The majority of the SPS research performed to date has concerned the technology required for the large-scale satellites that are to be used to collect solar energy in space and transmit microwave energy to users on Earth. As a result of this work there is now fairly good understanding of the technological requirements of such systems - at least at the scale of pilot plants of several MW output. However, the main reason why the world electricity industry continues to give little support to SPS research is that the probability of SPS becoming competitive with other sources of energy is considered to be low, because all space activities are very costly. In addition, although the electricity industry is one of the largest industries in the world, it is quite separate from space engineering, and it is understandable that electrical engineers should not understand the great potential for reducing the cost of space activities.

The main reason why space activities are so costly today is because launch costs are very high - more than $10,000 to place 1kg in low Earth orbit. At such launch costs the construction and operation of SPS units with masses of thousands of tons would cost tens of $billions, which is much too expensive to be able to compete with other electricity generation systems.

However, with the end of the Cold War, taxpayers' willingness to pay for the activities of government space agencies has been declining, and their budgets are being cut. This has led space agencies to acknowledge that launch costs are too high: Mr Goldin, the administrator of NASA, even stated that the US space industry should "...hang their heads in shame" because they have not developed a new rocket engine for 25 years (1). This revival of interest in developing re-usable launch vehicles with much lower launch costs has created a growing body of opinion that, with appropriate technology development, reusable launch vehicles (RLVs) could be developed with operating costs of 10% of today's costs or less. Some of the more important projects under way are described briefly in the following section.

Most notable of the reusable launch vehicle projects to date was the McDonnell Douglas "DC-X" and "DC-XA" VTOVL (Vertical Take-Off Vertical Landing) rocket test-vehicles which were re-flown 15 times between 1993 and 1996, and which demonstrated many important steps towards low-cost reusable launch-vehicles, including rapid turn-around, low-cost operation, rapid proto-typing, and other achievements. The total budget of DC-X and DC-XA was about $100 million, most of which came from the US Department of Defense.

More recently the " X-33" reusable VTOHL test-vehicle is now under construction, using a budget from NASA of about $1 billion. It is intended that the X-33 will demonstrate several technologies relevant to developing a reusable Single-Stage-To-Orbit ( SSTO) vehicle, and will lead on to the development of the "Venture Star" commercial launch vehicle, with a payload of 20 tons to low Earth orbit, at costs in the range $1000-$2000/kg.

Whether the Venture Star will be commercially viable is controversial, even among those who are confident that the technological challenges will be solved, due to the difficulty of raising and repaying a commercial investment of some $5 billion (2).

NASDA is currently planning the development of a fully-reusable SSTO VTOHL "Rocketplane" launch system aimed at achieving a launch cost per ton of payload about 10% of that of their H-2A expendable rocket with a payload of 10 tons (3). The dry mass is estimated as 50 tons; the mass at take-off is estimated as 500 tons; the propulsion system uses LH2 and LOX to achieve specific impulse of 450 seconds; and the target lifetime is 100 flights. It is described in more detail in (3).

Figure 1: Reference concept of NASDA RLV (3)

In ESA, studies undertaken as part of the FESTIP program have led to the preparation of a "Central Reference Configuration" for a near-term fully-reusable launch vehicle. The design is currently being refined, but, using near-term technology, the favoured configuration is a rocket-powered SSTO VTOHL vehicle.

In addition to these government projects a number of private companies and organizations are working on RLV designs. Kistler Aerospace Corporation is currently developing the "K-1" reusable 2-stage VTOVL vehicle for launching small satellites described in (4).

A more innovative approach is Rotary Rocket Company's " Roton" concept of an "orbital helicopter" (see Figure 2), described in detail in (5). Roton offers the possibility of very low launch costs provided that its development can be completed successfully.Since 1993 the Japanese Rocket Society has been studying the design and development of the SSTO VTOVL " Kankoh-maru" passenger launch vehicle, as shown in Figure 3, which would carry 50 passengers to LEO (6). A cargo version is currently being designed (7), and is expected to have a payload capacity of some 5-6 tons, with an unusually wide cargo bay.

Figure 2: Rotary Rocket Co's VTOL " Roton" (5)

Figure 3: Kankoh-maru reusable launch vehicle (6)

In contrast to the above vehicles, Bristol Spaceplanes in Britain and Pioneer Rocket-plane in the USA favour horizontal take-off and landing ( HTOL), the former in an innovative 2-stage con-figuration (see Figure 4) and the latter in a "1.5 stage" configuration using in-flight re-fueling to permit take-off with minimal propellant load. Both use existing jet and rocket engines. These staged designs are necessary to make HTOL possible since SSTO horizontal take-off vehicles are not feasible with currently known technology.

Figure 4: Spacecab 2-stage HTOL RLV using high-altitude separation and existing engines (8)

SPS launch systems

For SPS studies this recent growth of interest in sharply reducing launch costs is very promising because it means that the central cost problem of SPS is beginning to be tackled. This is a major change from the past 20 years during which this problem was largely ignored by government space agencies, which used less than 1% of their budgets for research aimed at low-cost reusable launch vehicles.

The increased activity on reusable launch vehicles is promising for SPS for a second reason. Even in their early flight-test phase reusable launch vehicles could be useful for launching SPS pilot plants such as the "SPS 2000" system currently being designed in Japan, shown in Figure 5.

Figure 5: "SPS 2000" solar power satellite pilot plant (9)

The SPS 2000 satellite is currently estimated to have a mass of some 200 tons, and would be launched as a number of modular units. Thus it will require 20 launches of 10 tons each, or 40 launches of 5 tons each. Using ELVs with present-day launch costs this would

cost some $2 billion, which is several times the estimated cost of the SPS 2000 satellite.

Even at a cost of several $billions, the SPS 2000 project would still cost less than other comparable energy projects, such as a fast-breeder nuclear-reactor test plant. That is, an impartial cost-benefit analysis comparing the potential benefits of SPS 2000 with those of other such future energy projects such as fast-breeder reactors and nuclear fusion reactors would justify using a budget of even several $billions for SPS 2000. However, if such funding was available, it would be more cost-effective to use it for the development of a reusable launch system, and to use that for launching SPS 2000, rather than to purchase 20 or more high-priced expendable launches.

Synergy between SPS and reusable launch vehicles

A major problem facing RLVs is that they need many payloads to launch in order to achieve economical operation. Unfortunately the demand for launch of satellites and other payloads to orbit has low "price elasticity". That is, it is not expected to increase significantly even if launch prices fall considerably. Demand for launch can be divided into commercial demand for satellite launch, and government demand for research missions. Total launch demand is not expected to grow substantially beyond its current level of about $3 billion per year; Arianespace has estimated it to be $34 billion over the next 10 years, as shown in Table 1 (10).

Geostationary satellites	14-16
LEO	< 6
Space station	8
Earth observation	3
Science & technology	1
Total launch market	< 34

Although there are plans for several "constellations" of LEO satellites (66 in the Iridium system, and 288 in the Teledesic system) these will be launched at lower cost due to their large volume, and there is no likelihood of many such systems being built, due to saturation of demand. In addition, govern-ment budgets for launch are not expected to increase, but rather may be expected to decrease with a fall in launch prices.

In the face of such a limited market, the development of reusable launch vehicles creates a severe problem for manufacturers - the "50:50 Pinch" - illustrated in Table 2.

In round figures, manufacturers of expendable rockets throughout the world make about 50 rockets per year, which are used to launch about 50 satellites, earning revenues of some $3-4 billion per year worldwide (see Table 2 (a)). Although this business is quite small (for example, in car manufacturing, many individual companies earn more than $30 billion per year) it is relatively stable. However, if a single reusable launch vehicle is made which can launch one satellite per week, expendable launch vehicle makers' business would fall to zero (see Table 2 (b)). There is no other way to escape from this "Pinch" than to find a new launch market that will grow many times larger than the existing launch market (see Table 2 (c)).

Seen from this economic perspective, in the absence of major new launch markets, it is clearly against the corporate interests of existing makers to develop a reusable launch vehicle. For example, a company such as Lockheed-Martin, which currently earns $billions from expendable launch vehicle sales, faces a real conflict of interest working on " X-33" and a possible follow-on launch vehicle. If successful, its own business will decline substantially.

In the SPS studies by the US Department of Energy during the 1970s, Heavy Lift Launch Vehicles were conceptualized with payloads of as much as 400 tons to LEO. Present-day thinking about the appropriate development path for SPS has changed. While there will be a need for large-scale launch capability for commercial SPS units with power output of 1 GW or more and masses of thousands of tons, it is now understood that it will be more economical to launch these using many flights of vehicles with relatively small payloads. This has the major advantage of enabling economical, routine, airline-like operations to be achieved. That is, although building larger launch vehicles with more massive payload capability may achieve economies of scale, it also reduces both the number of vehicles that need to be manufactured, and the number of times that they are operated. But in reality, economies of large volume production and large volume operation are more important than the advantages of large unit size, and they can be achieved only with relatively smaller vehicles.

The aviation industry demonstrates this clearly: for decades aircraft have been developed that are small enough for the demand for air travel to justify making and operating them in large numbers. It is only as the demand for air travel has reached its present high level of tens of millions of flights per year that very large aircraft have come to be built. In theory, larger aircraft could have offered some economies of scale from an earlier stage, but these were less important than the economies of large volume manufacture and operations that have been achieved by successful models. Indeed, certain giant aircraft such as the Brabazon in Britain and the "Spruce Goose" in USA were famous failures of their time.

It is now believed that if reusable launch systems are operated in large volume - hundreds of flights/year each by tens or hundreds of vehicles - then they could achieve similar cost reductions to aircraft, of which the operating costs have fallen continuously for decades as the demand has grown by several orders of magnitude. However, this of course depends critically on the demand for launches being sufficiently large, as shown in Table 2.

It is in order to exploit this possibility that the Japanese Rocket Society has been studying the design and development of the " Kankoh-maru" launch vehicle, and a cargo version of it, due to the very large potential demand for passenger travel to orbit that is known to exist (11). As in the aviation industry, the cargo version of Kankoh-maru will benefit from the economies of scale in both manufacturing and operation achieved by making and operating the passenger version on a large scale.

In this situation, SPS 2000 and other SPS pilot plants can be excellent customers for a new reusable launch vehicle at an early stage of its operational history, and commercial SPS units will be very important customers for mature reusable launch systems requiring launch of thousands of tons of equipment and components of many sorts. Thus in the near future there could be valuable synergy between the design of the new generation of reusable launch vehicles and the design of SPS pilot plants which will require launch services on a relatively large scale.

The shape of future commercial SPS systems cannot be known in detail today,　before a single pilot plant has been built or operated. However, the overall scale of future SPS systems is already well understood since it depends on the demand for electricity. Over the next 100 years, global demand for electricity can be expected to grow from its current level of some 1,000 GW of capacity to about 10,000 GW, based on the assumption that most countries will aim towards the average supply capacity in developed countries of about 1 kW per person (12). Thus, in order to make a significant contribution to world electricity supply, it will be necessary for SPS to generate hundreds of GW of power - which will require thousands of square kilometers of solar cells. Even assuming that structures are developed which are significantly lighter than are possible today, satellites of this area will have a mass of 1 million tons or more.

From both economic and engineering points of view it is probable that, rather than lift such large masses of components from Earth to high orbit, it will be more profitable to collect raw materials from the lunar surface, and from asteroids and comets in convenient orbits. That is, even if the cost of launch to LEO falls to $100 / kg, the price of materials in orbit will still exceed $100,000 / ton, which is 30 times the price of aluminium, and more than 100 times the price of steel or liquid oxygen at the Earth's surface, using existing technologies. This provides a wide cost margin within which companies could extract raw materials from non-terrestrial bodies and deliver them profitably to users in Earth orbit (13).

In order to use liquid oxygen or other raw materials from the Moon, electric launch from the lunar surface using linear motor propulsion has been under investigation for years, and there is clear potential to achieve very low launch costs based on the use of solar-generated electric power.

Research on a wide range of subjects related to this possibility is under way at various research centers in Japan, including NASDA, ISAS and major corporate laboratories, including detailed surveys of the lunar surface, a lunar lander mission (14), lunar base design and construction, extraction of oxygen from Moon-rock (15), processing of lunar materials, asteroid missions, and other subjects. These and related activities are likely to play an important role in the future transportation requirements of SPS.

Recent activity aimed at developing and operating low-cost, reusable launch vehicles have improved the prospects for sharply reducing launch costs. Which vehicles will be successful is still uncertain, as several different concepts are still in competition. However, all successful work towards reusable launch vehicles improves the prospects for SPS, both by improving the prospects of low launch costs, and in convincing the electricity industry and energy policy makers of SPS's feasibility.

In addition, plans for SPS pilot plants such as SPS 2000 have a potentially important role to play in encouraging prospective manufacturers of reusable launch vehicles to continue their work. By creating and making visible potentially large demand for a new generation of vehicles, and by providing details of the launch services that they require, planners of SPS pilot plants can help launch vehicle designers to design vehicles for which there will be large-scale demand. Without this launch cannot become much cheaper, and will continue to depend on taxpayers rather than becoming a commercially profitable business.

1. D Goldin, 1996, " Slash costs to open the space frontier", Aviation Week & Space Technology, Vol 144, No 9, p74
2. J Grey, 1996, " The rocky road to space launch heaven", Aerospace America, Vol 34, No 11, pp 20-25
3. H Nanri, K Kato and H Taniguchi, 1997, " The investigation status of NASDA RLV system", Proceedings of 7th ISCOPS, AAS in press
4. Kistler Aerospace Corporation, 1996, " An introduction to Kistler Aerospace Corporation", KAC, Kirkland, WA
5. http://www.rotaryrocket.com/
6. K Isozaki et al, 1994, " Considerations on vehicle design criteria for space tourism", IAF paper No. IAF-94-V.3.535
7. P Collins and K Isozaki, "The JRS Space Tourism Study Program Phase 2", Proc. 7th ISCOPS, AAS in press, 1997
8. D Ashford, 1994, " A preliminary feasibility study of the Spacecab low-cost spaceplane and of the spacecab demonstrator", ESA Contract No. 10411/93/F/TB
9. M Nagatomo et al, 1994, "Conceptual study of solar power satellite SPS 2000", Proc. ISTS, Paper No. ISTS-94-e-04
10. P Selding, 1997, " Civil launch rivals discuss cooperation", Space News, Vol 8, No 2 22, p 17
11. R Akiba and M Nagatomo Eds, 1996, Special Issues on Space Tourism, Space Energy and Transportation, Vol 1, No 1, pp 9-65
12. P Collins, 1993, "The promise of electricity from space for world economic development," Proc. 5th Int. Energy Conference, Vol 3, pp 50-59
13. M Sonter, 1996, "On the technical and economic feasibility of mining the near-Earth asteroids", University of Wollongong, Department of Civil and Mining Engineering
14. S Sasaki et al, 1997, " Scientific objectives of SELENE mission", Proceedings of 7th ISCOPS, AAS in press
15. Shimizu Corporation, 1996, " Research on lunar oxygen production", Space Technology Division Brochure