Looks like there were some errors in the design process according to this NIST report:
1. Using enclosed building vs. partially enclosed.
2. Using the wrong wind method (roof was > 60 feet but they used the <60 ft provisions).
3. Using the wrong roof slope in the calcs.
4. Depending on the fabric for bracing.
5. Using the wrong unbraced lengths in design.
6. Using k = 0.5 instead of k=1.0.
"Using k = 0.5 instead of k=1.0"
This has long been a point of contention. I have seen plenty of engineers use a K of 0.65 when designing compression diagonals/verticals in trusses with big gusset plates at the ends or welded.
I had posted about this on the forum and it was discussed that in order for the end moments to develop the chords have to be stiff in-plane and significant out-of-plane bracing will need to be provided.
I recently reviewed a design drawing where a K of 0.65 was used in the design, but the analysis model had pinned ends (no end moments). This, in my opinion, is double-dipping and incorrect.
Are there ways to quantify the use of lower effective length values based on the amount of fixity provided? I know it is conservative to use K = 1.0.
Add "improper reinforcement" to JAE's list above. Reinforcing with self-tapping screws? Really?
Maybe the larger failure is allowing the overwhelming influence to make the cheapest building possible compete with the integrity of the structural engineering. Every dollar the manufacturer skinnies out of the structure is a dollar he puts in his pocket. Isn't this a conflict of interest? What engineer can properly look out for the Owner (and Public)'s best interest when his company is pressuring him to make a buck for them?
As far as 1), this could be somewhat understood based on horrible way Enclosed and Partially Enclosed are defined in ASCE 7. And 2) might be case of a little engineering judgement being exercised. If there is that big of a difference between a 60 ft. tall structure and a 66 ft. tall structure, there's soemthing wrong.
I'm not trying to make excuses for the design engineer as I think he made some sloppy assumptions, but I suspect these type studies are similar to an full IRS audit combined with a colonoscopy.
I'm going to agree with JedClampett on this issue of enclosure.
Page 70 of this document shows all the openings used in the calculation of enclosed or partially enclosed. At first glance at the figure, I would have guessed it to be enclosed.
The report states that the building failed due to a transverse wind, but I've got a serious issue with the definition of partially enclosed and the vents on the gable walls of this building in relation to a longitudinal wind.
The ASCE definition of "opening" leads me to believe that only vents, not doors and windows, should be used in the calculation to determine enclosure.
Left gable wall - six 5'x5' vents
150 ft2 opening / 13,700 ft2 total
Right gable wall - ten 5'x5' vents
250 ft2 opening / 13,700 ft2 total
Roof - four 4'x4' vents
16 ft2 opening / 82,800 ft2 total
Long walls - no vents
0 ft2 opening / 18,800 ft2 total
You're telling me that 416 ft2 of vents over a total surface area of 129,000 ft2 makes a structure partially enclosed? That's 0.3% of the total surface area of the structure. That just doesn't sit right in my stomach.
Technically, the structure is partially enclosed by the letter of the law, but when you step back and look at the big picture does it really make sense.
I would point to the use of k=0.5 instead of k=1.0 for web compression members as the major culprit, not the wind enclosure classification.
[\quote vandede427]The ASCE definition of "opening" leads me to believe that only vents, not doors and windows, should be used in the calculation to determine enclosure.[/quote]
The commentary for Section C6.5.9 reads that openings that allow air flow through the building that could be open such as doors, operable windows, air intake exhausts, ventilation systems, gaps around roors, operable louvers.
As pointed out recently in another thread, a structure can be enclosed in one primary direction, yet partially enclosed in the transverse primary direction.
Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask
My issue with the definition for partially enclosed is that it only compares openings to openings. There's nothing that factors in the size of a building.
By the definition, this building with 129,000 square feet of surface area could be considered partially enclosed if it had a 2'x2' opening in each of three walls and a 4'x4' opening in the fourth wall. It could have fabric covering 99.98% of it's surface area and be considered partially enclosed by ASCE7.
If the size of a building isn’t considered in the definition of partially enclosed, then the same configuration of openings could qualify a very small house as partially enclosed as a very large building.
What has to take place in order for that internal pressure coefficient to go from 0.18 to 0.55? The wind has to enter the building through the openings and pressurize the inside of building. For a very small house, that could take place within the 3 second gust we design for. But how long do you think it would take for that to happen to this building with over 5 million cubic feet of interior volume? I’m betting that it takes longer than a 3 second gust to take place.
And did the structural designer even know there were vent openings? A lot of times the HVAC engineers don't complete their design until the last minute.
Possibly the code requires partially enclosed design. But who's to say that the designer didn't make the same judgement as vandede above and that his judgement and experience tell him that these internal pressures aren't reasonable. Was a wind tunnel test performed on this structure?
