We explore two high-profile product reliability problems that have been in the news recently: the destruction of the Oceangate Titan submersible that catastrophically imploded near the wreck of the Titanic and the reliability failures of the Siemens Gamesa wind turbines that suffered from vibration landing them with a $2 billion price tag to fix and denting their stock prices. What caused them, why did they happen, and what could the companies have done to avoid these unfortunate cases? Find out here.

Prefer listening to reading?

Listen to the audio here.

1. The destruction of the Titan submersible visiting the wreck of the Titanic.

The Titan submersible was used to ferry paying customers to the wreck site of the Titanic and it went down with all 5 people aboard dead on June 18th 2023. But why did it implode?

The New York Times reported on the ‘maverick design choices that may have doomed Titan:’

The best way of outwitting the many dangers, the experts agreed, would have been to subject Titan to rigorous testing under the anticipated conditions and stresses. Fatigue of various materials also would have had to be considered and continually monitored. Manufacturing defects or any damage could build up over time as the Titan endured the cycles of stress associated with repeatedly going down miles to the bottom of the Atlantic and back.

Most deep-sea craft undergo costly rounds of inspection and testing by reputable marine organizations that specialize in certifying the deep-diving craft as safe. But Mr. Rush (the Oceangate CEO) obtained no certification for Titan, saying it stifled innovation. In a documentary, he said: “You are remembered for the rules you break, and I’ve broken some rules to make this. The carbon fiber and titanium — there’s a rule you don’t do that. Well, I did.”

Most submersibles exclusively use Titanium for their living compartment, but by using a mixture of carbon fiber and Titanium, the Oceangate submersible was able to accommodate 2 more (paying) passengers than others. It did this by having a pill-shaped design with a carbon fiber body bonded to Titanium bands and a front section with a large porthole. Others use a full Titanium spherical design which can only fit 3 people. The issue here may well have been that two different materials were mixed that behave differently when compressed by great forces experienced miles below the ocean. Over time on previous dives the carbon fiber repeatedly compressing even slightly more than the stronger Titanium would have weakened the bonded joints, eventually leading to the catastrophic results of this fateful dive.

Not doing enough testing and going through recognized certifications to oceanic standards where the material choices, for example, would have been scrutinized, was a huge mistake and was done to save on time and costs. But ANY product that is not put through reliability testing runs enormous risks of breaking down when in use, and in this case that proved to be fatal as at that depth every square inch of the sub’s surface is subjected to a pressure of 3 tons.

In fact, there is a list of issues that suggest ‘low reliability’ and here are just a few:

• They used a commercial wireless video game controller to control the sub – these are made (and tested) to be used in a living room, not in a pressurized sub cabin and wireless switches are more prone to connection problems than a wired switch, for example.
• It used circuit boards that were not engineered to be used in a corrosive saltwater environment and some water may have gotten into the sub each time it was opened and closed before and after dives.
• The master power control was not even on the main power bus.
• Substandard underwater guidance via GPS which is not reliable underwater.
• No location beacon.
• Using a porthole rated for far less depth than the sub was used at.

These design issues suggest that the sub was designed on a budget and only one was built and used, rather than two or three which each would have been put through a relevant testing and certification process. Of course, the latter would mean the cost of several subs and testing to the business which is a much greater cost than just building and using one. Ethically is it even OK to send passengers down to those depths in a submersible that is not certified for use in that high-pressure environment and is arguably far less scrutinized for safety than, say, even rollercoasters at a theme park? Arguably not.

A standard engineering process is to do reliability testing on the product and component level reliability testing, too. If Oceangate had built a semi-functional prototype and tested it to destruction they could have put the learnings into just one final sub because they would have found and fixed the problems that led to the implosion. Costs would have been higher, for sure, but do they compare with the cost of the sub’s failure? (02:23)

Do strict standards actually stifle innovation and make it hard for new companies to enter a market?

This is somewhat true as standards often have guidelines that are more general and don’t suit specific products, therefore it makes it hard to bring out a new product that exactly complies with them. But they do at least provide some direction as to how the product should be designed and tested for safety and reliability. An innovative manufacturer may instead create their own reliability standard that is based on the existing standard but is far more relevant to their own product and so being constrained by an existing standard is not really an excuse not to innovate. Oceangate could have produced an unmanned scale model to test and perhaps used monitors to assess if the joints between carbon fiber and Titanium were remaining sound, and this could have been in their standard.

Also, SpaceX has been able to innovate in the field of space travel even though they had to test to military and NASA standards. They found ways to use more common materials and components, but have sent up unmanned launches when assessing safety and have also gone through the required testing. Something that Oceangate seemingly failed to do. (22:14)

2. Vibrating Siemens Gamesa wind turbines.

Siemens Gamesa produces wind turbines for renewable energy. A wide-ranging issue occurred with a lot of their onshore turbines, according to Reuters, specifically with the:

108 gigawatt (GW) onshore turbine fleet, where the group has discovered quality issues in certain components, including rotor blades and bearings.

Siemens Energy said that 15%-30% of the fleet could be affected by the problems, which were exposed during a review which noted “abnormal vibration behaviour of some components” and unspecified problems around product design.

Reliability issues with one or two products from a fleet of probably thousands are to be expected, but 15-30% both damaged Siemens’ stock prices and it casts doubt on the entire renewable industry that is scrambling to quickly transition away from fossil fuels…perhaps trying to move too quickly.

This could be a case where a large corporation has cut corners during product development and is likely caused by design defects that could and should have been caught by appropriate reliability testing on components and the turbines themselves. The result is a cost to Siemens of around US$ 2 billion to correct the issues instead of a cost-saving.

The product itself is so huge that it may have been found too difficult to test properly during development. Siemens may have gotten around this issue by hiring military-grade reliability engineers who are used to theoretically designing large devices like airplanes to fulfill specific reliability goals without having physical samples. Their theoretical design would be tested to withstand the environment, lightning strikes, etc, and would be close to the final physical design. Then a physical product would be made, put through a wind tunnel, thermal cycling, etc, in order to thoroughly analyze it. (27:46)

Comparing possible testing solutions for the Titan sub and the wind turbines.

The Titan sub could have been tested correctly by building a scale prototype and sending it down to the correct depths, monitoring its results, and using those results to improve the final full-size sub. In the case of the turbines, the solution would be far more theoretical as the turbines are so huge it is hard to build and test a lot of physical prototypes so computer modelling would be used instead and perhaps individual components would be tested as with airplanes rather than trying to test a finished turbine. Siemens may have done modelling and component-level testing, but it may have been one of their suppliers who didn’t do adequate testing, for example. (34:52)

Summary: What it takes to manufacture and keep manufacturing a reliable product.

To make a reliable product you need a reliable design, manufacturing controls, and regular ongoing reliability testing in place to ensure that continuing products keep meeting reliability objectives. If these things were done right, the two cases we’ve covered probably wouldn’t have occurred. (39:52)

Related content…

• • The Oceangate Titan submersible
  • The Siemens Gamesa wind turbines
  • Get help from Sofeast to do Reliability Engineering & Testing (in a China lab) – we provide reliability engineering services including reliability testing on component and PCB levels as well as the product level on various types of consumer products. Some of these tests include drop, vibration, temperature and humidity and package testing. We can customize the right tests for your product.
  • How Reliability Testing Is Critical To Obtaining Great Mass-Produced Products
  • The Design for Reliability Process for Launching Reliable Products

Editor’s note: images courtesy of OceanGate, via Getty Images and REUTERS/Dado Ruvic/Illustration.