Saturday, January 24, 2015

BFR and MCT: Some thoughts on what they will look like, and be able to do.

BFR and MCT; some thoughts on what they may look like, and what they will be able to do.

There’s been much speculation on what SpaceX’s “BFR” (a non official designation meaning “big freaking rocket”, or other variants of the middle word). What’s known is it will use Raptor methlox staged combustion engines, and the design intent it to make it totally reusable. The specs on Raptor keep changing, and I highly doubt SpaceX will settle on a BFR design until the engine design is finalized (doing otherwise would be insane).

So, let’s work with what we’ve got; BFR will be big, and be reusable, and use Methlox. It will have a low cost per pound to orbit (or it won’t be built). It will also carry MCT (Mars Colonial Transport) to orbit on some missions.

Now lets get into the specs, and the first spec to look at is, of course, $$$. Developing BRF and Raptor will cost one heck of a lot. It’s been theorized that SpaceX will pay this out of pocket and use BFR for nothing but its own Mars launches. I consider this preposterous; we’re talking billions, and there’s no reason to do it that way, so why not save that money for other purposes? SpaceX has always optimized for cost.

BFR is planned to be a very large, fully reusable, system, and thus should have a very low cost per pound to LEO. I’ve long argued that they’ll offset R&D and construction costs the way they always have; by selling launches. Why wouldn’t they? The counter argument has always been that there’s no call for that much capacity, but SpaceX’s recent announcement of a 4025 satellite constellation, plus other companies being interested in such large constellations, blows that argument out of the water; there may well be, and soon, plenty of demand for a large low-cost-per-pound launcher. Selling such launches (or using them internally to launch revenue-generating sats) is IMHO how SpaceX will offset the R&D and construction costs of BFR. One result of this will be a far lower per-Mars-mission launch cost – because the infrastructure will already exist, and will have been paid for. (A current example of this dynamic; SpaceX is using paid-for expendable launches to develop its reusable F9R).

Now, this gets us to MCT, the Mars vehicle itself. SpaceX has released little but the name and payload capacity, so there has been much speculation. Much of the speculation claims that MCT will be able to go, on its own power, from LEO to a landing on the surface of Mars, and then (after tanking up ISRU) return from the Mars surface to Earth’s surface. The inherent problem with this concept is the rocket equation and the fuel fraction required. Building a craft able to land on Mars, as well as fly a reentry and land on Earth, requires one heck of a lot of dedicated mass (heatsheild, airo surfaces, landing engines for Mars, and above all structural strength). Let’s use a Boeing 747 airliner as an example; its cargo capacity (cargo version) maxes at 123 tons, within the ballpark for MCT’s claimed 100 tons payload capacity to Mars. The 747, bare bones empty (without fuel or cargo) masses 128 tons. So, one ton of cargo for roughly 1 ton of aircraft. Not too bad… but if you start adding things like heat shielding and life support, plus the heavier pressurized compartment needed for space, you’re increasing it by a lot.
For comparison, the Space Shuttle orbiter had a dry mass of 110 tons, and could loft 25 tons, a ratio of greater than 4 to 1 (and the Shuttle didn’t carry its primary fuel internally, so didn’t have the mass of the external fuel tank counted). So, let’s say that, via some magical engineering a liberal use of unobtainium as a structural element, you can get a better mass/payload ratio than Shuttle (with its tiny pressurized volume and no significant internal tankage) when scaled up to the size of a 747’s internal volume and cargo capacity. Let’s call it a 4-1 ratio.     

Where does that leave us with a postulated surface-surface MCT? It has to be able to land on Earth and on Mars, as well as have all the life support and other equipment needed for long-duration deep-space missions. Let’s be optimistic and assume the 4-1 ratio above, and we get 400 tons. But, we’re forgetting something; the fuel tanks. Shuttle didn’t include them in the airframe (except the small hypergolic supply) but airliners do (and doing so adds a lot of mass). The heat shielding shielding, and structure, has to protect the fuel tanks too, which means it’s huge, and thus heavy. However, we’ll be super optimistic and say the airframe can be enlarged to hold it all by simply adding 147 tons (If that sounds like a lot, it isn’t, as we’ll see in a bit). So, a surface-to-surface MCT with 100 tons of cargo masses, unfueled, at a very optimistic 647 tons. That’s our dry weight – everything but fuel.

There’s also the matter of internal pressurized volume. Again using an airliner as an example, the 747 has around 32,000 cubic feet – and so, for comparison, does ISS. ISS has a crew of 6, though could handle a few more. However, let’s use the 747 – ISS internal volume and assume it’d be enough for 100 people. (for comparison, the 747 carries 500, but does so in such cramped confines that a 6 hour flight in one is bad, and a 15 hour flight is unmitigated hell). Even if we postulate horrific crowding, you’d need at least that much volume (probably more) to accommodate 100 people for a months-long journey. However, we’ll assume the aforementioned magic engineering and say that a surface to surface MCT (which has to deal with reentry and landing on both earth and Mars, as well as long-duration deep space capability) will mass 647 tons.

So, let’s say you have your surface-to-surface MCT in LEO, and want to go to Mars. How much fuel do you need? Fortunately, that’s easy to figure out; the rocket equation. The optimal Trans Mars Injection burn (least delta/V, but a rare window) is 4.7 KPS. This assumes essentially zero propulsion for Mars orbital insertion (which can be done via aerobraking or multi pass aerocapture) and the landing itself. For the latter, we’ll be optimistic and say .3 kps (which is less than the F9R recovery profile, even without full boostback).  That gives us a needed delta/V of 5kps, so time to tank up! Calculating how much is easy; we’ll assume 380 ISP for Raptor Vac, the most likely engine choice. We already have our dry mass of 647 tons. The rocket equation gives a fuel requirement of 1833 tons (, but there are always boiloff losses, margins, etc, to consider, so add 10%, and we need 2016 tons of fuel. That gives us a fueled MCT mass in Leo of 2663 tons (about 3X the mass of ISS).

Now, how do we get a 647 ton MCT to Leo? This is what’s called, in engineering terms, a bit of a problem. BFR is going to need to be really, really, really big. A Saturn 5 could put 118 tons into LEO. However, MCT will take a performance hit due to reusability, so at best you’ll need a BFR significantly larger than a Saturn 5 just to equal a Saturn 5’s payload. But, our postulated MCT, unfueled, is 6 times the capability of a Saturn 5. A BFR that could launch it thus can’t be 3 cores, each the size of a Saturn 5. You’re going to need something massing 10 times the Saturn 5 – far larger than any estimate I’ve ever seen for BFR, and also beyond the realm of the plausible (either fiscally or physically).   

You’re also going to need the equivalent of 17 Saturn 5 launches just to fuel up one MCT. You’ll also need one hell of a lot of ISRU fuel production on Mars to refuel it once it gets there.    

All this begs the question; why do it that way and spend so much fuel boosting, for example, the earth-entry structures all the way to Mars and back? Or the Mars entry and landing systems all the way to Earth and back? Why not do it much more economically from every perspective, and do so in a way that gives you a far more versatile system? It’s the same problem that makes Orion such a pathetic design; you’re hauling along a huge mass because you’re treating your living space as the reentry vehicle. Far, far better to use a very small RV, like Soyuz’s, for the RV, and use a lightweight hab for the rest.

So, given the implausibility above of a surface-to-surface MCT, what might an optimized MCT-BFR system look like? MCT would not land on either earth or mars; it would be a space-only vehicle, thus saving enormous mass. A good example of such a craft would be a space station module; an inflatable one, such as Bigelow is building… let’s use their BA-330 design for a starting point. Once inflated, it’s big; 11,654 square feet internal volume (for comparison, ISS has 32,333, roughly akin to a 747) It’s a space station module, thus has life support, etc, included. Mass? 20 tons. It’s not big enough though… so you’d need more. Let’s assume 4 linked together. That gives you redundancy too, plus an internal volume of 46,616 square feet. Mass? 80 tons empty – which interestingly, is 4/5th the mass figure SpaceX gives for MCT- 100 tons empty. That’s also far more realistic an internal volume for 100 people, plus the needed life support equipment and consumables. It’s designed as a space station, so it has the ability, inherently, to exist in space long-term, no need to land. 

However- we need propulsion and fuel. So, add a 5th inflatable module, because inflatables would be ideal for fuel storage in space – why waste the mass needed for a rigid tank like the Shuttle ET? I’ll assume 20 tons (it’d be a lot lighter than a hab module, but it’d need a Raptor engine and thrust structure – perhaps two Raptors, for redundancy)

Okay, we have our 5-module MCT. Now, we need to get it to Mars. We’ll do the same rocket equation as above, and so our 200 ton (100 ton empty mass, plus 100 tons of cargo or humans plus supplies) MCT needs 565 tons of propellant to push it through TMI from LEO. However, this MCT doesn’t land on Earth or Mars, so there’s no reason to waste Delta-V by going deep into the gravity wells of Earth and Mars; high-energy orbits will be far better. Let’s use geosynchronous transfer orbit as an example for Earth, and a similar orbit for Mars. That has a major impact on the needed Delta/V. Instead of 5 kps propulsive ability (Leo to Mars landing, assuming aerobraking) we need 1.3 kps (assuming multipass airocapture into Mars GTO). Now, what does that do to our fuel requirement? It reduces it from 565 tons to 85 tons.  (Quite a big difference from the 2016 tons of fuel a surface-to-surface MCT would need to get from LEO to Mars! It’s reduced our fuel (all of it very expensive upmass) needed by 96%). We also save on margins by omitting the need to haul decent fuel (with resulting losses) all the way to Mars.

However, we still need to get to and from the surface of Mars. The surface-to-surface MCT could do it, but the space-only one can’t – it’s limited to orbit. Fortunately, the answer lies in the launch vehicle, the BFR; the reusable upper stage, to be exact. Any upper stage that can return from orbit to Earth is going to be light, about the density/volume ratio of an empty beer can. Thus, even in Mars’ very thin atmosphere, terminal velocity should be in the low supersonic range. That makes for an easy propulsive landing (it already has a heat shield, due to needing one to reenter on Earth). It already has landing legs, too. With some minor modifications (incorporated into the original design), it should be able to land on Mars, and carry a payload while doing so. Once on Mars, it can fuel up from ISRU, and function as a very capable SSTO – with at least the same payload as the full BFR’s max capacity from Earth – Mars’ far lower gravity, and thus orbital speed, makes SSTO easy.  The MCT would arrive in an eccentric (Basically, Martian GTO) Mars orbit, and be met by a BFR upper stage from the surface. If the stage had a payload shroud, cargo could be placed within it for the trip to the surface. A pressurized compartment would do the same for people, or, something akin to Red Dragons could function as landers (and then be returned to orbit by the BFR stage, which could loft more than enough to carry 100 people down – for a trip of less than an hour, 10 could fit in a Dragon).

So, a BFR upper stage, you have your needed Mars ascent/decent vehicle.

Now what about getting the MCT from Mars to Earth? Easier. The BFR stage brings fuel and cargo and/or passengers. You then need 1.1 kps, with multipass aerocapture (which does not require heat shielding) to get from Mars GTO to Earth, and brake into GTO at Earth. From there, high capacity (10 seat) Dragons could deliver any crew to earth, or a simple, small, orbital tug could transfer any cargo to LEO (again using areocapture). A likely cargo (seeing as how MCT would otherwise be coming back empty) would be fuel for the fuel depot in GTO, or one in LEO – it takes a lot less delta/v to get to Leo from the surface of Mars than it does from the surface or Earth. 

Given the low delta-V requirements from GTO to Martian GTO, you could add 1 KPS to do a fast transfer. (something else SpaceX has mentioned). This would also ease the launch window timeframe significantly (my calcs in this post are based on the once-in-2.2 years optimal Mars window).

Therefore, my hunch is that the MCT will either be, or be very similar to, four or five BA 330 modules. Going further, the four manned ones could be linked in pairs, separated by a tether, and spun up to generate artificial G. Generating Mars G would be even easier, as it’s 38% of Earth’s. This would acclimatize crew to Mars G en route, while avoiding the debilitating effects of prolonged weightlessness.  Also, artificial G will probably be required in order to get the breeding stock of food animals (which any Mars colony will have to have) to Mars. No need to take a whole flock of chickens or drove of pigs, for example, but you’ll need to take two or three plus a few hundred frozen embryos.

That brings us back to the BFR; how big does it need to be? For the space-only MCT, it really only has to get, at most, about 100 tons to GTO, which makes it about one and a half Saturn 5 class in capacity – well within the speculated range for the BFR. That’d also give its upper stage the capability (assuming ISRU refueling on Mars) to function as a Mars ascent/decent SSTO vehicle with 100 or more tons of payload.

As a further piece of evidence supporting my hunch that that’s what they have in mind for some BFR upper stages, SpaceX has said one of the reasons for choosing methane was the ability to obtain it via ISRU on Mars. So, unless part of the BFR is intended to go to Mars, that makes little sense (otherwise, only a surface-to-surface MCT would need Methlox, and could use smaller engines).

Using this architecture, the result would be a fairly low-cost, reusable multi-vehicle Mars transportation system. It would have cargo capacity in both directions, allowing for the Mars colony to export ISRU-derived commodities (Fuel, oxidizer, water, food) to Earth orbit (a good fiscal basis for an economically self-supporting colony), due to the fact that it takes far less propulsive delta-V to get from Mars surface to LEO than to do so from Earth.  Granted, there isn’t currently a demand, but the near future should see such a demand, in the form of several LEO and higher space stations, fuel depots, etc. (the advent of the cheap-per-pound launch capacity BFR promises would help create that market).  

Major caveat; I’m writing this post in the belief (If I’m in error, please correct me) that it contradicts nothing SpaceX has recently officially announced regarding BFR and MCT. I do however discount some SpaceX announcements from the more distant past, due to SpaceX's penchant for changing plans due to encountering physical and fiscal limitations during R&D. They have always done this. For example, their recovery method for F9R looks nothing like their announced parachute-based splashdown concept they tried with F9 1.0. Also, Falcon Heavy will look very different from the F9 1.0 based FH they originally announced. They’ve also changed the specs on Raptor massively, more than once. The only thing I’m accusing them of is having a sane approach to engineering, one that’s not needlessly bound by their past estimates. I consider this very commendable. However, it does means that outside speculators, such as myself, sometimes need to assume that some of SpaceX’s announcements may have a short shelf life. Thus, I’m taking such liberties in my speculation here.

A couple of definitions for terms used here; Aerobraking is using a partial entry to dissipate velocity and enter orbit (or set up for full entry). Aerocapture is multiple passes through the atmospheric fringes, such as MRO used to enter Mars orbit. Aerobraking requires a heat shield, while aerocapture does not. Both save on propulsive delta-V.



1 comment:

ken_anthony said...

Bravo. Now I need mission cost estimates to compare it with my 13 ton dozen crew 40 ton TMI concept.

We are in total agreement that surface to orbit to surface needs to be separate ships or the MCT becomes just too ridiculous.

Do you have some ball park dollar amounts?