This sounds like one of those "and also" things. I'd say you add fire suppression AND ALSO try more to reduce leaks. It's got to be really difficult to build huge massive tanks that hold oxygen and other gases under pressure (liquid methane too will have some vapor of course). Are leaks inherently going to happen?
This is meant to be a human rated ship of course, how will you reduce this danger? I know this stuff is hard, but you can't just iterate and say starship 57 has had 3 flights without leaks, we got it now. Since I have no expertise here, I can imagine all kinds of unlikely workarounds like holding the gas under lower pressure with humans on board or something to reduce the risk.
This might be one of those components where it just needs to be built without problems, and improved safety means fixing individual design and manufacturing flaws as you find them, until you’ve hopefully got them all.
This can work. Fundamental structural components of airliners just can’t fail without killing everyone, and high reliability is achieved with careful design, manufacturing, testing, and inspection. I’m not sure if a gigantic non-leaky tank is harder to pull off that way, but they might have to regardless.
We’re going to have to accept that space travel is going to be inherently dangerous for the foreseeable future. Starship is in a good position to improve this, because it should fly frequently (more opportunities to discover and fix problems) and the non-manned variant is very similar to the manned variant (you can discover many problems without killing people). But there are inherent limitations. There’s just not as much capacity for redundancy. The engines have to be clustered so fratricide or common failure modes are going to me more likely. Losing all the engines is guaranteed death on Starship, versus a good chance to survive in an airliner.
All other practical considerations aside, I think this alone sinks any possibility of using Starship for Earth-to-Earth travel as has been proposed by SpaceX.
High reliability of airliners is achieved by having redundancy of all critical parts. The idea is no single failure can cause a crash.
For example, if system A has a failure probability of 10%, if A is redundant with another A', the combined failure probability is 1%.
That of course presumes that A and A' are not connected.
Yes for systems, not always for structure. A failed wing spar means everybody dies. For real-world examples, there were two 747 crashes caused by improper repairs to a rear pressure bulkhead or aircraft skin. When the repairs eventually failed, the explosive decompression caused catastrophic damage to the tail in one instance, and total structural failure resulting in a mid-air breakup in the other.
The response to this was to make sure repairs are carried out correctly so the structure doesn’t fail, not to somehow make two redundant bulkheads or two skins.
The wing spar is dual, too.
The idea is to design the airplane to survive an explosive decompression failure, not pretend that explosive decompression doesn't happen. For example, on the DC-10, the floor collapsed from explosive decompression, jamming the control cables and causing a horrendous crash.
The fix was not preventing explosive decompression. The fix (on the 757) was to locate the redundant set of control cables along the ceiling. Also, blowout panels were put in the floor so the floor wouldn't collapse.
It's not always practical to fix an older design like the 747. When it isn't practical, a stepped-up inspection protocol is added.
P.S. The 747 was designed to survive a decompression. The oversight was nobody realized that a failure of the rear bulkhead could destroy the tail section. Things like that happen in complex systems, and an airliner is incredibly complicated.
P.P.S. When I was a newbie at Boeing, I asked about the wing spar, too. That's how I know it is dual!
Lindbergh's Spirit of St Louis had the main fuel tank directly in front of him. This was in spite of his primal fear of being burned alive. In some airplanes you sit on the fuel tank.
Given that a) most human rated rockets have had 0 flights before use, and b) I'd expect each starship to have at least 10 flights, and at least 100 in total without mishap before launching, the statistics should be good
I don’t think (a) is true. The Shuttle flew with people on its maiden voyage, but that’s the only one I can think of.
(b) is true and should make it substantially safer than other launch systems. But given how narrow the margins are for something going wrong (zero ability to land safely with all engines dead, for example) it’s still going to be pretty dangerous compared to more mundane forms of travel.
Most rockets flew test flights before sticking people inside the same model, but most rockets are also single use and so each stack is fundamentally new.
A future starship could plausibly be the first rocket to fly to space unmanned, return, and then fly humans to space!