You're Gonna Kill Astronauts - Reliability and Crew Safety at NASA
Вставка
- Опубліковано 29 тра 2024
- NASA is viewed as the go to organization for anything to do with flying humans? Is this reputation earned? What is the history of NASA in this area?
@Eric_Gunnerson on Twitter
Triabolical_ on Reddit - Наука та технологія
11:28 “and most people are human” under rated joke
I always hope when I put those in that somebody will notice. Thanks for the note.
Holy shit I am so happy I found your channel. Crazy how high quality your stuff is
Thanks.
5:50 i know this is just a quick comparison to get the concept accross, but the cause of the concorde crash was lacking maintanance quality on a different airplane which took off just before the concorde. it is my understanding that any plane that used the runway right after that would have had an increased risk of an incident
This is a really interesting question, and one I didn't think I really had time to dive deeply into.
The concorde crash was definitely caused by FOD on the runway. The open question is whether the Concorde design is complicit in the FOD leading to a crash that killed everybody on board, or whether it was just a freak accident that could have happened to any airplane.
I'm not an expert on FOD damage to airplanes, but my understanding is that while it's not uncommon to see FOD damage to engines, see meaningful damage elsewhere on the airplane is pretty rare.
Is it possible that concorde just got unlucky? Sure
But given that we don't see that sort of failure across the hundreds of millions of flights done by other types of planes suggests that there's something about the design that made it more likely that a FOD issue would lead to a fatal fuel tank puncture.
@@EagerSpace The thing is that safety analysis needs to be performed on the entire operation, and not just the vehicle. NASA's two LOCs on Shuttle were of course caused by vehicle deficiencies, but it required the complicity of the rest of the operation. Namely, yes, SRBs are by definition death traps, and the seal design on the SRBs was awful. BUT Challenger wouldn't have happened had NASA actually listened to the engineers and paid proper attention to temperature. Had they done the responsible thing and not launch when the vehicle had been exposed to temperatures outside its operational range the night before, Challenger wouldn't have happened. Columbia is harder, but it shouldn't been an Apollo 13 situation, not a Challenger situation. NASA decided to roll the dice and tell them to reenter, instead of informing astronauts of the failure. There were a lot of things that could be done to save Columbia. So, both failures were due to intrinsic issues in the vehicle, but also required borderline criminal neglect on the part of management.
Concorde was pretty much the opposite. There were a million things that could go wrong in that vehicle, but it was operated far less than other aircraft, it always flew the same route between major airports, it was always operated by a few crews that were highly trained in that aircraft, and were already the most experienced crews, only the top senior pilots ever got to get type rated for Concorde. Airliners fly constantly, and maintenance and inspection windows are short. The Concorde, on the other hand, got more hangar hours than flight hours, and more maintenance hours than hours in the air.
We see the same kind of discrepancy in the A380. So many flights without a single fatality. Is it because it's the safest plane flying? It's certainly a safe airplane, but there are other things at play. It's only operated by top-notch airlines, that mostly use it to sell quite expensive tickets, and so the plane gets special treatment. It can't land at just any airport, so it only goes to the best airports in the world, with wide and long runways and top safety. It gets flown by the best crews, maintained by the best mechanics, and always special attention is paid to it.
Same as ULA's reliability. The Delta IV Heavy is a horrible design, with a lot of things that are high risk and could go wrong at any time. But it has a very low cadence, an insane amount of hours go into building it and an even more insane into QA, and every actual launch often takes a bunch of scrubs. It also always flies quite important payloads. How much of that reliability is due to the vehicle itself, and how much due to ULA having a heavy bias towards "scrubs are cheaper than RUDs"?
@@almafuertegmailcom Lots of topics there and I can't really do justice to them in a reply, so I'm going to be lazy and pick out a few points. I'll note at the start that I'm l largely in agreement with you.
>BUT Challenger wouldn't have happened had NASA actually listened to the engineers and paid proper attention to temperature.
Launching that day was clearly a very bad decision, but shuttle was headed for a serious SRB problem at some point; they were seeing regular O-ring erosion and Thiokol believed they did not have two-point redundancy and was advocating with NASA to do a redesign. Do they figure out how to reduce erosion without an LOC and make it through the program? Maybe, but I don't think it was promising; I think they were going to get unlucky and get a burn through eventually given the culture at the time. It's really the same issues as Columbia - you have a system that isn't meeting design parameters but you choose to keep flying it.
Concorde, I don't feel strongly about, but I do find it exceptional that a plane with so few flights just happens to have such a bad reaction to FOD when there just isn't history of that kind of issue across so many other aircraft - at least, not history I could find. It could be they just got really unlucky, or it could be that they hit a situation that other designs would of shrugged off but their airframe could not.
WRT the A380 and Delta IV, the sample size is simply far too small to really come to any conclusion about their reliability. There's a whole branch of analysis that looks at what you can conclude - well, "guess" is probably a better word - about reliability from small sample sizes, but I don't think it's very illuminating.
@@EagerSpace
If you want a good video breaking down the Concorde crash, I highly recommend Mentor Pilot's video on the subject.
ua-cam.com/video/C-nALYF73hU/v-deo.html
Basically, known design issues did meaningfully contribute to the crash.
For some reason, I thought that the title to this video said “How to safely unalyve astronauts”
good video with your usual dry wit. thanks
Great video, just found your channel today. I am definitely going to watch more.
I'm beginning to see why NASA is slow to experiment with new tech. The collected data from existing space systems is handy to access probabilities in future craft composed of modifications of existing components. Whereas SpaceX is not afraid to stick their finger in the fan and plow ahead with completely new designs & technology. As seen with their destruction of rockets, even launchpads.💥
Congress very much like NASA the way it is, and since they decide what gets done, that's the NASA we get.
How much of SLS's risk do you think is inherited from the solid boosters? If I recall right, flying with SRBs has more modes where aborts are not possible? I love me some APCP but something tells me it's not something to mix with humans.
Interesting question, and I don't recall seeing a detailed PRA for the SRBs. Shuttle couldn't abort until the solids were done burning because they couldn't be jettisoned, but IIRC the Orion escape motor can be triggered during that period.
Solids can have very energetic failure modes, however, and they can toss flaming chunks of propellant very high into the air. Not good if those happen to come down on your parachutes.
Eric, as of 2018 the total mission LOC estimate for SLS/Orion is 1/345 and the ascent LOC estimate is 1/3500.
The EDL LOC estimate is 1/720.
These numbers are for the Artemis 2 crewed lunar flight, but they are a pretty good reflection of the overall system.
Overall this is subtantially better than what the Space Shuttle could acomplish in the last decade of operation
source: 21-JSC-F-00590 - Orion LOC/LOM Requirements and Current Estimates - 2018
Thanks.
Can you point me to the actual document?
Or can you explain to my why the the ascent LOC estimate is so much better than shuttle?
@@EagerSpace SLS has far better abort modes than the shuttle had with its launch escape system. SLS having far better ascent LOC estimates than the Shuttle had seems only natural to me, but that is only my default assumption. I don't have any numbers.
Good one again. 17:36 a constellation video would be quite cool
It's on my list; there's lots of fun (okay, weird) stuff there.
Awesome in-depth analysis #TeamEager
Fascinating analysis.
The cultural element is key here.
NASA had its formative years and culture created in a war environment, when loss of crew was regarded as an acceptable risk for national warfighting during the race to space.
Something of that culture remains.
It comes down to incentives and a bureaucracy will do anything to realise its true motives, whatever they are. If you’ve spent a few billion dollars on a launch vehicle, ya gotta launch.
Yes, it's all about incentives.
Part of the reason Apollo worked was that there was an overriding goal - *everybody* wanted to beat the Russians, and nobody wanted to be the ones who fell short or killed an astronaut. Given what they accomplished with Apollo, it's really surprising that they didn't kill anybody in flight - though Apollo 13 came close.
Then post Apollo, the incentives were all set up by "the people who matter", in this case the congresspeople, the managers at NASA centers and headquarters, and the management at the contractors. Their incentives were all about career & money.
I'm the middle of a video that talks about this as it pertains to Artemis, and it does *not* look good to me.
3:14 I’d recommend instead this expression:
[(square root of x))-1]^2]. So square root of 36 is 6, minus one to get five and square 5 to make 25.
It's not quite right to be comparing SLS's 1 in 75 to Commercial Crew's 1 n 270, because much of the risk for SLS is during the in-space mission.
That said, I think this does understate the safety difference because failures from space vehicles tends not to be randomly distributed, but caused by specific targetable deficiencies in the design. For a design that only flies humans, every such deficiency will first fail in a way that directly puts lives at risk. For a design that flies humans only on a tenth of flights, any such deficiency only has a 1-in-10 chance of failing on a human flight. As these risks tend to retire after a single incident, this directly implies a factor-10 safety improvement.
One can say a similar thing about reuse, in that most failures occur during production. You could even imagine that after SpaceX has a decent fleet of space capsules all with a significant flight record, then the first flights of new vehicles might be done unmanned to retire that extra upfront risk that you don't have with the old ones.
I thought about the difference in risk scenarios, and decided not to dive in. SLS does have quite a bit of risk because of the lunar portion of the mission, but that mission is relatively short in duration. The hardest risk part for commercial crew is not ascent and descent, it's the requirement to be on-orbit for 90 days because that exposes the capsule to a *lot* of risk from impact damage, and it's really hard to manage that risk - it was clear early in commercial crew whether 1 in 270 was possible because of that risk.
I agree about early program failures and decided not to dive too deeply into that, other than the late mention that Soyuz had issues up front and not later. This wasn't really true with Shuttle - I was really surprised to see not only how bad the risks were but how they mostly persisted throughout the life of the program; to only make it from 1:10 to 1:90 over 30 years is a real indictment of how poor the shuttle design was from a safety perspective.
Good video. Unfortunately, NASA got complacent, and put statistics before common sense.
The "chance" of failure with rockets and space flight is and always will be 100%.
Successful missions are not "accidents".
They are the result of a lot of people working together to eliminiate the 100% "chance" of failure, knowing that Murphy and his law are always present.