tl;dr The development processes of SLS vs. New Shepard and specifically Falcon 9 are vastly different.
I don't understand how this is possible. How can a hyper-complex rocket, that has literally never flown before, be considered safe for human spaceflight after just a single test flight?
This is the first test flight in this exact configuration, but it is by far not the first test.
- All of the main engines have flown before on multiple Space Shuttle missions.
- All of the main engines have been static fired multiple times.
- The whole core stage has had a full mission-duration "green run" static fire test.
- The second stage is a slightly modified Delta IV upper stage, which has been extensively tested.
- The European Service Module is derived from the ATV vehicle which has flown many times.
- The ESM engine is a reused Space Shuttle Orbital Maneuvering System engine that has already been to space multiple times.
- Orion has already had a test flight on an Atlas V as well as landing tests and abort tests.
- The Solid Rocket Boosters are 5-segment variants of the 4-segment Space Shuttle SRBs. The 4-segment ones have been to space many times. The 5-segment boosters were extensively tested for OmegA.
- Obviously, you cannot static fire an SRB and then use it, but you can static fire identical SRBs, and this has been done multiple times.
- The Launch Abort System is closely derived from Apollo.
For example, Falcon 9 flew 84 times before carrying people for the first time in Demo-2.
This is a misleading way to count. Falcon 9 has had multiple extensive design changes over the course of its lifetime: Falcon 9 1.0, Falcon 9 1.1, Falcon 9 1.2 aka Falcon 9 Full Thrust, which itself had 5 different block upgrades.
What is currently flying is Block 5 of Falcon 9 FT, commonly known simply as Falcon 9 Block 5. It is, in total, the 7th version of Falcon 9.
But even within Block 5, SpaceX has made constant upgrades.
In fact, NASA required only a small number of flights without changes to crew-rate Falcon 9. If my memory is correct, NASA required nine flights in order to certify Falcon 9 for crew.
New Shepard flew 15 times before its first crewed launch. Frankly, it seems crazy to declare that you're satisfied with the safety parameters of a system so complex after a single test, even if that test goes well.
And after how many tests do you declare that you are satisfied with the safety parameters? Based on you citing New Shepard as an example, would you think 15 times is enough? Before you answer, keep in mind that it exploded on the 23rd flight!
The production cadence for SLS is one rocket per year. The contractors have said, they could push for 2 rockets per year, but that is the absolute maximum and cannot be sustained indefinitely.
So, if you were to require 15 flights of SLS before crew-rating it, that would put the first crewed mission at no earlier than 2037.
How does NASA adequately manage so many unknowns with just a single test, while the competition seems to think that way more tests are required before you're ready to put people at risk?
The development processes of SpaceX and practically every other rocket manufacturer on the planet (except maybe a few of the smaller newSPACE ones) are vastly different.
NASA and the "old space" contractors use a very traditional engineering process, where everything is specified in as much details as possible, the specifications get reviewed over and over, subsystem are designed, the designs are reviewed over and over, the interfaces between subsystems are designed, the designs are reviewed over and over, and so on and so forth.
Typically, you have reviewed and discussed and specified all aspects of the system so extensively that you are typically confident that everything will work, even before you build the first piece of hardware.
Additionally, all the contractors have extensive experience with such projects (many of the contractors involved in Artemis were already involved with Space Shuttle, ISS, or Apollo) and a lot of the hardware and systems have design heritage or even flight heritage.
Most importantly: this is how it was always done and thus everybody at NASA knows exactly how to evaluate the risks.
Note that the Space Transportation System (STS) (colloquially known as Space Shuttle) had zero flights before the first crewed flight, since it was incapable of flying without crew. And neither of the two tragedies were caused by a lack of testing, in fact, both of the problems (the O-rings becoming brittle at low temperatures and the heat shield being susceptible to FOD damage from the ET) were well known, simply based on design reviews.
The problems were found before they occurred, simply by thinking through the problem. The accidents happened because the problems were deliberately ignored, not because of a lack of testing.
Whereas SpaceX's approach is more like:
- Guess what will work.
- Build it.
- Launch it.
- If it works, go to step #1 and guess what will improve it. If it doesn't work, go to step #1 and guess what will fix it.
- Repeat for every single launch and every single vehicle.
One of the problems with crew-rating Falcon 9 was simply that NASA didn't know how to even begin evaluating such a process since they had never encountered such a process before.
Note that both NASA as well as the US military in the beginning required that SpaceX use only brand-new boosters. Then, they relaxed the requirements and said that SpaceX can re-use the boosters, but there must be specific boosters set aside only for NASA or only for DoD. But now, they have no problems with re-used boosters, even boosters that have flown for other customers in between.
Similar for Crew Dragon. In the beginning, NASA didn't want them re-used. Then NASA certified them for 2 flights, if I remember correctly. Now, they are certified for 3 or 4. Whereas SpaceX has always certified them for slightly more flights, and SpaceX launched Inspiration 4 on a re-used booster then NASA was still not allowing it.