Capacity Design of Roof Deck

[Gumpmaster]

In the last couple of years, several professor types have been throwing out the idea of designing roof decks (diaphragms) to exceed the capacity of the seismic lateral force resisting system. Apparently, this is already the case in Canada. A couple of questions:

-Are there any instances, worldwide, of a deck failing in an earthquake and that failure being the root failure for the structure?
-If the answer to the above question is "no", then why introduce a new requirement where there is apparently no real-world problem?
-Are there any Canadians out there who've dealt with this? What are the practical effects? Is it an onerous requirement?

 

[GalileoG]

Canadian designer here. I have not seen many people here take that requirement to heart (including myself.) Would be interesting to see what people have to say about this...

Clansman

If a builder has built a house for a man and has not made his work sound, and the house which he has built has fallen down and so caused the death of the householder, that builder shall be put to death." Code of Hammurabi, c.2040 B.C.

 

[msquared48]

I don't get this idea, for we always design the diaphragm to equal or exceed the required seismic force already. After all, when do we actually design to the exact force without some inherent reserve in the diaphragm capacity? Answer is that we don't.

What percent overdesign are they talking about?

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[Gumpmaster]

The concept is that your braced frame is your seismic fuse. It's designed to yield at the EQ level (hence an R greater than 1). The actual EQ force is higher than the force we design for because we take into account post-elastic response with the response modification factor. In order to transmit that actual force to the braced frame, you need to design the connected element for the actual force, hence the omega factor. With an SCBF you would need to design the deck for 2 times the load with which you design the braces.

The professor types are taking a different tack and say design the deck for a capacity greater than the CAPACITY (in tension) of your brace. This can often be a huge number, but it would ensure that the brace is the failure mechanism.

 

[steellion]

Yep, Gumpmaster has it. I actually just went to a diaphragm design seminar yesterday (in America). There are already provisions in ASCE 7-05 requiring diaphragm design for forces that exceed the equivalent lateral force requirements of seismic design of the braced frames/shear walls. If the diaphragm is disabled, the lateral FRS is useless because it has no way to transmit the shear. Better to overdesign the diaphragm so the (ductile) failure is in the lateral FRS and not the diaphragm.

 

[slickdeals]

I am not sure whether a diaphragm needs to be designed for omega level forces. The diaphragm to braced frame would make sense to be designed for the omega level forces and the connections of the braced frame would too.

But would you be designing your diaphragm to remain elastic during a seismic event? That seems like an overkill.

Thoughts?

 

[sandman21]

Omega does not account for the actual loads to a brace, it is a code factor to ensure that certain items in the LFRS stay in there elastic ranger so that the inelastic range is reserve capacity, because code earthquake loads are several times smaller than the expected loads on a building. I just took a quick look at a SCBF job I just did, the expected yield of the brace is 3.6 times higher than code earthquake. Good luck trying to get a deck to work for the expected load of a brace.

 

[JedClampett]

If I understand this, and I'm not sure I do, you're going to overdesign the diaphragm so it's not the weakest link. But I'm willing to bet that the connection between the diaphragm and the rest of the lateral FRS is the weak link. For CMU, you have a bunch of very brittle, potentially poorly installed anchor bolts tansferring the forces to the diaphragm. And that's if they do it per design. Most of the time the contractors ask to replace the CMU AB's with some epoxy contraption with questionable properties. Making the diaphragm stronger isn't going to improve their capacity.
If they're worried about the diaphragm capacity, I would prefer to add a horizontal truss instead of making the deck heavier and adding welds.

 

[Gumpmaster]

So, If you're Curious as to where my question came from, there's an article in the First Quarter 2010 AISC Engineering Journal about it. Give it a read for a more in depth look.

My first question was "Are there any instances, worldwide, of a deck failing in an earthquake and that failure being the root failure for the structure?". If I had to guess, I'd guess likely not. I never heard about anything like this in Northridge or Loma Prieta. Never heard about it in Hanshin, or Sichuan. Are they trying to solve a non-existent problem?

I agree, designing your roof deck (and the connections to the vertical bracing) for 3.6 times higher than the EQ would be near impossible. The resulting structure would cost significantly more than one designed under the current code. Are we codifying ourselves out of a job? What about when the next owner says "Gee, I'd really like a new building, but $500 per square foot? I'll stick with what I have".

 

[slickdeals]

I believe there has been precedence where the diaphragm to LRS has failed. As I stated before, and I agree with Jed that if anything needs careful attention, it is the connection of the diaphragm to the LRS.

 

[Gumpmaster]

There is precidence for the roof to wall connection failing, but I think it's for the out of plane wall to the roof. One instance is in CMU structures with wood diaphragms. Some of these failed in cross grain bending in the California earthquakes. I'm not sure if the same thing happened with metal deck diaphragm buildings or not, but it's definately possible. The current ASCE 7-05 has several clauses to preclude this now.

Do you know of any in-plane diaphragm/wall connections that failed though? I suppose it's possible.

That's not the point of the Engineering Journal article though. It's pointing to the diaphragm as a whole as the thing that needs to be strengthened.

 

[jheidt2543]

A question has occured to me, why not call for some axial capacity in roof joist to carry some of load that would normally go into the diaphram? If joist are spaced at 4', 5', 6' the loading would not accumulate too quickly and then be small enough that you might be able pick up some of that load. Of course, that only works in one direction, unless you change direction of the joist in the last bay.

Just wondering.

 

[steellion]

In ASCE, the overstrength factor only applies to chords, collectors, and their connections, not the diaphragm itself. The force in the diaphragm is a function of the weight of the floor and all floors above it. Since the roof is the top "floor", the equation works out so that on the roof, the EQ force for the lateral FRS and the design force on the diaphragm are equal. The diaphragm force will be greater than the braced frame/shear wall force on elevated non-roof floors, but on those floors you will have concrete to help the diaphragm meet code requirements.