Glass Fin Wall Joints and Lateral Deflection

[StrucPEng]

Hi all,

We are in the process of designing a glass curtianwall system that is supported off of vertical glass fins. Currently the face glass will be structurally siliconed to a hanger bar between the glass and the fin and dead loaded at each lite via setting blocks and steel tabs (shelf). My question is more of a conceptual one as it relates to seismic and drift. The system is in a high seismic zone in Canada but the seismic loads will not govern the design. What will have an impact is the drift as this system spans 5 stories high and has drifts of 100 mm or more.

To get to the question, how do you tend to think about the drift and geometry of the panels as it relates to the structural silicone? The joints need to be reasonably sized but it really comes down to the shears and the elongation if your silicone joints which is dependent on the geometry and how the panels rack (or translate) during building drift.

Conceptual we have kicked around the idea that the glass will essentially stay square and translate laterally as the fins away but once you throw in any fixity from the silicon it becomes a more difficult problem.

Has anyone designed these systems before and if so how do you size your joints in relation to lateral deflection.

Thanks!

 

[KootK]

Has anyone designed these systems before and if so how do you size your joints in relation to lateral deflection.

I cannot claim to have done anything quite like that. Smaller scale and much less drift for me.

how do you tend to think about the drift and geometry of the panels as it relates to the structural silicone?

Like you, I see the glass lites being essentially rigid in plane and any movement coming about via deformation in the silicon.

Conceptual we have kicked around the idea that the glass will essentially stay square and translate laterally as the fins away but once you throw in any fixity from the silicon it becomes a more difficult problem.

1) I don't see the drift being accommodated primarily via shear slip in the silicone. I see each lite rotating to accommodate drift sort of how we imagine that sheathing in wood shear walls rotates. So mainly perpendicular to joint tensions and compression in the silicone.

2) The trouble with any model where drift is accommodated through lite rotation or translation is that those models imply strain in the silicone that had to be induced by the fins acting on the lites to produce that effect. And that road would seem to lead to weak axis moments and shears in your fins that may be unpleasant to deal with.

Ideally, I'd like to see the thing suspended from the top and the lites somehow allowed to translate both vertically and horizontally in plane at the bottom. That said, at 100 mm drift, I could see how that could be a problem.

Some additional questions that might help us to help you:

3) Fins suspended from the top?

4) How often, vertically, do the connections to the lites occur?

5) Four connections to each lite, two per fin?

6) Any chance you could send schematic connection details? Definitely fin to lite. Top and bottom would be useful too.

7) Will there be connection, top or bottom, that will inadvertently deliver in plane shear / movement directly to the lites?

Five stories is... tall. Your path forward might be to somehow estimate out how much weak axis curvature can be expected in the fins and then demonstrate that the fins are flexible enough to accommodate that.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[StrucPEng]

Koot,

Yes a lot of the issue is coming from trying to avoid relying on the silicone in shear. more than that I suppose is the requirement that the silicone only be allowed to elongate by 15% so as the panels rack we have to figure out how much movement to expect which then drives the joint size for the silicone bite. I guess that can be relatively easy or relatively hard depending on how you look at it since the more variables you add the less reliably you can predict the deformation I am finding. We would also like to conceptually find a way to do this without FEA. We can do that but it is a pain for each project to model it and want to develop a better mouse trap as it were.

I had a similar thought about allowing for more than just vertical movement in the fins but you're also right it can get quite large over that height.

I should point out that 5 stories is the worst case height as the wall slopes to one side. best case is 3 stories worst is 5 stories.

3) Fins are suspended from the top. Vertically slotted at bottom.

4) Vertical connection occur at the bottom of each lite at each corner. Lites are typically 5m tall by 2.6 m wide and the wall is three lites high with the upper most lite being smaller at 3m tall.

5) (2) connections at each lite for dead load, one shelf at each bottom corner. other three sides are structural silicone.

6) See image

7) Top and bottom are continuous aluminum channels and should flex as they are separated by continuous gaskets from the actual IGUs

We also kicked around your last point about maybe finding the weak axis bending in the fins acceptable and moving on but is seems like a bit of a cop out. Maybe not though.

 

[KootK]

Thanks for the added info.

I see what you mean, with silicone horizontal joints, I don't see that you have much choice other than to allow that deformation in shear by sizing the joint suitably if that can be done.

Is it accurate to say that, at the top, drift is delivered directly to the IGU without passing through the fins? It looks that way but the graphic is too blury for me to tell definitively. If you could somehow do the same at the bottom, you could at least take the fin weak axis bending out of the picture. That's no mean feat however if you're to retain vertical slip at the bottom at the same time.

HELP! I'd like your help with a thread that I was forced to move to the business issues section where it will surely be seen by next to nobody that matters to me:

http://www.eng-tips.com/viewthread.cfm?qid=456235

 

[glass99]

StrucPeng

Are the fascia panels only structurally siliconed on? If yes, then you need to keep the silicone within 20psi for all load cases including story drift.

If there is some kind of mechanical capture, the usual thing is to keep within the elongation limit of the silicone for an elastic seismic or wind event (usually 50-100% strain in tension depending on the silicone), and allow sealant failure for an inelastic event. The ratio between elastic and inelastic is usually around 5x, so its quite a big difference.

 

[StrucPEng]

Koot,

That is not exactly true, the IGUs are supported on a steel hanger from the top bracket but there is some fixity between the bar and the glass fin along the length. I see what you are saying and that might be an option but you are right keeping the vertical slip and adding horizontal slip may be difficult.

Glass,

Yes the panels are only siliconed on, no capture so we are relying only on the silicone in a seismic event. We have come across literature from Dow that allows for 25% movement in the joints under a seismic event while still maintaining about a 2.5 safety factor so that is helpful but doesn't get us all the way there probably.

I am surprised there is not more literature or guidance on this given the large use of glass for facades.

 

[klaus.]

...well most companies keep their stuff and knowledge 'in House'

 

[glass99]

story drift normal to the wall is usually easy bc the wall just tilts. Large in plane story drift will likely require a sliding joint at the top and the seismic inertia is transferred through the silicone to the base.

25% strain is in the ballpark of 20psi in shear, but you should look at the silicone stiffness and stick to 20psi. You need to do an analysis of how the force and imposed displacement is distributed. I would do an FEA given your high seismic load.

The structural side of the facade world is definitely not documented to the same degree primary structures are. Just ASTM E1300 alone expressly prohibits its use for structural glass, including every fin wall in America. The ASTM for structural glass is still in committee (and has been for a while). There is a patchwork of guidelines though. AAMA, ASTM, NGA, IBC, DOW manuals etc. The DOWSIL technical manual is the most relevant to this.

 

[glass99]

Also, the rough approximation of is that the horizontal joint shear per joint is the story drift divided by the number of joints. But you need to think about the stress on the silicone to the fin in particular bc you are using structural silicone.

 

[glass99]

Also also: the ASTM says 20psi, and NOT a factor of safety of 2.5 or any other FOS number. i.e. if you use stronger silicone it doesn't help (which is madness and frustrating beyond belief when you have a product with strength in the 200psi range). The Eurocode folks kinda talk about a FOS of 6 or a FOS of 4 if you take into consideration joint rotation, but even that is very wonky and basically limited to 20psi. Structural silicone is a whole thing...