Q: reinforce existing beam for new RTU? 2

[CPBe]

Looking for some input on an existing roof structure.
Field-gathered info indicates the existing structure was designed for a 30 psf roof load.
When the added weight of a 3.1k RTU is introduced beam deflection exceeds L/240, so I'd like to reinforce per the following schematic:

Trying to keep the weight off the existing joists (w/kickers down to WT) so that those don't have to be reinforced as well.
Would the WT need to run the full length of the beam to achieve the desired effect? Or is welding the member along the length of the RTU (+/- 2/3 length of the existing beam) enough?
Also, would the ends of the beam (and columns) need to be reinforced as well? Or are the increased shear loads (less than 15%) small enough to negate that?
Any feedback would be greatly appreciated.
Thank you in advance for your input!

 

[KootK]

Would the WT need to run the full length of the beam to achieve the desired effect? Or is welding the member along the length of the RTU (+/- 2/3 length of the existing beam) enough?

I suspect that 2/3 of the beam length would be enough but, by the book, you'd want to run a deflection analysis looking a the reinforced beam as a variable Ix member. One usually extends the WT a bit beyond what would be required by this analysis to account for shear lag.

It will also be prudent to consider the loading condition of the beam at the time of reinforcement with respect to how things will play out for deflection. Will you unload it? Jack it to horizontal?

Also, would the ends of the beam (and columns) need to be reinforced as well? Or are the increased shear loads (less than 15%) small enough to negate that?

15% is creeping up a bit higher than my "do nothing" value so I'd recommend evaluating those items.

Trying to keep the weight off the existing joists (w/kickers down to WT) so that those don't have to be reinforced as well.

I realize that this is a pretty popular strategy but I've always been a bit queasy about it. My concerns:

1) Unless the RTU is dead symmetric wrt weight distribution, the beam will twist and tend to load the joist a bit anyhow.

2) I don't love the thought of the joists hanging up on the kickers when they try to deflect. This can create negative moments at the joist ends and an increased need for bridging etc.

3) In my area, snow drift tends to matter more than the RTU weight anyhow.

I'm not saying that you ought not do this but, rather, that there are some nuances that I feel warrant consideration.

 

[Aesur]

I agree with KootK here and want to add: I disagree with your idea of keeping the loading off the joists, if the angles are attached to the joists, the joists will see the loading. Additionally, look at the components, you will be relying on the steel joist for tension loading in the top chord to keep the brace from rotating. Being that this is how these joists work, I suggest you look at the tension in the top chord or add some kind of tension tie between the braces.

 

[CPBe]

Maybe I'm making this more difficult than it needs to be...(?)
Would reinforcing each of the joists the RTU is supported by be a more straight-forward approach?

 

[EngDM]

Would reinforcing each of the joists the RTU is supported by be a more straight-forward approach?

Yes. You'd have to site confirm the chords and depending on if they are some proprietary shape you'll have to do some research on how to analyze them accurately, but you've drawn them as B2B angles so I doubt you have that issue. You'd need to check the webs too as the loading can increase in those as well.

However, your reinforcing may be required to be nearer to the middle of the joists, as thats where the compression maxes out. I've had cases like yours before where I don't need to reinforce below the RTU, but the compressive load at the middle exceeds the Cr of my top chord so I need to add the round bar for say, X ft extent centered on the top chord.

I will say, I'd usually provide a steel channel where you are showing the angle, and then have a clip angle off the top chord of the joist to pick up the channel.

Another note, in cases where the RTU is really heavy in lieu of evaluating the existing beam-column connection I have just added a beam below the existing (similar to your T) and spec'd a new connection.

Not sure if snow drifts are an issue, but if they are you need to check the steel deck to support this increased snow load. You can add intermediate an intermediate channel the length of the drift, between the joists to cut down the deck span.

 

[JLNJ]

At first glance, it looks too much like a teeter-totter. If you put some eccentric snow and wind as KootK says, where do the forces go?

 

[bones206]

Another note, in cases where the RTU is really heavy in lieu of evaluating the existing beam-column connection I have just added a beam below the existing (similar to your T) and spec'd a new connection

Im trying to visualize this. When you add the beam below, do you stitch it to the existing beam above?

 

[EngDM]

Im trying to visualize this. When you add the beam below, do you stitch it to the existing beam above?

Yes, and if you really want to be cautious I've seen the new beam be designed for the full load, and cut the existing connections so the load transfers directly down to it. Can't really work if there is any axial loading in the joists since it would create quite a bit of overturning on the low beam, but for a typically roof beam located in the "field" of the roof it shouldn't be a problem.