Rag and tube construction and computer models?

Hi all...

Many moons ago I worked at place doing finite element modeling....and the work
I did was with solid objects...dont worry..I get back to the rag and tube
question eventually...

The way that works (for a solid object, say like a crankshaft, a bracket,
etc,,,) is you build your big object outa lots of little objects...lets call
them bricks.....Now, with lots of nasty math, physics, and engineering you can
develop equations that say if you put this much force or displacement (or
whatever) on this part/side of one brick, then such and such will happen here
and there and there on the "brick".....now the corners/sides of that brick are
mathmatically tied to the next brick and so on and so on.....and what you end
up with in the end is a GIGANTIC mass of equations (often thousands if not tens
of thousands of them) that the computer works hard to find the solution to...

Now, its not quite as bad as it sounds (as long as you werent the poor soul who
had to write the program in the first place) because what you generally did was
use another program to make a geometric model of the object of interest (kinda
like a fancy autocad)....and that spit out another nasty file that got feed
into the first program I described above...

The nice thing about this was you could input ALL kinds of material properties
describing each brick like strength, rigidity, fracture toughness (and
MANYmore) and as importantly it allowed an object to "constructed" with
different materials....and then you told the program where to put forces, or
displacements or whatever....and off it went to crunch numbers....

Once all the computing was done, you used a third program to visualize
stresses,deformations etc etc...and with that you could "see" where you had
more material than you needed, or where the stresses were too high, or where
something was likely to buckle etc etc....so it allowed you to optimize a part
in ways standard textbook engineering equations never could...

The other cool part is you could even do things like create a "crack" here and
see if was likely to propagate....or "break" a part there and see how the load
was redistributed among the remaining parts.....allowing you to check out lots
of "what if" scenarios you'd never have the time or money to do otherwise)

Now, I never used this capability of the program...but it also had the
capability to construct objects out of plates, shells, infinitely thin rods
(wires?), and hollow tubes....

It occured to me the other day that virtually all the rag and tube designs
being built today were designed before this computer capability existed (or at
the time only at the real high tech computer power houses of the day)...

So, 2 questions...does anyone know of any small plane "rag and tube" designs
where any significant computer modeling was used?

And secondly....any guesstimates on how much weight percentage wise you could
shave of the typical tube structure by using such modeling and still maintain
the same structural margins....?

take care

Blll