Bracing of Wide Flange Beam 1

[hdn32]

Hello fellow Engineer,

Our office is asked to evaluate an existing roof framing structure that was modified previously due to the addition of a higher roof. Please see attached photos below.

You will notice that, in order to handle the new drifting snow load, the following modifications were done:
• Introduce an additional support to the original steel bar joist (at mid-span)

• Reinforce the bottom chord of the original roof joist with steel angle “Br 1”. Braces “Br 1” are not attached to Beam “B1” ” (There is a 1” space between the end of brace “Br 1” and the top flange of beam “B1”)

• Introduce a WF beam “B1” to pick up the reactions from the new support of the roof joist.

• Introduce a WF beam “B2” to support (1) end of beam “B1” and support the other end of beam “B1” on a hanger connected to a braced frame.

• Introduce steel angle “Br 2” (welded to the bottom chord of the roof joists) to brace beam “B2” along its span. However, the steel angles “Br 2” are not attached to Beam “B2” (There is a 1” space between the end of brace “Br 2” and the top flange of beam “B2”)

Engineers in our office agree with the followings:
1. Capacity of the roof deck (Fiber board roof deck) needs to be checked for drifting snow.
2. Top chord of roof joist (at new support of roof joist) needs to be checked for bending.
3. Capacity of the new joist seat needs to be evaluated.
4. Stress reversal of web members of the bar joist needs to be evaluated.
5. Check capacity of beam “B1” and “B2” to support reactions from new bar joist seat.
6. Check the strength and stiffness of brace “Br1” and, especially, brace “Br2”

Finding information to work on #1 (Allowable capacity and as-built thickness of the fiber roof deck) and #6 (out-off plane stiffness of steel bar joist) are some of the challenges for us. However, the #5 is the one that has caused some debates and discussion among us.

One side argues that it is because the steel angles are not attached to the top flange of the beams, they (steel angles “Br1” and “Br2”) are not bracing the beam at all per AISC Steel Manual

The other side argues that if beam “B1” and “B2” buckle laterally, the top flange of the beam would be in contact with the vertical leg of the steel angle. Hence, the beams are being braced. This is engineer’s judgment

Would you consider that the beams “B1” & “B2” are being laterally braced by “Br 1” & Br 2” respectively?

Do you have any information on the fiber roof decking (Building circa 1965’s)? I am guessing that the fiber roof deck will not be able to handle the additional drifting snow of 60 psf (on top of the balanced/ flat roof snow load of 35 psf), what you do?

Thank you!

 

[WesternJeb]

I could buy the argument that there is lateral bracing from the angles to the top flanges if there is the expectation that it is lateral buckling with large deflections. I could also see the beam laterally buckling in more of a rotational manner, in which the angles would not provide any bracing. I would consider myself 50/50 based on how critical the bracing is to the capacity.

However, I could see someone coming along one day and not realizing this was your assumption since there is no connection between the two. That is too big of a risk for me to take to rely on it for my design. In order to rely on this, I would need some form of actual connection just for the security that no one down the road doesn't realize what my assumptions were.

 

[KootK]

I vote no on the beams being braced by the angles with those 1" gaps in play. Firstly, I feel as though the look of it would terrify the villagers should the beams buckle enough to mobilize that mechanism. Secondly, that amount of movement would increase the required brace force and stiffness in a manner that would be a bit of a pain in the butt to evaluate. And there's Jeb's valid concern for the permanence of that bracing. That's a whole lot of "yuck" for my liking.

This is a big stretch but the triangle things on top of the joist supporting beams might be envisioned to provide constrained axis LTB restraint. Problems there:

a) Triangle things probably not designed for this purpose.
b) Connection of triangle things to beams probably not designed for this purpose.
c) I'd like at least partial depth stiffeners for this purpose.

Why are those 1' gaps there? To accommodate joist end rotation? A simple and common bracing technique for these situations is to run diagonal kickers from the top of the beams to the top of the joists. This, too, restricts joist end rotation but is done ubiquitously.

 

[human909]

The other side argues that if beam “B1” and “B2” buckle laterally, the top flange of the beam would be in contact with the vertical leg of the steel angle. Hence, the beams are being braced. This is engineer’s judgment

I don't know about you, but I like bracing my members BEFORE they buckle, not waiting until they begin to buckle!

Like Kootk says that is a whole lot of yuck. Though it could still be effective in preventing total failure, in my book the member already has failed serviceability and there are a whole bunch of other nonlinear aspects to consider.

 

[canwesteng]

I think you could make the case for those braces acting as top flange restraint in a vacuum, but it turns into a bit of a science project, and would be cheaper to get the owner to connect them properly. In this actual situation, you have these triangle post things supporting a load several feet above the shear centre providing no rotational restraint. I think if the top flange goes 1" laterally to contact the brace, it will just dump the joist above it and the roof comes down. Plus hard to account for the load being that high above the shear center in AISC I believe, though some EC design guides have info that can be used to account for it. Long story short, I would require the braces with an actual connection in this case and not think of it twice.

 

[hdn32]

Thank you for sharing your thought!

Fyi, I am thinking of creating braced frames (between two joists, see below photo) at existing steel angle locations and use the frames to brace the beam. I would detail the connection to allow the joists to deflect vertically.

Again, thank you!

 

[bones206]

Introduce steel angle “Br 2” (welded to the bottom chord of the roof joists) to brace beam “B2” along its span.

Are you sure this is the intended purpose of the angles? Are they noted as such on the framing plans? Same question applies to the Br 1 members. Just saying they could be just for the joists and never intended to brace the WF beam.

 

[lexpatrie]

This looks like it might be Tectum or a similar product for the roof deck. If you get capacity issues here it goes to the original engineer, but perhaps going with some kind of "bulb T" reinforcement approach might alleviate some of the excess loading, failing that, modify the roof to add a foam wedge here to prevent the accumulation of the snow drift (probably blocks windows, can't tell from these photos).

If your "braces" are not in contact with the thing they are supposed to be bracing, they aren't braces. What you potentially have is some kind of "it buckled but it can only go so far" contact stress problem, which would be a real challenge to model realistically. Once it buckles and your assumed buckled shape comes into contact with the "braces" then sure, sort of. You could look at is to see if there was some post-buckling additional strength. To me this is something for an Abaqus model, and I've never used that software, I'm just going off it's reputation and where I've seen others publish research and thesis work that's in the neighborhood. This is a more advanced topic and most engineers lack the background in stability, because it's a good ways beyond "normal stability" at the graduate level. How would you even go about doing a simplified model to check your computer results?

If others have already mentioned, sorry, but to me those added angles look like bottom chord reinforcement for the open web steel joists. The center of the joist span typically has limited shear capacity and they added a center support there, that's the beam you are trying to justify. If the original engineer for the modification picked something that looks like a column more than a beam section, to me that's a clue it's unbraced or mostly unbraced, so a W12, W14, or W16 section would suggest that.

 

[hdn32]

@bones206: yes & by calculation. The beams would fail badly if those are not considered bracing points.

@lexpatrie: Thank you! Do you have any info on the roof deck? (the building was built around 1964)

 

[lexpatrie]

Well Tectum is still made, you'd want to determine the thickness and width of the panel and then poke around. I've seen it on roofs in manufacturing facilities in the 1960-1980 range. They were also fairly popular for swimming pool buildings.

 

[bones206]

yes & by calculation. The beams would fail badly if those are not considered bracing points.

I see. That's very bad.