Reinforcing for renovations

[EngDM]

I'm wondering what the general opinion is when reinforcing members for a renovation. For instance, we do a lot of RTU fitup jobs. I just got one that the existing drawings shown a 20psf roof assembly and loading, but when we went to measure up joists on site we noticed that it is now a ballasted roof with quite large river rock, which has been determined to be in the ballpark of 27psf. Further, the code has increased snow loads and the original build did not design for any snow drift cases. It's also littered with cantilevered beams that have nothing by way of bottom flange bracing. As a result of this, the existing members under this new loading are WILDLY under capacity at the negative bending locations, and the positive bending is exceeded on a simple span just due to the new dead load.

I have reinforced any member that is directly influenced by our new RTU to support the new loads, but I'm wondering what the consensus is about what should be done about the remaining structure now that it is a ballasted roof.

 

[lexpatrie]

I'm getting more and more of these RTU jobs too.

You cannot turn a blind eye to the existing apparent overstress. Either that's a letter to ownership, a letter to the building department or something to disclose the issue.

This sounds a lot like something I read lately, (ignore the interpretation of the code, because in my opinion the guy is totally incorrect on it), but it did have something similar - an extra 10 psf from a ballasted roof (I guess I'll make the leap and presume it actually is ballast, at 10 psf, a lot of engineers call a built-up roof at 5.5 - 6.0 psf to have "ballast" and that's incorrect (e.g. a certain roof inspection manual calls gravel "ballast" constantly), anyway, then they put spray foam over that, may have encapsulated a wet or partially wet roof, and then snow/rain collapsed it. The writer makes some highly incorrect assertions regarding design standards (Steel joist versus AISC), but the underlying elements are common to what you're describing, so it's worth some consideration.

Roof Collapse, Whitley, EDT Engineers web site, Oct 1, 2020.

I don't think you should provide a reinforcing design for this situation as the roof is already failing on paper. Either something has to be done in daylight to reduce the roof dead load globally, or the RTUs should not be placed on the roof. It almost sounds like you've got an existing structural deficiency as well as excessive dead load... which was potentially added after the fact, perhaps not. I suppose if the original roof was designed correctly (very speculative at this point here), it's potentially "okay" as an existing condition, but with the excess 10 psf that seems unlikely.

Or did you reinforce the areas that support the new RTUs to accommodate the existing dead load issue at 27 psf as well as resolve the bracing, and reinforce the steel beams as well, what about the shear splices at the suspended beams? If that's the case, I think you're in a better situation, but it should still be disclosed to the owner/building department that the roof seems to be overstressed due to excess dead load. So long as they keep the scuppers and roof drains clear, there's at least the potential the roof structure will survive until it can be reinforced or the overload removed (I say potentially, because the scupper and roof drains can be undersized, particularly if the rainfall intensity has gone upward since the original design). Also, an older roof may have effectively zero slope which puts it in a more vulnerable situation regarding pooling water which can produce ponding.

I think you'd also want to see a web stiffener at the roof beam where it lands on the column below.

 

[EngDM]

Or did you reinforce the areas that support the new RTUs to accommodate the existing dead load issue at 27 psf as well as resolve the bracing, and reinforce the steel beams as well, what about the shear splices at the suspended beams? If that's the case, I think you're in a better situation, but it should still be disclosed to the owner/building department that the roof seems to be overstressed due to excess dead load. So long as they keep the scuppers and roof drains clear, there's at least the potential the roof structure will survive until it can be reinforced or the overload removed (I say potentially, because the scupper and roof drains can be undersized, particularly if the rainfall intensity has gone upward since the original design). Also, an older roof may have effectively zero slope which puts it in a more vulnerable situation regarding pooling water which can produce ponding.

I reinforced the joists and beams being affected by the RTU for the 27psf and the RTU, not just for the additional load due to RTU. Lukily I have all of the existing sizes so I can check pretty accurately. For the cantilever beam adjacent to my space I also added bottom bracing to resolve it's negative bending deficiency. I brought that whole area "up to code", but I haven't addressed anything in other areas of the building.

The only thing I didn't take a look at is the shear splice, but what we ended up doing to reinforce was add a clearspan beam below this whole area to pick up 100% of the load (a little conservative I know, but alternatively I'd be reinforcing beams that are not in my space with full on angle's top and bottom instead of just the bottom flange bracing) so the connection is unloaded. New connections to the column are going to be provided.

I think the building owner just got their roof done one time, and the roofer did a ballasted roof under the guise of it being cheaper.

a lot of engineers call a built-up roof at 5.5 - 6.0 psf to have "ballast" and that's incorrect

Agreed, a lot of the time a tar and pea rock roof is referred to as ballasted when in reality it's much lighter. This is fine when going from ballasted to tar, but going from pea rock to full on river rock ballast is asking for trouble.

 

[canwesteng]

I'd reinforce the areas that see snow drift from the RTU as well, and then a letter explaining to the owner that the ballasted roof exceeds the original design dead load. I don't believe you have any responsibility to address increases in snow load that come with new code revisions. Also, for cantilevered beams, you brace the top (tension) flange, except for the case of wind uplift, do you mean no bracing at the support locations they run over the columns?

 

[lexpatrie]

I think you're in a decent zone, considering, as you addressed the affected area, and considered the overload in that area, at least.

Well, you can overload in both directions, the gravity load case when you increase from builit-up to ballasted, and if you drop the dead load that can produce uplift in the beams in a somewhat pathological sense. It may not happen... but a drop from 10 psf (x 0.6, say, for ASD) to 5.5-6.0 (x 0.6, again, for ASD), may create some uplift that wasn't present before. It's a net change of 4 x 0.6 (2.4 psf for me, today), I suppose I'm omitting the 0.6 psf for the membrane, too.

 

[EngDM]

I'd reinforce the areas that see snow drift from the RTU as well, and then a letter explaining to the owner that the ballasted roof exceeds the original design dead load.

Luckily 2020 NBCC allows us to neglect snow drift for projections smaller than 3m.

Also, for cantilevered beams, you brace the top (tension) flange, except for the case of wind uplift, do you mean no bracing at the support locations they run over the columns?

They are braced with tie joists at the column, but for whatever reason the beam line is designed such that the double cantilevered spans are shorter than the drop in sections, so the entire beam is in negative bending in some load cases, so my unbraced length is clearspan of columns. I need to add bottom bracing to cut this span down to actually work. I suspect this is because of the higher dead load being a massive point load at the cantilevered end, and not enough clearspan between columns to effectively counteract the induced negative bending.

Well, you can overload in both directions, the gravity load case when you increase from builit-up to ballasted, and if you drop the dead load that can produce uplift in the beams in a somewhat pathological sense. It may not happen... but a drop from 10 psf (x 0.6, say, for ASD) to 5.5-6.0 (x 0.6, again, for ASD), may create some uplift that wasn't present before. It's a net change of 4 x 0.6 (2.4 psf for me, today), I suppose I'm omitting the 0.6 psf for the membrane, too.

Lately what I've been providing for uplift is just what I know will stay. Deck, joists, and a membrane. Anything else gets omitted from my uplift calcs. Conservative yes, but if I design for ballast and they ever change it, I haven't designed my building for its design life (in my opinion).

 

[JAE]

Bottom line you need to tell (shout) to someone that there is a serious issue with the existing roof.

 

[Greenalleycat]

We've had a very similar situation recently
We were providing engineering advice to a company installing abseil anchors to an existing timber roof
It turns out that half the rafters had no fixings at all to the supporting walls so we've written a letter that we will be passing to our client to pass to the building owner
It highlights that we've identified this issue and we think it is a concern because XYZ so we recommend that you do ABC to address it

Note that I don't see this obligation as something that you should do every time that you see a noncompliance in old buildings
However, we do have an ethical obligation when we consider that the failure poses a legitimate risk to safety (rather than say, to comfort from an undersized joist causing a serviceability issuue)
I would consider your case a potential safety issue and hence it should be documented with a letter to the building owner

 

[dik]

With RTUs the biggest problem is often snow accumulation. This can easily exceed the weight of the RTU. The web stiffener denoted as a single shear plate should be located on each side of the web. It's purpose is not normally for web crippling, but is to maintain the cross section shape in the event of overloading.

You tell the owner about the issue not the building department. If the owner does nothing and the situation is dangerous... then contact the building department, only after informing the owner that you are obligated to.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[dik]

This was a recent project where the EOR had provided the details below. The W beams were W12x35 with simple spans of 20'. My proposed framing using continuity and splices at 3' from c/l of column allowed a W12x30 to be used with the OK of the EOR. My client had already sourced the W12x35. I'm not sure how the web stiffeners on the EORs details worked. New framing worked and is a lot easier.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[dik]

and even more load if they have changed it from conventional to an inverted roof, where the river stone gravel is used to ballast the insulation.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[lexpatrie]

Least weight is not lowest installed cost.

What the EOR showed, with the stiffeners is probably there to brace the bottom flange and stabilize the column. (See AISC/SSRC Basic Design for Stability, Lecture 4, slide 56).

As far as snow always being an issue with an RTU, you must be dealing with much larger RTUs than I do, mine are in the 4 ton range and about 3' x 3', if that. Usually it's the chillers and the larger units that create snow drifts as they are 15+ feet long in one direction (ASCE 7, here). There's the option of elevating them (required in Florida, not that it's a snow region), so the drift doesn't occur. Why would you bother with a moment splice on that revision anyway?

I don't see much logic in contacting the owner, waiting, and then contacting the building official. Let alone creating extra effort. Does one schedule something on their calendar so they remember? Why bother. If this is something you think is a dangerous, hazardous or a condition worth writing a letter, why is the building department being kept in the dark for an unspecified period of time? To allow the owner to make a "fix" that's not designed, permitted or inspected? If it can be resolved by removing the ballast, then say so.

Interesting point on the protected membrane roof, as the insulation could get wet and store water, adding to the dead weight.

 

[CDLD]

They are braced with tie joists at the column, but for whatever reason the beam line is designed such that the double cantilevered spans are shorter than the drop in sections, so the entire beam is in negative bending in some load cases, so my unbraced length is clearspan of columns. I need to add bottom bracing to cut this span down to actually work. I suspect this is because of the higher dead load being a massive point load at the cantilevered end, and not enough clearspan between columns to effectively counteract the induced negative bending.

Quick tip: you can use an increased Cb factor (w2 in canada) for this situation (Assuming you have twist prevented at supports and top flange support).
Yura - Bracing for Stability

AISC commentary

 

[JoshPlumSE]

I have reinforced any member that is directly influenced by our new RTU to support the new loads, but I'm wondering what the consensus is about what should be done about the remaining structure now that it is a ballasted roof.

Here's what I did for one project in the past. There was a relatively low sloped roof that had a lot of moisture damage / dry rot in certain areas. So bad that you had to avoid walking on those areas when you were on the roof because it felt like it would collapse under a single person's weight. Awful, awful stuff.

There were some very poorly designed trusses that were common for the whole roof. They were at least 30% overstressed, if I remember correctly. My design drawings included two details for the new trusses:
[ul]
[li]The first detail was required for the locations where we were removing a damaged truss. This detail was fully up to current code.[/li]
[li]The second detail was labeled as an "optional" and less expensive upgrade of the undamaged trusses that would improve their performance to be close to what the current code requirements would be. It might have just been doubling up the top chord and ensuring that the web members were properly connected.[/li]
[/ul]


I'm not saying that this "two detail" design should be used in all cases. Just that it worked out well for that project. And, that I only REQURIED upgrades on the members subjected to new conditions, but informed the owner about how the existing members (that were not required to be upgraded) did not meet current code standards and that they might want to consider upgrading them.

 

[dik]

Least weight is not lowest installed cost.

Concur... but two 60' W12x35 would normally be more costly than two W12x30, with connections and details being the same.

What the EOR showed, with the stiffeners is probably there to brace the bottom flange and stabilize the column. (See AISC/SSRC Basic Design for Stability, Lecture 4, slide 56).

I'll take a look...

As far as snow always being an issue with an RTU, you must be dealing with much larger RTUs than I do, mine are in the 4 ton range and about 3' x 3', if that.

RTUs are often 6'x8'x12'... I guess it depends on the project.

Usually it's the chillers and the larger units that create snow drifts as they are 15+ feet long in one direction (ASCE 7, here). There's the option of elevating them (required in Florida, not that it's a snow region), so the drift doesn't occur. Why would you bother with a moment splice on that revision anyway?

Beams designed plastically and you need to develop 25% of Mr. The section is Class 1. A W12x26 works, but is not Class 1... else it could be smaller.

Interesting point on the protected membrane roof, as the insulation could get wet and store water, adding to the dead weight.

They usually use insulation that does not absorb water. The water greatly reduces the insulation quality. Inverted roofs are not so common now but were fairly common two decades back.

-----*****-----
So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik