The Questionable Engineering of Oceangate

On June 19, the submersible Oceangate began its descent toward the fractured remains of the Titanic here in the northwest Atlantic. With a descent speed of 55 meters per minute, the submersible would take 70 minutes to reach its destination. With each passing minute, the equivalent of 5.3 atmospheres of pressure would be added to the submarine's hull. When it reaches the desired depth, the weight of 366 Earth atmospheres will try to crush it. Applied to just an area slightly larger than a sheet of A3 paper, that's equivalent to the thrust of a Saturn V F1 engine. Pressing from all directions. One hour and 45 minutes of immersion.

In the deep sea. 3800 meters deep. With this immense pressure putting pressure on her hoof. The Oceangate submersible lost contact with the surface. We now know that the submersible imploded, a sudden and catastrophic collapse. It was a tragic accident, but any composites engineer could have predicted this failure, and they did. I worked as a composite design engineer and completed my master's thesis in composite failure prediction. So let's delve into the

questionable

engineering

decisions Oceangate made that led to a tragic accident. That pressure comes with some extreme design challenges. Oceangate's Titan submersible consists of two titanium caps, one with a large acrylic window, attached to a filament-wound carbon fiber-reinforced epoxy hull.

The interior of the submersible is quite basic. Screens that are fed with data from outside cameras. The four electric thrusters are controlled, as far as I can tell, with a third-party Gamecube controller. A lot of people are focusing on this part of the design, but that's the least of my worries. We will focus on one aspect of the design of this submersible, the composite hull. Carbon fiber composites are not really known for their compressive strength. They work best in tension. Ideal for aircraft pressurized from the inside. Where the pressure inside the fuselage works to expand the circular cross section, putting the fibers in tension.

For a submarine, the pressure will work to compress the hull, placing the fibers primarily under compression. This immediately set alarm bells ringing in my mind when I heard about the missing submarine, so I began digging into his justification for using the material. Oceangate chose this carbon fiber composite primarily to benefit from its natural buoyancy. In reality, these types of submarines want to have as neutral buoyancy as possible. Which means that the weight of the water they displace weighs approximately the same as the vehicle itself. Minimizing the energy needed to rise or sink in the water. This means that to surface in an emergency they do not need power, they simply drop ballasts that make them positively buoyant and they simply float to the surface.

Carbon fiber composite can be lighter and therefore helps achieve this goal of neutral buoyancy. Submarine hulls are typically made of steel or, more recently, titanium, but to achieve the desired buoyancy they require outer layers of foam. Stockton Rush, CEO of OceanGates, who was inside the submersible, stated that they wanted to skip this layer of foam because it increased costs. This is not necessarily a bad thing if the compound is suitable for the job. This is where things get iffy, Oceangate had no idea if they were up to the task or not, and we know this because they admit in their own blog post that they justify their decision not to test their vehicle with a regulatory body. .

The now-deleted blog post said: Most major shipping operators require chartered vessels to be “classified” by an independent group Classing assures shipowners, insurers and regulators that vessels are designed, built and inspected according to accepted standards. Rating can be effective in filtering out poor designers and builders, but established standards do little to weed out poor ship operators. The vast majority of marine (and aviation) accidents are the result of operator error, not mechanical failure. As a result, simply focusing on classifying the vessel does not address operational risks. Now, I would have thought this wouldn't need to be pointed out, but the fact that this blatant red flag was written up on its own website for all to see, and several extremely wealthy and successful people didn't seem to think it was a problem. seems to suggest the opposite.

Perhaps most submarine failures are due to operator error precisely because they have undergone rigorous testing and certification. The comparison with aviation is also just a can of worms with which we have to compare very recent catastrophic accidents, where manufacturers intentionally circumvented regulatory procedures. These are not procedures created as some kind of unnecessary inconvenience to inventors. These are very useful testing guidelines and procedures to demonstrate that no catastrophic failure mode has been overlooked in your design, and that is exactly what happened. One failure mode is unique to composite materials in deepwater applications. Some of the first tests we got were fiberglass-reinforced deepwater pipes.

It is called instantaneous buckling failure mode or coupled delamination buckling. For thin-walled pressure vessels, the failure mode is simply buckling, where the pressure simply collapses the entire wall, but here, because the wall thickness is so great to cope with the immense pressure from the depths from the sea, the failure mode becomes more complicated. This failure mode is characterized by delamination of the inner layer of the pressure vessel; Basically, the interior of the pressure vessel suddenly detaches from the rest of the wall, leading to catastrophic failure of the overall structure. The exact mechanisms of the failure mode are not well understood, as noted in this 2022 research paper titled “A Review on Structural Failure of Deepwater Composite Pressure Hulls.” In his conclusion he states that “research on this topic has remained at the theoretical level. “There is still very little research in this area in recent years.” So I'm guessing whoever designed this pressure vessel did at least the same level of research as a YouTuber who wrote a script on the subject in two days.

They should have at least been aware of this problem. A failure mode like this will be exacerbated by gradual damage caused by cyclic stresses, such as those experienced when diving to depths of 4,000 meters and resurfacing. They have done it at least 50 times with this boat, without ever testing it due to fatigue. By their own admission, they didn't subject themselves to industry-standard ratings, but they did do their own rating test. A proof, stated in that blog post once again. “A licensed marine inspector will witness a successful dive to 4,000 meters, inspect the vessel before and after the dive, and provide a statement of facts attesting to the completion of the dive test plan.” That's simply not enough.

Anything that undergoes repeated stress cycles like this should be fatigue tested. Like test airplane manufacturers do with wings, where they flex the airplane's wings thousands of times on the ground. They designed this essentially with computer models without real-world testing. Composite materials are incredibly difficult to model in computer software. Composite materials are much more complicated than homogeneous materials like steel, where the material properties are the same in all directions and in all regions of the material. Composites are made up of fibers held together by a matrix material such as epoxy. If we zoomed in on a small cross section of this, we might see something like this.

Some fibers, some could be touching without any matrix between them, others could be spaced further apart and there may also be empty gaps that the epoxy failed to fill. We can simulate a small section like this independently to help analyze some failure modes, such as delamination. However, if we want to zoom out and test a larger section, we often need to average the material properties in some way or else the simulation will take too long to run. My own master's thesis attempted to combine the two by incorporating a fiber and a cellular matrix within an average composite material.

There are specialized softwares that can model different fiber orientations, but they are still largely an estimate. You must back this up with physical evidence. Something OceanGate didn't do. There are more signs of incompetent use of composites. For a critical structure like this, I would highly recommend curing it in an autoclave. It's basically a pressurized oven that helps get those holes out of the epoxy that I mentioned before. Autoclaves are expensive and it can be difficult to find ones big enough to fit something like this, but it's nice not to have air bubbles in a pressure vessel keeping you alive.

However, it was not used for this pressure vessel. It was simply bagged and cured, a cheaper alternative. Spencer Composites was commissioned to manufacture the composite hull and its CEO was interviewed in a 2017 compositesworld article. He seems well aware of the immense design challenge presented to him and mentions that the porosity of the cured pressure vessel was assessed to be less than <1%. “Less than 1% voids” doesn't exactly inspire confidence. Is it 0.99% or 0.0001% porosity? That is exactly the type of defect that will increase the risk of rapid buckling, since gaps can be sites of delamination propagation. The blog post goes on to dismiss classification tests as inadequate for innovative designs, which is complete narcissistic nonsense.

This is not innovative. They just made the pressure vessel from a carbon fiber composite, using known manufacturing practices, and started charging rich people $250,000 to risk their lives on it. The innovation would be to create that pressure vessel and test it. Discover the causes of rapid buckling and advance materials science for all humanity. This is speculation, not innovation. To remove all doubt that this company has intentionally ignored standard security practices. This 2018 court case was brought against Oceangate by a former employee who was fired shortly after raising concerns about the composite hull. Indicate that not enough was done to check for delaminations and gaps.

Oceangate argued that the technology did not exist to check for voids and delaminations in a composite of this thickness, and instead they would use their acoustic monitoring system, which essentially listens to the sounds of fracture to predict failures. But this is what a fracture curve for a carbon fiber composite looks like. It is the edge of a cliff. Sudden catastrophic collapse. An acoustic monitoring system like this is similar to setting up a camera to warn you that thunder is approaching. You will see the lightning before the thunder, but the time between them is minimal. Pure security theater.

It gets even worse. On a test dive in the Bahamas, a submersible expert could hear the cracking sound with his own ears and warned Stockton Rush that it was a sign that damage was accumulating. Stockton, once again, chose to ignore expert advice. So this acoustic monitoring system was solely for marketing purposes. To give customers, who don't have the

engineering

expertise to see it, a false sense of security. I would attribute this to malice and greed, but Stockton Rush decided to pilot this death trap himself. As Hanlon's Razor says: never attribute to malice what can be attributed to stupidity.

There is a growing trend in engineering to move fast and break things. Mainly driven by an overconfident CEO who, to be fair to him, managed to revolutionize an entire industry in that spirit. That damning blog post even mentions SpaceX, Blue Origin, and Virgin Galactic as bastions of innovation for this move fast and break things mantra. Virgin Galactic also has blood on its hands, with its in-flight breakup in 2014 blamed on inadequate design safeguards and poor pilot training, along with a lack of oversight by regulatory authorities. This attitude is dangerous and this will not be the last story of rich, narcissistic CEOs putting lives at risk.

When a doctor is negligent with the safety of his patients, he is liable for negligence, but the impact of his incompetence affects a patient; When an engineer neglects his responsibilities, he can affect hundreds, thousands and even millions of people. We need more people like David Lockridge, the employee who was fired for raising concerns about safety. The first video on this channel declared ourmission. To inspire the next generation of engineers, but it's more than just inspiring engineers, we need engineers working in the most impactful industries. Solve the biggest problems facing our civilization. You've probably heard a lot of career advice: "follow your passion", "do what you love", "take initiative", etc., etc.

But a lot of the advice out there doesn't take the best into account. The available evidence we have about what really makes a career fulfilling and high impact. Today's sponsor, 80,000 Hours, is a nonprofit organization that, for the past ten years, has been researching the question of how to find a fulfilling career that also does good. Everything they publish is based on careful consideration of data, consultation with experts, and taking into account the best evidence they could find on the topic. And because they take this evidence-based approach, they can make concrete recommendations on tough questions like: Where should I work if I really want to make a positive difference in the world?

What are the most pressing problems facing humanity today and how could we solve them? What can I do today to start planning a fulfilling career? They have published over 30 articles on an incredibly pressing topic: security in AI development. They have a website with tons of compiled research on high-impact careers, a job board that lists currently open positions that they think could help you make a positive impact, and a podcast where they have conversations with experts about the world's most pressing issues . Everything they offer is free, forever, because they are a non-profit organization - their sole goal is to help solve global problems by helping people find the most impactful careers they can.

If you join the newsletter now, you'll receive a free copy of your in-depth research-based career guide delivered to your inbox; simply register at 80000hours.org/realengineering or click the button on the screen now.