Analyzing and code checking a pre-engineered metal building for new rooftop RTU loading 8

[JMASE]

Does anyone have any insight on proper methods to analyze and code check a 1970's era pre-engineered metal building (tapered columns, moment haunch, tapered beams, etc.)? The client wants to add 4 -6 new rooftop RTUs. Will likely need to load the PEMB purlins too. I'v attached a typical section to get the gist.
Thanks

 

[SandwichEngine]

The odds of you getting it to work are low. I'm a PEMB guy and buildings we designed a few years ago wouldn't be shown to work for the newer wind loads. Not only has the wind loads gotten stronger according to the code, the wind maps are totally different. Because the IEBC required anything to be brought up to code if you add to the loading I'd just use beams and not load any purlins.

I've seen it reported here that RISA does PEMB frames right but can't verify.

 

[canwesteng]

Going back to the 70s you might actually have some luck, I've found PEMB to be designed a little more robust back then, and AISC is less conservative if you switch to limit states design. The purlins are almost certainly a wash though.

 

[JMASE]

Sounds like a good approach: bypass the purlins and span new dunnage to the frames. Has anyone used AISC Design Guide 25 for frame design?

 

[canwesteng]

Do you have access to a simple FEA program like STAAD or SAP? My method is to use STAAD to model the tapered frame, and then verify select results from the design module against design guide 25. Otherwise figuring out internal forces can be quite difficult.

 

[Eng16080]

Last time I took a closer look at a PEMB, I wondered how any of it worked, similar to prefab. wood trusses. I mean no disrespect to those industries, though. I think they just squeeze every last drop of capacity out of everything.

One thing I've done before is isolate the new loading to the column locations, reinforcing the columns as necessary. If you try to add load to a purlin, then you probably overload the purlin, the supporting frame(s), and the supporting column(s). It's easier to just deal with the column(s) and bypass the rest of it.

Of course, if adding the rooftop units means that you're now certifying the entire structure meets current code, good luck.

 

[JMASE]

Yes, I'll need to look carefully at the IEBC. Thanks

 

[phamENG]

DG 25 is the basis for the tapered design checks in RISA. It works well for this, though you'll be on your own for the connections.

 

[Aesur]

I would typically approach this by adding new purlins to support the units and span to the main beams and see if we can get within the 5%, if not start taking measurements and use RISA or some other program to check the girders. If you find you have to bring up to code, which I highly doubt would be the case for simple mechanical unit additions, you will find it hard to do so.

I am still puzzled at how they fine tune every little thing for PEMB's and push the cost to the foundations; the last package I reviewed now takes member utilizations up to 103 and 105% and they had some paper to justify it, I forget how or what it was, but I believe it had something to do with most engineers allowing 5% over based on IEBC. I have started adding to my GSN that the utilizations may not exceed 100% as I couldn't win the argument with them to make them change it to 100% as "it added too much cost to adjust a few members". I am of the opinion that PEMB's are garbage and just push the costs to the foundations and are not easily modified for future uses. I once did an exercise where I proved you could build a conventional framed building for sim cost to the PEMB.

 

[dik]

It's a real issue if you have snow accumulation. I like the engineering that goes into PEMBs, but consider them as 'throw away' buildings in the longer term.

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

-Dik

 

[bones206]

Last time I did this, I convinced the client to group all the new RTU's in one central location and provided a large dunnage platform with it's own independent columns, lateral system and footings. I got beat up by the owner, architect and GC, but in the end they are happy with it. It's hard to convince non-structural engineers how costly it can be to reverse engineer and upgrade these types of buildings. At least the OP has drawings to work with, which is a minor miracle. Although 9 Gauge webs do not inspire confidence.

 

[Eng16080]

the last package I reviewed now takes member utilizations up to 103 and 105% and they had some paper to justify it

It's hard to believe this is legal. Even if 5% over is allowed for an existing building, it's not existing when it's originally designed. How would a future engineer trying to add RTUs know that the 5% extra was already used? WTF!

 

[JMASE]

bones206. I was coming to the same idea last night...new central dunnage on new braced columns bypassing the existing frame, and structurally independent of the lateral PEMB frame. Might get the slab to work in lieu of new footings with relatively light RTUs. Sort of a mezzanine, but above the roof.

 

[WesternJeb]

How would a future engineer trying to add RTUs know that the 5% extra was already used?

That is not how the code reads to me. To me it is more of a, "well it has stood for this long without an issue. An extra 5% won't hurt it." It is not that engineers can take existing structures up to 105% of the design capacity estimated by the engineer in charge of renovation.

 

[bones206]

JMASE, if you can utilize some nice rigid moment connections and utilize a spring constant for drift calcs rather than a pure pin at the base plates, you might be able to omit braces. I had to go that route in my project, as braces would have destroyed the usability of the floor space.

 

[Eng16080]

"well it has stood for this long without an issue. An extra 5% won't hurt it."

It is not that engineers can take existing structures up to 105% of the design capacity estimated by the engineer in charge of renovation.

If a building is designed to 100% utilization, then aren't these both the same? If IEBC says it's ok to add 5% more load, presumably that means the structure will now be at 105% of the design capacity. No?

And based on what Aesur stated above, some of these PEMB are being designed at 105% from the start. How would you approach a renovation in that case? How would you even know it was originally designed at 105%? You're saying that another 5% is allowed by IEBC as long as the building is still standing at the time of the renovation? If you keep renovating, do you get to keep adding 5% every time? I know I'm getting carried away here, but I'm just curious how others approach this.

In any case, designing a PEMB for 105% utilization seems wrong.

 

[canwesteng]

The actual risk to a building of being designed to 103% is negligible. I'd argue nothing we do has that much precision, and it makes sense to round utilization down to two sig figs.

 

[WesternJeb]

If you keep renovating, do you get to keep adding 5% every time?

That is a rabbit hole I have gone down as well. This is an "engineering judgement" to me I guess, and evaluating what all others have done based on how nasty the scabs are.

I do my own evaluation to whatever the original design load was and determine the 5% rule based off of that.. I don't truly look at the capacities of the members unless I have a major red flag being thrown up due to undersized or already modified members.

edit: I agree with canwesteng.

 

[phamENG]

While I agree that we're not precise enough to worry that much...there still needs to be a line or we'll round and justify ourselves into oblivion. I like 100% as that line. 100.5%? Okay. 102%? No. But that's me.

If a building is designed to 100% utilization, then aren't these both the same?

As a matter of fact, they are not. What is a safety factor for? Material defects, defects in workmanship, loading above anticipated loading, etc., right? While a building exists on paper, these are all possible and have a certain probability that feeds into the determination of what that safety factor is/should be. Once the building goes from paper to real life, the math changes. The longer it stands, the more the math changes. If a building has stood for 40 years, the probability that there is a remarkable material or workmanship defect goes down a lot. It's not zero, but it's much smaller than before the work was done. As such, the necessary safety factor changes a bit. Codifying or creating a coherent and practical system for this is probably a bit beyond what anyone wants to attempt...so bumping the allowable capacity for proven systems up by 5% is probably good enough.

 

[Eng16080]

The actual risk to a building of being designed to 103% is negligible.

I agree that it's very unlikely that 103% will every be an issue. But if I'm using AISC 360, for example, to design a steel beam, 103% is technically NOT OK. If 100% is the limit allowed by code, I take it that that limit was chosen for a reason, based on a statistical analysis of worst case things that might occur (loading, material strength, occupant risk, etc.) and accounting for some safety factor.

For me, I generally trust these codes because, what do I know? I haven't studied structural failures. Maybe 103% is ok. Is 110%? How about 125%? I have no clue. Anybody can say that 3% more or less of something is a small amount and will probably have no consequence. I think that is probably true here, but again, I have no further justification for it.

If I design a critical component of a structure at 103% and then the structure fails, even if the failure is for an unrelated reason, how can I justify that my design was code-compliant?

(Sorry OP for going slightly off topic here)