[Update 2/10/09 11pm: I’m actually pretty sure I made a mistake here, and my entire idea may be more or less worthless. Basically, the problem is that for any elliptical trajectory, you have a “line of apsides”, which basically is a line connecting the apogee and perigee. In order to have a good lunar transfer, you need that line of apsides to basically pass somewhere close to where your orbital plane and the lunar orbital plane intersect, with the moon being somewhere near that line. Unless I’m missing something, this may invalidate the whole idea, and you may really be stuck with a situation where a given depot only has a 2-4 narrow windows per month for a good lunar departure…until I can get verification from someone who knows more about orbital mechanics, you should probably ignore this idea.]
I’ve had an ongoing debate with a friend (“vanilla” from NASASpaceflight.com’s forums) about the utility of LEO propellant depots compared to doing a Lagrange Point Rendezvous architecture. Vanilla is of the opinion that an architecture that launches crews, cargo, propellants etc directly to L2 (entirely bypassing LEO) is the way to go, while I favor having depots in LEO to aggregate propellants first. Vanilla is a smart guy, and he raised a legitimate concern about using LEO stations and depots as departure points for cislunar missions–infrequent launch windows. Mike Griffin also says the same in some of the presentations he gave before becoming a NASA admin, and various articles over the years about lunar transportation have also pointed out the issue in more detail.
The Issue with In-Plane Departure Opportunities
The basic challenge is that in order to keep delta-V requirements reasonable, you really want to depart for the moon when the moon is in the same plane as the orbit you’re departing from. When you’re in LEO, say at a depot or staging station, you’re in a specific orbital plane, and that plane and the Moon’s orbit precess at given rates resulting in the moon only crossing through a given orbit only a few times per month. How frequent that is depends on a bunch of factors, but the typical frequency is every ~9 days. Launching directly from the surface however doesn’t have this constraint, because you don’t establish your orbital plane until you’ve actually launched, and the plane you would enter from a given launch site will line up with the moon once every day. If you’re trying to go to a station in LLO as opposed to a lagrange point, your windows get even more constrained compared with a ground launch to L1/L2.
Now, I’d love to be in a situation where two-three flights to the moon per month was considered a stifling constraint. It’s a problem I’d really love to have. But there are some additional considerations that make the constraints a drag even for early lunar development. Infrequent launch windows means that scrubs are a bigger problem. When your opportunities are rare, and scrubs can delay things significantly, it’s a lot easier to get “Go Fever” and end up taking more risks. Also, earlier in our development of cislunar transportation systems (especially reusable ones), there will likely be a higher probability of scrubs as we work out the bugs in the system.
Multi-Burn Departure Options
So, for the last year or so since vanilla mentioned this constraint to me, I’ve had the problem in the back of my mind, trying to figure out if there was a way to give vehicles departing LEO stations more flexibility. Saturday morning, I rediscovered a good idea from 40 years ago that looks like it can help retire this issue, and maybe even give LEO-based vehicles more launch flexibility than ground based systems. All without adding very large performance or flight time penalties.
The idea I came up with is to move from a single-burn departure to a two or three-burn departure. In case you don’t know what I mean by that, a traditional Trans Lunar Injection, the lunar stack in LEO does a single long engine burn that places the stack in a trajectory that passes near or tries to impact the moon. Once you get close to the moon, you do a second burn (a Lunar Orbit Insertion burn or a Lagrange Point insertion burn) to deliver the stack to its destination orbit. My idea was to instead go with a three-burn departure that uses an intermediate transfer orbit, and an apogee plane change. For this concept the lunar injection process would take a form like this:
- Sometime after departing the LEO station, the stage would perform a burn into an elliptical orbit with a high apogee–possibly as high of an apogee as GEO (which would be ~2500m/s).
- At apogee a small plane change is performed (for GEO birds launched from Canaveral, the 23.8 change usually costs ~350m/s) that puts the stage into an orbital plane that will intersect with the moon at perigee.
- At perigee the remainder of the TLI burn is completed (~500m/s for a GEO apogee orbit in step 1), and the mission is continued as normal from there.
Now, there are plenty of subtle nuances involved with all of this (probably less than half of which I fully understand myself), but that’s the basic concept. This is a complicated enough problem, that some enterprising PhD student could probably write a full dissertation on how best to balance the tradeoffs and timing issues for such trajectories. However, it would appear pretty likely that such a solution would give at least daily low-cost opportunities for departure, and for some orbits you might even be able to get more frequent opportunities than that.
As is often the case in life, there are some drawbacks to this approach compared to the traditional method, such as:
- Delta-V penalties–while the velocity expended to enter the transfer orbit wouldn’t be wasted, and would serve as the first part of the TLI burn, the plane change delta-V would be wasted. So, there will be a tradeoff of performance vs. longer trip times.
- Added mission complexity–while transfer orbits and apogee plane changes are actually quite common (all GEO missions launched from the US have to do this), it does add more things that can go wrong in a mission. Depending on the reliability of your transfer system, and if you have engine-out capabilities, this could have a non-negligible impact on the probability of mission success
- Van-Allen Belt Transits–you end up passing through the Van Allen belts two more times. While this isn’t going to be a big issue for propellants or cargo, manned vessels might require a “storm shelter” to keep the radiation levels down–but such “storm shelters” are good ideas anyway for trips beyond the earth’s magnetosphere.
- Trip Time Increases–this one’s pretty minor. For a GEO altitude transfer orbit, you’re talking about adding ~10hrs to a 70-100 hr trip (ie less than 15% increase).
Of these concerns, the main two I’m worried about are the added complexity and the extra Van Allen crossings. But the added risks, with reliable engines like the RL-10 are not that great, and the complexity is on a level that seems to be acceptable for owners of billion dollar satellites, so it can’t be that bad of an added risk.
One of the key benefits of this approach is that it allows you to get better utilization out of your cislunar transportation infrastructure. Coming from a manufacturing background, its often easy to underestimate the value of being able to level the load, reduce lead times, allow for smaller more frequent deliveries and smaller move batches. While turn-time isn’t as absolutely critical for interorbital reusable vehicles as it is for earth-to-orbit ones, it does make a difference. The more flights per year you can get out of your vehicles and your people, the lower that portion of the mission cost gets. And with more demand for propellant, that part of the cost equation can be driven down in time as well.
It is also worth mentioning that staging a transfer stage at the point where it’s at a GEO level apogee would make it a lot easier to reuse your cislunar transportation hardware. I’ll have to run the numbers on that sometime. But basically, without aerobraking, it’s a challenge to come up with a reusable cislunar architecture that closes easily. But just like in terrestrial applications, breaking up the TLI/LOI/TEI/EOI delta-V into two stages can possibly make a big difference (round trip delta-V for a non-aerobraking stage is on the level of an SSTO).
Lastly, this is an option that can actually be used both coming and going from LEO and lunar orbit. If you’re going to a lunar polar orbit station for instance, you can split your LOI into three burns. One slows you enough to capture into a highly elliptical lunar polar orbit, followed by an appropriately timed plane change, and your final braking burn to bring you down tot he LLO station’s orbit. You can do the reverse process to give you anytime return from an LLO outpost, and getting up to an altitude where the plane change is cheap is a lot quicker on the moon.
Basically, if I’m understanding this right, these tools can help take a lot of the orbita-mechanics-induced pain out of moving goods between LEO and LLO transportation nodes.
What do you all think?