Lifting your rocket from Terra's surface into circular orbit takes an unreasonably large amount of delta V. As a matter of fact, if your missions use Hohmann trajectories, the lift-off portion will take about the same delta V as does the Hohmann from Terra to the destination planet. As Heinlein put it:
How much delta V does it take to go from Low Terra Orbit to Mars orbit? About 5.6 kilometers per second.
How much delta V does it take to go from the surface of Terra to Low Terra Orbit? 7.6 Freaking kilometers per second, that's what! In other words it takes more delta V to travel the pathetic 360 kilometers up to Low Terra Orbit as it does to travel the 228,000,000 kilometers to Mars!
From Low Terra Orbit, where can you travel to with 7.6 km/s? Oh, only to the Planet Saturn, 1,433,000,000 kilometers towards the edge of the entire solar system.
But the delta V cost breakdown is interesting. Getting into orbit takes just a little bit of delta V. It is making sure you stay in space that takes a freaking lot of delta V.
A little sounding rocket can easily rise from 50 to 1,500 kilometers above Terra's surface, where outer space starts about 150 kilometers up. Then the propellant runs out, and the poor little rocket finds itself unsupported hundreds of kilometers up. So it plummets to its doom.
How do you support the sad little rocket? If it uses propellant it will eventually run out, sooner more than later. You can't build rocket legs that are hundreds of kilometers long. You can't use a helicopter blade because there is no air.
But what you can do is put the rocket in an "orbit". An orbit is a clever way to constantly fall but never hit the ground. The trouble is that entering an orbit takes a freaking lot of delta V, about 8 kilometers per second around Terra.
Of course, once you have torchships you can stop all this child's play with wimpy Hohmann transfers and start doing some big muscular Brachistochrone trajectories. Brachistochrones typically require delta Vs that are hundreds of times more than the equivalent Hohmann. So any ship that can handle a Brachistochrone is not going to even notice the delta V cost for lift-off.
But even with torchships, the real bottle-neck restricting developing space resources remains the cost to boost payloads into Earth orbit.
For some cold hard reality read When Rocket Science Meets The Dismal Science.
There are other ways besides rocket boosters and space shuttles to get payloads into orbit. These might take the form of rockets climbing rails set up the side of a mountain, a laser thermal launching facility (in THE MILLENNIAL PROJECT, Marshall Savage calls this a "Bifrost Bridge", that is, a bridge to space composed of colored light), launching loops, space fountains or the base of a Space Elevator.
Polar orbits pass over both the north and south poles, with an inclination close to 90 degrees with respect to the equator. But the important point is a satellite in polar orbit will eventually pass over every single spot on Terra. Heinlein calls these "ball of yarn" orbits, since the path of the satellite resembles wrapping a strand of yarn around a yarn ball. This is why such orbits are used for Earth-mapping, Earth-observation, some weather satellites … and reconnaissance satellites aka "spy" satellites.
For communication satellites, space stations, resupply missions, space exploration, and pretty much everything else, you launch into equatorial orbits.
When deciding where to put a launch site, you have to plan around the Launch Corridor. This is the path the rocket will take when launching which will  allow the rocket to reach the desired orbit and  if the rocket engines fail, the rocket (or the remaining flaming rocket debris) will only fall on uninhabited areas as long as it stays inside the launch corridor. The standard practice is to arrange launch corridors to be over the ocean. Failing that, you need land areas where a rain of flaming rocket bits is unlikely to result in lawsuits or negative publicity. And of course ones that do not violate another nation's sovereign airspace.
During launch, the range safety officer will be watching the rocket like a hawk. If the rocket shows signs of failing to reach orbit, the officer will make a note to dispatch a rescue/cleanup team. If the rocket shows signs of leaving the launch corridor, the officer will hit the panic buttons. Unmanned rockets will shutdown their engines and vent their propellant. Manned rockets will have the on-board pilot take action, but if they are ineffective the range safety officer might have to shoot the rocket out of the sky.
Obviously polar launch corridors have to be along the north-south axis.
The United States uses Vandenberg AFB Space Launch Complex 6 (SLC-6 aka "Slick Six") to launch into polar orbits. Rockets launch due south so the launch corridor is thousands of miles of uninhabited Pacific ocean. The alternative is to launch due north, but that puts the launch corridor right across California, the long way.
Equitorial launches have a second consideration besides the launch corridor.
When you are dealing with feeble launch vehicles using chemical propulsion you need to use every trick you can find. They have grotesque mass ratios which really cut into the payload mass. The most important trick is one to reduce the delta V the rocket needs to achieve orbit.
Since Terra is spinning on its axis, when the rocket is sitting on the ground it is actually already moving. At least it is moving relative to the desired orbit, which is the important thing. If you are standing in New York City; you, the ground, the skyscrapers, the taxi cabs, and everything else is moving at 356 meters per second. The only reason everything seems stationary is because everything is moving together. Now remember that on Terra everything is moving due east because that is the direction Terra is spinning on its axis.
The technical term is the tangential velocity of Terra's surface. It is equal to
tangentialVelocity = ((2 * π * planetRadius) / siderialRotation) * cos(latitude)
tangentialVelocity = tangential velocity at planet surface (m/s) (Terra = 465 m/s)
π = pi = 3.14159…
planetRadius = radius of the planet (meters) (Terra = 6,371,000)
siderialRotation = siderial rotation period (seconds) (Terra = 86,164 seconds, which is actually 23 hours, 56 minutes, 4 seconds)
cos(x) = cosine of x (do not make the mistake of giving your spreadsheet or calculator "x" in degrees when it is expecting radians or something)
I gave you the entire equation in case you wanted to do the calculations for an extraterrestrial planet. If you are just trying to place launch sites on Terra, the equation is:
tangentialVelocity = 465 * cos(latitude)
The point is that the delta V the launch vehicle needs to achieve orbit is reduced by the tangential velocity of the launch site. Bottom line is the closer you can put the launch site to the equator, the better.
For Terra, the pure orbit delta V is about 9,700 m/s (would be 7,800 m/s except for air-drag, gravity-drag, and vertical acceleration). But when launching from New York the delta V is only 9,700 - 356 = 9,344 m/s. And launching from the equator it is 9,700 - 465 = 9,235 m/s. That kind of delta V reduction can buy you lots of extra payload.
Keep in mind that since Terra is spinning due east, the rocket has to launch in an easterly direction in order to take advantage of the bonus. By the same token, if the stupid rocket launches west, the bonus turns into a liability. Launching westward on Terra's equator means the rocket needs an additional 465 m/s to reach orbit.
The important point is that on Terra the equatorial launch corridor is going to point due east.
The better science fiction novels put Terran equatorial launch sites as close to the equator as possible, and where an eastward launch corridor passes over lots of ocean (i.e., on the east coast, near the equator).
- The North Maluku province of Indonesia has parts right on the equator. It has pretty much the entire Pacific Ocean to use as a launch corridor, except only scattered tiny islands in the launch corridor. Possible launch site.
- There is a part of the coast of Brazil that is right on the equator. It has pretty much the entire Atlantic Ocean to use as a launch corridor. Possible launch site.
- Parts of the Galápagos Islands are right on the equator. Unfortunately it only has 906 km of Pacific Ocean launch corridor before flaming rocket bits start raining down on Ecuador. Possible launch site.
- In ARTEMIS by Andy Weir the launch site is in Kenya, with parts right on the equator. It has pretty much the entire Indian Ocean to use as a launch corridor. However, the part closest to the equator that does not include Somalia in the launch corridor is located at 1.7° S latitude.
- In ISLANDS IN SPACE by Arthur C. Clarke the launch site is at New Guinea, with point closest to equator at about 2.6° S latitude. It has pretty much the entire Pacific Ocean to use as a launch corridor, except for the Solomon Islands.
- The real world Guiana Space Centre in French Guiana is at about 5° N latitude. It has pretty much the entire Atlantic Ocean to use as a launch corridor.
- Palmyra Atoll is at about 5° N latitude. It has pretty much the entire Pacific Ocean to use as a launch corridor. And it is a US unorganized incorporated territory. Drawbacks include it is pretty much on the opposite side of Terra from the continental US so that logistics is a nightmare, and the highest point is (currently) only 10 meters above sea level.
- The US Virgin Islands are at about 17.7° N latitude. It has pretty much the entire Atlantic Ocean to use as a launch corridor. Possible launch site.
- In High Justice by Jerry Pournelle the launch site is at Cabo San Lucas, Mexico. It is at an unhelpful 22.8° N latitude. And it only has 390 kilometers of launch corridor.
- The real world Kennedy Space Center Launch Complex 39 is at an ugly 28.5° N latitude. But the United States does not get that much closer to the equator. It has pretty much the entire Atlantic Ocean to use as a launch corridor.
- The real world Baikonur Cosmodrome is at an almost utterly worthless 45.6° N latitude. What's worse it it has to launch at a 51.6° inclination, since China takes a very dim view of being in the launch corridor. Sadly Baikonur is probably located at the best out of Russia's poor selection of launch sites.
THE CURIOUS CASE OF THE ISS INCLINATION
Now one would have expected that the International Space Station (ISS) would be in a 28.5° inclined orbit, which is the orbit you get when launching due East of Kennedy Space Center (latitude 28.5° N).
But it isn't, the ISS is instead in a 51.6° inclined orbit. Why? So that Russian cargo rockets from Baikonur Cosmodrome can reach it. Launching into a different inclination than the space port's latitude costs rocket propellant and reduces payload.
Changing the ISS planned inclination to 51.6° was in retrospect a very good decision. When NASA stupidly cancelled the Space Shuttle program before the replacement vehicle was online, they assured everybody that the replacement would be flying by 2014 at the latest. This would make a small three-year gap in NASA's ISS transport ability. Unfortunately and predictably when 2014 arrived NASA has not even started work on deciding which of the many proposals will be used, much less bending metal and cranking out functional rockets. This leaves NASA at the mercy of the Russians for access to the ISS, but without the Russians there would be no access at all and the station would have long ago burnt up in reentry like Skylab. But I digress.
Clever readers will say but wait! Baikonur Cosmodrome is at latitude 45.6°, should not that be the inclination?. In a perfect world, yes, but there is a problem. When a spacecraft is launched from Kennedy Space Center the lower stages fall into the Atlantic Ocean. And if something goes really wrong, the entire spacecraft can abort and ditch into the ocean as well. If Baikonur Cosmodrome did the same thing, large spent lower stage boosters and/or huge flaming aborting Russian spacecraft would crash into Mainland China, and the political situation would rapidly deteriorate. To avoid that unhappy state of affairs, Russian spacecraft launched from Baikonur go at a 51.6° inclination, so falling rocket bits will miss China.
The Russians already have an annoying problem with the lack of warm-water ports for seagoing vessels. They really dislike having much the same problem with respect to space launches. Therefore they are in negotiations for launch privileges at the ESA's Guiana Space Centre, which is optimally located quite near the Equator and to the West of the Atlantic Ocean.
For comparison purposes, here are the masses of a few sample payloads. This is to give you a mental image of the capabilities of the following booster systems. It will also be useful if the cargo space could accommodate standard cargo cannister sizes.
|GPS satellite||0.8 metric ton|
|Communication satellite||1 metric ton|
|Weather satellite||1 metric ton|
|Hubble Space Telescope||11 metric tons|
|KH-11 spy satellite||13 metric tons|
|TransHab habitat module||34 metric tons|
|Skylab||77 metric tons|
|Space Station Mir||124 metric tons|
|International Space Station||287 metric tons|
|1 gW Solar Power Satellite||1,900 metric tons|
|Lunar Mass Driver||2,750 metric tons|
|Lunar Base (150 crew)||17,050 metric tons|
|10 gW Solar Power Satellite||19,000 metric tons|
|5 gW Solar Power Satellite (Rockwell International estimate)||37,000 metric tons|
|2001 Space Odyssey Station V||145,000 metric tons|
|1 tW Solar Power Satellite||1,900,000 metric tons|
|1.5 tW Solar Power Satellite||2,800,000 metric tons|
|L5 Colony||10,000,000 metric tons|
|Heavy Lift Launch Vehicle (HLLV)||Payload mass delivered to LEO||Cost per payload kilogram|
|Zenit 2 (Ukraine)||13.7 metric tons||$3,093/kg|
|Zenit 3SL (Sea Launch)||15.9 metric tons||$5,354/kg|
|Japan H2B||16.5 metric tons||?/kg|
|Ariane 5G (ESA)||18 metric tons||$9,167/kg|
|Atlas V 551||18.51 metric tons||?/kg|
|Ariane 5 ES (ESA)||20 metric tons||?/kg|
|Titan IV-B||21.69 metric tons||?/kg|
|Falcon 9 v1.2 (SpaceX)||22.8 metric tons||$2,720/kg|
|Delta IV Heavy (ULA)||28.79 metric tons||?/kg|
|Proton-M (Russia)||23 metric tons||$4,302/kg|
|Space Shuttle (NASA)||24 metric tons||$10,416/kg|
|Falcon Heavy (SpaceX)||63.8 metric tons||$2,968/kg|
|Saturn V (NASA)||118 metric tons||??|
|System||Payload mass delivered to LEO||Cost per payload kilogram|
|Black Horse||0.45 to 2.3 metric tons (est)||$227/kg (est)|
|Black Colt||0.45 metric tons||??|
|Rocketplane XS||1.5 to 3.0 metric tons||??|
|The Rocket Company DH-1||2.2 metric tons||$440/kg|
|SASSTO||2.8 metric tons||$11/kg (1968 dollars)|
|Douglas ASTRO||16.9 metric tons||$88.38/kg (1964 dollars)|
|Collier's space ferry||25 metric tons||??|
|ASPEN Nuclear SSTO||36 metric tons||??|
|SERV/MURP||53 metric tons||$95/kg (1971 dollars)|
|Star-Raker||91 metric tons||$22/kg to $33/kg|
|Nuclear DC-X||100 metric tons||$150/kg|
|Rombus||450 metric tons||$2.30 to $5.40/kg (1964 dollars)|
|Sea Dragon||550 metric tons||$59/kg to $600/kg|
|GCNR Liberty Ship||1,000 metric tons||??|
|Uprated GCNR Nexus||1,500 metric tons||??|
|Space Elevator x1||2,000 metric tons/year||$3,000/kg|
|Planetary Orion||3,000 metric tons||??|
|Laser Launch (HX)||3,000 metric tons/year||$550/kg|
|Space Elevator x2||4,000 metric tons/year||$1,900/kg|
|Super Nexus||4,600 metric tons||??|
|Space Elevator x3||6,000 metric tons/year||$1,600/kg|
|Aldebaran||27,000 metric tons||??|
|Lofstrom loop small||40,000 metric tons/year||$300/kg|
|Rocket Sled (StarTram)||150,000 metric tons/year||$43/kg|
|Bifrost Bridge||175,200 metric tons/year||$20/kg|
|Verne Gun||280,000 metric tons||??|
|Lofstrom loop large||6,000,000 metric tons/year||$3/kg|
|Super Orion||8,000,000 metric tons||??|
Since the dawn of the space age, building a vehicle that can fly to orbit using only a single stage has been the holy grail of astronautics. The problem is that single stage to orbit flight is really hard to do because the velocity change necessary to achieve orbit ("Delta-v," 31,000 ft/sec, typically), including losses due to aerodynamic drag, gravity, back pressure on the engines, steering, and so forth, imposes vehicle-full to vehicle-empty mass ratios that are difficult to achieve with current structural technology. The usual approach is to seek more energetic propellants with high specific impulse ("Isp" the number of seconds a pound of fuel can be made to deliver a pound of thrust) values. Alternatively, people have tried airbreathing approaches, which are also an attempt to achieve large Isp. The first approach, ultra-high Isp rocket engines, tends to involve propellants that are not very dense and are difficult to handle, such as liquid hydrogen. The second way, using hypersonic airbreathing jets, imposes surpassingly difficult design and operations problems such as those that have afflicted the National Aerospace Plane program.
Three major configurations have been proposed for single stage rocket vehicles: vertical take takeoff/horizontal landing (VTO/HL), such as the SSTO/R vehicle proposed by the NASA Access to Space Study; vertical takeoff/vertical landing (VTO/VL), such as the McDonnell Douglas Delta Clipper; and horizontal takeoff/horizontal landing (HTO/HL), such as the Boeing Reusable Aerospace Vehicle (RASV) or British HOTOL designs. Between the first two of these, there is no obvious distinction in terms of empty weight. Credible design studies appear to give similar weight estimates for similar vehicles. Horizontal takeoff and landing vehicles, however, tend to be much heavier for a given payload because of the unique requirements imposed by runway takeoff, with wing loads at rotation and the weight of landing gear being of particular concern. Because of these inert mass hits, horizontal takeoff and landing vehicle designs generally tend not to be pure single stage to orbit, but rely instead on sled launch or auxiliary boosters to reduce gross weight.
Our purpose is this article is to discuss another approach for operating spaceplanes off conventional runaways with conventional facilities: Using in-flight propellant transfer to reduce the takeoff gross weight of a rocket-powered aircraft, and hence its size, weight, and cost. This is not an attempt to solve the single stage to orbit problem by means of increasing Isp, but by decreasing Delta-v. It turns out that if you begin the mission to space from tanker altitude and airspeed, the amount of propellant that must be expended overcoming drag and gravity losses is greatly reduced, the vehicle tankage, wings and landing gear all become smaller, and everything becomes a whole lot easier.
The Aerial Propellant Transfer Spaceplane
The general concept of the operation of an aerial propellant transfer (APT) spaceplane is shown in Fig. 1. The spaceplane, carrying only a fraction of its required propellant, takes off a runway in a conventional manner using either rocket power or a set of air-breathing engines and climbs to rendezvous with a tanker, typically at an altitude between 20,000 and 40,000 ft., depending on the spaceplane design. The tanker transfers the remainder of the required propellant and departs, after which the spaceplane fires its main rocket engines at full throttle and accelerates to low Earth orbit. Upon reaching orbit, the payload is released, possibly to be propelled to higher orbit by its own propulsion system, while the spaceplane re-enters the atmosphere, glides to the vicinity of an airport, and then lands in either an unpowered or rocket or jet powered mode.
There are many variants possible to this basic plan, including selection of propellants, propulsion systems, and refueling scheme. For example, the possibility of a spaceplane using the leverage offered by employing very high specific impulse air-breathing propulsion above the tanker's maximum velocity of Mach 0.85 to cut the rocket's required Delta-v to orbit, needs to be considered and traded against the large inert mass penalties and System complexity associated with such jet engines. Use of smaller jet engines for takeoff, loiter, self-ferry and landing offer many operational advantages, but decrease performance.
In the case of refueling scheme, in principle both fuel and oxidizer could be transferred from the tanker to the spaceplane. However, for the propellant combinations of interest, between 72% and 88% of the vehicle's propellant is oxidizer, and therefore the lion's share of the benefit of aerial propellant transfer can be achieved by transferring oxidizer alone, with all fuel loaded on the ground. As such a scheme offers significant gains in simplicity and moderates the amount that the aircraft weight multiplies during refueling at a small cost in system performance, it was the method of choice for all systems included in this article.
Comparison with Alternative SSTO Concepts
It is useful to compare the general characteristics of the APT spaceplane to the three SSTO types that have been considered in the past. These are the Vertical Takeoff/Vertical Lander (VTO/VL), the Vertical Takeoff/Horizontal Lander (VTO/HL), and the Horizontal Takeoff/Horizontal Lander (HTO/HL). A summary of all four options is given in Table 1.
Table 1. Comparison of SSTO Options VTO/VL VTO/HL HTO/HL APT Airframe Minimal Small Large Small Landing Gear Small Small Large Small Engines Large Large Medium Small Expans Varies 1 bar 1 bar 1 bar 0.2 bar New Infrastruct. Much Much Little Little Payload Medium Small None Medium P/L Integration Hard Hard Easy Easy Pad Site Flex Low Low High High Inclination Flex Low Low High High Abort Capability Low Low High High Self Ferry Maybe No Yes Yes Crossrange Low Medium Medium High Evolve w/ Jets? No No Yes Yes Launch Vehicle? Yes Yes No Yes Military A/C? No No Yes Yes Passenger A/C? No No Yes Yes
Overall, the Aerial Propellant Transfer (APT) Vehicle Rates the Highest.
The airframe of the VTO/VL is clearly the lightest, but the VTO/HL is not far behind, and because it takes off and lands light, the airframe of the APT spaceplane will be comparable to that of the VTO/HL. The large airframe required by the HTO/HL to take off at runway speeds with a full load of propellant severely penalizes this option by SSTO application.
In the case of landing gear, the VTO/HL and APT pull about even with the VTO/VL, but once again the HTO/HL system is severely penalized due to the excessive weight its gear must support.
In the case of engines, the APT spaceplane is a clear winner. This is because as a horizontal takeoff System, it only needs a system T/W (thrust/weight) ratio of about 0.8, while the vertical takeoff VTO/VL and VTO/HL needs T/W ratios closer to 1.5, effectively doubling their engine mass requirement relative to the APT spaceplane. The HTO/HL will also need larger engines than the APT because it is carrying inert mass penalties in its airframe and landing gear, and because its required Delta-v to reach orbit is greater. Furthermore, the VTO/VL, VTO/HL, and HTO/HL rocket engines all must be designed to expand their exhaust gasses to a sea level environmental pressure, while the APT spaceplane rocket engine need only face a maximum back pressure of the air at tanker altitude, which can be as little as 0.2 bar. This makes it much easier to achieve optimal rocket engine performance on the APT spaceplane system.
The vertical takeoff options require a lot of expensive new infrastructure in the way of launch pads, while the HTO/HL and APT spaceplane, which use existing airports, require little or none.
The vertical takeoff systems' payload integration is much more complex, and pad site flexibility, payload orbit inclination flexibility, and abort capability are much more limited than either the HTO/HL or APT options. The vertical takeoff systems require specialized pads, the HTO/HL and APT systems can take off and land at conventional airports anywhere in the world, and can ferry themselves around the world for payload integration, launch operation, or vehicle servicing, as required. The winged options have much greater cross-range capability upon reentry than the VTO/VL, with the APT spaceplane having the most, since compared to the VTO/HL or HTO/HL it has the fewest conflicting requirements with the demands to optimize the airframe and trajectory for effficient hypersonic flight. If hypersonic air-breathing propulsion systems should become available, the HTO/ HL and APT can evolve to take advantage of them, while the vertical takeoff systems cannot.
As a launch vehicle the VTO/VL and APT have the highest payload capability, the VTO/HL less, and the HTO/HL none. On the other hand, the APT and the HTO/HL have the potential of functioning as a revolutionary military or civil aircraft while the vertical takeoff systems cannot. Thus, of the four options considered, only the APT has the potential of functioning effectively both as a launch vehicle and as a revolutionary ultra-high speed aircraft, and can therefore be rated as having the highest potential overall.
We've looked at four candidate propellant combinations for use in an APT vehicle; LH2/LOX, CH4/O2, RP/O2, and JP-5/H2O2. All of these are non-toxic and non-polluting.
LH2/LOX offers the highest specific impulse (450 s) of any realistic chemical propellant option. It also offers a relatively high oxidizer/ fuel mass ratio (6:1) which is advantageous in an APT spaceplane system, and several off the shelf engines are available. The primary disadvantage of LH2/LOX is the very low density of hydrogen, which creates the need for large volume tanks which are difficult to incorporate into a reasonable airframe design, and which also creates tankage mass penalties that counter much of the benefit of LH2/LOX's high Isp. An additional disadvantage is the necessity to handle the very cryogenic hydrogen, which stores at 20 K, and is not available to support vehicle operations in many parts of the world.
CH4/O2 offers the second highest Isp (385 s), with the benefit of a great increase in propellant density over LH2/LOX (CH4 is seven times as dense as hydrogen), allowing tankage sizes and masses to be brought under control. The O2/CH4 mixture ratio of 3.5:1 is not too bad, and while both CH4 and O2 are moderately cryogenic, they both store at the same temperature (around 90 K) so that compact tankage arrangements are possible. Liquid methane, while not a staple at today's airports, is available in most places around the world. In fact, CH4/O2 is by far the cheapest rocket propellant combination there is, so that if APT spaceplane operations were to expand to the point where fuel costs were an important factor, this would be a real plus. A significant disadvantage of CH4/O2 is that no flight-rated engine is currently available; development of one based on Pratt and Whitney RL-10 engine technology would probably take three years and cost on the order of $30 million.1
RP/02 (RP is rocket engine grade kerosene) offers a specific impulse of 355 s (using Russian NK-31 engines), and a rather dense propellant combination, only one of whose members is cryogenic. Existing engines achieve their highest specific impulses at mixture ratios of about 2.6:1, which is on the low side from the APT point of view, but the fact that such engines are available off-the-shelf could make RP/02 the propellant choice for a near-term APT system.
JP-5/H2O2 only offers an Isp of about 330 s, but it is by far the densest of the four propellant combinations considered (H2O2 is 1.43 times as dense as water), and burns with a mixture ratio of 7.3:1, very satisfactory from the APT point of view. JP-5 (ordinary jet fuel) is cheap and available at airports everywhere; H2O2 is ten times as expensive as LOX but still a lot cheaper than N204, and is available in many parts of the world. The big advantage of the JP-5/H2O2 propellant combination is that it is entirely noncryogenic, so that the required tanker modifications needed to support APT operations will be much less involved than for the other propellant combinations considered. A significant disadvantage of the JP-5/H2O2 combination is that no high-performance engine utilizing it is currently available. Modifying a Russian NK-31 might produce a rocket engine with an Isp of 330 s and a T/W of 60 at a program cost of about $25 million 2. Developing a clean sheet engine to improve the T/W above 60 could cost significantly more.
We conducted a trade study to examine the performance of these four propellant combinations across a range of APT options and compared them against a VTO/HL baseline. The VTO/HL assumed in our study has no jet engines and must perform a 31.1 kft/s Delta-v to reach orbit from a launch pad. The VTO/HL has a pad liftoff T/W of 1.5. The three APT spaceplane options included:
The "Mach 0.8" option, in which small turbofans are employed to bring the APT fully loaded with fuel (but no oxidizer) up to rendezvous with the tanker. As the APT is loaded with oxidizer during refueling, the increased thrust required to maintain flight is provided by placing a towing load on the refueling boom. After separation the rocket engines are lit and provide the full 27.9 kft/s required to bring the APT spaceplane to orbit from the tanker's Mach 0.8, 20,000 ft level flight condition. The Mach 0.8 APT has a T/W of 1.0 at rocket ignition.
The "Mach 3.0" option, in which larger jet engines are employed that not only can maintain the APT on the tanker until fully fueled but also accelerate the APT spaceplane up to Mach 3.0, prior to ignition of the rocket engines. This reduces the required rocket Delta-v to orbit to 24.0 kft/s. The Mach 3.0 APT spaceplane has a T/W of 0.8 at rocket engine ignition.
The "Mach 5.5" option in which subsonic combustion ramjets are employed to bring the APT up to Mach 5.5 prior to rocket ignition. This reduces the required rocket Delta-v to orbit to 20.7 kft/s. The Mach 5.5 APT has a T/W of 0.6 at rocket engine ignition.
We assumed that the subsonic L/D (Lift/drag) for APT vehicles was equal to 10, supersonic L/D was set equal to 3. Rocket engines were assumed to have a T/W of 60, jet engines a T/W of 8. Tanks were assumed to have a mass fraction of 4% of the propellant they contain if the propellant was water density; this fraction scaling inversely with the 2/3 power of propellant density.
The results of the analysis are shown in figs. 2, 3, 4, and 5. Fig. 2 shows the ratio of "payload" to vehicle wet mass for the options considered. "Payload" here is defined as that portion of the vehicle dry mass not consumed by tanks, wings, landing gear, rocket or jet engines. You can see that under the VTO/HL mode, only the high- performing LH2/LOX and CH4/02 propellants can deliver any payload at all to orbit. On the other hand, if the APT mode is employed, all of the propellants analyzed can be used to achieve SSTO operation.
The ratios of payload to wet mass in Fig. 2 appear to indicate large advantages for the higher performing propellants, however a more relevant basis of comparison is the ratio of payload to day mass, and this is shown in Fig. 3. Here you can see that CH4/O2 offers equal performance to the operationailly much more cumbersome LH2/LOX, and that both of these propellants more than double their payload when used in APT mode compared to their VTO/HL utilization. You can also see that RP/O2 and JP-5/H2O2 offer about equal performance, with JP-5/ H2O2 holding a slight edge for the Mach 0.8 (more or less pure rocketplane) incarnation. Interestingly, the data also shows that using airbreathing propulsion to go to Mach 3 offers little benefit over lighting the rocket engine in the subsonic. However if the airbreathers can be used to drive the APT up to Mach 5.5 (the upper limit of subsonic combustion ramjet operation), they more than pay for their mass penalty with reduced rocket Delta-v, and the payload gains start to become impressive. Since the lower performing RP/O2 and JP-5/ H2O2 propellants are hurt more by having to deal with a large Delta-v, they benefit more by airbreathing augmentation; in the Mach 5.5 APT their payload delivery comes close to matching the CH4/O2 and LH2/LOX options.
Dry mass allocations for the Mach 0.8 and Mach 5.5 APT systems are shown in Figs. 4 and 5. Both tankage and rocket engine masses drop sharply as we go from the Mach 0.8 to the Mach 5.5 versions, and while jet engine mass greatly increases there is still a net gain in payload. Interestingly, the tankage mass of the JP-5/H2O2 system is slightly less than the LH2/LOX system; it's carrying much more propellant but the propellant is much denser, resulting in rough equality. It should be noted, however, that no mass penalty was assumed in this analysis for insulating the hard cryogenic hydrogen, nor was any vehicle L/D penalty assumed for the non-optimal airframe shape that hydrogen use may require. If these effects are taken into account, hydrogen's performance evaluation could turn out to be much less favorable.
Therefore, to summarize, what we've shown is that the APT mode offers large improvements in payload delivery over the VTO/HL mode, so much so that an APT using low-performing non-cryogenic JP-5/H2O2 actually outperforms a VTO/HL using cryogenic LH2/LOX (fig. 3). Within the APT family, the CH4/O2 system offers the highest performance, but if operational considerations such as the desire to avoid cryogens entirely or the need to employ off-the shelf engines in a near-term vehicle are taken into account, then JP-5/H2O2 or RP/02 both offer attractive options with real near-term capability.
The "Black-Horse" Study
During the winter of 1993-94, the U.S. Air Force's Phillips Laboratory conducted a contracted six-week study 3 with WJ Schafer Associates and Conceptual Research Corporation which developed the APT concept further. Since the emphasis was on maximizing the use of existing components and keeping the design simple, existing tankers, landing gear, and conventional technology were used as much as possible. For reasons which cannot be explained here, the rocketplane design was named "Black Horse."
The ground rules for the study were:
- Horizontal takeoff like an aircraft
- Two engines firing at takeoff
- Propellant transfer at 40,000 - 43,000 ft
- Hydrogen peroxide and jet fuel propellants
- Power-off landing
- LEO mission
- Throttling during propellant transfer
- Maximize use of existing facilities and support equipment
- Conservative design assumptions
- Tanker Aircraft Selection
The first problem addressed in the study was how to do in-flight refueling of rocket propellants. In-flight refueling of jet aircraft is commonplace in the US Air Force and Navy today. Two systems are used: the Navy's probe and drogue system and the Air Force's boom refueling system. The probe and drogue system requires the pilot of the receiver aircraft to do all the work, and transfers about 250 gallons per minute. The boom system requires some cooperation between the boom operator and the receiver aircraft pilot, and can transfer 1,200 gallons per minute. The boom refueling system was selected for this design because of its high rate of propellant transfer. Two types of tankers use the boom system today: the KC-10 and KC-135. Of these, the KC-135 is smaller, less expensive, and more readily available. Of particular interest is the KC-135Q and KC-135T. These aircraft have an isolated fuel system, from which the tanker's own engines cannot draw. This will allow dedicated rocket propellant tankers to op erate with only minor impact on the tanker's own systems. To avoid a costly development program, and the need to completely re- engineer the transfer system, it was decided that the propellant carried by the tanker serving Black Horse should be noncryogenic and non-toxic.
There are only a few non-cryogenic oxidizers available: red fuming nitric acid, nitrogen tetroxide, and hydrogen peroxide are the obvious choices. Of these, only hydrogen peroxide is non-toxic. It has other advantages as well. It is very dense (1.432 g/cc in 98% concentration). It has a vapor pressure about one-ninth that of water. It is relatively inexpensive because it is an ordinary industrial chemical rather than a dedicated rocket propellant. Because it is a good coolant, ordinary JP-5 rather than expensive RP-1 can be used as the fuel. Although some special precautions must be taken to prevent it from decomposing in the presence of impurities, it is a stable molecule, and once those precautions have been taken it essentially handles like water. Detailed analysis of a hydrogen peroxide/jet fuel engine indicates the performance figures shown in Table 2 for a engine running with a mass mixture ratio of 7.30:1 (oxidizer fuel). The two columns in Table 2 are for the two versions of the engine. (Black Horse carries seven engines: two for take-off and five for climb.) The "takeoff" version is operable at sea level and permits the aircraft to take off, rendezvous with the tanker, and transfer propellant. The "climb" version is only operable at tanker altitude or above, and is optimized for the climb to space.
Table 2: JP-5/H2O2 Engine Performance Climb Engine Takeoff Engine units Chamber pressure 3000 3000 psia Exit plane pressure 1.0 5.7 psia Expansion ratio 240 70 - Ideal Isp (shifting equilibrium) 354 340 sec Losses due to geometry 2.4 2.4 sec Chemical Inefficiency Losses 1.8 1.0 sec Viscous drag 7.8 6.6 sec Energy release efficiency 6.7 7.3 sec Delivered Isp (in vacuum) 335 323 sec Thrust 19930 19210 lb Weight 310 280 lb
The advantages of the Black Horse aerial propellant transfer concept are threefold. First, the propellants are at a very high density — 1.32 g/cc of propellant at the mixture ratio given. This leads to a smaller vehicle and the capability of transferring up to 155,000 pounds of hydrogen peroxide from the tanker to the receiver. Second, they are non-cryogenic, so that the modifications to the KC-135Q or KC-135T model tanker will be minimal. Finally, the mixture ratio is unusually high. At a mixture ratio of 7.30 to 1, 88% of the benefits of aerial propellant transfer is available if one propellant only is transferred. This helps with keeping the operation simple and removes some safety concerns with simultaneous propellant transfer.
The Black Horse mission profile begins with a takeoff from a conventional runway using the two takeoff rocket engines for thrust. The aircraft is loaded with all the fuel it needs for the climb from the tanker to orbit. It also has fuel and oxidizer aboard sufficient for 15 minutes of atmospheric flight. The total weight of the vehicle at takeoff is about 50,000 pounds, but by the time it achieves tanker rendezvous at 43,000 feet and 0.85 Mach number its weight has dropped to about 38,000 pounds. When the aircraft meets the tanker it takes on about 147,000 pounds of hydrogen peroxide. It then disconnects from the tanker and climbs to space. As it inserts into orbit, its weight has dropped to about 16,500 pounds. After performing its orbital mission, the aircraft reenters and then glides to a normal landing at a runway. A drawing of the Black Horse vehicle is shown in Fig. 6.
The weight buildup of a rocketplane vehicle will determine whether it is possible to enclose the required volume of propellant in an aircraft that weighs little enough to permit that propellant to launch it into space. Table 3 indicates the assumptions for each of the major weight components and the total weight of the Black Horse system. The basic assumptions made for the vehicle were to apply conventional structurait technology by forming the blended wing/ body of the aircraft from ordinary aluminum alloy. The thermal protection system technology deemed suitable for this application is carbon/silica carbide for the nose cap, DuraTABI for broad areas on the lower surface, and a lightweight blanket insulation for the upper surface. The crew cabin accommodations are austere, as in the U2 reconnaissance aircraft.
Unlike most spaceplane designs, Black Horse needs to have a particularly high subsonic lift to drag ratio. This is necessary for two reasons. First, the requirement to fly in the atmosphere on the rocket engine impels the designer to minimize thrust required, so that the rocket propellant load at takeoff remains small. Second, the vehicle's gross weight changes by a factor of about 4.5 during propellant transfer. The maneuver will be very difficult for the pilot to fly if the aircraft does not have a good cruise lift-to-drag ratio. To accomplish these objectives, the Black Horse aircraft features a highly blended design to maximize volume. The doubledelta platform was adopted to provide minimal change of the aerodynamic center over a broad speed range, and also to provide a large strake to hold fuel and oxidizer so that the center of gravity does not move as the propellant is consumed.
Black Horse Weight Breakdown
Structures Group 6,686 Wing 1,572 Vertical tail 739 Fuselage 2,924 Main landing gear 916 Nose landing gear 243 Engine mounts 292 Propulsion Group 3,091 Engines 2,120 Fuel system 971 Equipment Group 1,181 Flight controls 372 Instruments 142 Avionics 567 Furnishings 100 Mission-specific Group 4,000 Reaction controls 400 Life support 800 TPS 2,800 TOTAL EMPTY WEIGHT 14,958 Load Group 33,494 Payload 1,000 Crew 440 Propellant 32,054 TAKEOFF GROSS WEIGHT 48,452 Tanker rendezvous weight 37,380 Oxidizer transfer 146,870 GROSS LIGHT-OFF WEIGHT 184,250
The overall wing area is 780 square feet. The wing loading is sufficiently low that no liift devices such as flaps or slats should be needed for takeoff or landing, especially with the enormous thrust available from the rocket engine The low wing loading also helps a lot in moderating the thermal environment during reentry.
The "Black Colt" Study
Another study of a somewhat different APT concept was done at Martin Marietta during January through May 1994, this one of a near-term suborbital X-Plane that could serve as a demonstration vehicle for the APT concept. Because the vehicle was about half the size of Black Horse, it was decided to call it "Black Colt." A drawing of this vehicle is shown in Fig. 7.
The spirit of Black Colt was to design something that can be buiit and flown today. Thus, an existing NK-31 RP/O2 rocket engine was chosen for primary propulsion, with two Garrett F 125 turbofans used for takeoff, loiter during aerial propellant transfer, and landing propulsion. Also, rather than push for the very high performance required to achieve true SSTO operation, it was decided to settle for only flying Black Colt to haif orbital velocity, with the 1000 lb payload then being delivered to orbit by means of a Star 48V upper stage. By adopting this strategy, it was found feasible to design a reusable satellite delivery system with a wet mass fraction of 0.78 which represents a large relaxation in requirements relative to the 0.92 needed for Black Horse. Since the expended Star 48V only costs $1.4 miillion and allows Black Colt to deliver a larger payload than a $12 million Pegasus, the trade was considered quite favorable. The Mach 12 reentry required of Black Colt is also much less demanding on the technology of reusable thermal protection systems than the Mach 25 reentry of any true SSTO.The NK-31 burns oxygen and kerosene at a mixture ratio of 2.61. Thus about 19,000 lbs of kerosene are required to burn with the oxygen. The remaining 2,000 lbs are used for climb to the tanker and for 30 minutes of subsonic powered flight after reentry. When fully fueled with kerosene and carrying no LOX or payload, the vehicle has a fuel mass fraction of 0.53, with a subsonic L/D of 9, this gives it a subsonic seif-ferry cruise range of about 3,300 nautical miles.
While the Black Colt's F125 engines have a sea level static thrust of 10,000 lbs each, thrust falls off to 6,300 lbs each at 20,000 ft and Mach 0.8. Since the vehicle has a subsonic L/D of about 9, implying a total thrust requirement of 10,500 lb when fully loaded, this limits the refueling altitude to about 23,000 ft if the refueling is to be done under jet engine power alone.
Rather than modify the fuel tanks of a KC-135 to carry LOX, it was felt that the simplest way to enable aerial LOX transfer would be to carry a dedicated LOX tank within the fuselage of any suitable large subsonic aircraft. Tank trucks of the Liquid Air company currently carry tanks with 70,000 lb LOX capacity, greater than that required for Black Colt. Such tanks have boiloff rates of less than 0.25% per day, and with a diameter of less than 8 ft, could be installed within many candidate aircraft. This leaves a cryogenic refueling boom as the primary refueling related development item. While transferring LOX at altitude has been raised by some as a safety concern, in many respects such transfer is safer than handling LOX on the ground, as any LOX spills will be instantly dissipated to the environment and there is no water around to freeze things up.
A cost estimate was done on the Black Colt program. Current estimates indicate that such a vehicle could be up and flying within three years of program start for a total cost of less than $100 million.
Unlike most space vehicles, it will be possible to test the aircraft proposed here in a conventional flight test environment. No special range requirements beyond what is conventionally available at, for example, Edwards AFB should be required. Because there are aviators aboard the vehicle, no destruct package is needed. Aside from storage areas for the new propellant, it should not prove necessary to construct any new facilities for any phase of this program.
The flight test program could begin in a conventional build-up fashion, starting with taxi and ground tests, first flight, performance, and flying qualities testing. This phase of the program would emphasize handling qualities while connected to the tanker boom, because the oxidizer transfer could as much as quadruple the weight of the aircraft when it takes place. Once the flight control system has been qualified, transfer of steadily increasing amounts of oxidizer would support envelope expansion and flight to increased altitudes and airspeeds. Exoatmospheric flight and reentry could be investigated, and the operational envelope of the thermal protection system could be determined. The capability of the system to perform long distance ballistic transfers (in the case of Black Horse, to anywhere on Earth within one hour) could be demonstrated. Loading the aircraft with fuel and oxidizer, up to the maximum takeoff weight, could also permit exoatmospheric flight without propellant transfer. The ballistic ferry range of the Black Horse aircraft under these conditions is about 3,200 nautical miles, allowing for some aerodynamic range extension at the end of the trajectory.
Black Colt Weight Breakdown
Turbojets (2 F125's installed, 10klbf each) 3,458 Rocket Engine (1 NK-31, 90 klbf) 1,540 Structure 3,200 Wings 2,490 Propellant Tanks 500 Landing Gear 1,621 Thermal Protection 1,760 Pressurization system 50 RCS 150 Avionics 450 Contingency (20%) 3,043 TOTAL DRY WEIGHT 18,263 Kerosene 21,000 Oxygen 50,000 Star 48V 4,840 Payload 1,000 TOTAL GROSS WEIGHT 95,103
An orbital flight attempt would follow the envelope expansion phase. Investigation of on-orbit flying qualities could proceed at this point, as well as an experimental determination of reentry cross range. One sub-phase of the orbital flight test program of particular interest would be on orbit propellant transfer. If the aircraft were completely refueled in low-Earth orbit, it would have enough Delta-v to visit anywhere in cislunar space, such as geostationary orbits, or to perform multiple plane changes and visit many different points on a single mission. Reentry from increased altitudes and entry speeds could be tested, yielding an assessment of the capabiitity of a high temperature reentry capability in realistic conditions.
Military and Civilian Applications
Beyond their application as SSTO payload delivery systems, there are many potential uses for Black Horse/ Black Colt type APT spaceplanes. With performance requirements considerably less than that needed for SSTO flight, such vehicles could fly from any point on Earth to any other point in less than an hour, with most of the flight being exoatmospheric. In the NASA arena, such APT vehicles flying routinely in a suborbital mode could be used as hypersonic test vehicles, medium duration (15 minutes per flight) zero gravity labs, ionospheric/thermospheric sounding vehicles, and even short duration astronomy platforms. Within the military sphere, APT vehicles could be used for a number of interesting applications, many of which are enhanced by the ability of an APT to bounce out of the atmosphere after an initial reentry. This would allow it to drop in over a designated area from space at hypersonic (Mach 10-20) velocities, release a payload, and then pop out of the atmosphere and get away to space (Fig. 8). Within the civil sector, such vehicles could be used for fast passenger transport, fast package delivery, prestige business travel, space barnstorming rides, and as a vacuum or zero gravity industrial research platform.
It should be noted that most of the range of suborbital vehicles is provided by aerodynamic lift after reentry, not by the ballistic hop itself (Fig. 9). This is the fundamental reason why for all applications involving surface-to-surface distance capability, winged vehicles such as the APT offer much greater performance potential than strictly ballistic vehicles of the VTO/VL type. It can also be seen (Fig. 10) that the mass ratios required for such vehicles to achieve substantially global range for surface to surface travel on Earth are much less than those required for SSTO flight, which means that the payloads delivered can be much greater. It may be objected that while a CH4/02 APT might only require a mass ratio of four to travel from New York to Tokyo, which is clearly achievable in the engineering sense, a subsonic airliner can perform the same feat with a mass ratio of two, and thus despite its (much) faster flight time, the APT spaceplane would fail economically. This argument neglects the fact that almost 80% of the propellant of the CH4/02 spaceplane is LOX, which sells for $0.05/lb, compared to $0.20/lb for jet fuel. When the difference in the per pound price of propellant is taken into account, the spaceplane's fuel economy is quite competitive.
A Speculative Idea
Beyond their baseline mode of operation, certain speculative operabonal modes exist that could signiticantly enhance the capability of APT vehicles in the future.
Consider the case where we have two Black Horse type vehicles, each using JP-5/H2O2 with an Isp of 335 s. The vehicles have a dry weight of 15,000 lb and a propellant load of 180,000 lb, which assuming a required Delta-v to orbit of 27 kft/s, allows them to deliver 1,000 lb to LEO. Now, let's say that we fly the two of them off together, accelerating them jointly not to orbit, but rather to a suborbital trajectory with a velocity of 18.5 kft/ s. The two space planes are now outside the atmosphere, in free fall (i.e. zero gravity) in the immediate vicinity of each other. Let's say we now bring the two together and extend a refueling boom, allowing the 20,000 lb of residual propellant from one to be transferred to the other. The two then separate, the empty vehicle to return to Earth, the enriched vehicle to ascend to orbit with a payload of 12,000 lbs. Without any new hardware, the orbital delivery capability of the system can be increased by a factor of 12.
Such a non-material enhancement by teamwork would allow even an APT spaceplane that was designed for suborbital flight to achieve orbit. Or put another way, let's say that it turned out after the construction was done that the actual Black Horse dry weight came in not at 15,000 lb, but at 24,000 lb, a 60% mass growth over the estimate. The vehicle would now only be capable of suborbital flight to 23 kft/s. However, if two such vehicles were flown, performed a suborbital propellant transfer at 15.5 kft/s, the enriched vehicle would be able to make orbit with a 1000 lb payload. Since the propellants being transferred are non cryogenic, such a suborbital zero-g propellant transfer could be done using bladders. If the APT in question used LOX for its oxidizer, the transfer would require a weak gravity field, which could be created by both vehicles firing their RCS systems continually during the transfer.
The plan certainly sounds incredible, and to be frank, we don't expect such maneuvers to be done anytime soon, but it's not impossible. On a suborbital trajectory with a velocity of 16 kft/s and a 120 nautical mile apogee, the vehicles in question will be out of Earth's atmosphere for about six minutes. The actual propellant transfer can be done in less than two minutes. With sufficient training, good pilots could eventually do the job.
The idea of refueling an airplane in flight must have seemed bizarre to anyone witnessing the Wright brothers first flights. By the 1920s it had been demonstrated, and today it is done routinely. Doing it on the way to space, and in space itself are just the next steps up the ladder.
What we've shown is that using inflight propellant transfer to reduce the Delta-v needed to fly to space makes it possible for a fighter sized aircraft to achieve orbit. Furthermore, such an aerial propellant transfer spaceplane outperforms alternative vertical and horizontal takeoff SSTO concepts, and offers a much greater range of potential alternative suborbital applications as well. If such a flight mode is used, a completely noncryogenic SSTO employing nontoxic rocket propulsion based on H2O2 and JP-5 is feasible, and even higher performing systems can be built using soft cryogenic CH4/O2 propulsion. Perhaps most importantly, the development of an aerial propellant transfer X-plane spaceplane could be done quickly and cheaply and would permit a variety of revolutionary military and civil capabilities to be demonstrated. Why most importantly? The answer is this:
If the real space age is ever going to be opened up, we will need to have a market for rocket vehicle technology that supports the manufacture of spacecraft components not in lots of ones or twos, but in hundreds or thousands or tens of thousands. If travel to orbit is ever to be as cheap as air travel, we will need a worldwide infrastructure that supports not hundreds of flights per year, but hundreds of flights per day. The only market that can support that kind of development is global long-distance surface to surface passenger travel - like it or not, a lot more people want to go to Tahiti than want to go to orbit. If you want to serve that market, you want to use winged rocketplanes, not vertical takeoff vehicles, because trillions of dollars worth of infrastructure is available all over the world - in the form of existing airports and surrounding facilities - to support their operation. Now it won't happen all at once, and for the same reason that military and then postal aircraft preceded passenger aircraft, military and fast package delivery rocketplanes will precede passenger rocketplanes. But once it does happen, once thousands of rocketplanes are crisscrossing the globe daily, serving business and vacation travelers from New York to Sydney, it won't be to hard to set some of them aside to establish passenger lines to orbit. In fact it won't be too hard to specially outfit a few to be refueled on orbit for expeditions to the Moon or Mars.
That's why Black Horse is an airplane worth fighting for.
The authors wish to acknowledge the important work done by Dan Raymer and the W. J. Schafer team of William Nurick, Frank Kirby, Ed Nielsen, Robert O'Leary, Rett Benedict, and Ray Waish in developing the Black Horse design concept. We also wish to acknowledge the contributions in helping to analyze the Black Colt performed by the Martin Marietta team of Sid Early, Jim Greenwood, Grady Romine, Robert Humphries, Greg Velasquez, Lars Onsager, Elizabeth Sholes, Ann Palen, and Jeff Schnackel. Useful comments and suggestions on Black Colt from Burt Rutan, of Scaled Composites and Lt. Col. Steve Brandt and Lt. Col. Doug Beason of the U.S. Air Force Academy are also gratefully acknowledged. The drawings shown in figs. 1 and 7 were done by Robert Murray, of Martin Marietta. The drawing of Black Horse shown in fig. 6 was done by Dan Raymer.
- R. Parsley, Pratt and Whitney, Private communication, Sept. 1993.
- T. Fanchiullo, Aerojet, Private Communication, Feb. 1994.
- M. Clapp, W. Nurick, F. Kirby, E. Nielsen, R. O'Leary, R. Benedict, R. Walsh, and D. Raymer, "Aerial Propellant Transfer to Augment the Performance of Spaceplanes," Phillips Lab Report, Feb. 1994.
|Payload mass delivered to LEO||Cost per payload kilogram|
|2.2 metric tons||$440/kg|
While the design is fictional, it would actually work. The authors have patented it. The small payload means the rocket is intended more for "space access" instead of heavy lift to orbit. The business model for the developers was more to sell the rockets (at an attractive price of $250 million) rather than selling cargo boost services.
|Payload mass delivered to LEO||Cost per payload kilogram|
|2.8 metric tons||$11/kg (1968 dollars)|
|Gross Mass||97,976 kg|
|Empty Mass||6,668 kg|
|LEO Payload||2,812 kg|
|Thrust (vac)||1,558,100 N|
In 1966 when winged space shuttle designs were being studied, the Douglas Aircraft Company was doing a cost-benefit analysis. They were comparing reusable space shuttle costs to throwaway two-stage ballistic boosters. Somewhere along the line they took a look at whether it was possible to make a reusable single stage ballistic booster. The SASSTO was the result. The payload was not much, but it was enough for a Gemini space capsule. A Gemini would transform the SASSTO into a space taxi or even a space fighter, capable of satellite inspection missions. Without the Gemini it could deliver supplies and propellant to space stations and spacecraft in LEO.
Bono pointed out how inoperative satellites could become space hazards (although the concept of the Kessler Syndrome would not be created until 1978). A SASSTO could deal with such satellites in LEO (Bono called this Saturn Application Retrieval and Rescue Apparatus or SARRA). Even better, such satellites could be grabbed and brought back to Terra for refurbishment and re-launch. This would be much cheaper than building an entire new satellite from scratch, which would interest satellite corporations. Only satellites in LEO though, communication satellites in geostationary orbit would be out of reach.
The interesting part was on the base. Conventional spacecraft trying to do an aerobraking landing need a large convex heat shield on the base (for example the Apollo command module.). Unfortunately a reusable spacecraft has a large concave exhaust nozzle on the bottom, exactly the opposite of what you want. Tinsley's artist conception for the "Mars Snooper" had petals that would close over the exhaust nozzle sticking out of the heat shield, but that was impractical.
Douglas' solution was to use an aerospike engine with the spike truncated (which they confusingly call a "plug nozzle", contrary to modern terminology). The truncated part became the heat shield, the untruncated part around the edge was the aerospike engine.
delivered to LEO
|16.9 metric tons||$88.38/kg (1964 dollars)|
$1,329.42/kg (2020 dollars)
|Engines||2 × RL-10|
1 × J-2
|Inert Mass||14,000 kg|
|Payload Mass||16,851 kg|
|Dry Mass||30,851 kg|
|Fuel Mass||74,842 kg|
|Wet Inert Mass||89,290 kg|
|Wet Mass||106,141 kg|
|Engines||1 × M-1|
2 × J-2
|Dry Mass||32,558 kg|
|Fuel Mass||269,434 kg|
|Wet Mass||302,183 kg|
|Wet Mass||407,870 kg|
ASTRO is an acronym for Advance Spacecraft Truck/Trainer/Transport Reusable Orbiter
This was a Vertical-Takeoff/Horizontal-Landing (VTHL) Two-Stage-To-Orbit (TSTO) spacecraft for transporting space station crews and cargo. One of the study requirements was that off-the-shelf technology was to be used: meaning M-1, J-2, and RL-10 rocket engines (though the M-1 was never actually built because the Nova was never actually built). The RL-10, J-2, and M-1 engines had thrust levels of 67,000 N, 890,000 N, and 6,700,000 N, respectively (as mentioned by Douglas, some of the engines had higher thrust when actually built).
In its surface-to-orbit cargo transport role it was a two-stage vehicle. Conventional rocket are stacked nose-to-tail (on top of each other like a totem pole) for staging, winged STO rockets are usually stacked ventral-to-dorsal (piggyback fashion). ASTRO was unusual in that it was stacked.
Douglas figured it would be easier to make a recoverable and reusable design if they aimed for a smaller spacecraft with lower payload but capable of frequent flight schedules. Larger spacecraft with higher payloads with infrequent flight schedules are more difficult to design to be reusable, actually more difficult to design regardless of re-usability.
Both the orbiter and the booster were lifting bodies (since putting wings on a craft that can aerobrake from orbit and perform hypersonic flight is a bit of a challenge). Wings/lifting body allowed the spacecraft to be reusable, i.e., landing on a landing field instead of ditching in the ocean. Both landed on skids with steerable nose gear.
Orbiter had a pilot and co-pilot. Booster just had a pilot. The crew compartment of each are abortable, jettisoning from the spacecraft in case of a catastrophic malfunction. For missions with only pilots on-board, only the cockpit is jettisoned and the cargo compartment is not (and may not even be pressurized). For missions with more crew, some will ride in the cargo compartment. In that case both the cockpit and the cargo compartment is jettisoned and both are pressurized.
When boosting cargo both the booster and orbiter were VTHL. The orbiter alone was capable of horizontal take off and landing (HTHL), allowing it to be used as a sub-orbital trainer to educate space pilots. This solved the problem of how do you train ASTRO pilots before it is operational. They would quickly build the orbiter section and use it to train the pilots while the booster section was developed. In addition, the orbiter alone was a useful suborbital craft with a range of 8,000 kilometers and a cargo capacity of ten people or 2,100 kilograms of cargo.
In Phase II the booster prototype was made by taking the orbiter, adding a second J-2 engine, redesigning the cabin so it only holds one pilot with zero cargo, and altering the nose so that the orbiter tail bumper can perch on it. Douglas figured this early version could put one tonne and two crew into orbit.
In Phase III (final phase) the booster was beefed up, making it larger than the orbiter, and adding a freaking M-1 between the two J-2s. As previously mentioned, the M-1 was a monster rocket intended for NASA's Nova. It made the F-1 engines on the Saturn-V's first stage look like a child's bottle rocket. The Phase III could put nine tonnes and two crew into orbit. The cargo compartment has a volume of 14.9 cubic meters. If that is not enough volume (the cargo bulked-out before it grossed-out), some cargo can be carried on the skin in detachable pods. Meaning if your cargo has a mass of nine tonnes but a volume bigger than 14.9 m3 you will be forced to put some of the cargo metaphorically on the roof rack. In space slang, your cargo has "bulked-out" before it "grossed-out".
The maximum 9 tonnes of payload is only if the ASTRO is launched from the equator with a launch azimuth of 90° (due east). The maximum payload decreases as the launch site is displaced from these conditions.
But to hedge their bets, Douglas also wanted the ASTRO to be useful for any mission the military had for a space plane: defense bombardment, reconnaissance, satellite inspection, satellite interception (SAINT II), satellite logistic support (refuelling station-keeping thrusters, swapping out spy satellite camera film, etc.), recoverable space laboratory, astronaut training, maintenance, rescue, and supply.
For reconnaissance missions the payload is cameras and film. For satellite surveillance missions the payload is inspection, kill, and capture equipment. For repair, cargo transfer, and general service missions the payload is maintenance equipment including modular spares, tether reels, space pods, and the cargo. With some missions maintenance personnel are part of the cargo.
The ASTRO stack took off from a mobile launcher-erector. This was to eliminate the need for a large Cape Canaveral type launching gantry, and the need to use huge cranes to piggyback the orbiter on top of a cargo aircraft to fly the damn thing back to the launch site. NASA actually did use the piggyback method with the Space Shuttle, transporting the Shuttle from its landing site to the launch site on the back of a modified 747.
The mobile launcher erects the stack so that it points skyward (or you will not go to space today). The booster blasts off and carries the orbiter and itself to an altitude of 82 kilometers. There stage separation occurs, and the orbiter continues upward to a 555 km orbit. The poor booster pilot has to glide the corpulent booster to a dead-stick landing 830 freaking kilometers from the launch site. Get it right the first time because you can't turn around to make a second attempt.
The proposed ASTRO system would have a fleet of 12 boosters and 24 orbiters with a turnaround time between missions of less than 18 days. This would allow about 240 flight per year. The planned service life was 100 flights for each orbiter and 200 flights for each booster. The engines were rated for 50 firing before needing a major overhaul. The orbiter and booster airframes were rated for up to 300 flights.
|Payload mass delivered to LEO||Cost per payload kilogram|
|25 metric tons||??|
If one looks at a large orbit-to-orbit transport ship to be the "freighter" for colonization efforts at Mars, then there must be some way to ferry-up cargo into Earth orbit to load this ship. And, there must be some way to ferry-down cargo from this ship in Mars orbit, to the surface. This is the analog to the way ocean-going ships were loaded and unloaded, for centuries here on Earth.
Ferry at Earth
Consider: the Spacex "Starship" is first and foremost a large-payload transport from Earth's surface to low Earth orbit. It uses a recoverable booster to deliver payload to orbit in 100+ ton lots, without any refueling. One flight, one 100+ ton payload delivery. I am not aware of how much it can transport down from Earth orbit, but for colonization, that is not so relevant (later on with interplanetary trade it is).
If one believes the published numbers for "Starship", this is an impressive ferry vehicle. It could easily transport a payload to a colonization vessel in Earth orbit in those 100+ metric ton lots. So I think we have that end covered. 1 flight for a 100 ton colony ship payload, 10 flights for a 1000 ton colony ship payload, etc. Cost per ton delivered will be whatever it turns out to be for actual "Starship" operation. We won't know what that really is until it has been tested and begins regular flights.
Ferry at Mars
Now, what about at Mars? What might "Starship" do for us, to unload that orbiting colony ship and bring its payload to the surface? I took a look at that with data for low Mars orbit (3.55 km/s), a small landing burn allowance (Mach 1-ish 0.33 km/s), factors on the delta-vees, and a generous rendezvous allowance on-orbit at Mars (1 km/s). The factors were 1.02 for gravity and drag to reach orbit, 1.5 on the min landing allowance, and 1.0 on the rendezvous allowance.
For the ship, I used the published figure for ship inert mass 85 metric tons (about which I have serious doubts until I see it actually fly at that inert weight), and a "typical" but conservative figure for vacuum Raptor engine performance of 350 sec Isp. It holds up to 1100 metric tons of propellant that must be produced on Mars from local materials, and in quantities and rates to support the flight rates. All that is assumed for this investigation.
I did the mass-ratio-effective dV thing to estimate performance vs payload and propellant load, done as the same large payload masses transferred both ways (both up, and down). Those results were surprisingly good:
100 650 200 990 234 1100
This result is driven by the low inert mass reported so far for the "Starship" design. If you believe that really will be achieved, then it looks like a "Starship" stationed on Mars, and locally refuelled there, can serve very well as a reusable ferry for colonization ships sent to low Mars orbit. That covers the Mars ferry needs for an orbit-to-orbit colonization ship.
Rough-Field Considerations on Mars
Now, the vast bulk of Mars's surface resembles fine, loose sand. Here on Earth, safe bearing load pressures for fine, loose sand (based on many decades civil engineering experience) is 0.1 to 0.2 MPa. Period. You must use the lower figure to design things in the absence of real soil test data from your actual site. So you must use the 0.1 MPa figure.
I looked at various ignition masses of "Starship" that are more-or-less appropriate to the Mars orbital ferry role, calculated their weights at Mars 0.384 gee, and divided by the soil bearing strength figure, to find the total tail fin landing pad areas that are required to support rough-field takeoff operations. They fall in the 45-55 square meter range. That's just the nature of rough-field operations on Mars, and it will have to be dealt with in the "Starship" design. They currently have a single handful of square meters, at best. Spacex will have to address this issue, sooner or later.
Heat Shield Considerations
The final thing to worry about is the "Starship" heat shield. I presume there will be PICA-X ablative on the windward surfaces, nosetip, and leading edges. Entry from Mars orbit is about half the speed from Earth orbit, so the heat shield will likely fly 3 or 4 times (maybe more!) before being used-up. There will have to some way to refurbish this on Mars (in the cold and the near-vacuum) using materials brought from Earth. Spacex will have to eventually address this issue as well, if they ever use the "Starship" for this purpose.
That still begs the question of landing stability, since the vehicle is tall and narrow, and the gear is tripod, not quadruped. Landing fields will have to be very level, very flat, and very free of big boulders. Period! Spacex will have to face up to that issue eventually, regardless of what purpose their "Starship" gets used for, on Mars.
That's the problem with me being a real engineer. I tend to worry about the damndest real-world things. How very inconvenient!
All that being said, it looks to me like "Starship" would make a very good ferry to load and unload orbit-to-orbit transports, at both ends of the Earth-Mars journey. That means we are NOT barking up the wrong tree in looking at design approaches for orbit-to-orbit colonization transports.
Figure 1 is an image of the spreadsheet worksheet I used to figure these numbers. Inputs are highlighted yellow.
Figure 2 is a plot of the significant results for payload carried and propellant required. Figure 3 is a plot of the results for sized landing pad areas.
|Payload mass delivered to LEO||Cost per payload kilogram|
|53 metric tons||$95/kg (1971 dollars)|
|Wet Mass||2,040,816 kg|
|Payload Bay||7 m dia|
18 m high
|Service Life||x100 flights|
over 10 years
|Total Thrust||25,795,300 N|
|Specific Impulse||347 sec|
|Total Thrust||111,796 N|
SERV stands for Single-stage Earth-orbital Reusable Vehicle. This was a Chrysler study produced when NASA asked for proposals for a Space Shuttle. However, NASA made it clear that it was mostly interested in winged Shuttles. Chrysler was the only one who bothered with a non-winged proposal, and NASA returned the favor by not giving the SERV any serious consideration at all.
The SERV was shaped like an Apollo Command module magnified to a seven times larger scale. Just like the SSASTO it surrounded the aerobraking heat shield on its butt with an annular aerospike engine. Unlike the SSASTO the SERV's heat shield had hatches for the landing gear and turbofan lift engines. The aerospike engines had hatch covers, but they did not penetrate the heat shield.
The ballistic aerospike flight could aim for a landing site within a 15 kilometer diameter circle, but that was not good enough for NASA's specifications. That's whe the turbofan lift engines were added to the design. This allowed it to get within 75 meters of its aim point. The ability to use the atmosphere for oxidizer made the difference.
For uncrewed missions the SERV would carry a cylindrical payload module with a tiny nose cone on the top, and deliver it to orbit. For crewed missions the SERV would also carry on top a Manned Upper-stage Reusable Payload (MURP) spaceplane, capable of an aerobraking re-entry.
As interest in the SERV wained, Chrysler desperately invented new modules for it to carry. There was a tiny modifed Apollo Command module so the cargo version could also carry crew, a long aerodynamic spike that would lower the drag and increase the payload, and a plan to use the SERV as a sub-orbital airliner capable of carrying passengers from Heathrow to Sydney in three hours instead of twenty-two. Oh, and a solid-core nuclear upper stage suitable for a Mars mission or transporting outrageous payloads to Luna.
The MURP was based on the Northrop HL-10. It had a spray-on silicon ablative skin which was peeled off and refreshed after every mission. There were two MURP designs: the the D-10 and the D-34. Since the cylindrical cargo pod is a more efficient use of space, the D-10 has a lower mass than the D-34.
|Internal cargo||5 m3||85 m3|
|Cargo Pod||80 m3||n/a|
|Mass||11,640 kg||16,150 kg|
|Payload mass delivered to LEO||Cost per payload kilogram|
|91 metric tons||$22/kg to $33/kg|
Star-Raker is from a 1970's Rockwell International study, one of the many proposals on how to boost into orbit the outrageous payload requirements of a multi-kilometer solar power satellite (SPS). They were figuring on about 37,000 meric tons per SPS, and they wanted a constellation of 60 of them. For the project they estimated boosting 74,000 metric tons per year (2 SPS/year).
Star-Raker was a single-stage-to orbit airbreathing horizontal takeoff and landing craft (HTO-SSTO). The gross mass would be about 2,268 metric tons, the payload mass was about 91 metric tons, and it was claimed it would have a boost turnaround time of about a day and be really really cheap. Keeping in mind that at the time Rockwell was also claiming that the Space Shuttle would have a two-week turnaround and be really really cheap, which turned out to be somewhere between irrationally optimistic and an assurance from a used-car dealer. It was to be capable of delivering its payload into a 550 kilometer equatorial orbit.
To manage the proposed schedule of boosting the payload for two SPS per year would need about 815 flight per year, or 2.2 flights per day. This assumes a fleet of more than one Star-Raker.
Horizontal takeoff and landing, and single-stage were design choices due to the need for rapid turnaround. Having to fish stages out of the ocean, haul them to the launch site, refurbish, and re-stack them would make it impossible to have a single-day turnaround. To save mass the take-off wheels would be jettisoned at the end of the runway and recovered. For landing lighter internal landing gear is used, since by then the craft will be lighter by many metric tons of absent payload and burnt fuel.
It has a "wet-wing" design, that is, the wing is the fuel tank. The body of the craft is reserved for the payload. It was to be capable of taking off and landing on a 2,500 meter runway.It is an air-breather using atmosphere for oxidizer up to the point where the air is too thin at thirty kilometers altitude (ten supersonic-turbofan/airturbo-exchanger/ramjet engines with a combined thrust of 6.2×107 newtons thrust). For the last portion of the boost it switches over to rocket engines (three rockets with 1.4×107 newtons thrust each). The jet engine air inlets will be closed by retractable ramps while the craft is under rocket flight and during ballistic re-entry. From zero to 1,800 m/s it will be using airbreathing propulsion, from 1,800 to 2,200 m/s it will use both airbreathing and rocket propulsion, and from 2,200 m/s to orbit it will use only rocket propulsion.
It would also be capable of making trips as a conventional cargo aircraft. For instance, from the launch site to a site where the payload had been assembled, and back to the launch site. It saves on having to ship the payload to the launch site, but I question the wisdom of risking an expensive HTO-SSTO craft when a less expensive and more expendable cargo plane would suffice. The entire nose (including crew compartment) swings open to expose the cargo hatch (which must be scary for the crew when the playload is released into orbit). This allows it to be loaded from a conventional cargo platform. Cargo floor is designed similar to a C5-A military transport aircraft.
There was another design tailored for delivering payload into polar orbits, which would reduce the payload mass. Polar orbits are expensive in terms of delta V, but are necessary for Department of Defense spy satellites.
Report can be found here.
fictional surface-to-orbit reusable shuttle featured in the movie 2001 A Space Odyssey (1968).
This is bitterly ironic, since Pan American World Airways went bankrupt in 1991, before many of our younger readers were born. For that matter, the movie was suppose to take place in the far-flung future year of 2001, Clavius moon base and all.
Actually as it turns out, the Pan Am clipper was called "Orion" because originally it was going to be an honest-to-Pournelle surface launched nuclear pulse vehicle. I read that while Arthur C. Clarke was working on the movie, he was contacted by some scientists who were still angry that Project Orion had be canceled in 1964 (they were only teeny-tiny A-Bombs, honest!). They asked Clarke if an Orion drive spacecraft could be used in the movie, to promote the concept.
So the Orion III was actually going to be a real Orion. Sadly Stanley Kubrick thought that sending Dr. Floyd into orbit on a series of nuclear detonations was hard to take seriously, so the Space Clipper was downgraded to a conventional liquid hydrogen - LOX rocket. The Discovery was considered for an Orion Drive as well, but that too was vetoed.
|Payload mass delivered to LEO||Cost per payload kilogram|
|450 metric tons||$2.30 to $5.40/kg|
|Gross mass||6,363,000 kg|
|Specific Impulse||455 s|
The Reusable Orbital Module-Booster & Utility Shuttle (ROMBUS) is from Frontiers of Space by Philip Bono and Kenneth Gartland (1969). This is a reusable plug-nozzle powered booster. It used an aerospike engine with the spike truncated and turned into an aerobraking heat shield.
Bono also created a passenger carrying variant named Pegasus, and a military troop carrier called Ithacus. When the concept lost support at NASA, Philip Bono designed a more modest concept, adding an aerospike engine to a Saturn V to create the SASSTO concept.
The vehicle is staged in the sense that it jettisons external hydrogen fuel tanks during the ascent phase. The tanks have parachutes to increase the chance they can be reused.
After delivering its payload, the vehicle would typically spend 24 hours in orbit before the ground track passes close enough to the landing site. It lands using parachutes and rockets, with the final touchdown burn delivered by four engines running at 25% thrust for twelve seconds. The vehicle turnaround time would be about 76 days.
1. Payload 0.8 to 1.0 million pounds to orbit
2. Roll-control nozzle pairs
3. Vent lines for liquid hydrogen tanks (8)
4. Propellant utilization probes (8)
5. Booster centre body
6. Fuel tank support fittings (16)
7. guidance and electronic package
8. Attitude-control propellant tanks
9. Spherical oxidizer tank
10. Anti-slosh baffles
11. Fuel feed lines (18)
12. Quick-disconnect fittings (8)
13. Propellant turbopumps (18)
14. Peripherally arranged combustion chambers (36)
15. Oxidizer feed lines (18)
16. Liquid hydrogen tank for entry cooling
17. Turbine discharge lines (18)
18. Turbine discharge port
19. Oxidizer-tank-pressurization helium bottles (4)
20. Propellant tank for retro-thrust
21. Isentropic-expansion plug nozzle
22. Retractable landing legs (4)
23. Regeneration-cooling tubes
24. Liquid Oxygen Tank sump
25. Solid motors for thrust augmentation (4)
26. Liquid hydrogen manifold
27. Fuel manifold valve for liquid hydrogen tanks (8)
28. Attitude-control propellant tanks (4)
29. Centrebody recovery components
30. Cylindrical liquid hydrogen fuel tanks (8)
31. Tank recovery thermal protection (4)
delivered to LEO
|550 metric tons||$59/kg to $600/kg|
|Wet Mass||12,799,000 kg|
|Dry Mass||1,333,000 kg|
|Max Accel||4.21 g|
|Wet Mass||4,823,000 kg|
|Dry Mass||465,000 kg|
|Max Accel||5.2 g|
|Stage 1||12,799,000 kg|
|Stage 2||4,823,000 kg|
Sea Dragon was designed by Robert Truax in 1962 to be a low-cost heavy lift launch vehicle. A "big dumb booster", emphasis on "big". To reduce costs for launch pads and gantries, the vehicle was to be launched from the ocean. It would be towed out to the watery launch site, and the ballast tank in the first stage exhaust nozzle would be flooded. This would drag the tail down and the nose up, orienting the rocket into launch position.
At 150 m long and 23 m in diameter, Sea Dragon would have been the largest rocket ever built. To lower the cost of the rocket itself, it was designed to be build of inexpensive materials, specifically 8 mm steel sheeting.
The contruction techniques would be quite different than modern-day rockets. The latter are horribly damaged if they are touched by sea water, especially rocket engines. This is why SpaceX goes to the trouble of landing their reusable rockets on robot barges instead of letting them splash down in the ocean.
The design ground rules mandated a minimum payload of 450 metric tons delivered to a 600 kilometer orbit. For the reusable version of the vehicle, a 10 year useful life for the system was assumed.
The Sea Dragon project was shut down by NASA in the mid-1960's due to budget cuts.
delivered to LEO
|13 metric tons||$13,000|
(with zero re-use)
|Wet Mass||72,600 kg|
|Propellant Mass||35,700 kg|
|Inert Mass||36,900 kg|
|Structural Mass||7,260 kg|
|TSP Mass||7,260 kg|
|Average Isp||1,662 sec|
|NTR and shield||7,080 kg|
|Air breathing||2,270 kg|
This is from The Nuclear Thermal Turbo Rocket: A Conceptual High-Performance Earth To Orbit Propulsion System by John R. Bucknell. John Bucknell was Senior Propulsion Engineer for the Raptor full-flow staged combustion methalox rocket at Spacex and is currently the Senior Propulsion Scientist for Divergent3D in Torrance, CA developing additively manufactured vehicle technologies. Slides from his talk are here.
Mr. Bucknell notes that the only practical method of dramatically bringing down the cost of boosting payloads into low Earth orbit (LEO) is to lower investment and realize a large return on that investment. The implication is you want a low dry mass Single Stage to Orbit Resuable Launch Vehicle with a high payload mass fraction. This is challenging.
Nuclear thermal rockets (NTR) have the highest specific impulse and thrust of available rockets. But the thrust-to-weight (T/W) ratio is poor since the blasted thing needs heavy radiations shielding. This really cuts into the payload fraction.
NERVA had a T/W of 5:1, particle bed had T/W of 15:1, and Miniature Reactor Engine (MITEE) managed 23:1. Unfortunately chemical LOX/RP-1 engines can achieve 150:1 easy.
Air-breathing propulsion has much higher specific impulse than NTR. But air-breathing propulsion don't work if there isn't any air. Long before LEO is reached the air pressure will drop below the level required for the air-breathing engine. Air breathers can only operate for the first 25% of the ascent, after that you need a rocket.
Therefore Mr. Bucknell's concept is to have a hybrid engine that can start in air-breathing hypersonic turbine mode and switch to NTR mode when the air runs out. This is called Nuclear Thermal Turbo Rocket (NTTR).
From Mach 0 to 8 the engine is in air-breathing subsonic ramjet mode. Combustion is subsonic. The nuclear rocket hot-hydrogen thrust is used to spin the fan rotor, driving the turbines. The hydrogen escapes via the trailing edge of the thrust fan blades. The turbine thrust fan blade vary their pitch and the variable nozzle throat geometry adapt to the changing atmospheric conditions. The turbine compresses the atmosphere from the inlet cone and the hydrogen from the thrust fan blades into the combustor, where they are burned for ramjet thrust.
From Mach 8 to 14 the engine is in air-breathing scramjet mode. Combustion is supersonic. The thrust fan blades lock into the neutral position aligned with the vehicle axis (depitches). The variable inlet cone expands, as does the PYBB variable nozzle.
From Mach 15 on up, the engine is in nuclear thermal rocket mode. The variable inlet cone contracts shut. The only thrust is rocket thrust from hot hydrogen escaping the trailing edge of the thrust fan blades.
Late breaking news, Mr. Bucknell has an updated paper out: The Turbo Rocket - A high performance air-breathing rocket propulsion system with nuclear and chemical variants.
Among other things the payload mass fraction calculations have been updated. The payload fraction has risen from 19% to 44.8%, for the 11 meter core version with a thrust of 1,150,000 Newtons and a mission average specific impulse of 1,695 seconds. The paper presents a sample lunar mission for comparison purposes.
The paper also discusses a totally non-nuclear version, citing the lack of available nuclear thermal propulsion hardware. Because that version has sigificantly poorer performance, and because this is the ATOMIC rocket website, I'm going to ignore it.
Improvements to the Turbo Rocket Concept
The aspects of the design that have been improved from the first paper are:
- Trajectory Optimization
- Scaling Sensitivity
- Increasing reuse through improving aerobraking performance
- Extending Airbreathing Operation
The first paper had plain vanilla unoptimized trajectory called Turbo Rocket Reference Trajectory MkI. This paper has the new and improved Trajectory II, which maintains inlet conditions for best air-breathing performance up to Mach 14. It also minimized airframe drag in pure rocket mode from Mach 15 to 25.
The first paper had a wet-mass (GLOW or gross lift-off weight) 74,400 kg (160 klb) spacecraft with a core diameter of 3.66 meters, since with Trajectory MkI increasing the core to 5 meters reduces the payload fraction from 25.6% to 19.9%. Not good. The reduction is due to aerodynamic drag.
However, with the new and improved Trajectory MkII, increasing the core to 5 meters actually increases the payload fraction from 30.6% to 32.4%. So it is a win-win.
Now for a 445,500 kg (982 klb) GLOW spacecraft, the optimal core diameter is 11 meters. Payload fraction is 44.8%, which is fantastic! Table above includes some SpaceX boosters for comparison. SpaceX is nowhere near as good, but by the same token Elon Musk is not allowed to use nuclear rockets.
The NTTR was analyzed assuming a nuclear rocket designed around the Miniature Reactor Engine (MITEE) using highly enriched uranium (HEU, 98% Uranium 235). Actually I'm not sure that is accurate. 20%-85% U235 is highly enriched uranium. 85%-100% U235 is Weapons-Grade Uranium.
Which explains the report seriously looking into several other nuclear engines which use low enriched uranium (LEU or < 20% U235). Report says The availability of these reactors allows development with conventional nuclear fuel and doesn’t require the oversight required for highly enriched fuel. Translation: those reactors use the relatively cheap off-the-shelf commercial nuclear fuel, and you do not need an army of on-site killer SWAT teams to prevent terrorists from ripping off some HEU and making their very own terrorist nuclear bombs.
Sample Lunar Mission
A NTTR launch from Terra into LEO consumes about 42% of the GLOW, with 44% remaining for payload. Which obviously means a second NTTR could boost a complete propellant refueling load for the first ship (refuel load needs 42% GLOW of second NTTR, and it can boost 44%. 2% to spare). That would give the first ship enough delta V to go to Luna, land a large payload (68,000 kg) on the lunar surface, then lift-off and travel back to LEO (with 18,300 kg payload).
Table above has the details about the mission.
|Payload mass delivered to LEO||Cost per payload kilogram|
|1,000 metric tons||??|
Anthony Tate has an interesting solution to the heavy lift problem. In his essay, he says that if we can grow up and stop panicking when we hear the N-word a reusable closed-cycle gas-core nuclear thermal rocket can boost huge amounts of payload into orbit. He calls it a "Liberty Ship." His design has a cluster of seven nuclear engines, with 1,200,000 pounds of thrust (5,340,000 newtons) each, from a thermal output of approximately 80 gigawatts. Exhaust velocity of 30,000 meters per second, which is a specific impulse of about 3060 seconds. Thrust to weight ratio of 10. Engine with safety systems, fuel storage, etc. masses 120,000 pounds or 60 short tons (54 metric tons ).
Using a Saturn V rocket as a template, the Liberty Ship has a wet mass of six million pounds (2,700,000 kilograms). Mr. Tate designs a delta V of 15 km/s, so it can has powered descent. It can take off and land. This implies a propellant mass of 2,400,000 pounds (1,100,000 kilograms). Using liquid hydrogen as propellant, this will make the propellant volume 15,200 cubic meters, since hydrogen is inconveniently non-dense. Say 20 meters in diameter and 55 meters long. It will be plump compared to a Saturn V.
Design height of 105 meters: 15 meters to the engines, 55 meters for the hydrogen tank, 5 meters for shielding and crew space, and a modular cargo area which is 30 meters high and 20 meters in diameter (enough cargo space for a good sized office building).
A Saturn V has a dry mass of 414,000 pounds (188,000 kilograms).
The Liberty Ship has seven engines at 120,000 pounds each, for a total of 840,000 pounds. Mr. Tate splurges and gives it a structural mass of 760,000 pounds, so it has plenty of surplus strength and redundancy. Add 2,400,000 pounds for reaction mass, and the Liberty Ship has a non-payload wet mass of 4,000,000 pounds.
Since it is scaled as a Saturn V, it is intended to have a total mass of 6,000,000 pounds. Subtract the 4,000,000 pound non-payload wet mass, and we discover that this brute can boost into low earth orbit a payload of Two Million Pounds. Great galloping galaxies! That's about 1000 metric tons, or eight times the boost of the Saturn V.
The Space Shuttle can only boost about 25 metric tons into LEO. The Liberty Ship could carry three International Space Stations into orbit in one trip.
Having said all this, it is important to keep in mind that a closed-cycle gas-core nuclear thermal rocket is a hideously difficult engineering feat, and we are nowhere near possessing the abilty to make one. An open-cycle gas-core rocket is much easier, but there is no way it would be allowed as a surface to orbit vehicle. Spray charges of fissioning radioactive plutonium death out the exhaust nozzle at fifty kilometers per second? That's not a lift off rocket, that's a weapon of mass destruction. However, see the Nexus.
There is an interesting analysis of the Liberty Ship on Next Big Future.
|Payload mass delivered to LEO||Cost per payload kilogram|
|1,500 metric tons||??|
This is from some fragmentary circa 1964 documents uncovered by The Unwanted Blog.
A Convair concept for an all-chemical Nexus SSTO launch vehicle with a second stage using open-cycle gas-core nuclear thermal rockets. Presumably the designers thought that the chemical stage would loft the second stage high enough so that the twin plumes of incandescent radioactive death would be diluted into plausible deniabilty.
|Payload mass delivered to LEO||Cost per payload kilogram|
|4,600 metric tons||??|
This is from some fragmentary circa 1964 documents uncovered by The Unwanted Blog.
This monster is the Uprated GCNR Nexus grown to three times the size. The document says that it can deliver 453 metric tons not to LEO, but to Lunar orbit. Doing some calculations on the back of an envelope with my slide rule, I estimate that it can loft 4,600 metric tons into LEO. And also with a proportional increase in radioactive exhaust.
A bit over 122 meters tall with the second stage having a diameter of 37 meters. Total wet mass of 10,900 metric tons. Second (nuclear) stage wet mass 5,900 metric tons for the Lunar orbit configuration. Dry second stage at Lunar orbit has a mass of 450 metric tons. The LEO configuration will be different.
The chemical stage has a total delta V capacity of 2.4 km/s. The gas core engines have a specific impulse rating of 2,220 seconds. The gas core stage in Lunar orbit configuration has a total delta V capacity of 21.8 km/s.
|Payload mass delivered to LEO||Cost per payload kilogram|
|27,000 metric tons||??|
This extreme heavy lift vehicle appears in Beyond Tomorrow by Dandridge Cole of "Macrolife" fame (Amherst Press 1965). The best place to watch lift-off is from an adjacent continent. That engine looks like it could accidentally vaporize Florida. They better work on the cargo handling system, though. Loading it crate by crate by helicopter is too much like eating a bowl of rice with tweezers one grain at a time.
Mr. Cole assumes that the economies of scale would dictate such a huge rocket to keep up with the orbital boost demands of the far-flung futurstic year 1990. The wet mass would be 50,000 tons. If the propulsion system had a specific impulse of 3,000 seconds, it would have a propellant fraction of 0.7 and a payload mass of 60 million pounds (27,000 metric tons). Or it could soft-land a smaller payload mass of 20,000 metric tons on Luna. If the propulsion system was weaker, say a specific impulse of 1,500 seconds, it would have a propellant fraction of 0.5 and a payload of 20 million pounds (9,000 metric tons). That propellant fraction doesn't make sense to me, I'll have to do the math.
The design is winged, for controlled aerodynamic Earth landing (now that would be a sight to see). Water take off and landing because there isn't a runway in the world that could survive that monster.
|Type||Payload mass delivered to LEO||Cost per payload kilogram|
|Planetary||3,000 metric tons||??|
|Super||8,000,000 metric tons||??|
|Payload to LEO||Cost per kilogram|
|1,000 metric tons||??|
|Engine Type||Clean Fusion Orion|
|Engine Thrust||3,000,000 N|
|Propellant Mass Flow||10 kg/sec|
|Total Thrust||30,000,000 N|
|Total Propellant Mass Flow||1,000 kg/sec|
|Exhaust Velocity||30,000 m/s|
|Specific Impulse||3,060 secs|
|Payload to Orbit|
|Payload to Orbit|
|Inert Mass||600,000 kg|
|Dry Mass||1,600,000 kg|
|Propellant Mass||1,100,000 kg|
|Wet Mass||2,700,000 kg|
This is a species of Orion drive, including the useful ability to boost absurdly huge masses of payload into orbit. But with the attractive difference of not using dirty fission explosives for propulsion. It uses fusion explosions, triggered by convergent shock waves from chemical high explosives. Meaning there is zero radioactive fallout and arguably no problems from the Nuclear Test Ban Treaty. Yes, there will be some neutron radiation but you can't have everything.
The performance is very similar to the gas-core nuclear rocket Liberty Ship. But without the Liberty Ship's huge load of highly enriched uranium fuel, aka flying nuclear disaster waiting to happen. The Thermonuclear Orion's fuel would be non-radioactive deuterium and/or tritium. Both ships have approximately the same thrust (about 30,000,000 Newtons), approximately the same exhaust velocity (about 30,000 m/s, Isp around 3,060 secs), and approximately the same propellant mass flow (about 1,000 kg/sec).
Since they have the same exhaust velocity, both could manage the delta V for orbit (8,000 m/s) with a reasonable mass ratio of 1.3, or the delta V for orbit plus a powered landing (15,000 m/s) with a still reasonable mass ratio of 1.65. Which among other things means you don't have to deal with the design and maintenance nightmare called multi-staging, unlike pretty much all chemical rockets.
The amount of payload that can be carried depends upon design assumptions. As an example: the Liberty Ship was scaled to have the same mass as a Saturn V, but instead of the Saturn's top-notch payload of 118 metric tons, the Liberty Ship could boost a jaw-dropping 1,000 metric tons! Eight and a half times as much payload in one trip. And be resuable to boot. The Thermonuclear Orion's payload would be similar. Meaning a single launch could boost into orbit three International Space Stations and have enough spare payload capacity to squeeze in one Mir.
However, unlike the Liberty Ship, the Thermonuclear Orion will have severe design problems when it comes to landing the blasted thing. You see, when an Orion propulsion charge explodes in normal operation, the ship moves away from the explosion. Sadly, when landing, the ship will move into the nuclear explosion. For a conventional Orion using nuclear fission charges this would be suicide. The Thermonuke Orion might be able to get away with landing, since the fusion detonations are more like micro-explosions inside a mass of liquid hydrogen propellant.
|Payload mass delivered to LEO||Cost per payload kilogram|
|17 metric tons||??|
This appears to be an early version of Dr. John Slough magneto inertial fusion rocket.
The critical limitation for the exploration and development of space stems from the fact that existing propulsion technology has not achieved cost effective payload delivery to low earth orbit, let alone deep space. This is largely due to the low exhaust velocity provided by chemical combustion compared to that required for spacecraft orbit and planetary travel. The large increase in velocity can be obtained by ionizing and heating ambient air, or an onboard propellant in space The large increase in specific energy required for such a system can only come from a nuclear fuel, and it is proposed here that it be provided by the direct coupling of energy from a fusion based reciprocating engine to the propellant stream. This heating is a natural byproduct from incorporating the propellant into a plasma liner that is used to magnetically compress a magnetized target plasmoid via an oscillating compression coil. The reciprocating nature of the system also provides for an efficient, direct method for extracting the electrical power needed for target plasmoid formation and heating. An experiment is currently underway to produce the target plasmoid at the conditions required. 2D MHD calculations based on plasma liner compression of this plasmoid were carried out. A description of the operation of the fusion engine and the results from 2D MHD calculations of the fusion cycle are presented.
In the NTR a cooling fluid or propellant is passed through a core of material that has been heated by fission. This makes the NTR effectively a heated gas rocket. Since the NTR is a heat transfer rocket, the propellant can be selected to maximize performance of the propulsion system. With the present limitations of materials, NTR gas temperatures cannot exceed chemical propulsion gas temperatures but the use of a low molecular mass propellant provides for an exhaust velocity much greater than that of chemical rockets. The importance of a higher Isp is made evident by the NTR. The use of hydrogen provides for an increase in Isp from ~ 300 s for a high Isp chemical rocket to 900 s for an NTR based on the particle bed reactor (PBR).
With Δv for a typical orbit velocity with losses (~ 9 km/sec) this would reduce the propellant mass Mprop by an order of magnitude for a given spacecraft mass Msc. Unfortunately the spacecraft mass (payload, structure, avionics, tankage etc.) increases due to the increase in tank mass required for the low mass density propellant (H2). The specific gravity of liquid hydrogen is around 0.07, compared to 0.95 for an O2-H2 chemical engine.
Ultimately nuclear fission propulsion concepts have too many disadvantages when compared to chemical rockets for the Earth to Orbit (ETO) mission. While NTRs have significantly improved Isp over chemical rocket engines, chemical rocket engine thrust-to-weight ratios greatly exceed the NTR values, and NTR concepts do not have sufficient engine thrust-to-weight ratios to compete. The relatively low power density and specific power of NTR concepts is due to the mass required for efficient turbulent convective heat transfer in addition to propellant tankage.
Fusion nuclear, at least in the form that the world has pursued with virtually all its resources – the tokamak, is wholly inappropriate for the ETO mission. The primary reason is that the singular objective of fusion effort to date is the generation of electric power in the form of ~ 1 GWe power stations. The threshold size of a steady state fusion reactor to achieve the required power for ignition, while maintaining safe, protective shielding is quite large. This has driven the scale, capital costs and time for developing fusion power to levels that are well beyond what would appropriate for propulsion.
The straight forward application of a reactor based fusion-electric system creates a colossal mass problem for space application. A detailed analysis for the most compact tokamak concept – the spherical torus, spacecraft masses of 4000 MT were projected. The maximum launch mass would need to be less than 200 MT if current chemical rockets are used.
A practical path to fusion propulsion can only be achieved by creating fusion conditions in a different regime at much smaller scale (r ~ a few cm). For small scale fusion systems, such as the reciprocating fusion cycle based on the magnetic compression of plasmoids considered here, the possibility of a near term application to propulsion becomes feasible.
The fusion concept to be employed in the fusion engine to be described takes advantage of the very compact, high energy density regime of fusion employing a compact toroidal plasmoid commonly referred to as a Field Reversed Configuration (FRC). The reciprocating pulsed fusion of the FRC has several attractive features for space propulsion applications. This particular system can be made electrically very efficient, which allows for operation at the lower fusion gain appropriate for space. The ability to compress and expand the FRC plasmoid after fusion burn provides for a mechanism to extract electrical power for the driver by direct conversion from flux compression/expansion. For the ETO application of fusion power, the FRC plasmoid compression is achieved via a magnetically driven plasma liner (see Fig. 1). The exhausting of the liner plasma through a magnetic nozzle provides for efficient conversion of the fusion particle energy (4He ion at 3.5 MeV) into directed propulsive power at high thrust. A description of what is entailed in this process will now be covered in more detail.
II. Reciprocating Fusion Engine
A basic requirement for any propulsion system to be employed for the ETO mission is the achievement of a thrust to weight ratio greater than unity for take-off. If one is restricted to employing a low Isp propellant for thrust, the problem becomes one of fuel mass as this will then dominate the lift-off mass. The big payoff for the ability to achieve a higher Isp comes in the form of a much higher payload to lift-off mass ratio. The problem in the past for non chemical based propulsion has been to keep the engine mass low, otherwise the advantage of lower liftoff mass quickly disappears. For a fusion based system, this problem becomes even more difficult. The complication of the requirements for achieving and sustaining a fusion plasma must be solved as well as an efficient mechanism for converting the fusion energy into directed propulsive energy flow. The exhausting of the fusion plasma itself would not be a suitable propellant for launch due to the mismatch in Isp. A 10 keV fusion ion would have a directed velocity of vd ~ 106 m/s. This is far more than that desired for ETO (vd ~ 104). The consequence is a much reduced thrust for a given engine mass. With a fusion electric system, while the production of the plasma propellant could be made efficient, a very large mass would be required for the reactor conversion of the thermal energy of the fusion products into electric energy.
What is proposed here is a system that can provide for the direct conversion of the fusion energy into the propellant flow at the desired range of thrust and Isp, as well as provide a direct method for the production of the electric power needed for the formation and sustainment of the fusion plasma. The fusion engine part of the thruster is illustrated in Fig. 1. The plasmoid to be “combusted” is introduced via an axial guide magnetic field indicated by coil 1. The proper initial field, density, and temperature for the plasmoid is determined by the requirements for fusion gain as well as the range of compression that can be achieved. Both factors will be discussed in the next sections. It is believed that the compression ratio (initial plasma radius/final plasma radius) that can be achieved by the magnetic “piston” can be a large as 10 making the initial plasma requirements no greater than what can be accomplished currently in the laboratory. The piston is comprised of the flowing plasma sheath driven by an oscillating axial magnetic field produced by coil 3. This coil is part of an oscillating LC circuit formed by the coil inductance and a set of capacitors. The oscillation in the coil current produces an axial magnetic field that drives the plasma liner by inducing a large azimuthal image current in the plasma. The resultant JΘxBz radial body force drives the liner in and out. The magnitude of the field between the plasma liner and plasmoid can be much larger than the field experienced by the coil due to the large inertia of the plasma liner. In this way the high densities and temperature required for fusion can be achieved during the compression stroke.
The liner oscillation is damped primarily by ohmic dissipation in the plasma liner itself. The oscillation can be made to be self-sustaining or even increase in amplitude if the energy (pressure) created by the fusion reaction is sufficient to overcome these resistive losses during the compression cycle. This dissipation in the liner is of course precisely what is desired as the liner is also the propellant. It was found that the resistive damping is not large with only a small fraction of the total oscillatory energy thermalized per cycle. This means that plasmoid energy gain from fusion required to drive and maintain the oscillation need not be much greater than unity. Since only about a quarter of the fusion energy in the D-T reaction is in the form of plasma energy (the fusion alpha), the gain from fusion, Qfus ≥ 4. The goal would be to have a larger Qfus (~ 6) to be able to extract electrical energy directly from the fusion driven oscillator. This extraction would most likely be through a transformer coupling to the coil - capacitor circuit. The electrical energy is required for the formation and maintenance of the fusion plasmoid and other magnet and fueling systems.
The analogy can and has been made that the operation of the fusion engine is similar to that of an internal combustion engine. The analogy extends further in that the fusion burn (combustion) should occur at the end of the compression stroke to drive the oscillation efficiently and obtain the maximum work from the burn. This occurs naturally in the fusion engine driven by the reciprocating plasma liner (RPL). After injection into the burn chamber, the initial plasmoid ion temperature is in the range of 1 keV. After compression, the ion temperature climbs to 8-14 keV. After burn the temperature drops due to adiabatic expansion of the FRC plasmoid. Since the fusion yield scales as the plasma density squared, the density modulation from the RPL would vary the fusion output by an order of magnitude or more. Even more significant is the rapid change in the fusion cross-section as the ion temperature changes. The variation in fusion cross-section is plotted in Fig. 2, and it can be seen that the effect of compression and heating over this temperature range will act very much like the rapid ignition in a combustion engine.
The air intake inlet (coil 2 in Fig. 1) provides for new plasma for the RPL as plasma is lost out the magnetic nozzle. The RPL density is very high (~ 1025 m-3) so that the incoming air is instantly ionized as it enters. A high inlet pressure is maintained as in a ramjet by the high mach flow. The fusion engine is in effect a supercharged engine. A large mass flow however is not required (2-10 kg/sec). Large thrust can be generated by the high exit velocity (10-20 km/sec) after heating. The magnetic field is diverted into the coaxial inlet gap to entrain the new plasma on to the axial magnetic field of the compression/burn region. The dwell time of the plasma in the RPL determines the total heating. Even for sonic flow through the region, the plasma would experience several cycles of heating before exiting. In the MHD calculations it was possible to follow the RPL for a few oscillations, and there is a marked increase in liner energy and mass flow with each cycle. It should be possible to vary the Isp and thrust by the modulating the flow rate through the compression region This can be accomplished by varying the magnetic nozzle field strength (coil 4) or the inlet mass flow.
There are some significant departures in combustion engine analogy being made there. For one, the fusion fuel is not spent after a single burn cycle. The time of compression is short (a few microseconds) compared to the FRC plasmoid lifetime (several milliseconds). The fusion burn at the peak of the compression cycle can thus occur many times before the plasmoid is mostly spent and a new plasmoid needs to be injected and merged into the burn chamber. It is the liner plasma that flows continuously out the nozzle end to be lost with every compression cycle.
The magnetic insulation of the engine is also unique to the RPL fusion engine (RPLFE). The plasma liner is hot with respect to the engine wall (several eV). There is a very significant magnetic field however between the RPL and the wall to insulate the wall (several Tesla). The rapid flow of the plasma along the magnetic field through the chamber assures that cross field radial diffusion of the liner plasma is negligible. Similarly, the hot FRC plasmoid is insulated from the relatively cold RPL by a strong magnetic field barrier as well. The axial magnetic field insulation is maintained throughout the entire engine – from intake to nozzle exit. This magnetic insulation is what allows for such a high power density.
IV. Discussion and Conclusion
After two cycles the plasma liner produced an axial momentum flow of roughly 40 kN and a mass flow of nearly 4 kg/sec. This implies a mean flow velocity of roughly 104 m/s which satisfies the target Isp even without a nozzle. It is clear that both thrust and Isp will continue to increase from continued liner oscillations, so that it is interesting to compare the performance of the RPLFE based on the steady flow estimation made above to that of two historically significant chemical rocket engines (see Table I). While the RPLFE may not be as powerful as the larger chemical rocket engines, it doesn’t need to be as there should be essentially no significant propellant mass penalty at launch. In fact, due to the capability of operation at high Isp, it makes little difference whether the ambient air is used for the liner mass or an on-board propellant. For the space-like leg of the launch, a switch to an on-board propellant will be necessary in any case. There is a wide range of propellant options. A good choice would be lithium as it is easily stored as a solid and easily ionized.
Table I Launch
6,800 310 25 0.039 24 1,000 425 Atlas II
386 316 22 0.032 1.2 73 449 RPLFE 250 2500 > 500* 0.77** 17
# per engine *based on estimated engine mass (see text) **assuming negligible propellant mass
There is certainly the possibility for increased jet power for the RPLFE, as there has yet been no analysis of scaling the engine to larger size or power. The desire here was to stay within the range of target FRC plasmoid parameters that can currently be achieved. Given the limited amount of experimentation with pulsed magnetized plasmoids, there is quite likely room here for further gains in output if needed.
In the table the mass for the RPLFE was very crudely estimated as follows. The main source of mass comes from the requirements for the on-board energy storage for both the fusion engine start-up and FRC plasmoid formation. The capacitor energy storage in the LC oscillating drive coil is 250 kJ. There would be a similar amount in the FRC formation system. Assume an additional factor of two for other electrical systems and a safety margin for a total stored energy capacity of 1 MJ. Modern high voltage pulsed energy storage capacitors store up to 0.4 kJ/kg for a mass of 2.5 MT. Even though the magnets and coils are small, the ancillary cooling and shielding systems may not be. A reasonable assumption would be another 2.5 MT of mass for the rest of the engine for a total of 5 MT.
There are several other factors that must be considered in determining the suitability of the RPLFE for ETO. Principle among them must be a discussion of the shielding requirements. The idea here is to basically allow the fusion neutrons to escape. There will be absorption in the liner and magnets, and the effect of these neutrons needs to be understood. There can be significant standoff (up to 20 meters) for the FRC plasmoid formation system, but there will need to be shielding introduced to protect the crew and sensitive electrical equipment.
There are other fusion cycles with much better performance in terms of fusion product energetics. D-D fusion has two pathways for the fusion reaction that are roughly equal in probability. One results in two charged particles (H and T), and the other with a charged particle (3He) and a neutron. More than three quarters of the fusion energy would thus be available for conversion, and the neutron production would be reduced by half. The fuel would also be much more readily available, and easier to handle. The D-D fusion cross section is lower, and increases at a higher temperature so that a larger system and stronger liner compression would be needed. At even higher ion temperatures (~ 70 keV) it becomes possible to use the D-3He fusion cycle where all of the fusion products (4He and H) are charged particles. This fusion system would however not be completely aneutronic due to the much smaller but significant D-D fusion reaction rate at this temperature. As ideal a fusion cycle as this would be, there is a considerable problem in the availability of the 3He isotope. There are only trace amounts on the Earth since the only source is the solar wind which is deflected by the earth’s magnetosphere. The moon however is believed to contain significant quantities embedded in the regolith (Nope). If the RPLFE proves out to be a viable propulsion vehicle, going to the moon to get fuel should be no problem.
|Type||Payload mass delivered to LEO||Cost per payload kilogram|
|Small||40,000 metric tons/year||$300/kg|
|Large||6,000,000 metric tons/year||$3/kg|
|System||Payload mass delivered to LEO||Cost per payload kilogram|
|StarTram||150,000 metric tons/year||43/kg|
The report properly points out that NASA's Space Shuttle did many wondeful things, but lowering costs sadly was not one of them. NASA proudly predicted that the proposed shuttle could boost payload into orbit for $260 US per kilogram. In practice the accurséd thing cost $18,000 per kilogram of payload, which was pathetic compared to the $5,000 per kilogram price of the non-resusable simple-as-dirt 1966-vintage Russian Proton booster.
Naturally researchers were motivated to find some alternative boost method that might lower the cost by a couple of orders of magnitude.
In theory electromagnetic acceleration should be far more efficient than using a disintegrating totem pole made of high exposives. However in the past applying electromagnetism to space launch took the form of guns, as in railguns and coilguns. Both of those are still not ready for prime-time, despite the military throwing lots of money at the project of turning them into weapons.
But the study author John Mankins said "What about magnetic levitation trains?" Good old MagLev. You know, the kind that was patented in 19-freaking-37 and which currently holds the speed record for rail vehicles? Technology that is actually being used in the real world in bullet trains is certainly mature technology.
The concept is called "MagLifter".
The bottom line is the MagLifter can provide the launch vehicle with a free 300 m/s of launch delta-V. Granted this is only about 3% of the total delta-V needed, but the cost savings are huge. It cuts the delta-V from the start of the launch, when the propellant cost per meter/sec is at its most expensive.
The paper has an analysis, comparing a sample single-stage to orbit rocket with the same rocket scaled down but using MagLifter. The scaled-down version saved 327 metric tons of wet mass, 24 metric tons of dry mass, and required only 4 rocket engines instead of 6.
Railguns and coilguns are typically short, since they have to fit on some sort of military vehicle. This means all the velocity has to be jammed into the projectile within the short length of the gun, meaning that the acceleration will be strong enough to smash an astronaut like a cockroach. It will also do nasty things to unliving cargo.
On the other hand since MagLifter is based on a railroad train, the accelerating segment can be, say, four kilometers long. This means velocity can be added at a much more leisurely pace and gentler acceleration. The advantage is that the astronauts don't die and the inert payloads do not need expensive reengineering.
Another advantage over railguns is that MagLifter does not expend lots of hardware with each launch (sabot, projectile heat shield, orbit insertion propulsion module).
And unlike the Space Shuttle, MagLifter does not require very high launch rates in order to achieve economical operations. Railgun launches are even worse, some concepts can only bring the price down to the goal by doing four launches per day.
The report estimates that current (1997) maglev train in the 300 miles-per-hour range costs about $10 to $20 million US per mile and $3 to $5 million US per train (payload of about 23 metric tons). Annual operations and maintenance cost around 1% of capital cost.
The MagLifter system has five major elements: Catapult, Structural Support Systems, Power Systems, Supporting Systems, and Launch Vehicles.
The catapult has thee major elements: Maglev Guideway, Accelerator-Carrier Vehicle, and Accelerator-Carrier Staging Facility.
This is the "rails" of the maglev railroad. It will be about 3 to 4 miles of maglev rails. 2.5 miles where the payload is accelerated, and 0.5 to 1.0 mile where the accelerator-carrier is frantically decelerated after the payload is launched. You have to be able to reuse the accelerator-carriers, those things are expensive. The acceleration segment is enclosed in a pressurized tube full of helium gas; since helium has low density, low drag forces, and a high speed of sound.
These are the "cars" that are accelerated by the railroad track. The launch vehicle is strapped to the accelerator-carrier with rapid precisely-controlled release mechanism. If the launch vehicle is extra-long, several accelerator-carriers will have to be linked like cars on a choo-choo train.
Each accelerator-carrier has cradles to give structural support to the launch vehicles during the acceleration phase. Mostly on the "rear" of the launch vehicle, so the carrier does not go shooting ahead while leaving the launch vehicle hovering in mid-air like Wile E. Coyote.
Accelerator-Carrier Staging Facility
This houses the operation control center, and the accelerator-carrier management center. This is where the launch vehicles are strapped to their carrier, and also contains the carrier servicing and maintenance facilities.
Structural Support Systems
This is the part that supports the maglev guideway. It is assumed to be mostly composed of a mountain, since building support towers two kilometers tall is a bit of a challenge. The guideway will either be on trestles set on the exterior of a mountain, inside a 'cut' made into the side of a mountain, or inside a tunnel in the mountain's interior.
It has three elements: Tunnel, Tunnel Environment Monitoring and Control Systems, and Launch / Exit systems.
The acceleration section of the maglev guideway is encased in a tunnel, to smooth things as the launch vehicle furiously accelerates. The deceleration section of the guideway has no encasing tunnel, but still needs trestles or something to support it.
Tunnel Environment Monitoring and Control Systems
The tunnel will be filled with a normal oxygen-nitrogen atmosphere at the start, but near the exit it will be filled with gaseous helium. This will provide a low-density low-drag medium as the launch vehicle exceeds Mach 1. The speed of sound in helium is also about 2.6 times what it is in ordinary atmosphere. This is a good thing because you do not want a sonic boom inside the tube.
The tunnel will need sensors and gas injectors to ensure the gaseous environment is arranged properly and the tube is clear of foreign objects.
Launch / Exit systems
This is the system that manages the separation of acceleration-carrier and launch vehicle, and their exit from the tube gaseous environment.
This is a bank of batteries, probably a superconducting magnetic energy storage system (SMES). It will be gradually charged up from the local power grid, and used to power the launch. It would be nice to generate the required power during the launch. But since the blasted thing sucks 10 gigawats for a whopping 20 seconds, generating the power during launch is out of the question. Unless you have an antimatter power plant up your sleeve.
Power Management and Distribution
This system has to manage the massive discharge of the SMES and direct each watt to the proper component over the 30 second launch. It better not slip up or the maglev is in for tons of high-voltage fun that will make a thunderbolt look like scuffing your shoes on the carpet.
The second law of thermodynamics says there will always be waste heat. If the maglev system is 80% efficient, this means the thermal management system has to deal with 40 gigaJoules of waste heat over three miles of catapult. Otherwise the entire thing turns into three miles of molten lava.
This is mostly the stuff crammed into the accelerator-carrier staging facility.
This handles staging for the launch vehicles, the payload, and the accelerator-carrier. It also handles mating the launch vehicle (including payload) with the accelerator-carrier, and performing maintenace on the launch vehicle following each flight.
Operations Control Center
The crew here control both the staging and launch operations. MagLifter operations rely upon rapid turn-around, low-cost (submarine-style) launch operations.
This is in charge of all those behind-the-scenes details vital to the operation. This includes maintenance on the access roads servicing the entire operation and temporary housing for the launch passengers.
The small-, moderate-, and large-scale rockets that transport the payload the rest of the way to orbit. These are winged like the Space Shuttle, so they can return to the launch site and be re-used.
(Prometheus Alpha Dry)
|Wet Mass||450,000 kg|
|Exhaust Velocity||6,700 m/s atmo/|
6,318 m/s methane
|Specific Impulse||683 sec atmo/|
644 sec methane
(min ramjet 165 m/s)
This is from a science fiction novel by Sir Arthur C. Clarke. Keeping in mind that Clarke was the Chairman of the British Interplanetary Society from 1946 – 1947, and again from 1951 – 1953. The performance data for the nuclear stage was taken from the classic paper The Atomic Rocket by A.V. Cleaver and L.R. Shepherd, published in the Journal of the British Interplanetary Society in a series of articles September 1948–March 1949.
|Payload mass delivered to LEO||Cost per payload kilogram|
|280,000 metric tons||??|
Brian Wang has come up with an innovative concept. He mulled over a couple of his articles from his blog The Next Big Future (specifically this one and this one) and had an idea. Remember that one of the best propulsion systems for boosting huge payloads into orbit is the Orion drive; were it not for the fallout, the EMP, and the Nuclear Test Ban Treaty.
Then Mr. Wang thought about Jules Verne's novel From The Earth To The Moon, and the giant cannon Columbiad.
You set off one solitary ten megaton nuclear device in a deep underground salt dome. Perched on top is an Orion type spacecraft. All the EMP and radiation is contained in the underground cave (as has been done with historical underground nuclear tests). And 280,000 TONS of payload sails into low Earth orbit. Not pounds. Tons.
I say "sails into orbit", but of course it is more like "slammed by thousands of gs of acceleration", so this has to be unmanned (any human beings on board would instantly be converted into wall-gazpacho). But 280,000 tons? That's about one thousand International Space Stations, an entire Space Elevator (see below), an entire Lunar colony, an orbital fuel depot that would make future NASA missions ten times cheaper, a space station the size of the one in the movie 2001 A Space Odyssey, or about one-tenth of a ecologically clean 1.5 terawatt solar power station.
I know that nuclear-phobes will have a screaming fit, but this concept deserves close consideration.
Karl Schroeder analyzes the concept here.
Mass Drivers are a way to use electromagnets to hurl, well, pretty much anything. But with respect to Surface To Orbit maneuvers, they can be used to accelerate spacecraft to assist their boost into orbit. They can also accelerate engine-less cannisters of cargo into orbit, if the mass driver is powerful enough.
They do have the side effect of turning a spaceport into an impromptu planetary fortress. After all, they are basically huge coil guns. This was popularized in the classic Robert Heinlein novel The Moon Is A Harsh Mistress.
The acceleration track has to be in vacuum, or air friction will do unfortunate things to the cargo cannister. Mass driver launchers on Terra have to be encased in a vacuum chamber, such a in the Bifrost Bridge. On Luna or other airless world they already have all the vacuum needed, you just have place a series of acceleration rings every few meters.
|System||Payload mass delivered to LEO||Cost per payload kilogram|
|Pournelle||? metric tons/year||$1.9/kg plus power plant amortization|
|Jordin Kare HX Laser Launch||3000 metric tons/year||$550/kg|
Details about Laser Launch can be found here.
Matter Beam points out that the system will also work with an orbiting spacecraft equipped with a powerful laser battery, sending a beam to assist a surface-to-orbit shuttle lifting off. This can come in handy if the planet does not have a ground based laser launch facility, for instance an exploration spacecraft orbiting an uninhabited planet helping one of its landing craft return to the ship. A warship could also use its laser weapon batteries to give a boost to its fighters and missiles during a space battle, but I digress.
An important thing to keep in mind is that a laser-launch site is functionally equivalent to a planetary fortress. It can hurl projectiles and use laser beams directly at any invading spacecraft.
This is from Laser Propulsion (1972)
This is a fairly standard laser launching setup. A high-powered laser is located at the launch pad. The spacecraft uses hydrogen propellant. Since hydrogen is regrettably transparent to most laser frequencies, it is seeded with some sort of powder to make it opaque. Otherwise the laser bolt would go sailing through the transparent hydrogen, not heating it at all, and fry the engine nozzle.
The laser energy heats the seed powder, which heats the hydrogen propellant. This is converted into high specific-impulse thrust by expanding the hydrogen through a nozzle. The report figures that the specific impulse that yields the highest payload boosted into LEO per total energy consumed lies in the range of 1,200 to 2,000 seconds.
The report figures that for a thrust-to-initial-weight ratio of 1.2 to 4.0, and with a specific impulse of 2,000 seconds, this will allow the rocket to have a whopping payload fraction of 0.20 to 0.40. This is fantastic! Most chemical rockets have payload fractions that are a miserable one-tenth of that, which implies the cost of a laser-launch vehicle per kilogram of payload could also be one-tenth of a chemical vehicle.
As with all laser-launch vehicles they require large electrical power plants to feed the hungry lasers. This will drive up the cost of the launch system, unless the power plant amortization is shared with other purposes (seawater desalinization or something valuable like that). But keep in mind that the launch vehicles are relatively low cost, each needs no expensive engine since the laser "engine" is based on the ground and can be shared by all the vehicles.
At theoretical maximums, the minimum energy required to transfer payload from Terra's surface into LEO is about 3×107 joules per kilogram. At 1972 prices this comes to about $0.044 per kilogram. Naturally the laser launch system will be nowhere near this cheap, but there is plenty of room between that and the Falcon 9's $2,720/kg
The hydrogen propellant along with the seed particles are injected into the exhaust nozzle through the porous walls, and are heated by the high powered laser beam. In other systems the injection system is replaced by a large slab of solid propellant, which is ablated by the laser beam into hot gas.
This system can also be used to propel spacecraft in LEO into trajectories to various destinations, or for ballistic transcontinental passenger services.
The report just looks at the performances and the simple costs of such a system. An over-all cost effectiveness study is beyond the scope of the report (translation: there was no funding for such a study).
The high-energy laser beam is directed to the propellant injection plate located inside the nozzle. The hydrogen propellant is seeded with something like carbon or natural uranium particles to render it opaque so it can absorb laser energy (this technique was previously studied for gas-core nuclear rockets). The propellant enthalpy and specific impulse are determined by dividing the beam power by the propellant mass flow rate. The hot propellant and vaporized seed material is expanded into space. A nozzle skirt directs the expansion to provide a more efficient conversion of thermal power into thrust.
Note that the laser beam does not have to be parallel to the thrust axis. The report figures that the rocket can be canted up to 45° off the laser beam and still work. The only limit is that the beam hits the propellant injection plate.
The skirt is protected from the high-temperature propellant by an opaque boundary-layer film, composed of more of the propellant+seed mix. This is injected through the porous or slotted walls of the skirt. This will reduce the specific impulse somewhat but it certainly would be a bad thing if the skirt was incinerated. The report figures that the skirt can be protected up to about a specific impulse of 5,000 seconds or so.
It is very important to keep the high-powered laser beam directed at the propellant plate on the rocket, otherwise the engine goes dead and the rocket plummets out of the sky. One solution is to have the laser beam direction slaved to tracking information sent by the rocket. The big laser has low-powered laser guidance beams parallel to the main beam, aimed at the rocket. The rocket skirt has laser detectors that can see the guidance beams. If the guidance beams start to drift off the detectors, the rocket sends radio signals to the big laser which allow it to get back on target. If the guidance beams drift too much off target, the high-powered beam will be reduced in power so it doesn't slice the rocket into bits.
Remember that laser launchers can probably be used as impromptu planetary defense weapons.
The intensity of the laser beam (kW/cm2) depends upon the desired propellant injection enthalpy and flow rate per unit area. It will probably on the order of megawatts per square centimeter. Naturally if this is beyond the capability of a single laser, there is no reason that an array of several laser cannot be used. In fact this is probably a good idea anyway, to allow redundancy and permit gracefull degradation if one laser malfunctions. Relying upon a single huge laser means if it malfunctions there is no way to prevent the rocket from doing an imitation of Icarus.
Obviously the laser frequencies will have to be ones that can penetrate Terra's atmosphere (no vacuum frequencies) and the spaceport should be located where the weather is not prone to clouds, smog, turbulance, or other atmospheric things that can attenuate the beam (and also subject to the standard spaceport location restrictions). Laser will need a line-of-sight to the rocket during boost phase.
The report figures that the thrust-producing part of the system (the part actually attached to the rocket) will be no heavier nor more complicated than a conventional rocket engine.
The report assumes that the propellant absorbs all the laser energy. So the propellant enthalpy is
The specific impulse is related to the enthalpy by:
where they assume an overall nozzle expansion coeffcient CN ≈ 0.64 for various reason. Rearranging the equations gives:
which is plotted in Figure 2.
The laser beam power per thrust is calculated by:
and plotted in Figure 3.
The propellant mass fraction α is determined from the classic rocket equation
Here ΔVideal is the ideal mission velocity, e.g., 8,080 m/s for a low-orbit mission. It assumes a value of 1,070 m/s to account for atmospheric drag, and the fact the orbit will be elliptical rather than circular (because the rocket has to be visible to the ground laser during the entirety of the boost phase). The gravity drag is approximately g = 0.8 g0 which is conservative for a thrust-to-initial-weight ratios of 1.2 to 4.0.
Since tB = Ispα/k and substituting in equation (5) results in
Sove for α in terms of Isp and k gives Figure 4.
The payload mass fraction is calculated by using this propellant mass fraction and assuming a rocket structural weight fraction of 0.20. This results in:
which is plotted in Figure 5.
The laser beam energy per payload mass delivered to LEO is calculated using equations (3), (7), and (8) in the following equation:
which is plotted in Figure 6.
The electrical energy per payload mass depend upon overall beam efficiency Ee = Eb/ηb. The value of ηb includes the ground-based electric-laser conversion efficiency and the laser beam transmission efficiency. The ground-based electric energy per payload mass is shown for various efficiencies in Figure 7.
The required electrical powerplant capacity per unit mass of payload is calculated from
and plotted in Figure 8.
The dollar cost of energy per payload mass in LEO is calculated by assuming an electrical cost of $1.39×10-9 per joule ($0.005/kW-hr)The energy cost per payload mass becomes
and is plotted in Figure 9.
The dollar cost of liquid hydrogen propellant per payload mass is calculated by using a projected future (+1972) cost of $0.22 per kilogram. The propellant cost per pound of payload becomes:
A Lightcraft is a type of laser launch vessel. Air enters in through vents at the waist. A laser beam is shined at the parabolic mirror on the base where it flash-heats the air there into plasma. The plasma rapidly escapes out of the base creating thrust. More air enters in through the waist vents and the cycles start again.
Since it is using atmospheric gas for propellant instead of on-board propellant, and the mass of the engine is at the spaceport instead of being on-board, most of the mass of the spacecraft will be payload. Instead of being mostly non-payload like most other booster vehicles.
This is from Beamed-Energy Propulsion (BEP) Study. The report looks at three different types of laser launch systems which are reasonably mature. This means the payloads are pretty small, forty to eighty kilograms as most (40 kg ` six cubesats). The payload mass will rise with technological advancement.
The hope was that using optical and millimeter wave lasers to power propulsion systems would give high exhaust velocity and high thrust. Plus the advantage that the power plant is on the ground instead of adding penalty mass to the boost vehicle.
- LASER OPTICAL: mirrored cowl intercepts and focuses laser light from a ground-based installation to heat atmosphere or water propellant.
- LASER THERMAL: propellant is circulated inside a large heat exchanger (HX). The exchanger is heated by a visible-light laser beam from a ground installation.
- MILLIMETER WAVE THERMAL: same as laser thermal, except instead of a visible light laser a microwave laser is used instead.
All the designs use two ground laser installations. The first is the "boost" beaming station, it is optimized to propel the spacecraft from the launch pad to high altitude as fast as possible. The "main" beaming station located downrange is optimized to delta-V the spacecraft up to orbital velocity and orbital height.
This engine operates in air-breathing laser ramjet mode from launch up to the time it reaches Mach 7 and an altitude of 35 kilometers. Then is switches to rocket mode using water propellant.
In both modes visible laser light is interceped by the mirrored base of the boost vehicle and funneled into the cowl. There the laser energy heats up either atmospheric gases or water propellant. The hot propellant exists through the bottom of the cowl, providing thrust. In ramjet mode the spacecraft forebody directs atmospheric gases into slots on the top of the cowl. The gases are compressed and injected into the laser cavity. In rocket mode the slots are closed, and water from onboard tanks is injected into the cowl.
The vehicle assembly building and the laser boost beaming station are a single building, unlike the other two concepts. This is because the laser beam has to be directed upwards into the base of the vehicle. The other two concepts direct the laser beam at the ventral side of the vehicle.
Risks and issues:
If the mirrored surface of the base and inside the cowl is damaged or degradated, the 3000 watts per square centimeter of laser energy will quickly burn through and destroy the launch vehicle. Mirror damage can come from debris impact or erosion by the propellant plasma.
If the propellant plasma comes witin a few centimeters of the mirror surface, there will be excessive heating. This is because the mirror surface is not as refective to the heat frequency from the plasma as it is to the laser beam frequency.
Laser light reflected off the mirror can possibly reach the surface of Terra, which could damage the eyesight of people on the ground watching the launch.
A large external heat exchanger (HX) is heated by the ground-based laser installation. Liquid propellant (water and liquid hydrogen) from onboard tanks is heated inside the HX, and escapes through a conventional rocket nozzle to provide thrust.
Risks and issues:
In order to achieve the required heat transfer capabilities, the heat exchanger walls are very thin. This means the HX is very fragile. It can be broke by:
- Aerodynamic loading during ascent
- Thermal stress due to large temperature variations during launch
- The temperature gradient across the HX during normal operation
- The high internal pressure due to the superheated and expanding hydrogen propellant
MILLIMETER WAVE THERMAL
Risks and issues:
Basically the same as the laser thermal: problems with the heat exchanger.
|Payload mass delivered to LEO||Cost per payload kilogram|
- Does not require materials with extreme strength
- Can be located at any point on a planet's surface instead of just the equator
- Can be raised to heights lower than the level of geostationary orbit
- Requires large constant amounts energy
- If the power is interrupted, the entire tower comes crashing down
|Payload mass delivered to LEO||Cost per payload kilogram|
|1||2,000 metric tons||$3,000/kg|
|2||4,000 metric tons||$1,900/kg|
|3||6,000 metric tons||$1,600/kg|
You can find details about space elevators here.
The big limitations are: it must be sited exactly on the equator, and it is absurdly vulnerable to sabotage.
You can read all about the complicated equations required to calculate the annual payload lifting capacity of a space elevator here. A baseline Edwards-Westling 20 metric ton space elevator powered by a bank of solar panels could boost about 272 metric tons a year. If powered by a large nuclear reactor it could boost about 2,720,000 metric tons a year.
8.21 The beanstalk. Suppose we have a space station in geostationary orbit, i.e. an equatorial orbit with period exactly 24 hours. A satellite in such an orbit hovers always over the same point on the Earth's equator. Such orbits are already occupied by communications satellites and some weather satellites.
Now suppose a strong loop of cable runs all the way down to the surface of Earth from the space station. The cable must be long as well as strong, since geostationary orbit is more than 35,000 kilometers above the surface. We defer the question as to how we install such a thing. (A geostationary satellite has a period of 24 hours, and hovers above a fixed point on the equator. A geosynchronous satellite simply has a period of 24 hours, but can be inclined to the equator and reach to any latitude.)
Attach a massive object (say, a new communications satellite) to the cable down on the surface. Operate an electric motor, winding the cable with the attached payload up to the station. We will have to do work to accomplish this, lifting the payload against the downward gravitational pull of the Earth. We do not, however, have to lift the cable, since the weight of the descending portion of the loop will exactly balance the weight of the ascending portion.
Also, suppose that we arrange things so that, at the same time as we raise the payload up from the surface, we lower an equal mass (say, an old, worn-out communications satellite) back down to the surface of the Earth. We will have to restrain that mass, to stop it from falling. We can use the force produced by the downward pull to drive a generator, which in turn provides the power to raise the payload. The only net energy needed is to overcome losses due to friction, and to allow for the imperfect efficiency of our motors and generators that convert electrical energy to gravitational energy and back.
The device we describe has been given various names. Arthur Clarke, in The Fountains of Paradise (Clarke, 1979), termed it a space elevator. I, in The Web Between the Worlds (Sheffield, 1979), called it a beanstalk. Other names include skyhook, heavenly funicular, anchored satellite, and orbital tower.
The basic idea is very simple. There are, however, some interesting "engineering details."
First, a cable can't simply run down from a position at geosynchronous height. Its own mass, acted on by gravity, would pull it down to Earth. Thus there must be a compensating mass out beyond geosynchronous orbit. That's easy enough; it can be another length of cable, or if we prefer it a massive ballast weight such as a captured asteroid.
Second, if we string a cable from geostationary orbit to Earth it makes no sense for it to be of uniform cross section. The cable needs to support only the length of itself that lies below it at any height. Thus the cable should be thickest at geosynchronous height, and taper to thinner cross sections all the way down to the ground.
What shape should the tapering cable be? In practice, any useful cable will have to be strong enough to stand the added weight of the payload and the lift system, but let us first determine the shape of a cable that supports no more than its own weight. This is a problem in static forces, with the solution (skip the next half page if you are allergic to equations):
In this equation, A(r) is the area of the cable at distance r from the center of the Earth, A(R) is the area at distance R of geosynchronous orbit, K is the Earth's gravitational constant, d is the density of cable material, T is the cable's tensile strength, and f is the function defined by:
The form of the equation for A(r) is crucial. First, note that the taper factor of the cable, which we define as A(r)/A(R), depends only on the ratio of cable tensile strength to cable density, T/d, rather than actual tensile strength or density. Thus we should make a beanstalk from materials that are not only strong, but light. Moreover, the taper factor depends exponentially on T/d. If a cable originally had a taper factor from geosynchronous orbit to Earth of 100, and if we could somehow double the strength-to-density ratio, the taper factor would be reduced to 10. If we could double the strength-to-density again, the taper factor would go down to 3.162 (the square root of 10). Thus the strength-to-density ratio of the material used for the cable is enormously important. We note here the presence of the exponential form in this situation, just as we observed it in the problem of rocket propulsion.
We have glossed over an important point. Certainly, we know the shape of the cable. But is there any material with a large enough strength-to-density ratio? After all, at an absolute minimum, the cable has to support 35,770 kilometers of itself. The problem is not quite as bad as it sounds, since the Earth's gravitational field diminishes as we go higher. If we define the "support length" of a material as the length of uniform cross section able to be supported in a one-gee gravitational field, it turns out that the support length needed for the beanstalk cable is 4,940 kilometers. Since the actual cable can and should be tapered, a support length of 4,940 kilometers will be a good deal more than we need. On the other hand, we must hang a transportation system onto the central cable, so there has to be more strength than required for the cable alone.
Is there anything strong enough to be used as a cable for a beanstalk? The support lengths of various materials are given in TABLE 8.1.
Strength of materials
Material Density Tensile strength Support length (gms/cc) (kgms/sq.cm.) (kms) Lead 11.4 200 0.18 Gold 19.3 1,400 0.73 Aluminum 2.7 2,000 7.40 Cast iron 7.8 3,500 4.50 Carbon steel 7.8 7,000 9.00 Manganese steel 7.8 16,000 21.00 Drawn steel wire 7.8 42,000 54.00 Kevlar 1.4 28,000 200.00 Iron whisker 7.8 126,000 161.00 Silicon whisker 3.2 210,000 660.00 Graphite whisker 2.0 210,000 1,050.00
The conclusion is obvious: today, no material is strong enough to form the cable of a beanstalk from geostationary orbit to the surface of the Earth.
However, we are interested in science fiction, and the absolute limits of what might be possible. Let us recall Chapter 5, and the factors that determine the limits to material strength. Examining TABLE 5.1, we see that solid hydrogen would do nicely for a beanstalk cable. The support length is about twice what we need. It would have a taper factor of 1.6 from geosynchronous orbit to Earth. A cable one centimeter across at the lower end would mass 30,000 tons and be able to lift payloads of 1,600 tons to orbit.
Materials, potential strengths
Element pairs* Mol wt Bond Strength to (kcal/mole) strength weight ratio Silicon/carbon 40 104 2.60 Carbon/carbon 24 145 6.04 Fluorine/hydrogen 20 136 6.80 Boron/hydrogen 11 81 7.36 Carbon/oxygen 28 257 9.18 Hydrogen/hydrogen 2 104 52.0 Muonium/muonium 2.22 1,528 9,679 Positronium/positronium 1/919 104 95,576
* Not all these elements exist as stable molecules.
Unfortunately, solid metallic hydrogen is not yet available as a construction material. It has been made as a dense crystalline solid at room temperature, but at half a million atmospheres pressure. We need to have faith in progress. There are materials available, today, with support lengths ten times that of anything available a century ago.
Beanstalks are easier for some other planets. TABLE 8.2 shows what they look like around the solar system, assuming the hydrogen cable as our construction material.
Beanstalks around the solar system.
Body Radius of stationary Taper factor satellite orbit (kms) Mercury 239,731 1.09 Venus 1,540,746 1.72 Earth 42,145 1.64 Luna 88,412 1.03 Mars 20,435 1.10 Jupiter 159,058 842.00 Callisto 63,679 1.02 Saturn 109,166 5.11 Titan 72,540 1.03 Uranus 60,415 2.90 Neptune 2,222 6.24 Pluto* 20,024 1.01
* Since Pluto's satellite, Charon, seems to be in synchronous orbit,
a beanstalk directly connecting the two bodies is feasible.
Mars is especially nice. The altitude of a stationary orbit is only half that of the Earth. We can make a beanstalk there from currently available materials. The support length is 973 kilometers, and graphite whiskers comfortably exceed that.
Naturally, the load-bearing cable is not the whole story. It is no more than the central element of a beanstalk that will carry materials to and from orbit. The rest of the system consists of a linear synchronous motor attached to the load-bearing cable. It will drive payloads up and down. Some of the power expended lifting a load is recovered when we lower a similar load back down to Earth. The fraction depends on the efficiency of conversion from mechanical to electrical energy.
So far we have said nothing about actual construction methods. It is best to build a beanstalk from the top down. An abundant supply of suitable materials (perhaps a relocated carbonaceous asteroid) is placed in geostationary orbit. The load-bearing cable is formed and simultaneously extruded upward and downward, so that the total up and down forces are in balance. Anything higher than geosynchronous altitude exerts a net outward force, everything below geosynchronous orbit exerts a net inward force. All forces are tensions, rather than compressions. This is in contrast to what we may term the "Tower of Babel" approach, in which we build up from the surface of Earth and all the forces are compressions.
After extruding 35,770 kilometers of cable downward from geostationary orbit, and considerably more upward, the lower end at last reaches the Earth's equator. There it is tethered, and the drive train added. The beanstalk is ready for use as a method for taking payloads to geosynchronous orbit and beyond. A journey from the surface to geosynchronous height, at the relatively modest speed of 300 kilometers an hour, will take five days. That is a lot slower than a rocket, but the trip should be far more restful.
The system has another use. If a mass is sent all the way out to the end of the cable and then released, it will fly away from Earth. An object released from 100,000 kilometers out has enough speed to be thrown to any part of the solar system. The energy for this, incidentally, is free. It comes from the Earth itself. We do not have to worry about the possible effects of that energy depletion. The total rotational energy of the Earth is only one-thousandth of the planet's gravitational self-energy, but that is still an incredibly big number.
The converse problem needs to be considered: What about the effects of the Earth on the beanstalk?
Earthquakes sound nasty. However, if the beanstalk is tethered by a mass that forms part of its own lower end, the situation will be stable as long as the force at that point remains "down." This will be true unless something were to blow the whole Earth apart, in which case we might expect to have other things to worry about.
Weather will be no problem. The beanstalk presents so small a cross-sectional area compared with its strength that no imaginable storm can trouble it. The same is true for perturbations from the gravity of the Sun and Moon. Proper design will avoid any resonance effects, in which forces on the structure might coincide with natural forcing frequencies.
In fact, by far the biggest danger that we can conceive of is a man-made one: sabotage. A bomb exploding halfway up a beanstalk would create unimaginable havoc in both the upper and lower sections of the structure. The descent of a shattered beanstalk was described, in spectacular fashion, in Kim Stanley Robinson's Red Mars (Robinson, 1993). My only objection is that in the process the town of Sheffield, at the base of the beanstalk, was destroyed.
The central station of a space elevator is 35,786 freaking kilometers from the surface of Terra, way out in geosynchronous orbit. The orbital ring is more like 300 to 600 kilometers, in LEO. The difference is that we have yet to find a material we can manufacture which will support a 35,786 km strand of itself, while a 600 km strand is orders of magnitude easier.
The ring is spinning at 8 kilometers/sec or whatever, thus preventing itself from collapsing onto the surface of Terra by centrifugal force. The "elevators" are tethers extending from the ring down to Terra's surface. The tether is attached to the ring indirectly by superconducting magnets. So the tether stays "stationary" over one spot on the ground moving at a speed of zero km/sec, while being magnetically attached to a ring moving at 8 km/sec.