Saturday, January 06, 2024

Complex ordnance fail

Dud rates for various ammunition calibers from a USAEC report: From 0 to 11.7%, average about 3.45%. This was part of a paper about a model for ground water contamination from unexploded ordnance. Conceptual Model for the Transport of Energetic Residues from Surface Soil to Groundwater by Range Activities

Back in World War 1 from 10-25% of the explosives failed to detonate. (One source suggested that the British dud rate went up to as much as 30% when a manpower shortage required that they draft more of the craftsmen at home.) In the American civil war duds could be 50% of a batch (not sure about statistics batch to batch).

Rifle ammo is quite simple, and except for rimfire rounds, quite reliable. Artillery isn't quite so simple, as the higher failure rate attests. How about anti-aircraft or anti-missile missiles? Several sources have informed me that these much more complicated systems were used with a "fire two to be sure" philosophy in naval operations, and that sometimes they had to "float test" a missile that didn't launch--though I can't speak from any experience myself, and dramatic incidents tend to stick in the memory more than "no problem" ones. But you might guess what sort of "dud" rate that entails.

I'm not sure how many missiles Iron Dome uses per intercept. A wikipedia image shows 6 trails, though there's no way to see how many targets there were. If anybody knows how many missiles they use (it might vary depending on the percieved risk), I'd be interested in knowing. I'm pretty sure it doesn't average more than 3, from reported cost estimates.

Of course that's not a perfectly fair way to evaluate their dud/miss rate. For an existential threat, firing more than once isn't a bad idea.

I wonder about ICBMs. I assume the details about ours are classified, so don't get me in trouble here.

Some of Trent Telenko's twitter threads are about comparing complex turnkey weapon systems with cheaper systems (not just individual weapons), and the drone/communications arms race in progress right now in Ukraine. (I'm not as sanguine as he about Ukraine's chances for victory.)

Take a machine with a million parts, each of which can wear (even computer chips wear out over time), be distorted, corroded, or just out of tolerance to begin with. Which failures will cause the system to fail to work precisely as you want?

1 comment:

  1. Can't speak to the complex stuff, but doctrine circa 2005 said that the failure rate on AT4 rockets, which are quite simple, was assumed 10-20%. Sure enough, the only time I saw a whole bunch get fired on a range day, we had 19 samples and 2 failed to fire.

    ReplyDelete