|
Bingo... the big question! Too bad there isn't a useful
answer.
As you noted, everything in the real world factors into the actual
limits. It's easy enough to plug the material properties in to a formula,
assuming a perfect cylinder or hemisphere, but that's really just a good
starting point. There are lots of academic papers that try to validate or modify
the simple equations with FEA simulations, but again, the results are still
quite theoretical, depending on uniform material properties and exact
construction.
I believe the difficulties of factoring in the myriad variables dictates
the lack of complexity in the shape of submersibles. If the shape is
simple, the construction can be verified, and the assumptions made in the
calculations can be 'trusted'. I've been thinking of a bent cylindrical
shape, that would envelope a person sitting in a 'car seat' orientation... nice
and comfy. Acrylic dome oriented at head level, tilted at about 45 degrees
for good visibility. Hmm... I haven't seen any calculations on a bent
cylinder :) The mating surface of the dome and the pressure housing
would have an interesting shape, and who knows the stresses. I'm not an FEA
expert either, so it's either trial and error, or revert back to combinations of
cylinders and hemispheres.
I once designed a cylindrical glass housing for a camera, using a trusted
program for the basic calculations, but deviated from theoretical by using a
flange. Seemed like a good design... unfortunately the flange completely changed
the stresses at the interface from compressive to tensile, so the housing failed
in testing, WAY sooner than expected. Luckily, non have failed in the field, but
it was embarrassing nonetheless. OK, I got wiser, and now I have someone do
FEA modelling for me. Not cheap, but well worth the cost. With materials
like steel, where the tensile and compressive yields are somewhat equivalent,
you can probably get away with some deviations from perfect geometry as long as
the material thickness is increased, but for materials like glass
and concrete, don't count on it!
Mark Roberts
----- Original Message -----
Sent: Friday, January 05, 2007 5:10
PM
Subject: Re: [PSUBS-MAILIST] Crush Depth
Calcs
Hello psubbers.
I have a question on the whole "crush depth" thing.( actually, a couple
of questions.)
First, how can you factor in all the variables of hatch design,
window design, thru-hull pens, weld attachments, stresses built in from
weldments, additional reinforcing around penetrations, and god knows what else
I might put on this thing, and still come up with an accurate guess ( and
guess it must surely be ) of when my negative pressure envelope is gonna come
crashing in on me?
Doesn't any one of these ( or other ) changes to a tank type shape create
a need for a separate calculation on each area's configuration and ability to
withstand pressure until the time of failure?
And doesn't the plastic deformation of an area of the hull, due to
reaching that moment of failure, change the whole calculation for the adjacent
areas, and this must be factored into the failure threshold of each
area affected by the point of failure.
I'm a welder, not a mathematician. Does anybody have a simple
calculation of the " plain" tank for the Kittredge design?
If a guy has a tank 48" round, quarter inch thick, hemisphere ends,
no stiffeners, basically your average propane tank, how deep can it go before
it crushes.
My thoughts on this are, Find the weakest point in the envelope, and
that's your crush depth.
Pumping a pressure tank down with a vacuum pump won't work. The dynamics
of the steel, method of fabrication, weld quality, inherent
stresses, and flaws in the material or welds, all change the way steel reacts
to external or internal pressure variations.
Destructive testing is the best way to determine "crush depth" but
it's just too expensive to do.
And even if you do it, and get the golden guarantee, knowing just where
that hull is gonna implode, now we have to calculate what the new stress of
the dive cycle has had on the individual hull components, to find the new
crush depth.
What percentage of safety factor is acceptable, ( how much risk are you
willing to accept )
If you're cruising along, and crash into a rock, cause the surge/current
pushed you, and get a dent in your hull, do we have to recalculate
everything?
Too much math.
I'm looking for a depth, that a simple tank will go to, without getting
really flat.
Anybody got any ideas?
Frank D.
|