Any flaw (such as a dent) would be pushed out by the internal pressure or if it ruptures, the pressure is almost instantly reduced.
Whereas, with external pressure, the dent would provide a weakened focal point for a rupture to allow the entire ocean to try to rush in.
Maybe it was I read that the PSI on Titan was 6,500 and the PSI for SCUBA is around 3000.
Big difference.
For a vessel pressurised on the inside, nearly all of the material will be in tension - like the skin of a balloon); with a vessel resisting pressure from the outside, it’s more complex - the outer parts are subject to compressive forces - the inner parts might be in tension as a result of that - and to an extent, this can all be compensated for by pre-tensioning the material during manufacture, although maybe not enough to compensate for the sorts of forces this thing had to endure.
Sloppy analogy: String is great for making a hammock, which will resist force/pressure on the concave side. String is not great for building an arch, which will need to resist force on the convex side.
There are reports claiming that the porthole was only certified for less than half the depth that the vessel was diving to - if this is true, then they were relying very heavily on the engineered safety margin every time - that’s the sort of excessive use where a tiny flaw (like a microscopic scratch added when it is cleaned) could easily be all it takes to trigger a catastrophic failure.
Not exactly. The carbon-fiber cyclinder had two titanium rings glued to the ends of the cylinder, with titanium hemispheres bolted to the titanium rings. There was no hatch per se. Instead the hemisphere with the the porthole was unbolted each time to allow for entry and egress. The occupants were then bolted inside, with no capability of opening the “hatch” from the inside.
This presented another failure mode. If they had surfaced and were bobbing around, they would have eventually suffocated if they were not found in time by the support vessel.
In addition to simple material overstress, vessels with higher pressure on the outside (e.g. submarines) are prone to buckling failure:
Buckling is a failure mode you won’t see in vessels with higher pressure on the inside (e.g. scuba tanks). To prevent buckling, a submarine’s walls need to be considerably thicker than a simple material stress calculation would suggest.
Most likely the former. Sort of. For a thin-walled cylindrical pressure vessel, hoop stress is twice the axial stress; this is why, for example, copper pipes almost always fail with a rupture running parallel to the pipe axis. So, assuming the Titan’s hull had comparable compressive strength in the circumferential and axial directions, it probably experienced a radial compressive failure, being crushed radially instead of axially. So it wouldn’t be flattened like a tin can stomped on its side, it would be squeezed from all directions radially into something with a smaller diameter.
Because the Titan’s hull wall was thick compared to its diameter, it has to be analyzed as a thick-walled pressure vessel. From here, the ID was 66 inches, and I recall reading elsewhere that the hull was five inches thick (so 76 inches OD). There’s an online calculator for stresses in a thick-walled cylindrical pressure vessel:
I used:
You can specify the radial distance to the point of interest in the hull material, e.g. 33 inches will give stress values at the inner surface of the hull, and 38 inches will give you stress values at the outer surface. No matter what location you choose in the hull wall, the hoop stress is much greater than the axial stress, and at the inner surface, it’s fully twice the axial stress, almost 53,000 psi.
The strength of the actual carbon fibers themselves (reportedly about 500K psi)is only half of the picture. The other half is the matrix, i.e. the epoxy/polymer that’s gluing all of those layers of carbon fiber together. So the actual strength of the composite material (not just the fibers) is probably quite a bit less than 500K psi. Exactly what that strength is depends on the choice of matrix material, the construction methods, how well the matrix actually bonds with the fibers, and what’s happened to the finished product during its lifetime.
Obviously there are a lot of variables that might make things different in the case of the Titan submersible, but I have seen a very small scale pressure test of a carbon fibre cylinder with metal end caps - it resisted significant pressure, then there was a sharp cracking sound and it failed instantaneously; when retrieved from the test rig, the entire tube was split along its length - probably initially from buckling, then this ran away into a tear along the grain of the material.
It’s a wobbly and quite informal video, but interesting and possibly relevant, despite the difference in scale:
The water would have flowed in through this crack to fill the inside of the vessel in a fraction of a second. In the case of a larger object such as the Titan sub, the mass of the total volume of moving water is much greater (as well as the pressure) - the water may have rebounded once or more after imploding, which would shred anything nearby - as I understand it, only small pieces of debris from the carbon fibre part of the vessel have been found.
From the screenshot of that video, the wall thickness is pretty small compared to the cylinder radius, maybe 1:20. With that in mind, it’s not surprising that it failed by buckling, resulting in localized bending and fracture. Titan’s wall cylinder is much thicker compared to its radius, about 1:6.6, so far more resistant to buckling failure.
I think @GIGObuster’s video upthread is probably more relevant to how the material of Titan’s hull failed, i.e. simple compressive hoop overstress resulting in delamination and shredding rather than a clean fracture like the one seen in your video.
One of those pieces looks like a big ring - woul that be the piece glued to a cylinder end?
My thought is that rather than a one-horse-shay total failure, there was a weak spot on the cylinder somewhere - as it buckled, that spot would cave in first. Whatever the hammer blow of inrushing water would be, that would cause further destruction; perhaps rushing in then the sudden impact hitting all sides from inside being like setting off an explosion (or hammering) the rest of the cylinder.
A failure of a SCBA tank would be less catastrophic because the pressure would be relieved fairly quickly as the tank vented; whereas the pressure of the water in a sub failure would be essentially constant.
Isn’t that the tensile strength? Does carbon fiber have any compressive strength even when held firmly by the matrix? A string is great for pulling on something but not so great for pushing. Carbon fiber would be good for a pressure vessel (hoop stress is tensile) but not for what is in effect a vacuum vessel (hoop stress in compressive). Seems like most of the compressive strength is provided by the matrix. The tensile strength of the carbon fiber might be useful is some ways, but it doesn’t help resist compression.
Apparently it does, according to this PDF:
COMPRESSIVE STRENGTH OF A CARBON FIBER IN MATRIX
Compressive strength of a single carbon fiber in epoxy matrix was investigated. Apparent compressive strength of carbon fiber in matrix was measured by four-point bending test. Actual compressive strength was calculated based on the apparent compressive strength. It is concluded that the fiber compressive strength is almost same to the tensile strength.
As noted earlier, the strength of the composite material (carbon fiber + matrix) depends on many factors beyond just the strength of the carbon fibers themselves. Thinking of the 5" thick Titan hull, I wonder what percentage of the cross-sectional area was actually composed of carbon fiber, and what percentage was the epoxy/matrix.
When you consider a bigger arc, compressive strength for any small area actually is resistance to bending, more like shear strength. If one area is weaker (or weakened) due to flaws in construction, or starts to delaminate, the problem will only get worse with each successive stress event. Any flaw will spread, until there’s that one point where it’s weak enough to buckle, which then causes the adjacent areas to buckle until the entire area fails completely. This is where it’s important to know what level of quality control was in effect during manufacturing, and whether any measures were taken to check for impending failure after each dive.
This seems to me to be one possible scenario, unless the ends, or the hatch or the window failed first. I suspect one area buckling is more likely than the entire circumference basically being “choked” compressively to a smaller diameter and then total failure.
Report that human remains have been found.
One nitpick: In a thick-walled composite structure, any bending/buckling force will put the fibers on top in compression, but the fibers on the bottom will be in tension. So the tensile strength of the fibers comes into play when the wall is buckling.
However, any delaminations or air gaps or water intrusion between layers will totally screw things up.
I remember a composite Long-Eze that had a wing failure in flight, and the inspectors discovered that the builder had not attached the wings to the spar at all. The airplane wings stayed together by the sheer strength of the foam and fiberglass composite skin. Until they didn’t.
In the footage where the front dome is being lifted, the webbing is strung through the porthole - this means the transparent window is not in place. I suppose it could have been blown out by the forces rebounding from the implosion though.
The BBC’s video:
Here’s a possible failure mode: the immense differential pressure on the composite tube has water trying to find a way through it, a pore that lets water intrude. As a sneaking finger of water penetrates along a pore or weak path, it starts to separate the wall in the radial direction (which might be quite weak). A patch of wall delaminates, making a popping or banging sound, which has been reported by previous passengers. But this isn’t a runaway failure because the pore still throttles the water entering the delamination, and the delamination relieves the high pressure in the water filling it. Successive dives gradually increase the size of these delaminations until one lowers resistance to buckling enough that the tube starts to buckle, and the process rapidly runs away.
Some have said there was monitoring of the hull, such that they might detect a change before the catastrophic runaway. There was a report that recovery folks had found ballast was dumped before the catastrophe, as if people inside reacted to a sudden but limited incremental change before the runaway.
From what I’ve seen, the ballast was only attached using open S-hooks anyway (something about if the mechanical release mechanism failed, the occupants of the sub would be able to rock it back and forth to try to unhook the ballast), so it’s possible the ballast just shook loose in the turmoil of the implosion.
I assume a catastrophic implosion created some interesting shock waves, which in water are going to bounce all over. Basically, giant hammer blows everywhere. Any weak points will also fail.
My earlier point was that if the initial implosion follows the length of the tube, then the resulting shock wave will concentrate as it hits the hemispheric end piece(s) and possibly rebound to affect the other end. A wavefront channeled into a constriction like a hemisphere will be heavily magnified. To answer the OP, if it pushed the contents of the sub ahead of it, then this resulted in a concentrated hammer-blow into the small apex of the hemisphere where they could be "presumed… " found. It sounds not unlike the process used to create artificial diamonds by concentrating pressure into a diminishing area.
