When Physics, Economics, and Reality Collide
The Challenge of Cheap Orbital Access

by

John M. Jurist, M.D.
Sam Dinkin, Ph.D
David Livingston, DBA


Contents - Section Two

Discussion
Merging the Rocket and the Payload
Market Decision-Making and Economic Factors
Interest
Insurance
Research, Development, and Fabrication
The Range Problem


Discussion

It is apparent that the previous analyses are highly sensitive to some variables: Payload mass, range and insurance costs, and, in the case of ELVs, vehicle production costs. In the case of RLVs, R&D costs must be higher because of vehicle complexity, designing and testing for multiple flights, and incorporation of recovery systems. Then, recovery, inspection, and refurbishing costs add into the mix.

Flight volume (production run in the case of ELVs or the product of fleet size and vehicle flight lifetime in the case of RLVs) is a very relevant factor. Past history may put various volume projections into perspective. Launch histories are available [Ref. 18]. In the 48 years since the first orbital launch in 1957, there have been 4,700 orbital launches in addition to about 22,000 suborbital space (exoatmospheric) launches. Considering a specific commercial vehicle line, 144 Arianes were launched in the 25 years between 1979 and 2003. Only 21 Ariane-5 vehicles were launched in the nine plus years between 1996 and February, 2005.
The effects of using RLVs over a flight lifetime of 100 flights has been discussed, but a 1,000 or more flight RLV remains an unproven concept.

Payload mass is a critical variable. For example, assume that demand for launch services is much higher than in the previous analyses – 100,000 pounds of payload per launch over 500 flights. Because of the economies of scale, the payload fraction as a percentage of dry vehicle weight is increased. In this case, it is assumed to be 40 percent for a large ELV. This is roughly comparable, but slightly better than, the overall figure for the first two stages of the old Saturn-V system. Then, the required propellant mass (assuming LOx/RP-1) is 1,018,981 pounds per launch of a 100,000 pound payload by a two stage vehicle with 150,000 pounds of dry structure. The assumed ELV fabrication cost was reduced from $75 per pound to $48 per pound (big things are cheaper by the pound than little things). This results in a vehicle structure cost of $7.2 million. Assuming that the launch facility cost is quintupled to $10 million for the larger vehicle and that program R&D is $400 million, the model results in a direct launch cost of $75 per pound of payload to LEO and indirect costs of $56 per pound for a total of $131 per pound. This figure includes interest of 12 percent annually over a 10 year program lifetime. Thus, developing an ELV roughly 20 percent of the size of the old Saturn-V system could markedly reduce the direct costs of payload delivery to LEO, but the overall program costs would be about $6.5 billion to deliver 50 million pounds to LEO. This appears to be beyond the ability of the alt.space community to fund without governmental participation. For $1 billion, we could get a respectable six million pounds to LEO with an RLV for 60 flights versus one half million pounds for our 500 smaller rockets, but $670 million in R&D, vehicle and startup costs is a lot of money to raise and pay interest on. For $1 billion of big dumb ELVs, we could get 76 flights and 7.6 million pounds. The maximum capital in this instance would only be about $418 million to get started. For the big boosters, insurance is the key issue again. Self-insuring the $60 million RLV will pay for itself in six flights. Self-insured for first party property, the big RLV gets 148 flights for our $1 billion -- about 2.5 times as many. This analysis leaves out interest, which will be expensive for the big vehicle given the radically lower flight rate (one percent of the small vehicle if payloads are interchangeable).

The production run of an ELV program is also significant. Figure 1 on the next page shows the flyaway costs (including 12 percent interest in a 10 year program) for various production runs and payload sizes as discussed previously.

The extreme cases of 500 units produced for payloads of 1,000 and 100,000 pounds were discussed in the text. Increasing production past 1,000 units has very little effect on cost to LEO. A simulated production run of 10,000 units differed from the 1,000 unit case by less than 21⁄2 percent which is indistinguishable on the graph. If a given production run is planned for amortization of, for example, R&D cost, and the program is cut prematurely, significant losses can occur. This is a very real business risk.

Very small payload capacity leads to different economics. By going to a 100 pound payload, higher reliability can be demonstrated leading to lower insurance premiums for expensive payloads. The capital requirements for self-insuring are also lower. A 10,000 flight program would get lower insurance costs after more than 100 successful launches. Smaller payloads might change on-orbit construction requirements or might change bulk cargo economic considerations. Relatively small payloads might be most useful for flights to a manned station such as the International Space Station or a Nautilus. There are fixed weight elements regardless of the size of the vehicle such as the avionics and communications. These take a heavy toll on small launchers.

Examination of the ELV and RLV factors leads to another consideration. The difficulties of designing, building, and maintaining reusable vehicles are well established. The recovery process and required systems for a reusable first stage are potentially much simpler, and therefore less expensive, than for the orbital stage. It appears that the STS got it exactly backwards by having the heavy RLV components in the orbiter and using an expendable external tank and relatively short-lived SRBs. That is, the first stage of a cheap system should be recovered and reflown, and the upper stage(s) should be expendable. Note that this is the approach to be used in the SpaceX Falcon series.



Figure 1. Flyaway cost per pound to LEO versus payload size and production run for a 10 year program.

The RLV analysis resulted in some interesting findings. First, payload delivery costs are relatively independent of RLV flight lifetime if the vehicles can be reused more than 100 times. Second, fleet sizes of more than about 20 vehicles do not provide much economy under the analytical assumptions used. As in the case of ELVs, larger payload vehicles lead to significantly reduced flyaway costs per pound of payload. This economy comes at the expense of a significant increase in research and development costs. Furthermore, a large RLV capable of launching 100,000 pounds into LEO is an unproven concept. These findings are illustrated in Figures 2 and 3 on the next page.

Figures 4 and 5 show the reduction of flyaway costs with larger payload vehicles. These figures also show that the cost improvements with increased vehicle flight lifetime are greater for smaller fleets than for larger fleets.

These analyses are simplified in that they assume that all development costs are amortized over the life of the specific vehicle program. They do not consider the effect of vehicle evolutionary processes on development, manufacturing facilities, or manufacturing costs.



Figure 2. Flyaway cost versus vehicle lifetime and fleet size for 1,000 pound payloads.



Figure 3. Flyaway cost versus vehicle lifetime and fleet size for 100,000 pound payloads.



Figure 4. Flyaway cost versus payload size and vehicle lifetime for a fleet of 5 RLVs.

Figure 5. Flyaway cost versus payload size and vehicle lifetime for a fleet of 40 RLVs.


Merging the Rocket and the Payload

There are several ways to rewrite the rules of the game to make the economics friendlier. If the top stage is the payload, then substantially more mass is devoted to payload. The residual fuel that would normally be unused could be salvaged as could the engine and the fairing. For orbital construction, these parts could be smelted or otherwise recycled. Avionics and communications may be useful as is. If not, clever usage of standard electronic components may allow the avionics to have more general computational uses in space. This approach necessitates a manned orbital facility for payload processing and recycling.If the structural components are not desirable as payload, it may be possible to modify the structural components to make them into useful payload. For example, if the cargo is water, the cargo could have some sawdust added to it and frozen. At that point, it would have the structural strength of a battleship (Pykrete) and could potentially be used to replace a portion of the structural mass of the upper stage. Another possibility might be the use of edible materials for sound or thermal insulation.

Market Decision-Making and Economic Factors

Many factors influence market analysis. They include costs for interest, insurance, research, development, and fabrication. In addition, potential market share risks must be analyzed in the context of the anticipated market. Regulatory factors including range costs and insurance requirements must be considered and the associated costs and uncertainties of attempting to influence these factors must be evaluated. These factors will be discussed below.

Interest

The straight line amortization of R&D and RLV production costs is a simple approximation to a more detailed business model. R&D and a large fraction of total production costs must be paid for up front. Unless all flights of the program are conducted in the same year, there will be interest costs to pay if the money was borrowed. If the money was not borrowed, inflation has reduced the value of the dollar received versus the dollars invested and the rate of return on investment has to beat the return for similarly risky investments or else few will want to invest. Patrick Collins calculated comparable activities at 18 percent annual cost of capital [Ref. 19].

If 100 flights are spread out over 8 years, the total of amortization and interest costs for R&D are approximately the percentage cost of capital times the R&D cost divided by the number of flights per year (i.e., the principle payment is negligible at the beginning of paying off R&D). This is the point where Ted Taylor’s first law applies: “There shall be no more than one flight per month.” If 1,000 flights were stretched out to 83 years at one per month, interest would cost $1,125,000 per flight. Even in an eight year program, interest expenses or capital costs would require $7 million per year on a $75 million investment with all of the R&D costs and production costs incurred prior to the first year. This results in an interest cost of $56 per pound. Even for big dumb RLVs launched monthly, the interest would be about $10 million per flight, cutting in half the number of flights for $1 billion.

If the interest is compounded every flight at regular time intervals, the relevant relationship is:


where A is the amount of R&D principal and interest paid per flight, R is the total R&D expense, i is the annual interest or discount rate, n is the number of flights in the program, and y is the program lifetime in years.

Insurance

Insurance costs can be reduced by extending the flight testing program. Assume that the vehicle is 99.9 percent reliable -- the recommended goal of Henry [Ref. 9]. It will be difficult to prove that the vehicle exhibits that reliability level. Insurers are likely to assign a high loss rate until the vehicle has shown many successful flights. Insurers are likely to use something akin to the following equation:


where p is the premium paid per flight. MPL is the maximum probable loss, f is the number of failures, s is the number of successes, and π is the gross profit ratio. If a successful 500 flight test program is conducted, with a π of 3, the insurance premium will be about 0.6 percent of MPL. Assuming the MPL is about the same as for Pegasus ($40 million), the third party liability insurance costs would be $240,000 per flight. The trouble is that adding 500 flights would cost $1 billion mostly because they still need to be insured! As an alternative, a clever study may be able to show that the vehicle is safer than the rate implied by the flight history [Ref. 16]. Using direct auctions to obtain insurance may reduce insurance costs by ten to 25 percent.

Research, Development, and Fabrication

The R&D costs used in this analysis are almost certainly low. Ashford presents a graph of historic development cost trends against dry vehicle mass for demonstration and prototype aircraft, airliners, advanced aircraft, ELVs and manned spacecraft [Ref. 20]. Table 5 on the next page compares the R&D costs used in our model compared to those derived from Ashford’s data for a prototype (ELV) and an advanced aircraft (RLV).

In current dollars, the research and development cost of the X-15 suborbital rocket plane program was estimated to be $1.415 billion [Ref. 9]. The cost of the Mojave Aerospace Ventures Space Ship One development is proprietary, but is generally considered to be in the $20 million to $40 million range.

A similar comparison of our model with established fabrication costs can be made. Table 6 compares the structure or fabrication costs used in our model compared with similar costs of various vehicles expressed in 2004 dollars [Ref. 9].

Extrapolation of what is not forbidden by science and technology to what may be possible involves a risk that does not exist with already proven technology. That risk usually results in cost overruns and can occur in relatively mundane activities. The published R&D cost of the Airbus A380 (approximately $12 billion) was between $1 billion and $2 billion over budget. Keep in mind that the A380 was an evolutionary end product that benefited from the development of prior generations of vehicles. Those prior generations subsidized A380 R&D to an unknown extent.

The rationale for use of markedly lower vehicle costs than past history would suggest comes from drawing on an existing base of knowledge and experience, using existing technologies and components whenever possible, and running lean organizations rather than organizations based on long histories of cost plus government contracting. Yet, the analysis does not incorporate the cost of maintaining the organization during the R&D phase or the costs associated with the time during which demand is created or grows to a steady-state level. The analysis also does not account for co-development of a family of vehicles such as Falcon-I and Falcon-V.

As a general proposition, demand increases as cost per unit decreases. Thus, it is reasonable to conclude that launch service demand will increase as price per pound to LEO decreases. Henry concludes that the demand will be relatively inelastic (one percent decrease in price leads to no more than a one percent increase in demand) until cost decreases to around $1,000 per pound. Elasticity may occur in the $1,000 - $2,000 range if orbital tourism materializes, but that market remains unproven. If launch demand proves to be lower than anticipated, an ELV program has R&D and manufacturing investment at risk. An RLV program has R&D and already-built vehicles at risk. This risk component can be abated somewhat by keeping the production line open and waiting for additional demand to develop before fabricating more RLVs.

A very important consideration in creating a large scale business with large capital investment is to clearly identify the products or services to be supplied rather than adhering to the “build it and they will come” philosophy of hoping for demand to manifest itself.

The Range Problem

Range costs are a large component of the total flyaway costs of launching into orbit. Comprehensive analysis of orbital launch economics shows that cost-effective launch operations are inhibited by high (and perhaps unwarranted) range fees imposed on a launching company for use of a federal range. While the adverse economic impact of range fees has been discussed, those range fees are only a symptom of the underlying disease. The fundamental problem is badly outdated federal launch and operational systems and an infrastructure that will not support high flight rates and commercial operations.

One of the best concise explanations of this problem was provided by James Muncy [Ref. 21]. Muncy pointed out that the federal range launch infrastructure was historically developed for missile (IRBM and ICBM) launches. This governmental infrastructure is now maintained, operated, and updated by numerous government contracts. For example, the U. S. Air Force has a contract for operations, another company has the contract to repair broken equipment, another company handles R&D needs, and still another company handles the procurement of new or replacement equipment. There is no single entity with the authority to make rational economic decisions about the launch infrastructure. The systems are outdated and they are not user friendly. It is challenging if not impractical for a federal range to cycle more than one launch daily. Therefore, the existing federal range launch infrastructure cannot support the commercial launch needs anticipated by private launch companies with their commercially designed rockets and the projected markets.

Discounting range fees does not solve the underlying problem although such discounting may allow a more cost effective economic profile. However, discounting does nothing to modernize the obsolete infrastructure, systems, and management of federal ranges. In contrast, the Mojave, California civilian flight test center and spaceport was designed to test new systems and to utilize modern operating methodologies -- something federal ranges are unable to do. The Mojave Air and Spaceport works by establishing a flight and user friendly environment. The federal ranges cannot do this given their legacy and operational profile.

If the goal is to have multiple space flights carrying many passengers, it is essential to run the spaceports or ranges as rational economic entities. It may be possible to privatize and streamline federal range operations, but there is not much interest among the range operating entities or the U.S. Congress to do this. Lower range fees can certainly help a private rocket company, but the federal range will still be operated uneconomically and require taxpayer subsidy.

Pressure to overhaul the federal ranges and launch infrastructure might increase as private spaceports receive licenses for vertical launches. A good example of potential pressure can be seen by examining the developing private range in New Mexico. This range plans to offer private rocket companies state of art launch and range services designed to be commercial from the ground up. When the Southwest Regional Spaceport of New Mexico does begin offering launch services, what will happen to the Florida Spaceport which is stuck in the federal range environment? If Florida wants to compete for private launch business, it will have to modernize. The modernization program will involve far more than just lower range fee pricing or the use of available military facilities at discounted prices. Simply put, without comprehensive modernization, the existing Florida launch infrastructure will not easily support competitive commercial launch services. While Wallops Island and Kodiak Island may be able to launch private rockets with lower range costs, their infrastructure is still not suitable for anticipated commercial operations. There are also issues with these lower cost ranges that relate to orbital mechanics which make them less desirable.

At present, private rocket companies must launch vertically and use a federal range with negotiated range fees, use air launch to avoid federal ranges, attempt to launch from a barge in the ocean, or launch from outside the United States.

As previously discussed, air launch has technical limitations. Furthermore, air launch still incurs range and tracking fees if the launch trajectory intrudes on a federal range. This is precisely the case with Pegasus and the necessity to pay high Vandenberg Air Force Base range and tracking fees even when launching over the Pacific Ocean [Ref. 22].

Range fees could be potentially reduced by launching outside of the United States. This is the rationale behind the SeaLaunch concept, which launches in international waters, but use of foreign ranges has a significant regulatory risk: the International Traffic in Arms Regulation (ITAR). This puts regulation of space launch activities involving foreign governments or foreign nationals by U.S. entities under the U.S. State Department. ITAR has proven to be extremely burdensome to academic as well as commercial international collaborations. Technology-related export control issues also exist. At present, Europe and others are many years behind the U.S. in modernizing regulation, but it is entirely possible that the foreign regulatory regime will be more advantageous in the future.

Given the present status of the private rocket industry, it is possible to negotiate lower range fees and operate a vertical launch system profitably since flight rates and demand are still low. However, for the industry to develop and eventually mature, federal ranges and launch infrastructure issues must be addressed. This requires broad modernization of facilities and systems. Furthermore, private spaceports must be established and licensed for sustainable commercial operations. Failure to modernize the federal ranges may result in the development of private spaceports and the ultimate abandonment of federal ranges by private rocket companies.

At present, the impact of range fees must still be considered in business planning. Thus, this paper considers the present impact of the high range fees on potential commercial orbital launch operations.

<< Return to Section 1 / Go to Section 3 >>