SpaceX to the moon: mission profiles

[sevenperforce]

During the Apollo program, a variety of mission architectures were considered, including earth-orbit rendezvous, earth-orbit rendezvous with assembly, direct ascent, and finally LOR. There were other sub-parts, however; crasher stages, uncrasher stages, and so forth. A lot of these have already been discussed here or there on these forums.

Falcon 9 and Falcon Heavy aren't designed for prodigious BLEO performance, so they are automatically less than ideal for a lunar landing. Dragon V2 was designed as an LEO crew ferry with the possibility of Martian landings. SpaceX has no direct intentions for the moon; Elon has been pretty vocal about setting his sights on Mars. However, if SpaceX could support a manned lunar landing using the platforms they already have (or will have in the next year), they would have a substantial headstart on other options and be able to roundly demonstrate the legitimacy and flexibility of their platform. So I wanted to run some numbers and see what would be required for SpaceX to put men back on the moon.

A couple of things I eliminated right off were on-orbit assembly and on-orbit refueling. Neither have ever been used for a transfer vehicle, and I don't want to propose completely untested mission steps. Plus, those sorts of steps encourage an unrealistically large number of launches and aren't limiting at all. "Let's launch eight Falcon Heavys, use propellant transfer in orbit, then strap the full ones together! Wait, no, let's launch forty Falcon Heavys and take the whole ISS to the moon!" Not realistic.

I'm also eliminating ridiculously dangerous stuff like slapping landing legs on a Falcon 9 upper stage and setting it down on its tail. No matter how good SpaceX gets at suicide-burn Falcon 9 RTLS recovery, it's way too risky to consider for a manned landing on another world.

This makes lunar ascent a challenge, because you need a minimum of 2 km/s to get off the moon. No way around it.

The pressure-fed kerolox Kestrel engine used on the upper stage of Falcon 1 has a decent specific impulse. But trying to use it would require SpaceX to design a completely new platform for what they consider to be a dead-end mission profile, which won't really work. This means the only engines potentially available for ascent are the SuperDracos on the Dragon V2. Now, this isn't too far-fetched; Elon has spoken specifically about the use of the Dragon V2 platform for landing on a variety of worlds, including the moon, suggesting that the heat shield and parachutes could be exchanged for a larger fuel tank to increase dV.

Unfortunately, the SuperDracos on the Dragon V2 are extremely under-expanded, resulting in a really pathetic specific impulse of 240 seconds. Getting 2 km/s out of such low-impulse engines requires a propellant mass fraction of 57%, which means an empty Dragon V2 (with a dry mass of 4.2 tonnes) would need at least 5.6 tonnes of propellant in order to get off the surface of the moon, 370% of its current fuel capacity.

However, the SuperDracos run on MMH/NTO, which has an optimal vacuum ISP of 336 seconds. Theoretically, then, SpaceX should be able to design a SuperDraco Vacuum engine with an extended exhaust bell capable of operating in a vacuum at a much higher expansion ratio. Even though the engines are pressure-fed, they have a combustion chamber pressure of 1000 psi, which should allow maximum specific impulse. The thrusters are mounted at an angle of 15 degrees, but if the extended exhaust bells can angle that downward by half the angle, that's an effective specific impulse of 362 seconds and a propellant mass fraction of 43%.

Thus, retrofitting a Dragon V2 frame with larger tanks and extended-nozzle SuperDraco Vacuum engines should allow it to act as a Lunar Dragon descent or ascent vehicle. This means that a variety of mission profiles can be generated using only Falcon 9, Falcon Heavy, and the Dragon V2 platform.

So far, I've come up with the following:

1. Dual series launch to double LOR. Falcon Heavy with unmanned Lunar Dragon to LLO, followed by Falcon Heavy with stock manned Dragon V2 to LLO. Orbital rendezvous; crew transfers to Lunar Dragon and uses the first Falcon Heavy's upper stage as a crasher stage to the lunar surface and the Lunar Dragon hovers to a landing. Lunar Dragon then makes ascent, docks with the stock Dragon V2 for crew transfer, and the second Falcon Heavy upper stage returns the stock Dragon V2 to Earth.
2. ISS to EML-1; crasher-stage direct ascent. Single Falcon 9 launch takes crew and Dragon V2 to ISS; docks. Falcon Heavy lifts Lunar Dragon to the ISS; crew and consumables transfer; Falcon Heavy transfers to EML-1 and then burns crasher-stage to drop a Lunar Dragon with additional fuel on the lunar surface. Lunar Dragon makes ascent and transfer to ISS and crew returns to earth via docked Dragon V2.
3. Single launch to direct ascent. Standard Dragon V2 is retrofit with detachable nozzles and an internal auxiliary tank; launches with crew on a single Falcon Heavy to EML-1 and then is delivered via crasher stage to the lunar surface. Dragon V2 makes direct ascent to Earth aerobraking trajectory.

All are within dV limitations.

Any other possibilities?

[redliox]

SpaceX is probably considering some Moon plans, although I'm pretty sure they favor Mars more considerably.

Considering they'll have better experience in LEO, they could have something like a "traditional" D2 rendezvous with a Lunar Dragon to transfer crew (and maybe fuel) before essentially flying the Lunar Dragon for the remainder.  So my guess would be at least 3 launches: 1 apiece for crew, lunar vehicle, and fuel.  Anything cargo could be 2 launches sending a one-way vehicle.

SpaceX and the other commercial companies might wait until NASA makes a decision on if they want a new commercial vehicle, particularly for either Lunar or Martian landings, before investing further than LEO activity and crewless launchers.

[nadreck]

Orbital propellant transfer is on the critical path to Mars for SpaceX so I don't think it is fair to rule it out.

More than 20t could be delivered to the Lunar surface using an on orbit refueled Falcon upper stage in either crasher mode or with legs. I think the lowest risk lunar program with a reasonable cost using Falcon hardware, might involve lunar surface rendezvous with the return vehicle and most surface use equipment brought by 20t loads either in crasher stage mode or with landing legs and that the crew(s) initially be brought by one time dragon landers.

However, that is for a limited mission that someone other than SpaceX foots the ~$500M bill. And that is for the program up to the first crew.

It makes even more sense to wait for a raptor based architecture either with the MCT or the potential Falcon/Raptor upper stage.  With either of those a more substantial first mission for a similar price tag could be run.

I am all for 3 priorities for missions to the moon:

Polar ISRU, relatively low polar orbit station/depot/gateway along with high inclination communications constellation, EML-2 station/observatory and deep space communications node.

[IainMcClatchie]

If Falcon Heavy gets crossfeed, can a Lunar Dragon launched by FH land on the moon with a trunk and some useful payload in it?

[Rocket Science]

I toyed with this a while back for fun, have a look!

http://forum.nasaspaceflight.com/index.php?topic=30567.0

[launchwatcher]

1. Dual series launch to double LOR. Falcon Heavy with unmanned Lunar Dragon to LLO, followed by Falcon Heavy with stock manned Dragon V2 to LLO. Orbital rendezvous; crew transfers to Lunar Dragon and uses the first Falcon Heavy's upper stage as a crasher stage to the lunar surface and the Lunar Dragon hovers to a landing. Lunar Dragon then makes ascent, docks with the stock Dragon V2 for crew transfer, and the second Falcon Heavy upper stage returns the stock Dragon V2 to Earth.

open questions this inspires:

What's the on-orbit endurance of the Falcon upper stage?   How fast will its LOX boil off?   Will the mvac engine be able to restart after several days in vacuum?

Will the combined Dragon + upper stage have enough maneuverability to be able to dock with another Dragon?

[sevenperforce]

 

Considering they'll have better experience in LEO, they could have something like a "traditional" D2 rendezvous with a Lunar Dragon to transfer crew (and maybe fuel) before essentially flying the Lunar Dragon for the remainder.  So my guess would be at least 3 launches: 1 apiece for crew, lunar vehicle, and fuel.

SpaceX and the other commercial companies might wait until NASA makes a decision on if they want a new commercial vehicle, particularly for either Lunar or Martian landings, before investing further than LEO activity and crewless launchers.

I don't see anything more than two launches per mission, at least as far as essential components are concerned. A two-launch manned landing with a third launch for auxiliary/optional support payload is possible.

SpaceX has lunar contracts already. These are probably going to primarily be LLO missions, but some could be one-way unmanned surface missions using the Dragon V2 platform. So if SpaceX can generate a near a terms realize able mission profile for a manned return, I think they will.

Orbital propellant transfer is on the critical path to Mars for SpaceX so I don't think it is fair to rule it out.

Perhaps not rule it out entirely, but that would be something to test on a second manned mission with a secondary payload, rather than a first-time thing. That's another value of SpaceX servicing manned lunar landings: it gives them practice for what they will need to go to Mars.

If Falcon Heavy gets crossfeed, can a Lunar Dragon launched by FH land on the moon with a trunk and some useful payload in it?

By my calculations, FH full thrust with crossfeed can deliver 73 tonnes to LEO, 28 tonnes to EML-1, 20 tonnes to LLO, or 12 tonnes to the surface of the moon as a crasher stage.

1. Dual series launch to double LOR. Falcon Heavy with unmanned Lunar Dragon to LLO, followed by Falcon Heavy with stock manned Dragon V2 to LLO. Orbital rendezvous; crew transfers to Lunar Dragon and uses the first Falcon Heavy's upper stage as a crasher stage to the lunar surface and the Lunar Dragon hovers to a landing. Lunar Dragon then makes ascent, docks with the stock Dragon V2 for crew transfer, and the second Falcon Heavy upper stage returns the stock Dragon V2 to Earth.

open questions this inspires:

What's the on-orbit endurance of the Falcon upper stage?   How fast will its LOX boil off?   Will the mvac engine be able to restart after several days in vacuum?

Will the combined Dragon + upper stage have enough maneuverability to be able to dock with another Dragon?

The LOX shouldn't boil off at all. And yeah, particularly with the SuperDracos, docking with a Falcon upper stage attached shouldn't be a problem.

[Alf Fass]

Could SpaceX's pressure fed Kestrel engine be adapted for use as a descent stage engine? Thrust and Isp are similar to the descent engine used on the Apollo LEM.

[sevenperforce]

Could SpaceX's pressure fed Kestrel engine be adapted for use as a descent stage engine? Thrust and Isp are similar to the descent engine used on the Apollo LEM.

The SuperDracos on the Dragon V2 already each have twice the thrust of the Kestrel, and that's without a nozzle extension on the SuperDracos.

I suppose that if SpaceX wanted to design a single-purpose lander then they could use a Kestrel or a cluster of them, since they already have the designs. But if they want to use the Dragon V2 platform like Elon has talked about, it makes more sense to develop extended vacuum nozzles for the existing SuperDracos.

Using a single manned Falcon Heavy launch to EML-1 with gravity capture, free fall, and a Falcon upper stage crasher to land a direct ascent vacuum-optimized Dragon V2 would be....awesome.

[Eagandale4114]

If Falcon Heavy gets crossfeed, can a Lunar Dragon launched by FH land on the moon with a trunk and some useful payload in it?

By my calculations, FH full thrust with crossfeed can deliver 73 tonnes to LEO, 28 tonnes to EML-1, 20 tonnes to LLO, or 12 tonnes to the surface of the moon as a crasher stage.

Are those numbers with the Raptor upper stage? If not, could you run numbers with what we know of one?

[Chris_Pi]

Since there's much more fuel needed anyways could it make sense to put those tanks in the trunk along with one Vacuum SuperDraco? Leave the ones on the capsule as-is and possibly make up for the extra motor weight with a better expansion ratio and no cosine losses? It would avoid having to deal with detachable nozzle extensions on the capsule or extra fuel plumbing to those motors. They wouldn't be needed until Earth landing.

[QuantumG]

I had this silly idea the other day.. if we were to offer to pay SpaceX to design a reference mission to send crew and cargo to the Moon, would they do it? I know they like money and they have engineers and stuff, but I can imagine they might not want to "waste their time" if you couldn't show you were serious. I wonder how much it'd cost if they did.

[IainMcClatchie]

The idea of two launches for a moon-orbit rendezvous sounds pretty impressive, in the sense that you could put a couple guys on the moon for ~$200m launch cost.

According to Wikipedia, the LEM was 15 tonnes when it detached from the CM for the ride down to the surface.  20 tonnes is a bit better than that!

How did you get an effective ISP for BigBell SuperDraco of 362 seconds when the optimal vacuum ISP is just 336 seconds?  If 1730 m/s takes 43% of start mass as fuel, then you used something like ISP=314 s, right?  The LEM's stages were 311 s, so that sounds reasonable.

[sevenperforce]

By my calculations, FH full thrust with crossfeed can deliver 73 tonnes to LEO, 28 tonnes to EML-1, 20 tonnes to LLO, or 12 tonnes to the surface of the moon as a crasher stage.

Are those numbers with the Raptor upper stage? If not, could you run numbers with what we know of one?

Those are the numbers for the Falcon Heavy Full Thrust, back-calculated from the updates to GTO payload. The quoted FH payload to LEO has not changed since FH was first announced, but the GTO payload has gone up significantly.

The only difference between FH FT and F9 FT in GTO performance is the amount of fuel the upper stage has in the LEO parking orbit, so I used the performance from the SES-9 mission to determine how much fuel the FH upper stage would need in LEO in order to meet the quoted GTO payload flying fully-expendable. That's what gave me the 73-tonnes-to-LEO figure. The Full Thrust update gives the upper stage Merlin 1D Vacuum a specific impulse of 348 seconds, so everything else was just basic Hohmann transfer calculations from that starting point. It's possible that a bi-elliptic transfer on an unmanned segment of the mission could increase the payload significantly; that might be useful for two sequential launches where the first unmanned launch places the lander and crasher stage in stable LLO before the manned FH is cleared for launch.

I don't know the Raptor upper stage performance.

My two preferred mission profiles are the two-sequential-launch with double LOR and the single manned launch to EML-1 with a crasher stage and a direct ascent to aerobraking at Earth. The first one is safer than Apollo and has a huge payload for only two launches without any on-orbit assembly or refueling; the second one has the beauty of being a single launch and of landing the same spacecraft on more than one world.

Since there's much more fuel needed anyways could it make sense to put those tanks in the trunk along with one Vacuum SuperDraco? Leave the ones on the capsule as-is and possibly make up for the extra motor weight with a better expansion ratio and no cosine losses? It would avoid having to deal with detachable nozzle extensions on the capsule or extra fuel plumbing to those motors. They wouldn't be needed until Earth landing.

It doesn't make sense mathematically to drag engines down to the surface that you aren't going to use for ascent. Leave those engines in LLO and descend in a purpose-built lander.

Having separate descent and ascent stages can work, though I would still stick with the same engine set. The "descent stage" could consist of the landing legs and a drop tank. The only reason Apollo used two separate engines for the LM descent and ascent stages was that they hadn't gotten the hang of deep throttling (the descent stage engine was throttled by literally killing specific impulse and keeping the same fuel flow) and they wanted a descent stage with separate plumbing because they were worried about trying to do a drop tank. We have come a long way since then.

I really prefer the crasher stage architecture, though; if you have that big beefy transfer engine and its nice roomy tanks, why not use them? Then your lander only needs enough for hover and ascent rather than needing separate budgets for descent and ascent. It's also good because you get the dV-saving advantages of a suicide burn without the risks (at least, without the splat-landing risks; you have to add a stage separation event but that is basically the same as a launch abort so you should be fine). It's also good practice; eventually, you can use on-orbit propellant transfer and that crasher stage will have enough dV to return to orbit on its own for reuse.

If a separate engine is used, then SpaceX should dust off the Kestrel and build a purpose-built lander around it. But they've expressed a specific desire to use the Dragon V2 platform for these kinds of landings so that's probably the route they'd prefer.

I had this silly idea the other day.. if we were to offer to pay SpaceX to design a reference mission to send crew and cargo to the Moon, would they do it? I know they like money and they have engineers and stuff, but I can imagine they might not want to "waste their time" if you couldn't show you were serious. I wonder how much it'd cost if they did.

Actually this whole idea/thread was my attempt to personally do exactly that. If I can openly publish a series of mission architectures within high enough tolerances that SpaceX engineers can refine them into actual mission proposals, then they have every incentive to do so. Also to hire me. Though that last bit is more conjecture.

The idea of two launches for a moon-orbit rendezvous sounds pretty impressive, in the sense that you could put a couple guys on the moon for ~$200m launch cost.

More impressive if we can put a couple of guys on the moon with a single launch, though the tolerances for that mission profile are really tight.

How did you get an effective ISP for BigBell SuperDraco of 362 seconds when the optimal vacuum ISP is just 336 seconds?  If 1730 m/s takes 43% of start mass as fuel, then you used something like ISP=314 s, right?  The LEM's stages were 311 s, so that sounds reasonable

Whoops, it was supposed to be 332 seconds, not 362 seconds. I think I must have mistyped.

[shooter6947]

1. Dual series launch to double LOR. Falcon Heavy with unmanned Lunar Dragon to LLO, followed by Falcon Heavy with stock manned Dragon V2 to LLO. Orbital rendezvous; crew transfers to Lunar Dragon and uses the first Falcon Heavy's upper stage as a crasher stage to the lunar surface and the Lunar Dragon hovers to a landing. Lunar Dragon then makes ascent, docks with the stock Dragon V2 for crew transfer, and the second Falcon Heavy upper stage returns the stock Dragon V2 to Earth.

Sounds like this one would need propellant refrigeration -- not the end of the world, but a tech to develop.

[sevenperforce]

1. Dual series launch to double LOR. Falcon Heavy with unmanned Lunar Dragon to LLO, followed by Falcon Heavy with stock manned Dragon V2 to LLO. Orbital rendezvous; crew transfers to Lunar Dragon and uses the first Falcon Heavy's upper stage as a crasher stage to the lunar surface and the Lunar Dragon hovers to a landing. Lunar Dragon then makes ascent, docks with the stock Dragon V2 for crew transfer, and the second Falcon Heavy upper stage returns the stock Dragon V2 to Earth.

Sounds like this one would need propellant refrigeration -- not the end of the world, but a tech to develop.

If I recall correctly, LOX won't need refrigeration for a couple of weeks, particularly when the tank isn't very full and has room to self-pressurize a bit.

[nadreck]

Orbital propellant transfer is on the critical path to Mars for SpaceX so I don't think it is fair to rule it out.

Perhaps not rule it out entirely, but that would be something to test on a second manned mission with a secondary payload, rather than a first-time thing. That's another value of SpaceX servicing manned lunar landings: it gives them practice for what they will need to go to Mars.

The only values for SpaceX by their stated goals for servicing any lunar missions are for revenue first, then mission and craft development and refinement.

So if we go with my thesis of a lunar surface rendezvous, we send the return craft and support material/supplies via a refueled S2 launched on an FHR refeuled by three FHR launches. Once it has successfully landed the manned craft launches via FHE and lands beside the return craft. Total 4 FHR flights at total $320M one FHE $125M + 2 dragons and development to modify them adequately probably in the order of $200M.


Oh and I think freezing RP-1 will be the issue to worry about with propellant instead of losing all the LOX.

[sevenperforce]

So if we go with my thesis of a lunar surface rendezvous, we send the return craft and support material/supplies via a refueled S2 launched on an FHR refeuled by three FHR launches. Once it has successfully landed the manned craft launches via FHE and lands beside the return craft. Total 4 FHR flights at total $320M one FHE $125M + 2 dragons and development to modify them adequately probably in the order of $200M.

For a first mission, I think they would fly FH expendable all the way (except, of course, for the F9R if they staged at the ISS). Save FHR and on-orbit refueling for a second or third mission with non-mission-critical elements. If you look at SpaceX's track record, they always take a conservative approach and then attempt testing after all other mission elements have been satisfied...like the way they tested a second restart on the upper stage with the Jason mission in order to get ready for SES-9.

But a lunar surface rendezvous off a double FHE is an interesting possibility that I hadn't considered. It should allow for even more payload than a double LOR. It just lacks any sort of landing abort provisions and requires EVA to the ascent vehicle, which seems really risky.

[CuddlyRocket]

I don't think SpaceX will be going to the Moon until after the BFR/MCT has been developed. A system capable of getting to Mars and back will in all likelihood be capable of going to the Moon and back (perhaps with some modifications). I expect SpaceX would be willing to utilise the system on lunar missions - they would be useful test flights - and especially if someone else is paying!

[sevenperforce]

I don't think SpaceX will be going to the Moon until after the BFR/MCT has been developed. A system capable of getting to Mars and back will in all likelihood be capable of going to the Moon and back (perhaps with some modifications). I expect SpaceX would be willing to utilise the system on lunar missions - they would be useful test flights - and especially if someone else is paying!

I would imagine that if SpaceX can go to the moon now, with existing platforms, they would jump at the opportunity to test the tech and operations needed for Mars. The moon is a lot closer than Mars, and if they can use non-mission-critical legs of lunar missions to test technology like orbital propellant transfer, repeat rendezvous, uncrasher stages, hoverslam landings, and so forth, they can get to Mars that much earlier.

And as far as the bill is concerned...if SpaceX can offer a return to the moon 5-8 years earlier than the closest competitors, I am sure someone high-ranking at NASA would at least consider it.