Building Drift and Glass Fallout/Facade Serviceability

[jimmytwotimes]

Curtain walls are required to withstand the lateral story drift of the building. The ultimate limit state is the glass in the curtain wall can not fall out. ASCE 7-10 13.5.9.1 requires Δfallout ≥ 1.25 * Ie * Dp

Therefore, the curtain wall needs to accommodate 1.25, 1.56, and 1.875 times the design drift of the building structure for risk categories I/II, III, and IV respectively. I understand the design drift is divided by Ie, so Ie in the Δfallout calc is not redundant. But the values seem rather high, particularly when computing the actual drift values, ex. a 15ft story height gives ~4in.+ of drift. So for those knowledgeable with buildings in seismic zones, here are a couple questions:

- Is it typical to design a structure right to those limits? Reason being most curtain wall systems are tested to H/100 (for serviceability - ie. remain air/water tight), and H/66 (for ultimate - glass can't fall out). These are recommended values from architectural test methods (ex. AAMA 501.4), and apparently not very close to the upper code limits.

- How does the drift impact other structural parts of the building, and in checking those other parts, do you include the importance factor and an additional 1.25 factor (or similar)? For example, structural separation calcs in ASCE 7-10 12.12.3 remove the importance factor, and do not have an additional 1.25 factor. So this analysis does not align with the Δfallout analysis. As a note, the 1.25 factor comes from the recommendation that drift calcs can be underestimated by up to 30% (from 6.2.10.1 in NEHRP 2000, which is referenced in the ASCE 7-10 commentary for 13.5.9.1).

- Do you ever consider seismic events at a service level? If a small earthquake hits, how is structural integrity of the building structure verified? I've yet to see wind drift values exceed H/100, so having a curtain wall system that remains air/water tight at H/100 seems pretty arbitrary.

 

[NorthCivil]

These are all good questions, and the "engineering" part of facade engineering is interpreting code requirements, project specific requirements, the behavior suggested by your calculations, and the behavior seen during testing/performance of similar or same construction types.

Now on to your last question. Curtainwall systems do need to acheive a level of performance after accomodating certain deflections. As far as serviceability requirements - and the particular maths and formulas behind justifying certain construction methods - how to apply and approach these problems is up to you.

Facade engineering is the art of making a practical product from:

-somewhat conservative and well developed codes for the main structure
-complex yet likely very inaccurate wind load calculations
-building movements that are not fully understood, even by the primary structural engineer
-a value for building movement provided by the building structural engineer that may or may not reflect the true movements (often the project engineer will cite code maximum deflections, and elaborate no further)
-project specifications, often held up as a holy grail, that were likely copy-pasted by a junior architect buried in a closet somewhere responsible for the entire 3000 pages of spec-writing for the entire building
-well defined and refined, complex codes for aluminium, that we often make calculation mistakes with while utlizing (admission of guilt here!)
-old, rusty and somewhat useful design codes for glass products
-a complete lack of code guidance for how curtainwalls respond to building movement. you really have to see it yourself to comprehend it.

the engineering part of this is making sense of it all. If you don't get it, and there is nobody in your office to assist you, highly suggest hiring and external peer reviewer to assist you who has experience in the field.