Engineering assessment of the OceanGate Titan failure
captainsjournal.com
Engineering assessment of the OceanGate Titan failure
1–10 of 100 posts
Re: Engineering assessment of the OceanGate Titan failure
#2I know nothing, literally nothing, about material engineering, past what I can skim off wikipedia and ingest on cycling websites.
But I always thought composite nonmetal materials are hard to test nondestructively, which is the main limiting factor for their use in aerospace. Metals can be X-rayed to assess their condition, carbon not.
Is that wrong? What am I missing? Just curious.
Re: Engineering assessment of the OceanGate Titan failure
#3> From all available sources, it is apparent that the viewport was designed and certified down to a depth of 1300 meters, not 4000 meters.
Wow. What kind of hubris could look past that huge red flag?
Apparently the kind that thinks it is interesting that it deforms by several inches and not scary
Re: Engineering assessment of the OceanGate Titan failure
#4I know nothing, literally nothing, about material engineering, past what I can skim off wikipedia and ingest on cycling websites. But I always thought composite nonmetal materials are hard to test nondestructively, which is the main limiting factor for their use in aerospace. Metals can be X-rayed to assess their condition, carbon not. Is that wrong? What am I missing? Just curious.
I don't think you're missing anything. Others, including those in-the-know, were left scratching their heads too.
https://www.nytimes.com/2023/06/20/us/oceangate-titanic-miss...
> Years before OceanGate’s submersible craft went missing in the Atlantic Ocean with five people onboard, the company faced several warnings as it prepared for its hallmark mission of taking wealthy passengers to tour the Titanic’s wreckage.
> OceanGate’s director of marine operations, David Lochridge, started working on a report around that time, according to court documents, ultimately producing a scathing document in which he said the craft needed more testing and stressed “the potential dangers to passengers of the Titan as the submersible reached extreme depths.”
> Two months later, OceanGate faced similarly dire calls from more than three dozen people — industry leaders, deep-sea explorers and oceanographers — who warned in a letter to its chief executive, Stockton Rush, that the company’s “experimental” approach and its decision to forgo a traditional assessment could lead to potentially “catastrophic” problems with the Titanic mission.
Re: Engineering assessment of the OceanGate Titan failure
#5I know nothing, literally nothing, about material engineering, past what I can skim off wikipedia and ingest on cycling websites. But I always thought composite nonmetal materials are hard to test nondestructively, which is the main limiting factor for their use in aerospace. Metals can be X-rayed to assess their condition, carbon not. Is that wrong? What am I missing? Just curious.
well, a carbon frame is trash after a single crash even if it seems alright due to microscopic fissures. that's also the point cameron's making. the compound carbon material will progressively deteriorate with every dive due to the immense pressure difference from surface to target depth. extremely small deformations cause tearing and delamination. so small you can't see it. but one dive too many and it just cracks.
Re: Engineering assessment of the OceanGate Titan failure
#6> From all available sources, it is apparent that the viewport was designed and certified down to a depth of 1300 meters, not 4000 meters. Wow. What kind of hubris could look past that huge red flag? Apparently the kind that thinks it is interesting that it deforms by several inches and not scary
It's pretty much 1/3 the depth which is suspiciously close to a 3x safety factor which is what I've always heard engineers use. They certainly seemed cavalier with safety, so it wouldn't surprise me if they reasoned about it in that way.
Re: Engineering assessment of the OceanGate Titan failure
#7I know nothing, literally nothing, about material engineering, past what I can skim off wikipedia and ingest on cycling websites. But I always thought composite nonmetal materials are hard to test nondestructively, which is the main limiting factor for their use in aerospace. Metals can be X-rayed to assess their condition, carbon not. Is that wrong? What am I missing? Just curious.
You can ultrasound and radiograph CF to do NDT. I think the challenge is just more that we have over half a century headstart working with Al in aircraft to understand their failure modes and how to adequately test for them. Our understanding of CF (+CF's inherent nature of having more sudden failures) makes designing adequate NDT regimes more challenging.
That said, we are seeing much more CF (and CF composites) being deployed into commercial aerospace (famously the 787 wings, but also the A350,
Re: Engineering assessment of the OceanGate Titan failure
#8After an initial reluctance to follow this story, mostly due to the hysterical media coverage of a pointless rescue operation, my engineering half of my brain got interested. I am by no means a materials engineer nor a naval architect, but I was involved in some Design FMEAs and wrote thesis about the manufacturing of structural, pressurized aerospace carbon fiber parts.
Some things I came across, which are in line with the article and the first spurce video:
- The view port was not rated for the intended operating depth, if true I am out of words...
- the pre-preg material was apparently beyond shelf life, if true almost as bad as the view port. When pre-preg is beyong shelf life, the resign doesn't reliable bond anymore. That means basically delaminated layers before the "vessel" saw water for the first time
- 480(!) layers of carbon fibre, that carries the risk of prematurw dryong of the resin during production. Impact, see above.
- not curing and no auto-clav, again with the same results as above
- The mating rings for the emd caps have been glued on. Nothing wrong with that per se, but a bond supposed to sustain repeated dives to 4,000 m needs to be very precise. Ocean Gate applied the glue manually by hand, with inconsistent thickness in a non-clean enviroment. No idea to which tolerances the CFC hull and rings have been machined, but based the video of the bonding (they used chains attached to gate crane) they didn't seem to be too precise. The results are a high potential of foreign objects and debris being glued in, non-defined wiggle room between the CFV hull and the titanium rings and end caps and yet another strucural weakness of the vessel.
- they also used titanium fibre. If trued, it doesn't matter because the preasure was applied from the outside, all the tensile streangth of carbon and titano fibers only worls against preasure from inside the vessel.
- No back-ups, exposed wiring at the outside, flamable material at the inside, no telemetry, no emergency beacon, only wireless (!) controls (and they didn't even do enough testing to find key mapping errors of their Logitech controller...)
- Despite claiming otherwise on their homepage (archive.org still has it) they did attach equipmwnt directly to the carbon hull on the inside by screwing (!) to the hull. Result, all the fibers cut by the screws don't take any load whatsoever anymore, increased risk of delaminations (the hull is compressed, the screw on the inside isn't)
- Safety culture: completely absent, to the point of being proud about it. Unable to distinguosh between vessel certification and proper operations (there is a reason that aerospace certifies the design, production, maintenance and operations organisation seperately), firing people raising concerns, I am out of words here..
- The vessel can only be opened from the outside (!), so even a perfect dive can lead to suffocation is the vessel isn't located fast enough...
I could go on, but calling this abomination "engineering" is an insult to the universe.
Finally, some potential failure modes:
- the view port (obviously)
- the hull (delamination, stress, bad material, intentionally induced defects...)
- the bond between the rings for the end caps and the hull (weakened bond due to different deformation of the two materials, weak bond due to foreign objects, inconsistens bond due to bad application of the glue...)
All in all, from design over manufacturing to maintenance and operations, this thing was a death tral whos only upside was a quick death in case of an implosion.
Re: Engineering assessment of the OceanGate Titan failure
#9One can take comfort from this in a perverse way. The failure of the submersible was entirely predictable and should surprise no one. The alternative would be more troubling.
Re: Engineering assessment of the OceanGate Titan failure
#10> From all available sources, it is apparent that the viewport was designed and certified down to a depth of 1300 meters, not 4000 meters. Wow. What kind of hubris could look past that huge red flag? Apparently the kind that thinks it is interesting that it deforms by several inches and not scary
It's pretty much 1/3 the depth which is suspiciously close to a 3x safety factor which is what I've always heard engineers use. They certainly seemed cavalier with safety, so it wouldn't surprise me if they reasoned about it in that way.
And there is a reason it is called safety factor. Whether it is 10, 5, 3 or 1.1 doesn't matter, it is not be used.