Wind codes and unrealistic velocities/pressures

[pwht1]

Hello,

I was told to design some purlins by a senior engineer that must be Stramit Z250-15 at 1500 centres. These things are spanning 11m and after a few checks I find they fail under the wind pressure from AS1170.2(the Australian wind code), no big surprise. I go back to the engineer and report what he needs to get it to work and he rejects it. He proceeds to tell me that the wind code is unrealistic and I should "reverse-engineer" the purlins. So here's the question, what wind code does give realistic pressures and what's you view on AS1170.2 (if you're familiar with that code)?

Thanks,
Paul

 

[rscassar]

I can't find a Z250-15 in the strammit catalogue and yes, a Z250-19 would be struggling to span 11000mm. The wind code is based on a 5% probability of being exceeded in 500yrs (or something like that). Personally, to have a senior engineer say that the wind code is unrealistic so it can be ignored is unprofessional in my opinion. The reality is that the structure will never experience the design wind load, the design earthquake load or even the design live load, but that is no reason for the engineer to ignore these actions.

Tread carefully with your senior engineer (do not rock the boat!!). But at the same time, do what is ethical and is in the best interest of the community. If your certifying engineer ever finds them-self in a lawsuit because someone was injured and their defense is "the code is unrealistic", they will more than likely have their licence removed.

The pressure coefficients of AS1170.2 are upper bound. I was submitted wind tunnel tests once where the maximum pressure zone on a windward wall was C=0.5 whereas the code stipulates a co-efficient of C=0.7.

What do you mean by "reverse-engineer", I have never heard that saying before?

 

[pwht1]

Kikflip,

I think they've got manufacturers mixed up, I know Bluescope makes a Z250-15. Sorry, it didn't register when I posted earlier.

I don't sleep when I think things haven't been done right so I'd prefer to rock the boat

Your example suggests AS1170.2 is conservative (which is lower bound not upper, yeh?). How do you work out the Cpe from a wind-tunnel test? It doesn't sound like a very straightforward exercise.

You've never seen Paycheck? Reverse engineering is when you're given a solution to a problem and you have to work out how the solution was achieved.

Thanks,
Paul

 

[paddingtongreen]

Reverse engineering was coined for certain Asian countries that took goods made in the developed world and took them apart to figure out how they were made. This allows pwht1's version.

Along time ago, on agricultural open sheds etc. we allowed the girts/sheeting rails to deflect and act as combination beam and catenary. I admit we didn't design the corner columns for the tension, but I never heard of any complaints.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[weab]

I don't know what a Z250-15, but I'll assume that it is 250mm deep. For a 11000 span, this is not even close. Have you checked deflections? I'd be concerned that purlins may hit diagonal bracing or something else in the wall.

-Are you in consulting? If so, then what your boss told you to do sounds really bad on its face.

-Or are you working for a shipyard or contractor or some other group and you are building this building for your own company for a building of minor importance? If so, then your company may be willing to accept additional risk. Just make sure that you let everyone know the limitations on the design.

With regard to what loads your building will experience, I would add this. It is unlikely for your building to experience design loads, but my experience is that you may see relatively close to wind design loads several times. If, as I suggest, your design isn't even close, you will likely overstress this particular purlin. No amount of good experience with any design is proof that the design is satisfactory.

 

[Tomfh]

AS1170.2 is good. Follow it.

Use all the reduction factors if you have to, e.g. direction multipliers, shielding, surface combinations, etc, but don't just ignore the code...

 

[Tomfh]

By the way, are your 11m spans continuous or simple?

 

[rscassar]

Full scale testing is usually performed on large-scale projects such as tall buildings. It will need to be performed in an accredited laboratory. Generally the report from such testing will only be available for the person/company that has paid for it.

http://www.jcu.edu.au/cts/

http://www.windtech.com.au/

For some roof purlins such as your doing, I would use the pressure co-efficients set-out in the code. I don't see any basis to justify that a lower wind pressure than that derived from AS1170.2 can be used.

There is a good text called "Wind Loading on Structures" by J Holmes who sits on the Wind Code Committee for AS1170.2. It gives some figures which show the average, upper bound and lower bound pressure distributions as established from wind tunnel testing.

 

[MikeHalloran]

We rode out Hurricane Wilma a couple of years ago.

After a few hours of banging and crashing and strange noises associated with shingle removal en masse, the wind abated quite a lot, so I opened the front door to see what was going on.

The (now greatly reduced) wind was blowing a tree down the street. Not a branch; the whole tree, root ball included.

I closed the door.

Follow the code.




Mike Halloran
Pembroke Pines, FL, USA

 

[pwht1]

Thanks all.

I'm more interested in the background to the code and how they get their velocities/coefficents, not the actual design of the roof. Thanks for the text reference kikflip, looks like a dull but necessary read.

FYI it's 7 internal spans of 11m with 7m endspans over an office with a span/300 deflection criteria.

 

[ooox]

PWHT1, The lysaght capacity tables for 4 lapped Z25015 is -0.78kN/m with 3 bridging points so if you 'reverse engineer' your maximum wind load should be +/-0.52kPa. This is based on a span/150 deflection criteria however. Depending on your location and building geometry this might be OK.

Stick to the code, if your design is not compliant and something happens you've dug a massive hole for yourself.

 

[apsix]

AS/NZS1170.2 is referenced in the BCA and therefore cannot just be ignored if your building has to comply with the BCA.
The usual deflection limit under wind loading is L/150.

 

[ishvaaag]

Also, the lifespan. When I had the occasion to analyze the tallest brick and mortar tower in town (Zaragoza), from the early XVIII century, close to 100 m tall, that by then (2004) had some 6 cm cracks (standing for years) at some arches around 60 m height, I was able to convince myself that wind forces were the more likely cause (overcoming other lesser solicitations by earthquakes, between them the destructive one at Lisbon in the early life of the tower, also felt here). Limestone decorative statues of people or animals at the height were eroded to round shapes in the windwards side. So you have windstorms of say 100 miles per hour cracking quite massive columns where the tower had a 10 m inscribed octogonal shape with 4 arch insets.

 

[csd72]

AS1170.2 is a good code and is less conservative than some others. I agree that you should follow it.

Try and get your hands on the code commentary as this will explain how the loads were derived. Basically they are from probabilities of wind gust turbulence and the smaller the area the higher the likely maximum average pressure on that area.

Anyway, I have been in a similar situation before so I know a few tricks that can be done.

1. if the load limit is based on deflection, can you accept a higher deflection and base it purely on strength calculations. I would only recommend this if you are looking at very unlikely or infrequent wind events.
2. you could specify a longer lap length and then analyse it on the computer along with using the lysaught tables.
3. have you used the actual stepped down wind loads on the end span. The manufacturers give guidance on how to reduce this to an equivalent uniform load but it can also be analysed directly.
4. a reduced end span if this is an option.
5. use thicker purlins only on the end span.
6. How have you calculated the internal pressures? have you done the actual calculation of open areas e.t.c. for the critical direction and used a permeability for metal clad walls and roof of 0.04% as per the code. This can make a significant difference.
7. Check your shielding and terrain assumptions in the critical direction.

Any of the above may/may not help depending on the exact situation.

Ignore kikflips comment regarding wind testing, if the client is not prepared to pay for thicker purlins then they definately will not pay for wind testing.

Take this as an opportunity to really get to know the wind code as this is a very important code to know and can really save money if you know how to shave every item down.

 

[pwht1]

Thanks csd72, a few good points.

Where in the code does it give a permeability coefficent for metal clad walls and roof? I've never heard or seen that.

 

[frv]

I agree with others on here that the advice given to you by the more serious engineer is, at best, negligent--primarily due to his reasoning. You can't simply state that the code is unrealistic.

On the other hand, csd72 gives good advice. The exceptions and the provisions to "shave" loading are there for a reason. In the US, for example, there is a "simplified" method that is fairly conservative, but very simple. Then there is an "analytical" method that is a bit more accurate but a pain in the ass to do from scratch.

As for the deflections, limits depend on the material in question. For example, the steel code in the US doesn't really have any hard limits, but there are some good rules of thumb.

For cladding, there are also rules of thumb; the more flexible your material the more it is allowed to deflect. Also remember that deflections are a serviceability issue. In the US, it is common practice to reduce the wind loads for the serviceability check. ASCE recommends multiplying wind loads by .7 to obtain an approximate 10 year wind load (AISC recommends .75).

Be aware that these factors are approximate and need to be adjusted based on design wind speed and recurrence period (the two I just mentioned correspond to a 50 year wind event converted to a 10 year wind event for, I believe, 100 mph wind load or less)

 

[Ron]

pwht1...to your question regarding code development...

Wind load structural affects are based on statistical predictions of the wind pressures. Which ones apply?

That's a bit difficult to discern, but in short, just follow the code.

Code mandated wind loads are based on statistical probabilites of occurrence. For instance, in the US, the most popularly used codes are ASCE 7 and the International Building Code. Both are based on a 3 sec gust with a probability of occurrence of 0.02 (once in every 50 years). The factors that are used for modifying the wind loads are based on occurrences of once in 500 years, or thereabout. Further compounding this problem is that the wind load considerations of the code to not match directly with the Saffir-Simpson scale of wind loads for hurricanes. As an example, a category 4 hurricane would have a Saffir-Simpson scale would actually have design wind speed of about 10% higher.

Confusing at best! Follow the code, whether Australian, American, European, or Asian.....there's a reason for their development (protection of the health, safety, and welfare of the public). Respect and abide the requirements.

 

[pwht1]

Thanks Ron. I think I'll be leaving the wind theory alone for now and getting on with the roof design.

I don't ignore codes, but I do think it's important to question them.

 

[apsix]

pwht1

It's not clear, I assume you are using the reduced velocity V25 from Table 3.1 for your deflection calcs.

 

[pwht1]

Apsix, that's because I didn't say what I was using...I just wanted to understand wind codes a bit better.

Anyhow, I'm use a 20 year velocity for serviceability and 1000 year velocity for strength. I have used all of the reduction factors that I can but the local pressure factor is causing problems, the pressure zone width "a" is 13m. Increasing the lap length has solved the strength issues (awesome idea csd72), but I'm still getting a 70mm deflection.

Paul