Lateral Torsional Buckling of Composite Plate Girder during Bridge Demolition

[question_asker]

Looking into the stability of a composite plate girder during bridge removal. The span in question is a 85' simple span and has 4 girder lines. Prior to girder removal, the overhangs and deck will be sawcut and removed, with sawcuts made near the flanges. Steel cross frames/diaphragms will be left in place. Due to the presence of shear studs on the top flange, it is preferred to keep the concrete over the flanges left intact rather having to chip out.

For picking and removing the first three girders, stability is not an issue, as the cross frames are left in place. Each girder will be supported at the 1/4 points by a crane prior to cutting free the diaphragms, and cutting the end connections. However, the final girder will be left free-standing, and I am checking the stability of this plate girder with an unbraced length equal to the full span length.

I am checking allowable stresses in accordance with AASHTO Standard Specs 17th Edition, Table 10.32.1A. If it was just the girder alone (no concrete on top flange), my calculations show the bending stress due to self weight is lower than allowable stress. However, adding the dead load of the concrete strip on top of the girder produces self weight stresses that exceed the allowable stress of the steel plate girder alone.

My question is - with regards to stability/LTB, how do I properly account for the contribution of the composite concrete section? My first instinct was to convert the concrete into an equivalent steel area. Using a modular ratio, n=0.1, I re-produced the calculations using a top flange that was now an equivalent section that went from 5/8" thick to 1 5/8" thick. Using the revised section properties, stresses, and Iyc, the stresses were lower than allowable.

Thoughts?

Notes:
-Can't make the final pick a double due to crane capacity/availability

 

[question_asker]

Attached is a cross section of the composite section.

 

[BridgeSmith]

I assume n=0.1 is the Ec / Es? We generally do it the other way - Es / Ec = n, with resulting values generally around 8, maybe as high as 9 for low strength concrete, but I never seem n=10. Your area increase for the top flange by including the concrete seems like a reasonable value

Anyway, as long as the concrete over the flange is intact until the crane is supporting it, I don't see why it can't be considered part of the section.

 

[centondollar]

Are you removing the concrete slab before lifting? I assume you will, since the weight of the slab is probably greater than the weight of the adjacent girder, and in that case you cannot assume it to contribute to section stiffness. How was the bridge originally built? If there were no props, one would assume that two girders were joined and lifted simultaneously, leaving you with one practical and safe option: acquiring more cranes or sturdier cranes for the job.

If all else fails, you could consider installing a temporary prop which restricts torsional deformation at mid-span (two columns and some type of cross beams and/or trusses) and calculating LTB for the steel section assuming an unbraced length of half the full span. Another expensive and difficult solution is to connect flange edges with continuous plates (or trusses) of appropriate thickness to make torsional and warping stiffness large enough that LTB is no longer an issue.

 

[question_asker]

@ BridgeSmith:
Yes, you are correct, I have my modular ratio shown backwards, and it should be Es/Ec like you stated. I assumed a conservative estimate of 2,500 psi for the existing concrete from the 1960s (in reality expect the concrete to be closer to 3 to 4.5 ksi compressive strength). The result is a high n value roughly equal to 10.

@centondollar:
The slab between the girders will be removed before any girders are removed. However, the small portion of slab (roughly equal to top flange width) will remain atop the girders (due to the difficulty in removing and chipping out around shear studs) and will be picked as one composite section. Larger crane is not an option due to site constraints. If we are not comfortable with the final girder's standalone stability, we may provide a temporary uplift force (nominal 5k to 10k) at the midspan of the section with another crane, reducing the demand on the section until the larger crane is rigged and fully supporting the girder at the 1/4 points. We have considered use of a stiffening truss welded or bolted to the top flange, but are leaning away from this option due to the additional work and fabrication required.

Thanks for your responses.

 

[centondollar]

If you have access to more than one crane, you could just make it a 4 or 6-point lift, increasing the LTB capacity considerably.

If you assume a composite section, make sure to cut the slab near the girder flange in a controlled manner. A quick and dirty demolition might damage the shear studs, and you are relying on those and the integrity of the longitudinal and transverse reinforcement to provide a composite section.

 

[BridgeSmith]

I don't think the stability for lifting is the concern. It sounds like it's the stability of the last girder after the diaphragms have been removed to allow removal of the adjacent girder.

 

[Lomarandil]

Agreed that the remaining concrete will reduce the flange's tendency to buckle. I think it's fair to increase LTB capacity accordingly.

If your web is very slender (late 1960s, early 1970s), note that your (mostly) non-composite girder now may have a very large portion of the web in compression. Hopefully your loading is low enough by this point in the sequence to avoid any web buckling issues. Or as a demo job this may not be a concern for you.