Tourist submarine exploring Titanic wreckage disappears in Atlantic Ocean

[gt1guy]

Yes and no. Big issue with a cylinder is loading of the entire section as a column by the end coverings, and interaction of forces at the junction between cylinder and end closure. If you look at a thin ring of the cylinder in isolation it's the same as a ring of a sphere. The different through is the axial forces. This places the entire cylinder under load like a column, and the structure is consequently susceptible to failure in buckling. The ring frames and deep frames in a navy sub are to mitigate this, not to support directly against water pressure. Designer do take advantage of these to thin the skin, but even if the hull itself was thick enough to handle the water pressure you still need significant internal structure to protect against the entire hull buckling from axial load.

Long and short, there's no inherent reason why a tube type hull cannot work at deep depths. The reason it hasn't been done is that weight would be considerable with practical materials and there's no real reason to need a large hull capable of that depth. USN has built at least 2 boats that were cylindrical steel pressure hulls and went down to at least 3000 ft and some of the soviet boats could go quite deep. A sphere is the most effective shape from a strength/weight perspective, but other options are possible.

Much much does syntactic foam compress at that depth?

I'm following you on the tube being loaded like a column due to the end caps. I guess it's up to the smart people to figure out how long the tube can be before the buckling pressure wins out.

HZ Grade Microsphere Syntactic Foam​

PREMIUM PERFORMANCE
HZ Grade syntactics are formulated to survive crush pressures greater than 20,000 psi for safe operation in the Hadal Zone. Produced from the most advanced resin and hollow glass microspheres available, these homogeneous materials provide the industry’s lowest density for this extreme depth range.
INDUSTRY APPLICATIONS
Ultra-buoyant HZ Grade materials are predominantly used for deep water AUV, ROV and HOV applications, where reliable performance, increased payload and longer mission duration at extreme depths are critical for successful operation.
PRODUCT AVAILABILITY
HZ Grade syntactic foams are available in block form and may be readily used for customer assembly or cut to fit application requirements. Blocks or trimmed parts may also be bonded together to form larger structures.
Standard block size is 4” x 12” x 24” (102 x 305 x 610 mm).
Custom sizes and shapes are available upon request.

HZ Grade Data Sheet



PROPERTIES:
Properties provided below are typical for the cast block:

Product	Density lb/ft3 (g/cc)	Service Pressure psi (Bar)	Service Depth feet (Meters)	Uniaxial Compressive Strength psi (MPa)	Uniaxial Compressive Modulus ksi (GPa)	Hydrostatic Crush psi (Bar)	Weight Gain 24 hours @ depth
HZ-34	34 ± 2 (0.55 ± .03)	10,150 (700)	23,000 (7,000)	12,370 (85)	545 (4)	15,325 (1,057)	1% Max
HZ-43	42 ± 2 (0.69 ± .03)	18,000 (1,241)	36,000 (11,000)	16,000 (110)	650 (41)	24,000 (1,650)	1% Max
HZ-45	45 ± 2 (0.72 ± .03)	18,000 (1,241)	36,100 (11,000)	17,000 (118)	675 (5)	24,000 (1,650)	1% Max

I made the mistake of mistaking one of these type buoys for a bouncy Norwegian buoy one time. Thinking I'd kick it across the deck..................that fucker didn't even move, was hard as a rock and I though I broke my foot. Instantly did the look side to side to see who saw me being stupid.

 

[Weasel]

OceanGate Titan Submersible Failure: Consequences of Dodgy Engineering

Exploring the tragic OceanGate incident, this article deeply delves into the critical engineering flaws and dire consequences of neglecting safety.

www.electropages.com

from one of the mags I get

edit: not that great of an article, nothing much new just a overview of what's already known.

 

[3nuts*]

All the yellow you see on this ROV is flotation. It doesn't crush and the buoyancy doesn't change with depth.

Filled with diesel or other oil yeah?

 

[Thefishguy77]

The principle is not sound when you're 12,000' underwater.

They only made 3000’ish

 

[3nuts*]

Anyone been guessing where the actual point of failure was?

Can’t just be “ YeAh, iT SqUaShEd lIke a SarDiNe CaN”

More like, it imploded in the centre of the cf cylinder/ the wiring penetration hole failed/the monitor was actually mounted via screws into the cf hull…etc





I look at this and my brain…

Can’t explain it, because likely wrong but…




Good to do a poll or something and see everyone’s guess
It’ll probably take a good 12mths before we know

 

[gt1guy]

Filled with diesel or other oil yeah?

Negative. It's that fancy foam I posted above.

 

[aczlan]

Good point and it very well could have been deeper when they all got really small. Would be nice to know what the currents were at the time.

I would be curious to see what comes of the rumor that they were taking on water and had dropped ballast to try to surface.

Real time monitoring with strain gauges should be quite doable. You will have some element of elastic deformation which would be measurable and would give an indication of how the hull is behaving. In all likelihood you would be able to see deformation far enough in advance of failure to abort the dive if you know enough about the structural behavior to understand expected and unexpected responses. The key here though is you need a considerable amount of instrumentation in the right spots and detailed understanding of the hull behavior, probably informed by destructive testing.

Additionally, as I understand it, the 'monitoring system' was based on audio sensors and the assumption that hull stress would be correlated with sound somehow. While this might be possible, it seems much more likely to have been someone's innovative idea than something that was actually developed

IMO the bold part is what they were missing, they needed to take a full size version to test depth in a controlled environment (ie: the big tank in Russia) repeatedly until it broke and note the changes showed by their instrumentation from dive to die preferably with a inspection (ultrasound?) of at least a selection of marked critical points between each cycle to see if they changed from dive to dive.

While I think the concept of their launching/retrieval sled is neat, it's no substitute for being able to pull it completely out of the water and get it inside to work on something, I can't imagine trying to work on electronics or anything else that you have to trust to work that hard down on what is essentially a small submersible barge with no shelter to speak of from the waves where if you want to get back to the ship you have to get in a RIB to get run back.
I get that it's a good concept for being able to launch when it's nastier out, but giving up the ability to pull the sub completely out of the water to work on it seems as sketchy as a lot of their other shortcuts, but maybe that's just me.
gt1guy any thoughts on the sanity of their concept of servicing the sub on the platform rather than pulling it back onto the ship?


Aaron Z

 

[sdmuleman]

I'm following you on the tube being loaded like a column due to the end caps. I guess it's up to the smart people to figure out how long the tube can be before the buckling pressure wins out.

Yep. Very very complicated, especially since it's highly desirable to save as much weight as possible but simultaneously critically important it not fail ever, even when explosions happen next to it or you run into things. Lots of complex failure modes where things look like a wavy mess. Making it even more complex is the fact that you also have various tanks and internal structures as well as penetrations for hatches, tubes, shafts, valves, etc. All of these are both weak spots, but also don't behave the same way as the main structure which causes all sorts of weird behavior to figure out. Even today there's lots of unknowns and lots of very smart people who's job it is to deal with this stuff.

Fun fact - the head designer of most of the US early nuc submarines (CAPT Harry Jackson) did a bunch of experiments with model hulls made from soup cans submerged in the Chesapeake to study how pressure hulls fail.

 

[sdmuleman]

Anyone been guessing where the actual point of failure was?

Can’t just be “ YeAh, iT SqUaShEd lIke a SarDiNe CaN”

More like, it imploded in the centre of the cf cylinder/ the wiring penetration hole failed/the monitor was actually mounted via screws into the cf hull…etc

Would be interesting. My bet is failure of the CF at or near the aft titanium mating ring.

I would be curious to see what comes of the rumor that they were taking on water and had dropped ballast to try to surface.

Seems highly unlikely to me. If they dropped ballast they should have immediately started to ascend, which would reduce pressure and make it increasingly less likely to suffer catastrophic failure. Any significant leakage volume would probably flood the pressure hull to the point where it would become negatively buoyant before they got to the surface, so end result would likely be boat stuck on the bottom slowly filling with water and eventually dead people, but hull would be intact. Maybe there's a mode where leakage leads to deterioration and collapse, but it seems unlikely.

gt1guy any thoughts on the sanity of their concept of servicing the sub on the platform rather than pulling it back onto the ship?

No reason it should be possible. Pulling the boat back onboard really doesn't fundamentally gain you much. Certainly easier to work on, but if nothing needs to be worked what's the difference? No amount of operational service is really going to tell you if there's structural issues - if it's noticeable by inspection it would have failed during the dive.

 

[aczlan]

Seems highly unlikely to me. If they dropped ballast they should have immediately started to ascend, which would reduce pressure and make it increasingly less likely to suffer catastrophic failure. Any significant leakage volume would probably flood the pressure hull to the point where it would become negatively buoyant before they got to the surface, so end result would likely be boat stuck on the bottom slowly filling with water and eventually dead people, but hull would be intact. Maybe there's a mode where leakage leads to deterioration and collapse, but it seems unlikely.

Any idea how much weight they carried in ballast?
I was thinking more of one of the "control pods" leaking, shorting out and them going down, hitting and imploding, but I have no idea how close they normally were to neutral buoyancy

No reason it should be possible. Pulling the boat back onboard really doesn't fundamentally gain you much. Certainly easier to work on, but if nothing needs to be worked what's the difference? No amount of operational service is really going to tell you if there's structural issues - if it's noticeable by inspection it would have failed during the dive.

From that the one guy who ended up not going was saying, it sounded like they were spending pretty much all the time on the surface working on it.
Was thinking more of working in a warm/dry garage vs in the elements for hours on end and the resultant risk of human error.
The overall impression I got was a HS or College project where you don't have all the bugs worked out, so you pull all nighters for a week to get it working before the deadline.

Aaron Z

 

[plym49.2]

Would be interesting. My bet is failure of the CF at or near the aft titanium mating ring.

Why aft?

 

[gt1guy]

gt1guy any thoughts on the sanity of their concept of servicing the sub on the platform rather than pulling it back onto the ship?

Aaron Z

I don't understand their thinking at all with that contraption. Basically pulling a trailer with your expensive toy on it is a great way to loose it if the weather turns to shit. Also, I have never seen anything remotely like it off shore.

That said, I believe there was footage showing them overboarding it off the back of the mother ship. So I assume that was always an option, which isn't a bad idea. Probably really sketchy if it wasn't flat calm seas.

It also says that they took the platform down to ~30' before they released the sub. That make sense and is very common with other submersibles to get them through the splash zone.

I just get the feeling that that sub was not stable at all at the surface. The fact that they needed divers to release it and reattach to the platform tells me it either couldn't make fine moves and/or it didn't have any/enough external cameras to be able to do it themselves. If I had to guess, I'd say they couldn't see shit.


They certainly could have made a self centering cage that could lock it in place to come up and down. If the seas are rough, you put heave compensaters on the winch. All you need to do is to swing the whole deal over the side with a u-boom or even a crane.


Here's how the ROV's do it.

ROV is connected to the TMS with a bullet socket.
Both get lowered into the water.
TMS contains the teather for the ROV
When at depth the TMS releases the Bullet and the ROV goes to work
Do the reverse when you pick it up.

Fun fact. What looks like just a winch wire to do the lifting, actually has a core of fiber optics wires and 13k volt power wires.

 

[WLDWUN]

Why aft?

my guess would be mid point between the ends

 

[SLOWPOKE693]

How many trips to the bottom had this sub taken before it finally failed?

 

[ProjectTwin]

If you're interested, check out Ghosts of the Abyss. James Cameron and Bill Paxton spent six weeks exploring.

79 cent rental.

Amazon.com

www.amazon.com

 

[Provience]

How many trips to the bottom had this sub taken before it finally failed?

5 or something if I remember.

 

[TiTRD]

How many trips to the bottom had this sub taken before it finally failed?

One too many

 

[Mikel]

 

[aczlan]

I don't understand their thinking at all with that contraption. Basically pulling a trailer with your expensive toy on it is a great way to loose it if the weather turns to shit. Also, I have never seen anything remotely like it off shore.

That said, I believe there was footage showing them overboarding it off the back of the mother ship. So I assume that was always an option, which isn't a bad idea. Probably really sketchy if it wasn't flat calm seas.

From what I was reading, this year was the first year that they used the barge to launch from, my bet was that they were hoping to be able to use the launching barge and avoid needing to get a vessel that had a crane or A frame/u boom hoist heavy enough to pick it up.
Cheaper that way (completely ignoring the usefulness/safety factor of having a crane that can pick your sub up out of the water if you have a problem).

It also says that they took the platform down to ~30' before they released the sub. That make sense and is very common with other submersibles to get them through the splash zone.

I just get the feeling that that sub was not stable at all at the surface.

The added stability part makes sense, don't want to have your customers barfing all over the inside of the sub on their way back up or to start your trip.

The fact that they needed divers to release it and reattach to the platform tells me it either couldn't make fine moves and/or it didn't have any/enough external cameras to be able to do it themselves. If I had to guess, I'd say they couldn't see shit.

One story said they had a dive recently where they lost controls coming back in and had to be manually set in the platform by the divers.
The videos from the guy who ended up not going down made it sound like they are having frequent issues with controls not working or only working partially and that having people out working on the sub the whole time it was above water (specifically to get it working well enough that they could send it back down) was considered normal, not sure I would want to go underwater that deep on something that needed that much work, seems like it did not have all the bugs worked out, but maybe that's normal for submersible?

They certainly could have made a self centering cage that could lock it in place to come up and down. If the seas are rough, you put heave compensaters on the winch. All you need to do is to swing the whole deal over the side with a u-boom or even a crane

That sounds like something a "uninspiring 50-year-old white guy" would design

Here's how the ROV's do it.

ROV is connected to the TMS with a bullet socket.
Both get lowered into the water.
TMS contains the teather for the ROV
When at depth the TMS releases the Bullet and the ROV goes to work
Do the reverse when you pick it up.

Fun fact. What looks like just a winch wire to do the lifting, actually has a core of fiber optics wires and 13k volt power wires.

You mean people who do this for a living don't want to have to fiddle around at the water/air interface any more than absolutely necessary and want everything to work as expected all the time?
You'd think that they treated the sea like it was trying to kill them or something...

Aaron Z

 

[sdmuleman]

Any idea how much weight they carried in ballast?
I was thinking more of one of the "control pods" leaking, shorting out and them going down, hitting and imploding, but I have no idea how close they normally were to neutral buoyancy

Aaron Z

Probably a pretty small amount. Rigorous design practice would consider recoverabilty in case of a leak in one of the aux volumes, but we already know how likely that was. Nonetheless, does not seem likely. Unlikely an impact with the bottom would cause implosion based on likely low sinking rate and soft bottom. Also, the smaller external pods were likely fairly straightforward from an engineering perspective, so less likely to screw up.

Why aft?

The junction between the 2 different shapes is a discontinuity, and the difference in material only makes it worse. Thus to me one of the other end junction is the most likely spot for the failure to have initiated. We know the structure was strong enough to survive several dives, so a general collapse or similar major error is unlikely. Instead, some form of material degradation likely occurred. Again, the junctions are the most likely areas for this.

The photos of the wreckage is what makes me say aft. The fwd end bell looks perfectly intact, and I would expect there to have been some amount of damage/distortion if the failure was at that end. In contrast, I have not seen any pictures of the aft bell, and the aft non pressure hull section is ripped off the end bell. This suggests the end bell itself suffered damage, which points to collapse at that end of the hull.

 

[CDA 455 II]

nOOB question:
In essence;
Is it demonstrating stainless steel would be the ideal material to build a deep submersible vehicle?

 

[CrustyJeep]

nOOB question:

In essence;
Is it demonstrating stainless steel would be the ideal material to build a deep submersible vehicle?

Looks like aluminum would have been better than CF. I wasn't expecting that.

Also, the mild steel in the video was "seam pipe", which is pretty shitty. I wouldn't expect stainless to be that much stronger than a DOM steel tube, but..?

 

[sdmuleman]

Also have to consider things like ability to fabricate, low temperature behavior, elasticity, fatigue behavior, toughness, etc. etc. Also consider that there's a whole heck of a lot of different alloys to use. USN subs use HY100 steel - high yield strength, high ductility and relatively easy to fabricate. Somewhat like 4340.

 

[Tacodriver]

Anyone been guessing where the actual point of failure was?

Can’t just be “ YeAh, iT SqUaShEd lIke a SarDiNe CaN”

More like, it imploded in the centre of the cf cylinder/ the wiring penetration hole failed/the monitor was actually mounted via screws into the cf hull…etc





I look at this and my brain…

Can’t explain it, because likely wrong but…




Good to do a poll or something and see everyone’s guess
It’ll probably take a good 12mths before we know

Also not very deep sides for the bond I would have thought the sides would be equal to the thickness not 1 inch depth. Atherton bikes from the uk builds a bike with 3d printed Ti lugs and Cf tubes bonded, and the tubes have way more bonded surface per diameter.

 

[animator]

Also not very deep sides for the bond I would have thought the sides would be equal to the thickness not 1 inch depth. Atherton bikes from the uk builds a bike with 3d printed Ti lugs and Cf tubes bonded, and the tubes have way more bonded surface per diameter.

Well does Atherton bikes in the UK hire 50 year old white guy engineers??

 

[ProjectTwin]

nOOB question:
In essence;
Is it demonstrating stainless steel would be the ideal material to build a deep submersible vehicle?

nOOB question:

Looks like aluminum would have been better than CF. I wasn't expecting that.

Also, the mild steel in the video was "seam pipe", which is pretty shitty. I wouldn't expect stainless to be that much stronger than a DOM steel tube, but..?

The trick is that video is showing compression along the length of the cylinder.

A better representation would be compression across the diameter but there are still more variables at work. End caps and such.

 

[gt1guy]

Also have to consider things like ability to fabricate, low temperature behavior, elasticity, fatigue behavior, toughness, etc. etc. Also consider that there's a whole heck of a lot of different alloys to use. USN subs use HY100 steel - high yield strength, high ductility and relatively easy to fabricate. Somewhat like 4340.

I have a friend who was a welder/welding inspector on Navy subs for years. The procedures they follow are insane.

 

[Firstram]

I have a friend who was a welder/welding inspector on Navy subs for years. The procedures they follow are insane.

I have few friends that worked at General Dynamics as welders, they say the same thing. One of them has some pretty cool stories about destructive testing sections in a deep pool with explosives

 

[shortbus4x4]

I have a friend who was a welder/welding inspector on Navy subs for years. The procedures they follow are insane.

And they haven't had a sub implode for several decades now because of it.

 

[bdkw1]

Anyone been guessing where the actual point of failure was?

Can’t just be “ YeAh, iT SqUaShEd lIke a SarDiNe CaN”

More like, it imploded in the centre of the cf cylinder/ the wiring penetration hole failed/the monitor was actually mounted via screws into the cf hull…etc





I look at this and my brain…

Can’t explain it, because likely wrong but…




Good to do a poll or something and see everyone’s guess
It’ll probably take a good 12mths before we know

Showed the vid of them installing the ring to a buddy that does a lot of aerospace carbon work. He said the same thing about not enough space for the glue. The surfaces should have been concave. There should have also been primer on the TI ring. They also should have bead blasted the bonding surfaces on the ring.