Tuesday, February 22, 2011

A Project Plan for Space Based Solar Power


In my post, Space Based Solar Power, I present an overview of SBSP and its value to humanity. I also propose the use of asteroids for raw materials to reduce the cost to profitable levels. Still, the construction of the first Solar Power Satellite (SPS) is special, as we must incur all of the up-front (bootstrap) costs in addition to building the SPS itself.  I will demonstrate that it still can be a profitable venture. However, this project plan targets the initial construction of a dozen SPS’s, because the slag by-product of that much steel is the correct mass of shielding needed for a Kalpana-One style habitat for the workers (see my post, designing a space habitat).

The approach I recommend is to capture an asteroid (see  Asteroid Capture into Earth Orbit) and use its resources to produce the steel and other materials needed to construct and operate a constellation of SPS’s.

This approach has a relatively high up-front investment and results in a relatively low per-SPS cost. The project has many phases:

  1. Research & Development
  2. Capture a suitable asteroid into Earth orbit
  3. Launch Mining & Manufacturing Tools
  4. Launch Construction Shacks & Workers
  5. Construct & Deploy the SPS(s)
  6. Construct the Kalpana-One style Habitat(s)
  7. Repeat 5-6 as long as asteroid resources remain

For the purposes of this post, I make several assumptions:

  • we have chosen the asteroid 99942 Apophis, which also locks us into a 2029 timeline (project start in 2027, launch capture mission in early 2028).
  • use of SpaceX’s Falcon 9 Heavy launch vehicle (announced with pricing, but not yet built, yet alone flown).
  • pricing in 2011 dollars which must be adjusted for inflation.
  • Worker launch costs initially average $24M/person, but in a 2031 timeframe human-rated transports will be built for the Falcon 9 Heavy (or eventually Falcon X Heavy) that can transport 30 to 100 people per launch into LEO at a price of $1,000,000 per ton (also per person).
  • I ignore in-space worker salaries (small compared to launch costs)
  • For income, I only consider Solar Power Satellites, ignoring tourism, the sales/rentals of housing and retail space, sales of propellants to outside parties, product placement, tie-ins, naming rights, or any other potential revenue sources.


It may be an unreal oversimplification, but I will ignore the R&D costs for this project although they are large and unknown. I believe that NASA, the various governments, and the network of universities should bear this cost as part of their normal operations (supported by tuitions, taxes, and donations), although the more directed research will be borne by the corporations that will profit from the sales of related equipment and supplies. Known areas of research & development include:

  • Heavy lift launch vehicles
  • Space taxies / trucks (using ion thrusters to cheaply change orbits and to capture the asteroid)
  • Long-term life support (recycling & farming in space)
  • Zero Gravity Mining, Smelting, Refining, and Manufacturing (note need to recycle reagents and to capture and utilize bi-products such as CO2).
  • Manufacturing Solar Photovoltaic Panels from asteroid materials
  • High-efficiency Muilti-Gigawatt Microwave Transmitters / Receivers
  • Large-scale in-space construction techniques (of SPS and habitats) with limited resources


How to capture an asteroid is largely covered by my previous post, Asteroid Capture into Earth Orbit, and a future post will detail the components of the capture project, including a proposed project breakdown of the needed components, launches, delta-v budgets, and a project timeline.

The Asteroid Capture Project begins in 2027 with construction (in orbit) of a space tug. The tugship launch from LEO to intercept Apophis occurs in April of 2028, and the final Lunar slingshot of Apophis effecting capture into Earth orbit occurs 18 months later, in October of 2029. These dates are driven by orbital dynamics.

A simplified overview of the Apophis Capture Project (using current or near-term launch vehicles) is:

  1. Launch tugship, fuel, solar panels to LEO: total 150 tons, needs 5 Falcon 9 Heavy launches ($500M); cost of tugship, etc. another $1.5B. This cost is lower than many space-based projects because of the extensive use of modular components assembled by humans in space(reducing the need for first-time perfection).
  2. Launch in-orbit assembly crew: 1 @ $100M, probably another $1B for training, tools, supplies, support staff.
  3. Launch mission crew & supplies: 1 @ $100M
  4. Intercept mission uses ion thrusters and a lunar slingshot to enter Apophis intercept trajectory in April 2028.
  5. 2 months to intercept; 3 months of thrust to change Apophis orbit; 1 month of contingency; 6 months of coasting. After the slingshot around the Earth in April of 2029 there is another 6 months of coasting (with likely trajectory tweaking) before the final drop into Earth orbit via a Lunar slingshot in October, 2029. Capture Mission duration is a little over 18 months; with recycling of water & oxygen, 600 days of food & water & hydrogen (to recycle CO2) for a crew of 4 masses four tons; a similar mass is needed for the associated equipment (triply redundant).

Total capture cost: $3.2B (ignoring R&D and ground support costs).


This is a big unknown, in that we don’t yet know how to best do mining, smelting, refining, and manufacturing in space (microgravity, necessity to recycle & save everything) from asteroid resources. We can make good guesses, however. Also, the facilities should grow over time, but the initial launches should suffice for starting manufacturing, especially since we are largely launching tools to make tools. After the initial launches, most of the needed additional equipment will be manufactured in space. These launches would begin in 2029; I expect the tugship will be used to transport these components from LEO to the asteroid now in Highly Eccentric Earth Orbit.

  • Solar Power generator: 1 launch, $200M
  • Mining facility (ore movers, grinders, separators): 2 launches, $400M
  • Solar smelter, gas collection & refining, metal purification: 5 launches, $1B
  • Steel production: 5 launches, $1B
  • Rolling mill (girders, rods, sheet metal): 3 launches, $600M
  • Finished metal product plant (nuts, bolts, rivets, connectors, pipes, tanks, etc.): 2 launches, $400M
  • Silicon refinery, solar panel manufacturing (3 launches, $600M)
  • Slag processing, shaping, rock wool production, & slag handling (1 launch, $200M)

Total estimate is $4.4B for equipment and 22 launches, and this phase begins in 2029, ends in early 2030.


The initial workforce for all of the mining & manufacturing  equipment, even assuming significant automation, is considerable. I estimate 36 workers, needing 6 launches using Falcon 9’s and Dragon X capsules (total $600M) just to get started. Each crew member will likely require a ton or more of life support equipment and supplies.

The Construction Shacks are the living quarters, recycling equipment, supplies, and solar panels for the initial mining and manufacturing workforce, intended to be buried beneath asteroid regolith for radiation and meteorite shielding (alternately regolith may be bagged and packed around the shacks). This will need to be repeated as the workforce grows (until a permanent habitat is at least partially completed).

The Falcon 9 payload faring may be equipped to house up to 12, with room for supplies and equipment. Three of these would provide adequate housing and first-year supplies for the initial 36 workers, at a cost of $600M bringing the total cost to $1.2B ($33M/person).

Subsequent workers initially have similar high launch and support costs, ramping down to $15M per-person as capacity grows to produce living quarters, water, and oxygen in space. When the interim habitat is ready and larger capacity people launchers (32 workers/launch) become available, per-worker launch costs should plummet to roughly $3M per person (in 2031), and as low as $1M per worker in subsequent years (with use of dedicated people-carrying spacecraft ferrying 100 workers into space per launch). Note that a 1 ton per worker mass allowance includes some essentials that cannot be easily fabricated in space, such as high-efficiency LEDs for the farms,  some fertilizers, and medical supplies.

The workforce & construction shack launches could begin in late 2029.

This initial 36 worker contingent will be sufficient to prove the feasibility of mining, refining, and manufacturing in space, but will be inadequate to achieve sufficient volumes of production for a timely project completion.

A ramp up to 500 workers during 2030 should result in a steel production rate of roughly 100 tons per day, which in turn allows construction of the habitat pressure shell plus containers for the volatiles (oxygen, CO2, and H2O) in one year. At this point (an interim habitat consisting only of a spinning pressurized shell with partial radiation shielding), many of the workers can work without pressure suits, gravity prevents long-term health problems of zero-G living, and in-space production of food can begin. 

Doubling the population and production annually allows the first SPS to be completed by the end of 2031 (when steel production is 200 tons per day, comparable to a small mill on Earth). 2032 should see 2 more SPS’s completed, and 2033 results in 4 more and completion of the habitat structure (with room for 8,000). By the end of 2033 the habitat capacity is reached, and the associated manufacturing facilities should be capable of building 8 SPS’s per year. Remember, it takes all of the slag associated with building the first 12 SPS’s to fully shield the habitat, which occurs during 2034. By the end of 2034, the first permanent habitat (8,000 person capacity) is complete, along with 15 Solar Power Satellites.

The total worker cost through completion of the first SPS and the habitat shell would be $15B (by the end of 2031, a total of 1,500 workers). Adding 2,000 more workers in 2032 and 4,000 more workers in 2033 would cost only another $6B.

This analysis has ignored the cost of salaries for space workers (and the cost of ground support). One reason to ignore the salaries is that they are insignificant compared to launch and equipment cost in the early years.

Another reason to ignore them is that portions of each “salary” may be offset: workers may be charged for rent (or purchase of a condo), for food, water, recycling, even oxygen. The cost of living is very high in space, and salaries should reflect that, but the net cost of a worker is difficult to estimate. Many of the readers of this blog would likely volunteer to work for little more than the right to permanently live in space and be a part of humanity’s spacefaring future – I know I would. Very few of the workers would view these orbiting factories and habitats as temporary jobs and expect to return to Earth in a few years, especially if they were afforded the opportunity to buy a home and raise a family in the habitat.


The construction of a Solar Power satellite requires multiple components, most of which will need to be built in orbit. For more details, see Solar Power Satellite Design Considerations. The components are:

  • Solar Panels (from high-grade silicon, in thin sheets such as commonly manufactured today)
  • Steel structures to frame and aim the panels.
  • Steel structures and motors to maintain alignment of the panels with the sun, and the antenna with the target Earth station.
  • Microwave transmitting antenna
  • Microwave transmitter
  • Shield mass to protect the electronics from meteorite damage.
  • Fuel to insert the SPS into the target geosynchronous orbit

I expect that the motors and electronics (including the microwave transmitter) will be launched from Earth, for which I’ve budgeted $200M per SPS.

A considerable amount of fuel is needed to move each newly constructed SPS from the near-Lunar location of its birth and insert it into an appropriate geostationary orbit. The total delta-V needed is roughly 2 km/s, the total mass is 25,000 tons, and even using 5,000 ISP VASIMR ion thrusters, 1,000 tons of fuel is needed. Luckily, that is a tiny fraction of the 15,000 tons of oxygen produced as a by-product of the smelting of iron ore into steel for each SPS. Also luckily, we’ll have a lot of power available, so the time required for this orbit change is limited largely by the number of ion thrusters (and thus the fuel flow rate) we wish to dedicate to this task. We’ll have plenty of oxygen to spare, so our ion thrusters should be optimized for this most available rocket fuel.

Each SPS will require a matching ground receiving station consisting of a large (10+ kilometer) rectenna plus power conversion and distribution. These costs are only a guess, but I’d estimate that $1B each is more than enough. I have not included this portion of the per-SPS cost in my summary, as this cost would be born by the receiving local electric utility.


Another post details the Design of a minimum Kalpana-One style Habitat: a 100-meter radius cylinder, 130 meters wide, spinning at 3rpm for 1G along the outer rim, lined with 3 meters of slag which provides excellent shielding against radiation and meteorites.

The outer pressure shell is built using average strength (easily welded) steel plates nearly 1 inch thick, and masses approximately 28,000 tons. Completed first, once spun up it provides a shirt-sleeve environment with gravity, boosting the productivity and health of the workers.

Over time, blocks of slag are brought inside to line the exterior walls, and interior structures are built as housing, workspaces, and farmland. Most of the volume is left open to eventually become a 12 acre cylindrical park. The core will be available as a low-gravity industrial and research space. Agriculture space (primarily for crops) is more than 5.5 million square feet; residential space is over 2.5 million square feet, there is 1.5 million square feet of offices / light industrial and a million square feet of storage and overhead space. The total internal structures including the outer 15 meters along the endcaps and perhaps 6 stories of rooms lining the cylinder rim, plus a 40 meter diameter core structure, masses 126,000 tons (a generous allowance). This is less tonnage than an Oasis Class Cruise Ship (225,000 tons carrying 8,000  passengers and crew) yet much more spacious.

Power (6 kw/inhabitant or 50 megawatts total) is provided by solar panels lining the exterior surface of one endcap, plus a ring extending 100+ meters beyond the cylinder (effectively a 425-meter diameter solar panel array).  Additional power needed for exterior factories (such as the smelter) is included in their structures and allowances. The spin axis of the habitat points to the sun, requiring a massive gyroscope at the center of the habitat to slowly precess the spin axis at one revolution per year.

Speaking of spin, it takes a lot of fuel and energy to spin up the habitat to 3rpm: about 1150 tons of fuel (oxygen) at an ISP of 5,000. Using 25 megawatts (half of the habitat’s available power), the spin up takes nearly 2 years. However, the empty shell (no shield mass) spins up very quickly; the majority of the energy is needed to maintain the spin as roughly a million tons of shield mass is gradually added to the habitat’s periphery. 

Once complete, this habitat houses 8,000 workers / colonists, and provides all life-support needs. All of the steel (154,000 tons) and shield mass (1,600,000 tons of slag) are from the asteroid, along with the oxygen, water, and carbon needed for the life support system. The known components to be launched from Earth include the LED’s and other light sources for the farms and interior (see Lighting our Space Habitats), various electronics, seeds, and perhaps 50 kilograms of nitrogen (per inhabitant) needed as fertilizer.

There are no additional launch or equipment costs not included elsewhere (possibly excluding the ion thrusters needed for spin-up).


This project is an eight-year plan (5 years to the first SPS) culminating in the deployment of a permanent, self-sustaining habitat with a population of 8,000 and fifteen 5-gigawatt Solar Power Satellites in geostationary Earth orbit. By completion, the project will have consumed approximately 10% of Apophis, leaving  90% available to build additional habitats and Solar Power Satellites.

The capacity will be in place to build additional Solar Power Satellites at the rate of 8 per year, but note that demand is much higher: current global electric demand would need 400 five-gigawatt SPS’s to fill, and total projected energy demand might require more than 4,000 by 2050 (most energy is used for heating, generally by burning fossil fuels). Apophis alone has sufficient resources to build up to 150 Solar Power Satellites (while building 10 habitats accommodating 8,000 people each), but in the long run additional asteroids will be needed.

We should build additional habitats, including larger ones, as part of an ongoing project. We’ll likely need a “mining town” for each asteroid we capture, and an SPS construction / maintenance habitat just outside of geostationary orbit, and even a large habitat in low Earth orbit (as low as possible without serious atmospheric drag) as an Earth to LEO launch target and tourist destination. Good arguments can be made for additional habitats in various Lagrange points such as L5, and for transition orbits such as LEO to Geostationary, or to the Lagrange points. There may even be a 3:1 Lunar synchronous stable Earth orbit that visits L3, L4, and L5 in succession, 9.1 days apart – an ideal platform from which to host interplanetary missions, including to Mars.


It all sounds time consuming and expensive: $22.8 Billion before that first SPS is operational. But remember, in 2032 that first SPS will generate $1.3B in cash flow (at a wholesale price of $0.03/kwh), in 2033 the 3 deployed SPS’s generate another $3.9B, in 2034 the 7 deployed SPS’s generate $9.1B, and subsequently the 15 SPS’s generate nearly $20B/year. The steel production rate is 800 tons per day, allowing 8 or more additional SPS’s to be built each year. The peak cash outlay is $25B with all investment repaid in 2035 – and huge net profits after that.

Even considering only that first SPS, the total cost is comparable to the $5 billion per gigawatt that my local power company may spend on a nuclear power plant, without any need for radioactive waste disposal or refueling, and without quite as much local objection to the installation.

Yes, I’ve been accused of being an optimist. Yes, it is likely to take longer than this project plan, due to unforeseen setbacks. Yes, I’ve ignored the cost of money in this analysis. I’ve even ignored the cost of bringing all these workers back to Earth (assuming they’d want to return). But even if it takes another year or two before breakeven, even if the costs might run 20% (or 100%) higher, even if you factor in the up-front R&D costs, even if you add a generous salary for every in-space worker plus free room & board, it is still a wildly profitable venture.

And did I mention that Apophis has enough raw materials to repeat this project another nine times, at a fraction of the cost each (and completely funded by free cash flow)?

Is there an entrepreneur listening that likes the sound of $200 Billion per year of free cash flow for a measly $30 billion investment? And with the bonus of earning a permanent place in the history books as the party responsible for bootstrapping humanity’s move into the Solar System?!