Steel box section flexural analysis

[atrizzy]

Hi all,

As the title says, I'm analyzing a steel box section for flexural capacity. For sake of argument lets say the section is 20" deep with continuous plates top and bottom, and vertical 'webs' every 24". (It is actually a little more complicated than that, but this is really just for illustrative purposes at this point)

There are two approaches I'm considering:
1) considering an equivalent I beam shape where the length of the flanges is determined by compactness limits
2) considering the entire cross section with yield stress reduction to account for slenderness of the flanges

My concern is that the calculated capacity using approach 1 is much higher than that calculated when using approach 2. This suggests, counterintuitively, that removing material from the section would improve the capacity.

Secondly, if the calculations are correct what would this mean in reality? That the box section would have the ultimate capacity of design 1, but that the middle of the unsupported flange sections would actually buckle locally (at design 2 capacity) before reaching this ultimate design 1 capacity?

Any ship builders here that might lend a thought?

 

[bugbus]

I think it boils down to whether or not the slender/unsupported parts of the flange could induce a failure at a load that is potentially less than that if they were ignored completely (a series of I-sections as you said).

It does seem counterintuitive, and my feeling has always been that removing/ignoring slender parts of a section will always reduce the capacity (even if it is marginal), but would be very happy to be shown otherwise.

I think it is probably possible (especially for a very slender flange) that you could get some local buckling away from the webs, and still have additional capacity in the section. I'm not sure if the formation of a local buckle necessarily is the point at which a section is deemed to have 'failed'.

 

[atrizzy]

Hi gusmurr, I agree that if my suspicions are correct it shouldn't 'fail' the system, just buckle the unsupported part of the flange.
For the loads I'm working with, even a 3" extension from the web technically has the capacity to resist the loads. I'm not sure if the amount of buckling of the interior of the plate would cause significant visual or deformation issues though.

 

[bugbus]

The other thing to consider is that the stress in the flange is probably by no means uniform. I bet that midway between webs, it is quite small compared to near the web. This would probably have the effect of delaying the point at which local buckling occurs. Also, if local buckles were to occur, this would probably result in an additional redistribution of stress towards the unbuckled parts of the flange.

FE modelling would be helpful in this case in my opinion.

 

[atrizzy]

Excellent point, gusmurr.