Results of Parametric Analysis

Although the first two reports, each carried the title, "Cost Optimization of Large Booster Systems," dated November 1959 and July 1961, were not approved and released, the third report, "Proposed Minimum Cost Space Launch Vehicle System," July 1968, incorporated significant material discussed in the unreleased reports. The third report is available from The Aerospace Corporation and contains much of the work, some in greater detail, that will be discussed in these columns.

For those concerned with the political aspects of this work: since minimum-cost launch vehicles may continue to be unwanted by the aerospace community, the second report was reviewed for technical accuracy by three, outside consultants. In my mind this was a delaying action by management since the report had received the usual internal reviews. When the consultants submitted their reports in which they expressed whole-hearted agreement and I pressed for release of my report, I was told to drop the project.

I quit Aerospace in August 1962 and accepted employment as a consultant to the new, NASA Associate Administrator for R&D; he was one of the consultants who reviewed my report and believed that NASA would greatly benefit from using the MCD criteria. My assignment was to brief personnel at Washington Headquarters and Wernher von Braun's team in Huntsville to gain their concurrence. (I plan to cover this experience in a later column.) After many months of waiting for a reaction, I was told that NASA programs are too far along to permit making a far-reaching change in design criteria. Not able to secure employment in my former discipline of structures with some of the local, prime aerospace contractors, I returned to Aerospace in May 1963. I was assigned to the ballistic missile division.

I was indeed surprised when, after several months, I was asked to design a survivable ballistic missile weapon system using the MCD criteria. Several years later I was given another assignment: apply the criteria to an MCD space launch vehicle, and after that, to a redesign of an existing payload. Between these programs I worked on the "high-priced line" in the Titan III program office where I gained valuable, relevant experience. It is noted that the released report also contains a description of the MCD/SLV design.

The parametric analysis, initiated in the first report, led to the following conclusions:

• The cost of typical, multi-stage, expendable space launch vehicles could be appreciably reduced.
• The optimum hardware cost, in $/lb, increases with each upper stage; that is, the optimum sophistication of the second stage is higher than that of first stage, and so on with each, successive upper stage. However, the final stage that reaches LEO should have a lower sophistication level than current hardware.
• Lower values of specific impulse increases optimum hardware cost.
• Minimum-weight, expendable stages that have high structural factors do not lend themselves to much in cost savings when designed for minimum cost.
• The optimum sophistication of a single stage increases rapidly (more quickly with a flyback stage that has a higher structural factor) as its velocity approaches that required for LEO; therefore, in the interest of minimizing cost, at least two stages are called for in reaching LEO.
• Considerable R&D is required to devise new forms of hardware that range in cost, in $/lb, from close to commercial to near the state-of-the-art of aerospace hardware. This calls for the development of an array of materials, forms of construction and fabrication techniques, together with appropriate manufacturing facilities and the methods of transportation to the launch site. Research and development may be required to increase the current level of hardware sophistication for very high-velocity missions.
• Kick stages, which impart small, incremental velocities to payloads, should use unsophisticated hardware because the increase in stage weight is negligible.
• It is not clear whether MCD pressure-fed or pump-fed propulsion systems would be less expensive. A design example of a simplified, higher-weight turbopump should answer this question.
• Payloads should have the same sophistication level as the final stage. It should be realized that in the boundary condition, when launch costs approach zero, payload costs could approach zero as well.
• It may be economical to recover a first stage if this can be done by simple means and if only minor refurbishment is necessary. If this could be devised, the cost-optimum velocity of the stage should increase as well as its sophistication level, particularly if propellant costs are significant as a result of reuse. If recovery significantly lowers the stage cost, the sophistication levels of the upper stages and payloads should be somewhat less, thus providing additional cost savings.
• It would be uneconomical to retrieve and reuse stages that reach orbital velocity. Such stages, because they carry recovery gear and hardware to shield against the thermal and aerodynamic environments, must be designed for minimum weight. These hardware configurations would result in much high costs than achievable by simple, expendable stages.
• Most importantly, in designing a launch vehicle for minimum nonrecurring and recurring (life cycle) costs, all elements of the complete system including such items as facilities, operations, readiness, and reliability must be considered simultaneously. This is in contrast to the strict design of a minimum-weight launch vehicle that, by definition, need not be related to the rest of the system.

• MCD launch vehicles, of the same basic design, should be manufactured in a variety of sizes. An easily scaleable propulsion subsystem could make this possible.
• Minimum cost payloads may be more efficiently designed if such a family of launch vehicles exists. It would provide assurance of the existence of a launch vehicle large enough to carry the final design that generally experiences weight growth.

• Consider a first stage. If one pound of weight is added to the stage to reduce its hardware cost in $/lb or to increase its reliability, the weight of the stage must increase by more than one pound in order for the stage to have the same performance. Say, the weight increases by a factor is five; this is known as the "growth factor." It is composed of the added pound of weight and four pounds of incremental tankage, propellant, and other subsystems. Since hardware cost is many times more than propellant cost in $/lb, it is reasonable to expect a minimum-cost stage to be slightly heavier but cost much less a minimum-weight stage.
• Consider a second stage. If the growth factor of the second stage is the same as the first stage, a pound of weight added to the second stage will increase the weight of both stages by a factor of 30. This explains why the optimum sophistication of a second stage is somewhat higher than that of a first stage.
• It follows that subsequent stages and payloads should be designed to higher levels of sophistication.

In 1962, while waiting for NASA's decision on whether they wished to pursue the MCD criteria, I prepared a paper in which I fully described the criteria and its application. The paper was submitted to the AIAA for publication and it was rejected.
 
 

Based on the criteria above, what do you think an ideal minimum cost vehicle would look like? [clicking on the above opens link to discussions that followed the initial posting of this column]

Next Column: A description of the minimum cost design criteria.

Return to Published Columns