At 12:56 PM 11/21/2006, Dan Zimmerman N3OX wrote:
>And remember, if you're going to do your own engineering on an
>aluminum tower, make sure you understand fatigue limits.
>
>What do you suppose the manufacturers of aluminum towers do? Do they
>just design the tower so that stresses are very *unlikely* to exceed
>the fatigue limit of the alloy they're using?
This is a fairly subtle problem, and it's good that you bring it
up. Both steel and aluminum can fail from repetitive loading, but
for steel, there is a load that if you never exceed it, you can't get
a fatigue failure, so you can just do a conservative design, and walk
away, knowing that it won't fail from cyclical loading. Aluminum,
though, doesn't have a lower bound on the load for fatigue failures.
(something that the designers of the DeHavilland Comet weren't
totally aware of...) The fatigue life for aluminum depends on a lot
of factors, but the net result is that every load/unload cycle takes
away some fraction of the life (big cycles take more). Every little
gust of wind, every rotation of the antenna, etc.
As Dan says, the designer just makes some reasonable assumption about
the number of cycles and the load behavior and does the fatigue
calculation using that. Or, they build a prototype and test it (the
usual ham approach.. make it wicked strong, and if it falls down,
make it bigger next time)
>That's pretty significant over-engineering; if you know the wind speed
>probablilities and projected service life of the tower, you could
>probably allow the stresses to be higher, but if you're selling your
>product to hams who might want to overload it and hope it lasts for 20
>years, you're probably going to be more conservative. Otherwise,
>maybe you need a disclaimer like "sways 10 million times and then
>falls down"
Bringing up an interesting question.. "What would be a reasonable
number of cycles?"
Say a 20 year life and a daily temperature cycle gets you started
with 7000-8000 cycles of one form, but that's not a lot. Probably
all those little wind induced vibrations is what would get
you. Every minute would be 60/hr or 1440/day, so in 20 years, you'd
be at around 10 million.
Jim, W6RMK
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|