Starship On-orbit refueling - Options and Discussion

[Nevyn72]

are you trying to SSTO? Because that never works out as well as you think it does.

Not at all, the SH (tank) is a second stage using a conventional SS thrust structure (you could even delete the Sea level Raptors).

You use a conventional SH for the first stage.

[rakaydos]

...
Musk hates the idea of an orbiting tanker. What has been stated is one Starship fueled up by an additional 3 to 4 other Starship launches.
...

"Hate" seems a bit strong; do you have a cite?  Pragmatic is likely a more appropriate description.  Not to mention that Musk appears to be trying to stay out of the "no depots" debate.

I believe (my opinion, no citation) that the issue Musk has with an orbiting tanker is that it's located for where you need it initially but the desired location changes with each mission so quickly becomes useless.

My suggestion means that when needed you can reposition the unit to the next ideal staging orbit.

You will be leaving a SS in orbit in this role anyway but with the SS(tanker)-SS you are launching a whole lot of additional weight to allow the SS(tanker) to return to Earth, just so you can then re-launch the whole lot again to a different orbit for the next mission.

the "extra weight" of a tanker let's you use Earth's atmosphere to adjust inclination and Earth's resources to reset the rocket equation for each mission.

[aero]

I know that there is a very large, if not an infinite number of possible orbits for the tankers, but practically speaking, how many tankers would be needed to cover 90% of the orbits that would be used? For example, would the LEO parking orbit for all Mars-bound Starships be the same or does the parking orbit depend on the Earth/Mars locations so strongly that the Mars Transfer Orbit can't be efficiently adjusted during or after TMI? By efficiently adjusted, I mean "adjusted without serious impact to the payload or fuel reserves."

[rakaydos]

I know that there is a very large, if not an infinite number of possible orbits for the tankers, but practically speaking, how many tankers would be needed to cover 90% of the orbits that would be used? For example, would the LEO parking orbit for all Mars-bound Starships be the same or does the parking orbit depend on the Earth/Mars locations so strongly that the Mars Transfer Orbit can't be efficiently adjusted after launch? By efficiently adjusted, I mean "adjusted without serious impact to the payload or fuel reserves."

I recall reading that Earth Mars cyclers could reasonably adjust their orbits to match the double flyby every 6 synods. That would imply six reasonable ejection angles for a Mars transfer.
But what about the moon? What about outer solar system destinations? What about different launch inclinations to Mars for that matter.

[xvel]

How much propellant mass a tanker Starship could haul into LEO? Regular SS carries 1200 t of prop and 100 t of payload, so tanker SS would only carry 1300 t of prop? I mean, since it can carry more prop it can burn more prop, and it has internal space for carrying 2100 t of propellant without compromising fairing space for header tanks, batteries, actuators, etc. but could a booster even lift that up? and in the end, would a 2100 t prop tanker have more fuel left when it gets to LEO than say a 1300 t prop tanker?

This is a tanker with capacity for 2100 t:

A simple example with a tanker that takes 200t of fuel instead of 100t:

SH 280t dry, 3300t propellant
SS 220t dry (with payload), 1200t propellant
stack weight 5000t
final mass for SH 1700t

SH with SS assuming isp 350 (simplification, isp increases during flight when the ambient pressure drops) has dV 3,7km/s
SS assuming isp 380 has dV 6,95km/s

now with a tanker that has 200t of fuel as a load so it weighs 1520t
stack weight 5100t, final mass for SH 1800t

SH dV 3,57km/s (130m/s less and slightly more gravity drag)
tanker dV  must be 7,08km/s to compensate and you have to burn 1293t of fuel to do that, so you're left with 107t of fuel instead of 100t in orbit

and the more fuel you take, the more gravity drag (which is not taken into account in this calculation) eats

Orbital mechanics is brutal

Only real way to increase tanker capacity is to shave dry weight, or you need bigger booster (which MUST be larger diameter)

[50_Caliber]

A depot at EML-1 makes sense for Lunar missions, if they build a reusable Lunar lander with similar mass ratio to SS, then they can refuel at EML-1, go to the moon, land and come back ready to refuel for next mission. The SS has a 6km+ delta V, so a lander with similar delta V could be refueled with perhaps 3-4 tanker flights to EML-1 station to refuel the lander. It can make the trip to the moon, land and come back to EML-1 for around 5km/s delta V. If it uses methane\lox , then it's a pretty straightforward refueling exercise that meshes perfectly with the SS architecture. We bypass any need for ISRU mining on the moon and it fast-tracks lunar development.

This also gives us an excuse to make a legit 2001 space station at that point. 

[Nevyn72]

Orbital mechanics is brutal

Only real way to increase tanker capacity is to shave dry weight, or you need bigger booster (which MUST be larger diameter)

Interesting, so you take the first phase of In-orbit refueling to be SS to SS with a goal of having a delivered weight of around 100t of fuel. We already believe this will be the case so can use it as a starting point.

Now in my speculated second phase where you have a dedicated, and movable, In-orbit refueling tanker/depot you now have a starting target weight.

The question then becomes how much would my postulated SH (tank) weigh?
Would it be more or less than a fully functional SS with 100t of fuel?

Alternatively, if you stripped the aero-surfaces, landing structures, thermal tiles and sea-level Raptors off a SS how much less would it weigh and hence how much extra fuel could it carry to orbit?
(This option wouldn't give you the additional capacity of the SH (tank) but would function in the same way)

[zodiacchris]

Not quite sure where the SH as tanker second stage would be required. Wouldn’t it make more sense to simply build the tanker SS with extended tanks by fitting a few extra hoops in the tanks and deleting the payload bay? With the exception of flight relevant systems no fit out of the nose, no decks, air locks, life support, windows etc. That would be the most KISS solution and not change SS aerodynamics or mould line.

The weight savings of the reduced equipment plus the normal SS payload capacity would give you around 150 -200 tonnes of refuelling capacity. Send Tanker 1 up, refuel that from Tankers 2 -6, then launch the Mars SS last to take on the fuel from Tanker 1.

That avoids loiter time for the manned Mars SS in orbit and possible hang ups with bad weather.

Of course Elon did say a while ago that the tanker version would look weird, so yeah, what do we know?

[libra]

How much propellant mass a tanker Starship could haul into LEO? Regular SS carries 1200 t of prop and 100 t of payload, so tanker SS would only carry 1300 t of prop? I mean, since it can carry more prop it can burn more prop, and it has internal space for carrying 2100 t of propellant without compromising fairing space for header tanks, batteries, actuators, etc. but could a booster even lift that up? and in the end, would a 2100 t prop tanker have more fuel left when it gets to LEO than say a 1300 t prop tanker?

This is a tanker with capacity for 2100 t:

A simple example with a tanker that takes 200t of fuel instead of 100t:

SH 280t dry, 3300t propellant
SS 220t dry (with payload), 1200t propellant
stack weight 5000t
final mass for SH 1700t

SH with SS assuming isp 350 (simplification, isp increases during flight when the ambient pressure drops) has dV 3,7km/s
SS assuming isp 380 has dV 6,95km/s

now with a tanker that has 200t of fuel as a load so it weighs 1520t
stack weight 5100t, final mass for SH 1800t

SH dV 3,57km/s (130m/s less and slightly more gravity drag)
tanker dV  must be 7,08km/s to compensate and you have to burn 1293t of fuel to do that, so you're left with 107t of fuel instead of 100t in orbit

and the more fuel you take, the more gravity drag (which is not taken into account in this calculation) eats

Orbital mechanics is brutal

Only real way to increase tanker capacity is to shave dry weight, or you need bigger booster (which MUST be larger diameter)

Excellent post, really. Another way of putting it is as follow.

What is the tanker ?

Basically, it should be the usual Starship, its propellant mass, and its payload.

Then in the case of a tanker, the payload, obviously, should be the maximum amount of propellant you can haul into orbit by sacrificing everything but the kitchen sink, to rise payload.

That is, the cargo area becomes another fuel tank.

Then there is the problem of mass: what part of the "standard Starship"are you willing to sacrifice and turn into propellant mass, that is, useful tanker payload ?

Do you want a dumb prop tank inside the cargo area ? or do you start modifying or removing cargo area structure ?

There it depends of two things a) commonality with a "standard Starship" and b) reusability or not.

The more you sacrify commonality with the standard Starship, the more propellant raw mass you get into orbit, but of course tanker development costs rise...

An interesting historical exemple are the Boeing 707 innumerable variants, tankers included. While the KC-135 was a separate development of 707 from day one (different fuselage), there were also civilian 707s turned into makeshift, cheaper tankers - also far less efficient.
Well then the kerosene plumbery, the distribution of kerosene between the aircraft own tanks and the tanks dedicated to refueling, where vastly different.
Musk will face the same issues with tanker Starship.

Orbital mechanics is brutal

That's an understatement. The basic Starship has 1200 mt of propellant in the tanks when separating from SH, yet most of this vanishes during ascent. If it carries zero payload instead of 150 mt, then somewhat automatically, 150 mt of props remain in the tanks. That's a standard Starship. Then by scrounging whatever can be scrounged of structure and weight, maybe you can rise this to 200 mt. Maybe. Since Starship needs to be fully-fueled for Mars departure, then you need at least 6 and probably 7 tanker flights to fill that thing. Brutal numbers, sure.

How about starting from this number - 7 tankers - and start stockpiling 10 - 12 of them in orbit, docked to a dumb truss ? Would not be that expensive and would give margin.

But that's for 1 flight; Imagine for 1000 or more, the sheer number of tankers that will be needed...

As for different orbits with different inclinations - no panic. There is a simple formula to calculate plane changes delta-v
https://en.wikipedia.org/wiki/Orbital_inclination_change

Musk advantage here is the high specific impulse of his methalox engines. I can see "Starship depots" making reasonable plane changes by firing their engines.

[Ultrafamicom]

How much propellant mass a tanker Starship could haul into LEO? Regular SS carries 1200 t of prop and 100 t of payload, so tanker SS would only carry 1300 t of prop? I mean, since it can carry more prop it can burn more prop, and it has internal space for carrying 2100 t of propellant without compromising fairing space for header tanks, batteries, actuators, etc. but could a booster even lift that up? and in the end, would a 2100 t prop tanker have more fuel left when it gets to LEO than say a 1300 t prop tanker?

This is a tanker with capacity for 2100 t:

A simple example with a tanker that takes 200t of fuel instead of 100t:

SH 280t dry, 3300t propellant
SS 220t dry (with payload), 1200t propellant
stack weight 5000t
final mass for SH 1700t

SH with SS assuming isp 350 (simplification, isp increases during flight when the ambient pressure drops) has dV 3,7km/s
SS assuming isp 380 has dV 6,95km/s

now with a tanker that has 200t of fuel as a load so it weighs 1520t
stack weight 5100t, final mass for SH 1800t

SH dV 3,57km/s (130m/s less and slightly more gravity drag)
tanker dV  must be 7,08km/s to compensate and you have to burn 1293t of fuel to do that, so you're left with 107t of fuel instead of 100t in orbit

and the more fuel you take, the more gravity drag (which is not taken into account in this calculation) eats

Orbital mechanics is brutal

Only real way to increase tanker capacity is to shave dry weight, or you need bigger booster (which MUST be larger diameter)

On the other hand SH now needs 130m/s less dv for boostback and stages closer to launch pad, so more propellant can be used to boost SS, which gives a further increase in payloads.

[Nevyn72]

I think there needs to be a distinction made between the tanker that remains in orbit and stockpiles fuel for the mission SS and the tanker that is used to transport fuel to the orbital stockpile.

The one that remains in orbit just needs to be delivered there (whether it is designed to return to Earth or not) and the amount of fuel it retains on arrival in orbit is largely academic. Design for the function becomes the priority.

The one(s) that deliver the fuel to orbit need to be optimised to maximise the fuel delivered and expedite launch cycle frequency, that becomes the design priority for these units.

The differing requirements mean that each unit are not necessarily identical in design.

[Eka]

I've never been able to pencil a scenario where an orbital fuel depot is better than just sending up the SS tankers. It all comes down to flexibility, the cost of putting that infrastructure into space, and how refueling is done. That last one, how refueling is done, is the killer. Once part or all of the propellants are available from outside Earth's gravity well, then the math may change, but not until then, and not always.

On how in orbit refueling is done. With the current Starship design, it MUST be done under acceleration. You must take that into account for any system you develop. Will that inflatable tank be able to handle the repeated accelerations and decelerations needed? Think of how much more mass you are accelerating and decelerating when you have a propellant load in that depot greater than than one SS can receive.

An fuel depot who can be one tanker or an dedicated tanker with an sort of sun shade, think an air mattress on an beam makes sense in that you can top it up during slow days and the Mars or Moon mission only need to refuel once.

Acceleration is needed to settle the fuel at the bottom of the tank, this does not to be an high acceleration, The dedicated fuel depot tanker will have pumps doing most of the work the trust is just to have the liquids settle at the bottom.

With a pump, you need enough acceleration and propellant head or tank pressure to keep the pump from cavitating.* Might as well just use a bit of acceleration, and tank pressure alone. Keep it as simple as possible. One thing to remember is both the source tank, and the receiving tank need the acceleration. The propellant must also be settled in the receiving tank to vent the gasses being displaced by the fuel. Want to transfer propellant without acceleration, one still needs the pressure, but another form of containment is needed. Fuel bladders in both source and destination tanks would work, but they add weight and can fail. Yes, many gotchas for fuel transfer in orbit. Gravity makes a lot of things easier here on earth.

* This is why 6 bar is the design goal for SS tanks. They need that pressure to keep the Raptor turbopumps from cavitating as they suck in huge amounts of fuel.

[Ionmars]

To synthesize from previous posts, I agree that the tanker version of SS has the  function of delivering maximum feasible propellant to LEO. We should minimize structural mass and eliminate any feature that does not contribute to this function. It could be perceived in stages:

First, consider the standard SS with extra propellant tanks in the cargo area. We note that cargo SS features a large payload volume to accommodate a variety of payloads, such as equipment that occupies only part of the volume. In this case the mass carried in the cargo bay leaves some empty space in the cargo bay.  Liquid propellant, however, could occupy every cc of available cargo volume, provided propellant density allows propellant mass to be less than cargo mass capacity of 1100 tonnes. (I am no longer adept to calculate the excess volume.)

Second iteration, shorten tanker SS cargo bay. Determine the exact length required to carry the mass of propellant that is (edit:) to be transferred to another SS. The resulting mass savings allows propellant load to increase by an equal amount.

Third iteration, eliminate the cargo bay. Increase the length of main propellant tanks to accommodate both the fuel required to reach LEO and the fuel to be transferred.

Fourth iteration, increase diameter of fairing. Note that the most efficient shape to carry liquid is a sphere, which has less surface area than a cylinder for equal volume. In this step we would widen the nose section of tanker SS to be more like a sphere. It would be similar to the wide fairings used for some missions of Falcon 9, only more so. Tanker SS would assume the appearance of a Lollypop, which may be why Musk has said tanker SS would look kinda weird.

[rakaydos]

To synthesize from previous posts, I agree that the tanker version of SS has the  function of delivering maximum feasible propellant to LEO. We should minimize structural mass and eliminate any feature that does not contribute to this function. It could be perceived in stages:

First, consider the standard SS with extra propellant tanks in the cargo area. We note that cargo SS features a large payload volume to accommodate a variety of payloads, such as equipment that occupies only part of the volume. In this case the mass carried in the cargo bay leaves some empty space in the cargo bay.  Liquid propellant, however, could occupy every cc of available cargo volume, provided propellant density allows propellant mass to be less than cargo mass capacity of 1100 tonnes. (I am no longer adept to calculate the excess volume.)

Second iteration, shorten tanker SS cargo bay. Determine the exact length required to carry the mass of propellant that is feasible to launch to LEO. The resulting mass savings allows propellant load to increase by an equal amount.

Third iteration, eliminate the cargo bay. Increase the length of main propellant tanks to accommodate both the fuel required to reach LEO and the fuel to be transferred to another SS.

Fourth iteration, increase diameter of fairing. Note that the most efficient shape to carry liquid is a sphere, which has less surface area than a cylinder for equal volume. In this step we would widen the nose section of tanker SS to be more like a sphere. It would be similar to the wide fairings used for some missions of Falcon 9, only more so. Tanker SS would assume the appearance of a Lollypop, which may be why Musk has said tanker SS would look kinda weird.

This all begs the question, why?

A tanker exists to refuel Starship. Therefore the most fuel it needs at any one time is enough to fully fuel one starship. Is there an existing ship with tanks that size, one you're already mass producing? Well, yes... Starship itself.

Any permanant depot must consider the trade between reentry mass, which allows the rocket to come home and resupply, and plane chance fuel, which must be lifted to orbit and docked to the existing depot in order to move the depot and any residual fuel to the new orbit.

If a plane change requires more than a mass ratio of 1- that is, if the mass of fuel needed for a useful "next plane" change of orbit is more than the (already reduced, by the removal of reentry mass) dry mass of the tanker, without fuel, it's easier to launch an entirely new tanker in the new orbit than move the old one.

[Ionmars]

This all begs the question, why?
...
...

The reason that first comes to mind is an insurance policy. If there are unexpected fuel leaks during launch to orbit or minor spills during fuel transfers, we would still be able to deliver enough propellant over the planned number of launches to complete the mission.

Nevertheless, your point is well taken and SS tanker version is low priority.

Added: We may be addressing two different types of “tanker.” I am addressing the version of SS used to deliver propellant to orbit where it transfers fuel to another SS in orbit or to a “tanker-in-orbit,” which is a substitute for a propellant depot. I would call the tanker-in-orbit an SS Depot or some different name to avoid confusion.

[ZChris13]

...
Musk hates the idea of an orbiting tanker. What has been stated is one Starship fueled up by an additional 3 to 4 other Starship launches.
...

"Hate" seems a bit strong; do you have a cite?  Pragmatic is likely a more appropriate description.  Not to mention that Musk appears to be trying to stay out of the "no depots" debate.

I believe (my opinion, no citation) that the issue Musk has with an orbiting tanker is that it's located for where you need it initially but the desired location changes with each mission so quickly becomes useless.
My suggestion means that when needed you can reposition the unit to the next ideal staging orbit.
You will be leaving a SS in orbit in this role anyway but with the SS(tanker)-SS you are launching a whole lot of additional weight to allow the SS(tanker) to return to Earth, just so you can then re-launch the whole lot again to a different orbit for the next mission.

Is that an issue?

[dnavas]

If a plane change requires more than a mass ratio of 1- that is, if the mass of fuel needed for a useful "next plane" change of orbit is more than the (already reduced, by the removal of reentry mass) dry mass of the tanker, without fuel, it's easier to launch an entirely new tanker in the new orbit than move the old one.

Just so.  In LEO there are a large number of possible waypoints, each of them mutually "inaccessible."  The question is whether there are a small enough "significant" set that the idea is tenable.  Because it's likewise a small iteration from fuel depot to electric LEO/GEO tug dock, but that iteration only makes sense if there are reasonable targets for LEO.  If only we had SST-lagrange point, but we don't.

Is there some way to determine, given a single SS tanker load of fuel, what LEO coverage exists for about 100mt of fuel?

[wannamoonbase]

Only real way to increase tanker capacity is to shave dry weight, or you need bigger booster (which MUST be larger diameter)

I would not be surprised to see SH have a larger diameter than SS. 

The engines under the 9 meter SH have been crammed in and little or no room for growth.

The number of refueling flights seem like they could be reduced with a larger booster or more staging (3rd stage or FH style.)

If these two vehicles are going to be so cheap to build then why not have a different diameter for each? 

Edit: If the vehicles can eventually be as cheap to build and easy to operate as projected then the cost of fuel for each launch drives cost then at some point there are more efficient ways to get 100 tons of fuel on orbit. 

I suppose there is a long way to go and there is plenty of time to optimize.

[TrevorMonty]

If SS needs refuelling it is going to GEO, moon or mars. Which should reduce number of possible tanker/depot orbits.

[aero]

If SS needs refuelling it is going to GEO, moon or mars. Which should reduce number of possible tanker/depot orbits.

Remember, as pointed out above, plane change delta-V is linear with velocity. So if it is going beyond LEO then the plane change maneuver becomes cheaper the lower the velocity of the Starship. My orbital mechanics is beyond rusty but I think I recall that doing a plane change maneuver during the Mars transfer is the most economical. That means all this discussion of doing LEO maneuvers to calculate plane change delta-V is a sub-optimal topic.