Submersible missing while visiting Titanic wreck

[johansen]

this forum is one of the worst, for people nit picking every perceived falsehood in someone's comments.

It's an engineering forum full of engineers.

It's full of nit-picking pedants who like to fell trees instead of seeing the woods, point to the scratch on the display while missing the info it's displaying, argue the pronunciation of an ingredient instead of savouring the taste, etc.

(Some exceptions, though)

It's also full of people who are incapable of interpreting metaphors, and wondering why one would talk about chopping down forest trees in an electronics forum.

its the autism....

[G0HZU]

A lot more images have been released showing the remains of the Titan. The youtube video below contains a lot of fresh images showing how the carbon fibre layers have become delaminated.  Also, some remaining parts of the carbon fibre hull are quite large.

[tom66]

It's incredible that they had data showing that the hull had deformed somewhat (based on the strain gauge readings) and didn't immediately decide to pressure-test the vessel.

[dietert1]

Yes, the nonlinear strain gauge curves indicated that the hull was no longer responding to pressure as expected from a rigid body. It was in pieces.

[Gyro]

Looking at the non-linearity increasing in every dive after 80 should have been a total red flag. That it was happening at relatively shallow depths makes it look as if it needed some increasing pressure to force the (now partially unglued) 5 layers back together.

Looking at the way the flanges, which look surprisingly thin, on the end rings were sheared off, inside and out, there could have been no way that the ends of the hull were still cylindrical when the ultimate failure happened. The reverse bends in some of the hull fragments are terrifying.

[nctnico]

It's incredible that they had data showing that the hull had deformed somewhat (based on the strain gauge readings) and didn't immediately decide to pressure-test the vessel.

Probably because all the people who had actual knowledge had already left or had been fired long before.

[Siwastaja]

Looking at the non-linearity increasing in every dive after 80 should have been a total red flag

It's the opposite: because it has been increasing every dive, it was the new normal. Why would it implode this time?

Every detail in this story sounds incredible in isolation, but if you think about the big picture where nothing mattered and everything was ran by luck only, then it all makes sense.

[NiHaoMike]

It's just like Challenger and Columbia where concerns were ignored since the Shuttles made their way back. That is, until they didn't...

[Phil1977]

With the subtle difference that the space shuttle had thousands of possibly lethal failures and NASA spent huge - though still too little - effort to keep it safe.

This sub had practically one main function: To withstand the pressure. Usually I really don't like criticizing work that I don't completely understand, but here it seems to be a combination of bad engineering and completely reckless safety culture.

[nctnico]

With the subtle difference that the space shuttle had thousands of possibly lethal failures and NASA spent huge - though still too little - effort to keep it safe.

This sub had practically one main function: To withstand the pressure. Usually I really don't like criticizing work that I don't completely understand, but here it seems to be a combination of bad engineering and completely reckless safety culture.

Reckless safety: yes because the tell-tale signs where there the hull was compromised. Bad engineering not so much. The V2 hull was much better compared to the V1 hull and some testing has been done. Nevertheless it looks like the V2 hull had manufacturing defects as well which should have been mitigated using the monitoring system which did seem to work as intended and the design seemed to have enough margin to make quite a few dives before total failure. So from an engineering perspective I would call that a fit-for-purpose design. IMHO the conclusion is that they should have tested the V2 hull after the dive where the 'loud bang' occured or at least after the strain gauges indicated the stress on the hull has become uneven (for whatever reason). Instead they choose to ignore the warnings and continued to dive. But that is human error, not a design flaw.

[PlainName]

Looking at the non-linearity increasing in every dive after 80 should have been a total red flag.

That's only apparent if you compare previous dives by overlaying the data as shown, and even then it's not that noticeable unless you're looking for something - anything - to explain a catastrophe. I bet they didn't compare, or if they did only superficially, and instead were looking for something much more obvious - a spike or dent, perhaps - and at depth rather than wallowing about in the wave field.

That it was happening at relatively shallow depths makes it look as if it needed some increasing pressure to force the (now partially unglued) 5 layers back together.

Again, obvious in hindsight. But that it was only at shallow depths perhaps mitigated. Where it really mattered, at proper depths, it looked normal.

[dietert1]

No the design wasn't good enough. The last video explains that the minimum safety margin to get any kind of certification would have been 1.25 and they didn't achieve that. They didn't have any certificate that allowed them to dive with humans inside.
Then there are the doubts about leaving the submersible in a parking lot exposed to winter weather for several months. Did they realize it was bad and were giving up? If they changed their minds later for some reason (probably money), then the whole story is already close to criminal intent.

[nctnico]

No the design wasn't good enough. The last video explains that the minimum safety margin to get any kind of certification would have been 1.25 and they didn't achieve that. They didn't have any certificate that allowed them to dive with humans inside.

This is just legal nitpicking. What if they tested it to 1.25 times the pressure? Would that make the hull work any better? Would the hull have lasted forever? IMHO the idea of the realtime monitoring system is a good one. A pressure test is just 1 data point and without data to extrapolate an estimated lifespan, the data becomes irrelevant very quickly.

Then there are the doubts about leaving the submersible in a parking lot exposed to winter weather for several months. Did they realize it was bad and were giving up? If they changed their minds later for some reason (probably money), then the whole story is already close to criminal intent.

People keep tripping up over this but so far I have not seen any hard evidence pointing towards storage and towing actions being harmfull to the hull in a way it would weaken it. Maybe further investigation can answer that.

[Phil1977]

Bad engineering not so much.

I can't help: Good engineering would have relied on a DOE that at least proofs the durability of your material and proofs the stability of your production process by some destructive tests. Good engineering would have demanded independent certifications of the design and of the manufacturing process. Good engineering would have included a risk management in the style of a FMEA.

Here it seems a single guy just wanted to get famous. You could say he succeeded in this - sadly for his passengers.

[nctnico]

Bad engineering not so much.

I can't help: Good engineering would have relied on a DOE that at least proofs the durability of your material and proofs the stability of your production process by some destructive tests. Good engineering would have demanded independent certifications of the design and of the manufacturing process. Good engineering would have included a risk management in the style of a FMEA.

AFAIK they did all that up to a certain point and I will agree they didn't do enough due to a pig headed CEO. OTOH the testing of scale models and building a V2 hull shows that there has been (destructive) testing going on and the design has been improved that way over a few iterations. Getting certifications is not a guarantee a submersible is actually safe to operate (just as passing CE testing is not a guarantee your product is reliable and/or 100% safe). Maybe a carbon composite hull is uncertifyable due to lack of data despite the fact that others have succesfully build carbon composite hulls for submersibles. There really is a grey area between engineering and legal nitpicking where it comes to certifications.

The reality is that certification always lags behind technical progress. Some of the products I work on can't be certified to be compatible with other products because there simply isn't a certification lab (yet) which can verify compatibility indepedently.

And do not be lead astray by Youtube videos made by non-experts which draw all kinds of premature conclusions. Also keep in mind that witness statements from former employees may be colored by anger, resentment or having a different view which is not necessarily right or true (and can be included out of context in Youtube videos for dramatic effect; the video linked in a few posts above is a good example of doing that). The investigative team has access to all data, debris and independent expertise; give it time and let them draw the final conclusion. So far they have been collecting facts and data.

[TimFox]

Looking at the non-linearity increasing in every dive after 80 should have been a total red flag.

That's only apparent if you compare previous dives by overlaying the data as shown, and even then it's not that noticeable unless you're looking for something - anything - to explain a catastrophe. I bet they didn't compare, or if they did only superficially, and instead were looking for something much more obvious - a spike or dent, perhaps - and at depth rather than wallowing about in the wave field.

That it was happening at relatively shallow depths makes it look as if it needed some increasing pressure to force the (now partially unglued) 5 layers back together.

Again, obvious in hindsight. But that it was only at shallow depths perhaps mitigated. Where it really mattered, at proper depths, it looked normal.

One of the great engineering disasters was the collapse of a bridge over the St Lawrence River in 1907.  Apparently, there was evidence that the structure was yielding (in the technical sense) before the cantilever was completed, but management blamed bad measurements and put a crew out on the incomplete structure, which failed and resulted in the death of 75 construction workers.
https://engagedscholarship.csuohio.edu/cgi/viewcontent.cgi?article=1020&context=encee_facpub
In one account, the engineering error was the use of weight calculations based on the overall architectural drawings, rather than on the detail drawings of the individual components, underestimating the weight.

[Andy Chee]

Maybe a carbon composite hull is uncertifyable due to lack of data despite the fact that others have succesfully build carbon composite hulls for submersibles. There really is a grey area between engineering and legal nitpicking where it comes to certifications.

Lack of data is a big problem for engineering as well, not just lawyers!

[floobydust]

Engineers at some point need to stop technical fault-finding and being the fall guy for everything that fails, blaming themselves to no end. It's that goddamn failed bridge the Iron Ring is made from lol.
As I keep saying "engineers carry all of the responsibility, yet have none of the authority".
The best engineering can be ruined, subverted by corrupt leadership, management or manufacturing. It's becoming much more commonplace.

These sociopathic CEO's will see a Speed Limit 50 sign and rip past at 85 or 100. The rules, laws don't apply to them because they are special. Every time they do this without getting a ticket, it reinforces their belief system as superior human beings, innovators etc.
Eventually you have a monster that only lives to weasel around regulations and flogs engineers and staff to get what they want. Ref. Boeing, Theranos, Nikola, FIU bridge etc.

[nctnico]

While everybody is concentrating on the delamination of the hull, what about the viewport? As far as I know the viewport has not been found / recovered so far. Also the viewport seems to be a critical part of the submersible which was severely underrated for the depth. What also puzzles me is why the outer flanges which held the hull in place on the front and end titanium rings have sheared off as well. With an implosion of the hull I would expert the outer flanges to stay intact (or at least not shear off as much as they did).

Based on the above (assuming the viewport hasn't been recovered), maybe the following happened:
- Viewport cracks and gets pushed in by the water
- Compression of the air causes the temperature to rise quickly and vapourising whatever is inside (people, plastics - including the viewport-) forming a combustible mixture.
- An explosion follows (*) which blows the hull apart outward, tearing the flanges off the front & end rings and seperates the lid fromt the front section.
- After the explosion, the water pressure pushes some debris into the rear section.

So what I'm proposing is A) that the hull didn't break by itself and B) the viewport no longer exists as it has been burned up.

The way I see it, the viewport is an important piece towards finding out what happened.

* In a diesel engine the pressure for self-ignition is around 50 to 60 bar. So the pressure would be high enough to self-ignite a combustible gas mixture.

[johansen]

Anyone who thinks that the veiwport failing blows up the sub, does not understand physics.

The water rushes in, bounces off the rear hull. A momentary jolt happens due to the kinetic momentum transfer, the sub fills up. And sinks in one piece.

18 million pounds force axially on both ends. For a brief moment there is 18 in one direction and 16 in the other assuming a perfect pelton wheel type reaction as the water hits the back of the sub and changes direction while maintaining 800 feet per second velocity.

There is still a net 16 million pounds force holding thr sub together

I dont think there was actually 1 million pounds force on the viewport. A 12" diameter circle is around 600,000 pounds.

[pqass]

The (wedge-fit) viewport likely popped-off like a champagne cork after water entered at the front ring-CF interface (opposite that remaining large CF plank).  As the hull shatters and collapses into the rear dome.

If water entered via the viewport then that would have behaved like a immense jet and popped the rear dome clean off since the CF hull (now as a pressure vessel) is stronger than the CF-rear ring interface and may have remained intact.  Yes compressed gases do get very hot but that would only be after the volume is much reduced.  The CF-rear ring would fail before the hull in this scenario.

Just my guess.

[EEVblog]

[johansen]

The (wedge-fit) viewport likely popped-off like a champagne cork after water entered at the front ring-CF interface (opposite that remaining large CF plank).  As the hull shatters and collapses into the rear dome.

If water entered via the viewport then that would have behaved like a immense jet and popped the rear dome clean off since the CF hull (now as a pressure vessel) is stronger than the CF-rear ring interface and may have remained intact.  Yes compressed gases do get very hot but that would only be after the volume is much reduced.  The CF-rear ring would fail before the hull in this scenario.

Just my guess.

If the viewport is a cannon ball and you have a barrel and a bust disk.. yes in theory you can pop off the back hemisphere. the initial peak pressure of a solid mass 1 foot in diameter stopping instantly could but is unlikely exceed 18 million pounds force. But the energy of 6 million foot pounds (600,000 pounds force from the water being pushed 10 feet into the back of the sub) Could only theoretically push the rear hemisphere a theoretical maximum of 4 inches, because its being held by 18 million pounds force. This neglects the momentum of the water behind the rear dome which is also holding it in place!

so the reality is at best a cannon ball accelerated to 650 feet per second will probably crack the glue joint, but not materially separate the hull to the point that anything else fails. if you had a cannon ball and a barrel, the hydraulic peak pressure generated inside that barrel when the cannon ball stops at the rear dome would break the barrel and separate the back dome, but again, this is not what existed in real life. and, there's a good chance such an experiment would result in the cannon ball passing through the rear dome, if the barrel was stronger than the dome.

so in real life the idea that an initially broken stream of water passing through fragments of the viewport and hitting everyone inside, has enough kinetic momenum (again its only traveling at a maximum of 800 feet per second) to blow the back dome off.. is just fantasy. the reaction force never exceeds 2 million pounds force which is 1/9th that which is required to separate the rear dome from the front dome.

this also neglects the dampening effect of the uh, people, breaking up the initial stream.

remember that once the water passes through the view port, it is only slowing down from there, it does not collect any more energy on the way, You need a toroidal deflector spike to get the doubling effect of the force as the water flows back to the view port at 700 feet per second instead of 800. it didn't have one.


think of the viewport and the sub as a 6000 psi feynman sprinkler. it doesn't move when the viewport pops...

[dietert1]

Machining the wrinkles means cutting fibers, so between every 2 of those 5x 1" layers there is a layer of fiber cut into pieces about the size of one revolution or even less. Probably a thickness of some millimeters, where the fiber won't behave as one big knot anymore.
Once two of the five layers delaminated in some portion the process continues until implosion. When all five layers delaminate at their interfaces, mechanical resistance diminishes by a factor five. In civil engineering we cannot substitute a 15" beam by five 3" beams.
They also machined the hull to fit the end flanges. I'd guess that green glue was 100x weaker than undamaged carbon fiber.

Regards, Dieter

[5U4GB]

It's the opposite: because it has been increasing every dive, it was the new normal. Why would it implode this time?

The technical term for this is deviation of normalcy.  It happens an awful lot in practice, which is why it's good to either have experts around who can say "this isn't normal any more" or to occasionally bring in fresh faces who can look at it and say "should that gauge really be reading that high?".