Is there any reason not to think any control lines were run through the metal end caps instead of the carbon fiber? I would think that a known solution.
But wouldn’t that create a pressure risk? Anything that carries those wires from inside to outside has to resist the pressure.
Isn’t carbon fibre first made into (crossweaved?) mats? So would this be more like rolling a moderately wide ribbon spiralling around a tube, so some longitudinal strength?
It’s highly variable. But no, not necessarily. For an application where absolute strength is crucial, you’ll want to directly use “tow”, which is basically a CF ribbon. And you’ll want to place that ribbon in a way that makes sense for the application, since the strength is only along one axis (it’s “anisotropic”). Usually that means some kind of cross-weaving to give it strength along multiple axes (which might not all have the same strength along them). The most sophisticated designs use a robot to get the most precise and uniform placement.
The question about electrical connections has been answered above. There isn’t anything new about passing signals into a pressure vessel. There are off the shelf solutions. None of them involve wires in flexible insulation. Rather solid conductors in ceramic or high strength solid polymer.
The pressures involved on the Titan were huge, but not worse than are managed in a lot of other industrial applications. The problem was the cowboy approach to engineering.
Thanks, I obviously didn’t understand it the first time. Ignorance fought!
‘Here’s how you are supposed to do it’ has been answered.
‘Here’s how they actually did it’ has not, as far as I can see.
That’s what they used for the Titan - here’s footage of it being wound on OceanGate’s YouTube channel - the only sense in which there was crossweaving was that they wound it spirally one way along the steel tube then spirally back in the other direction - so the fibres are crossed by a couple of degrees maybe.
Aha, so it’s much like the unidirectional fibreglas used for homebuilt aircraft (specifically the wing spar stiffening), with minimal cross-threads to keep the cloth intact and 90% in the one direction.
It seems like this construction would lose integrity in the axial direction after repeated cycles of compression and release - the whole thing must be shrinking a little bit when put under massive pressure then expanding when it comes back up and with no fibres running lengthways, it’s only the resin bearing that stress - there’s potential for it to decompose somewhat toward becoming a huge slinky
Not quite what I was thinking about for “robot.” Is that even CNC? This is the type of CF placement robot I’m thinking of:
But thanks for the link; that’s nuts that they have no CF in the axial direction. That’s usually about half the hoop stress in a cylindrical vessel, so that’s still quite a lot. I’d think you’d want windings at ~27 degrees (arctan(1/2)). And cross-wound to prevent it from coming apart like a spring.
Wikipedia has some more details on the CF winding, with this cite:
Specifically relevant:
Spencer opted for a layup strategy that combines alternating placement of prepreg carbon fiber/epoxy unidirectional fabrics in the axial direction, with wet winding of carbon fiber/epoxy in the hoop direction, for a total of 480 plies
“Prepreg” is basically CF fabric with the binder already impregnated. At any rate, they did have some axial strength at least. So things aren’t quite as bad as the video makes it look.
I found this an interesting bit of trivia:
It is believed that the first time carbon fiber composites were applied to the hull of a deep-diving, manned submersible was for the one-person DeepFlight Challenger , commissioned by adventurist Steve Fossett in 2000 for a dive to the bottom of Challenger Deep
Fossett was of course killed while flying. Seems like both these domains are catnip to the high-risk-tolerance adventurer type.
Yeah, I think the OceanGate thing is a robot in the same sense my washing machine is a robot. Technically the truth.
Their manufacturing processes also all seem a bit un-clinical; people standing over stuff, probably contaminating it with hairs, dust and sweat, clothing fibres etc; assembly of critical components in spaces with open windows and doors for the wind to blow in dirt and insects. https://www.youtube.com/watch?v=WK99kBS1AfE
Agreed. Not that everything needs a cleanroom, but all of their processes look pretty uncontrolled to me.
My money is still on a progressive delamination mode. After a few cycles, some of the layers lost their adhesion and eventually buckled relative to each other. Could have started with a minor defect or contamination, but in a large object there’s always going to be some weak point. But the nature of this application means that small defects aren’t self-limiting in scope. They just keep propagating until there’s a large defect. And they can’t be observed without sophisticated equipment.
+1 … resulting eventual in a delamination-runaway (walkaway? creepaway?)
kindalike a crack in a windshield of a car … but it gets stressed twice in each dive … going down and going up
It wouldn’t matter so much if the primary purpose of using carbon fibre was just weight reduction of a component for more ordinary use, where the absolute limits of material strength would never come into play.
Whatever method of communicating with the thrusters the sub used, it didn’t work well. (Who would have guessed?)
Wow, I’m really shocked by that bonding process video. All sorts of things could be going wrong there, especially creating voids and weak spots with no apparent way to test for them, some due to capturing bubbles between mating surfaces, some due to contamination (the guy wiping with a dirty looking rag while also touching the surface with his other hand), and maybe even some due to nonuniform strength in the adhesive getting mixed in a tub and scooped out in a way that would collect the mix right against the tub wall where it’s least well blended. I assumed that they had mounted the two end rings onto whatever mandrel they wound the fiber with, and incrementally wound the fiber in ways that captured more and more design features built into the rings. But there’s NONE of that! They just glued the ends together. I’d expect voids or weak spots in the joint to shift radially due to different moduli and/or thermal coefficients of expansion, and gradually grow, providing a path for water to probe deeper and deeper into the assembly.
As far as I can tell, the mandrel is a permanent fixture - the end flanges might have been cut off after the composite was built up, to allow the titanium ring to be bonded to the carbon fibre, but the steel tube/mandrel is just left in place to be the inner wall of the pressure vessel.
And yeah, the epoxy preparation and application is especially shocking - a guy just mixing it in a tub with a wooden stick (and struggling to mix it, which doesn’t bode well); apparently no measures to degas or even to verify uniformity of mixing. The guys who pour resin to make river tables in YouTube videos have a more clinical process than that.
Bingo.
Look at mangetout’s winding video at the :40 sec mark, that texture looks to me like some bad inclusion.