All page references to Bradley Edwards' and Eric Westling's "The Space Elevator" below are to the 2002 Spaego paperback, ISBN 0-9726045-0-2 .

Space Elevator Platform

This raised platform has six advantages:

It has five drawbacks:


Reference Design

Assume the 30MYuri, 20T, 33% safety factor(!) design in Ben Shelef's "Reference Design".

20,000

kg

Tether rider weight

9.613

m/s2

gravity at 50km

192,000

N

Tether rider load at 50km

300,000

N

Tether load above rider at 50km

108,000

N

Tether load below rider at 50km

1300

kg/m3

Density of tether at 50km

39

GPa

Design strength of tether

52

GPa

Yield strength of tether

910

GPa

Young's modulus

7.69

mm2

Tether cross section at 50km

0.01

kg/m

Tether mass per meter at 50km

2,200

km

Maximum space debris height

100

m

Lateral movement for debris

910

s

Radar horizon visibility time

500

s

Movement time


Problem's with Ben Shelef's "A Solar Based Space Elevator Architecture"

Do not assume Ben's power systems - the supposed "PV panel" he shows is an ESA-DLR solar sail, with a reflective coating on kapton to thrust with light pressure, NOT to convert light to electricity, or to conduct electricity to a load. Ben also assumes some kind of electrostatic ring structure to ride on a curled tether - but an object inside a closed ring is not centered, the fields don't work that way (think of a particle inside a faraday cage).


Optimum Altitude

Most dimensions scale with altitude. Energy per meter also scales with altitude, thus velocities scale with the square root of energy and altitude. So the total system energy goes up with the square of the altitude. Parsimony suggests the lowest practical altitude consistent with the goals.

Space Debris Increases With Altitude

One non-obvious risk is collisions with space debris. Assume a future time when we are putting enormous amounts of material into orbit. With prudent debris capture and recycling, only a small fraction of material escapes to become debris. Some of those escaped objects will be in orbits that intersect the earth; only one pass through platform altitude. That debris is not a problem. But some debris objects will be in slowly decaying orbits, providing repeated opportunities for collisions.

Even low debris escape rates become significant in a "megatons per day" future. We can expect to get better and better at cleaning up after ourselves in orbit, but we will never be perfect. That changes expectations.

Because reducing perigee velocity lowers apogee, we can expect that debris objects in elliptical orbits will tend to circularize and then spiral into the denser atmosphere. The drag increases inversely exponentially with altitude, so at lower altitudes the vertical decay rate will be exponentially higher. Thus, within a few hundred kilometers of the surface, we can expect the density of decaying objects (objects per volume, or objects per vertical meter) to increase exponentially with altitude, because the lower objects are decaying downwards faster. This debris will tend to impact structures from the side. Debris with a high vertical velocity component will not be around for long.

The actual rate that any particular item decays is dependent on mass and density and shape, so the scaling factor for the "collision flux rate" will depend on the nature of the debris stream. Dense objects will decay more slowly than hollow objects, and pose a higher risk per kilogram. Very large and dense objects would hypothetically pose the most risk, but they are also easiest to track and avoid (or capture in orbit and remove from the debris stream). Very small objects can be stopped by "meteor bumpers" at the side, turned into a spray of harmless fragments that cannot penetrate the tube/track. But regardless of size and scaling factor, the sum of all the individual objects will still follow the density altitude exponential.

A table of density-altitude, micrograms per cubic meter:

Altitude

Density

Normalized

km

\mu g/m^3

Collision Flux

0

1225000

1.0

5

736000

1.7

10

414000

3.0

15

195000

6.3

20

88900

13.8

30

18400

66.5

40

4000

307

50

1030

1190

60

310

3960

80

18.5

66400

100

0.56

2190000

120

0.022

55000000

140

3.8ng/m3

3.2e8

160

1.23ng/m3

9.9e8

180

0.52ng/m3

2.4e9

200

0.25ng/m3

4.8e9

Numbers from "U.S. Standard Atmosphere, 1976"

A few years into working on launch loop, I dropped the altitude of the launch loop from 120km to 80km - almost 3 orders of magnitude diminution of collision flux. While a higher altitude launch saves some drag, that is "only" energy. At 120km, the extra cost of materials for longer stabilization cables and inclines, and the higher collision failure rate, is much more expensive.

Recent work drops the launch track still further at the west end, to 50km at the west station. See LowerWestIncline.

Is a Tall Dynamic Structure Platform more Collision Resistant?

Assume there is some way to build a very tall dynamic structure, hundreds of kilometers high (I don't know how to do this without tons of unobtainium). The structure will need stabilization cables. The cross section of track and cables is a lot higher than the space elevator, so there will be more collision risk. It is possible that the main track can be equipped with "meteorite bumpers" which can deflect small stuff, and the structure itself can support radars to see the big stuff. Rather than just passively getting out of the way of the big stuff, it may be possible to "drop water vapor balloons" in the paths of derelict objects in some manner that will slow them down without fragmenting them. In the short term, and for objects that must remain in orbit, the platform must move, and components like cables must move independently, or there will be too much cross section to move at one time.


Climber Lasers

Using a 20 metric ton climber, and 200 km/hr (56m/s) climb rates, the initial climbing power is 2e4 * 9.8 * 56 watts = 11 MW mechanical. Assume 22% of the laser output is captured and converted into mechanical energy (which is incredibly optimistic, but that is the "community norm"), then the laser beam output is 50MW for each climber, 150 MW total for 3 climbers spaced 2.5 days apart. Jordin suggested 2 to 3 MW of waste heat per MW output. Lets optimistically assume 2MW - it makes the calculation easy. So we need 450MW of power, and 300MW of cooling.


I too calculated the initial power as 11 MW. I don't know where Brad Edwards gets 2.4 MW; he doesn't say. I thought he might assume a lower speed to begin with, but the times don't add up. 200 km/hour is already very slow to travel 36,000 km.

KHL: A question for Peter, I suppose. If we assume all the laser power is captured to about 1 radii out, then the speed will slowly increase as gravity is reduced. But with 2.4MW mechanical power and 20 tons to lift, the speed does not increase to 200kph ( 55.55(5) m/s ) until 3 days 2 hours and 13374km out. At 200kph, it is another 6 days to GEO. The assumption is that the laser beam will diverge beyond the photovoltaic panel at higher distances, with the diminishing collected power matching the diminishing thrust requirements. Fair enough.

Elapsed time 9.1 days, not 7.5, at 2.4MW and 200kph max. The elapsed time drops to 7.5 days to GEO if we up the max speed to 240kph and the power from 2.4MW to 2.9MW, which is much less than the 11 MW you and I computed. The Shinkansen goes at 300 kph, and the TGV has reached 575 kph in tests. Against air drag, but not straight up, and not for 30,000 km! Again, I suppose we should ask Peter.

More Climber, laser, etc.

There are some big issues with heat and efficiency in Edwards' "The Space Elevator". On page 52, he assumes 10% of the incident energy will be converted to heat. This is WILDLY too optimistic; it is even unrealistic to assume 90% efficiency from photovoltaic panel output to mechanical climber energy. The main problem is the photovoltaics.

Edwards refers to the 1992 D'Amato paper on page 63. In that paper, a GaAs photovoltaic cell reached 59% in the lab with an optical input power density of 54W/cm2. The actual experiment combined the 1.2W optical outputs of five 826nm AlGaAs laser diodes, shined them through an aperture at a 4mm diameter AlGaAs/GaAs test cell, to produce a nonuniform, "pie shaped" 1.7W spot with 54W/cm2 power density. The D'Amato paper is a little vague about how the power or spot size or power density was actually measured - the power density and power numbers imply a circular disk 2mm diameter with uniform illumination. However, without independent measurement (say with a bolometer) the power actually hitting the cell could be a larger fraction of the supposed 6W input. If the laser diodes were running at higher power than expected (they might be rated for 1.2W end-of-life, which means they produce significantly more optical power at beginning of life) then the actual power reaching the PV cell may be underestimated, and the efficiency of the cell overestimated. Just because something is in a letter to an IEEE Journal does not mean it is accurate.

The window material is reported as Al_{.23}GaAs_{.77}, which does not make chemical sense. AlGaAs is actually a blend of Al and Ga with As, and Al_{.23}Ga_{.77}As is probably the proper stoichiometric (?) mixture the paper refers to. That has a bandgap of 1.59V (estimated), while the GaAs has a bandgap of 1.42V.

The efficiency numbers are higher than reported elsewhere. The power delivered was 0.98W at 1.02V at the maximum 0.82 "fill factor" (defined as the largest area box that can be drawn within the I/V curve, divided by the product of the open circuit voltage (1.20V) and the open circuit current (not given, but implied from 0.59A/W) . With an 826nm source, the incoming photons have energies of 1.50 eV ( 1240 eV-nm / 826nm ), so an open circuit voltage of 1.20V seems high but not wildly out of line for these high illumination levels - hundreds of suns.

This was a tiny cell mounted on a thick molybdenum block heat sink. The temperature was not given, but we can assume it was probably around 300K . The difference between the bandgap of the photocell (est 1.42V, like GaAs) and the max-fill diode voltage (1.42 - 1.02 = 0.40V ) will be approximately proportional to the absolute temperature, so if the cell was heated to 400K the cell might have an output voltage of (1.42-0.53 = 0.89V), and the efficiency would drop by 0.89/1.02 or 87% of the room temperature case. If the paper is accurate about the efficiency (doubtful), it would drop would be from 59% to 51%. There will also be increased resistive losses in junctions and wiring, and increased leakage losses. The loss to mechanical power ratio will be more like 140% rather than 11%. Since climber power is limited by the black body heat dissipation of the 12m2 solar cell, that is really bad news for the maximum climber power.

Can cell efficiency get a lot higher? That requires a very high bandgap material and high energy ultraviolet photons. Such a material must be highly doped to be conductive. Perhaps the highest that could be expected would be some variant of Aluminum Nitride, with a laser wavelength of 262nm and photon energies of 4.73 eV. If that drove a PV material with a bandgap of 4.6eV and a diode voltage of 3.9V, we might expect more like 82% efficiency out of the PV, and 75% system efficiency. That says a ratio of 33% power loss to mechanical climber power - not as good as Edwards' outlandish 10%, but a lot better than 150%! However, the stratosphere is opaque to UV, so the UV lasers would need to be located above the stratopause, at 50km altitude, and cooling them would be very difficult.

What can 12m2 dissipate? Assuming top layer infrared emission, with an emissivity of 0.8, with a zero K environment, we get the following table:

lift power

lift speed at 0km

radiated

11%

33%

140%

11%

33%

140%

Temp K

watts

watts

watts

watts

m/s

m/s

m/s

300

4409

40082

13361

3149

4.54

1.51

0.36

350

8168

74256

24752

5834

8.42

2.81

0.66

400

13935

126678

42226

9953

14.4

4.79

1.13

450

22321

202914

67638

15943

23.0

7.67

1.81

500

34020

309273

103091

24300

35.1

11.7

2.76

550

49809

600

70544

74160

8.41

My brisk walking speed is about 5km/hr, or about 1.4 m/s ... though not straight up.

Note that these numbers are for an absolute zero environment. A climber with the earth underneath will get about 620 W/m2 of Earth heat and reflected sunlight on the underside, and 1360 W/m2 of sunlight on the top side. That will result in black body heating of 364K, so at noon the PV cells will be quite hot, even with zero laser power.

400K is the hottest most military grade electronics are allowed to get. Failure rates in electronics tend to double with every 10K increase in temperature (Arrhenius curve for 1eV activation energy).

If some of the electrical power is driving a heat pump, which pushes heat "thermodynamically uphill" to a radiator, then perhaps the cells can stay cooler. But even a perfect heat pump is limited by Carnot efficiency. More likely, at 80% of perfect Carnot efficiency, we probably can't go much above 600K radiant temperature for a 400K photovoltaic temperature. That would require 1kW for every 2kW pumped; if the efficiency is 70% for UV photocells at 400K, that results in 74160 climbing watts.

MORE LATER

Climber motor

BTW, Edwards worries that the motors will overheat in the beginning, starting from stall. Not true. Photovoltaic cells look like current sources until the forward voltage reduces current, so a stalled motor (with no backEMF) will not draw a lot more current than a turning one, and will absorb relatively little extra power. Motor 101 - the torque is proportional to current, the voltage (backEMF) proportional to speed, the mechanical power proportional to speed times torque or voltage times current. The losses are mostly proportional to current squared - resistive heating and magnetizing hysteresis. That is why a very fast motor can be a lot more efficient than a slow one. This was one of the exciting early discoveries about the launch loop. In any case, stall is no problem - if the current (and the torque) is slightly higher because the photocells are near zero voltage, the laser power can be reduced a few percent to match. What makes most motors overheat in stall is that they are voltage-fed (very high current!) and their cooling fans are stalled also.

The motor will still need electronic commutation, because it will be operating at higher torque and lower speed until it reaches the 3 day point. This can also control the current and power levels in the motor.

Cooling the Lasers

Water Cooling: Assume we will cool by exhausting low pressure steam on the downwind side of the platform. We start with 20C liquid water, and output steam at 50C and 100 Torr ( the atmospheric pressure at 20km is 40 torr ). That is 30 kcal/kg (120kJ/kg) for heating the water, and about 2380kJ/kg for latent heat of vaporization at 50C, for about 2500 kJ/kg cooling . If we need to make 350MW of vapor, that is 140 kg of water per second. We are lifting that 20 km, or about 200kJ/kg, or about 30MW of of pump power at 100% efficiency. Rather than push that water up a pipe with pumps, it would probably be easier to move it up in a conveyor belt of pressurized tanks. It would be cheaper if we could use seawater for the cooling, though I suspect corrosion would be an issue. There may be a "two loop" cooling system, with pure water and heat pipes cooling the lasers, feeding a heat exchanger to vaporize sea water. Lifting water to a higher altitude would involve a correspondingly higher cooling power.

Air cooling, 15km: Let's assume 15km altitude instead of 20km, denser air and high wind speed. The wind speeds are perhaps 180 kph ( 50 m/s ) and density is around 0.2 kg/m3 and the heat capacity is around 1kj/kg-C . If we efficiently mix heat into the air, through a radiator 5 meters tall and 200 meters wide, that radiator will "see" 10000 kg of wind driven air per second. 350MW of heat (including coolant circulation friction ) will heat that air by about 35C, from -56C to -21C. Nice! That is why I like a lower altitude, with a fast steady wind. The drag on the radiator will be significant, though, and wind turbulence would vibrate the platform.

Air cooling, 20km: The wind speeds are perhaps 20 m/s and density is around 0.09 kg/m3. Assume a radiator 5 meters tall and 1200 meters wide, for the same 10000 kg of wind driven air per second, resulting in cooling similar to the 15km case.

Rotor cooling: If the heat was transported by the rotor to a sea level cooling system, that would not require a radiator or lifting a lot of water. On the other hand, pumping heat into and out of rotors by black-body radiation would be very difficult unless the temperature of the rotor was quite high. Assume that the rotor is carrying away 300MW, that it has a surface area of 0.2m2 per meter, and that about 2km of rotor is exposed to heat pump surfaces. That is 400m2 of total rotor surface. Assuming a 100 percent emissivity (ha!) and no back-radiation (ha!!), then the power per square meter is 750kW/m2. Divide that by the Stefan-Boltzmann constant (5.67E-8) and take the fourth root to get the emission temperature of the heat cavity - 1900K! Getting the heat from the laser to that temperature pumps it uphill, working against Carnot efficiency, which means much more heat output, which means higher temperature ... Obviously, rotor cooling will not work unless we figure out a way to heat a MUCH larger rotor surface. So radiator or steam cooling at altitude is probably our best bet.


Air cooling at 20 km involves a radiator of about 300 cubic meters, presumably weighing about 1000 tons. I see the attraction of the lower altitude. The air at 15 km should still be dry, so laser dispersion should be quite small. I think it's important to point out that there is a tradeoff with altitude, and the Lofstrom Loop technology is capable of 120 km or more.

At higher altitude, a radiator of 6600 square metres, emissivity 0.8 and temperature 1000 deg K can lose 300 MW. The option of higher altitude is important. Correct me if I'm wrong, but at 20 km we're still subject to electric storms at the platform, and one of our aims is to lift the SE ribbon above that zone. Peter Swan originally suggested 100 km as a round-number altitude, although I'm sure we can argue the case for other heights. (50 km is my favorite, but that's not important.)

KHL: There are two tasks - the space elevator, and the lasers. All the lasers need is some elevation and transparency, so 15km might work for them, while a much higher loop might connect to the space elevator. The problem with "higher" is that costs go up supralinearly with altitude - not only a taller and bigger footprint, but more rotor speed. I am curious about lightning at 20km+ altitudes. Cumulonimbus clouds stop at the tropopause, and those are typically the drivers of the atmospheric "van de Graff" generator. I would expect little charging above that, and very low horizontal conductivity, so there would be nothing to feed current to an arc. Perhaps the worry is high voltage and arcing in the climber.

JK: There is a wiki at http://en.wikipedia.org/wiki/Upper-atmospheric_lightning It seems they occur at many altitudes, even above 100 km in some cases, and are often associated with thunderstorms lower down. Apparently, they can also be triggered by powerful lasers. I think there is a dearth of hard facts because of the difficulty of observing them, something that the Lofstrom Loop may be able to help with.

Using Launch Loop to Build the Space Elevator

KHL: I'm uncomfortable with promoting launch loop for building a space elevator. Too much of the camel in the tent, too soon. Anyone with two neurons to rub together will realize that a system capable of a thousand vehicle launches a day will probably not be used to build a less efficient system capable of 3 vehicles a week. A 20km platform will be a lot cheaper, lower energy, and lower speed than a full launcher capable of orbital speeds.

KHL: The launch loop I envisioned launches 5 ton vehicles, up to 80 per hour. Perhaps it could be tuned for one or two 10 ton vehicles per hour. Launching a vehicle puts a large perturbation on all the spacing control systems along the track, and the power those controllers must handle is proportional to the perturbation. A launch also puts a tangential velocity perturbation on the rotor, with a maximum near release velocity. This means that the involved regions of rotor are "moving backwards" in relation to the uninvolved regions. Unless we have enough expensive power plant to put half of the velocity in in advance of acceleration, and half afterwards, then the rotor will get thin in front of the involved area, and denser in back of the involved area. Putting 5 tons at 3 gees into a 10km/s transfer orbit from a 14m/s rotor means feeding something like 3.7 GW into the input rotor and 3.7 GW into the output rotor all at once, or more likely moving a large fraction of that from one part of the rotor to another to manage the stretch and compression. Assuming power plant costs of \$5/watt (to input the power) and electronics costs of \$0.10/watt (to convert between electrical and mechanical power) then moving 3 GW from one part of the rotor to another might cost \$600M (two conversions), while inputting it directly might cost \$15B . Moving to 10 tons would double those already high costs - that is a lot of sofa cushions to look under for lost coins!


I see the problem. My thought was to propose more than one rotor, essentially adding to the launch mass linearly, just as I have always proposed several pairs of tubes in the space cable. This has the advantage of redundancy, whether or not we talk about using the loop to start the space elevator. Of course, it's more expensive.

One of the active areas of discussion in the SE community is to use weaker material by having a greater taper ratio, e.g., 10 times as thick in the middle as at the end rather than twice as thick. This requires a greater initial launch mass. Maybe it's sufficient to point out that the loop was originally envisaged for launching and we are now talking about its alternative use as a support platform.

Using Rockets, and Lasers, but with a Raised Platform

The space elevator is assumed to be built as thin as possible, then used to lift more cable and exponentially increase in thickness and capacity. A space elevator cable terminating at an elevated platform can be thinner than one terminating on the surface, so the initial mass in orbit can be smaller. Assume something like a Russian Proton booster launches one 10 ton cable payload with a 10 ton laser-powered-hydrogen upper stage (A lot cheaper to build than a launch loop!). When the vehicle passes overhead, we boost it with the 150MW laser, driving the hydrogen to perhaps 10km/sec . Assuming 67% conversion efficiency to thrust (rockets are easier than solar cells and motors), we can heat 2kg per second of hydrogen, and provide 20kN of thrust, accelerating our upper stage at 1m/s . If the stage remains in view for 1000 seconds, we can add 1000m/s of velocity at every orbit, with 3000 m/s for GTO perigee insertion from LEO, and 3000 m/s for GTO to GEO apogee insertion. It would be good to check those numbers with Jordin. Note that a Proton launch out of Baikonur Cosmodrome, at 45 degrees north, will not pass visibly over an equatorial or 8S platform very often. The LEO to GTO transfer may take many orbits. The overhead time will be better from KSC ( 28N ) and CSG Kourou ( 5N ), but times won't be really good anywhere until the vehicle is above the horizon more often. (The horizon is a lot farther away at 20km altitude, though!). The long wait times suggest the use of a lower-ISP but more storable laser propellant.


This is interesting stuff. How confident are you about introducing laser propulsion? That's a whole new technology that is little tried, though very promising, I agree. I thought the use of ion propulsion from orbit was a reasonably safe bet, as it's already proven.

KHL: You are probably right, we should assume magnetoplasmadynamic thrusters. We want to stay within the mythos of the space elevator, though there are assumptions about power conversion and cooling that are not sound (if solar cells and power electronics were really that easy, there are fortunes to be made!). I did send an email to Jordin about direct laser launch, mostly about the ISP issue. However, the same visibility issues apply for the MPD thrusters.

Support structure

Assume the lasers and associated hardware weigh 3.3kg/kW, 500 tons, adding into the mass of the whole platform plus stabilization cables of 2000 tons (WAG). Assume a separated bolt rotor, with a velocity change of about 20% rising 50km. The rotor speed will be around 1220m/s on the ground, 1000m/s at altitude. Assume a deflection radius of 100 meters, with four up and four down ramps at 45 degrees. The vertical deflection is 8 \rho v^2 \sin( 45 ) ~ = 9.61*2E6 N , resulting in a rotor mass per length \rho of about 3.4 kg/m. The total length of rotor is about 250 kilometers including ambits and motors. The rotor mass is 850 tons, and the total kinetic energy is 5e11 Joules, or 140 MWh. Rotor power (from four sides) is 4 \times 0.5 \rho v^3 ~ or 1.7GW; tapping off 500MW for laser power slows the rotor at the top by 100 m/s ( say 1050 m/s coming up, 950 m/s going down ).

Ambit Turnaround Forces: The ambits push outwards with a force of 2 \rho v^2 = 8E6 N . Assume they are 25 km from the center. Steel has a tensile strength of 2000 MPa - with a safety factor of 4, the steel cables supporting this force would have a cross section of 160cm2 and a weight of 126 tons per kilometer. Anchored to a 4km deep sea bottom with 45 degree slope ( \sqrt{2} length, \sqrt{2} force ), cables and floats might weigh 8000 tons. At \$2000/ton formed (WAG), that is \$16M for the steel.

However, if deployment involves pulling the ends inwards by 8km to maintain constant density, we might need double this amount of steel for the cable. We will also need a fairly hefty amount of steel to build the incline ramps and the floats, and to build the mast that holds the platform during rotor startup.

Design parameters

Design Tensile Strength

1.04

MegaYuri

Platform Height

50.00

km

Surface angle

38.00

degrees

Track to Rotor ratio

2.000

Ground

All rotors, lineal dens.

3.48

kg/m

Rotor ground speed

3.500

km/s

Gnd horizontal velocity

2.76

km/s

Gnd vertical velocity

2.15

km/s

ground ramp depth

0.131

km

grav. and centrif. accel

9.764

m/s2

Incline

Incline horiz. distance

81.69

km

Incline vert. distance

50.00

km

Incline length

96.03

km

Incline time

28.05

sec

Incline total track mass

1731.16

tonnes

Incline total rotor mass

683.19

tonnes

Platform interface

track tension

4.29

MegaNewton

track lineal dens.

11.086

kg/m

rotor lineal dens.

3.627

kg/m

angle

23.40

degrees

rotor speed

3.358

km/s

rotor compression

1.0430

horizontal velocity

3.082

km/s

vertical velocity

1.334

km/s

grav. and centrif. accel

9.612

m/s2

Platform

platform length

6.752

km

platform center height

50.699

km

platform mass

3000

tonnes

Estimated costs, arbitrary units

cost of ground magnets

819

cost unit

cost of inclines

668

cost unit

cost of platform magnets

504

cost unit

cost total

1991

cost unit

computed with plat03.c

50.00 km platform edge height

1.040 MYuri design load, 2.000 Track/Rotor mass ratio

ramp

depth

Vrotor

Mrotor

platform

platform

cost

degrees

km

m/s

kg/m

leng.km

tonnes

relative

38.00

0.131

3.500

3.480

6.75

3000

1991

38.00

0.268

5.000

1.173

18.47

3000

1232

38.00

0.386

6.000

0.755

28.04

3000

1106

50.00

0.221

3.500

2.086

11.27

3000

1393

50.00

0.452

5.000

0.864

25.10

3000

1090

50.00

0.651

6.000

0.575

36.85

3000

1019

Rotor Startup: Assume motors on 4 legs of the support structure. If the segmented rotor has a maximum tensile support length of 500m during deployment, and each motor pulls (and pushes) 63 km of rotor, the acceleration rate is 9.8*0.5/63 m/s2, or 280 m/s-hr . Assuming a 500MW power plant, the system will ramp up to 1000m/s in perhaps 4 hours, and to 2000m/s in about 30 more minutes, assuming the motors add 30 meters per second every time 63 kilometers of rotor passes by. Initial startup will be faster if the track magnets are operated in "AC repulsive" mode, with a traveling AC wave providing lift until the rotor is moving fast enough.

Incline Curvature:

MORE LATER

Rotor Design: Since the rotor is not directly accellerating payloads to speed, it does not heat nearly as much as the rotor of a launch loop. That means it can be dense, close to solid, and the cross section ...

Note, if the cross section is too small, the ambit magnets will not be able to exert sufficient pressure ...

MORE LATER

SpaceElevatorPlatform (last edited 2011-07-19 23:07:10 by KeithLofstrom)