Reinforcing Existing Roof For New Drift Loads?

[SteelPE]

I am in the process of designing an addition to an existing building. Building Code is IBC 2015/ASCE7-10. The new addition is higher than the existing structure which is going to impose a significant snow drift onto the existing. By some miracle, we have a water stained version of the existing design documents.

Normally we would install new joists in-between the existing joists in order to support the new load imposed by the drift. This is great, however, there is a specific space where getting into the space to do this style of reinforcing is prohibited (any reinforcing in this area is not allowed). So now we are required to come up with a different method to reinforce the existing roof. We do have access to approximately 18’-0” of the 38’-4” long existing joists (just under ½ of the joists length):

Lowering the addition is not allowed.
Framing a new roof on top of the existing roof is not allowed.
Any type of reinforcing inside of the prohibited space is not allowed due to access issues.

In talking this over with my colleague, we have come up with a method that may work… although it is a little sketchy. We are proposing to introduce new beams in-between the existing roof joists. These members will only be applied above the space that we can access (which is roughly the length of the snow drift). These new members will be supported by stub columns and new support beams that will be 4” below the existing joist bottom chord (this is to allow the existing joists to deflect and not be accidentally supported by the new support beam and cause a stress reversal in the existing joist).

I think this is the best solution that we can possibly come up with. It checks off lots of boxes, however, I am slightly nervous about introducing a new support system in-between the existing joists that will be supported at mid span of the existing joists (thinking that we may still end up with some stress reversal due to increase stiffness of the support system at this location). I can easily design the new support system to support all of the loads in the bay if necessary.

Does anyone have any comments on the proposed plan? Has anyone attempted a similar method of reinforcement in the past?

 

[milkshakelake]

If you're cutting the joist span in half, it might be worth evaluating the negative moment. It would be 20-25% of the positive moment (maybe 60% due to snow drift). Assuming open web steel joist, it's also possible to reinforce top chord by welding rebar, though I don't know how to calculate it. If it checks out and the roof deck can support increased snow drift, it might be cheaper to put a new WF girder under the existing joists and reinforce/stiffen them at the bearing location.

 

[r13]

If top of the roof space is open for utilization, and free of obstruction, have you considered external reinforcing from top?

 

[SteelPE]

If we apply a support at mid span of the joists then we will need to add bridging in the space we are not allowed to access. It was an idea we thought of but dismissed due to reinforcing requirements.

Reinforcing from above is not possible

 

[phamENG]

SteelPE - I think it's a good option. Do you have a good way of running an analysis on the existing joists? I think the potential for load reversal due to the interstitial, half-span beams is valid. Especially with the tight joist spacings for snow loads I've seen in Mass. With the stiff beams, the deck is going to be holding the joist up rather than let it deflect freely, and this effect will be the most noticeable at the new line of supports near midspan.

I'd try modeling the existing joist and then add spring restraints in the vertical equal to the flexural stiffness of the deck spanning the distance between the interstitial beams. If you really want to get fancy, model fake beams with the flexural stiffness of the deck and support them on the interstitial beams in the model for a really accurate look.

Oh, and do this with your monitor turned off, wearing an eye patch, and probably with your mouse unplugged for good measure.

 

[SteelPE]

We know the size of the joist but not the configuration at this time. We may not be able to take measurements for some time. We could figure the moment of inertia of the joist and model a fake beam with similar moments of inertia.... then run your spring analysis with the moment of inertia of the deck (if I can figure that out).... and then check for a negative moment on the beam (which is where stress reversal would take place).

This isn’t going to be easy.

 

[phamENG]

That could work.

I'm really curious about this space that can't be accessed. Unless it's a Covid sample testing lab, what could possibly so vital that it not be touched? (I know, you probably can't tell us - the question is mostly rhetorical.)

 

[milkshakelake]

You could install bridging in the area you can access. I don’t know the code specifics but bridging helps even if it doesn’t go wall to wall. The series of joists bridging each other help lateral torsional buckling.

The solution in your sketch works but it’s theoretical. By the time you have a major blizzard, the deck and old joists have to deflect 4” before the new joists take over, which is plastic deformation and damage of insulation and membrane. It will save the occupants of the building but the roof will be destroyed. You can reduce the gap. It’s hard to analyze.

 

[r13]

 

[SteelPE]

By the time you have a major blizzard, the deck and old joists have to deflect 4” before the new joists take over, which is plastic deformation and damage of insulation and membrane. It will save the occupants of the building but the roof will be destroyed. You can reduce the gap. It’s hard to analyze.

I have no idea what you are talking about. The cross beam is installed 4" below the existing joists in order to eliminate the possibility of the joists "accidentally" bearing on the new beam framing perpendicular to them. We would be installing new beams in-between the existing joists that would be tight to the underside of the existing deck. these beams would be supported by stub columns down to the beam 4" below the existing joists. Is it what I want to do??? No. It's not what I proposed 4 months ago when I took the job. Now I am being told no reinforcing is possible.

What is in the room? Not exactly sure, but it's some sort of data room for local internet or phone service (as in, the entire area not just the building).

Retired.

I am trying to eliminate the possibility of stress reversal. If I were to attempt a similar repair to what you are showing I would just put the beam to the underside of the existing joists.

 

[structSU10]

The only problem I see is deflection compatibility at the new support line. the new reinforcing beam want to move zero at its new columns, or a relatively small amount midspan of the new support beam, but the joists next to them want to drop more due to it being near midspan - you will be binding things up which may be effectively providing support for the joists at midspan via metal deck 'bridging'. Perhaps the deflections can be managed, but it may be worth a study to see if you get load reversal in the joists.

 

[Celt83]

I can easily design the new support system to support all of the loads in the bay if necessary.

This would be my approach rather than spending the time to do the relative stiffness analysis and potentially end up at the same beam/joist sizes anyway. Could always write a letter/email to accompany the reinforcement documents outlining the increase in snow drift and that the imposed access restrictions have severely limited the reinforcement options.

My Personal Open Source Structural Applications:

https://github.com/buddyd16/Structural-Engineering

Open Source Structural GitHub Group:

https://github.com/open-struct-engineer

 

[SteelPE]

Hopefully I can explain this properly.

I decided to run an approximate stiffness analysis assuming the joist was a beam by calculating SJI moment of Inertia (MOI). I ended up using a W14x30.

I calculated the deck stiffness and made a beam with springs every 2’ equivalent to the stiffness of the deck (units were a pain as my software was using stiffness as kips/ft). I used 2 deflection equations to figure the stiffness AISC table 3-23 diagram 7 and AISC table 3-23 diagram 16.

I figured that all of the dead load is currently on the roof joist so the only additional load the system was going to see was the additional snow loading (uniform and drift). I also figured that there is a positive moment already built up in the joists that the snow load moment needs to overcome before the BC goes into compression.

Using diagram 7 I figured no negative moment was being built up in the W14x30
Using diagram 16 I figured there is a slight negative moment in the bottom chord but the initial dead load moment is far greater than this moment, so the bottom chord should never see compression.

This is probably the best analysis I can do. Also I am thinking figure 16 is never going to be possible as the whole system is going have some stiffness reduction because nothing is infinitely rigid.

So it seems like the proposed system I outlined above is feasible and the concern regarding stiffness of the new system introducing negative moment into the existing joist was not valid.

 

[r13]

I gave up, it was a modification of your original plan. Without touching the existing joists, where is reversal to occur?!!!

 

[SteelPE]

retired,

The thought was, introducing a set of beams in-between the existing joists with support for those beams at mid span. This reinforcing with support mid span would be much stiffer than the original system accidentally causing a support at mid span of the existing joists.

 

[r13]

The members around the stiffer joint will deflect together, and the stress in the existing joist will alter, but to the better. There was a thread a few days ago about strengthening the existing joist with midspan support for additional load. A quick evaluation has found that all members are well within allowable with inevitable stress reversal. Also, a compressible material can be add to the top chord of the new assembly to ensure it picks up load only after the existing joist has had its share.

 

[phamENG]

SteelPE - glad to hear there are no major complications. I agree it's not a first choice, but it sounds reasonable and feasible - especially after at least a basic relative stiffness check.

 

[milkshakelake]

I can't figure out how to do quotes, so bear with me.

"I have no idea what you are talking about. The cross beam is installed 4" below the existing joists in order to eliminate the possibility of the joists "accidentally" bearing on the new beam framing perpendicular to them. We would be installing new beams in-between the existing joists that would be tight to the underside of the existing deck."

Thanks for clarifying. I totally misunderstood your idea. I didn't know that the new joist beams would be tight with the deck. Somehow I thought they'd be 4" below. This can totally work. However, I'd be worried that the deck would "pick up" the existing joists in a way as phamENG mentioned. Also, I didn't see anything wrong with my idea so far besides the negative moment. Also retired13 had a good idea about dealing with the negative moment. Anyway, if you're set in your way, good luck.

 

[Ron247]

****
The person with a cloud over their head shown at the 4 asterisks--------^
You still have to type the name of the person you are quoting and have to cut and past the comment you are quoting. Unless there is an easier way I do not know how to do

I can't figure out how to do quotes, so bear with me.

 

[r13]

milk,

1) Copy the quote.
2) Open the reply window, and click the icon

.
3) Provide the person's name in the box, or just click on "OK" if you don't wish to provide the name.

4) Paste the quote in between the square brackets,

.
5) Submit post.