Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Appendix C LOGISTICS CONSIDERATIONS The shuttle in l9b6 has accomplished a long series of successful operational flights. Nonetheless, from a technical point of view, it is not yet out of the development phase. Design flaws, expected in a system of this complexity, are still being corrected. The spares complement is just being developed for main engines, solid rocket boosters, other line replacement units (LRUs), and so forth. In that respect, the past operational flights have been essential to determining the real, as opposed to the hypothetical, logistics needs for various sustainable flight rates. Consequently, a more efficient logistics program can now be set up than was possible some years ago. By the l990s, most major development should be completed and the logistics picture should have stabilized. This appendix endeavors to estimate what that picture will be. The panel accepts NASA's estimates of the immediate logistics needs. However, NASA has not systematically examined the consequences of the eventual loss, through use or accident, of Orbiters. Such losses must be expected, as NASA itself has stated in testimony on the Challenger loss. The complexity and uniqueness of the shuttle (critical elements, design margins, "rebuild" for every flight, etc.) reinforce that point. Well-recognized calculations relating system reliability, confidence level in accomplishing the mission, fleet size, and flight rate (equipment lifetime) have been used for years to determine the buy rate for aircraft, satellites, and other fleets. Figures C-l and C-2 show the relationships among these factors for the range of parameters applicable to Orbiters. Figure C-l shows the situation assuming no Orbiter losses: the upper curve gives an upper bound on flight rates per year so that there is a 50 percent chance of no Orbiter loss over the period. The lower curve represents an upper bound for a 90 percent chance of no loss. Figure C-2 shows the situation assuming one loss, i.e., the upper bounds on flight rates for no more than one Orbiter loss during the period. ("Loss" may be through wear out, severe overstress, or any accident that precludes further use.) An alternate way of showing the information in Figure C-l is given in C-3. Based on all experience to date, one would have to have unachievable reliability to have a high confidence manifest without some planned backup, workaround and/or replacement Orbiter. 33
34 SPARE PREFERRED SPARE DESIRED 0.97 0.98 0.99 SHUTTLE MISSION SUCCESS RATE 0.04 0.03 0.02 0.01 EQUIVALENT MISSION FAILURE RATE 0.00 Figure C-l The Chances of Losing Zero Orbiters l988 through 2000 Time Period.
35 pflfc *-â SPARE DESWEDS2 0 0.96 PRESENT 0.97 0.98 0.99 ACHIEVABLE? SHUTTLE MISSION SUCCESS RATE I 0.04 0.03 0.02 0.01 EQUIVALENT MISSION FAILURE RATE 0.00 Figure C-2 The Chances of Losing No More than One Orbiter l9b8 through 2000 Time Period.
1.0 0.1 m 0.01 0.001 1988-2000 (incl.) 0.96 0.97 0.98 0.99 MISSION SUCCESS RATE 1.00 Figure C-3 (An Alternative View of the Information in Figure C-l.) The shuttle system reliability to date is 0.96 (a failure rate of 4 percent). Given NASA's vigorous efforts at improved safety, this failure rate might be cut by a factor of 4; reducing it by a factor of l0 to a reliability of 0.99b is most unlikely in a short time or for costs less than the development cost to date (based on aircraft development experience).
37 It is important here to distinguish between having sufficient LRU spares and replacing a lost Orbiter.* The former is planned by NASA with appropriate budget requests; it is crucial for maintaining an acceptable Orbiter turnaround time. But, it assumes infinite life for the airframe, major structural components, and fittings. A noncatastrophic failure of an LRUâeven if the cause were a design problem that grounds the Orbiter fleet for a whileâis unlikely to affect the flight schedule for much more than a year. (The Challenger solid rocket problem will take somewhat longer but past ELV experience suggests that this is an exception.) Such schedule disruption can be accommodated through increased satellite design life of more than 5 years, as well as spares on orbit; planetary and space station programs clearly have more difficulty accommodating delays. The loss of an Orbiter, however, as demonstrated by the Challenger accident, has long-lasting effects on all missions. Missions are placed in jeopardy. Mission and industrial teams are in danger of breaking up, especially university scientific teams ana subcontractor teams. Time-critical satellites may be mothballed or discarded. The collateral costs are already a major fraction of the cost of another Orbiter. Yet this kind of disruption is inherent with small fleets of reusable, fully-booked vehicles with long replacement times. There are several possibilities for minimizing the severity of impacts. First, a full-up spare Orbiter, with another ordered later, would help alleviate the buildup of backlogs if the anomaly is not a systems design-related one. Satellite constellations, for example, frequently have spares on orbit, but such an alternative for the shuttle may be too expensive. Another, albeit riskier, possibility is a flight rate sufficiently below the maximum rate where a multiyear "surge" could alleviate schedule conflicts until a replacement Orbiter is brought on line. Shifting of "standard" dual-compatible satellites to ELVs along with a sufficient inventory of "ready-to-go" vehicles, might help. In any case, without an agreed strategy to accommodate Orbiter loss, long-term confidence in any shuttle manifest is certainly limited. This could be crucial for the space station, planetary launches, and some national security flights. The Report of the Presidential Commission on the Space Shuttle Challenger Accident and the statements of individuals who met with the panel identified a number of relatively short-term problems that the panel believes NASA is likely to solve by the early l990's: o The cannibalization of LRUs from one Orbiter for parts or repair of other Orbiters (on the order of 5U percent on past flights). o Limitations on nondestructive testing. *The NASA spares program does not supply a full "ship set" of spares, some of which have very long lead times. Consequently, a replacement Orbiter cannot be built just out of spares.
38 o Main engine and landing gear replacement and repair provisions. o Flight simulator limitations (technological and capacity). o Critical skills, spares, and maintenance requirements at KSC. o Critical skills, simulators, data storage and software development at JSC. o Establishment of a long-range logistics support plan, updated annually or as experience dictates. General concurrence in the plan and its financing by the Executive and Legislative Branches. o Agreement by the NASA and DoD as to which payloads would be compatible with both Orbiter and ELVs. In this connection, it is not necessary that all payloads be dual compatible in order to assure access to space. But those that are will require advance planning of launch support if changeover is to be practical. The panel also notes the continuing trend toward the formation of a Shuttle Operations Organization, which could be an important factor in achieving confidence in any shuttle manifest. A serious limitation to a reliable manifest is the maintenance of an industrial base to support the spares and replacement needs of the shuttle fleet. The shuttle fleet is a small one, with a low replacement rate. But according to Rockwell International, to ensure that a replacement Orbiter is available within, say, an l8-24 month call-up would require a production capability of 2 Orbiters on order all the time (or l every 2 years). However, an efficient industrial base can be maintained with a production rate of one Orbiter about every 3 years along with continued production of spare parts. But with an Orbiter fleet operating at about l0 flights per year and a loss rate of l to 2 percent, the replacement order rate would be one every 5 to l0 years, too low to retain an efficient industrial base. Increasing the shuttle's reliability to the point of not needing any further replacements nor an industrial base is not practical (see Figure C-l). A compromise would be a shuttle production rate of about one Orbiter every 4 years along with a national commitment to aggressive use of shuttle-unique capabilities.
