Aviation week article on SpaceX reusability

[LouScheffer]

There is a most excellent article on Aviation Week, SpaceX Building Airline-Type Flight Ops For Launch that is chock full of information. It's behind a paywall, but here are a few points:

They are now aiming at 15 flights per booster.  They do this vibrating, heating, etc. each component to 4x the expected time of exposure at 15 flights.

They are aiming at a flight every 5 days with a 21 booster fleet.

The engine boots are water-cooled and get replaced proactively.

The oxidizer transfer tube has seals (thermal expansion, I'd guess) and is being looked at carefully.

SpaceX has an airline type inspection system.  'A' check after every flight, 'B' check every 6-7 flights, 'C' check for life leaders and crewed missions.

Static fires used to be run after an engine was pulled or a turbine wheel changed.  Now only if 3 engines are pulled; turbine wheels are now monitored by accelerometers during startup, and don't force a static fire.

Even Block 5 is still evolving; examples include better thermal shields and a way to drain the second stage tanks even if the QD pops off (which would otherwise be fatal). For most changes they identify the potential improvement, then test it on the bench, then on a StarLink launch, then it becomes general.

SpaceX has had two recent-ish booster engine shutdowns (March 2020 and Feb 2021).  Both still made orbit.  The 2020 failure was cleaning fluid left in the oxygen section of the gas generator, the 2021 failure was hot gas penetrating the engine shield and eating a harness.  This was the center engine so no landing was possible.

They keep tweaking the time interval between MECO and second stage start.  The smaller the delay, the better the performance, but the more wear on the first stage.

On Starlink launches they are experimenting with dropping the fairings earlier, allowing up to 10x the normal heating (which I think is one sun).

They have two software systems tuned for re-use.  One called Hyperion looks at the telemetry and points out areas that need a closer look.  Warp, their production system, also has modifications for re-use such as pointing out similar problems that have occured on prior flights.

Lots of other tidbits.  Overall a fascinating article in the best Aviation Week tradition.

[ThatOldJanxSpirit]

It’s so good to see how well Gerst appears to have adapted to SpaceX culture, and fascinating how frustrated he had become with the NASA way.

[Navier–Stokes]

Thanks for the summary. It should be noted that the full article is accessible with a free Aviation Week account.

[Norm38]

Why would dropping the fairings earlier mean more heating? 
Isn’t that a lower velocity and closer to shore?

[gongora]

Why would dropping the fairings earlier mean more heating? 
Isn’t that a lower velocity and closer to shore?

More heating of the uncovered payload on Stage 2

[rpapo]

Why would dropping the fairings earlier mean more heating? 
Isn’t that a lower velocity and closer to shore?

More heating of the uncovered payload on Stage 2

It's not just heating.  If you uncover the payload too soon, you also risk damage due to physical drag.

[Jim]

It’s so good to see how well Gerst appears to have adapted to SpaceX culture, and fascinating how frustrated he had become with the NASA way.

He was the reason for the "NASA" way and was in position to affect it.

[meekGee]

It’s so good to see how well Gerst appears to have adapted to SpaceX culture, and fascinating how frustrated he had become with the NASA way.

He was the reason for the "NASA" way and was in position to affect it.

Yup and either he found it impossible to change, or more likely it was impossible for him to see how stagnant it is while looking from the inside.

If that was all he saw and knew, then that's what he thought the universe was like.  And after a bit of time somewhere else, he came around to realize maybe there is a better way, and he now gets frustrated at the very things he used to accept.

It just shows that he's capable of changing and growing.

[john smith 19]

They are now aiming at 15 flights per booster.  They do this vibrating, heating, etc. each component to 4x the expected time of exposure at 15 flights.

This is a very interesting.

So 7 years after the start of reuse a booster can manage maybe 15 flights, compared to the previous 0 reuses of all previous systems. It's very impressive what can be achieved with a constant focus on improvement.

So linearly that's a smidgeon over 2 increases per year.

[Edit Just to put this in perspective the Space Shuttle Discovery made 39 flights during its opearting lifetime. I'm not sure if the structure was at its retirement limits at that point or if they could have continued flying if the programme continued. Stage and pad turnaround times are remarkable, but F9 is still far from breaking that number of reuses but with current rate of prgress should be there by 2033]

IOW into 10s of reuses of a first stage to M6, but nowhere near the 100's (1000s?) Musk thought was possible.  That suggests the booster takes a hell of a beating

Obviously a clean-sheet design incorporating all the lessons learned from F9/FH ops will be considerably better

But how much better? On the current basis a 100 flights sound like a massive improvement.  Actual airliner relaibility is into the 10s of 1000s.

Static fires used to be run after an engine was pulled or a turbine wheel changed.  Now only if 3 engines are pulled; turbine wheels are now monitored by accelerometers during startup, and don't force a static fire.

There was a NASA project using these techniques connected to the SSME's. Seemed an excellent idea but I was never sure if they got it deployed before the programme ended.
Obviously the processors running the Merlins are much more powerful than the dual M68000's that ran the 2nd generation SSME controllers.

Even Block 5 is still evolving; examples include better thermal shields and a way to drain the second stage tanks even if the QD pops off (which would otherwise be fatal). For most changes they identify the potential improvement, then test it on the bench, then on a StarLink launch, then it becomes general.

TBH I always thought the claim that "Block 5" would be the end of F9 development was rather strange and always expected them to go on tweaking it, wheather it was announced as  new version or not.

They keep tweaking the time interval between MECO and second stage start.  The smaller the delay, the better the performance, but the more wear on the first stage.

Now this is surprising. I'd have guessed a longer coast would have given them more altitude and be beneficial. OTOH more booster damage due to earlier sustainer ignition makes perfect sense due to greater exposure to engine plume.

They have two software systems tuned for re-use.  One called Hyperion looks at the telemetry and points out areas that need a closer look.  Warp, their production system, also has modifications for re-use such as pointing out similar problems that have occured on prior flights.

NASA Goddard have done a lot of work in this area. The application of AI ("Deep learning" to some, advanced pattern recognition to others) to space operations as a way to eliminate routine eyeballing of the data and automated management-by-exception

It's good to get an idea of how an SoA launch operations centre functions.

One key area that has proved very expensive in the past (IIRC about 30% of shuttle launch costs) were misison planning. IE trajectory design and things like coupled loads analysis IE making sure the payload didn't excite dangerous resonances in the LV, and vice versa. Historically this has been very time consuming, especially when the a dangerous mode can't be fixed by changing the engine burn, or the fuel loading and the payload needs to be rebuilt.

[FutureSpaceTourist]

IOW into 10s of reuses of a first stage to M6, but nowhere near the 100's (1000s?) Musk thought was possible.  That suggests the booster takes a hell of a beating

100s of reuses requires 100s of flights. I think SpaceX have just been taking a cautious / methodical approach as they’re not sure where the weak links might be and at what point they’ll hit a limitation (at least without significant refurbishment). So they’ve built up a reasonably sized fleet of boosters to support many flights even if 10 - 15 per booster proves a limit.

100s of flights may be possible but with Elon switching SpaceX’s focus to Starship, I don’t think he’s particularly interested in finding out any more (for F9, Starship is a whole other issue).

[john smith 19]

100s of reuses requires 100s of flights. I think SpaceX have just been taking a cautious / methodical approach as they’re not sure where the weak links might be and at what point they’ll hit a limitation (at least without significant refurbishment). So they’ve built up a reasonably sized fleet of boosters to support many flights even if 10 - 15 per booster proves a limit.

15 flights (IIRC the fleet leader has done 13) seems pretty low. They could do 2 more launches on it and find out what happens. That would still leave them with 20 boosters to cycle through

100s of flights may be possible but with Elon switching SpaceX’s focus to Starship, I don’t think he’s particularly interested in finding out any more (for F9, Starship is a whole other issue).

Maybe, maybe not.

The question is wheather what causes failure is down to the materials or the launch mode itself.

If the former then (plausibly) switching to steel for Starship will radically raise the threshold as Aluminum just can't take it.

But if it's the launch mode itself that's the issue that's not really going to change.

Imagine you drove to work each day like this.
 You'd stand on the brakes, build up the revs then screech out your driveway.
At work you'd stop by coasting into your parking spot, starting up you engine in neutral, then engaging in reverse at the last possible moment so it stops just before it hits the kerb.


Sounds like it could be a bit tough on your engine?

Time will tell how much higher SX can get their stage limit.  Aircraft, any aircraft can handle 1000s of take offs and landings.  SX's launch manifesto should kick at least one stage up to its (current) life limit by the end of the year.

You can argue that this is not that important, with Starship taking over F9 launch duties but the question for that is when will it do so, given a) SS has not achieved orbit yet b)its last test flight was 3 months ago.

Maybe it's next flight be an "all-up-test" straight to orbit with full TPS, but we'll see.

[Hobbes-22]

Now this is surprising. I'd have guessed a longer coast would have given them more altitude and be beneficial. OTOH more booster damage due to earlier sustainer ignition makes perfect sense due to greater exposure to engine plume.

Coasting leads to an increase in gravity losses.

[Hobbes-22]

Why would dropping the fairings earlier mean more heating? 
Isn’t that a lower velocity and closer to shore?

Lower velocity, but also a lower altitude so more atmospheric drag. The threshold for dropping the fairing is the amount of heat generated by drag.

[rpapo]

Why would dropping the fairings earlier mean more heating? 
Isn’t that a lower velocity and closer to shore?

Lower velocity, but also a lower altitude so more atmospheric drag. The threshold for dropping the fairing is the amount of heat generated by drag.

Both of you are right, but talking about different things.  It looks like Norm38 is referring to the heating on the fairings, particularly upon reentry.  And that you (Hobbes-22) are talking about the heating on the payload. 

[john smith 19]

In another description of the article it's stated that 15 is the limit as they want to phase out F9 launches and that in fact the real limit is higher.

That may be true, but given that Starship hasn't reached orbit yet that is a very bold plan, never mind going to Mars (still 130 days to go to 2023).

I wonder if someone who's been tracking number of launches across the 21 booster fleet can say how many of those 315 possible launches (15*21) the fleet has left in total?

That's basically how long SX has to get SS orbit capable. Once that happens they can fix any outstanding issues with the design.

Obviously SX could use the last flight of each as a "special" with maximum payload due to no landing gear (even for the fairings, as they are phasing out the whole vehicle). Really crank up the payload. Outer planets probe? I don't think Neptune or Uranus has had a visit since Voyager.

[M.E.T.]

In another description of the article it's stated that 15 is the limit as they want to phase out F9 launches and that in fact the real limit is higher.

That may be true, but given that Starship hasn't reached orbit yet that is a very bold plan, never mind going to Mars (still 130 days to go to 2023).

I wonder if someone who's been tracking number of launches across the 21 booster fleet can say how many of those 315 possible launches (15*21) the fleet has left in total?

That's basically how long SX has to get SS orbit capable. Once that happens they can fix any outstanding issues with the design.

Obviously SX could use the last flight of each as a "special" with maximum payload due to no landing gear (even for the fairings, as they are phasing out the whole vehicle). Really crank up the payload. Outer planets probe? I don't think Neptune or Uranus has had a visit since Voyager.

How do you know that they won’t build any more boosters beyond the current 21? Every additional booster gives them 15 more launches as may be needed.

[john smith 19]

How do you know that they won’t build any more boosters beyond the current 21? Every additional booster gives them 15 more launches as may be needed.

It's the logical thing to do. Just as shutting down F1 production and the CFRP SS were the logical things to do after they outlived their usefulness. I don't think anyone has ever accused Musk of being a sentimentalist.

I agree it would be sensible for the time being, to continue low rate production of F9 boosters until an SS makes orbt and lands. But I'm not Elon Musk. It's the article we are discussing that says SX is retiring F9's after 15 flights because they want to be ready for SS.

For maximum financial benefit you shut down the booster mfg line, leaving the US line (are they still making new Dragons as well?) running, then build enough for the remaining agreed launches and shut that down as well.

The key items that can throw big spanners in the works are a)NASA contracts to supply ISS and b)SS is not orbit ready before the last F9 flies out.

Of these two I think NASA is the bigger issue. I think SX would like to go with a fully reusable system to support the ISS IE phase out Dragons. I think that's a step too far for NASA, and they will want to continue with dragons. Docking SS to ISS is a very different beast to docking a crewed Dragon.   If they really decided to be awkward they could insist on them launching on F9's, but I think they will be reasonable (and I suspect the contract clauses cover this issue in some form).

Which leaves what happens if the have shut down the F9 boosterline, run out of F9 booster launches and (for whatever reason) SS has still not reached orbit.

Hence my interest in finding out what the state of the "clock" actually is. With the fleet leader at 13 that gives at least another 42 launches across the fleet, which sounds quite a lot (by the standards of other launch service companies) but for SX not so much.

So I guess the next question is which has the least number of launches on it so far?

Of course if it's just a business decision to retire after 15 flights (and F9 boosters have lots more flights left in them) then they just have to raise the "retirement age" and keep them flying and the problem goes away.

[Robotbeat]

I agree it’d be smart to prepare a much bigger buffer of F9 capacity than Elon would probably prefer, but…

Docking Starship to ISS is not very different from docking Shuttle to ISS.

[john smith 19]

I agree it’d be smart to prepare a much bigger buffer of F9 capacity than Elon would probably prefer, but…

Keep in mind 42 is a minimum figure for available launches. I'm not sure how evenly the launches have been spread throughout the fleet. I think some of them are quite a bit lower, but it's not something I've tracked in detail, hence my question.  [EDIT Checked the F9 launch list on Wiki. It reckons they are at about 178 launches so far, so between 6-7 launches per booster left on average. Given some are within 2 launches of their 15 launch limit that confirms my feeling that the launches have been quite unevenly allocated, for whatever reasons]

As I noted for other launch companies having 42 launches available would be good for years of launches but SX's cadence is much higher than most (all?) others.

Docking Starship to ISS is not very different from docking Shuttle to ISS.

Given that it's about 13m longer than shuttle and designed to carry about 4x the mass of Shuttles payload bay (even if it came to orbit with only Shuttle's payload level) it's mass properties, such as 2nd moment of area, are going to be very different than Shuttles.

That's important because of the loads put on the docking adapter. I think of it as a pool cue. The sharp end is at the adapter but the heavy end is at the other end of the cue, only the heavy end is now 13m further away, able to excert much more torque on that interface.

Obviously with enough RCS control authority and fast enough acting control systems these issues can be overcome, if the will is there to do so.

We'll find out when Starship makes orbit. Starship docked to ISS will be quite a sight. 

[AmigaClone]

Currently, SpaceX has 13 active boosters. Additionally, there are 10 more boosters that have been spotted in either testing or transit.

It has been speculated that B1049 will be expended after it's next mission. Based on the available mission profile, it appears the 4 FH core boosters will be expended after their first launch as well.

The true number of launches with just the current boosters likely is between 80 and 160 - without SpaceX deciding to certify some of the later ones to 20 flights.

Based on the number of FH launches scheduled in the next 5 years, SpaceX will have to make up to 9 core boosters in that time frame, possibly more. I would not be surprised if SpaceX also makes several boosters in that time period.

I personally don't see the F9/FH being retired prior to Starlink Gen2 being at least close to full deployment. That easily could take up to 9 years taking into account the shear number of launches needed.