Lu for pemb

[calculor1]

I'm trying to determine Lu for the beam in a rigid frame (tapered web), the top flange is braced laterally with purlins at every 2.5 feet and the bottom flange is braced with diagonal braces spaced at approx every 9 feet. For LTB Lu is 9 feet since the bottom flange is in compression; however it would seem to me that the inplane buckling strength for the beam should govern for the axial capacity and Lu would be much higher. The span for the rigid frame is 120 feet with splices located approx 35 feet from each end column. Am I missing something?

 

[JAE]

Wouldn't you just compare your KL/r values in each axis and use the larger? Or check both directions with their applicable KL/r values and associated axial/bending forces?

Am I missing something?

 

[calculor1]

I undertand that, I guess my question would relate to what you would use for Lu for the x axis, 9 feet, 2.5 feet or I believe it should be the full length of the beam/column. This should be so obvious but I've been looking at too long.

 

[msquared48]

As both flanges can go into either tension or compression depending on the load combinations, you would have to specify which load condition governed for each flange, and where it occurred along the frame.



Mike McCann
MMC Engineering

 

[calculor1]

Okay I get that, I have only one load combination (DL+SL), see attached moment diagram. What would you use for Lu to determine the inplane capacity?

 

[JAE]

Perhaps you would keep a distinction between the Lu for KL/r axial capacity vs. the unbraced length Lb for flexure. They are dealing with two different things.

 

[hokie66]

9 feet for weak axis buckling, probably the distance from eave to ridge for strong axis buckling, but strong axis buckling is a rarety.

 

[calculor1]

Why would you say strong axis buckling is a rarity, not doing any calcs I would think having a column 50 feet long combined with a moment.

 

[hokie66]

If it is a column, yes, buckling of a long column about the strong axis can certainly occur if braced well in the other direction. But members of a rigid frame as you depicted are primarily bending members, with the axial component of stress relatively quite small.

 

[JoshPlumSE]

AISC Design Guide #25 (newly published) is a good resource that you should probably check out for this situation. My take on your issue:

1) SA Buckling (and frame buckling): The Design Guide would have you use the full length of the member (with a K=1.0 assuming you are using the Direct Analysis Method). But, the calcualtion of the buckling load would be more complex since the properties vary along the length of the tapered member. Though if you have approximately uniform axial load, you can just assume an "equivalent" moment of inertia for buckling calcualtions.

One of the tricky things with tapered frames is that it should be easier for frame buckling to control over weak axis buckling. Therefore, the Direct Analysis Method becomes (IMHO) a more viable option than traditional hand calcs.

2) Beam Buckling (LTB): This should be calculated for each unbraced segment using the properties of the beam at the halfway point of the unbraced segment. This should be calculated separately for each flange which experiences compression. In your case, therefore, the 2.5 feet would be used when investigating top flange LTB buckling. The 9 feet would be used when investigation bottom flange LTB.

FWIW: I am not an expert in metal buildings or tapered frames. So, it's not like I speak from years of knowledge. My experience is relatively new as my company (RISA) is considering incorporating this design / code check procedure into our programs.