Fatigue Failure In Industrial/Commercial Building

[JohnnnyBoy]

I'm a junior engineer looking to get a little clarification on when to be calculating for fatigue failure in Commercial Applications. In school we learned about fatigue in more mechanical situations where we have significant force being cycled quite often but when it comes to building design in a lot of application/design manual it seems to be neglected completely.

Now assuming a building is loaded to capacity at least twice a day (will discuss actuality later) for 365 days a year and a design life of 50 years. We would be looking at a cycle of 36500. According to a simple fatigue failure curve of steel that would give us a reduction ration of approximately 0.6-0.7 roughly.

Now when calculating live load from the building code does this already assume a factor of safety so that the structure is never loading to that limit and is more in the endurance limit of the steel? Just assuming we use 100psf for example which is then factored to 150psf that would assume in a 10x10ft area we have 75 people weighing 200 lbs (nearly impossible).

If anyone has any insight or design guides on when I should/shouldn't design for fatigue it would be very helpful.

 

[WARose]

I normally never consider fatigue in a commercial building......or really in a industrial one either (unless we are talking loading from something like machinery, crane/monorail runway, vehicular traffic, etc).

 

[JohnnnyBoy]

In this case I would solely consider normal Live loads causing the fatigue but moreover wondering why we don't consider it at all.

 

[rb1957]

First, I'd've thought buildings (though not bridges) were pretty immune to fatigue.

Second, I'd've thought your allowables were derated so this shouldn't be a problem.

Third, assuming 2 limit load cycles per day seems awfully severe. It'd make more sense to say your everyday loading is maybe 60% of limit (ie design) which offsets the reduced allowable.

Fourth, I don't know if you included a safelife factor in determining your allowable discount. If you want a service life of 36,500 cycles, then you should have a fatigue life of between 100,000 and 365,000 cycles (depending on safelife factor.

another day in paradise, or is paradise one day closer ?

 

[WARose]

In this case I would solely consider normal Live loads causing the fatigue but moreover wondering why we don't consider it at all.

As you observed in your OP, it's mainly because the odds of the framing seeing that full live load again and again (enough to be in a fatigue situation) are typically pretty low.

It's not a bad thing to be on the lookout for though.......especially if you are designing framing (getting beyond the situations I mentioned above) where a large amount of human traffic is expected every day.

 

[KootK]

I've often wondered this myself. I've not seen an answer in print but my take on it is this:

1) Most commercial building elements spend most of their time being relatively lightly stressed.

2) Commercial buildings generally do not seem to suffer from fatigue issues in practice, based on experience.

I remember reading an article once about prestressing anchor bolts in industrial applications (utility poles etc) to ameliorate fatigue issues. They cited a handful of instances where it made sense prestress. Once of them was "anchors subject to cyclic tension as a result of wind induced overturning loads". You know, kinda like every commercial building braced frame ever.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[TehMightyEngineer]

I've designed certain elements of industrial buildings for fatigue; monorails, steel supporting reciprocating or vibrating machinery, steel supporting travel ways in warehouse or other heavy material handling. I've only ever seen it control the design for heavily used monorails; everything else it had a service life well beyond the design structure life. The only major industrial steel I've ever seen fail in fatigue was truck dumper frames. You can probably see why:

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

http://www.americanconcrete.com

 

[MotorCity]

Fatigue usually only controls for frequent, repetitive loads resulting in high stress levels. Typically this kind of loading is induced by equipment, vehicles, cranes, or other moving loads. Live loads, wind, and seismic are of course technically "moving" loads, but they only satisfy half of the criteria....possible high stress levels. They do not occur at such a frequency to cause concern for fatigue. They are not like a press stamping out 10,000 widgets a day or a bridge crane going back and forth 100's of times a day.

 

[SwinnyGG]

To start, you need to understand a little more about fatigue. Fatigue life charts for materials (where you got your 36,500 cycle strength reduction from) show the mean failure rate for cyclic fatigue, where the load is fully reversed.

Building components, with a few exceptions, are almost always loaded zero-max-zero or x-max-x. Think of a cable stay on a bridge- it is installed with some preload, and bears that tension all the time, plus added tension if you drive a big truck (or whatever) across the bridge. That cable stay, throughout its life, will NEVER be loaded in compression. Most components of a building are loaded in a similar mode. A floor joist in a house, when dead and live loads are in place, is subject to some bending load 'x'. Unless something very, very bad (or weird) happens to the structure, that joist will never experience a bending load of -x.

Assume you have a sample part subject to some sinusoidal load, such that the amplitude between maximum and minimum loading is constant. As the mean load moves away from zero in either direction (meaning that the minimum load approaches zero), the entire fatigue curve stays the same shape but moves up the y-axis, meaning that for any number of cycles the fatigue limit stress approaches the yield stress of the material.

A perfect real-world example of this is shot peened shafts. Consider two rotating shafts subject to some bending load, one simply machined and one machined and shot peened. In the first shaft, the load on the outer fibers is sinusoidal and the mean loading is zero- the load is fully reversed. In the shot peened shaft, the load on the outer fibers is also sinusoidal, but the mean loading is NOT zero. The load in the tension area is the tension load due to bending, minus the compressive load the shot peening process left behind. The load in the compression area is the compressive load due to bending PLUS the compressive load due to shot peening. The amplitude between the tension and compression load values is the same as in the first shaft, but the mean is shifted away from zero, and any millwright will tell you that the shotpeened shaft will last longer.


This is in addition to everything KootK just said, but in the case of most components of a modern structure, it seems to me that the ways in which the structure handles loads pushes the true, real-world fatigue limit stress high enough that other failure modes will tend to dominate.

 

[DaveAtkins]

The design of crane runway support beams is the only time I check fatigue in building design.

DaveAtkins

 

[JohnnnyBoy]

Thanks for all the responses and that is what I had originally thought as well. I did fine a reference in the NBC (Canada) as well in Division B Part 4. A-4.1.3.3 which states:

Failure due to fatigue of building structures refereed to in section 4.3 and designed for serviceability in accordance with Article 4.1.3.6 is in general, unlikely except for girders supporting heavily used cranes on which Article 4.1.5.11 provides guidance.

 

[Archie264]

My recollection from a strength of materials class years ago is that steel has threshold stress below which it is not susceptible to fatigue though aluminum does not. Whether that threshold for steel falls below .9/(1.2D + 1.6L) I don't know but I suspect it falls below thousands of cycles of such, which should take a long time to achieve. Also, to get to that point I suspect concrete and tile would chip and crack, pipes and roofs would leak, ducts and boilers would be replaced, etc.

I wonder if the bridge guys have some insight?

 

[SwinnyGG]

As a very general rule of thumb, the cyclic fatigue limit for most non-exotic grades of steel below 160 ksi UTS is approximately 50% of UTS.

I suspect that some parts of some buildings are stressed beyond this level, but without some significant vibration load or something similar, you'd never get enough cycles built up in the part's history for fatigue to be a real problem.

I'd also bet that if we got a bridge guy sucked into this thread he or she would be able to give us some additional insight.