OceanGate Submersible Goes Missing During Titanic Dive

[toucana]

The process of conducting an underwater search for a lost submarine has been based on the use of Bayesian Search Theory ever since 1968, when the technique was first successfully used in the hunt for USS Scorpion (SSN-589).

This nuclear powered submarine had gone missing while returning from combat patrol to its base in Norfolk Virginia, and might in theory have sunk anywhere between there and its last known location near the Canary Islands. A team of mathematicians and acoustic specialists led by John Piña Craven calculated an optimum search box area, and subsequently located the wreck at a depth of 3047m, and within about 500 metres of the central X of the primary search box near the Azores -  (The US Navy thought it had sunk off the Eastern Seaboard).

https://en.wikipedia.org/wiki/Bayesian_search_theory

In the case of the Titan submersible, the search teams could define an optimum ‘box’ based on the known time of the loss of contact at 1h 45m into the dive, and the planned descent route to the wreck of the Titanic.

[Genady]

4 hours ago, mistermack said:

It's a bit surprising that they didn't hear anything on the mother ship, when the sub imploded. Sound carrys well in water, and they were only about 3km from the sub at the time. With the pressures involved, there should have been a tremendous bang when the hull failed, but no sound was mentioned in reports.

There was a bang, and it has been detected: Secret US Navy underwater microphones detected Titan sub implosion | The Independent

[StringJunky]

The descent takes 4-5 hours to Titanic. They imploded 1 hour 45 minutes from the start. What do any of you with knowledge think was the likely depth range?

Edited June 23, 2023 by StringJunky

[Genady]

14 minutes ago, StringJunky said:

The descent takes 4-5 hours to Titanic.

That is what I thought, too. But it turns out to be wrong:

Quote

It usually takes two hours to descend to the wreck.

How does the missing Titan submersible work? Here's a look inside (usatoday.com)

[StringJunky]

1 hour ago, Genady said:

That is what I thought, too. But it turns out to be wrong:

How does the missing Titan submersible work? Here's a look inside (usatoday.com)

Do you think it's a ten hour round trip, including Titanic search time? 10 hours there and back is the likely dive aim?

Edited June 23, 2023 by StringJunky

[Alex_Krycek]

• Not hiring 50 year old white guys with decades of submarine experience because they're not "inspirational"
• Piloting the sub with a Logtech controller that can be used by a 16 year old
• buying "off the shelf parts" and using an experimental carbon fiber hull

For me this a good example of the Overconfidence Bias as outlined by Amos Tversky and Daniel Kahneman.  Unfortunately it seems all those years of never being contradicted by yes men went to Stockton's head.

Great interview here with James Cameron and Bob Ballard on why the hull failure should have never happened.  Since 1960 there has never been an accident of this kind at these depths.  It was only when Stockton "got creative" and deviated from decades of proven engineering that this tragedy occurred.  

 

 

[John Cuthber]

2 hours ago, Alex_Krycek said:

Not hiring 50 year old white guys with decades of submarine experience because they're not "inspirational"

Some people are citing this as an argument against diversity.
Those are the people that don't  realise that "everyone is young" is no more diverse than "everyone is old".
They aren't helping.

 

11 hours ago, toucana said:

The process of conducting an underwater search for a lost submarine has been based on the use of Bayesian Search Theory ever since 1968,

Interestingly, they also searched the surface of the sea quite intensively in spite of  the fact that ... that's not where submarines "should" be.

An application of this algorithm
https://en.wikipedia.org/wiki/Streetlight_effect

[Genady]

6 hours ago, StringJunky said:

Do you think it's a ten hour round trip, including Titanic search time? 10 hours there and back is the likely dive aim?

Yes, 8-10 hours round trip. E.g.,

Quote

 Each full dive to the wreck, including the descent and ascent, reportedly takes around eight hours.

Titanic tourist submersible goes missing with search under way - BBC News

[mistermack]

3 hours ago, Alex_Krycek said:

For me this a good example of the Overconfidence Bias as outlined by Amos Tversky and Daniel Kahneman.  Unfortunately it seems all those years of never being contradicted by yes men went to Stockton's head.

Eerily like the Titanic disaster itself. The "unsinkable" Titanic having too few lifeboats, and not cutting it's speed when icebergs were about. 

And the phenomenon of one disaster causing follow-on disasters happens somewhere every day on motorways and freeways, when people slow down and look at an accident that's happened in the opposite carriageway. It's called rubber necking, and you don't have to be a billionaire to fall victim to it.

One snippet of info that James Cameron came up with in an interview was that in the last communication that the ship had with Titan, they said that they had "dropped the weights" meaning that they were coming back up, so if that's right, they had some sort of clue that something was wrong. 

Maybe the use of carbon fibre was a mistake in this case. I'm aware that it's a very strong material in tension, but I haven't heard the same about strength in compression or shear forces. But that's just my speculation. Others were wondering about the window material. It might be that we'll never know. 

Normally, these things are tested to destruction at the design stage. Sending an unmanned craft down to much deeper depths multiples of times would give you more confidence, but it would probably make the whole project uneconomic to do that. 

[Danijel Gorupec]

21 minutes ago, John Cuthber said:

Interestingly, they also searched the surface of the sea quite intensively in spite of  the fact that ... that's not where submarines "should" be.

I would expect more likely for such a sub to surface than not. Hull implosion seems to me like (it should be) the least expected fault.... In fact, I considered the underwater search like some sort of wasting resources that could be used for the surface search.

That said, I am unpleasantly surprised that the sub did not have any orange-colored part. Cannot think of any reasonable explanation.

[Genady]

I think that rather than Bayesian Search Theory, this information was more helpful:

Quote

the navy had detected "an acoustic anomaly consistent with an implosion" shortly after the Titan lost contact with the surface.

The official said the information had been relayed to the US Coast Guard team, which used it to narrow the radius of the search area.

Titanic director James Cameron accuses OceanGate of cutting corners - BBC News

[StringJunky]

2 hours ago, mistermack said:

Eerily like the Titanic disaster itself. The "unsinkable" Titanic having too few lifeboats, and not cutting it's speed when icebergs were about. 

And the phenomenon of one disaster causing follow-on disasters happens somewhere every day on motorways and freeways, when people slow down and look at an accident that's happened in the opposite carriageway. It's called rubber necking, and you don't have to be a billionaire to fall victim to it.

One snippet of info that James Cameron came up with in an interview was that in the last communication that the ship had with Titan, they said that they had "dropped the weights" meaning that they were coming back up, so if that's right, they had some sort of clue that something was wrong. 

Maybe the use of carbon fibre was a mistake in this case. I'm aware that it's a very strong material in tension, but I haven't heard the same about strength in compression or shear forces. But that's just my speculation. Others were wondering about the window material. It might be that we'll never know. 

Normally, these things are tested to destruction at the design stage. Sending an unmanned craft down to much deeper depths multiples of times would give you more confidence, but it would probably make the whole project uneconomic to do that. 

I feelthat carbon fibre stays in conformation until there's an asymmetric load on the hull, like bumping  into something. I thought carbon fibre had super tensility, but crush resistance was weak? Broke too many fishing rods myself. Failure is instant.

[Genady]

@StringJunky, the article that I've linked in the previous comment, contains an answer to your earlier question:

Quote

Within about an hour I had the following facts. They were on descent. They were at 3,500 metres (11,483ft), heading for the bottom at 3,800 metres.

 

[TheVat]

There were safety concerns raised several years ago, about the carbon fiber hull.   And the viewport being only certified to a much shallower depth.  

https://www.npr.org/2023/06/21/1183408455/titan-missing-submarine-oceangate-submersible

The rating for the viewport seems especially a red flag. From the NPR report...

Lochridge said he first raised his safety and quality control concerns verbally to executive management, which ignored them. He then sought to address the problems and offer solutions in a report.

The day after it was submitted, the lawsuit says, various engineering and HR executives invited him to a meeting at which he learned that the viewport of the submersible was only built to a certified pressure of 1,300 meters, even though the Titanic shipwreck lies nearly 4,000 meters below sea level.

Lochridge reiterated his concerns, but the lawsuit alleges that rather than take corrective action, OceanGate "did the exact opposite."

"OceanGate gave Lochridge approximately 10 minutes to immediately clear out his desk and exit the premises," it said.

1 hour ago, StringJunky said:

I feelthat carbon fibre stays in conformation until there's an asymmetric load on the hull, like bumping  into something. I thought carbon fibre had super tensility, but crush resistance was weak? Broke too many fishing rods myself. Failure is instant.

Some years ago I would hear a cycling friend and his fellow cyclists debating carbon fiber.  Everyone agreed you could shatter it with a hammer so it wasn't good for bikes that got rough use like mountain bikes.  The plus was that it didn't fatigue like metal and it was light, so in theory you could race it forever.

Edited June 23, 2023 by TheVat
spmgor

[StringJunky]

10 minutes ago, TheVat said:

There were safety concerns raised several years ago, about the carbon fiber hull.   And the viewport being only certified to a much shallower depth.  

https://www.npr.org/2023/06/21/1183408455/titan-missing-submarine-oceangate-submersible

The rating for the viewport seems especially a red flag. From the NPR report...

Lochridge said he first raised his safety and quality control concerns verbally to executive management, which ignored them. He then sought to address the problems and offer solutions in a report.

The day after it was submitted, the lawsuit says, various engineering and HR executives invited him to a meeting at which he learned that the viewport of the submersible was only built to a certified pressure of 1,300 meters, even though the Titanic shipwreck lies nearly 4,000 meters below sea level.

Lochridge reiterated his concerns, but the lawsuit alleges that rather than take corrective action, OceanGate "did the exact opposite."

"OceanGate gave Lochridge approximately 10 minutes to immediately clear out his desk and exit the premises," it said.

Some years ago I would hear a cycling friend and his fellow cyclists debating carbon fiber.  Everyone agreed you could shatter it with a hammer so it wasn't good for bikes that got rough use like mountain bikes.  The plus was that it didn't fatigue like metal, so in theory you could race it forever.

Another thing is that it has a limited life in fishing due to uv degradation. The speed at which a rod recovers from being bent slows down over time. In fishing parlance it's called 'recovery time'.  It basically goes from springy to floppy over time.

[toucana]

3 hours ago, Genady said:

I think that rather than Bayesian Search Theory, this information was more helpful:

Titanic director James Cameron accuses OceanGate of cutting corners - BBC News

 

In the case of the 1968 search for the USS Scorpion which I cited, it was the triangulation of recordings of an underwater explosion recorded by several secret US Navy listening stations which persuaded John Craven’s team to relocate their search area to the Azores. The analysis of these audio recordings was a key input into drawing up the Bayesian probability density maps which led to the discovery of the wreck.

Subsequent surveys of the wreck site disclosed that the USS Scorpion had suffered a devastating explosion in its forward torpedo room, and that uncontrollable flooding had caused the vessel to sink below its crush depth and implode - which was the event recorded by the  listening stations.

 

[Genady]

1 hour ago, toucana said:

 

In the case of the 1968 search for the USS Scorpion which I cited, it was the triangulation of recordings of an underwater explosion recorded by several secret US Navy listening stations which persuaded John Craven’s team to relocate their search area to the Azores. The analysis of these audio recordings was a key input into drawing up the Bayesian probability density maps which led to the discovery of the wreck.

Subsequent surveys of the wreck site disclosed that the USS Scorpion had suffered a devastating explosion in its forward torpedo room, and that uncontrollable flooding had caused the vessel to sink below its crush depth and implode - which was the event recorded by the  listening stations.

 

I see. Thank you. Such input is a significant factor. However, it is not a Bayesian update. The latter would've been an update based on intermediate results of an ongoing search, i.e., the step 6 in Bayesian search theory - Wikipedia:

Quote

Revise all the probabilities continuously during the search. For example, if the hypotheses for location X imply the likely disintegration of the object and the search at location X has yielded no fragments, then the probability that the object is somewhere around there is greatly reduced (though not usually to zero) while the probabilities of its being at other locations is correspondingly increased. The revision process is done by applying Bayes' theorem.

 

[John Cuthber]

On 6/20/2023 at 11:53 AM, toucana said:

“It’s got one button, and that’s it. We run the sub with this game controller - It’s made by Logitec, but it’s basically a Sony PS style controller”  he says  -

What could possibly go wrong ?

How many other systems were as well tested as a game controller?

[StringJunky]

1 hour ago, John Cuthber said:

How many other systems were as well tested as a game controller?

But built to a price, probably not adequate for a mission as critical as that. The price you pay in high-end mission critical equipment is the testing of the individual products. ie The test sampling rate on the production would be much higher than a shop bought controller.

Edited June 23, 2023 by StringJunky

[toucana]

4 hours ago, Genady said:

I see. Thank you. Such input is a significant factor. However, it is not a Bayesian update. The latter would've been an update based on intermediate results of an ongoing search, i.e., the step 6 in Bayesian search theory - Wikipedia:

 

Bayesian Search Methods are based on Bayesian Statistics (as your own citation points out in its first paragraph).

Bayesian Statistical modelling  (as far as I understand it ) doesn’t just refer to the process of making computational updates of probabilities in the light of new data -  it also takes account of the degrees of belief in an event - which may be based on prior knowledge about an event - such as the results of previous experiments - or the personal beliefs about the event held by the observer.

https://en.wikipedia.org/wiki/Bayesian_statistics

[studiot]

26 minutes ago, toucana said:

Bayesian Search Methods are based on Bayesian Statistics (as your own citation points out in its first paragraph).

Bayesian Statistical modelling  (as far as I understand it ) doesn’t just refer to the process of making computational updates of probabilities in the light of new data -  it also takes account of the degrees of belief in an event - which may be based on prior knowledge about an event - such as the results of previous experiments - or the personal beliefs about the event held by the observer.

https://en.wikipedia.org/wiki/Bayesian_statistics

Well yes and no.

Since Bayes theorem gives a conditional probability, that is a probability given so other result, it is necessary to start with a value and apply the condition.
Conventionally in the absence of any other information (you called it belief) we start with P = 0.5.

There is one other point to mention about Bayesian searches.

The Scorpion was travelling basically horizontally

The Titan was travelling basically vertically.

 

Here are some thoughts as to the the failure. First I must stess that at this stage we really don't know, we must just draw up a list of factors to investigate.

1) The carbon fibre issue may be pointing the finger in the wrong direction.

2) Yes carbon fibre is very strong and does not fatigue appreciably.

3) But the fibres are embedded in a matrix which is a different matter and this may easily creep, although the sub was lass than 10years old.
I remember the 1960s fibreglass kit cars, the Fairthorpe in particular.
A fellow student was killed in one in those days where the body shell had crept appreciably over its mounting points.

4) This bring me to another point. It is all very well having superstrong materials, but it has to be mated with other parts and so may fail at the union. I understand that an intact titanium nose cone has been found.

5) I remember also similar problems besetting the RB211 Rolls-Royce aero engine, the first with carbon fibre fan blades,  eventually leading to shortened life and a couple of disasters. The blades were eventually replaced with metal ones.

So was it a material failure or a naval architecture vessel design failure ?

 

Edited June 23, 2023 by studiot

[iNow]

1 minute ago, studiot said:

So was it a material failure or a naval architecture vessel design failure ?

Yes

[Genady]

One more bit of information regarding the search and statistics, Bayesian or not:

Quote

"We detected its implosion with our underwater listening system, and that's what the Navy told the Coast Guard the moment it happened. They picked up the implosion, passed it on to the Coast Guard, so we immediately knew where to go and where to look, but it just took time to get the ROVs down there, and they finally got there yesterday," Ballard said.

Robert Ballard, who discovered Titanic wreck, said authorities 'immediately knew' where to look when they heard the Titan's implosion, but it took days for an ROV to get there (yahoo.com)

And another:

Quote

“For me, there was no doubt. I knew that sub was sitting exactly underneath its last known depth and position, and that’s exactly where they found it. There was no search. When they finally got an ROV down there that could make the depth, they found it within hours. Probably within minutes.”

'Titanic' director James Cameron says the search for the missing sub became a 'nightmarish charade' - ABC News (go.com)

[mistermack]

It's becoming clearer that there has been a lot of "news management" in this incident. 

I posted earlier that there would have been a very loud bang when the sub imploded, and that they would surely have heard it, or at least detected it on the ship. We now know that the coastguard was told of that noise immediately, and they would have informed those on the ship. So even if they didn't hear it themselves, they knew something had gone bang, and that could only have meant one thing, unfortunately. 

And that's why they found the wreckage so quickly. They knew it had gone bang. They knew exactly where it was when it went bang. And as the sub was only about 15 minutes off the bottom, they knew that the wreckage would be directly below that spot, exactly as James Cameron said. So as soon as the ROV arrived, it went straight to the wreckage. 

So all the talk about oxygen supplies, and the scanning the surface, was just to cover the one-in-a-million chance that the sound of the collapse was a freak false lead, but they all knew in their heart of hearts that the worst had happened. 

I'm quite claustrophobic, I wouldn't even go in a cave or a mine, unless I had to. I sometimes look up in wonder when I'm on the ground floor of an office block, and picture the lot coming down on my head. So I can't imagine how people can go down to those depths voluntarily. 

If I was going down 300 feet, I would like to see the reports where it had previously been tested down to 1,000 feet. Minumum. But that doesn't seem to be the case with Titan. The testing seems to have been only to a small percentage over the Titanic depth. I just couldn't handle the thought of that. 

But the weird thing is that the people who died on the Titanic would have felt no hint of danger, dressed up for dinner or dancing on the ballroom floor. But they were actually in just as much danger as these five. 

It's a strange world. 

[Alex_Krycek]

This article at PM offers further analysis on the engineering of the craft, and how the composition differed from other battle tested submersibles.

https://www.popularmechanics.com/science/a44305386/titan-submersible-vs-deepsea-challenger/

Excerpt:

"The Deepsea Challenger—made famous by film director and underwater explorer James Cameron—was mostly built with a special syntactic foam, with the crew compartment composed of a steel orb attached to foam beams. “Roughly 70 percent of the vehicle's volume is made up of this foam, which is composed of glass spheres embedded in an epoxy resin and which provides both flotation and structural support for the vehicle,” according to the Woods Hole Oceanographic Institution, which now operates the vessel."

"The Titan differed in other ways. Unlike the spherical shape of the Deepsea Challenger that distributed pressure equally, the Titan’s cylindrical shape meant that some parts of the craft were subject to more pressure than others. At a depth of more than two miles beneath the surface, even a small hull breach would cause an instantaneous implosion."