Dave,
Well, whatcha know.
I just did a quick calculation, based on a single beam, 444" long, 13" thick x 14" wide on the outside, 12.5" thick x 13.5" wide cut-out interior, yield strength 70 Ksi, cantilever.
The force you can put on it and keep the outer fiber stresses below Sy is about 1,850 lbs with a deflection of 5.3". This gives about 4,900 in-lb potential energy.
Converting this entirely into linear velocity gives about 83 ft/sec = 56 mph.
A quick check of the rotational energy (0.5 Ixx omega^2) shows the rotational energy for one of these (almost hollow) box beams to be a trivial percent of the linear energy. So, figure it drops the linear velocity to about 54 mph.
Note that I figured this for a cantilevered beam, which stores only half as much energy as a beam that is bent into a uniform arc throughout its length. Twice as much energy gives you 41% more velocity, or 76 mph. Not bad.
Bill, what this calc does show is that, contrary to what you were "certain of", the box beams can take quite a bit of deflection (over 5") before they start to take a set, and substantially more than this before they are in danger of kinking. And they can store an enormous amount of energy in them in the process.
As expected, until one runs the numbers, you don't know what you're gonna get. And you, bill, have precisely ZERO "feel" for any of this.
You have simply your wishful thinking, based on your foregone conclusions.
tk