Starship On-orbit refueling - Options and Discussion

[edzieba]

The HSS Starship doe snot move the tanks up into the nose. the nose there is where the habitation and cargo areas are. The HSS Starship significantly lengthens the overall stack to achieve the greater tank capacity.

The reason that's a problem for a depot variant is that the HSS Starship also omits the flaps and the ability to re-enter and land. That means once you launch it, you don't get it back and can't use it for anything else. If you want to make a similarly stretched vehilce but add the flaps and TPS back on, you now have a bunch of unique structural work (longer vehicle 'spine'), EDL behaviour modelling, and testing, for one very limited use variant, since this stretched depot is only ever actually  needed[/i[ for HSS.
On the other hand, two regular-sized Starships used as depots with prop transfer hardware in the cargo bay only need to play depot for a week or two per year to handle HSS Starship, and the rest of the year can spend their time launching payloads just like any other Starship. They are also the right size for any regular Starship that needs in-orbit prop transfer (e.g. for direct GTO injection).

Or using your numbers: instead of spreading making 20x, 10x and 10x of 3 different variants, you make 40x of one variant and keep and modifications within the payload bay.

[OTV Booster]

Largely agree here, although I still think that we're going to see an LSS with 1500t-1600t tanks, and you might as well use that layout for lift tankers and depots.  But if everything remains at 1200t, then yes, you could conceivably have a single propulsion section layout that was good for everything.

The Lunar Starship is the only vehicle that would need more than one 'standard' Starship's worth of propellant to be transferred, and those will be launching maybe once a year, if that. The rest of the time, a standard capacity Starship is sufficient as a tanker or depot for all regular launches (albeit most launches will likely require zero tankers, let alone a depot).
Rather than using a custom stretch depot Starship that may-or-may-not be able to ever deorbit (and even if it can, that's a lot of extra R&D work to make sure it actually works) and require s abunch of unique handling equipment and procedures to accommodate its greater size; for that once-a-year occasion that actually requires that capacity you can launch two regular sized depots instead.

In terms of cost and time, it's always preferable to make more copies of the same item than to make more one-off custom item variants whenever you can possibly help it.

The 1600t tanker can work without a change to the OML. The common dome moves up and the upper dome moves up into what would otherwise be the cargo bay. All associated plumbing and wiring does need to stretch to match. Total tank stretch would be 6.4m for boiling props. A bit less for sub chilled. Landing characteristics should be very little different than an empty SS of standard internal dimensions.


There is one issue that might make a 1600t tanker unworkable. The static head pressure at the bottom of the tanks will be higher than normal. Ullage pressure and acceleration loading will need to be rethought. It's possible that this has already been baked into the design. Maybe it can be handled by ullage pressure and the acceleration profile. If it calls for a major redesign of the tanks and domes it probably won't happen. At least not until SSv2.


On the issue of special builds, it seems more of a non issue. If your building 10k widgets a day customizing even one is a big deal. The automotive industry routinely works with 'build sheets' that show every option available. The build sheet for a semi truck goes on for pages. It's low volume, almost kit like, assembly.


If SS were built the same way as SLS I'd agree with you on this. If the Pez dispenser works out we'll see that as a cargo variant with a very specific build sheet and IMO we'll see any other variant that's useful and can build on the core design. An unexpected benefit of rapid iteration at this level of sophistication is an unprecedented ability to be easily modify. What we have today is SN1 modified inch by inch and step by step.

[TheRadicalModerate]

The HSS Starship doe snot move the tanks up into the nose. the nose there is where the habitation and cargo areas are. The HSS Starship significantly lengthens the overall stack to achieve the greater tank capacity.

Do you have a source for this?

Crew area doesn't need to be in the cylindrical section.  Indeed, the bottom of the cylindrical section needs to be in vacuum, because that's where the cargo hatch, airlocks, and elevator have to be.

You only need about 2m of cylindrical height to do this.  And the ogive portion of the nose provides more than enough volume to support the pressurized crew module, especially since there are no header tanks.

That gives you at least 6m of room for the LCH4 dome to be moved forward.  At 3.6:1 O:F and using boiling densities, you get an average density of 885kg/m³.  So 6m of extra tankage gives you about 340t more prop than the vanilla tankage, i.e., 1540t total.

The reason that's a problem for a depot variant is that the HSS Starship also omits the flaps and the ability to re-enter and land. That means once you launch it, you don't get it back and can't use it for anything else. If you want to make a similarly stretched vehilce but add the flaps and TPS back on, you now have a bunch of unique structural work (longer vehicle 'spine'), EDL behaviour modelling, and testing, for one very limited use variant, since this stretched depot is only ever actually  needed for HSS.

But this argument is based on the assumption that the outer mould line is really stretched.  I don't think it is.  You can put header tanks into a long propulsion section with no problem, and you'll wind up with exactly the same balance and dynamics as the short propulsion version during EDL, when the mains are mostly empty.  The only time the CoM will be substantially different is when the thing is loaded to the gills with propellant--but that only happens in vacuum, where you don't have aerodynamic issues.

On the other hand, two regular-sized Starships used as depots with prop transfer hardware in the cargo bay only need to play depot for a week or two per year to handle HSS Starship, and the rest of the year can spend their time launching payloads just like any other Starship. They are also the right size for any regular Starship that needs in-orbit prop transfer (e.g. for direct GTO injection).

Or using your numbers: instead of spreading making 20x, 10x and 10x of 3 different variants, you make 40x of one variant and keep and modifications within the payload bay.

Again, this is all based on the eloneron dynamics of the short and long propulsion sections being very different.  If the dynamics are the same, then, since you've already conceded that the LSS version needs the long propulsion module, why wouldn't you use it and save yourself the cost of an extra depot (even if the depot turns out to be an ordinary lift tanker, which I think we both think will be more likely than not)?

The other thing a long propulsion tanker allows is that you can easily play with how much you fill it for launch, optimizing usable prop delivered to orbit.  There's no fooling with auxiliary tankage in the payload bay, which would need to be vented and managed at launch.  Just fill the tanks a little more or less until you find the optimum.  Then every time the Raptor people manage to eke out a few extra kN from each engine, you can just fill the lift tanker a bit more.

You can't do that with a 1200t tank.  If the thrust increases, you can still only put 1200t into it, not 1220t or 1280t or 1300t.  But if the tank will take up to 1500t, then all of those attempts to optimize things are easy.

PS:  There's also nothing that prevents you from supporting a short nose for third-party payloads.  Even with only 2m of cylindrical height, the ogive is so huge that any satellite that was designed to be dual-sourced for launch would be well inside the static envelope.

So if you're really worried about some horrible imbalance between long and short propulsion modules, use the long ones for payload when they're not needed for tanking or LSS duties.  Or expend them for StarKicker missions.

[Bryan Hayward]

2) Nose Section (long goes on short propulsion, short on long propulsion):
...
e) Jettisonable fairing for deep space expendable mission (probably short nose)

From these you make:

D) Deep-space "StarKicker"
...
Short jettisonable fairing nose
...

 (... is snippage)
One nit I am going to ask about.  I realize that the fairing is likely to be extra mass you don't want to push around. However, it is another layer between the crew and space hazards. Do we really want to lose it?

Otherwise, I agree it seems reasonable that making versions of SS is relatively simple, once the base version is proven. The one exception may be the docking structure/mechanism on the depot as that may not be restricted to the nose, per other posts in this thread.

[Paul451]

e) Jettisonable fairing for deep space expendable mission (probably short nose)

I realize that the fairing is likely to be extra mass you don't want to push around. However, it is another layer between the crew and space hazards. Do we really want to lose it?

Unless crew count as "expendable", I'm assuming he's talking about BEO robotic missions.

doe snot

Oh dear.

[Barley]

There is one issue that might make a 1600t tanker unworkable. The static head pressure at the bottom of the tanks will be higher than normal.

Pressure at the bottom of the tank depends mostly on the thrust and not very much on the amount of fuel.  Unless you add engines when you lengthen the tank this should be manageable.

The easy way to think about this is that pressure is force per unit area, and all the force comes from the engines.

The hard way us that the extra mass reduces the acceleration, which reduces the head.

[TheRadicalModerate]

doe snot

Oh dear.

Shouldn't that be "Oh deer"?

[OTV Booster]

There is one issue that might make a 1600t tanker unworkable. The static head pressure at the bottom of the tanks will be higher than normal.

Pressure at the bottom of the tank depends mostly on the thrust and not very much on the amount of fuel.  Unless you add engines when you lengthen the tank this should be manageable.

The easy way to think about this is that pressure is force per unit area, and all the force comes from the engines.

The hard way us that the extra mass reduces the acceleration, which reduces the head.

Yes, the greater load, the lower the T/W and the lower the acceleration which lowers the head. But, increasing prop load by 33% doesn't decrease acceleration by 33% unless you have a magical zero dry mass.


F=MA. Engines supply force. Propellant supplies mass. Head at the bottom very much depends on the mass of the propellant. The only reason we have static head pressure is because earth supplies a constant 1g acceleration.

[TheRadicalModerate]

There is one issue that might make a 1600t tanker unworkable. The static head pressure at the bottom of the tanks will be higher than normal.

Pressure at the bottom of the tank depends mostly on the thrust and not very much on the amount of fuel.  Unless you add engines when you lengthen the tank this should be manageable.

The easy way to think about this is that pressure is force per unit area, and all the force comes from the engines.

The hard way us that the extra mass reduces the acceleration, which reduces the head.

Yes, the greater load, the lower the T/W and the lower the acceleration which lowers the head. But, increasing prop load by 33% doesn't decrease acceleration by 33% unless you have a magical zero dry mass.


F=MA. Engines supply force. Propellant supplies mass. Head at the bottom very much depends on the mass of the propellant. The only reason we have static head pressure is because earth supplies a constant 1g acceleration.

Just do the math.  p = ρah.

You don't have to worry about loads at launch, because the stretched tanks won't be full.  In microgravity, I get a worst case 5.6kPa difference.  It's nothing like what the tanks would be subjected to, even with only 1200t of prop, at launch.  Everything's linear, and you know that the hydrostatic pressure drops to near zero at burnout, so the worst case is always going to be at startup with full tanks.

Here's an example:

[Bryan Hayward]

e) Jettisonable fairing for deep space expendable mission (probably short nose)

I realize that the fairing is likely to be extra mass you don't want to push around. However, it is another layer between the crew and space hazards. Do we really want to lose it?

Unless crew count as "expendable", I'm assuming he's talking about BEO robotic missions.

I was thinking "expendable" in terms of SS means no re-use of the ship, not necessarily a statement about the lack of return of the whole mission or all its components. But I do hope that we'll be able to figure out crew protection on long haul trips.

[TheRadicalModerate]

I was thinking "expendable" in terms of SS means no re-use of the ship, not necessarily a statement about the lack of return of the whole mission or all its components. But I do hope that we'll be able to figure out crew protection on long haul trips.

I'd expect a crew nose that can handle launch and EDL would evolve to handle at least trips to Mars and back.  Anything beyond that is probably science fiction for a while.

The "StarKicker" idea has been broached by Elon a few times, as a way to put heavy interplanetary (or even extrasolar) probes into the fastest transfers possible.  It doesn't necessarily have to have a jettisonable fairing, but it reduces the dry mass a bit and may have a couple of other advantages that are even more off-topic.  I'd expect SpaceX to do this eventually, but it's probably pretty low on the list--unless NASA drops by and says, "Hey, could you help us get 20t to Neptune right now?"

[OTV Booster]

Had an afternoon of profitable thinking. It brought on a prediction.


In 10 years, third party retanking will be a measurable portion of SX's SS 'for hire' launch income.


Here's my thinking. Given the current lamentable shape the world is in, the earth imaging market is bull. One major failing of sat imaging is that it is predictable. Activities can be timed around the sat's schedule and avoid exposure.


The biggest of these sats carry a good load of propellants and do have maneuverability but dV is limited and is a carefully managed resource. Refueling changes the game. Opinion: Once SX demonstrates on orbit transfer some agencies will want to have some serious discussions with them.


This says nothing directly about the technical issues we've been discussing. "nothing directly." If I hit the mark up above it changes assumptions and opens up the trade space. Not right away. But maybe in 5-8 years as a new generation of sats come on line with retanking baked in.

[OTV Booster]

There is one issue that might make a 1600t tanker unworkable. The static head pressure at the bottom of the tanks will be higher than normal.

Pressure at the bottom of the tank depends mostly on the thrust and not very much on the amount of fuel.  Unless you add engines when you lengthen the tank this should be manageable.

The easy way to think about this is that pressure is force per unit area, and all the force comes from the engines.

The hard way us that the extra mass reduces the acceleration, which reduces the head.

Yes, the greater load, the lower the T/W and the lower the acceleration which lowers the head. But, increasing prop load by 33% doesn't decrease acceleration by 33% unless you have a magical zero dry mass.


F=MA. Engines supply force. Propellant supplies mass. Head at the bottom very much depends on the mass of the propellant. The only reason we have static head pressure is because earth supplies a constant 1g acceleration.

Just do the math.  p = ρah.

You don't have to worry about loads at launch, because the stretched tanks won't be full.  In microgravity, I get a worst case 5.6kPa difference.  It's nothing like what the tanks would be subjected to, even with only 1200t of prop, at launch.  Everything's linear, and you know that the hydrostatic pressure drops to near zero at burnout, so the worst case is always going to be at startup with full tanks.

Here's an example:

Hmm did a sanity check. Took you're numbers for the weight for a cubic meter of methane, and the height of the tank. Multiplied by 1.5, assuming that is the T/W. Then consulted my oracle for 1bar in kg/m^2 (10197.16) and divided by that. Got  a .6bar head looking at the methane. A higher number but within BOE spitting distance. I was assuming it would me much higher. This is low enough that it retires my worries about pressure management. Nowhere near the margins.


Hmmm, Wait. Don't we need ~4bar for the pump inlets? Well, that margin just shrank.

[TheRadicalModerate]

Hmm did a sanity check. Took you're numbers for the weight for a cubic meter of methane, and the height of the tank. Multiplied by 1.5, assuming that is the T/W. Then consulted my oracle for 1bar in kg/m^2 (10197.16) and divided by that. Got  a .6bar head looking at the methane. A higher number but within BOE spitting distance. I was assuming it would me much higher. This is low enough that it retires my worries about pressure management. Nowhere near the margins.

Hmmm, Wait. Don't we need ~4bar for the pump inlets? Well, that margin just shrank.

You don't care about T/GLOW when looking at Starship (not SuperHeavy) hydrostatics.  You care about burnout T/W for the SuperHeavy.  That's something that's almost certainly acceleration-limited.  3g?  5g?  Whatever it is, it's way, way more acceleration than a full Starship will generate at startup.

If you need 4bar for the pumps, then you can reduce ullage pressure by however much the stretch in the tanks increases the hydrostatic head.  Bottom line:  The tanks are already strong enough to accommodate the stretch, as long as they're not filled all the way on the pad--which they can't be, if you want a reasonable T/GLOW.

[TomH]

One major failing of sat imaging is that it is predictable. Activities can be timed around the sat's schedule and avoid exposure.
......propellants.....dV is limited..... Refueling changes the game. Opinion: Once SX demonstrates on orbit transfer some agencies will want to have some serious discussions with them...... maybe in 5-8 years as a new generation of sats come on line with retanking baked in.

Well, an alternative to that might be that SS makes spy birds become as plentiful as Starlinks: there's always a dozen or so over every part of Earth, moving in multiple directions. No maneuvering and retasking then required. Of course, eventually LEO will become like a pinball machine, the debris field so dense that nothing can penetrate in or out without severe damage.

[OTV Booster]

One major failing of sat imaging is that it is predictable. Activities can be timed around the sat's schedule and avoid exposure.
......propellants.....dV is limited..... Refueling changes the game. Opinion: Once SX demonstrates on orbit transfer some agencies will want to have some serious discussions with them...... maybe in 5-8 years as a new generation of sats come on line with retanking baked in.

Well, an alternative to that might be that SS makes spy birds become as plentiful as Starlinks: there's always a dozen or so over every part of Earth, moving in multiple directions. No maneuvering and retasking then required. Of course, eventually LEO will become like a pinball machine, the debris field so dense that nothing can penetrate in or out without severe damage.

There's a lot that can be done with a lot of small sats. What they can't do is the resolution that comes with a 3m optical system. That's expensive. SX can drop the cost of launch but the glass is what it is.

[Bryan Hayward]

(snip)
The "StarKicker" idea (snip)  I'd expect SpaceX to do this eventually, but it's probably pretty low on the list--unless NASA drops by and says, "Hey, could you help us get 20t to Neptune right now?"

I'd say the most likely and important need for 20t in a hurry would be to "Asteroid X orbit" (where X is a heretofore undetected extinction level event sized body) for a re-direct mission. If I were planetary defense authority, that's what I would be asking someone deeply familiar with SpX intentions to consider. Paying for the difference between a Starkicker and their Mars concept would be considerably cheaper than making a bespoke system from the ground up (a la SLS).

[Robotbeat]

To deflect an extinction level asteroid with Starship would require a launch capacity comparable to what the full fledged Martian city would need, like at LEAST 1 million tons to orbit per year, if not 100 million or 1 billion. Even if you’re using nukes, it’d require quite a lot of lift.

[Zed_Noir]

To deflect an extinction level asteroid with Starship would require a launch capacity comparable to what the full fledged Martian city would need, like at LEAST 1 million tons to orbit per year, if not 100 million or 1 billion. Even if you’re using nukes, it’d require quite a lot of lift.

With sufficient lead time. Someone can build missiles using Starkicker frames armed with MT nuke boosted with about 20 tonnes of Tritium each. Drastically reduces the number of nukes and thus spacecrafts (lift) required especially if the missiles are use in salvos with simultaneous initiation of the payloads for a focused cone blast.

[Bryan Hayward]

To deflect an extinction level asteroid with Starship would require a launch capacity comparable to what the full fledged Martian city would need, like at LEAST 1 million tons to orbit per year, if not 100 million or 1 billion. Even if you’re using nukes, it’d require quite a lot of lift.

There are a lot of imponderables here, such as asteroid (or comet) mass/composition/orbit, but I'm assuming for the sake of in-space refueling discussions that getting out to "X" in a hurry is the point.  That means refueling as fast and as much as possible to get there in time for an infinitesimal nudge to matter. Would liquid refueling even be the point? Is Starkicker a robust solution for "get there ASAP" (neglecting the chances of mission success, admittedly small)?

That got me started on a different idea, which I've sent to Advanced Concepts so as not to send us off-topic.
https://forum.nasaspaceflight.com/index.php?topic=56871.0