this, correct?

It is hard to say without knowing more than I know now. It would certainly stiffen things up (locally)- where you do have stiffeners.

In a vibration problem, you are looking for a landmine....only calculations tell you where that is. (I.e. stiffening things could actually make the vibration problem worse....or better....you don't know until you grind the numbers. Coming up with a fix without that is problematic.)

 

[centondollar]

@centondollar - Im not sure I agree with your assessment that stiffeners + bearing plate wont solve this issue. Its not quite a torsion problem, but cantilever bending of the web plate

I wrote that welding vertical and horizontal continuous plates to the section will increase torsional stiffness and lateral inertia. If you have "cantilever bending" as you describe, you are dealing with both warping torsion and weak-axis-bending - both of which are greatly reduced with the proposed solution.

Discrete stiffeners do not change the global behavior of the beam and will not prevent lateral or torsional deformation or vibration.