The Next Generation of Human Spaceflight

[rjaypeters]

I wouldn't be surprised if he had the entire craft fold in half to expose a docking ring.

Yes, the Origami Orbiter. Much cooler sounding than Space Ship 3.

[93143]

Dragon is in many ways much more capable than the Orion capsule.

What ways?

[GIThruster]

Dragon is in many ways much more capable than the Orion capsule.

What ways?

Been some time since I compared them but for one, Orion was designed to carry 4 people and Dragon carries 7. It's lighter, they did some funky cool stuff double tasking the escape tower rocket, it's designed to have the heat shield replaced very quickly, is easily retasked to carry cargo or crew or a mix. Planning to launch it with cargo for the first bunch of trips in order to man rate it is pure genius. Basically, Musk was able to highly optimize Dragon over against Orion because Orion was designed to go to the Moon and beyond while Dragon is designed to go to ISS. If it ever goes further, they'll need to redesign some systems.

[GW Johnson]

There is a similar discussion to this one going on over at NewMars. A participant named "Bob" told me about this one. So, I registered.

You can get a very rough guide to the practicality of SSTO from the rocket equation. On chemical rocket propulsion, the Isp is in the 300-450 sec class. For 26,000 fps, even without gravity and drag losses considered, you are faced with mass ratios of 14.7 to 6.0. Those numbers correspond to fuel fractions (of gross liftoff weight) of 94% (kerolox Isp 300) down to 86% (LH2-LOX Isp 450).

At propellant percentages like that, there is very little room for a viable structure, much less any payload. Consider that structural (inert) mass fractions are typically in the 8-10% range for one-shot throwaway stages, and that we have had serious difficulties actually re-using items that fragile (shuttle SRB's and Falcon-9 1st stages).

If you are going to build some sort of reusable vehicle, it has to withstand some very punishing environments on ascent and especially descent, repeatedly. And, unlike shuttle, it needs to do this with minimal turnaround maintainence, and it needs to do it for far more missions than shuttles ever flew (under a hundred each).

In other words, it needs to look and act like a real airplane: tens of thousands of flights where you gas it up, turn the key, and fly, with only minor replacements like tires, and infrequent overhauls of engines and equipment. It takes significant amounts of any materials you care to consider to hold up like that. Our supersonic-capable bombers had inert mass fractions exceeding 50%. The most reusable rocket vehicle in all of history was the X-15 at 40% inert, and it was air-launched and sub-orbital.

I'd hazard a guess that a truly airplane-like reusable launch vehicle (winged or not) would have an inert mass fraction near 30%, and only if composite materials were included in appropriate ways. You cannot expose those materials to aeroheating. No organic matrix I am aware of holds up past 290 F. And none of the composite materials are as stiff as traditional metals, sometimes stiffness is necessary in design.

That being said, assume 30% inert, 5% payload, ignore the 10+% gravity and drag losses, and solve for the Isp necessary to reach LEO. You have 65% propellants, corresponding to mass ratio 2.857. As it turns out, you need an average Isp over 770 seconds. There is no rocket chemistry that can achieve that. That leaves you three choices:

stage the vehicle

use nuclear rocketry

use airbreathers, most likely in parallel with rockets

All of those things are being discussed by many folks, including here and over at NewMars. I'm quite sure there are many places, although I have not been out there looking.

The odd part of this is that when I was much younger, I was a full-capability ramjet propulsion engineer. All that stuff effectively died in this country 15 years ago. Why that was bad is another story. But, I have been looking at how best to improve the performance of rocket vehicles (winged or not) by adding high-Isp / low frontal thrust density ramjets while the vehicle is still deep in the air, where it does some good.

I don't have a trajectory code, but I have been able to make some pretty good back-of-the-envelope estimates, for a two-stage hypersonic launch airplane system. I posted these over at

http://exrocketman.blogspot.com

where they are mixed in with a lot of other stuff. There's a chronological navigation tool on the left, and the titles in that list will guide you to what you seek. Several of the NewMars guys have seen this stuff. It's been fun; I haven't done this sort of thing in 15 years now, and never for climbing accelerators before. Missiles were a different beastie.

Haven't done the ramjet-assisted vertical 2-stage rocket yet, but I will, and when I do, I will post it there, too. If there's anything there on "exrocketman" that you folks can use in your discussions here, be my guest. That's why I posted it.

GW Johnson
McGregor, Texas

[GIThruster]

Welcome to the forum, GW! It's always great to meet folks with serious expertise they can share. T-P is indeed a fascinating place in this regard.

I enjoyed your post but wanted to point out that X-33/Venturestar could have flown with much less than 30% inert mass-fraction had they not bungled the tank and especially if the aerospike hadn't needed the heavy heat sink. It really was an interesting design.

My contention thus far has been that we ought to be looking for a next generation thruster that can enable both cheaper launch and deep space exploration. A thruster with a 700's Isp certainly would do that, but a ram can't. That's one reason I've been championing the notion of a TRITON style aerospike. (BTW, the aerospike doesn't need quite so high an IsP since its chamber is more efficient. The key is to build a spike that weighs less, not more, than a standard bell. This is something that seems quite possible for a fission thruster where the bell heats the propellant rather than the other way around. Such a thing is easily air breathing as well. In fact, it could breathe atmosphere on most of the planetoids in our solar system.

[GW Johnson]

I'm not yet familiar with the TRITON-style aerospike of which y'all have been discussing. Although, I know what an aerospike nozzle is.

The ramjet stuff I have been looking at only helps with a part of the launch trajectory. For lifting winged launch vehicles, about Mach 6 at about 60 to 80 kft is looking pretty good. Of course, the beginning must be rocket-only: ram takeover speed is about Mach 1.5 for a realistic supersonic/hypersonic design.

Plain ramjet will work to Mach 6, we did it (accidentally) back around 1980 with the very first flight test of ASALM-PTV. That bird was really only intended to cruise at Mach 4 and dive at about Mach 5. But, it really was still accelerating at Mach 6 on that test. That record stood until the first successful test of the X-43 scramjet in 2004. And ASALM-PTV was plain subsonic-combustion (synthetic) kerosene-air ramjet.

For a vertical-launch stage rocket, I would add ramjet pods as "strap-on boosters". These would require integral boosters as in missile technology from the 60's-on, so as to provide good thrust from the launch pad. Takeover on ram is around Mach 1.5, and if you build it to accelerate hard, the vehicle could be doing near Mach 4 as it leaves useful air around 80 kft. Getting to Mach 5 or 6 by that altitude is a pretty challenging (but still do-able) acceleration averaging around 5 or 6 gees.

From there you continue burning the lower-expansion 1st stage to around 10,000 fps, just as now, and use a higher-expansion 2nd stage gravity turn to LEO, just as now. The blended thrust and Isp during the parallel-burn portion of the first stage burn (M1.5 to M4-ish) depends upon how much ramjet strapon you can physically add. Think something between 300-ish Isp kerolox and in the ballpark of 1300 sec ramjet, although that last figure varies a lot with speed.

Below Mach 3 at 80 kft, you could use turbine instead, and have higher blended Isp all the way from launch (no booster required). The effect of the higher blended Isp is to reduce required propellants a little and increase either payloads or structural weights a little, or maybe some of both.

The pod design I have been looking at features airplane-like high structure mass fractions, a flyback recovery mode with a folding wing, and thus airplane-like reusability. It would also have a very small logistical "tail" supporting it. That kind of reusability plus small logistical "tail" is where a huge savings could be made, leading to a big reduction in payload cost as delivered to LEO.

Crudely, it might cut costs in half to add a technology like this to the mix. I just dunno yet.

The basic ramjet combustor and nozzle technology is 1940's stuff. The spike inlet technology is 1950's stuff. The integral booster and ablative liner is 1960's stuff. The inlet port cover and eject nozzle concepts are 1960's and 1970's stuff. The sudden dump flame stabilization technique (instead of 1940's V-gutters) is 1960's and 1970's stuff. Use the same kerosene as the rocket for fuel.

If you use a different fuel, such as liquid methane, in the rocket, use it also in the ramjet. I don't recommend liquid hydrogen in a first stage: too low a density, too much volume, too much drag.

[GIThruster]

GW, there's no such thing as a TRITON aerospike. I was just ad-libbing that since the overweight problem of the aerospike might well be solved by using a TRITON style heat exchanger and the TRITON could have better performance with an aerospike.

I expect you're aware that any NERVA style thruster can use many different kinds of propellant, limited mostly by things like corrosiveness. This being the case, they'd make good air-breathers in all different kinds of atmosphere and of course can be run on carried propellant in space.

I'm an advocate of "simple is good." A single, simple, thruster that has many uses, all of which would outperform any existing tech sounds to me like a good way to go.

You'd know better if such a thing is possible. The heat exchange from the spike nozzle to the propellant seems to me to be the most critical. One might enhance this by for example, producing a strong magnetic field around the nozzle, PM preferred; and since O2 is strongly paramagnetic, you might get enhanced heat exchange. If you're pumping O2, you might do that magnetically in order to avoid heavy turbo-pumps.

However one might do it, I'm still saying a single thruster design that can be used for many purposes seems warranted, especially if we're to do any deep space exploration in a timely manner. If at the same time we get a next gen launch system, well great!; but that's not where I started from. Fact is, it's going to be fantastically difficult to build something that can compete with SpaceX.

Air breathers ARE awfully high performance. Had you thought about selling your ram strap-ons to Elon Musk? I bet he'd listen to what you have to say.

[GW Johnson]

Well, as soon as I do the ramjet-assisted stage rocket back-of-the-envelope study, I was going to vist Spacex's test site in McGregor. Their stand is 10 miles from my front porch, it being one of the stands of the old missile plant where I worked so many years ago. I know a couple of folks on the inside there.

Check out something called Project "Pluto". This was a project to develop a nuclear ramjet engine-powered cruise missile, back in the late 50's and early 60's. As it turns out, my Dad knew of it at LTV in Dallas. They were to supply the missile airframe, inlet and nozzle, and AEC was going to supply the reactor-heater for the "engine". This thing got to ground tests in connected-pipe mode. The hardware is still out at Jackass Flats / Nevada nuclear test site, along with all the Project "Rover" NERVA stuff.

As I recall from what I knew of NERVA back in the 70's, there were some severe Isp-loss issues as well as corrosivity issues, whenever something other than hydrogen was selected as fuel. Myself, I would prefer gas-core nuclear fission to solid-core nuclear fission. This was also done under Rover, although never tested as an all-up device. We got to within about a year or two of that, when "Rover" got shut down. Other good solid core design alternatives include things called Timberwind, and Dumbo. I'd also look at breeding U-233 from Th-232, instead of enriching natural uranium. Much safer cycle, less objectionable products, little or no weapons potential.

For deep space propulsion, I personally think nuclear is required anywhere beyond the moon. For really heavy payloads and/or longer missions (beyond Mars), I think the old Project "Orion" nuclear pulse propulsion is the way to go, although it could stand an update. That was early 1950's fission device technology, after all. It was never tested as a nuke, but did fly just fine at one-meter scale with pulses of high explosives. "Orion" died in 1965 when USAF's space program was turned over to NASA, and they failed to fund it as a "competitor" to "Rover".

BTW, I met Burt Rutan and his brother Dick back in the mid 80's, when he was building the Voyager round-the-world plane. Very inventive fellow. Knowledgeable enough to think way outside the box, and talented enough not to screw it up. He's been building very unconventional aircraft designs quite successfully for many, many years now. I think very highly of him.

Unbelievably enough, I had occasion to revisit Mojave CA again in late June, for the first time in a quarter century. I went to visit XCOR Aerospace, just down the flightline from Rutan's operation. Small operation doing great things. They have done things with liquid propellant rocket engines that few would believe: very safe, very reliable, long life. Comparable to recip, actually.

XCOR wanted to talk to me about ramjets. They couln't find any ramjet people left in California. It would seem I am one of the last surviving full-capability ramjet engineers left in America. That was a bit of a shock, although I knew I was the young one of the bunch way back when.

[GIThruster]

Check out something called Project "Pluto". . . .Other good solid core design alternatives include things called Timberwind, and Dumbo. I'd also look at breeding U-233 from Th-232, instead of enriching natural uranium. Much safer cycle, less objectionable products, little or no weapons potential.

For deep space propulsion, I personally think nuclear is required anywhere beyond the moon. For really heavy payloads and/or longer missions (beyond Mars), I think the old Project "Orion" nuclear pulse propulsion is the way to go. . .

Didn't we learn anything with the failure of solids for mains? The impulse of an Orion or a solid will shake the brain out of your skull. Bad idea, and there's no way to sell "bombs for rockets" to the public as a viable system.

TRITON is a winner. It needs an update to aerospike. Have you looked at the file included here in this thread?

[GW Johnson]

Solid core nuke rockets never failed, they got killed. When Nixon cancelled Apollo in the middle of the moon landings in 1972, he didn't just kill Apollo, he killed all human spaceflight outside Earth orbit. That was the wording of the order.

NASA killed Rover because "if we're not going, who needs the rocket?", as short-sighted as that sounds today. At the time of Apollo, the manned Mars mission was on-the-books, pushed back to 1987 from the original 1983. The baseline plan was a NERVA-style orbital transfer vehicle with habitat module and landing craft. Designs varied, and crew sizes varied, from 6 to 12.

The vehicle and mission plan they really were striving for in 1972 was not solid core NERVA, but a gas core nuke rocket that was almost ready to prototype. First test was scheduled for around 1974-ish, with a series of development improvements leading to a man-rated gas core nuke rocket by the time of the 1987 mission. NERVA was the backup. They already knew chemical was inadequate.

As for Orion pulse propulsion, USAF's contractor General Atomics had a design analysis verified from the subscale test flights, which showed pulse propulsion would not hurt a crew. Indeed, the bigger the vehicle design, the easier it is to smooth out the pulses. Very fortunate, because also the bigger the vehicle design, the higher the Isp. Think Isp > 12,000 sec at pathwise gees 2 to 4. That takes ship designs in the 10,000-20,000 ton class.

NASA indeed looked at an Orion pulse design for Mars in 1965. They couldn't break out of the small-rocket thinking mode, and insisted on a small ship. Lower Isp, harder to smooth the pulses. Not so very competitive with Rover nuke rockets, at only ranges to Mars.

Remember, all these techniques can reduce the one-way travel time to Mars to a mere month or two vs over-a-year. All are "obsolete" today in the sense that we could do a lot better the second time around, if we were to look at them again.

Once you build something powerful enough to travel like that, there is always a way it could be viewed as a weapon. Shoot, even a pointed stick is a weapon. My bare hands are weapons. Why let that potential weaponization possibility stop us from doing what we have to do, in order to go where we want to go? That makes no sense to me.

[GIThruster]

I have to agree with most of that but I can't agree Orion was ever a workable idea. When it was scrubbed no one knew anything about the EMP it would send out, frying every sat in direct line of sight, and pulsing again and again at 2-4 gees is certainly NOT safe for all aboard. Remember, they teach kids these days not to head the ball in soccer for a reason, and they didn't know of those reasons decades ago.

BTW, places like Sandia have continued to research the gas core reactor. I found a fabulous report of all their work over the decades about a year ago. there is a ton on the web to that issue. If someone were to decide to spend the billion dollars necessary to build a TRITON, I'd guess the question of solid, liquid or gas core would be an open issue.

http://www.engineeringatboeing.com/data ... 4-3863.pdf

[GW Johnson]

TRITON looks like exactly one of things I was talking about: updating the older idea to make it more capable.

Don't rule out pulse propulsion. There was a tuned shock absorber system between the pusher plate and the rest of the vehicle. By selecting a pulse rate that resonates with the shock-mass-spring correctly, you spread microsecond impact pulses out all the way to the period of the system. It does not feel like pulses. That's what they flew subscale, and it really did work. The explosives need not be simple fission devices, but they do need to be shaped charge directional devices. We had those in 1955-ish. What we didn't have then was fractional-kiloton devices. But, we've had those since the 70's, I think.

The Black Horse idea of mid-air refueling is very intriguing. I'm not sure that has ever been done with anything but gasolines and jet fuels, though. It might take a bit of effort to make that work reliably with cryogenics. Intuition suggests it can be done, though. Hydrogen would be the toughest. Actually, for atmospheric-stage use, I kinda like liquid methane-LOX. When drag is a problem, higher density is better. It tends to optimize more on (density)x(impulse) than it does impulse. It certainly did in missile work.

[93143]

Dragon is in many ways much more capable than the Orion capsule.

What ways?

Been some time since I compared them but for one, Orion was designed to carry 4 people and Dragon carries 7. It's lighter, they did some funky cool stuff double tasking the escape tower rocket, it's designed to have the heat shield replaced very quickly, is easily retasked to carry cargo or crew or a mix. Planning to launch it with cargo for the first bunch of trips in order to man rate it is pure genius. Basically, Musk was able to highly optimize Dragon over against Orion because Orion was designed to go to the Moon and beyond while Dragon is designed to go to ISS. If it ever goes further, they'll need to redesign some systems.

As you say, some of those advantages are due to lower capability in other areas. (For one thing, it's lighter mostly because it's smaller and doesn't have a service module.) Your original statement is unfair to Orion.

The crew capacity of Orion for lunar missions was 4, but for ISS it was originally 6, and if you crammed them in like a Dragon (two layers!) you could exceed 7 for sure, because Orion is bigger. Also, I'm not sure why you say it couldn't carry cargo; Orion is supposed to be capable of flying unmanned, so... granted it doesn't have a trunk, but then Dragon doesn't have a service module. Also, Orion was supposed to be reusable with land landing, and if it launches on SLS or perhaps Atlas V Heavy, that could still be added back in for Block II... along with the high-gain antenna and the toilet... maybe the notional RMS I've seen in presentations...

[GIThruster]

There are a lot of designs like Orion and Dragon. These are the only two to get built. IIRC, Boeing had a CEV Lite design which was a bit like Dragon.

Yes, most of the differences were in optimization, but I don't think Orion was designed around retasking the way Dragon was. Dragon really is going to be reconfigured on a regular basis and reused. I wish I could remember what it was they did with the escape rocket. . .are they going to leave it on and use it for part of the reentry cycle? Was a cool departure from the norm. . .

[93143]

I'm not sure TRITON would make such a great SSTO engine. Actually I'm pretty sure it wouldn't. The T/W is too low, even with LOX injection. What you want is the same general idea, but with a Dumbo-style core instead of a NERVA-style one.


Over the weekend, I threw together a trajectory analyzer for a fusion-powered LOX-augmented SSTO. Optimum LOX injection rates and overall jet power efficiency as functions of thrust-to-power ratio and ambient pressure were based on some fiddling around with CEA that I did during my free time last week. (Isp values are (2*equilibrium+frozen)/3, which is a painfully simple model, but I have no chemical kinetics resources.) Take the results with a grain of salt, but I think I more or less got it working.

Use of an advanced nozzle is assumed; pure hydrolox Isp (infinite thrust-to-power ratio) is 408 s at sea level (...yeah, the chamber pressure is 200 atmospheres) and 471 s in vacuum...

Optimum Isp is calculated based on the abovementioned performance surface, taking into account drag and required vertical T/W (pitch angle is a byproduct of this calculation), in order to achieve the minimum mass expenditure for a small increment in horizontal velocity. Vertical T/W is specified by extrapolators; the user (me) has to fiddle with the trajectory to get a decent result.

[My first attempt at a full run showed irregularities in the trajectory tracking near max-Q as the extrapolator went nuts; in a matter of seconds the vehicle was nearly side-on to the airflow, the drag was through the roof, and a complex number showed up and propagated through the simulation, bringing it down (ie: it blew up, in the best tradition of late '50s ICBMs). The second attempt, with what I thought was a bugfix, showed a normal ascent (with very high drag due to the stubby SSTO fineness ratio of 3) up until the rocket turned around and headed down towards the ground. The second-last drag-to-thrust value recorded before the simulation crashed was 0.9503. The very last value was 9212.2 (one guess what I exclaimed on seeing the graph)...]


The following simulation was done with a vehicle having a liftoff T/W of 1.5, and a liftoff fusion-power-to-mass ratio of 10 kW/kg. The engine thrust was capped at the liftoff value, leading to a sudden change in behaviour at the point where the optimum T/W drops below the maximum.

Fineness ratio is 7 (not 3), drag is straight Sears-Haack with geometric AOA correction assuming cylindrical symmetry and no lift. The density is similar to that of the Shuttle stack.

The target orbit was 200x200 km equatorial. The trajectory and engine sizing aren't completely optimized, though considering I made orbit with a little over 9.4 km/s total delta-V, I'd say it's not too bad. Still, I used a lot of propellant chucking the thing as high as possible as early as possible, with the result that the trajectory-averaged Isp isn't perhaps as high as it could be... On the other hand, gravity losses are gravity losses - you've got to get up there somehow, and even if you hold the jet power constant (which I don't), a T/W of 2.0 is actually optimal straight off the pad... with LOX augmentation, increasing the Isp entails reducing the LOX injection rate, leading to a precipitous drop in thrust, so reducing the T/W at liftoff by a significant amount doesn't buy you much Isp at all...

The mass ratio for the whole burn is 3.925, meaning that the dry power-to-mass ratio is 39.3 kW/kg. In other words, for a 10 GW core (and thus a liftoff mass of 1000 mT), the entire vehicle, including payload, has to weigh less than 255 mT dry, which is really pushing it for a Polywell. Or, to put it another way, the maximum vehicle T/W with nearly empty tanks has to be 5.89, which is getting uncomfortably close to TRITON's maximum of ~10 in LOX-augmented mode, particularly since TRITON can't match the high-Isp end of the performance curve.

Again, this trajectory was hand-optimized, and I didn't spend very long on it. Experiments with lower liftoff T/W gave worse performance overall, but I can't be sure I was doing it right; take the stated mass ratio with a grain of salt...

Oh, and the Isp was capped at about 1632 s (fueled T/W of 0.1). Assuming the H2 cooling loop can do 1800 K at 10 atm, this requires a direct-conversion efficiency of about 86%. If you can only do 80%, you get a maximum of just over 1500 s, and 75% puts you below 1400 s. A lower coolant temperature will hurt too. Shouldn't affect the mass ratio monstrously, since the bulk of the propellant is expended deeper in the gravity well, but still...

Liftoff propellant is 72% LOX by mass.

EDIT: Darn it, I somehow miscalculated the dynamic pressure. I outsmarted myself, read a graph wrong, and...

Max-Q for a 2.0 gee kickoff was actually over 90 kPa, which is 60% higher than for the Ares I. I've reanalyzed with lower thrust off the pad. Starting with 1.5 gees and tweaking the trajectory brings it down to 41.8 kPa, which is 873 psf.

Changes are in red.