But the fact remains that gas bottles do not spontaneously burst; they do so when they are being interfered with in some way, during handling or charging. This cylinder was sitting in its stowage out of the way of any normal "interference".
Without doubt the oxygen cylinder is a crucial element of the incident. As stated above, there had to be a 'discontinuity' (either physical or chemical) that induced the failure of the mounting bracket(s) and severed the valve and/or the regulator (with or without part of the cylinder) which then flew upwards through the cargo hold ceiling and the cabin floor. It is (IMO) inconceivable that an intact cylinder could inflict the damage seen in the cabin (ie there is no way a complete cylinder could levitate with sufficient force) therefore
at some stage either the valve became detatched from the cylinder
or part of the cylinder with the valve still attached was propelled from its original storage location.
Although plausible, I do not believe that a simple displacement of the cylinder (as shown in
would, in itself, cause sufficient damage to the structure of the cylinder to cause separation of the valve. Furthermore, (IMO) I don't believe that an external 'explosion' (or rapid burning) caused by leakage of oxygen from connecting pipework or valves could inflict destructive damage on an otherwise previously intact and undamaged cylinder. Such events usually occur when a flaw exists in a fitting fitted into the cylinder or an attempt to remove the fitting has been made. Failure of external pipework is unlikely to 'destroy' the cylinder itself, although dropping the cylinder
could 'knock off' the valve screwed into the cylinder.
If the cylinder was properly fixed into the support brackets and clamped in place then failure of any of the external connecting pipework is most unlikely (improbable - maybe impossible) to cause the catastrophic failure that occurred.
Therefore, either the cylinder was displaced from its mountings prior to the 'explosion', or the cylinder itself fractured. The former suggests that the mounting bracket(s) either simultaneously failed (unlikely) or that some mechanical force (such as part of the cargo) impinged on the cylinder, or that an 'explosion' displaced the cylinder from its stowage location. Subsequently (and maybe simultaneously) damage to the valve (and/or the cylinder) allowed the valve to detach itself.
Simple failure of the fuselage skin (ie
no primary failure of the cylinder) could cause rotation of the cylinder in its clamp as it was sucked out of the breach, but (IMO) the valve wouldn't detach until it struck something solid (such as the flange along the bottom of the 'curtain' or the edges of the remaining upper support bracket - there should be witness marks) and the state of the remaining pipework would indicate how the cylinder 'departed'.
I'd love to be able to view the scene (and it would help to know exactly what debris has been recovered).
It's easy to assume that 'cylinders don't fail spontaneously', yet should a cylinder fail . . .
. . . the result would be as seems to have occurred here.