Laminated Glass Deflection for Wind Load - Wind load duration

[Yao1989]

Here is a question that has been bugging me for a while.

For laminated structural glass, checking for wind load deflection, what load duration should we assume for the wind?

I am asking this because depending on the load duration, the stiffness of the PVB interlayer is drastically different. From my understanding:
Reference wind speed in ASCE is based on 3-second
Reference wind pressure in NBCC is based on hourly mean
design wind loads in both NBCC and ASCE, after all the funky formula, is hourly mean.

By this logic, since I design my wind load deflection for hourly mean, does this mean I would also have to assume the load duration is also one hour?

For example, under 40 degrees (the temperature I assume the IGU glass can reach on a windy day), if the load duration is 3 seconds, the shear modulus would be 11MPa, which would allow a laminated glass to behave fully composite. If load duration is 1 hour, then it would only be 0.6MPa, making it almost non-composite. However, this is not consistent with most literatures I have read, which suggest that laminated glass would behave as fully-composite under short-term wind load. Can anyone shed some light to my confusion?

 

[NorthCivil]

For wind loads, the peak loading will likely come in short gusts, so 3 second is appropriate.

This is the approach I always took.

 

[Yao1989]

I want to agree with this approach, except that the is the design load we calculate is not based on hourly mean?

 

[Settingsun]

Hourly *mean* isn't a sustained load at that magnitude so I don't think treating it as such is correct for your application. But I can't say what the true way to handle it is for your situation. Depends on how a time-varying load affects the glass. Does it recover during the 'trough' periods of low loading?

I'm not familiar with American loading codes but there must be some factor to get from hourly wind speed to short-term design load (1 to 3 seconds) otherwise buildings would blow over all the time. Ultimate load divided by 1.6 approx would be a reasonable estimate at a normal serviceability load for materials that don't have such a severe dependency on load duration.

 

[NorthCivil]

except that the is the design load we calculate is not based on hourly mean?

It has been a while since I dove into american wind load codes. I believe you are correct. however, this is the base wind load. there should be all sorts of additional factors you apply to get peak momentary windloads from this hourly mean.

Depends on how a time-varying load affects the glass. Does it recover during the 'trough' periods of low loading?

The recovery is near instantaneous. The glass will recover to its original position within a fraction of a second.

It may be a valid exercise to check what the mean loads would be over the course of an hour, and the response of the laminated glass over that entire hour,

and then check the peak windloads, and the laminated glass response to those loads over the course of those 3 seconds.

 

[Yao1989]

Can you please elaborate on:
check what the mean loads would be over the course of an hour, and the response of the laminated glass over that entire hour,
and then check the peak windloads, and the laminated glass response to those loads over the course of those 3 seconds.

By this, do you mean a dynamic time-history analysis? Would I be modelling a load-history of the wind that peaks and troughs, but average to the hourly-mean windspeed? Seems a bit complex as I've only ever dealt with static wind load for components and cladding before.

P.S. I found a table from the CGSB 12.20 M-89 (canadian glass design code) that appears to indicate wind load determined from NBCC (which is an hourly mean) should be treated as an approximately 1-minute load duration. I think that is the answer I will use, except I have no idea where that 1 minute came from. If anyone has any background on this that would be much appreciated.

 

[KootK]

1) Honestly, for Canadian design, I think that the most "correct" thing to do is to calculate your wind per ASCE7 and then compare those values to E1300 capacity checks. 1989 was a helluva long time ago. The statistical basis for NBCC wind loads has probably even changed since then. As long as your consistent between loads and capacity checks, I think that's the main thing. CISC actually instructed me to do this for an extended shear tab connection that had no official, Canadian design protocol. NAFTA forever!

2) The ASCE7 appendices contain provisions for converting between various wind speeds and duration. You might be able to use that to massage an NBCC hourly value into an ASTM1300 3s value.

3) Another wrinkle, if I recall correctly, may be that the reference heights used to measure wind speed may be different between the US and Canada.

 

[NorthCivil]

By this, do you mean a dynamic time-history analysis?

no. KISS. keep it simple.

Approach the standards with a dose of practicality. Dont just use the numbers, try to understand where they come from. Does it make sense? remember, these standards were written 30 years ago now, and are not perfect.

 

[Yao1989]

Thanks NorthCivil & KootK for the advice. Just want to close off with my findings.

From CGSB, if you read table 1 & table 3, it basically suggest that wind load ~ 1 minute load duration, and with that load duration if you look in table 3, you can assume "monolithic" even for laminated glass.

I read more into ASTM E1300-16, which is a lot more recent and contains a bit more commentary. It is pretty clear that we should be using 3 second and assume 50 degrees for the laminated glass. In practice, this ends up being a partially composite glass.

So in the end, ASTM E1300-16's approach as it is more understandable, and is more recent, so I will follow that. As for my project I'd have to switch up to using SGP for the laminates instead of PVB if I want to keep my glass thin. Thanks again!

 

[NorthCivil]

I've been out of the structural glass game for a while. but heard that there was talk that the NBCC might formally adopt ASTM E1300 as an allowable alternative to CAN CGSB 12.20 M89, as it has actually been developed since the issue date.

As an old structural glass guy, I could never fault a prudent engineer for using E1300 over M89 should you ever be subject to a peer review. hell, I have heard of peer reviews done where the engineer didn't even know which standards are available for structural design of glass...