Though my business partner, Rob DeCesare and I have left the industry, as two of the few people
in the world to have designed, constructed, and dove a carbon fiber/composite pressure section
submersible, I feel obligated to add my $.02 to the conversation. I’d like to throw praise towards
the industry professionals whose comments have been driven by fact and appropriate decorum
concerning still-unknown circumstances that cost the lives of 5 people. Everyone wants
immediate answers—partly out of curiosity, but certainly no one wants to see another life lost to
what we will eventually learn happened to Titan. I’ve written this post in the hope that some may
learn lessons that extend well beyond the submersible community.
Similar to Stockton Rush, my education and professional background were born in aerospace.
And though I have an MBA and taught business school, the majority of my experience is in
operations. I flew military combat aircraft before joining the airline industry, where I am
currently a captain at a major airline .
Forged in failure, the under-appreciated reality of commercial aviation is that it evolved from
being a fairly risky adventure into one of the safest activities on the planet, short of perhaps remaining still and breathing. It’s literally safer to fly than to
have a meal at the dinner table. This level of near-perfection didn’t happen by accident. The industry adopted a continuous improvement process decades
ago by which no victim of flight would die in vain. But even today, aviation professionals are continually having to defend that process from powerful
political forces that would upend it in a heartbeat. As an example, pilots involved in accidents have occasionally been thrown in jail. That response has often
failed to recognize there is a difference between a crime and a mistake. It’s convenient ignorance when the powers that be are in need of a villain. Where
there is legitimate intent and no gross negligence, a crime has not been committed. The pound of flesh taken from a well-intentioned pilot is nowhere near
as valuable as gleaning every last drop of information from that pilot regarding what lead to an accident (and therefore how to prevent similar accidents).
But you don’t get the latter if you pursue the former.
We need to let the Titan investigation provide the answers, but some facts have already come into focus. The Oceangate CEO appears to have had firm
control over most of the important operational decisions of his company. Where large mistakes may be found, they were most likely his. In spite of what was
likely some poor decisions, Mr. Rush appears to have had no ill-intent. The legal definitions of gross negligence vary. If any remaining OceanGate
stakeholders are shown to have been negligent with no ill-intent, I believe they should be given immunity from prosecution. The lessons they can teach us
are more valuable than their pound of flesh.
Much of the investigation will no-doubt focus on Titan’s design and engineering, and in particular, the use of carbon fiber in the sub’s pressure section. As an
MIT educated aerospace engineer with a working knowledge of composites, I initially saw the use of at least some carbon fiber in my company’s
submersible as a no-brainer. I knew the submersible community frowned upon it, but it was at first easy to discard that perspective as neanderthal thinking.
Commercial aviation was initially hesitant to adopt composite construction in airliners despite composites successfully flying in military aircraft.
Where OceanGate had aspired to great depth, my company had a less ambitious aspiration of speed. I
suspect our submersible still holds the unofficial world record for fastest personal submersible in the
world. Unlike OceanGate, my company’s business model required Classification. My recollection is that
Classification of a submersible employing composite construction was not technically forbidden,
however, for all practical purposes, it would not have been possible. It was an easy business decision
for us. After building our proof-of-concept sub from composites, we pivoted our prototype design to
employ an aluminum pressure section, with a composite fairing for speed. The OceanGate business
model did not rely on Classification. In my view, Rush was appropriate in considering the use of
carbon fiber. But there are at least two good reasons why he should have abandoned that decision.
For the curious, as I haven’t seen any good explanations, I’m going to offer the following to lay
persons: Implosion failures are a form of buckling, but buckling is, I believe the technical term is
‘weird.’ Those interested can do the following experiment. Grab some dry pieces of spaghetti. Take one
between the palms of your hands, and press your palms together. You’ll note that it takes almost no
resistance. You can keep on pressing with very little force until the spaghetti makes a big turning arch,
and eventually shatters. Next, take the a piece of spaghetti and hold it firmly with your fingers between both hands. Press the ends together in a straight line
and don’t let the ends rotate at all. You’ll find it takes far more force to shatter the spaghetti. Finally, take a third piece of spaghetti, and once again lock the end
in place with your fingers. This time, push it together, but rotate the ends by about 10 degrees. You’ll see the strength of this piece is far less than the piece that
was held straight (but still stronger than the piece whose ends were not constrained).
If we compare carbon fiber to metal, it is lighter, stronger and stiffer in tension. That makes it great for aviation. Unlike submersibles, aircraft are exposed to
low pressure externally, which puts the fuselage structure in tension. The external high pressure that submersibles are exposed to places their pressure
section in compression. In compression, the fibers behave similarly to the pasta. But real-world designs have to survive forces in multiple directions. In the
middle of a cylinder under pressure, the compressive forces occur both around the circumference from water pressure as well as longitudinally, from the end
caps pressing into the straight section. So how do you maintain strength? The solution is to be thoughtful in how the fibers are oriented. It takes some high-
level engineering to explicitly design for a carbon fiber pressure vessel subject to external pressure, but it can be done. Stockton Rush understood this.
Just as long as the fibers stay in place, it’s pretty strong. But if the resin were to let just a little section of carbon pop out of place, most of the fiber’s resistance
to buckling disappears. This is where it gets scary for submersibles, because water, particularly flowing water or water under pressure is an incredible force to
be reckoned with. Water’s small molecules can work their way into nearly any crevice, and then the water’s pressure will eventually begin to enlarge that
crevice. Unfortunately, composites are difficult to fabricate fully absent of small crevices (generally referred to as voids) held within the resin. This is a
problem that all adhesives face when submerged. You can seal your construction, but sealants all contain adhesives, so that doesn’t fix your problem. There
are always workarounds in engineering, and OceanGate did put certain workarounds in place, but
when lives are at stake, the reliability of such workarounds must be rigorously proven. The initial
indications suggest that sufficient design testing may have never occurred.
So would this be primarily an engineering failure? Many or most would say yes, but I would
disagree. Engineers make mistakes every day, just like everyone else. If you look hard enough, there
is an engineering solution to almost any problem. But like so many things, engineering is a tool,
and tools have to be used properly or they don’t work.
Where then, do we lay blame? Again, we need to let the investigation decide, but at first blush, this
appears to be a problem of management—particularly, the management of risk. In the midst of
sensationalist discoveries, there are clues. The choice of a game-controller to control the sub’s
thrusters might sound irresponsible, but the fact is that those ‘toys’ are incredibly resilient and
reliable. Not quite the level of an industrial rated joystick, but those are heavy and bulky. The game
controller would have been easy to hold in the cramped quarters, and even passed around for
occupants to ‘play with.’ It’s questionable that an equivalent industrial rated control device could
have been built without spending a small fortune. And even if the thing failed entirely, the risk
would have been minimal. Similarly, many large and reputable companies hire interns without
degrees to do simple engineering. Obviously, the work must be checked by qualified engineers, but there’s nothing inherently wrong with this practice.
And hiring a bunch of young staff is great…and inspirational, as Mr. Rush desired. But here’s where things start getting dicey. Because it doesn’t take an entire
company of “50 year-old white men” to be successful, but when you’re responsible for human lives, you have to have a team that includes fully qualified
individuals, no matter their age, race, or gender. Well run teams of experienced, qualified people are far less likely to make a costly mistake than any
individual, no matter how gifted that individual may be. Again at first blush, Mr. Rush seems to have surrounded himself with young, inexperienced, and
unqualified technicians that found it difficult to stand up to him – ‘yes-men,’ if you will. Where he did receive criticism, he appears to have done everything in
his power to silence opposing voices.
Experienced business people can feel the pain of this in their gut. Good leaders avoid working with ‘yes-men/women.’ Aviation had to learn this lesson the
hard way. It used to be that the captain of an airliner rang supreme. His (because they were all men then) word was the only word. Multiple accidents
revealed situations where other crew members KNEW the captain was making a critical mistake, but they felt powerless to correct mistakes which would
cost many lives. Initially in response, crews were trained that anyone who sees a problem needs to speak up, and the captain needs to pay attention. But
then human factors and adverse culture revealed that inexperienced crewmembers are hesitant to speak up, and even when they do, some captains
remained hesitant to listen. Only after years of working to change the culture of flight decks have airline crews become as effective as they are today.
Personal experience has taught me that one of the greatest challenges as a captain is ensuring that when we fly with somebody new, those newbies don’t
look at the gray-haired aviator to their left and think, just because they’re funny, charismatic, and probably know way more, that he or she isn’t screwing
up.
Bad engineering might have been a symptom of Titan’s failure, but initial indications are that the real problem was probably the company’s culture,
structure, and leadership style. And this is where it’s critical for all of us to pay attention. Because there are plenty of companies that make life and death
decisions. Take your iPhone. Steve Jobs liked to break rules, too. But he hired fully
qualified engineers so that if you get in a car accident and need to call for help, even if
your phone got banged around, it probably still works. As I taught my kids, if you have to
break a rule, you need to know exactly why that rule exists, and you have to be prepared
for the consequences which often catch up to you.
For those curious, even though I had meaningful interest in our sub, I couldn’t raise the
capital needed to get through the Classification process, and I refused to take a short-cut.
Sometimes the best decision you can make is to retreat. Avoid making the big mistakes,
and survive to fight again.
Learning the right lessons from Titan
The one that got away -- CGI of the Classed submersible we
couldn’t afford to build.