I’m speaking from a lot of ignorance here, not knowing the details of their “monitoring”. Both as a matter of tech and as a matter of the organizational behavior of what they did with the data gathered, be that great or meagre.
As a matter of principle, if they started with some factor of safety above 1.0, a certain amount of failure absolutely can be tolerated, even at max depth, before setting off a catastrophic cascade.
I suspect they did start with a factor of safety above 1.0. Maybe 1.5, maybe 2.0, whatever. Subject to the huge caveat that what they planned their factor of safety to be and what it actually was with their brand new sub net of all manufacturing and inspection uncertainties might be very different numbers. With considerable “ignorance bars” around what they thought versus the god-knowledge reality.
Of course with each dive, the factor of safety might change (probably did change) and if so only in a worse direction. Their ignorance bars would therefore be growing, except to the degree their monitoring could hold the ignorance factor constant over time.
I also suspect their monitoring was not as thorough as they thought / hoped it was. It’s unclear to me that they were able to monitor for the real problems they encountered, versus (in time-honored PHB practice) merely monitoring what’s easy to monitor and managing off that.
Lastly, from what we know of Rush, it’s a near certainty that any / all adverse monitor readings were ignored, downplayed, etc.
The sub clearly survived going to Titanic a few times. The factor of safety was at least 1.0 the first time and the last time they went down and came back up.
The evidence is they had some warning of structural problems well within human reaction times. The evidence also shows that by then the time remaining to cascade failure was less than the time to re-surface. Oops.
Not really. That’s basically the Challenger O-ring fallacy all over again, thinking that if only 1/3 of an O-ring has been burned through that there is therefore a 3x factor of safety. In fact, if the component is not supposed to burn through at all, there is zero safety factor and your component has in fact failed, and the entire system was perhaps only seconds away from a catastrophic failure with loss of life.
I can’t speak to carbon fiber pressure vessels (apparently, no one can–hence the problem) but when, for example, you use something like steel for a pressure vessel, failure begins at the microscopic level. By the time it progresses to the macroscopic level, catastrophic failure is already in progress. It goes faster than you can comprehend from “can’t detect a flaw with the naked eye” to “everyone in the engine room is now dead because we just dumped an entire steam generator’s worth of 1800 psi steam into the people space.”
My suspicion is that their fault detection system (being unprecedented in the industry, and so suffering from the same lack of testing data that the hull itself would suffer from) was not sensitive enough to detect the sort of microscopic flaws that precede an imminent catastrophic failure with enough time to arrest descent and surface before they could progress to a catastrophic failure at depth.
Honestly, the whole thing with the fault detection system seems to me like magical thinking on the part of Rush and his accomplices. How do you know that the untested real-time monitoring system will warn you of impending failures (in the hull built with similarly untested materials) with enough time to allow you to safely surface from your planned operating depth of 4,000 meters? Well because that’s what we said it should do!
This is problem that is worse with carbon fibre composites than just about any other material. They are notoriously hard to predict and test.
One area this is exemplified is in bleeding edge ocean racing yachts. Classes such as the IMOCA-60, VO65, VO70, MOD 70 and Ultim. These boats are carbon stem to stern. After every race leg they get ultrasound tested. They are filled with strain gauges to monitor loads. And they still fail. Masts can drop in well under design stresses, hulls delaminate unexpectedly. Cyclic loads are very difficult to understand and design around. CAD systems can really only implement rules based on empirical experience. Step outside of that experience and you are on your own.
Not just carbon that fails. Some years ago a boat sponsored by Safran (French aerospace company that amongst other things makes the landing gear for Aerobus) had its keel fin fail at half its design life. It was made of titanium and fabricated by Safran in their aerospace facility to great fanfare. (Since then the rules specify that keel fins are milled from a solid billet of forged steel. )
This issue here is carbon dioxide buildup, which can be scrubbed, with oxygen replenished from an interior compressed cylinder. No idea if that’s how they did it though.
I cannot imagine they would have been able to accommodate an external air tank. That’s another penetration in the hull, and something else (the O2 tank and any connecting tubing) that needs to be rated to 4000 meters (or at least 1600 meters and capable of surviving maybe a couple of descents to 4000 meters).
Whatever they were using, it was surely inside the pressure hull.
