Safe Instantaneous Impact Point Tracks for Starship Launch and Recovery

[TheRadicalModerate]

Update Dec 5, 2024:  Let's open this up to both EDL and launch IIP tracks.  Seems kinda silly to start a separate thread for launch. --TRM


This came up on another thread, and I suspect it's an interesting but off-topic rat hole there, so I'm creating this one for anybody who has any thoughts.

Here's the problem:

For any plausible set of inclinations, where Starships need to be recovered to either BC or LC-39A, the ground tracks are going to go over heavily populated areas, where any chance of falling debris will cause the FAA to wag its virtual finger at SpaceX.

Now, it's important to distinguish between the ground track (i.e., a normal projection of the locus of position vectors of the reentering Starship onto the ground) and the IIP, which is the locus of spots on the ground where debris would fall ballistically, then subject to drag in the atmosphere.  It's possible that the hypersonic portion of the the Starship's trajectory is so high and short that the scary parts of the IIP will only throw debris out to sea.  Then, once the Starship is in belly-flop mode, the IIP stays directly below the vehicle, which can then slide over to be caught by chopsticks on land at the last moment.

But I'm... skeptical that this is going to be the case.  ISTM that would require unreasonably high accelerations on the vehicle.  So some parts of the IIP with high probability of receiving debris are probably going down the Rio Grande Valley or Matamoros for BC, or Tampa and Orlando for LC-39A.

How is this gonna work?

[David GREENFIELD]

From my dumb point of view, can’t they avoid the issue by treating it like the X-37, or Orbiter? Or at some point they should have confidence that the likelihood of break-up is low enough to treat it more or less like an aeroplane.

[Brigantine]

How did shuttle do it?

Other than having some degree of steerability through those S curves / roll reversals.

Also, I suspect all the plans will change once we see some actual re-entry data

[TheRadicalModerate]

How did shuttle do it?

Other than having some degree of steerability through those S curves / roll reversals.

Also, I suspect all the plans will change once we see some actual re-entry data

I suspect that Shuttle did it by being in service before the FAA had the regulations in place.  I don't know how the X-37B does it; it could also be grandfathered, or maybe there's a military waiver of some sort.

[steveleach]

Be gentle with me if this is an obviously stupid question, but would Starship be able to do a u-turn?

As in, if it turns slightly to the left (or right) during re-entry, can it turn through 180 degrees and end up travelling east-to-west?

[Brigantine]

More generally, what's the best-case lift:drag Starship can get?

As for U-turn during re-entry... Actually a good question. Most of the maneuverability you do get is after you slow down to aeroplane speeds.
I found some info about Space Shuttle's ability to turn during re-entry, and that was a lot more aeroplane-like than Starship is.
(Note the loops near the runway)

[Twark_Main]

How did shuttle do it?

I suspect that Shuttle did it by being in service before the FAA had the regulations in place.

Welp, shut it down guys...

I guess humans can't (affordably) go to space. The Priests Glorious Regulators have spoken!!

[TheRadicalModerate]

Be gentle with me if this is an obviously stupid question, but would Starship be able to do a u-turn?

As in, if it turns slightly to the left (or right) during re-entry, can it turn through 180 degrees and end up travelling east-to-west?

This is that whole issue of "cross range," upon which USAF insisted and is widely blamed for a lot of the Shuttle's heavy aerosurfaces and generally poor performance.  I suspect that cross range will be extremely limited for Starship.

[edzieba]

In terms of public safety concerns with overflight during EDL, the Gordian solution would be to design for demisability in event of a RUD for as much of the flight as possible, overfly only in the demisable regions of flight (e.g. extend ground-track out past the coast and 'double back', doing so propulsively if absolutely necessary), and verify the accuracy of those models with real-world demisability tests over instrumented ranges.
There is nobody at the FAA twirling a moustache and trying to think of ways to 'block Starship'. If it meets the public safety requirements it meets the public safety requirements, if it doesn't it doesn't, and SpaceX are in the enviable position of being the ones who get to decide how they meet the requirements (same as for the Launch License).

How did shuttle do it?

I suspect that Shuttle did it by being in service before the FAA had the regulations in place.

Welp, shut it down guys...

I guess humans can't (affordably) go to space. The Priests Glorious Regulators have spoken!!

Oh no, a design constraint! No human spaceflight programme has ever survived a design constraint being imposed!

Except for the entire Commercial Crew programme (e.g. PLoC requirements),  STS (crossrange, HEXAGON payload bay volume, etc), Apollo (too many to count stemming from distributing the programme across multiple contractors and states), etc.

[steveleach]

More generally, what's the best-case lift:drag Starship can get?

As for U-turn during re-entry... Actually a good question. Most of the maneuverability you do get is after you slow down to aeroplane speeds.

It sounds like we can't say that a 180 degree turn is definitely impossible though.

How slow would the vehicle have to get before it could be considered safe enough to overfly populated areas, and how high would it be when it reaches that speed?

How fast could it turn through 180 degrees (or maybe even less, like 90), and how much horizontal and vertical speed would it lose in doing so?

Then how far could it get with the horizontal and vertical speed it has left, and would that be sufficiant to get back to the launch site?

If we know these then could we figure out if there's a viable profile that has the ship coming NW->SE over land at high altitude, slowing, descending and turning to an NE->SW profile while over the sea, and then landing back at the coast.

[catdlr]

More generally, what's the best-case lift:drag Starship can get?

As for U-turn during re-entry... Actually a good question. Most of the maneuverability you do get is after you slow down to aeroplane speeds.

It sounds like we can't say that a 180 degree turn is definitely impossible though.

How slow would the vehicle have to get before it could be considered safe enough to overfly populated areas, and how high would it be when it reaches that speed?

How fast could it turn through 180 degrees (or maybe even less, like 90), and how much horizontal and vertical speed would it lose in doing so?

Then how far could it get with the horizontal and vertical speed it has left, and would that be sufficiant to get back to the launch site?

If we know these then could we figure out if there's a viable profile that has the ship coming NW->SE over land at high altitude, slowing, descending and turning to an NE->SW profile while over the sea, and then landing back at the coast.

There is also the "upper node" approach over Mexico to the gulf and then into Boca Chica.

[Twark_Main]

In terms of public safety concerns with overflight during EDL, the Gordian solution would be to design for demisability in event of a RUD for as much of the flight as possible, overfly only in the demisable regions of flight (e.g. extend ground-track out past the coast and 'double back', doing so propulsively if absolutely necessary), and verify the accuracy of those models with real-world demisability tests over instrumented ranges.
There is nobody at the FAA twirling a moustache and trying to think of ways to 'block Starship'. If it meets the public safety requirements it meets the public safety requirements, if it doesn't it doesn't, and SpaceX are in the enviable position of being the ones who get to decide how they meet the requirements (same as for the Launch License).

How did shuttle do it?

I suspect that Shuttle did it by being in service before the FAA had the regulations in place.

Welp, shut it down guys...

I guess humans can't (affordably) go to space. The Priests Glorious Regulators have spoken!!

Oh no, a design constraint! No human spaceflight programme has ever survived a design constraint being imposed!

Fallacious reasoning by analogy.

Everything that can be called a "design constraint" is not created equal.  Some constraints are more constraining than others. 


And no, no fully-reusable (affordable) human spaceflight program has ever needed to survive this particular design constraint before. Gliding reentry is necessary for full reuse, but it's also fundamentally more risky.


Regulations can and do kill programs. It's weird to dismiss this causation as somehow absurd or unthinkable.

[TheRadicalModerate]

In terms of public safety concerns with overflight during EDL, the Gordian solution would be to design for demisability in event of a RUD for as much of the flight as possible, overfly only in the demisable regions of flight (e.g. extend ground-track out past the coast and 'double back', doing so propulsively if absolutely necessary), and verify the accuracy of those models with real-world demisability tests over instrumented ranges.

I'm skeptical that you can design a demisable vehicle with heavy engines and TPS that's designed to survive not only reentry but launch acoustics.

I hadn't thought about a "boostback" as part of the EDL sequence.  That's an interesting idea.  But the question is how far out to sea do you need to be before you begin the boostback.

The conops would look something like this:

1) Start west-to-east reentry close enough to the coast that the IIP starts over the ocean.

2) Slightly reduce downward lift, so the Starship pops up out of most of the atmosphere just as it reaches zero downrange speed.  (If you can have some vertical speed, that would be helpful, but I don't know how you'd have vx=0 and vy>>0.)

3) Yaw (or pitch) the vehicle 180º and do a boostback restart.  You'd want a 45º horizon angle, so the Starship would get as far back uprange as possible, making a second, pretty slow reentry, then belly-flopping just offshore.

4) Second reentry, which will have very low uprange velocity, and bellyflop pretty much from entry all the way to the ground.

5) Slide from just offshore onto the chopsticks if things look good for landing.  Otherwise, ditch into the drink.

Cheesy, flat-earth diagram attached.

The big problem here is that I suspect you need to be at least hundreds if not thousands of km out to sea when the Starship pops out of the atmosphere.  If you do an impulsive burn (which you can't), a 45º horizon angle will give you the most bang for the delta-v, but the boostback distance will be Δv²/g.  When you crank that through, 500km downrange will require more than 2200m/s of delta-v.  Pretty expensive.

I suspect that 500km is nearly far enough downrange to avoid debris falling on land in the event of an accident.


Another possibility:  Land significantly east of your launch site (e.g. the west coast of Florida for BC or the Turks and Caicos for LC-39A.  You then refuel the Starship and do a P2P trajectory to get it back to the launch site.

This is also a pretty terrible idea.  I don't have any good ones--hence the thread.

And no, no fully-reusable (affordable) human spaceflight program has ever needed to survive this particular design constraint before. Gliding reentry is necessary for full reuse, but it's also fundamentally more risky.

Regulations can and do kill programs. It's weird to dismiss this causation as somehow absurd or unthinkable.

I suggest we rule out discussion of how to get the FAA to change their regulations.  That strikes me as nearly impossible, and can only devolve into political ranting.

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.


PS:  Staring at my own cheesy diagram, I realized that the point from second entry to the just-offshore bellyflop point is a non-trivial distance.  That will help get the skip point (i.e., the point near entry interface where vx and vy are close to zero) farther downrange, which is what we want.  It's still a boatload of delta-v you have to waste, but you might get the IIP completely out to sea that way.

[Twark_Main]

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.

Sure. It's the same solution as launch abort.

"Don't explode."

[TheRadicalModerate]

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.

Sure. It's the same solution as launch abort.

"Don't explode."

The FAA thinks that exploding during launch is fine, at least for a cargo or tanker mission.  They just care about debris landing on people.  That's relatively easy to engineer during launch.  The question is how to engineer it during landing.

For purposes of this discussion, let's assume that crewed launch/EDL is off-topic.  There's another thread for that.

[sdsds]

Wacky idea: build a landing and relaunch pad at VSFB on the California coast, complete with its own set of SH boosters. Launch to the west out over the Pacific, and take the anti-Coriolis hit for the sub-orbital hop back to LC-39 or BC.

[eriblo]

In terms of public safety concerns with overflight during EDL, the Gordian solution would be to design for demisability in event of a RUD for as much of the flight as possible, overfly only in the demisable regions of flight (e.g. extend ground-track out past the coast and 'double back', doing so propulsively if absolutely necessary), and verify the accuracy of those models with real-world demisability tests over instrumented ranges.

I'm skeptical that you can design a demisable vehicle with heavy engines and TPS that's designed to survive not only reentry but launch acoustics.

I hadn't thought about a "boostback" as part of the EDL sequence.  That's an interesting idea.  But the question is how far out to sea do you need to be before you begin the boostback.

The conops would look something like this:

1) Start west-to-east reentry close enough to the coast that the IIP starts over the ocean.

2) Slightly reduce downward lift, so the Starship pops up out of most of the atmosphere just as it reaches zero downrange speed.  (If you can have some vertical speed, that would be helpful, but I don't know how you'd have vx=0 and vy>>0.)

3) Yaw (or pitch) the vehicle 180º and do a boostback restart.  You'd want a 45º horizon angle, so the Starship would get as far back uprange as possible, making a second, pretty slow reentry, then belly-flopping just offshore.

4) Second reentry, which will have very low uprange velocity, and bellyflop pretty much from entry all the way to the ground.

5) Slide from just offshore onto the chopsticks if things look good for landing.  Otherwise, ditch into the drink.

Cheesy, flat-earth diagram attached.

The big problem here is that I suspect you need to be at least hundreds if not thousands of km out to sea when the Starship pops out of the atmosphere.  If you do an impulsive burn (which you can't), a 45º horizon angle will give you the most bang for the delta-v, but the boostback distance will be Δv²/g.  When you crank that through, 500km downrange will require more than 2200m/s of delta-v.  Pretty expensive.

I suspect that 500km is nearly far enough downrange to avoid debris falling on land in the event of an accident.


Another possibility:  Land significantly east of your launch site (e.g. the west coast of Florida for BC or the Turks and Caicos for LC-39A.  You then refuel the Starship and do a P2P trajectory to get it back to the launch site.

This is also a pretty terrible idea.  I don't have any good ones--hence the thread.

And no, no fully-reusable (affordable) human spaceflight program has ever needed to survive this particular design constraint before. Gliding reentry is necessary for full reuse, but it's also fundamentally more risky.

Regulations can and do kill programs. It's weird to dismiss this causation as somehow absurd or unthinkable.

I suggest we rule out discussion of how to get the FAA to change their regulations.  That strikes me as nearly impossible, and can only devolve into political ranting.

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.


PS:  Staring at my own cheesy diagram, I realized that the point from second entry to the just-offshore bellyflop point is a non-trivial distance.  That will help get the skip point (i.e., the point near entry interface where vx and vy are close to zero) farther downrange, which is what we want.  It's still a boatload of delta-v you have to waste, but you might get the IIP completely out to sea that way.

Starship does not have anywhere close to the amount of lift you need to do that. Exit the atmosphere = still have most of your velocity. Low velocity = can not get enough altitude.

[steveleach]

I found some info about Space Shuttle's ability to turn during re-entry

That reference says that the shuttle entry interface was 122 km from touchdown, so I upped that to 150 km to be safe and used Google Maps to see what that might look like for a u-turn approach into BC.

I guess it all comes down to how much ability Starship has to turn during entry.

[edit] Ignore this, that's 122 km altitude, not downrange. Doh!

[Twark_Main]

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.

Sure. It's the same solution as launch abort.

"Don't explode."

The FAA thinks that exploding during launch is fine, at least for a cargo or tanker mission.  They just care about debris landing on people.

Of course, we all know that.

The way SpaceX will reduce the risk of debris landing on people during reentry is to reduce the risk of Starship exploding and/or disintegrating during reentry.

For purposes of this discussion, let's assume that crewed launch/EDL is off-topic.  There's another thread for that.

Launch, sure. But EDL is off-topic for a "Starship Recovery" thread? 

[edzieba]

I've been assuming that SpaceX has some kind of solution to this problem.  I just don't know what it is.

Sure. It's the same solution as launch abort.

"Don't explode."

The FAA thinks that exploding during launch is fine, at least for a cargo or tanker mission.  They just care about debris landing on people.

Of course, we all know that.

The way SpaceX will reduce the risk of debris landing on people during reentry is to reduce the risk of Starship exploding during reentry.

"Make it so safe it can't fail" has not been a winning move for all other forms of transport, even ones operating in a more benign than EDL.