Elastic Seismic Event (relating to curtain wall) 3

[CCWG]

A typical performance requirement for a curtain wall, as it relates to building drift, is: 1. nothing falls off the building in a design level earthquake, and 2. it remains air/water tight after a service level event (10 year wind event usually).

I've recently seen the requirement for the curtain wall to remain air/water tight after the greater of a 10 year wind event, or an elastic seismic event. From my end as the specialty engineer it's just a matter of requesting the values, but I am curious what an "elastic level seismic event" really means. My understanding would be this a performance-based design approach, and in essence, just a way of saying the curtain wall will remain air/water tight after an earthquake that is less severe than the design level event. Which makes sense to me in theory - no permanent deformation in the building, so no performance decrease with its facade.

Let's say I'm given the design level drift for an earthquake as H/100 - which I think is fair to say causes inelastic behavior in the building. Is there an easy way of converting the H/100 drift value to one that corresponds to an elastic seismic event? I'm also curious if there are any other parts of the building that must maintain performance after a lesser earthquake (assuming my understanding is correct)

 

[skeletron]

If you know the target deflection, could you back calculate the distributed force by assuming a cantilever beam type scenario. Then if you have the distributed force, you could calculate the total force applied which would be related to the spectral acceleration that could in turn be related to the POE or hazard level.

I guess the "elastic level" would mean that the force you calculate above would not have a reduction (RdRo) factor applied (ie. RdRo = 1.0).

This would be my first shot in the dark to get from deflection to force to a guesstimate of the seismic event.

 

[KootK]

I'm guessing too but one simple interpretation might be that your elastic drift is simply your design event total drift divided by the Cd inelastic drift amplification factor. So, if Cd were three, the elastic drift for a design event h/100 would be h/300.

 

[rowingengineer]

There is a lot of noise around this subject. The steel stud designers are being pushed to come up with low damage designs. Most are looking at top plates that don't fix to stud to get the drift levels from 0.3 to 1.5 ish. Personally I want cladding that has better displacement values.

 

[HTURKAK]

Just guessing based on my knowledge from other industries ;

This is a performance level for an EQ event could reasonably occur once in service life of the bldg ( 50 yrs ) which corresponds 50% probability of exceedance in 100 years.

Three level of EQ

- Fully operational , the structure and non str. components fully elastic ;50 yrs return period
- Operational but with small repairs etc (OBE level, DBE ) 475 yrs return period, (Operation Base EQ , Design Base EQ )
- Safe shut down or collapse prevention ( SSE level, MCE ) 2500 yrs return period.. ( Safe Shutdown EQ, Max. Con.. EQ)

I suspect , this is a new invented criteria to circumvent or say relaxing the requirements of DBE ..

My opinion ONLY !!

EDIT= Eurocode seismic EC- 8 has a similar level of EQ ,serviceability seismic action with 10% exceedance probability in 10 years (mean return period: 95 years).

 

[CCWG]

@skeletron - I'll have to digest that a bit more. I just do misc. small delegated stuff so building design is out of my realm. Thank you for the input

@KootK - I thought about this...reading through ASCE chapter 12 I understood the "elastic drift" gets multiplied by Cd to get a give more realistic value. But the "elastic drift" in itself is theoretical since it is assumed to be linear even if the displacements would cause stress beyond the yield of the building elements (or so I understand). So it's sort of like using the elastic drift corresponding to a useful definition, and not by it's value in reality. But I suppose that's true for many things we do

@rowingengineer - I think lateral slip between floors makes a lot of sense. A challenge I'd expect with that approach is offending architects with the large joints required at any corners/returns.

@HTURKAK - Thanks, those are acronyms I have not seen before. Looks like they lead down a bit of a rabbit hole I can go down

 

[XR250]

I'd expect with that approach is offending architects with the large joints required at any corners/returns.

When this tried to become a thing in my area a decade ago, my question to the Arch was exactly that - WTF are you going to do at the corners?
Happy to not be doing LGS design anymore!

 

[ggcdn]

If this is coming from a USA project, I can provide a bit of context about what I’ve done for performance based design projects in California. I can’t guarantee this corresponds to what you’re seeing though.

PBD in California tends to follow 1 of 2 guidelines, PEER TBI or LATBSDC. Both these guidelines require that the structure be analyzed for a service-level earthquake (SLE) with 40% probability of exceedance in 50 years with an elastic model (The collapse prevention nonlinear analysis is 2%/50yr and Code based design is 10%/50 yr). I haven’t really noticed a good rule of thumb between the SLE and collapse prevention drifts. In general the buildings I’ve schemed (40-50 story high rises) had peak SLE story drifts around 0.3% to 0.5%. The LATBSDC performance criteria is to be less than 0.5%.

It seems plausible that a peer review panel would ask the EOR to add components and cladding criteria for the SLE case, since we are checking the base structural components for that case as well.

-JA
try [link calcs.app]Calcs.app[/url] and let me know what you think

 

[KootK]

I thought about this...reading through ASCE chapter 12 I understood the "elastic drift" gets multiplied by Cd to get a give more realistic value.

Having slept on this, I feel pretty good about my recommendation for a building designed to ELF procedures. For buildings that are truly performance based design, I'd expect fancier definitions of the serviceability criteria than what you've expressed.

But the "elastic drift" in itself is theoretical since it is assumed to be linear even if the displacements would cause stress beyond the yield of the building elements (or so I understand).

I don't agree with that in the context of Cd. In that context, the elastic drift is that which would occur at the point in time at which the building elements ceased to respond elastically. So the assumption of linearly increasing drift should be as valid as it ever is.

So it's sort of like using the elastic drift corresponding to a useful definition, and not by it's value in reality. But I suppose that's true for many things we do

.

I certainly agree with the useful definition bit. My sense is that the EOR is basically just saying "Up to and including any seismic events that for which the primary structural would remain undamaged and elastic, we'd like also like the cladding to remain undamaged and functional". It's a useful definition of what the client & design team desires for the performance of the building. Sort of a poor man's performance based design.

I'm also curious if there are any other parts of the building that must maintain performance after a lesser earthquake (assuming my understanding is correct)

By extension of the same concept all of an ELF designed primary structure would be designed to the same criteria since, by definition, it will all be designed to remained elastic up to the intensity of this "elastic level seismic event". This is obviously circular reasoning after a fashion.

 

[CCWG]

@XR250 - yep, you end up debating about it for a month before the GC pushes and the design team decides it's OK to ignore the requirements they specified based on the first person to propose some feel-good logic. Which is probably not all wrong, but it can get annoying to have the same arguments every project.

@Sonofatkins - thank you that is exactly the kind of insight I was hoping to get here

@KootK - Figure 2 of the article I linked below illustrates what I was trying to explain. I understood the elastic drift (de corresponding to Ve) is theoretical in the sense that it seems to be calculated based on the assumption the structure behaves elastically indefinitely...then modified by other factors after to estimate the actual drift. If that figure could be considered accurate for this scenario then I'd understand "d" to be the elastic drift, and "de" to be theoretical - but probably still useful for the purpose of facade serviceability checks. But I'm just a misc. specialty stuff engineer and have never done this sort of analysis so I fully expect I misunderstood something there.

https://www.structuremag.org/?p=14696

 

[KootK]

@CCWG: how many stories are is the building that we're talking about? Do you know if it's been seismically designed using ELF or PBD? Everything that follows assumes ELF and not skyscraper.

But I'm just a misc. specialty stuff engineer and have never done this sort of analysis so I fully expect I misunderstood something there.

I've done a fair bit of miscellaneous specialty stuff on the west coast myself. Often, the hardest part is trying to figure out the EOR's intention for design criteria. Everybody seems to communicate that stuff a bit differently.

 

[hardbutmild]

I agree with HTURKAK, I think that it means "for a level of earthquake lower than the design one, wall should satisfy this requirement".

As mentioned before, some codes have this level of earthquake defined (for eurocode 8 this is a return period of 95 years).
From my experience with eurocode 8 (this is probably different for different countries) this usually leads to half of an inelastic drift at a design earthquake.
For example, if inelastic drift for a design earthquake is 80 mm and Cd is 4 (I never used Cd, but it was mentioned before in this topic so I used the same), then elastic drift would be 20 mm for a design level of earthquake, but drift would be 40 mm for this lower level of earthquake. I would check for that drift the crack width, or whatever serviceability requirement.
Note that this is for eurocode only and my interpretation of it!
EDIT: as mentioned before, there might not be a clear connection between the drifts in reality, but eurocode 8 does give this simplified "just multiply with 0,5" rule.
EDIT2: I messed up the numbers the first time, sorry.

 

[CCWG]

@KootK - this was just a general question. It could be anything from a convention center, 3 story hospital, or a basic (from my perspective) 30 story office building. I think your redlines on the diagram show we are saying the same thing. My confusion is largely in the fact that the code equations for drift produces de, which is not equal to d. I generally do not have much insight into the analysis method the EOR is using - we often have to start working in the early DD phase. But at least now I feel I know where I need to do some more research to be able to speak more intelligently about this when requesting clarification. The requirement seems to be driven by facade consultants (in project specs) and not the structural EOR...but it often seems those 2 do not talk much before projects get put out for bid. Thanks again for your engagement!

@hardbutmild - I think this is a bit more confirmation of the "just ask" approach. Seems pretty unanimous there is no clear/obvious way of determining this. Thanks for the input

 

[KootK]

My confusion is largely in the fact that the code equations for drift produces de, which is not equal to d.

We can certainly clear that up. The code doesn't actually produce de. When de is used in commentary-ish documents like the one that you posted, they only include that as a referential nod to Newmarks's equal displacement principle, illustrated in the snippet below. That principle basically says that, whether a building's lateral system is made to yield early or late in its seismic displacement history, the total drift will tend to be:

1) approximately the same and;

2) roughly equal to the theoretical drift that would have occurred if the building's lateral system had remained elastic throughout it's entire displacement history.

The reality is a bit more complicated than that, as they state in the the article that you posted, and that is why d_inelastic is not strictly equal to d_E. Again, though, d_E is simply used as a reference value and is not intended to represent the real state of affairs for any real structure at any point during any real seismic event (unless the building is designed to R=1.0).

But at least now I feel I know where I need to do some more research to be able to speak more intelligently about this when requesting clarification.

You might enjoy the technical guide shown below for that purpose. It's concise and highly readable. PBD bedtime stories as it were. It should provide you with ample detail to be able "talk the talk" when it comes to discussing performance based design with other project team members.

 

[CCWG]

Much appreciated KootK. The book will definitely put me to bed, but at least it would be more productive than falling asleep to Netflix