Looking at that full set of 33 engines burning almost brought a tear to my eye (and the controlled cutoff). That more than anything else tells me they've made it.
I'm going to assume you made this comment in good faith.
Most rockets (almost all) use just a few engines. The Saturn V had five engines, and the Space Shuttle had a total of five engines (including the Solid Rocket Boosters). The Atlas V uses one engine, New Glenn seven engines, etc.
Historically, there has only been one rocket that attempted the 'many engine' approach before now: the N1 by the Soviets, which they never managed to get working successfully.
SpaceX is thus attempting something unprecedented. They are not only trying to create a rocket with 33 engines, but they are also building the most powerful rocket ever made. Furthermore, it is the only rocket ever designed with fully reusable first and second stages. In this single machine, SpaceX is innovating in multiple ways that neither the US, the Soviets, nor any other country has managed to achieve.
Additionally, the Raptor engine used by SpaceX is the only full-flow staged combustion rocket engine that has ever flown. This design makes it more efficient and high-performing and has been considered the 'holy grail' of rocket engine design. Until now, no one has managed to build a successful rocket engine using full-flow staged combustion due to its complexity. The Raptor is also, pound for pound, the most powerful and efficient rocket engine ever made.
SpaceX is innovating throughout their space 'stack.' The entirety of the Starship project is immensely innovative in almost every way possible. Many others have tried and failed to achieve even a single aspect of what this project encompasses. The fact that SpaceX managed to have 33 Raptors firing perfectly today is already absolutely phenomenal.
So yes, I expect things to fail because much of what they are doing is literally the first attempt of its kind. It's entirely predictable that some aspects might fail. What you're asking is akin to questioning the pioneers of quantum computing: 'Hey, are you telling me the first quantum chip you built failed? That’s crazy! How can you build something so expensive and expect it to fail?'
Because of combustion instability modes, cascading failures, and finally the simple difficulty of high reliability operation of a piece of machinery that is so highly stressed.
There is a very interesting book about it, called Combustion Instabilities in Liquid Rocket Engines: Testing and Development Practices in Russia.
People genuinely don’t understand how statistics work. Which is part of why Vegas still makes so much money.
When you put together a bunch of equipment with a small error rate, the time between errors climbs very fast. Build a RAID array with 33 disks and you’d better have a vendor picked out for replacements because you’ll be doing replacements fairly often, instead of every four to ten years with a single disk.
And they don’t understand dependent statistics either. Every failing rocket engine can potentially damage its neighbors. Every failing hard drive requires a stressful operation on the remaining disks (resilvering) that may push the next drive to failure.
Yes, naturally 33 > 1, so you might expect 33 times as many failures of individual components.
But your analogy between arrays of rocket engines and disks is apt, because both have redundancy to survive the failure of individual components.
For example, in the high-altitude flight test of the Starship prototype in May 2021, three of the 33 Raptor engines powering the first stage failed shortly after liftoff. The vehicle still managed to continue flying, reaching an altitude of 40 kilometers before failing due to a variety of causes.
> People genuinely don’t understand how statistics work
Indeed.
If an individual disk has MTBF of 2 million hours, the probability of it failing in the first year is 0.437%.
But put 33 of those disks in a RAID 6 array, which can tolerate 1 or 2 failures without replacement, and the probability of the entire array failing in the first year drops by a factor of ten to 0.0413%.
The statistics say the array is even more reliable than a single component by itself.
> Every failing hard drive requires a stressful operation on the remaining disks (resilvering) that may push the next drive to failure.
I saw something similar: RAID array with 5 disks and 3 spare disks.
One night controller detected failure, ejected one drive and replaced with spare.
After rebuild of RAID it ejected another disk and then another. In an hour all spare disks were used and on fourth failure it stopped work
Everyday Astronaut had a couple of (IMHO) great videos which goes into why rocket engines are hard[1] and the Starship many-engine approach[2] specifically.
It’s literal rocket science. Liquid fueled rockets have to work in basically the most extreme environments we’ve ever made machines work. Extreme cold temperatures through extreme hot and from extreme high pressures through vacuum. You have to have the best metallurgy and amazing machining tolerances. And then on starship they got it to work more than thirty times simultaneously.
Things like vibration and Pogo oscillation (https://en.wikipedia.org/wiki/Pogo_oscillation). All the engines collectively have about twice the thrust of a Saturn S-IC first stage (on the Saturn V rocket) and they're all going to be trying to shake the crap out of each other just due to thrust oscillations. If they hit a resonant frequency of the rocket it could shake the rocket apart. So, you can have 33 engines that all individually ignite properly on the pad and are all well-built to tolerances so that they function, but when you put the whole vehicle together and have all 33 of them pushing against the same airframe they could spectacularly fail.
Yes but the reliability of modern car engines is a minor miracle, achieved through over a century of incredible engineering effort. Also remember that production lines always have yield, which means that not everything that rolls off it will actually work and the non-functional components are discarded. For example new chip production fabs very often have low yields, sometimes even the majority of chips don't work.
The limits of simulation (From Henry Spencer)
The 27May91 Aviation Week, reporting on the April 1 test-stand failure of an upgraded SRB (Solid Rocket Booster) for the Titan 4:
Investigators determined that extensive three-dimensional computer simulations of the [motor's] firing
dynamics did not reveal subtle factors that they now believe contributed to motor failure. [Program director]
Stifling said the full-scale test was essential precisely because computer analyses cannot accurately predict all
nuances of solid rocket motor dynamics. "That's why we test", he said.
For those who don't follow the space news, a few seconds into the test the motor pressure rose rapidly and
exceeded the limits of the casing, the result being a large, spectacular explosion that destroyed the motor and
much of the Edwards AFB test stand.
-ACM SIGSOFT SOFTWARE ENGINEERING NOTES vol 16 no 4 Oct 1991 Page 15
saberience’s reply should be read because it is more complete and insightful, but perhaps there is a simpler intuitive answer. Error rates increase exponentially.
Say you want to build something new, something that is beyond the bounds of what anyone has ever built before. In this case, a rocket that has more thrust than any built before. The basic mechanics say that it is feasible, so you start to work the details. You find the limit of knowledge in all kinds of directions, and determine which limits you can definitely build the rocket within, and which ones you will need to exceed. As the world's most powerful rocket, perhaps it will exceed current bounds of knowledge about things like material performance and turbulence under extreme temperatures and pressures.
Anything built outside of currently-known parameters might fail, so you design as much of the rocket as possible to operate within the limits of current knowledge.
Some of those limits turn out to be feasible to push in a laboratory. So you build and conduct experiments on those parameters, record the results, and design your rocket accordingly.
But for some parameters, the lab experiments required to test them are incredibly costly. Especially for parameters that vary strongly with scale (like turbulence), a system with the size and energy of the world's largest rocket can only be predicted by an experiment with the size and energy of the world's largest rocket.
Say it costs a billion dollars to build and launch your new rocket, but it will operate with ten unknown parameters, each of which will cost 200 million dollars to individually test on the ground. You could spend two billion dollars on experiments, then one billion dollars on a rocket you're confident will work. This is more-or-less the model that NASA used for developing the Space Launch System: do as much science as necessary, ask Congress for as much extra time and money as the contractors say they need, so they can be extremely confident that their new biggest rocket will work perfectly the first time. NASA doesn't like launching rockets that they don't know will work, because Congress doesn't like when rockets they paid for blow up, and NASA depends on Congress for all of their funding.
However, for the same money, you could build your rocket best-guess within the known unknowns and fly it. It's likely to blow up, because you guessed on ten unknowns. However, if the results of the flight allows you determine some of the parameters with greater confidence than one billion dollars in ground experiments, then it was worth the while, and you re-design your rocket according to the newly-understood parameters. You can do this three times for the cost of building the experiments and flying once. That's a decent chance at learning the parameters, and as a bonus, you get to practice building and flying the rocket, productively employing all of your staff and facilities.
That's more-or-less the model that SpaceX is developing Starship on, and previously Falcon 9. Blowing up prototype rockets only costs SpaceX time and money, and as long as it costs less time and money than running ground experiments (and keeping their production and launch crews on retainer, and maintaining their facilities between rockets), they're saving money by launching rockets they expect to blow up.
