I watched a nice video where the team from Blue Origin (Jeff Bezos’ hobby and side business) explains how they designed their series of prototypes.
It is a great example of validation of a totally new product that can have a catastrophic safety failure.
You can watch it here:
Some interesting takeaways:
- It is pretty clear that people can die if this product malfunctions. And, if there is one accident that results in deaths or near deaths, it will make their future customers hesitant. Safety is their top objective.
- They have had their own take on the way to design a way to send people to space. It was possible with the 1960s technology, but they wanted to do it their own way. I am pretty certain they estimated (from the start) that it would take a lot of time and investment.
- They designed a ‘flight test program’ (in other words, a testing plan), to have sufficient data to make the right decision. Each prototype function test is a rocket launch, so it is quite expensive.
- They stuck to the plan. They did not put people in their rocket in the first prototypes. Even though their first prototype test was “100% successful”, it was still far from sufficient evidence.
- They tried to learn by testing different things on each flight, including different environmental conditions.
- At one point they also started to test for failures of some components (parachutes that don’t open, something wrong on the propulsion module), to ensure it did not lead to a catastrophic failure.
- As they kept improving the product (e.g. the control systems for the capsule), they had to do more launches to validate the whole product design.
- Where user safety can be most impacted (the flight capsule), they made systems redundant, sometimes with a backup of a backup system!
In summary, that’s great product engineering, but it’s also great reliability engineering.
Some related reading: