Thunderfoot Is Wrong About Why the Titan Submersible Failed
Why the Titan Submersible Sunk
The Questionable Engineering Behind Titan
More about the Use of Carbon Fibre
He Tried to Warn the Titan Sub Of Its Downfall
While Thurderfoot’s points are interesting, and his theory is intriguing, it fails to account for the evidence we have available. There are three possible causes for this disaster: 1) The effects of repeated stress on carbon fibre, 2) the window, 3) the mismatch in compressibility between the titanium cap and the carbon fibre body.
Acrylic windows are common among submersibles, and the principles behind their design are well understood. What was novel about the Titan was its use of two different types of material, the use of carbon fibre, and its decision to combine them using an epoxy. All things considered it is the novel features of the submersible that are more likely to fail: Not the standard ones.
As thunderfoot points out, the use of two different materials that have different levels of compressibility could cause the seal to crack as one side contracts and the other, less compressible side doesn’t, producing a gap. [By the way, for those who wonder why OceanGate did not test the submersible with autonomous dives, the fact that carbon fibre fails after repeated stress is the likely reason. He probably hoped that after enough dives, with enough passengers, he could get the money to build a proper submersible—and he did not want to use up his dive cycles with tests. This, of course, implies he was aware of just how reckless his design was.] However, if this had been the cause of the collapse, you would have expected Titan to have failed in an earlier dive. Nor does Thunderfoot’s theory account for the submersible’s attempts to resurface or the loud noise picked up by the navy’s listening stations.
No, the most likely explanation is that the carbon fibre weakened over multiple dives: This would explain why the crew dropped its ballast, since they could hear the deafening noise made by the cracking of the carbon fibre hull. Weakening over repeated dives also explains why these problems didn’t occur on an earlier dive. Thunderfoot himself wonders how the ship managed so many successful dives. This is a tacit admission that his theory doesn’t hold up.
The deeper cause of the submersible’s failure is the poisonous effect of the Silicon Valley “move fast and break things” mythos. This mythos works well with software that is not carrying out any vitally important functions. However, it does not work in matters of life and death. Despite his brilliance, Elon Musk’s most dangerous effect on our culture is that he has encouraged the propagation of this mythos into engineering based industries.
Lastly, I have noticed that our culture seems to be moving towards thinking of simulations as a substitute for experiments and real world testing: While modeling is incredibly useful, confusing models with reality is a serious mistake. Models are not reality: Models are only as good as the assumptions that underlie them. They can show you unexpected consequences of your premises, but they cannot tell you if your premises are wrong. OceanGate, if it was not involved in overt fraud, may have believed its vehicle was safe because they ‘ran simulations.’ However, those simulations assumed that the carbon fiber, a very complex material, would behave a certain way. In short, carbon fiber is hard to model.
I have also noticed that Silicon Valley culture seems to think of expenditures on safety as wasteful. Again, this might make sense in the context of certain software applications. But in contexts where death or serious injury are possible, safety is not, wasteful. Waste is an expense that is unnecessary. Safety expenditures are necessary: They are intended to bring risks within limits normal people are willing to tolerate.
Stockton said, “Safety is waste.” Well, so is death.
Postscript:
If someone were to repeat this ill-fated idea, he should do the following:
Design the carbon fibre with the tubules running in criss-crossed directions; if this proves too expensive, then at least use two hules, one with the carbon running one way and the other with it oriented perpendicularly.
The carbon fibre hull should have been placed in an autoclave to ensure there were no gaps air bubbles. Furthermore, it should have been placed in one after each dive to “heal” the epoxy.
The seal between the two sections should have been done in a dust free location under more controlled conditions.
Some form of “gasket” should have been placed at over the location of the join to reinforce it.
A titanium alloy that matches the compressibility of the carbon fibre should be used for the end caps.
Obviously, a window rated to the relevant depth should be used.
It is likely the window necessitated the titanium cap: the carbon fibre would subject the acrylic to too much pressure. As such, the carbon fibre tube should have extended well beyond the attachment point for the titanium cap.
While this might seem to contradict point 7, it seems that Stockton joined the two ends in the wrong way: He had the titanium cap on the outside and the carbon fibre on the inside. If you have two different materials, you need the less compressible material on the inside so that it lends its compressive strength to the more compressible material. In short, there should have been two carbon fibre shells, an inner one and an outer one so that the cap could lie inside the outer shell while the carbon fibre still extended beyond the end cap.
Of course, in reality, no one should ever attempt this idea again. Just use titanium.