TL;DR, this was a triumph of stupidity and greed over good engineering.
Engineering design standards are often written in blood, in the sense that these standards often came into being after accidents involving serious loss of life. The CEO of Oceangate ignored them, apparently because they got in his way of saving costs. As it turns out, reality is not susceptible to CEO bullshit.
Unsurprisingly, “move fast and break things” is not a good strategy when lives are at stake.
Although it doesn’t show much on this blog, I am a professional mechanical engineer who has been designing and manufacturing products from composite materials for close to 30 years.
When the Titan went missing and I had reason to look into the details of its construction, there were three main reasons of concern based on my experience;
- a window that was only rated for 1300 m depth
- the wound carbon/epoxy pressure hull
- titanium bonded to carbon/epoxy in a saltwater environment
To be sure there are many more reasons for concern like life support, electronics, external cabling, ballast release. But those are outside of my expertise.
Depth rating of the window
When I was in training, a safety factor of 3 was usual for general engineering, and a safety factor of 10 for “heavy” engineering like cranes. The design loads are multiplied by this safety factor to account for things like production variances, unknowns in the real loads and general wear and tear during the service life of a product. And then the product is designed not to fail under this multiplied load.
Taking a window that is rated for 1300 m dive depth to three times its rated depth is at best plainly irresponsible.
It is well known that fiber reinforced composites are much less strong in compression than in tension.
The underlying reason is that while you can pull on a thread, you cannot push on it!
Under compression the epoxy matrix, which is much less strong than the fibers, is what holds the material together. How well that goes also depends on a lot of processing related properties. Things like void content, fiber alignment, fiber damage, degree of cure of the resin all play a role. In the end, compression strength of composites is more of a stability (buckling) issue that a material properties issue. For a carbon/epoxy unidirectional material my rule of thumb is that the pure compression strength of such a material will be only about 60% of the tensile strength. For fiberglass it is closer to 50%.
In contrast to metals that generally exhibit plastic deformation before failing, composite materials exhibit sudden brittle failure. And while you can sometimes hear noises when a composite is about to fail, there is generally not enough time to do anything about it. In the tests where I have made video’s from, complete failure occurred within 1/100 of a second for sure, and probably even faster.
For these reasons, I would say that carbon/epoxy is generally not suited for pressure vessels that have to withstand outside pressure.
Titanium bonded to carbon/epoxy in a saltwater environment
Bonding dissimilar materials can have a range of nasty effects. But what I want to focus on here is only one of them. Metals that come into contact with carbon fibers in a wet environment tend to exhibit galvanic corrosion. I’ve seen wet carbon fabric leave a visible etched surface on an aluminium mold after about 30 minutes.
So when metal has to be bonded to carbon/epoxy, we generally make sure that there is a non-conductive layer between the composite and the metal. Sometimes this is just the adhesive in case the layer is thick enough. But if needed a separate layer of fiberglass or unfilled epoxy is placed between the carbon/epoxy and the adhesive.
In the video clips that I have seen, only a very thin layer of adhesive was applied between the titanium and composite part. From the color it was clear that the adhesive has filler in it. But it is not clear what it was filled with. Hopefully something non-conductive. If not, it is another accident waiting to happen for a vehicle that already has too many of them.
For comments, please send me an e-mail.