The Unforgiving Sea: OceanGate, Titan, and Unforgiven Hubris

Michael Siegel

Michael Siegel is an astronomer living in Pennsylvania. He blogs at his own site, and has written a novel.

Related Post Roulette

23 Responses

  1. Damon says:

    “submersible exploration has actually had an excellent safety record because of the enormous care people put into it. We’ve just gotten a horrifying demonstration of what can happen when people don’t.”

    Exactly. Deep water and Space are, as you said, unforgiving. At least one guy, who’s video I posted elsewhere, had the sense to not go….Report

  2. Oscar Gordon says:

    Been a minute, I know…

    Comet: Fatigue, yes, but also crack propagation and sharp corners (Ever wonder why airlines have oval windows? The comet has square windows.).

    Titan: Oh, so much. I suspect that will be a subject of engineering classes for decades to come, and why engineering safety standards are not the same as bureaucratic red tape. I suspect it in the end, the carbon fiber will be the failure point, probably where it joined the titanium caps. At those pressures and temperatures (water is cold down there), it will be susceptible to embrittlement and delamination, and if they weren’t running a scan of the hull after every dive, they’d never have known.Report

    • This is precisely my thought. In Cameron’s video, they made a single sphere of steel so that it was uniform and a perfect shape for resisting pressure. Once you start mixing materials … it changes.Report

    • DensityDuck in reply to Oscar Gordon says:

      What I hope people take from this is that we ought to teach undergrad engineers more about CTE than that it exists, and to teach them more about designing for it than “it’s bad, mmmkay?”Report

      • Oscar Gordon in reply to DensityDuck says:

        CTE? Construction Testing & Evaluation?Report

        • DensityDuck in reply to Oscar Gordon says:

          Coefficient of Thermal Expansion. You can make a carbon-fiber structure strong enough to handle whatever load, but if it gets cold it shrinks, and if the metal fittings shrink at a different rate than the rest of the structure they’ll crack it to pieces.Report

          • Oscar Gordon in reply to DensityDuck says:

            Odd, discussions about thermal expansion were very much a part of my training, and I wasn’t even doing a lot of design work (mostly analysis).

            In my circles, we’ve been talking about how they clearly weren’t doing any CFD or FEA, or if they were, it was very idealistic analysis, with unrealistic variables.Report

            • Jaybird in reply to Oscar Gordon says:

              Wait wait wait wait.

              It’s now sinking in: This sub has made this journey before, hasn’t it? Like, this was the 4th or 8th or 12th trip it made down there.

              And the attitude was “well, it survived last time, it’ll survive this time”?Report

              • Oscar Gordon in reply to Jaybird says:

                It had dived to the target depth twice before (IIRC).

                Materials, when loaded to failure, have two key points of failure: Yield, and Fail.

                Fail is when it is very obviously broken/deformed/shattered/cracked, etc.

                Yield is when it’s loaded past the point when the material can return to it’s original shape and strength, but it’s not failed yet.

                Think of a metal bar. The bar can bend when loaded, and then spring back to it’s original flat shape when unloaded. This is called ‘elastic deformation’.

                But if you load that too much, it will yield, and suffer ‘plastic deformation’. It will have a permanent bend in it. Even if you bend it back to it’s flat shape, it’s lot a significant amount of strength, and can no longer carry the load it had before.*

                All materials that are loaded past their yield strength can no longer be trusted, even if they look fine.

                Oh, and temperature and pressure? Those can affect yield strength, a lot! Hot metals will bend sooner. Cold temperatures cause embrittlement, reducing the ability of the material to elastically deform. Hell, the cold and lack of pressure in space causes all sorts of strange material behavior (re: vacuum welding).

                Different types of materials have different yield characteristics. Ductile metals deform/bend. Ceramics tend to crack.

                Laminated composites don’t plastically deform like metals, or crack like ceramics. They de-laminate. The layers of fabric that are bonded together with resin, the resin cracks and the layers begin to separate. It’s a quiet failure, very little sound, and can start deep in the material, where it can only be seen with X-rays / MRI type imaging.

                So sure, you go down once, twice. During those trips, the cold, and the pressures, and the different rates at which the composite and the metal expand & contract against each other cause the CF to begin de-laminating, probably at the join of the end caps and the cylinder. Once it’s starts failing, it’s only a matter of time, and if they aren’t scanning the hull after every dive, they’ll never know.

                And any decent engineering school should have their students taking materials science courses, and more than one (I had 4 by the time I had my BS).Report

              • Oscar Gordon in reply to Jaybird says:

                Oh, and as an added feature, I heard that the guy had built the sub using expired pre-preg aerospace CF.

                Pre-preg CF is carbon fiber that comes pre-impregnanted with resin from the manufacturers. It’s easy to work with and the user doesn’t need to mess with mixing resins and applying them, just wrap your shape and shove it in the autoclave.

                But pre-preg has a use-by date. The resin begins curing right away, although very slowly at room temp and normal atmospheric pressure. But still, past a certain date, the resin is partially cured enough that the advertised strength is not possible. So places like Boeing will sell it cheap to hobbyists and smaller manufacturers, who use it to make bikes / kayaks / guitars / custom body panels for a car / anything that doesn’t need to really worry about having the stuff at full strength.

                Diving deep – you want full strength.Report

              • DensityDuck in reply to Oscar Gordon says:

                “I heard that the guy had built the sub using expired pre-preg aerospace CF.”

                wut dot jpeg

                ok so obviously the guy is beyond man’s justice but I think that the people who actually built the thing need to answer some hard questions here, because they were knowingly creating a hazard to human life! This isn’t just “well we’re sure he’s going to get the certification wait what do you mean he didn’t”, this is obviously not good material!Report

              • Oscar Gordon in reply to Oscar Gordon says:

                Another scary bit from the NYT:

                In the documents, Mr. Lochridge reported learning that the viewport that lets passengers see outside the craft was only certified to work in depths of up to 1,300 meters.

                That is far less than would be necessary for trips to the Titanic, which is nearly 4,000 meters below the ocean’s surface.

                https://www.nytimes.com/2023/06/20/us/oceangate-titanic-missing-submersible.htmlReport

              • Fish in reply to Jaybird says:

                “It’s now sinking in…”

                I see what you did there.Report

              • Burt Likko in reply to Fish says:

                As long as we keep giving this discourse oxygen, you can expect this sort of thing.Report

            • DensityDuck in reply to Oscar Gordon says:

              “discussions about thermal expansion were very much a part of my training…”

              I mean, less “it exists” and more “here are the features you put in a design to deal with it”. That sort of thing I didn’t find out about until a few years into my career, and mostly it came from me asking “why is that designed the way it is”…Report

              • Oscar Gordon in reply to DensityDuck says:

                Yeah, since my program was analysis heavy, I have pretty shallow knowledge as to how you design for it, but we sure as hell factored it into the analysis (part A will expand faster than B, causing these stresses to build up in the parts unless you deal with it; how to deal with it is an exercise for the mechanical/civil engineers).

                Still, I too didn’t learn HOW such things are done until after I got into industry. So yeah, the schools aren’t always good about that.Report

  3. Jaybird says:

    Saw this from eigen and I winced and nodded sadly.

    Report