Why did the Titanic tour submarine implode?

Much has been said about the void in the laminate. It can be measured but only after the manufacturing process but not after addition or assembly of other parts. They use NDT tester which bounces off a supersonic frequency on the backwall of the laminate. Any hole that is drilled or any dimples added would be interpreted as void. So to a test engineer, it is not possible to test it after you have drilled/milled/glued the laminate after fabrication.

To a designer, a void is counted as degradation of material property and adjustment of strength is considered in the calculations after the coupon test. A fraction to 1% is considered for autoclaved process, 2 to 3% for prepregs/out of autoclave. All other are considered bad process and is a cause for rejection. Low tech process such as infusion or wet bagging contains a lot of voids but the safety factor used is adequate to cover losses.

To the user, they consider it as a “God’s hammer” constantly expanding/contracting to find its way out when cycling thru extremes. Others call it “dieseling” effect.

To a designer, it is sometimes a boon as when it expands, it cracks the matrix. These cracks propagates until it reaches equilibrium (saturation) and no further propagation is possible. Manufacturers have been known to “crack” a pressure vessel to increase its modulus of rupture. Not much, but the increase counts. A reduction of yield strength is adjusted accordingly.

 

If the portlight failed wouldn’t the hull be intact?
Or visa-versa? Once one component failed wouldn’t the pressure be instantly equalized inside/outside, relieving the pressure off the surviving piece(s).

 

This business of cracking is good is disturbing.. The operator, believing so, hasn't even got the sense at say 8000 feet down to recognize an issue that could save them.

The difference between building the pv in a clean room or not seems the edge of lunacy as well. I mean, a clean room is expensive, but the tradeoffs also high.

Does a speck of dirt or air behave differently at different radial locations? I would think so. Like an air bubble in a 137mm laminate at 130mm from the inside would propagate cracks more than the same void at 18mm. Or is my intuition in error?

The idea, for me, is that the void or flaw closer to the surface encounters higher pressures than the same void deeper because the structure is reducing the pressure as you get toward the inside. Or no?

Another poster asked if there was an inner shell. I don't believe there was, but if so, then the pressure difference between two shells and the external environment would be a measure of main hull integrity. How to do it something else altogether...

It just seems like the idea was kind of working, but not designed away from catastrophic failure, or with any way for the operator to recognize a problem in the pv.

 

FG_Did I say any of what you are inferring?

READ carefully........
"Manufacturers have been known to “crack” a pressure vessel to increase its modulus of rupture. Not much, but the increase counts. A reduction of yield strength is adjusted accordingly."

Or do I have to cite the source(s) and author(s) of my statements every single line?

 

A lot of different industries are experimenting with carbon fiber. Sudden unexplained failures can be found on the internet on carbon fiber mountain bike and motorcycle wheels. Until fatigue failure can be better understood with carbon fiber, you either need to build in large safety factors or only use it in safer situations to study further.
Catastrophic Failure Of My BST Carbon Fibre Wheels https://www.r1-forum.com/threads/catastrophic-failure-of-my-bst-carbon-fibre-wheels.642478/

 

Because carbon fiber laminate is brittle. The allowable strain is 0.9 to 1.5.

While it is good for pressure vessel, It is a poor choice for compression vessel as its compressive strength is poor.

If you want it built to Class Rule, LR and DNV allows only 0.55 to 0.60 of its value of ultimate compressive strength which deems it impractical for it's intended use.

 

well, this is not all about you

There was a fellow on tv who rode in the sub to a shallower depth and said the sub made a lot of cracking noises..

The fact I find it disturbing is my personal opinion.

On a positive note, I completed all structural work today! It was a hard day. I did all exterior seam tapes.. thank you, again

 

Fair enough although you directed it to me. But as requested by others, I am attaching the source.

The stress/strain curve is what you will see from a typical coupon test. The curve reaches a point at at which the resin cracks, goes up a little then goes down again before going up. This is the "knee" where the resin cracks and reaches its equilibrium. Note that this is far from its ultimate strength, and far from rupture. This is proof stage where the stresses normalyses.

Attached are the excerpts from Filament Winding by S Mayers. Note that I refrained from "cut and paste" so it would not appear I am claiming the thoughts but referred to the third person(s).

 

I like that Mayers references Tsai, since a lot of Tsai's composite work goes back to things that he, the USAF, and MINS did for the DSSP in the 1970's to early 1980's. A lot of his initial work on CF was done for helicopter blades.

Oh, yeah!
But there is still some debate as both sides have proponents.
Carbon Fiber Is Safe for Submersibles When Properly Applied https://www.designnews.com/industry/carbon-fiber-safe-submersibles-when-properly-applied?
To quote the above Design News article

 

Victor Vescovo and James Cameron seems to have the same apprehension about the use of CFRP and the Hydraulic engineers seems to confirm this compression/decompression "dieseling" effect about trapped air bubble.

Following S Mayer's walk thru analysis of design, I did some analysis. It seems they did it correctly with 5" wall thickness with High Modulus CF. However the video showed only the cylinder being wound with radial/circumferencial/tangential fibers. Following S Mayers data, seems the shear strain is very low.

It is further aggravated that the form is a cylinder and have a much higher stress calculation requiring helical or polar winding to be included. Using standard engineering formula, it failed in the shear analysis if only hoop winding was used. The data I have is only 10-11 kpsi for hoop shear strength for a HSCF and I am getting 26.6kpsi calculated stress. A certain percentage of helical fibers would solve the problem.

 

Personally, I think they were expecting to get something out of the end fittings and squeeze (otherwise it makes no sense). So until everyone can see and examine their whole calculation package (if ever...), we just won't know what they were thinking.

Rickover's Rules.

https://www.allbusiness.com/seven-rules-of-safety-2-6789125-1.html

 

Fun read. But another disturbing fact which I will share is a 6 week design build window?

@rxcomposite can you explain the difference between hoop and helical in what appears to be tape? I believe the robot can change the winding angles, but for a wide tape; not sure how unless stops and starts, etc.

Composite submersibles: Under pressure in deep, deep waters https://www.compositesworld.com/articles/composite-submersibles-under-pressure-in-deep-deep-waters

 

Another thing in the video that bothered me that has not been mentioned is the ends of the cf tube appear to have been machined for the rings to fit?

I'm not sure this doesn't add a stress risor. I did not mention it because I wasn't sure about what I saw, but it looked that way to me.

And I'm not a composites expert and don't profess it.

 

Thanks for this interesting article, fallguy. What surprises me is the sentence "using a carbon fiber composite hull since 2010, primarily because it permits creation of a pressure vessel that is naturally buoyant and, therefore, would enable OceanGate to forgo the use — and the significant expense — of syntactic foam on its exterior". HD foams are quite expensive, I agree, but composite parts buildings is way more costfull...So, what kind of economical équation is that ? Also, I see that axial-oriented fibers have been used in the lamination. And, as far as I know, it is very hard to get a pure 0deg on a rotating build. That very 0deg that is required for buckling strenght. If FEA have been used for calculations, I wonder how the engineers have taken into account any fiber misalignement ? Even in the "micromechanics" calculations...

 

You speak nonsense.