Unbraced Length & LTB of a unitized shape (shape profile shown below)

[AnimusVox]

Hello all,

Background: This is regarding aluminum mullions (framing members) to be used within a glazing system. Our company mostly focuses on calculations ensuring that glazing systems (aluminum framing, glass, and steel reinforcing (if necessary)) are designed to code.

The Problem: We have noticed a methodology of analysis used within the industry (calculations signed and sealed by licensed engineers) to divide the unbraced length by 3 for the purpose of finding the allowable lateral buckling stress. (i.e. if a mullion is 90 in long and has an r[sub]y[/sub] of 1 in, the slenderness ratio is taken as 90 / (3 * 1) = 30 per this methodology).

Refer to the attached picture of the shape cross section as a reference.

We asked another company about the rationale behind dividing the unbraced length by three (since we couldn't find any reasoning for it within the code) and they replied with

"There is no code or design manual but it's based on theoretical column buckling. They are continuously interlocked and are assumed to buckle towards or away from each other... which creates a point of lateral support against lateral buckling. L/3 assumes the mullion is at the 3rd mode of buckling (same as column buckling) in which two points of contact have been attained at third points."

The response then goes on to say that they may divide by a number potentially greater than three due to their confidence that lateral buckling will not occur, because 'in my 22 years i've never seen a mullion buckle laterally'.

This explanation seems a little suspect, and we've decided to shelf the discussion until a week from now, and bring any engineering support for one way or the other. Obviously we'd like to adopt this methodology because we can increase the capacity of unitized mullions, but we want to make sure we have some reasoning to fall back on. We've also decided to contact The Aluminum Association, the body behind the aluminum design manual (the primary code resource we refer to for projects) for their input.

All of your feedback is greatly appreciated!

 

[oldestguy]

It's not my field, but have you ever run column type load tests on these units, possibly with partial and no lateral support (as for hinges, etc.)? If they are subjected to side loading, due to wind, that may be tested also.

 

[MotorCity]

Not my field either, and I have no idea what the response to your question is getting at, but here is my take. I see this shape behaving in 1 of 2 ways, either as a tube or 2 channels interconnected. If it behaves like a tube, LTB is a non issue. If it behaves like 2 channels, LTB is certainly an issue. Perhaps they are considering 1 channel braces the other channel, or the internal fins/walls have some sort of stiffening effect against LTB.

 

[AnimusVox]

@ oldestguy

That's a great suggestion, and tested values would something acceptable to fall back on.

However, there are several different manufacturers for these types of mullions, most of which have not tested their systems. Whenever we ask if there's testing to support the reduction in the unbraced length (so we could have something to reference and adopt the reduction), we're never given any supporting tests, at this point it seems to just be a rule of thumb.

 

[AnimusVox]

@MotorCity

We consider it to behave like two channels (the two halves are simply snapped in place and erected as shown in the picture).

We're having difficulty justifying that the interlocking flanges truly support each other for LTB.

 

[Shu Jiang PEng SE]

The member can experience flexural buckling and local buckling under compression load if its two components work together. the member can experience local buckling, flexural buckling and flexural-torsion buckling under compression if it is treated as two separated channels.

the member can experience yielding, flange local buckling and web local buckling under lateral load if its two components work together. if the member is treated as two separated channels. it can experience yielding, lateral-torsion buckling (bending along strong axial) or flange local buckling (bending along weak axial).

 

[dik]

You can't 'hang your hat on it', or quantify it. In my experience, it takes little force to provide stability for buckling. and it is likely that the 'forced' fit of the extrusions to provide a boxed section is sufficient and that LTB is not an issue.

Verification may occur by doing calculations on the complete cross-section and confirming this by physical testing.

Dik