As Mark Antony said:
I have come to bury Caesar, not to praise him.
The evil that men do lives after them;
The good is oft interrèd with their bones.
Should have aimed for “I shall return.”.
In principle, if your initial factor of safety is high enough, and your microfailure monitoring is sensitive enough, and you actually react promptly and diligently to microfailures by repairing the structure to new condition or by condemning the structure if it’s irreparable to new condition, then you can achieve safety by monitoring.
It appears Oceangate flunked all 3 parts of my syllogism. And therefore comprehensively and spectacularly flunked the conclusion.
Bravo.
Don’t forget, it must be resublimated thiotimoline
I always forget that part!
The problem with this as I see it is even if the RTM spotted issues how would they be pinpointed? They have already admitted there is no way to inspect carbon fiber this thick. And even then, so what? You can’t cut away a bunch of outer fibers and repair an internal flaw. That leaves condemning the structure.
'Zactly. It seems like they were planning to fail from the git-go. Whether they fully understood that or not.
More likely, they actually expected the monitoring system to never report a problem. Then when it did they pulled the “normalization of deviance” maneuver just like NASA did with their O-rings. But with even less engineering rigor and discipline.
“We got an alert and finished the dive and all was well. That proves there’s more structural margin than we thought. We are so-oo smart.” Lather rinse repeat for a few more dives while the margins get ever slimmer and totally unnoticed, or at least unappreciated.
Then a brief wail of alarming, maybe even a bit of cascading crescendo-ing alarming as fibers snapped and voids coalesced, then the big crunch-smoosh before they knew what hit them.
I strongly suspect that if they ever got an alert, it was both the first and last time. Because the entire idea that failure wouldn’t rapidly progress from minimum detectable the catastrophic in the course of a single dive seems not to have crossed their mind, and yet I strongly suspect that is exactly what happened.
No normalization of deviance. Just a substandard design, operating far to the left of the curve from the get-go.
Is this actually true? A very large industrial CT machine could do it. It would not be cheap.
Even properly built subs are only rated for so many dives. The hull is essentially a wear item. In principle, if the structure could be modeled properly, they could figure out the point where it becomes unsafe to use further. The progression of defects would have been apparent if they’d been able to look closely. But we’ll probably never know what it was that did this one in. At least not until and unless someone decides to repeat the experiment.
Their monitoring systems don’t inspire confidence. You might be able to detect snapping fibers acoustically, but that’s not going to detect delamination at all.
Still probably not. The size of defects would likely be smaller than a voxel. Probably laminar in shape with an extent wider than quite a few voxels but of essentially zero thickness. So no actual change in radio-absorption. Just a failure of the inter fibre bonding. If there were actual voids in the structure big enough to image the structure would have already failed at depth.
Ultrasound picks up the interface of a bond failure but is time consuming and in a wall five inches thick will likely have a lot of trouble seeing defects deep inside.
Hmm. The best CT scanners have a resolution of a few micrometers. I’d think that would be enough. Particularly if the assumption is that any defect will propagate from some starting point. Any voids or other inclusions will change shape if the defect starts to spread. You might not be able to track it past a certain point (delaminated layers might only be separated by a few nanometers), but you should be able to track the initial progression.
But not of a size that could hold the Titan. There are some fabulous scanners for small samples. But the limitations imposed by source extent and sensor size make scanners something that scales in all aspects.
Fair. That said, you don’t have to scan the whole thing all at once. You would need an emitter and receiver on the inside and outside of the hull, and some means (robot arm, etc.) of moving them to different parts of it. For example, something like this unit made for horses (!!!):
Recall that the lead engineer resigned, or was fired, because Mr. Rush declined to perform non-destructive inspection of the hull. OceanGate then sued the engineer for good measure.
There was also this business reported where the Titan allowable diving depth was derated for a time, and that the hull was either repaired or replaced, and then everything was copacetic again. Another reporter doubted any replacement of the hull took place during this time period, supply chain issues because of the lockdowns prevented securing the necessary carbon fiber or somesuch.
So that leaves “repair”. How was it done, and who performed the repairs? I wouldn’t want my signature on something like that.
Mainly I’m just wondering if “can’t be inspected” was a “stopped clock” moment for Rush. It’s bad either way: either there was a means of performing NDT and they lied/didn’t know about it, or they built something that couldn’t be inspected and just assumed it was safe. Or a little of each!
NDT is $$$ - it’s pretty spendy, this brought up here before. That seems to be a recurring theme.
Does anyone know what a ball-park cost for a carbon fibre hull would be? Or the sub overall? Given that they were getting $1M for each dive (4 “mission specialists” at $250K a pop) how many dives would it take before the sub was paid for? Or how many dives before replacing the hull would be economically viable?
Thanks ! Saved me the trouble of looking it up !..
In case anyone’s interested, Dr. Chris Raynor has weighed in on what the victims would look like.