Capacity of brace that extends 75% below top flange of W section to resist LTB

[mcustructureman]

Hi all,

I am currently working on a retrofit to a 50 year old steel building. After performing design calculations on an ex. W410x60 beam that has a laterally unsupported bottom flange, I realized it fails in negative bending. It is 2 span continuous with an 8 m and 6 m span. The condition is tricking since the intermediate support is a hanger which goes up to the floor above. I have proposed a detail bracing the beam at the hanger support which effectively restrains the bottom flange but because of existing mechanical, we are having difficulties bracing the bottom flange near the inflection points, but bracing can be achieved by extending the OWSJ bottom chord to the web of the W410. This bottom chord extension would brace 75% of the beam depth. Would this be considered enough to restrain the beam against twist if the bottom chords are extended on both sides of the web? I'd also be concerned about bottom flange buckling in plan since the bottom flange is in compression over the intermediate support and is not fully braced before or at the inflection point so it could still theoretically buckle from support to support. I also read a thread (which I can't find now) that suggested bending of the web in the weak axis from 2% of the UDL, which I'd also be concerned about if 15% of the beam depth is not braced.

*Fairly new engineer and still learning, please guide me to any reference material that can help me with this problem

Edit: 75% of the beam depth is braced

 

[KootK]

Would this be considered enough to restrain the beam against twist if the bottom chords are extended on both sides of the web?

Yes, it's generally accepted that a bracing connection that grabs 60% or more of the beam depth will suffice for roll bracing.

I'd also be concerned about bottom flange buckling in plan since the bottom flange is in compression over the intermediate support and is not fully braced before or at the inflection point so it could still theoretically buckle from support to support.

I'm confused by that. Your plan, ultimately, is to rotationally brace the beam at the hanger and at both adjacent inflection points, right?

To an extent, the hanger condition will be inherently stable. It's similar to bottom flange loading (below shear center) on a simple span beam and it will take some doing to get the beam to roll upwards at that intermediate support.

The hanger itself may be able to provide the rotational bracing at the intermediate support. It depends on what the hanger is and what it's connections to the beam are like. Can you elaborate on that, preferably with a sketch?

If both adjacent inflection points are rotationally braced, it is possible to make do without rotational bracing at the intermediate support. That's not my preference, however, so let's only go there if we have to.

Get yourself a copy of the AISC steel specification which is freely available online. Appendix 6 contains requirements for bracing which are, in my estimation, the best available in North America at this time.

 

[KootK]

Also, give some thought to the fire rating of that hanger. This is more of an issue for roof supported hangers but, perhaps, still something to take care with for your situation.

 

[mcustructureman]

[I'm confused by that. Your plan, ultimately, is to rotationally brace the beam at the hanger and at both adjacent inflection points, right?]

Sorry if I made that confusing but my question is, am I correct for considering the buckling in plan of the last 25% of the beam cross section due to the axial compressive force that develops because of the moment couple (See the second detail attached)? The bottom flange is only supported from buckling in plan from support to support regardless of the bracing at the inflection point, since the bracing at the inflection point only extends 75% down the beam depth and is rotationally restraining the beam up until that point.

The ex. hanger condition is also shown in bracing detail 1. The hanger is bolted to the top flange of the W410 (from what I saw on site).

[it will take some doing to get the beam to roll upwards at that intermediate support]

I guess I never thought of it that way, that does make sense. I've already got the go ahead for my intermediate support bracing detail so retracting it may be redundant.

 

[KootK]

Sorry if I made that confusing but my question is, am I correct for considering the buckling in plan of the last 25% of the beam cross section due to the axial compressive force that develops because of the moment couple (See the second detail attached)?

Don't sweat the 75% thing. Like I mentioned previously, 60% for a roll beam bracing scheme, which is what you have here even though it's OWSJ, is commonly considered adequate to brace both flanges.

I've already got the go ahead for my intermediate support bracing detail so retracting it may be redundant.

I'd do it as I've shown it below. You want the stiffeners run up to the hanger as web depth (not full depth) so that you can mobilize the hanger in flexure as the roll bracing element for the beam. Think of the stiffeners as a downward, flexural extension of the hanger. The kickers will work however:

a) They mess with the natural deflected shape of the joists.

b) They cost money.

c) Stiffening the top flange removes flexibility from the connection in a way that stiffening the bottom flange does not.

How sure are you that this is a hanging condition rather than a transfer beam connection? A bolted base plate to an unstiffened flange is a pretty crappy hanger setup in many respects.

Some engineers will prefer the kicker braces owing to their greater bracing stiffness.

 

[dold]

a) They mess with the natural deflected shape of the joists.

It's joist day I guess! OWSJ / OWSJ Girder designers use a really crummy similar detail to brace the bottom chord of joist girders - just an angle with realllllly janky welds to the girder and the joist bot chord. I've seen these welds fracture. No gusset/weld plates - just welded along the edge. They do note that these braces shouldn't be welded until all dead loads are applied though. Plus the tail of the joist seems pretty flex-y. Maybe flex-y enough to not affect the joist internal forces, but un-flex-y enough to brace the girder for uplift?

 

[mcustructureman]

How sure are you that this is a hanging condition rather than a transfer beam connection? A bolted base plate to an unstiffened flange is a pretty crappy hanger setup in many respects.

100%, the beam span is too long and tributary width too large for a W410x60.

I'd do it as I've shown it below. You want the stiffeners run up to the hanger as web depth (not full depth) so that you can mobilize the hanger in flexure as the roll bracing element for the beam. Think of the stiffeners as a downward, flexural extension of the hanger.

That makes a lot of sense. I'll pitch this to my boss and see where I can take it.

 

[KootK]

Plus the tail of the joist seems pretty flex-y. Maybe flex-y enough to not affect the joist internal forces, but un-flex-y enough to brace the girder for uplift?

The lateral stability of brace / joist chord combo is the part that bothers me most. Usually:

1) The first joist bridging line is a ways back and;

2) The installation is rough enough that you've got a fair bit of P-baby-delta baked in.

I wonder if anybody's tried to prosecute an appendix 6, bracing check on the setup.

 

[mcustructureman]

2) The installation is rough enough that you've got a fair bit of P-baby-delta baked in.

Is that assuming the chord deflects about its weak axis because of the kicker brace installation? Consequently, does the first mode shape then induce more weak axis bending because of the axial compression induced from flexure? If the kicker brace were installed to the bottom chord, I'd expect the axial tension counteract any p delta effects, is that safe to assume?

 

[KootK]

Is that assuming the chord deflects about its weak axis because of the kicker brace installation?

It would be assuming that the final configuration of bottom chord + brace is probably kinked a bit in plan. That would tend to introduce lateral bending into the bottom chord.

Consequently, does the first mode shape then induce more weak axis bending because of the axial compression induced from flexure?

As you drive flexurally induced bottom chord compression force into this laterally kinked thing, that creates additional lateral bending in P-baby-delta fashion.

If the kicker brace were installed to the bottom chord, I'd expect the axial tension counteract any p delta effects, is that safe to assume?

Sort of. The more nuanced version of that story would be:

1) When the joist is assumed to simple span, there is no tension in the bottom chord near the support. Rather, there will be compression as the the torsional stiffness of the beam attempts to resist the rotation of the joist end.

2) The joist bottom chord near the support will only be in tension once the rotation of the beam exceeds that of the joist end. In a way, this exacerbates the tendency of the beam to LTB buckle.

3) With a bottom chord extension to the beam on both sides of the beam, I see no way to get around your having created a two span, continuous joist across the beam for any load occurring after the braces are installed. This creates the demand for a tension connection at the joist top chord that crosses the beam.

Most engineers will consider this stuff "overthinking" but such is life / engineering. Everything that you've been proposing thus far is reasonable and with the range of what is commonly done in my opinion.

Do you know how far back your first line of joist bridging is?