Space Transport and Engineering Methods/Space Elevator
Space Elevators have been a theoretical transportation method since 1895. The original idea is impractical to build. This step adds a much more practical design as a transport hub for getting from one orbit to another quickly and efficiently. Initial construction can use materials from Earth, but in larger sizes or locations beyond Earth orbit using local materials is assumed.
Skyhook Concept 
The popular concept of a space elevator is based on the original design proposed by Tsiolkovsky in the late 19th century. It involves a single tower/cable extending all the way past Geosynchronous (24 hour) Earth Orbit (GEO). If the center of mass is at GEO and matches the Earth's daily rotation it will appear to hang motionless relative to the ground. Getting to space in theory then becomes an elevator ride. There are several problems with this simplistic design:
- The depth of the Earth's gravity well (6378 g-km) exceeds the scale length of the best available materials (350 km for carbon fiber) by 18 times, which then requires a structure-to-payload mass ratio of 65 million to 1. This would require more carbon fiber than the world makes to lift a reasonable cargo mass, and would never be economical because it would take too long to transport sufficient payload to justify the massive cost of that much structure.
- It is of no use for delivering cargo to low orbits. Release points somewhat below GEO result in elliptical orbits with a low perigee, but lower circular orbits cannot be reached. It also is of no use transporting cargo from the ground when partially built.
- Even with a magnetically levitated elevator car running at 300 km/hr, it will still take 5 days to deliver one payload to GEO, and you can only deliver one payload at a time.
- A single cable catastrophically fails when hit by natural or man-made debris. A cable over 35,000 km long has a lot of area exposed to such hits.
The Skyhook concept addresses all these problems. Instead of a static cable that stays over a fixed location, it is a rotating cable in a moving orbit, much like two spokes of an imaginary wheel rolling around the Equator.
- A full orbit velocity Skyhook reduces the structural requirement to about 2868 g-km, because only the tip sees the full Earth's gravity. The center is in orbit and thus has zero acceleration load. This immediately reduces the theoretical mass ratio from 65 million to 3,620:1. There is still an exponential relation of mass to tip velocity. Since conventional rockets also have an exponential relation of mass to velocity, it makes sense to split the job between both, because the sum of two exponents, for example e^2 + e^2 = 14.8 is less than a single exponent of the combined powers, i.e e^4 = 54.6. The optimum division of work between the Skyhook and vehicle coming from Earth will depend on technical details and costs, but a simple division of half to each results in a theoretical Skyhook mass ratio of 60:1. A real design will be heavier, but 60:1 is a feasible starting point, where 65 million is not. Reaching half of orbit velocity for a single stage rocket with a life of many flights is quite feasible.
- Assuming the tip is at 1 gravity, a Skyhook with a tip velocity of 30-50% of orbit velocity has a radius of 500-1400 km. The center point needs to be that altitude plus enough that the tip does not dip into thick atmosphere and create drag (100-200 km). Releasing from the center of the Skyhook at 600-1600 km altitude allows access to low Earth orbits.
- A partially built Skyhook can still function because the remainder of the velocity is provided by the Earth vehicle. During construction the velocity split is more towards the Earth vehicle. This reduces payload mass, but it can still deliver some. In particular, if part of your payload is more Skyhook structure, that payload pays for itself in increased payload on later trips. This is a literal version of "lifting yourself by your bootstraps".
- The same fast elevator car at 300 km/s can reach the center in 2 to 5.3 hours. If your destination is high orbit or Earth escape, you do not have to ride the elevator at all. You wait half a rotation of the Skyhook and let go, at which point you are going at orbit velocity plus tip velocity. To imagine this, think of the top point of a bicycle wheel. It moves faster than the center relative to the ground. A half rotation takes only 10 to 20 minutes.
- Space debris cannot be eliminated. Even if all the man-made junk in Earth orbit is eliminated, the natural flux of meteors will continue. Therefore the Skyhook design has to take that into account. The most practical way to do that is to use multiple redundant cables to distribute the load such that cutting one or two is not catastrophic. The cables should be spaced far enough apart that any single object will only hit one or two strands. The strands should also be cross-connected periodically to distribute the load around a break. Repairing a break then becomes replacing a short segment of one strand. Since you have the ability to install segments during original construction, you are able to replace segments as a maintenance job.
Skyhook Applications 
The Moon and Mars have smaller gravity wells than the Earth, by ratio of 22 and 5 respectively, so Skyhooks with the same materials can do more of the transportation job relative to Earth. But in this step-by-step combined system example, getting off the Earth comes first. We will discuss the other locations here, but the actual construction will be delayed until easy transport to those locations is needed.
Earth Orbit Skyhook 
Like the previous assembly and processing factory steps, the Skyhook is not a single fixed design. It is more like a road or an airport, which starts out small but can be put to use right away and added to over time. The initial structural materials would likely be brought from Earth, but once carbon fiber and other materials can be manufactured at the processing factory, there is the opportunity to use them. If better materials are developed over time, they can be incorporated into the Skyhook by incremental replacement. As traffic increases in size and frequency, the Skyhook can grow along with it by adding more or thicker cables and increasing the tip velocity. Higher tip velocity requires more structure, but if the Skyhook is used more often, that added structure can pay for itself faster, and so makes economic sense.
The long term effects of zero gravity are known to be bad for the human body, and some industrial processes work better with gravity, so a rotating habitat will be useful in any case. This can serve as the seed of a Skyhook, which can grow by simply increasing the radius of rotation and tip velocity. There will be a low gravity area near the center for tasks which require it.
Lunar Skyhooks 
There are two systems that are feasible because of the small gravity well of the Moon. The first is a catapult system to launch bulk materials off the Moon. The catapult uses a rotating cable driven by an electric motor to throw payloads directly into Lunar orbit, where they are picked up by a collector system. The second is an orbiting Skyhook which can deposit and pick up cargoes at zero velocity close to the surface.
Catapults - Basalt fibers are similar to fiberglass in that they are an extruded mineral. They have a strength of 4800 MPa, or 80% of carbon fiber, and a density of 2.7 g/cc, or 50% higher than than carbon. Thus the scale length of Basalt fiber is 178 km, or about half that of carbon fiber. The dark areas (Maria or seas) of the Moon are covered in basalt lava , so there is an a very large supply of raw materials. If a Lunar catapult delivers basalt to an orbital processing factory, or spools of fiber already spun on the Moon, it would be possible to build a Lunar Skyhook out of local materials. The choice of Lunar basalt for a Skyhook would have to be compared to the higher performance carbon fiber brought from Earth or made from NEO carbon. Certainly for Lunar surface construction it would have the advantage of being very local.
Catapults could also be built on the Earth or Mars, but for Earth it would need to be placed above the atmosphere on a tall tower to get significant velocity. It is probably not the best method when compared to the alternatives. Mars is much smaller, has less atmosphere, and very tall volcanoes that a catapult can be placed on. So it is worth considering placing a catapult there to deliver materials to orbit. Any catapult (Lunar or other) will need a significant power supply for the motor. In order to not waste the rotation energy when stopping to load the next cargo, it makes sense to consider two catapults, and use one as a generator to supply power to the other. Motor-generator efficiencies can be above 90%, so most of the energy could be recycled.
Skyhook - The Moon's gravity well is equal to 287 km at 1.0 Earth gravity. Thus even for the lower strength basalt fibers, the gravity well is only 1.6 times the scale length of 178 km. For the higher strength carbon fibers with 361 km scale length the ratio is 0.80. Note that scale length is based on breaking strength, actual designs will use lower loads and have overhead above a bare cable. For Earth the theoretical gravity well to carbon fiber scale length ratio is 18, so it is much easier to build a Lunar Skyhook on a relative basis. Another way to say this is the Lunar orbit velocity of 1680 m/s is less than the 2400 m/s tip velocity assumed below for the Earth Skyhook. That provides all of the velocity between the Lunar surface and orbit, while the Earth orbit version only provides 1/3. Since escape velocity is 1.414 times circular orbit velocity, and a full Lunar Skyhook is capable of releasing cargo at 2.0 times orbital velocity, it can handle cargo well beyond escape velocity. By climbing to a chosen radius from the center, and timing when you let go, you can get a wide variety of orbits.
Besides using it as transport to and from the Moon, there is also an opportunity for work crews to use the Skyhook at 1.0 gravity as a rest location, because we don't know the long term effects of 0.16 gravity on the human body. A lunar surface alternative is to use centrifuges to get 1.0 gravity or whatever level is needed. A full Lunar Skyhook would have a radius of 283 km and a rotation period of 17 min 40 sec at 1.0 gravity. Since Lunar orbit period is 108 minutes or longer depending on altitude, the Skyhook will make 6 or more rotations per orbit. If the orbit is equatorial, that allows it to service multiple locations around the Lunar equator, and transport cargo between those points at orbital speed at no cost.
It is an open question if an equatorial orbit is best. A polar orbit would let the Skyhook reach any point on the Lunar surface, but generally only twice a month. The Moon rotates very slowly, so the benefit of the rotation towards the orbit velocity is only 4.6 m/s, 1% of the Earth's contribution. A polar orbit can be arranged as a Sun-synchronous orbit, where the orbit plane always is in sunlight, while an equatorial orbit is in shadow about 40% of the time. Thus the solar arrays that power the Skyhook are more effective in the polar orbit. You can have both Skyhooks in orbit around the Moon, as long as you arrange their orbits to never intersect, such as by using different altitudes. In that case you might want to make the g-forces at the tips higher so the radius is smaller, and move humans quickly up the cable to a more comfortable g-level.
In any Lunar Skyhook, a lander vehicle will need some propulsion because the Moon is not a perfect sphere. So the tips need to stay high enough to miss any high points of the terrain, and some maneuvering is needed for an accurate landing. If you have two Skyhooks at different altitudes, the vehicle will need more fuel to land and take off.
Asteroid Skyhooks 
The largest asteroid, 1 Ceres, is 487 km in radius at the equator, with a day length (rotation period) of 9.074 hours. Therefore the equator is moving at 94 m/s. Orbit velocity is estimated at 360 m/s. The exact number will be found when the Dawn spacecraft arrives at Ceres in 2015 (it is in orbit around the 2nd largest asteroid, Vesta, as of 2012). A Skyhook thus needs only the difference of 266 m/s in order to land and pick up cargo and then toss them at more than escape velocity. The radius in this case at 1 g works out to 7.25 km. This is small enough that it could be built near the Earth, and then transported whole to Ceres. Setting it up in orbit would allow mining of the largest asteroid with easy access. A synchronous space elevator would be longer and not provide a 1 gravity environment, but could be used to launch cargo into transfer orbits away from Ceres or capture incoming cargo.
For small asteroids, a Skyhook isn’t necessary for surface access. Even low efficiency chemical rockets do not use much fuel to land, and you can just mechanically throw stuff into orbit or escape.
Mars Skyhooks 
Pavonis Mons is one of the large mountains on Mars. Since it is located on the Equator, it is an ideal location for some kind of transport system. Candidates include a centrifuge launcher like with the Moon, or a linear accelerator. The higher mass of Mars makes it more difficult than for the Moon, but a ground-based transport system can still do most or all of the job of reaching orbit velocity. Similarly, Mars orbit velocity of 3.6 km/s is within reach of a Skyhook, and there are two convenient former asteroids (Phobos and Deimos) as a source of building materials. A Martian Skyhook would likely be placed in low Mars orbit, with the ability to transfer down to the surface and up to Phobos, Deimos, or escape orbits.
Split Systems 
A catapult can be used in combination with a Skyhook to enable higher velocity missions with lower total mass ratios. Bodies as small as the Moon do not require very large mass ratios o reach orbit, so doing a split system will not gain much at the cost of the extra complexity. Conversely the Earth has a fairly dense atmosphere, so a high velocity centrifuge would see a lot of drag unless placed on a very tall structure. The best location for a split system turns out to be Mars, particularly with its tall mountains that are in near vacuum at their peaks. By dividing the velocity between two systems, it becomes 1.8 km/s each, which can be reached with existing materials and conservative mass ratios.
Design Parameters 
The Earth orbit Skyhook does not have a fixed design as noted above, but rather grows over time. We also do not as yet know what an optimum size will be for given circumstances. A concrete example, however, lets us examine the feasibility and understand what is needed for the various parts. We will assume a tip velocity of 2400 m/s for this example, or roughly 1/3 of orbit velocity, and derive the other characteristics.
Design Inputs 
Tip Velocity = 2400 m/s
Tip Acceleration = 10 m/s^2 - Earth surface gravity is 9.80665 m/s^2. We use 10 for simplicity. That provides normal gravity for any humans on the Skyhook.
Derived Values 
Skyhook Radius = 576 km - This is found by solving the centrifugal acceleration formula ( a = v^2/r ) for the radius.
Rotation Period = 25 minutes - We know the circumference of a circle by 2 x pi x r, or 3619 km in this case. Dividing by tip velocity gives the time. For convenience to reach the Skyhook from a launch site, the numbers can be adjusted so the period is an even fraction of the orbit time, i.e. 100 minute orbit with 25 minute rotation time. That way the landing platform will be in the same place each time relative to the launch site.
Orbit Altitude = 750 km - If the tips of the Skyhook reach deeply into the Earth's atmosphere, that will cause drag and heating, and eventually cause the Skyhook to fall down. By placing the tips at least 175 km altitude, then the center must be that plus the radius high. The exact height will be a trade off between less drag, and ease to reach from the ground.
Orbit Velocity = 7474 m/s - Found from the formula below where G is the Gravitational constant, M is the mass of the Earth, and r is the orbit radius, which is the Earth's radius plus the orbit altitude:
Launch Vehicle Payload = 13% - A good chemical rocket would have an exhaust velocity of 4.5 km/s and empty weight of 10%. Without a Skyhook, the total velocity required is about 9 km/s, which results in a payload fraction of 3.5%. Subtracting the 2.4 km/s provided by the Skyhook results in a payload of 13%, or 3.7 times higher. The exact numbers will vary depending on the launch vehicle design, but that gives an idea of the payload improvement the Skyhook can provide, and thus part of the reason to build it. The greatest advantage of a Skyhook is not from the increased payload it provides, but using some of the increase to increase the fatigue life of the vehicle, which is highly non-linear - typically ten times higher for a 10% addition in structure. Airplanes and rockets cost about the same per kg to build. This is not surprising since both are built by aerospace companies out of the same materials. The vast difference in transport cost is due to airplanes flying about 20,000 times during their service life, and rockets usually only flying once. By taking some of the payload increase from a Skyhook and applying it to giving the launch vehicle a long operating life, the operating cost will be vastly reduced.
Payback Time = 1 to 76 launches (average of 43) - If we remove the last 100 m/s of tip velocity, the launch vehicle payload falls to 12.56% from 13.07%. So the incremental benefit of the last 100 m/s is 0.51% of the vehicle mass. Assume we use Torayca T1000G carbon fiber as our main cable material. It has a tensile strength of 6370 MPa and a density of 1.8 g/cc. We allow 40% overhead mass above the bare fiber for a finished cable system, and a 2.0 factor of safety. Thus the working strength is reduced to 2275 MPa, and at a tip acceleration of 10 m/s, the working length becomes 126.4 km.
The 2400 m/s Skyhook has a radius of 576 km, and the acceleration varies from 0 at the center to 10 m/s at the tip, so effective length is half, or 288 km. The cable mass ratio is then e ^ (288/126.4) = 9.762:1. Subtracting the payload mass gives a theoretical cable mass of 8.762. Since this calculation is only for one arm of the Skyhook, we double it to 17.524. The payload of the Skyhook is the launch vehicle payload + the empty vehicle structure without fuel (10% of launch weight), for a total of 23.07% of launch weight. The Skyhook cable mass is then 404.3% of the launch vehicle weight. Doing the same calculations for the 2300 m/s Skyhook, we have a radius of 529 km, effective length of 264.5 km, mass ratio of 8.106, and cable mass of 16.211 times the vehicle arrival mass of 22.56% = 365.7%. The incremental cable mass is then 404.3 - 365.7 = 38.6%. Since we gain 0.51% in payload from this incremental addition to the Skyhook, it pays for itself in payload mass in 76 launches if we use the launch vehicle to deliver the added cable. If we get the extra cable from another source, such as our hypervelocity gun, or from NEO carbon, the payback could be much faster.
Doing the same type of calculation over the whole Skyhook, we have a gain in payload from 3.5 to 13%, or 9.5% of liftoff mass. The Skyhook cable mass is 404.3% of liftoff mass, so the payback time in increased payload is 43 launches. The first 300 m/s increment of the Skyhook increases payload from 3.53% to 4.46%, a gain of 0.93% of liftoff mass. It would have a radius of 9 km, and a cable mass of 7.25% times arrival mass, or 0.98% of launch mass. Thus the payback time is 1 launch, and the cable fits in roughly 1/4 of a payload on it's delivery flight. So the first part of the Skyhook has an immediate payback and is very desirable. Note that a Skyhook massing less than the arrival mass would not have enough orbital energy to give the vehicle on arrival. It would need to be attached to a larger "ballast" mass such as an assembly platform, bulk mined materials, or collected space debris.
Mass payback is not the same as cost payback, but if we assume they are for now, at a relatively low rate of one launch/month, payback takes about 3.6 years, which is reasonable from an economic standpoint. Since the cable mass grows exponentially with tip velocity, the earlier parts will pay back faster, and growth beyond this point will take longer at a fixed launch rate. A real payback analysis will have to take into account real cost instead of mass ratios, and actual launch rates. If less expensive launch systems from Earth are developed, then the cost benefit of a Skyhook goes down, even if the payload mass increase stays the same. On the other hand, if traffic rates go up, do does the cost benefit. No matter what numbers are used, the exponential growth of the Skyhook mass with tip velocity will eventually limit it's size for economic reasons. That is when the incremental cost of making the Skyhook larger becomes more than the incremental payload increase is worth. We emphasize that limit can change over time, however, as new materials become available, different methods of cable delivery are used, and traffic rates change. The Skyhook/launch vehicle interaction is a good example of why combined systems have to be looked at in their entirety, and not as single technologies or methods.
Design Components 
The main structural component of the Skyhook will be the tension strands. In addition there will be secondary structure holding the strands in position, and for the landing platform, propulsion system, habitats, and other items attached to the structure. Cables are not a stable structure unless they are in tension, so for the central portion of the Skyhook before it is set rotating, we assume a core rigid truss structure. The initial total radius is 2500 m, counting core plus cables. This allows 1.0 gravity acceleration at the tips with a rotation period of 100 seconds. The latter number is chosen so that humans are not dis-oriented by rapid rotation. Initial habitats would be placed at the 2500 m radius. Strands are installed both lengthwise and in parallel to expand the Skyhook.
Damage Tolerant Design - Man-made satellites and orbital debris are the largest impact hazard for the structure. For now we assume no cleanup of Earth orbit, though that is desirable. Doing so would reduce the risks by about ten times. If we assume there is a 0.5 mm protective sheath around the strand core, then objects smaller than half that thickness will merely make a crater and not penetrate the core. If we assume a large Skyhook supports 1000 tons of payload at 1.0 gravity, with an 8:1 mass ratio, we have a total load of 10 MN. Given the 2275 MPa working strength, that requires a total area of 44 cm^2 of cable. Assume each strand is 2 cm in diameter. Then we need 14 load carrying strands, and some number of extra strands for damage tolerance, which we will use 7 for now, giving 21 total strands.
Any debris object larger than 1/3 the strand diameter will likely cause enough damage that it fails. Based on a 1995 orbital debris assessment  there were about 1 million objects that size and larger. That produces an impact flux of about 10 ^ -4 per square meter per year. If we want a 1% chance of strand failure per year, then we are allowed 100 square meters of area. Since the diameter is assumed to be 2 cm, then the allowed length is 5000 meters. We place cross-connecting rings at that interval to distribute load around a failed strand. At each ring, there are 7 points where 3 strands each fan out. Since by design two are required to handle the load, the failure of the third by debris impact does not cause any reduction in the Skyhook's total load carrying ability.
We do not want a single object to impact too many strands at one time. In the worst case, the largest object in orbit besides the Skyhook, currently the Space Station, should not hit more than our reserve of 7 extra strands. In reality, 99% of damaging debris is smaller than 30 cm, and the Space Station is under active control, so it should not ever hit the Skyhook, but we are looking at a worst case. Since the Station is about 120m wide, if the strands are arranged in a circle, and there are 21 total, a 120m object should not intersect more than 120 degrees of the perimeter. Doing a little geometry yields 120m = 50% of the radius, thus the diameter of the circle is about 480 meters, and the 7 attachment points are spaced about 200 meters apart. A truss spans between each attach point, making a 7-sided ring. If a given strand is damaged, then it is simply replaced by the same construction method the Skyhook was built in the first place. As long as strands can be replaced at least as fast as they are damaged, the Skyhook can be maintained indefinitely.
The above calculations are an example. For a real design, you would find the optimum strand diameter and count, rather than just assume 2 cm and 21. The real debris population is not all in orbits that could intersect the Skyhook. For example, the Space Station is about 400 km altitude, and could only intersect with the bottom 225 km of the Skyhook when it is in the vertical position. Some efforts may be made to clean up orbital debris. But even if not, we can make a reasonable design that can withstand worst case damage and reduce expected damage to a 1% per year maintenance job.
Landing Platform - This functions somewhat like the deck of an aircraft carrier, in that it is a mobile platform which vehicles land and take off from. We assumed above that the Skyhook structure supports 1000 tons of load. This includes everything besides the main structure, including the landing platform. Arriving vehicles would have a smaller mass that is added to the load temporarily. Unlike zero-g docking, which is done slowly, the landing platform is rotating at 1 gravity, so landings will be similar speed to landing on Earth. The size of the platform will be governed by the accuracy of the vehicle navigation. The design can be a horizontal platform, or something like a latching hook or arresting cable, such as used on aircraft carriers. In that case the Skyhook name becomes a literal description. An alternate method is have a vertical capture net. It's as wide and tall as needed to make a good target. The vehicle has redundant capture latches deployed ahead of it, and arrives slightly faster so it runs into the net, and snags multiple cables.
It is assumed the landing will be automated for uncrewed cargo delivery, with radar, lidar, and other aids to getting within the landing target area. The landing platform is made a multiple of the navigation accuracy in size to have a high probability of hitting the target. The best design is an open question, but since landing at 1 gravity has been solved multiple times on Earth, it should be solvable for this task.
After delivering it's payload, the vehicle just needs to let go or be pushed off the platform at the right time. At it's low point, the platform is sub-orbital, so the vehicle will automatically re-enter. The vehicle will be moving at 4,600 m/s relative to the Equator, compared to 7,400 m/s for a rocket without a Skyhook trying to re-enter. Therefore the vehicle has to dissipate 39% as much kinetic energy, which makes the heat shield design much easier.
Low Gravity Platforms - Low- or Zero-gravity is desirable for some tasks in Space. You can place platforms or pressurized habitats at chosen distances from the Skyhook center and get any value between 0 and 1 gravity. For true zero-g, you would need to de-spin the structure so it does not rotate along with the rest of the Skyhook. Likely this will be a structure extending along the axis of rotation like the axle of a wheel, where the Skyhook cables would be spokes of the wheel.
Electric Propulsion 
Due to fundamental conservation of energy, transporting more payloads up than down causes the Skyhook to lose orbital energy and eventually re-enter if not corrected. We use the highly efficient electric thrusters developed in an earlier step to maintain orbit. In effect, the electric thrusters substitute for the lower efficiency chemical rocket engines on the launch vehicle. Electric type engines have too low a thrust to reach orbit by themselves, but by attaching them to a Skyhook, we can add orbital energy gradually, and then give that to the payload in a short time. The Skyhook becomes a very efficient battery for storing orbital energy, about 25 times the energy/mass of Lithium batteries on Earth.
Power Rquirements - For each kg of payload we are placing in orbit, we are changing the velocity from 5074 to 7474 m/s. This requires adding 15 MJ of energy. Since there are 31.5 million seconds in a year, then that means for each kg/year of payload, we need 0.477 watts of delivered orbital energy. In Earth orbit we are not in sunlight 100% of the time, so solar panels would need to be larger to average this power level, and electric thrusters are not 100% efficient, either. Using reasonable numbers of 60% sunlight time and 65% thruster efficiency, we get our solar panels need a peak output of 1.22 Watts/kg/year. If we want to deliver 1000 tons/year, then the power supply needs to be 1.22 MW. Nuclear power is excluded from consideration for political and safety reasons so close to Earth. Existing solar cells, allowing for 100% overhead, have a mass of 1.68 kg/m^2, and an efficiency of 29.5% vs a Solar constant of 1360 W/m^2. Therefore they produce 238 W/kg of output and their mass is 5.1 tons.
Thruster Type - For low Earth orbit, there are three types available with near-term technology: Ion, Plasma, or Electrodynamic. Electrodynamic uses less "fuel", but is not as well developed. Ion is well developed, but does not scale up to the high power levels as well. We will assume Plasma thrusters, but development of Electrodynamic should be pursued, and all three types considered as candidates. Current plasma thrusters in development are designed for 200 kW continuous power, so 6 units plus some number of spares would be needed for the Skyhook design. The estimated mass of the thrusters is 3 tons.
Fuel Requirement - With 1000 tons/year of cargo delivery to which 2.4 km/s of velocity is added, we need to expel 48 tons/year of thruster propellant at 50 km/s to maintain the Skyhook orbit. This can either come along with the cargo, serving as 4.8% overhead, or if materials are being extracted from nearby asteroids can come from them. The latter is preferred since it's more efficent for the launch vehicle.
Design Issues 
Space Environment 
Plasma Environment - The ionosphere can cause charge build up.