Allowable compressive stress due to bending

[jtseng123]

Hi all,
The question may be silly, but what is the allowable compressive stress due to bending during lifting or shipping ?
Please refer to the attached sketch.

Still 2/3 of the yield like on the tensile side ? or any other formula ?

 

[TGS4]

You are dealing with gross-section bending, which means that your vessel will be in membrane (longitudinal) compression on the top side. The failure mode for compression is buckling. Therefore, you would need to follow the rules in UG-28. Depending on the distance between lines of support, your allowable compressive stress could be quite low - certainly lower than 2/3 Sy.

 

[JStephen]

I hadn't really thought of that- but I note that while compressive stress is limited per UG-23, fabrication/erection/shipping loads/stresses are not included in the load cases considered for that compressive stress. With that in mind, I would think the normal allowable could be exceeded in that case.
The compressive stress in this case would be in the as-built thickness, ambient-temperature condition as well, which might increase it substantially.
I've seen data elsewhere to the effect that allowable compressive loading due to a moment could be considerably higher than compressive stress due to pure axial loading, and, if not required to be limited by the code, that might also allow some increase.

 

[TGS4]

It should be noted that the Code margin against buckling is only 2.0. Not much room to relax. And it is based on Code-required tolerances, too.

Definitely take advantage of the as-built thickness and room temperature - buckling is based on the Young's Modulus as the sole material property.

 

[racookpe1978]

Buckling stress due to bending will be concentrated at the stress points of the compressed side of the empty and hollow pressure vessel tube: Handholds, openings, nozzles, and vent or pipe flanges near the midpoint of the "bend" will cause stress concentration risers that a simple hollow "tube" analysis will not find. Doesn't mean the buckling failure will be at those points - they are, after all, reinforced against internal pressure at those same points!

But look for the stress risers for likely places where the buckling will start. AND , yes, construction failures while lifting up "simple" walls from horizontal to vertical ARE a failure mode.

 

[keyen]

TGS4: is there a generally accepted method of using UG-28 for loads that are not due to external pressure (such as the lifting case mentioned in this thread)? I assume it's not a case of simply keeping the compressive stress below the "B" factor, correct?

 

[TGS4]

Actually, just keep the compressive longitudinal stress below B.

 

[jtseng123]

I may not agree the use of UG23 and UG 28, which is for bucking stress across the entire cross section, that is also based on ‘line of support’, the “L” dimension. Cone can be defines as line of support both ends, either end or none, that will impact the dimension “L”, hence impact the allowable buckling stress. In my case I have two cones, such that no single value for the allowable buckling stress can be obtained because I can define line of support in many ways. But regardless how I define, it can be all wrong because there is no external pressure to justify the line of support during lifting or transportation. If using the entire length as the “L”, that will make the allowable buckling stress very small and to use that as the allowable compressive stress due to bending moment, may be overkilled.

Vessel can have many vacuum stiffing rings for external pressure. Does it means we can use “B” from UG23 and 28 to calculate the allowable buckling stress and treat it as the allowable compressive stress due to bending moment during lifting or transportation ? I wouldn’t think so.

In my sketch, the maximum compressive stress is close to 2/3 of the yield due to head heavy. If UG23 and 28 are a valid method, then I can simply add two stiffing ring between the maximum point, may be 1’ part, to make B exceeds compressive stress, or even I can define the small cone as line of support and make B large enough, and then I claim the maximum compressive stress is fine? Does not sound right.

The maximum compressive stress from bending moment will be very local at the farest point from the vessel center line. I do not think a gross buckling will occur even the compressive stress exceeds the “B” value from UG23 and UG28 if that can be calculated.

I could be wrong for all the thinking, but I am still not convinced the use of UG23 and 28.

 

[TGS4]

Do you use UG-23 and UG-28 for wind net section bending? Do you check it with zero internal (or external) pressure? Why is that any different from the lifting condition?

 

[r6155]

Allowable stresses per AISC. See Pressure Vessel Design Manual, by D. Moss 4th ed.

Regards
r6155

 

[racookpe1978]

Wind loading (cross section x wind force factor) would not be equal, nor comparable to, compressive stress loading due to gravity as it is lifted. The calc is roughly similar, since a bending moment is imposed from the side; and the bending moment does not depend on internal pressure. But the location of that bending moment and the cause of the moment, and the amount of the bending moment is different.

 

[victorpbr]

jtseng123

"Vessel can have many vacuum stiffing rings for external pressure. Does it means we can use “B” from UG23 and 28 to calculate the allowable buckling stress and treat it as the allowable compressive stress due to bending moment during lifting or transportation ? I wouldn’t think so. "

In my view there is no problem with that as we are addressing the same failure mode (buckling) on both cases.

 

[jtseng123]

What happens in my vessel currently under fabrication is, the compressive bending stress is 10% over the calculated "B" value. And I am seeking what is the correct allowable stress to be used. For sure if "B" is used since the beginning, (which in my view is too conservative because we are not talking about gross buckling from axial loading), then there won't have problem. In real situation, things always happen during design and fabrication and we have to resolve it. The lifting location can not be relocated. To thicken the shell become an impact and not desirable. A temporary 3' long I-beam to stiffen that local area during lifting may be the way.

In Dennis Moss 3rd edition page 389 (4th ed, page 664) for erection analysis, it says: " Compression (for short members only), allowable compressive stress = 1.33 x B. ". If that is true, that will save my vessel from any changes. Any opinion on that 33% increase on the B value ? And what is the "short members" the manual talking about ?

 

[TGS4]

My thought is that if you can't pass it with B, then you need to do a full-blown elastic-plastic buckling FEA to demonstrate that you have an adequate margin against buckling. Or, you could try the recommendation r6155 gave and figure out the AISC allowable compressive stress for your situation.

Either way, get someone who actually understands buckling failure to assist you.