Reducing Launch Operation Costs: NewTechnologies and Practices

September 1988

NTIS order #PB89-136402

Recommended Citation:U.S. Congress, Office of Technology Assessment, Reducing Launch Operations Costs: NewTechnologies andl%actices, OTA-TM-ISC-28 (Washington, DC: U.S. Government Print-ing Office, September 1988).

Library of Congress Catalog Card Number 88-600539

For sale by the Superintendent of DocumentsU.S. Government Printing Office, Washington, DC 20402-9325

(order form can be found in the back of this report)

Foreword

Reducing the costs and improving the reliability of space transportation are keyto making more effective use of the space environment for commerce, science, explora-tion, and defense. In order to achieve these objectives, the United States needs to givegreater attention to launch and mission operations, the collection of processes and pro-cedures used to ready vehicles and spacecraft for launch and insertion into orbit. Launchoperations make up a significant percentage of launch costs.

The United States already uses or has under development a variety of technologiesthat can make launch operations more reliable, efficient, and cost effective. However,as this technical memorandum explains, the United States has spent relatively little ef-fort in applying them to operations. Just as important as cost saving technologies areappropriate management methods, or strategies, to put these technologies to work. Insome cases, OTA has found, cost savings could be achieved by streamlining operationsand reducing the burden of documentation and reporting requirements that have slowlyexpanded over the years.

This memorandum is part of a broader OTA assessment of space transportationrequested by the House Committee on Science, Space, and Technology, and the SenateCommittee on Commerce, Science, and Transportation. In July, OTA released a com-panion volume, Launch Options for the Future: A Buyer’s Guide, which explores sev-eral possible options for space transportation systems.

In undertaking this technical memorandum OTA sought the contributions of a broadspectrum of knowledgeable individuals and organizations. Some provided information,others reviewed drafts. OTA gratefully acknowledges their contributions of time andintellectual effort. As with all OTA reports, the content of this technical memorandumis thesarily

. . .Ill

Advisory Panel on Reducing Launch Operations Costs:New Technologies and Practices

L.M. Granger Morgan, ChairHead, Department of Engineering and Public Policy

Carnegie-Mellon University

I.M. BernsteinProvost and Academic Vice PresidentIllinois Institute of Technology

Michael A. BertaAssistant Division ManagerSpace and Communications GroupHughes Aircraft Company

Richard E. BrackeenPresidentMartin Marietta Commercial Titan, Inc.

Edward T. GerryPresidentW. J. Schafer Associates, Inc.

Jerry GreyDirector, Science and Technology PolicyAmerican Institute of Aeronautics and

Astronautics

William H. HeiserConsultant

Otto W. Hoernig, Jr.Vice PresidentContel/American Satellite Corporation

Donald B. JacobsVice President, Space Systems DivisionBoeing Aerospace Company

John LogsdonDirector, Space Policy InstituteGeorge Washington University

Hugh F. Loweth *Consultant

Anthony J. MacinaProgram ManagerIBM Federal Systems Division

George B. MerrickVice PresidentNorth American Space OperationsRockwell International Corporation

Alan ParkerSenior Vice PresidentTechnical Applications, Inc.

Gerard Pie]Chairman EmeritusScientific American

Bryce Poe, IIGeneral, USAF (retired)Consultant

Ben R. RichVice President and General ManagerLockheed Corporation

Sally K. RideProfessor, Center for International Security

and Arms ControlStanford University

Tom RogersPresidentThe Sophron Foundation

Richard G. SmithSenior Vice PresidentJLC Aerospace Corporation

William ZersenProject ManagerSpace Flight SystemsUnited Technologies Corporation

NOTE: OTA appreciates the valuable assistance and thoughtful critiques provided by the advisory panel members. The viewsexpressed in this OTA report, however, are the sole responsibility of the Office of Technology Assessment. Partici-pation on the advisory panel does not imply endorsement of the report.

iv

OTA Project Staff on Reducing Launch Operations Costs:New Technologies and Practices

Lionel S. Johns, Assistant Director, OTAEnergy, Materials, and International Security Division

Peter Sharfman, International Security and Commerce Program Manager

Richard DalBello, Project Director

Ray A. Williamson, Principal Analyst

Eric O. Basques

Michael B. Callahan

Stephen W. Korthals-Altes

Gordon Law

Contractor

Trudy E. Bell

Administrative Staff

Jannie Home Cecile Parker Jackie Robinson

Workshop on Space Launch Management and Operations, Sept. 10, 1987

Jerry Grey, ChairmanDirector, Science and Technology Policy

American Institute of Aeronautics and Astronautics, Washington, DC

Col. William Anders, USAFWestern Space and Missile

CenterVandenberg Air Force Base, CAStewart BailyManager,Govt. & Commercial SystemsARINC Research Corp.Annapolis, MDHal BeckAssistant Division ChiefMission Planning and AnalysisJohnson Space CenterHouston, TXAldo BordanoBranch ChiefGuidance and NavigationJohnson Space CenterHouston, TXAlan DelunaProgram Manager,Operations IntegrationLockheed Space Operations Co.Titusville, FL

Cort DurocherProgram Manager,Space and Missiles GroupHughes Aircraft Co.Los Angeles, CADouglas HeydonPresident, Arianespace, Inc.Washington, DCJames HollopeterManager, ALS OperationsSpace Systems DivisionGeneral DynamicsSan Diego, CALyle HollowayDirector, Florida Test CenterMcDonnell DouglasCape Canaveral, FLAnthony J. MacinaProgram ManagerOnboard Software SystemsIBM Federal Systems DivisionHouston, TX

Allan McCaskillManager,Launch Vehicle ProgramINTELSATWashington, DCDouglas MorrisAerospace EngineerSpace Systems DivisionNASA Langley Research CenterHampton, VAWalter J. OverendGeneral ManagerPrograms & Performance

EngineeringDelta Airlines Engineering Dept.Atlanta, GAJ.D. PhillipsDirector,Engineering DevelopmentKennedy Space Center, FL

Acknowledgments

The following organizations generously provided OTA with information and suggestions:

Aerojet Corp. McDonnell Douglas Rowan Companies, Inc.Aerospace Corp. NASA Headquarters U.S. Air Force Cape CanaveralBoeing Aerospace Operations Co. NASA Johnson Space Center Air Force Station

General Dynamics NASA Kennedy Space Center U.S. Air Force Directorate of

Lockheed Space Operations Co. NASA Langley Research CenterSpace Systems

Martin Marietta NASA Marshall Space Flight CenterU.S. Air Force Space Division

This report has also benefited from the advice of many space transportation experts from the Governmentand the private sector. OTA especially would like to thank the following individuals for their assistance andsupport. The views expressed in this report, however, are the sole responsibility of the Office of TechnologyAssessment.

Harry BernsteinAerospace Corp.Darrell BranscomeNASAWilliam CaseMartin MariettaJohn DiBattistaNASA

W.E. FieldsMartin MariettaHardie FordKennedy Space CenterJohn P. FredricksMcDonnell DouglasJohn GainesGeneral Dynamics

Charles GunnNASADavid MooreCongressional Budget OfficeWilliam StroblGeneral DynamicsCol. John WormingtonU.S. Air Force

vi

Related OTA Reports

Civilian Space

●

●

●

●

●

●

●

●

●

●

●

●

●

Launch Options for the Future: a Buyer’s Guide. OTA-ISC-383, July 1988. GPO stock #052-003-01117-4; $5.00.

Commercial IVewsgathering From Space. OTA-TM-ISC-40, May 1987. GPO stock #052-O03-O1066-6; $3.00.

Space Stations and the Law: Selected Legal ksues. OTA-BP-ISC-41, September 1986. GPO stock#052-003-01047-O; $3.75.

lrtternational Cooperation and Competition in Civilian Space Activities. OTA-ISC-239, July 1985. NTIS ord-er #PB 87-136 842/AS.U.S.-Soviet Cooperation in Space. OTA-Th4-STI-27, July 1985. GPO stock #052-O03-O1004-6; $4.50.

Civilian Space Stations and the U.S. Future in Space. OTA-STI-241, November 1984. GPO stock#052-003-00969-2; $7.50.

Remote Sensing and the Private Sector: Issues for Discussion. OTA-TM-ISC-20, March 1984. NTIS order #PB84-180777.

Salyut: Soviet Steps Toward Permanent Human Presence in Space. OTA-TM-STI-14, December 1983. GPOstock #052-O03-O0937-4; $4.50.UNLSPACE ’82: A Context for International Cooperation and Competition. OTA-TM-ISC-26, March 1983.NTIS order #PB 83-201848.

Space Science Research in the United States. OTA-TM-STI-19, September 1982. NTIS order #PB 83-166512.

CiviZian Space Policy and Applications. OTA-STI-177, June 1982. NTIS order #PB 82-234444.

Radio frequency Use and Management: impacts From the World Administrative Radio Conference of 1979.OTA-CIT-163, January 1982. NTIS order #PB 82-177536.Solar Power Satellite Systems and Issues. OTA-E-144, August 1981. NTIS order #PB 82-108846.

Military Space

SD]: Technology, Survivability, and Software. OTA-ISC-353, May 1988. GPO stock #052-003-01084-4; $12.00.Anti-SatelZite Weapons, Countermeasures, and Arms Control. OTA-IX-281, September 1985. GPO stock#052-003-01009-7; $6.00.BaZZistic Missile Defense Technologies. OTA-ISC-254, September, 1985. GPO stock #052-003-01008-9; $12.00.

Arms Control in Space. OTA-BP-ISC-28, May 1984. GPO stock #052-O03-O0952-8; $3.00.

Directed Energy Missile Defense in Space. OTA-BP-ISC-26, April 1984. GPO stock #052-003-00948-O; $4.50.

NOTE: Reports are available through the U.S. Government Printing Office, Superintendent of Documents, Washington, DC 20401 (202)783-3238; and the National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161, (703) 487-4650.

vii

CONTENTS

Page

Chapter I: Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Chapter 2: Major Issues in Launch and Mission Operations . . . . . . . . . . . . . . . . . . 13

Chapter 3: Operations Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Chapter 4: Technologies and Management Strategies . . . . . . . . . . . . . . . . . . . . . . . . 53

Chapter S: Operations Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

Appendix A: Cost-Estimating Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Appendix B: Noneconomic Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Appendix C: Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

. . .Vlll

Chapter 1

Executive Summary

CONTENTS

Page

BoxBox

Page

I-A. Launch Operations ‘Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 3 Principal Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 4

Chapter 1

Executive Summary

INTRODUCTION

Achieving low-cost, reliable space transporta-tion is one of the most important space policychallenges facing the United States today. The Na-tion’s ability to assure timely access to space, toguarantee the general welfare of U.S. civilianspace activities, and to compete effectively withother countries depends on meeting this challengesquarely and thoughtfully.

Ground and mission operations processes arehighly complex and involve a wide variety of tech-nologies. As support functions, they only becomeobvious to the public and to Congress when theyfail to work properly. Because they constitute asignificant percentage of launch costs, reducingthe costs of these operations is crucial to lower-ing the overall costs of space transportation. Im-

could also lead to greater flexibility and respon-siveness to changing conditions in space activities.Yet these relatively mundane processes and proce-dures seldom receive close scrutiny from the Con-gress, or attention from the policy community.

This technical memorandum is intended to helpCongress understand the launch process and howthe use of advanced technologies and managementtechniques could reduce the costs of launchingpayloads. It does not discuss the management ofpayloads or crews for passenger-carrying vehicles.

The memorandum is part of an assessment ofadvanced launch technologies, which was re-quested by the House Committee on Science,Space, and Technology, and the Senate Commit-tee on Commerce, Science, and Transportation.It derives in part from a workshop held at OTAon September 10, 1987, which met to discuss is-

proed operations technology and management sues of launch operations technology and man-

Box l-A.—Launch Operations Processes

Launch operations includes the procedures necessary for launching payloads to orbit. It does not in-clude the management of payload or crew (for piloted vehicles) on orbit, which are generally consideredmission operations. Launch operations can be divided into the following overlapping steps:

● Processing and integration of vehicle: includes the assembly and testing of the launch vehicle, aswell as the integration of electrical, mechanical, and fluid systems. For reusable, or partially reusa-ble vehicles, this step also includes testing of refurbished components to assure that their character-istics remain within design specifications.

c Processing and integration of payloads: comprises the assembly, testing, and mechanical and elec-trical integration of payloads with the launch vehicle. Payloads must also be tested with the vehi-cle’s mechanical and electrical systems to assure they will not interfere with proper operation ofthe launch vehicle.

● Launch management and control: includes the preparation and testing of the launch pad, the con-trol center, and all of the other facilities critical for launch, as well as the actual launch countdown.During countdown, each critical subsystem must be continually monitored.

. Post launch responsibilities: includes the retrieval, return, and refurbishment of all reusable vehiclecomponents, and the cleanup and post-launch refurbishment of the launch pad. The launch of re-usable, or partially reusable, vehicles introduces an extra layer of complexity to the launch processand involves additional facilities and personnel.

c Logistics: encompasses the provision of spares, and replacement parts, as well as the scheduling oftasks, personnel, and equipment, which must be coordinated across the entire launch process.

3

4

agement, and from OTA staff research. OTA staffvisited Air Force and National Aeronautics andSpace Administration (NASA) facilities’ and

gathered additional information from a literaturereview and personal interviews with individualsfrom the major aerospace firms.

1OTA site visits included: Air Force Space Division, Los Angeles;Cape Canaveral Air Force Station; Edwards Air Force Base; John-son Space Center; Kennedy Space Center; Langley Research Cen-ter; Marshall Space Flight Center; Vandenberg Air Force Base.

PRINCIPAL FINDINGS

Finding 1: Because launch and mission operationsconstitute a sizable fraction of the cost oflaunching payloads to orbit, developing newlaunch vehicles will not, in itself, result in sig-nificant reductions of launch costs. If the UnitedStates wishes to reduce launch costs, systemdesigners and policymakers must give greaterattention to operations.

Because launch and mission operations are re-sponsible for up to 45 percent of the cost of eachlaunch, lowering these costs is crucial to reduc-ing the overall cost of space missions. Promptedby the needs of the spacecraft community, launchsystem designers have traditionally focused great-er attention on achieving high performance thanon operational simplicity or low cost. Recently,plans for a permanently inhabited space station,more extensive Department of Defense (DoD)space activities, and problems with existing U.S.launch systems have suggested the desirability ofattaining routine, low-cost launch operations.NASA and DoD have funded several studiesaimed at identifying technologies and manage-ment practices capable of reducing the costs oflaunch services.z The results show that a varietyof technologies, either new or in use in other in-dustries, could help to reduce operations costs.They also indicate that important reductions oflaunch costs are unlikely unless launch operationsengineers and facilities managers have a greaterrole in the design of future launch systems. Thedevelopment process should encourage a thor-

2The results of these studies are summarized in the report of theSpace Transportation Architecture Study: U.S. Government, Na-tional Space Transportation and Support Study 1995-2010, Sum-mary Report of the Joint Steering Group, Department of Defenseand National Aeronautics and Space Administration, May 1986.

ough and frequent interchange of information andideas among representatives from operations, lo-gistics, design, and manufacturing. It should alsocontain sufficient incentives for reducing costs.

Finding 2: Technologies capable of reducing therecurring costs of ground and mission opera-tions exist today or are under development ina variety of fields.

These include technologies for:●

●

●

●

●

●

●

●

●

built-in test equipment;management information systems;automated test and inspection;advanced thermal protection systems;fault-tolerant computers;adaptive guidance, navigation, and flightcontrol;automated handling of launch vehicles andpayloads;computer-aided software development; andexpert computer systems.

Some of these technologies could be incorpo-rated into the design of the launch vehicle. Forexample, built-in test equipment and softwarecould be used to detect faults in vehicle sub-systems, reducing ground operations labor andcost. Other technologies might find applicationin the launch and mission operations facilities. Forexample, management information systems couldsharply reduce the amount of human effort inmaking, distributing, and handling paper sched-ules and information. Such systems could also re-duce the number of errors experienced, and speedup sign-off procedures.

The amount of money such new technologiescould save, either from building new launch sys-

5

terns, or from enhancing existing systems, dependsstrongly on three factors: 1) the demand forlaunch services, 2) the non-recurring costs of tech-nology development and facilities, and 3) savingsachieved in operations. Unless launch demand forthe late 1990s increases sharply over current esti-mates, adopting new technologies could actuallyincrease the total (life-cycle) cost of space trans-portation.

Finding 3: Dramatic reductions in the costs oflaunch operations (factors of 5 to 10) could beachieved only under highly limited conditions.

Most experts OTA consulted thought that re-ductions in operations costs by five- or ten-fold,as suggested by the Space Transportation Ar-chitecture Study and the Advanced Launch Sys-tem (ALS) program,3 were unobtainable in prac-tice even with proposed new technology and newfacilities. They pointed out that although the largecapacity vehicles contemplated for an ALS mightsave costs by carrying more weight, they wouldnot be efficient for smaller payloads. In addition,new ground facilities (e.g., launch complexes,fabrication and assembly buildings) typically re-quire investments of several hundred million oreven billions of dollars. Such investments seldomlook attractive in the short run—the most rele-vant time period in a stringent budget environ-ment—and are therefore seldom adequatelyfunded. Finally, dramatic reductions in cost wouldrequire significant changes in the institutionalmechanisms of launch operations, which wouldbe very difficult to achieve without considerableinstitutional upheaval.

Such reductions would require high launch de-mand, a new generation of launch vehicles andground facilities designed to accommodate rapidturn around, and payloads of uniform design andorbital characteristics. In theory, it would be pos-sible to create new advanced-technology launchsystems, such as those proposed for the ALS pro-gram. These launch systems would be most ben-eficial for launching many payloads with similartechnical and orbital characteristics, such as com-

ponents of a space-based missile defense system,or perhaps fuel to send humans to and from Mars.Absent a decision to deploy SDI, or to increasesharply spending on civilian payloads, the num-ber and diversity of payloads NASA and DoDnow plan to launch through the late 1990s do notmeet the conditions necessary for dramatic costreductions.

Thus, under these conditions, the discountedlife-cycle cost—the total of recurring and non-recurring costs, appropriately discounted—oflaunching known or currently projected payloadsprobably can be reduced only marginally by de-veloping completely new launch systems. In addi-tion, because a revolutionary launch design suchas envisioned for ALS would involve new designapproaches and some new technologies,4 the tech-nological and economic risks would be higherthan for an evolutionary approach.

Finding 4: If the Federal Government wishes toinvest in new operations technologies, it shouldhave clear long-term goals and a well-definedplan for developing and incorporating newtechnologies in space transportation operations.Such a plan must be buttressed by data fromnew and more reliable cost models.

NASA and the Air Force are funding researchon new technologies for launch systems. NASA’sCivil Space Technology Initiative (CSTI) is pur-suing research on a number of technologies, in-cluding autonomous systems and robotics, thatcould improve some launch procedures and mighteven lead to cost savings. NASA and the Air Forceare collaborating on research in the AdvancedLaunch System’s Focused Technology Program,which may contribute to reducing the costs oflaunch and mission operations. Yet these researchprograms devote only a small percentage of theirbudget to space transportation operations. Inaddition, no well-organized or well-funded planexists to apply the technologies developed in theseprograms to launch operations procedures, or tocoordinate research being carried out through theexisting technology R&D programs.

3See especially U.S. Congress, Office of Technology Assessment,Launch Options for the Future: A Buyer’s Guide, OTA-ISC-383(Washington, DC: U.S. Government Printing Office, July 1988),for an extensive discussion of launch system costs and capabilities.

4U. S. Government, National Space Transportation and SupportStudy 1995-2010, Summary Report of the Joint Steering Group, De-partment of Defense and National Aeronautics and Space Admin-istration, May 1986.

6

In fiscal year 1989, NASA plans to start an Ad-vanced Operations Effectiveness Initiative, whichwould develop and carry out a plan for insertingthe results of technology R&D into launch andmission operations. However, NASA is allocat-ing only $5 million to this initiative in the 1989budget, an amount that will have a very smalleffect on reducing launch and mission operationscosts.

To complicate matters, the current restrictivebudgetary environment makes it difficult to spendmoney now on research and facilities that mightsave money later. To respond to this suite of tech-nical, institutional, and budgetary challenges, theUnited States needs a coherent long-term plan fordeveloping and incorporating new operationstechnologies into existing and future launch sys-tems. A technology development plan should in-clude work in all development phases:

● broad technology exploration (basic re-search),

● focused research leading to a demonstration,and

● implementation to support specific appli-cations.

Such a plan should be part of a more comprehen-sive National Strategic Launch Technology Planthat would develop and insert new technologiesinto U.S. launch systems.

Instituting a long-term research, development,and technology application plan will be extremelydifficult for three reasons. First, policymakers inCongress and the Administration have been un-able to agree on overall long-term goals for thepublicly funded U.S. space program. Operationsprocedures optimized for our current level ofspace activities would differ substantially fromthose designed to deploy space-based defenses ormount a mission carrying humans to Mars.

Second, current ground and mission operationsare partially controlled or influenced by the tech-nologies and management requirements from adozen or so different research centers, hundredsof technical projects, and thousands of individ-uals in NASA, DoD, and the aerospace industry.The Administration’s latest space policy statementdirects NASA and DoD to cooperate in pursuing

“new launch and launch support concepts aimedat improving cost-effectiveness, responsiveness,capability, reliability, availability, maintainabil-ity, and flexibility.”s This directive could providethe impulse for developing a national research anddevelopment plan. However, the institutionalstructure and will to focus the efforts of these in-terested parties on the common purpose of reduc-ing operations costs does not presently exist. UntilCongress and the Administration reach agreementon specific national space policy goals, develop-ing an effective, detailed, multi-year plan for de-veloping and incorporating new technologies intospace transportation operations will be extremelydifficult. Encouraging NASA and DoD to reduceoperations costs substantially may require mak-ing major institutional changes to these agencies,or developing a new agency for operations.

Finally, the lack of objective, verifiable cost esti-mation models makes it difficult to determinewhich technologies are worth pursuing or whichshould be discarded. Credible, objective opera-tions cost methods—similar to those of the air-line and other commercial industries—should bedeveloped, which would allow the Governmentto estimate the total cost of incorporating a newtechnology or management practice and the sav-ings it could generate. Current models haveproven inadequate, in part because data on pre-vious launch operations experience have neitherbeen collected in an organized way nor properlymaintained. Without adequate historical data touse as a benchmark, cost estimation involves toomuch guesswork. Congress may wish to directNASA and DoD, or some independent agency,to collect the necessary historical data and to de-velop better cost estimating methods for spacetransportation systems.

Finding 5: Although making evolutionary im-provements to exizting launch systems mayprove difficult and expensive, such improve-ments could reduce the cost of existing launchand mission operations.

Because launch vehicles and their ground sup-port facilities are highly integrated and interdepen-dent, it is difficult and expensive to incorporate

“’Presidential Directive on National Space Policy, ” White HouseOffice of the Press Secretary, Fact Sheet, Feb. 11, 1988.

7

new cost or time saving technologies. Neverthe-less, experts consulted by OTA agreed that itwould be possible to reduce operations costs byimproving vehicle subsystems such as onboardavionics, and many ground-based support activ-ities such as payload handling and fuel loading,through redesign, automation, and standardiza-tion. Technologies pursued for new launch sys-tems may have application to existing systems andvice-versa.

Finding 6: It will be difficult to improve the waythe United States manages its launch operationswithout making significant changes to the in-stitutions currently responsible for those oper-ations.

Current U.S. space management practices re-sult from a launch operations philosophy that em-phasizes long-lived, expensive payloads, high-performance launchers, very high reliability, andlow launch rates. The Soviet Union, on the otherhand–both by choice and as a result of its limitedtechnology base—has in the past relied on rela-tively inexpensive short-lived satellites, reason-ably reliable vehicles, and very high launch rates.As a result, the Soviet launch infrastructure ismore “resilient” than its U.S. counterpart, al-though not necessarily more effective at accom-plishing national goals.

The United States is now in the difficult posi-tion of attempting to retain its high-technology,high-performance approach to payloads and ve-hicles while attaining Soviet-style routine, lowercost access to space. This goal is probably un-attainable unless the U.S. Government substan-tially alters the way it conducts space transpor-tation operations. Such an alteration wouldrequire significant changes to the institutionalstructure and culture of NASA and DoD.

Congress could direct the Air Force and NASAto:

●

●

●

turn launch operations for all new launchsystems over to the private sector;establish operations divisions fully independ-ent of launcher development, including de-velopment of a Shuttle or an ALS; orpurchase launch services, rather than vehi-cles, from the private sector for existing ELVlaunch systems.

One way to manage the institutional challengeis to maintain separate institutions for launch ve-hicle development and operations by turning overoperation of new launch systems to the privatesector. Under such an arrangement, the launchcompany would assume control of launch oper-ations after the systems were developed andwould provide launch services to the Governmenton a contractual basis. In order to further reduc-tions of cost, the company would also be en-couraged to market its services to other payloadcustomers, either from the United States orabroad. The European Space Agency (ESA) andArianespace have demonstrated that such an ar-rangement can be highly effective. ESA fundeddevelopment of the Ariane launch system underthe management of the French space agency,CNES. Arianespace, S. A., a French corporation,which manages the Ariane operation and marketsthe Ariane launcher worldwide, set requirementsfor a successful commercial venture.

Although the European model may not be fullyapplicable to U.S. conditions, Congress must findways to give space transportation operations ex-tra visibility and “clout” so they will not be con-sidered a costly afterthought. Congress coulddirect NASA and the Air Force to establish oper-ations divisions fully independent of each agency'slaunch development organization, with the chargeof operating launchers on the basis of increasedefficiency and reduced costs. This would requireconsiderable congressional oversight to assure thatthe agencies carried out the will of Congress.

NASA and DoD could also reduce operationscosts by purchasing all expendable launch serv-ices, rather than launch vehicles, from the privatesector for existing systems. Recent Administra-tion policy directs the civilian agencies, includ-ing NASA, to purchase expendable launch serv-ices from private companies. However, the policyallows considerable latitude for DoD to continueits current practice of involving Air Force person-nel deeply in the launch process.

Finding 7: In addition to new technologies, adopt-ing new management practices and designphilosophies could increase the efficiency andreduce the cost of ground operations.

8

Management strategy may often be more im-portant than new technology for achieving lowcost

●

●

●

●

●

launches. Cost-reducing strategies include:

reduce documentation and oversight,create better incentives for lowering costs,provide adequate spares to reduce cannibali-zation of parts,develop and use computerized managementinformation systems, anduse an improved integrate/transfer/launchphilosophy.

Some management strategies could be enhancedthrough the appropriate use of technology. Forexample, OTA workshop participants pointed outthat operations costs will never fall significantlyunless ways are found to reduce the time con-sumed by human documentation and oversight.In many cases, automated procedures would re-duce the need for certain documentation, andcertainly shrink the necessary manpower to main-tain it. However, reducing the amount of over-sight significantly will be much more difficult.Since the Titan and Shuttle losses of 1985 and1986, the number of Government personnel re-sponsible for contractor oversight has increased.

Also needed are incentives to encourage loweroperations costs. The current institutional man-agement structure tends to penalize launch fail-ure, but is poorly structured to reward the lower-ing of launch costs or increases in launch rate.However, the Strategic Defense Initiative Orga-nization found in a recent project6 that it was ableto cut overall project costs in half by incorporatingsimple, common-sense management techniquessuch as reducing Government oversight, delegat-ing authority to those closest to the technical prob-lem, maintaining short schedules, and paying em-ployees bonuses for meeting deadlines. Althoughthe team was able to achieve some of its opera-tions cost savings as a result of a concentrated,narrow effort that would be difficult to maintainfor routine launches, the project nevertheless dem-onstrated that a management philosophy that in-cludes incentives for launch managers and tech-nicians can play a significant role in reducing thecost of launch operations.

bThe Delta 180 experiment. See ch. 2, Issues, for a discussion.

Vehicle design can also play a crucial part inthe ability to reduce launch and mission opera-tions costs. The accessibility of critical parts, theweight and size of components, and the abilityto change out modules quickly all affect the speedand effectiveness of operations. Several designprinciples are particularly important. One should:

●

●

●

engage all major segments of launch team inlaunch system design process;design for simplicity of operation as well asperformance; anddesign for accessibility, modularity, and sim-plicity of operation.

For example, considering all elements of thelaunch system, including the operations infra-structure and operations management, as a col-lection of highly interactive parts will allow sys-tem designers to anticipate potential operationsand maintenance problems and provide for thembefore the system is built. As was discovered withthe Space Shuttle main engines, certain sub-systems may pose unexpected maintenance prob-lems. All major subsystems should be designedto be readily accessible, and, as much as possiblewithin weight and size constraints, should also beof modular design in order to reduce maintenanceand integration costs.

Many concepts for improved launch operationstend to shift costs from operations to other stagesin the launch services process, such as payloadprocessing. For example, requiring payloads toprovide their own internal power, rather thanrelying on a source in the launcher, may reduceground operations costs, but could also increasethe cost of preparing payloads. In altering thestructure of space transportation operations, suchchanges in procedure or technology should notmerely send problems elsewhere.

Finding 8: Unless the Government can stimulatethe innovative capacity of the private sector,private sector contributions to reducing thecosts of space transportation operations willcontinue to be quite limited.

Almost all of the recent effort in improvinglaunch operations has been instigated by NASAand the Air Force in connection with the Space

9

Transportation Architecture and the AdvancedLaunch System studies.g Private sector initiatives,such as competitive bidding for components, andintroducing some new technologies, show whatcan be done in a modest way to reduce costs.However, these efforts are still relatively limitedand reflect the tenuous nature of the U.S. com-mercial launch industry.

The Government could use the talents of theprivate sector most effectively, and in the proc-ess encourage a more competitive industry, bypurchasing the services of expendable launchersrather than vehicle systems, and by offering strongincentives for decreasing costs. Although it istheoretically possible for the Government to pur-chase services for piloted launchers, such as theSpace Shuttle, private industry is unlikely to of-fer such services in the near future because thetechnologies of reusable vehicles are still imma-ture and the costs of change are great.

Congress could also enhance the developmentof new operations technologies and assist privatesector competitiveness by funding an “operationstest center” composed of a mock launch pad andfacilities. Such a center should be specifically de-signed to enable tests of new technologies for in-corporation into existing and new launch systems.The ability to try out new operations technologieson a working launch pad is limited. A center

‘U.S. Government, National Space Transportation and SupportStudy 1995-2020, Summary Report of the Joint Steering Group, De-partment of Defense and National Aeronautics and Space Admin-istration, May 1986.

8The results of seven contractor reports for phase I of these studieshave not yet been released. ALS phase II studies are scheduled tobegin in August 1988.

would give the private sector the opportunity totry out new operations technologies free from thedemands of routine operations. Such a centercould be a government-owned, contractor-oper-ated facility. Alternatively it could be partiallyfunded by the private sector, and operated by aconsortium of Government agencies, private sec-tor companies, and universities.

Finding 9: For certain aspects of launch opera-tions, the broad operational experience of theairlines and the methods they employ to main-tain efficiency may provide a useful model forspace operations.

Although airline operations face different tech-nical and managerial constraints than space launchoperations, certain airline methods used in logis-tics, maintenance, task scheduling, and otherground operations categories could make launchoperations more efficient and cost-effective.

The following airline practices could be of par-ticular interest for space transportation oper-ations:

●

●

●

●

●

●

●

●

involve operations personnel in designchanges;develop detailed operations cost estimationmodels;stand down to trace and repair failures onlywhen the evidence points to a generic fail-ure of consequence;design for fault tolerance;design for maintainability;encourage competitive pricing;maintain strong training programs; anduse automatic built-in checkout of subsys-tems between flights.

84-755 - 88 - 2 : QL 3

Chapter 2

Major Issues in Launchand Mission Operations

CONTENTS

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Major Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Issue A: Can New Technologies and Management Strategies ReduceOperations Costs?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1ssue B: Is the United States Devoting Adequate Attention To Reducing theCosts of Space Transportation Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Issue C: What Factors Impede the Introduction of New Technologies andManagement Strategies in Launch and Mission Operations? . . . . . . . . . . . . . . 18

Issue D: What Impediments To Reducing Operations Costs Are Unique tothe Space Shuttle?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Issue E: What Can the Operational Experience of the Airlines Contribute toSpace Operations? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Issue F: Does the United States Possess Adequate Techniques To Judge theRelative Benefit of Improvements in Launch and Mission OperationsProcedures? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Issue G: Are the Near-Term Launch Systems Under Study by NASA and theAir Force Likely To Generate Major Reductions of Launch OperationsCosts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Issue H: How Do Concerns for Launch System Reliability Affect LaunchOperations? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

BoxesBox Page

2-A. Required Changes to Other Shuttle Subsystems If Shuttle Main EnginesAre Altered Significantly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2-B. The Delta Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

FiguresFigure Page

2-I. Shuttle Recurring Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132-2. Titan IV Estimated Cost per Flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132-3. Delta 180 Project Schedule Reductions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152-4. Cost of Achieving Extremely High Reliability. . . . . . . . . . . . . . . . . . . . . . . . . . . 27

TablesTable Page

2-1. Cost Reduction Factors for the Delta 180 Project . . . . . . . . . . . . . . . . . . . . . . . . 162-2. Approaches to Low-Cost Launch Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172-3. Civil Space Technology Initiative Funding.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182-4. Advanced Launch System Focused Technology Program . . . . . . . . . . . . . . . . . 192-5. Steps in the Changeout of Defective Parts in the Shuttle Orbiter When

Replacement Spares Are Unavailable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Chapter 2

Major Issues in Launchand Mission Operations

INTRODUCTION

Launch and mission operations constitute a sig-nificant fraction of the cost of launching payloadsto orbit. For example, prior to the loss of Chal-lenger, Shuttle operations costs, including missionoperations, accounted for about 46 percent of thecost of a flight. Of that, ground operations to-taled at least 24 percent (fig. 2-l). Projected life-cycle costs of the Shuttle suggest that some 86 per-cent of the total can be attributed to the recur-ring costs of launch and mission operations.1

Because of recently mandated safety-related mod-ifications, recurring costs are likely to be higherwhen the Shuttle flights resume. For today’s ex-pendable launch vehicles (ELVs), operations costsare generally a smaller percentage of the total, inlarge part because these vehicles do not containreusable components and do not carry humans.However, they are still significant. For example,in the Titan series, launch operations costs canreach about 20 percent of total costs per flight (fig.2-2).

‘National Aeronautics and Space Administration, “Shuttle GroundOperations Efficiencies/Technologies Study,” Kennedy space c~rlt~r,NAS1O-11344, May 4, 1987.

Figure 2“1 .—Shuttle Recurring Cost(percent per flight)

Operations

Flight 1 0 “ Y “ ’ ”Ops ‘7

16 550/0

Launch ops

O~erations costsHardware

(percent of total recurring costs)SOURCE’ National Aeronautics and Space Administration.

Attempts to reduce operations costs must copewith the complexity of launch and mission oper-ations, and the relative lack of policy attentionthey have received over the years. Workshop par-ticipants and others who contributed to this study2

identified the following primary issues that shouldbe addressed in developing a sound Federal pol-icy toward reducing costs and increasing efficiencyof launch and mission operations.

‘The many interim reports related to the Space TransportationArchitecture Study and the Advanced Launch System effort pro-vided much of the initial information for OTA’s effort. In addition,the study team interviewed officials from the Air Force, NASA, andprivate industry.

Figure 2-2.-Titan IV Estimatad Cost per Flight(in millions)

64%

Hardwarea

$83

16% \20%

Warrtwre costs based on annual buy rote of WyrbGowrnment ~ts include propellant, r~go SUpp@, mcemtwes, Aerospace Corfx3-

ratlon support, etc

SOURCE: Aerospace Corp

13

14

MAJOR ISSUES

ISSUE A: Can New Technologies and Manage-ment Strategies Reduce Operations Costs?

Existing Systems

Evolutionary improvements to existing launchsystems appear to provide opportunities for mak-ing modest, but meaningful, reductions in groundand mission operations costs. Reducing operationscosts for existing launch systems generally meansreducing the size of operations staffs and short-ening the time it takes to prepare and launch avehicle. Vehicle subsystems, such as avionics, andmany ground-based support facilities can be im-proved through redesign, automation, and stand-ardization.3

It is extremely costly to shorten vehicle turn-around and processing substantially by makingincremental upgrades of the vehicle, because ve-hicle subsystems are highly integrated and inter-dependent. As a result, altering one subsystemoften requires changing others. For example, evensmall alterations of the orbiter outer structure mayrequire significant changes of parts of the ther-mal protection system. Box 2-A presents a list ofchanges that could be required in other systemsif the design of the Shuttle main engines werematerially altered. Such changes would involvemultiple NASA centers and contractors, and re-quire considerable coordination.

Commercial launch companies are investing inperformance improvements and exploring waysto reduce launch operations costs. For example,General Dynamics has developed a new avionicspackage for the Atlas-Centaur that reduces theweight of the avionics package and increases itsreliability. It also includes self-testing proceduresthat will reduce operations costs slightly. Otherlaunch companies are exploring similar ways ofreducing costs of the launcher and launch oper-ations.

Because changes in the design of vehicle sub-systems often have a direct effect on ground oper-ations or mission operations procedures, it is im-

3National Aeronautics and Space Administration, “Shuttle GroundOperations Efficiencies/Technology Study, ” KSC Report NASIO-11344 Boeing Aerospace Operations Co., May 4, 1987.

Box 2-A.—Required Changes to Other ShuttleSubsystems If Shuttle Main Engines Are

Altered Significantly

● Main Engine Controller (computer) hardwareand software.

● Engine interface Unit Hardware—device thatcouples the main engine controller computerto the General Purpose Computer network.

● Flight Software Applications executing in theGeneral Purpose Computer Complex.

The Pulse Code Modulation Master Unit dataaccess programs and telemetry formats—device that receives data from the main enginefor telemetry to Earth.

● Various Ground Checkout hardware and soft-ware at K!K—especially the Launch Process-ing System applications software.

c Mission Control Center software—used f o rmonitoring of engine performance duringlaunch.

● Main Engine Environment Models—used inthe following simulation and test facilities:—Software Production Facility—Shuttle Mission Simulator—Shuttle Avionics Integration Facility—Various flight design engineering simulators

portant for design engineers to work closely withoperations personnel to establish the best way toproceed in making changes appropriate to oper-ations and maintenance processes. Whether par-ticular changes will result in net reductions in life-cycle costs will depend on a variety of economic,technical, and managerial factors. Chapter 4 dis-cusses these factors for several specific cases.

Although new technology or design changesmay lead to reduced costs, management changesmay be more important. For example, a recentStrategic Defense Initiative Organization experi-ment called the Delta 180 Project used new man-agement techniques to achieve reduced projectcosts. Project managers found that decreasing theburden of oversight and review, and delegatingauthority to those closest to the technical prob-lems, resulted in meeting a tight launch scheduleand reducing overall costs.4 Maintaining a short

4Department of Defense Strategic Defense Initiative Office/ Ki-netic Energy Office, “Delta 180 Final Report, ” vol. 5, March 1987.

15

schedule (fig. 2-3) reduced overheadentire project by about 50 percent.was achieved by giving contractors

costs of theIn part thiscash incen-

tives to achieve the demanding project schedule.Company employees shared in bonuses paid tothe company for meeting deadlines, which gavethem strong incentives to increase productivity.Table 2-1 lists the major factors that led to lowercosts and shorter project schedules for the Delta180 project. Although the team was able to achievesome of its cost savings as a result of a focused,narrow effort, which would be difficult to main-tain for routine launches, the project neverthelessdemonstrated that management philosophy canplay a significant role in reducing the costs oflaunch operations.

Today launch system planners are focusingdirectly on reducing the labor and attendant costsof launch operations. Historically, the chief means

of reducing operations costs, relative to achievedlift capacity, is to increase vehicle performance.Over the years, NASA, the Air Force, and thelaunch vehicle manufacturers have made incre-mental improvements to launch system perform-ance and reliability that have also led to opera-tions cost savings. For example, in its early flightsin the 1960s, the Delta was able to launch only100 lbs. to geosynchronous transfer orbit (GTO).Today, the Delta can launch over 2,800 lbs. toGTO. Launch operations costs’ are now about 10

5The Space Shuttle presents a counter example. Because of thedesire to improve the safety of Shuttle crews and payloads, the pay-load capacity of the Shuttle has actually decreased over the years.Originally designed to carry about 65,000 lbs. to low Earth orbit(at 160 nautical miles), the Shuttle’s payload capacity is now onlyabout 48,000 lbs.

bThese costs include only contractor personnel and other recur-ring costs directly attributed to the launch. They do not include main-tenance and other general costs associated with the launch pad.

Figure 2“3.— Delta 180 Project Schedule Reductionsr

Contract startto launch

55% reduct ion

and fabri~ation*-

On standto launch* ●

20°\0 reduction

I 1 I, I 1 I Io 5 10 15 20 25 30 35 40

Months

Typical Delta schedule

● Represents design and fabrication of the PAS—essentiaily a new third stage.● ‘Includes the PAS which doubles requirements of a normal Delta launch.

SOURCE National Aeronautics and Space Admlnlstratlon

16

Table 2=1.—Cost Reduction Factors for theDelta 180 Project

● Given program autonomy—minimum programmanagement and reporting

● Short statements of work—2 pages● Organized in terms of working groups responsible for

specific tasks—given autonomy to solve problemswithin working groups

● Within working groups, contractors worked as anintegrated team from the beginning—close contactamong all team members, open discussion

SOURCE: Department of Defense Strategic Defense Initiative Office/KineticEnergy Office, “Delta 1S0 Final Report,” vol. 5, March 1987.

percent of the total costs per flight. Performanceimprovements to the Delta7 (designated a DeltaII) should increase its lift capacity to 4,000 lbs.to GTO, but are not expected to alter significantlythe complexity or the cost of ground operations,though the cost of the vehicle has certainly in-creased. s Hence, should the per flight costs di-rectly attributable to operations remain constant,operations costs of the Delta II per pound9 coulddecrease by about 40 percent compared to the cur-rent Delta launcher.10 11 Historically it has takenabout 150 resident McDonnell Douglas personnelat Cape Canaveral to perform the launch vehicleprocessing activity at a 6-launch-per-year rate.This includes all administrative functions, groundsupport equipment operation and sustaining,cedure preparations, payload integrationlaunch vehicle processing through launch.

pro-and

7McDonnell Douglas is making these improvements in connec-tion with the Air Force MLV program. The first Delta II launch isexpected in 1989.

‘Lyle J. Holloway, McDonnell Douglas Astronautics Co., 1987.9This example illustrates one kind of savings possible as vehicles

are improved. However, for many purposes, figuri, ~ costs of launch-ing payloads on a per pound basis may not be appropriate. Thelife-cycle cost of a launch system for a given collection of payloadsover the years is often a more appropriate measure. See U.S. Con-gress, Office of Technology Assessment, Launch Options for theFuture: A Buyer’s Guide, OTA-ISC-383 (Washington, DC: U.S.Government Printing Office, July 1988).

*“These performance improvements will be accomplished by im-proved solid rocket booster engines and an improved main engine.

“Concurrently, the Delta launch crew efficiency has also im-proved, resulting in a higher percentage of launch successes, andthe potential for a higher launch rate (box 2-B). Delta has improvedits launch success rate over the years from 93 percent (170 out of182 launches) in the 1960s to nearly 98 percent in recent years (onefailure in 48 launches since 1977).

New Launch Systems

Several recent studies’z have suggested thatstarting fresh and designing to cost rather thanfor performance would lead to significant reduc-tions in the costs of launch operations. Thesestudies identified several approaches to system de-sign. The OTA workshop generated its own listof design goals (table 2-2). The discussion in chap-ter 4 elaborates on these goals, and lists a num-ber of technologies that would serve them.

NASA and the Air Force are working on a va-riety of new launch system designs. In particu-lar, they are collaborating on a major study ofan Advanced Launch System (ALS), whose goalsare to increase the payload capacity per launchby a factor of 3 or 4 and to reduce the cost perpound of launching payloads to space by an or-der of magnitude.13 Although a “clean sheet ofpaper” approach to launch system design offerspotential benefits in reducing life-cycle costs, italso increases the technical and cost risk of launchsystem manufacturing and operations. In addi-tion, the non-recurring investment in new facil-ities, and research and development, will offsetpart of the savings in recurring costs anticipatedfrom such changes.14 Thus it is necessary to ad-dress the entire set of launch procedures, includ-ing aircraft, trains, barges, and other auxiliaryfacilities, which function as a single integratedsystem.

ISSUE B: Is the United States Devoting AdequateAttention To Reducing the Costs of Space Trans-portation Operations?

Both NASA and the Air Force are funding re-search on new technologies for launch systems.Yet only a small percentage of this research isdevoted to development of technologies for spacetransportation operations and only part of thisis directed toward improving existing operations.

“U.S. Government, National Space Transportation and SupportStudy 1995-2010, Summary Report of the Joint Steering Group, De-partment of Defense and National Aeronautics and Space Admin-istration, May 1986; Advance Launch System Phase I Study brief-ings, 1987, 1988.

13 See U.S. Congress, Office of Technology Assessment, LaunchOptions for the Future: A Buyer’s Guide, OTA-ISC-383 (Washing-ton, DC: U.S. Government Printing Office, July 1988), ch. 5.

141 bid., ch. 7.

17

Box 2-B.—The Delta Experience

The following illustrates one company’s experience in providing launch services to the Government.● Minimal oversight. Part of the key to lowering launch operations costs is to keep the number of Govern-

ment personnel devoted to overseeing contractor preparations as small as possible. Responsibility formanagement of the Delta program has recently shifted to DoD. When under the management of NASA,McDonnell Douglas’ main customer for Delta launches was the NASA Goddard Space Flight Center(GSFC), whose primary mission is the preparation and launch of NASA’s scientific payloads. GSFC em-ployed 15 to 20 engineers to oversee the Delta launch operations. The GSFC team was kept deliberatelysmall, to avoid the temptation to over-manage McDonnell Douglas’ launch preparations. McDonnellDouglas attempted to discuss launch problems and resolve them with GSFC immediately. GSFC person-nel worked with the contractor’s internal documentation, and if a Government or military specificationor procedure showed greater risk in cost than it was likely to return in increased reliability it was dis-carded or tailored. Documentation requirements were kept to a minimum.

Self-sufficiency. McDonnell Douglas has minimized the number of associate contractors or subcontra-ctors with their own independent documentation procedures and systems necessary to work on the vehi-cle or facility. In addition, the Delta team prepares the vehicle on the basis of a single Launch Prepara-tion Document, which includes inputs from all departments. It gives all requirements for assembly andtest of the vehicle, traceability and accountability of all flight and non-flight hardware, and of all testand operational requirements. Daily meetings near launch time with all the technicians, inspectors, testengineers, managers, and the customer for the launch, enables significant problems to surface. This re-sults in a single, informed team with a common objective.

Mindset toward economy. Although the Delta has always been operated on a budget typical of smallscientific or commercial payloads, in the late 1970s McDonnell Douglas began to explore new ways toeconomize on the Delta when it became apparent that Government use of all ELVS was to be phasedout after the Space Shuttle became operational. McDonnell Douglas funded (with RCA) the develop-ment of upgraded Castor IV strap-on solid rockets, which increased Delta payload capacity 50 percent,and also found ways to economize on launch operations procedures. Although each individual step hasbeen small, over time, such steps have made the entire set of procedures more cost effective.

SOURCE: McDonnell Douglas Astronautics Corporation.

Table 2-2.—Approaches to Low”Cost Launch Design prove some launch procedures and might even

● Include all segments of the launch team (includingmanagers) in the design of a new launch system

● Reduce launch system complexity● Increase maintainability

—Increase subsystem accessibility—Design for modularity—Include autonomous, high-reliability flight control

and guidance systems—Build in testing procedures, for mechanical and fluid

systems, as well as for electronic systems● Make payloads independent of launcher, with stan-

dardized interfacesSOURCE: Office of Technology Assessment, 19SS.

Through its Office of Aeronautics and Space Tech-nology, NASA has funded a Civil Space Tech-nology Initiative (CSTI), which is pursuing re-search on a number of technologies, includingautonomous systems and robotics~that could im~

lead to cost savings (table 2-3).

As part of the CSTI, all the NASA centers areinvolved to some extent in the Systems Auton-omy Technology Program, which has been de-signed to develop and demonstrate the feasibil-ity of using “intelligent autonomous systems” inthe U.S. civilian space program, and to enhanceNASA’s in-house capabilities in designing and ap-plying autonomous systems. Some of these sys-tems, if successful, will have direct applicationsfor launch and mission operations. For example,the Systems Autonomy Technology Program isdeveloping an online expert system to assist flightcontrollers in monitoring and managing SpaceShuttle communications. It is also developing thehardware and software for autonomous diagnos-tics and control for the KSC Launch Processing

18

System. Both systems would increase the relia-bility and capability of mission and launch oper-ations and could eventually lead to reductions inthe number of personnel necessary for these tasks.

The CSTI is designed to demonstrate the feasi-bility of selected technologies. However, withouta clear and focused plan for choosing which tech-nologies are needed for launch and mission oper-ations, and inserting them into existing proce-dures, they may not be applied effectively. TheNASA Office of Space Flight is planning an Ad-vanced Operations Effectiveness Initiative, to be-gin in fiscal year 1989, that would provide plansfor inserting new technology into launch and mis-sion operations. Though funded at only $5 mil-lion per year, this initiative should play an im-portant role in improving operations procedures,because it can verify, validate, and demonstratetechnologies developed under the CSTI. In thelong run, it could also lead to lower operationscosts. Congress could consider funding this pro-gram at a higher level.

Through the Focused Technology Program,funded within the Advanced Launch System pro-gram, NASA and the Air Force are working to-gether on research crucial to reducing operationscosts. Some of these technologies may also con-tribute to improving the efficiencies of existingsystems (table 2-4).

The National Aeronautics and Space Act of1958 gave NASA the responsibility of “the pres-ervation of the role of the United States as a leaderin aeronautical and space scienceogy. 15 Its role as a research and

15 Natjona] Aeronautics and Space Act, Sec.

2451.

and technol-development

102(5), 24 U.S.C.

Table 2-3.—Civil Space Technology Initiative Funding(in miiiions)

Program area FY 88 FY 89 (requested)

● Automation and robotics ... .$25.1 $25.9Propulsion . . . . . . . . . . . . . . . . 23.8 46.7Vehicle (aeroassist flight

experiment). . . . . . . . . . . . . . 15.0 28.0‘Information technology . . . . . 16.5 17.1Large structures and control. 22.0 25.1Power . . . . . . . . . . . . . . . . . . . . 12.8 14.0

Total ... ... ... ... ... ... .$1 15.2 $156.8“Technologies of importance to launch and mission operations.

SOURCE: National Aeronautics and Space Administration.

(R&D) organization is firmly imbedded in its in-stitutional culture. The Air Force is mission ori-ented; its launch systems organization is thereforeorganized to respond to the special transportationneeds of the DoD payload community. Both orga-nizations have developed different institutionalcultures applying different operational approaches,which occasionally lead to costly friction in pro-grams of mutual interest. For example, in the areaof launch vehicle R&D development, the two or-ganizations continue to compete for funding andfor program lead. Yet, especially in this era ofbudget stringency, the Air Force and NASA mustwork together more effectively on research to im-prove existing systems and develop the next-gen-eration launch systems.

ISSUE C: What Factors Impede the Introductionof New Technologies and Management Strat-egies in Launch and Mission Operations?

Existing launch and mission operations are ex-tremely complicated, and have unique require-ments for technology, facilities, and management.For example, operations procedures may neces-sitate airplane runways; test facilities for a widevariety of equipment; massive, environmentallycontrolled buildings for launcher assembly andcheckout; and fixed and mobile launch pads. Lo-gistics, including the provisions of parts and sup-plies, contributes its own complexities. Eachfacility adds additional complexity and distinc-tive management requirements. In addition, theGovernment is both financially and institution-ally invested in existing operations procedures.The following factors make it difficult to reduceoperations costs significantly for existing launchsystems:

Investment in Existing Infrastructure.—TheUnited States has already invested billions of dol-lars in facilities at Kennedy Space Center (KSC),Johnson Space Center (JSC), Cape Canaveral, andVandenberg.

From a near-term budget perspective, it is eas-ier to justify refurbishing old facilities than tobuild totally new ones because the short term costsare often lower. However, existing facilities thatwere built for earlier launch programs require con-tinued modernization and repair, and the result-ing inefficiencies become part of the work flow

19

Table 2-4.—Advanced Launch System Focused Technology Program (in millions of 1988 dollars)

Year

1987188 1989 1990 1991 1992Propulsion:

Engine definition/demonstration . . . . . . . . . . . . $ 12.00 $ 6.00 $ 16.50 $30.10 $ 31.40LOX/LH2 engine . . . . . . . . . . . . . . . . . . . . . . . . . .LOX/LHc engine . . . . . . . . . . . . . . . . . . . . . . . . . .Propulsion subsystems . . . . . . . . . . . . . . . . . . . .Solid rocket booster . . . . . . . . . . . . . . . . . . . . . .Propulsion facilities . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Avionics/Software:● Adaptive guidance, navigation and control . . .● Multi-path redundant avionics . . . . . . . . . . . . . .● Expert systems . . . . . . . . . . . . . . . . . . . . . . . . . . .● Electromechanical actuators . . . . . . . . . . . . . . .Flight simulation lab . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Structures/Materiais:Cryogenic tank(s) . . . . . . . . . . . . . . . . . . . . . . . . .Booster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .NDE for SRB . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Structural certification . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Aerothermodynamics/Fiight mechanics:● Precision recovery . . . . . . . . . . . . . . . . . . . . . . . .Multi-body ascent CFD . . . . . . . . . . . . . . . . . . . .Aero data base . . . . . . . . . . . . . . . . . . . . . . . . . . .Base heating codes . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Groundandfiight operations/Manufacturing:● Ground operations . . . . . . . . . . . . . . . . . . . . . . . .● Health monitoring demo . . . . . . . . . . . . . . . . . . .● ManTech . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17.6032.90

0.507.00

24.0094.00

26.8016.30

1.4012.0034.0096.50

45.6032.904.50

15.0020.00

134.50

18.2028.60

5.6015.507.00

105.00

9.6011.504.00

17.502.00

76.00

6.1010.303.506.502.00

28.40

6.400.000.000.002.508.90

7.000.000,000.003.00

10.00

4.000.000.000.003.007.00

5.000.000.000.003.008.00

15.006.002.005.70

28.70

15.006.004,00

10.0035.00

19.008.004.003.00

34.00

12.0011.002.001.00

26.00

14.003.001.008.00

26.00

2.500.500.500.504.00

4.000,000.000.004.00

5.000.000.000.005.00

5.000.000.000.005.00

2.000.000.000.002.00

14.104.004.50

22.60

7.004.635.20

16.83

13.005.167.22

25.38

12.004.635.70

22.33

7.003.584.03

14.61

Grand total . . . . . . . . . . . . . . . . . . . . . . . . . . . $175.00 $154.93 $209.88 $173.33 $126.61“Technologiesof importance in launch and mwslon operations

SOURCE: U.S. Alr Force.

and extend throughout the life of theprogram.16

For example, because the Vehicle Assembly Build-ing, used for attaching the Shuttle orbiter to theexternal tank and solid rocket boosters, was origi-nally built for the Saturn 5 program, it does nothave the optimum size and shape for the Shuttle,which leads to longer and more complicated ve-hicle assembly. Thus, the long-term costs maybegreater than if a new, more appropriate, facilitywere built.17

On the other hand, any investments in new fa-cilities, such as a new launch complex, must alsobe weighed against the expected savings to begained over the expected life of the launch sys-tem. If the up-front costs are great enough, theycould outweigh the total operational costs for cur-rent systems, even if some reductions in opera-tions costs are achieved. However, because facil-ities become obsolete and equipment wears outover time, and must be replaced, opportunitieswill arise for making program changes in thecourse of replacing outdated facilities. Programchanges that require either major alterations, orreplacement of otherwise usable launch facilities,may lead to greater life-cycle costs. Because theyinvolve projects requiring considerable manpower,the construction and geographical placement of

1bNationa] Aeronautics and Space Administration, “ShuttleGround Operations Efficiencies/Technology Study, ” KSC ReportNAS1O-11344, Boeing Aerospace Operations Co., May 4, 1987, p. 4.

IT]n addition, many replacement parts required for certain Shut-tle test or training systems are no longer being manufactured andmust be custom built or refurbished by NASA.

20

new facilities may also face political constraintsthat affect life-cycle costs.

Old Systems That Need Upgrading.—Becausethe United States decided in the early 1980s tophase out ELVS and depend solely on the Shuttlefor launch services, needed improvements to theefficiencies of ELV launch fleets and facilities werenot made. Many of these improvements, includ-ing performance upgrades, lighter and more ca-pable avionics packages, and higher performance,safer solid rocket motors, are being made todayas part of the Air Force’s competitive ELV pur-chases.

Because certain parts of the Shuttle system arenow more than a decade old, they need to be up-graded as well. For example, both the Shuttle’sflight computers and the Shuttle processing sys-tem computers are being replaced. These changesare unlikely to lead to cheaper operations, thoughthey will increase the capability of the Shuttle sys-tem and may contribute to greater reliability.

Excessive Documentation, Oversight, and Paper-work.—As one workshop participant charged, “itis the Government’s excessive oversight and docu-mentation that have kept the cost of space launchmanagement and operations outrageously high. ”Both the Government and the contractor incurhigh costs from extra oversight personnel andfrom reporting requirements such as the Cost andSchedule Reporting System (C/SCSC). Althoughthis system can provide useful information for re-ducing costs, “it must be tailored to the programand its true cost to administer must be carefullyweighed against its advantages. ”18

Excessive Government oversight and reportingrequirements generally develop incrementally asa response to real problems of quality control, aconcern for safety, and the desire to complete highcost projects successfully. Over time these smallincrements of personnel or paper build to thepoint that they impede efficient operations, limitcontractor flexibility, and add unnecessary costs.

Chapter 4 discusses several technological op-tions for reducing the paperwork burden through

18C@ce sy~tems and @eratjons Cost Reduction and Cost Credi-bility Workshop: Executive Summary (Wasl@@v DC: NationalSecurity Industrial Association, January 1987), p. 2-5.

installation of automated systems. It also exam-ines the inefficiencies introduced by excessive over-sight of contractors during the launch process.

Uniqueness of Launch Pad and Other Facilities.—Current U.S. facilities are often unique to a givenlaunch system, and therefore different facilitiescannot be shared. It may be possible to design fu-ture launch pads to accommodate several differ-ent launch vehicles in order to save on facilitiescosts. For example, the Aerospace Corporationhas explored the potential of using a universallaunch complex, which would be designed witha universal launch stand, a universal mobilelaunch platform, and a modular assembly integra-tion building.

19 A modular integration building,in which a variety of vehicle designs can be as-sembled and integrated, is particularly important.However, such designs would represent a majorchange in the way the United States manages itslaunch operations and would require strong in-teraction between launch vehicle designers andfacilities planners. These changes in operationsprocedures would also mark a step toward estab-lishing launch operations that functioned morelike airline operations.

Lack of Sufficient Incentives for LoweringCosts.—The current institutional and manage-ment structure provides few incentives for reduc-ing costs of launching the Shuttle or ELVS forGovernment payloads. “The system does not haveincentives built in for achieving low-cost, success-ful launches,” observed one workshop participant.“There is the incentive not to fail, but not the in-centive to lower costs. ” Several participants notedthat NASA lacked the administrative structure fortracking funds and responsibilities by item to re-ward managers directly for reducing costs and in-creasing efficiency. Participants also pointed tothe fact that although it is possible to fashion in-centives for top-level management, it is difficultto make suitable incentives “transfer down to theguys who do the work” on the launch pad.

A recent study echoed these points and foundthat contractors generally have little incentive toreduce costs because “their profit/cash flow is

“U.S. Air Force, “Strategic Defense Initiative Launch Site Con-siderations,” Report No. TOR-0084A(5460-04 )-1 (Los Angeles, CA:Air Force Space Division, July 1985).

21

reduced when they perform under budget. ” Inaddition, the program officers “do not have anincentive to reduce spending below the programbudgeted amount.”2°

ISSUE D: What Impediments To Reducing Oper-ations Costs Are Unique to the Space Shuttle?

The complexity of Shuttle and payload proc-essing, and crew training, require substantial an-nual investment in personnel and facilities. Thefollowing points illustrate the most important im-pediments to reducing the costs of Shuttle launchand mission operations:

● Shuttle still in development. —Although NASAdeclared the Shuttle system operational af-ter the fourth flight, it has as yet not achievedtrue operational status.21 Because the Shut-tle is still undergoing major design changes,it requires a larger launch operations staffthan an “operational”22 system. For example,NASA employs about 5000 engineers atKSC, Marshall Space Flight Center, and JSCwho work on Shuttle systems. They havestrong incentives to implement changes forincreasing safety and performance, many ofwhich increase the time and cost of prepar-ing Shuttle for flight. On the other hand,there are few incentives for increasing oper-ations efficiency and reducing costs.

● Safety requirements. —Because the Shuttlecarries human crews, and because it is ahighly visible symbol of American techno-logical prowess, safety issues receive un-usually great attention. As a result of the in-vestigation of Shuttle subsystems followingthe loss of Challenger and its crew in Janu-ary, 1986, the Shuttle system is now under-going many major safety-related changes,23

which have led to considerable system re-

●

●

‘“Space Systems and Operations Cost Reduction and Cost Credi-bility Workshop: Executive Summary (Washington, DC: NationalSecurity Industrial Association, January 1987), p. 2-2.

Z] George E. Mue]]er, Panel discussion, Space Systems productivityand Manufacturing Conference W (El Segundo, CA: AerospaceCorp., August 1987), pp. 232-235.

“The term “operational” irnp]ies that the vehicle in question iscapable of being launched routinely on a well-defined schedule witha minimum of unplanned delays.

23 Ma jo r alterations include improvements to the SRBS, modifi-

cations of the SSMES, and installation of an escape hatch in theorbiter.

design. These changes have also increased thetime and complexity of launch operations.Prior to the loss of Challenger, NASA hadreduced the turnaround time necessary toprepare the Shuttle orbiter for flight to about55 workdays (three shifts a day) .24 NASA ex-pects orbiter turnaround for the first fewflights to equal about 150 workdays, decreas-ing to an average 75 workdays only after 4years of additional experience.25 However,judging from the experience in preparing Dis-covery for the first reflight of the Shuttlesince the Challenger explosion, this sort ofturnaround may be extremely difficult toachieve.Lack of spares; cannibalization of orbiterparts. —The Shuttle program has had a con-tinuing problem maintaining a sufficientstock of major spare parts and subsystems.For example, 45 out of 300 replacement partsneeded for Challenger on mission 51-L hadto be removed from Discovery.2b This hassignificantly impeded the ability of launchcrews to refurbish and test Shuttle orbitersbetween flights. Each time a part must betaken from one orbiter to substitute for adefective part in another, the amount of la-bor required more than doubles (table 2-5).In addition, the process increases the chancesof damaging either the part or the subsystemfrom which it is removed. Although NASAhas improved its stock of spares for the Shut-tle, the budget allocated for spares continuesto be a target for reductions. NASA runs acontinuing risk of having to cannibalize partsfrom one orbiter to process another.Complexity of the Shuttle systems.—TheShuttle was a revolutionary step in launchsystems, and was not designed for opera-tional simplicity. As with experimental air-craft, many of its systems are highly com-plex, and made up of a multitude of parts

Z4This does not include time the orbiter spends in the Vehicle As-sembly Building and on the launch pad.

Z5Charles R. Gunn, “Space Shuttle Operations Experience, ” pa-per presented at the 38th Congress of the International Astronauti-cal Federation, Oct. 10-17, 1987.

ZbReport of the presidential Commission on the Space ShuttleChallenger Accident (Washington, DC: U.S. Government PrintingOffice, 1986).

22

Table 2-5.—Steps in the Changeout of DefectiveParts in the Shuttle Orbiter When Replacement

Spares Are Unavailable

A part is needed for orbiter A. It is not in the parts inventory,but is available in orbiter B, which is not scheduled to fly forseveral months. The following steps are necessa~:

++

+++

Document steps of part% removal from orbiter B.Remove part from orbiter B. (It takes longer to

remove a part from an orbiter than to take it fromstorage.)

Document installation in orbiter A.install the part in orbiter A.Test part in orbiter A.Document installation of replacement part in orbiter B.instaii replacement part in orbiter B.Test replacement part in orbiter B.

+ = Addition to standard procedure.

SOURCE: Office of Technology Assessment, 198S.

that need to be inspected or repaired.27 Forexample, each of the solid rocket boosters(SRBS), one of the simpler Shuttle elements,contains about 75,000 parts and components.Of these, about 5,000 are removed, in-spected, and replaced or refurbished aftereach Shuttle flight. A design that required in-specting and handling of fewer parts wouldrequire fewer launch personnel. However,the costs of redesign, testing, and acceptanceof such a simplified design must be taken intoaccount.

The thermal protection system, composedof over 31,000 fragile tiles, requires carefulinspection and repair, an extremely labor in-tensive operation. Although only about sotiles now need replacing because of damageafter each flight, all of them must be in-spected. 28 Not only must they be inspectedfor damage, they must also be tested foradherence to the vehicle, and the gaps be-tween tiles carefully measured to assure suffi-cient space for thermal expansion upon reen-try. (See ch. 4, box 4-A for a description of

ZTGeorge E. Mueller, “Panel on Productivity Issues for Space andLaunch Systems, ” Space Systems Productivity and ManufacturingConference W (El Segundo, CA: The Aerospace Corporation, 1987),pp. 232-35.

zeonly a few ti]es are interchangeable; most are unique threedimensional shapes that are fitted to the curved surfaces of the or-biter. Charles R. Gunn, “Space Shuttle Operations Experience, ” pa-per presented at the 38th Congress of the International Astronauti-cal Federation, Oct. 10-17, 1987, p. 2.

a semi-automated system for inspecting andreplacing the TPS. )

Finally, the Shuttle orbiter has about250,000 electrical connections which must betested for continuity. Each time one of the8,000 connectors is disconnected or removed,there is a chance that one or more pins willbe damaged or will otherwise fail to recon-nect properly.

ISSUE E: What Can the Operational Experienceof the Airlines Contribute to Space Operations?

Although the technical and managerial con-straints on airlines operations are quite differentthan for launch vehicles, certain of their meth-ods used in logistics, maintenance, task schedul-ing, and other ground operations categories mayprovide a useful model for making launch oper-ations more efficient and cost-effective. Becauseof the extreme volatility of launch propellants anda relatively low launch rate, other airlines meth-ods may not be applicable to launch or missionoperations. Chapter 4 discusses the specific ap-plications of airline operations practices to spaceoperations. Many of these lessons are being ap-plied in the Advanced Launch System program(see

●

●

ch. 4). The airlines:

. . . begin cost containment program at plan-ning stage. 29 New aircraft design takes intoaccount operational requirements such assupport equipment, logistics flow, and facil-ity design, as well as payload characteristicsand route structure, in the early planningstages.. . . involve operations personnel in designchanges. As one workshop participant ob-served, “the chief objective of the airlines isto move a seat from A to B as quickly andefficiently as possible. Safety is a primarygoal, but increased efficiency is a basic re-quirement for making any design change.”Increases in efficiency must outweigh anyshortcomings brought about by incorporat-ing such a design change in the entire system.

“’’Space Systems and Operations Cost Reduction and Cost Credi-bility Workshop,” Executive Summary (Washington, DC: NationalSecurity Industrial Association, January 1987), pp. 2-18-2-19.

23

●

●

●

●

●

. . . have developed detailed cost estimatingrelationships for operations. When an aircraftmanufacturer suggests improved equipmentfor an aircraft subsystem, the airline can gen-erally estimate the recurring and non-recurring costs and any potential savings tobe gained. The airlines also have an exten-sive historical database to assist them in test-ing the accuracy of their own cost estimationmodels.. . . stand down to trace and repair failuresonly when the evidence points to a genericfailure. Generally the airlines continue fly-ing when one aircraft has crashed unless thereis clear initial evidence of a generic fault inthe aircraft model. For example, in the No-vember, 1987 crash of a DC9 in a Denversnowstorm, other DC9s continued to fly.However, in the 1979 crash of a DC1O in Chi-cago, the entire DC1O fleet was grounded be-cause there was early evidence that the wingmounting of one of the engines had failed,and safety officials were concerned thatgeneric structural faults might have causedthe failure.. . . insist on aircraft designed for fault tol-erance. Commercial aircraft are designed tobe robust enough to fly even when they haveknown faults. Airlines, with thousands offlights per day, have developed a minimum-equipment list —a list of vital operationsequipment that is absolutely mandatory forflight; if any of this equipment malfunctionson pre-flight check-out, the plane is groundeduntil the problem is fixed. The existence ofsuch a list means that an aircraft can fly withknown faults as long as they are not on theminimum equipment list.

. . design aircraft for maintainability. Com-mercial airliners are designed to be inspectedperiodically and to have certain parts andsubsystems pulled and inspected after a givennumber of hours of flight. The airlines callthis practice “reliability-centered maintenance.”. . . encourage competitive pricing. For themanufacturers of aircraft and aircraft sub-systems, the ability to purchase competitivesystems from several suppliers acts as an in-centive not only to reduce costs, but to im-prove reliability.

●

●

. . . maintain strong training programs. Inthe airlines, at all times some 10 percent ofthe operations crews are in training. Train-ing includes all aspects of the operations pro-cedures. Extensive training contributes toflight safety as well as to lowering costs.. . . use automatic built-in checkout of sub-systems between flights. Many aircraft sub-systems can be checked from the cockpit orfrom mobile ground units between flights,and in some cases, even in-flight with the aidof ground-based data links. Because they in-volve minimum operator interaction, theseprocedures tend to be more accurate thannon-automated systems.

ISSUE F: Does the United States Possess AdequateTechniques To Judge the Relative Benefit of Im-provements in Launch and Mission OperationsProcedures?

When debating the relative merits of either im-proving current launch systems or developing amajor new one, the principal question for Con-gress is: will the investment in a new system beworth it? In other words, will spending moremoney now yield greater savings later in the lifecycle of a system? Answering this straightforwardquestion is impossible without adequate modelsfor estimating costs. OTA workshop participantsgenerally agreed the United States has not devel-oped adequate cost estimating models for launchand mission operations procedures.

Workshop participants noted that estimatingcosts of new or improved systems requires datafrom existing systems. Commercial airlines use ex-tensive historical data to help them create accuratemodels for estimating costs, but launch operationsdo not have a comparable database to draw upon.In addition, NASA and the Air Force have madeno focused effort to collect such information. Forexample, there are no detailed historical recordstracing the number and cost of personnel fordifferent components of space operations. Onereason is that this information, where gathered,often rests with the contractors, who regard it asproprietary, In addition, for systems in develop-ment, the technologies tend to be more fluid, andtherefore operations data that could be collectedare poor predictors of the applicable costs for later

24

routine launches. A deeper reason is that the fund-ing required to gather and analyze such histori-cal data is often the first thing to be eliminatedfrom a program to save money —’’but what is be-ing eliminated is the future capability to learnfrom mistakes,” stated one workshop participant.Today we do not have sufficient data on whichto base a more meaningful cost estimation model.

STAS contractors focused some effort on costmodeling, but their work was hampered by in-adequate historical data. In addition, the accuracyof cost models developed by the contractorsawaits validation. The ALS Phase 1 study teamshave continued work on developing better costmodels.

What is often unclear in policy debates over thechoice of a new space transportation system is thatestimating the costs of ground operations is nec-essarily uncertain and partially subjective. Pro-gram managers often fail to calculate and presentto policymakers the uncertainties in estimatedcosts. New cost models will reduce the uncertaintyand subjectivity of cost estimation but not elimi-nate them. If such uncertainties and subjectivejudgments were made more explicit, it would bepossible to give policy makers a clearer sense ofthe economic risks of alternatives.30

Appendix A contains a brief summary of cur-rent cost estimation methods that illustrates theuncertainty and subjectivity involved. Methodsused in the Space Transportation ArchitectureStudy (STAS) are typical and are used as ex-amples.

ISSUE G: Are the Near-Term Launch Systems Un-der Study by NASA and the Air Force LikelyTo Generate Major Reductions of Launch Op-erations Costs?

The goal of the Advanced Launch System(ALS) program is to design a low cost, heavy liftlaunch system to serve U.S needs at the turn ofthe century. al Chapter 4 examines many of the

30T, Mu]]in, “Experts’ Estimation of Uncertain Quantities and ItsImplications for Knowledge Acquisition,” LFEE Transactions on Sys-tems, Man, and Cybernetics [to be published].

JIThe program’s original goal was to design a heavy-lift launchsystem that would serve U.S. needs at the turn of the century, re-sulting in a so-called “objective” system, with an “interim” systembased on existing technologies for the mid 1990s. However, con-

operations technologies under study in the ALSprogram. However, because an ALS would re-quire some technologies not fully developed at thistime, and because NASA and the Air Force wouldlike to be prepared to meet any additional demandfor launch services in the mid 1990s, they are alsoconsidering options for new, interim, high capac-ity launch systems based on current technology .32

The following paragraphs discuss the launchoperations requirements of two options, onebased on Shuttle technology, the other based ona variety of other technologies, and exploreswhether they would lead to reduced operationalcosts. OTA’s analysis of these proposed systemsleads to the general conclusion that although care-ful design, which took into account the opera-tional requirements of such systems, could indeedreduce operating costs, the investment cost of add-ing the necessary operations infrastructure and itsattendant recurring costs might offset such gains,especially if launch demand remains low. Policymakers should carefully scrutinize estimated life-cycle costs of any new system.

Shuttle-C (Cargo Vehicle)

NASA’s version of a heavy-lift launch vehicleis the Shuttle-C, which in several respects com-petes with Air Force heavy-lift concepts. TheShuttle-C 33 would be an unpiloted cargo vehiclebased primarily on Shuttle technology. It woulduse the solid rocket boosters, the external tank,and the main engines (SSMES) from the Shuttlesystem. A large cargo canister, capable of trans-porting some 85,000 to 100,000 Ibs. of payloadto low-Earth orbit, would take the place of thecurrent Shuttle orbiter.34 If Shuttle-C is used toship major subassemblies of the space station toorbit, one Shuttle-C flight would replace twoor three Shuttle missions. These could reduce

gressional resistance to funding a system capable of launching aspace-based ballistic missile defense system caused Congress to forbidexpenditures for studies of an interim ALS system.

Szsee U.S. Congress, Office of Technology Assessment, LauncbOptions for the Future: A Buyer’s Guide, OTA-ISC-383 (Washing-ton, DC: U.S. Government Printing Office, July 1988), ch. 4,

331bid.34sTAs a]so considered in-]ine Shuttle-derived vehicles, but these

are considered to require too much development to be consideredas a low-cost alternative at the present time.

25

the need to fly all of NASA’s planned shuttlemissions.

Shuttle-C would have the advantage of usingmuch of the same technology and many of thesame parts that have already proved successfulin 24 flights. To keep fixed costs down, it woulduse the same launch pads, vertical integration fa-cilities, and launch support crews now used forthe Shuttle. It carries the disadvantage that a standdown of the Shuttle might well result in delayingShuttle-C flights for the same reasons.

However, the following considerations wouldaffect the costs of launch operations for theShuttle-C:

● Not having to process a Shuttle orbiter wouldlikely speed up launch operations and there-fore reduce total operations costs comparedto the Shuttle.35 However, it would still benecessary to assemble, integrate, and test theShuttle-C before flight. In addition, someShuttle-C designs call for employing a reus-able engine/avionics package that wouldneed to be refurbished after each flight. If theShuttle-C were not specifically designed forsimplicity and ease of operation, its opera-tional costs could grow to become a signifi-cant fraction of the cost of preparing theShuttle orbiter.

● Shuttle-C will affect the processing flow ofthe Shuttle orbiter. Because the Shuttle-Cwould be the same overall size and configu-ration as the Shuttle, it could be processedin the same facilities as” the Shuttle, and in-serted into the Shuttle flow, if the Shuttlelaunch schedule permits. This raises severalcost-related issues.

First, because NASA intends to use Shut-tle-C for transporting major components ofthe space station to orbit, which are likelyto be of substantially greater value than thevehicle, NASA would have considerable in-centive to process the Shuttle-C as carefullyas it processes the Shuttle, and with the samecrews and procedures.3b

●

JsBecause it takes ]onger than any other single ground operationsprocedure, processing of the Shuttle orbiter effectively sets the Shuttlelaunch rate.

3’It might even cost more to develop independent launch proc-esses for Shuttle-C because NASA would then have to train addi-tional launch crews.

Second, the Shuttle facilities themselves,including the launch pads, may constrainNASA’s ability to reach the 12-14 launchesper year projected for a 4-orbiter fleet, andsimultaneously launch Shuttle-C two or threetimes a year. Launching both vehicles re-quires either shifting some payloads, such asspace science missions, to the Shuttle-C andflying fewer orbiter missions, or building newfacilities to accommodate Shuttle-C. New orupgraded facilities might include increasedengine shop facilities, an engine/avionicsprocessing facility, and a mobile launch plat-form. Any necessary new facilities should betaken into account when costing the Shuttle-C system.37

Inserting Shuttle-C into the Shuttle proc-essing flow could actually increase costs forlaunching the orbiter because of the risk ofslipping Shuttle-C schedules. Experience hasshown that nonstandard tasks, such as engi-neering modifications or special instrumen-tation, imposed on the Shuttle processingflow can contribute as much as 50 percentto an orbiter’s turnaround time.38 Because,except for the orbiter, Shuttle-C hardwarewould be quite similar to Shuttle hardware,we should expect Shuttle-C to experiencesimilar delays for several years after beingintroduced into NASA’s fleet. Modificationsto non-orbiter Shuttle subsystems wouldhave to be made on the Shuttle-C as well.

Finally, because Shuttle-C would sharemany parts with the Shuttle, delays may oc-cur in one or the other launch system shouldthe parts supply system become choked. Onthe other hand, because parts become cheap-er when purchased in quantity, the existenceof a Shuttle-C might reduce the costs of someparts.Mission Operations would be simpler andtherefore less costly than for the Shuttle or-biter. Because the Shuttle-C would not carryhumans, mission operations would consist

JTRecent NASA estimates for the cost of additional facilities atKSC range from $60 to $3oO million depending on the number ofprojected Shuttle and Shuttle-C flights.

38 Charles R. Gunn, “Space Shuttle Operations Experience,” pa-per presented at the 38th Congress of the International Astronauti-cal Federation, Oct. 10-17, 1987.

84-755 - 88 - 3 : (JL 3

26

●

Air

primarily of control, navigation, and guid-ance, and releasing the payloads at the propertime and in the proper orbit.Payload manifesting. For non-monolithicpayloads, the costs of payload manifestingand integration are likely to be comparableto the Shuttle. NASA’s experience on the or-biter has shown that payloads may interactin unforeseen ways, and require considera-bly more testing. If payloads were requiredto be self-contained, as suggested for the Ad-vanced Launch System, it might be possibleto reduce some of these costs. However, be-cause the payloads would be required toprovide services normally provided by thelauncher, such payloads would likely weighmuch more. (See ch. 4 for a discussion of thispoint. )

Force Near-Term Launch Systems

The Air Force is considering building larger ca-pacity vehicles to carry spacecraft designed forongoing Air Force programs, which are slowly butsteadily growing in payload mass and size as theygrow more capable. Modifications to existing ve-hicles to increase payload capacity offer no greatoperations savings .39

A new high capacity interim launch vehicle spe-cifically designed for rapid ground operationsmight reduce launch operations costs sufficientlyto pay for the necessary R&D. However, thiswould also require high demand for launch serv-ices Q

40 If high demand for cargo launch failed tomaterialize during 1992-1997, it would not be pru-dent to invest in a high capacity launch systemdesigned to respond to high launch rate.41’ Devel-oping a new system, even from existing technol-ogy, also incurs substantial risk that the new ve-hicles could not be processed for launch as quicklyas planned, whether for technical or “cultural”

3Whe Aerospace Corp., Space Transportation Development Direc-torate, Air Force-Focused Space Transportation Architecture Study(El Segundo, CA: The Aerospace Corp., Report TOR-O086A(2460-01)-2, August 1987).

dOThe only program currently under consideration that could re-

quire this sort of payload demand is the deployment of space-basedballistic missile defenses (SDI).

41see u. S. Congress, Office of Technology Assessment SpecialReport, Launch @tions for the Futm: A Buyer’s Guide, OTA-ISC-383 (Washington, DC: U.S. Government Printing Office, July 1988).

reasons. In fact, OTA workshop participants wereextremely skeptical that launch processing timescould be dramatically accelerated in the near term.They argued that only if radical changes weremade in the methods of launching space vehiclesand the institutional culture surrounding launchprocessing could costs be brought down signifi-cantly.

ISSUE H: How Do Concerns for Launch SystemReliability Affect Launch Operations?

The reliability of a launch vehicle (see app. C)has always been of concern to payload managers,because the cost of a payload may amount to twoto eight times the cost of launch services. Al-though some commercial communications satel-lites are relatively inexpensive compared to re-search or national security spacecraft, companiesmay stand to lose potential revenue amountingto hundreds of millions of dollars if their space-craft fail to reach orbit. Payload owners regardthe incremental costs of additional procedures toincrease launcher reliability, even for unpilotedvehicles, as worth the cost. The built-in conflictbetween the desire to reduce launch operationscosts and the desire to increase the success oflaunching spacecraft typically results in increasedattention to detail and a consequent increase incosts.

For the high visibility Shuttle, national prestigeand leadership as well as safety and reliability areat stake. Losing an orbiter, or even encounteringnon-catastrophic failures, are blows to nationalprestige. Nevertheless, in the view of some launchmanagers, launch operations currently assumemore than their share of the cost burden for im-provements in the reliability of an operational sys-tem. Launch operations tend to be complex andtime consuming because vehicles have been de-signed to achieve high performance rather thanrapid, inexpensive launch turnaround, and be-cause launch managers perceive they can improvethe chances of launch success by repeatedly test-ing every possible subsystem before launch. Theirconfidence in the reliability of a launch systemis generally lower than the calculated engineer-ing estimates .42

4zIbid., p. 85.

27

Some experts argue that it may not be possibleto lower overall launch costs (including the vehi-cle, payload, and other subsystems) significantlywithout increasing system reliability because thecosts of losing launch vehicles and payloads aretoo high. A reusable system, especially, dependsupon successful recovery and easy refurbishmentof expensive components. On the other hand, theexperience with recent improvements to the Shut-tle system demonstrates how expensive it can beto improve a launch system’s reliability. Althoughexperts disagree about the estimated reliability ofthe Shuttle with these improvements, they doagree that instituting and carrying out a test pro-gram capable of substantially improving confi-dence in the reliability of the Shuttle is likely tobe costly. Figure 2-4 illustrates the expected rapidrise in costs as engineers attempt to design vehi-cles for reliabilities above the 99 percent level.

Because of Titan and Shuttle launch failures in1985 and 1986, the time it now takes to assemble,integrate, check out and launch these vehicles ismuch greater than before the failures. Increasedemphasis on safety and subsystem testing andquality control to catch potential failures contrib-ute most of such increases. Managers now knowmuch more about the vehicles, though they canseldom point to a specific example of a fault that

would likely have led to a launch failure unlessrepaired .43

430TA staff were told at a briefing at Vandenberg Air Force Basein July 1987 that the non-destructive testing of Titan solid rocketmotors instituted after the 1986 Titan launch failure had revealeda number of imperfections in the propellant of solid rocket motorssegments, which had been stored for some time. These were re-worked to eliminate such imperfections. However, it is not at allclear that they would have caused a launch failure if they had goneundiscovered.

Figure 2-4.—Cost of AchievingExtremely High Reliability

80l I

0

OperationsProductionFacilitiesDesign, development,testing and engineering i

0.98 0.99 .9999

Based on: Reliability

. Expendable vehicle, 20 years of flight, 1 launch site

. 5M lb to LEO, 125K capability

. 40 flights per year, 100% load factor

SOURCE: Mmlin Marietta

Chapter 3

Operations Procedures

CONTENTS

Page

The Titan 34D and Titan IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Launch Operations—Cape Canaveral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Launch Operations—Vandenberg Air Force Base ..,... . . . . . . . . . . . . . . . . . . . . 39

The Space Shuttle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39Payload Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40Pre-launch Processing, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Mission Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Post-launch Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

BoxesBox Page

3-A. Titan 34D and Titan IV Launch Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333-B. The Space Shuttle Launch System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

FiguresFigure Page

3-I. Overview of Vandenberg Titan Launch Facilities . . . . . . . . . . . . . . . . . . . . . . . 313-2. Kennedy Space Center Payload Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323-3. Titan 34D Launcher . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343-4. Titan IV Launcher . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343-5. Geosynchronous Orbit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343-6. High Inclination Orbit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353-7. Typical Titan Receipt-to-Launch Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363-8. Typical Titan Launch Schedule, Cape Canaveral. . . . . . . . . . . . . . . . . . . . . . . 373-9. Titan Integrate/Transfer/Launch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3-10. Titan 34D Checkout and Launch Control Equipment . . . . . . . . . . . . . . . . . . . 383-11. Launch Complex 40 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393-12. Orbiter Ground Turnaround Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403-13. The Orbiter Challenger Rests on the Mobile Launcher at Pad 39A With

the Rotating Service Structure Containing the Payload Canister MovedBack From the ShuttIe (January 1983) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

3-14. Space Shuttle Payloads Processing Classifications . . . . . . . . . . . . . . . . . . . . . . 423-l5. Payload Processing Flows at the Kennedy Space Center . . . . . . . . . . . . . . . . . 433-16. Kennedy Space Center Space Shuttle System Ground Flow.... . . . . . . . . . . 443-17’. Orbiter Ground Turnaround Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453-18. Aerial Photograph—Important Facilities at Kennedy Space Center . . . . . . . 463-19. The Orbiter Columbia, Undergoing Launch Processing for Its Second

Voyage Into Space, Is Gently Lowered Toward the Solid Rocket Boostersand External Tank for Mating (August 1981) . . . . . . . . . . . . . . . . . . . . . . . . . . 47

3-20. The Space Shuttle Columbia Begins Its Roll Up the Ramp to Pad 39AAfter Completing the 3% Mile Journey From the Vehicle AssemblyBuilding (September 1982). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

3-21. The Space Shuttle Challenger Lifts Off in the First Nighttime Launch ofthe Shuttle Era (August 1983). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

TableTable Page

3-1. Shuttle Mission Operations Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

Chapter 3

Operations Procedures

Existing operations, which are enormously var- ment techniques (fig. 3-1; 3-2). Mission operationsied and complex, offer important lessons for the includes all activities associated with planning anddesign of future vehicles. Launch operations in- executing a mission —flight planning, documenta-cludes all the resources required to maintain and tion, mission support, and training. This chap-launch space vehicles, including unique facilities, ter summarizes operations procedures and sched-specialized equipment, computer systems, sched- ules as they are currently followed for the Titanu-ling, documentation, personnel, - and manage- and the Shuttle. -

Figure 3-1. —Overview of Vandenberg Titan Launch Facilities

Oxidizer farm

SOURCE: Aerospace Corp.

THE TITAN 340 AND TITAN IV

The Titan 34D (fig. 3-3; box 3-A), an expend- several models of the Titan III launcher the Airable launch vehicle which evolved from the Ti- Force developed to launch a variety of militarytan II intercontinental ballistic missile,l is one of and civilian spacecraft. In January 1988, the Air

Force took delivery of the first Titan IV launcher

1A crew-rated version of the Titan 11 served as the launch vehicle (fig. 3-4; box 3-A), which is larger and deliversfor the Gemini program of the early 1960s. a heavier payload to orbit than the Titan 34D.

31

32

Figure 3·2.-Kennedy Space Center Payload Facilities

LEGEND

• PAYLOAD PROCESSING FACILITIES

• HAZARDOUS SERVICING; PROCESSING FACILITIES

A. PA"lOAD INTEGRATION FACILITIES

SHUTTLE LANDING FAClllITY /'

/{7

* AIR FORCE FACILITIES

ORBITER PROCE

i

\1 <Am" S"N~/. MD , \

_ ~ ~M"WA' ~~~., ~'~.~' .~ ~ ~ =-====-- /-= .. .z.. COMPC" " •

"'" "lin ~DQS 'SC Kenned ,CC -' • ~~."",~&C "DG. Space ~ente'r / -~ PAD A

'" .,= ~..' //_ 41 ,

\JIll .SA" 2 ~~~. /-' -.' _.....,.-::::~;;..=-Q:..,~~ ~.q.ql( /--- J' ~.-3~-1~,. . ~ ... > •. ~~ ....... ~. ..* .. ~ .. ~M-.. AB.~. _ n ~ ... ,'}-l PHSF. :~ -." p - ---------~ ."" . " _ . ~ :.-- ""-' ~~... _.,3). CD~PC" " 5----~./:----p ~ ~ ~~~;;/' "'CJ~j

"G 5'D"G' 'wo. ~ t=....-~/ - 0 .:~~GA-~O ... 'f>~ESA60A. /\/1 .. sd1?: '-../ I - 0

r&D~- H

PAnDAD ~"-../ = V PRO"'''~~'i.~~ / g I

____ --~ .D'..7.:.<-.... :AS <A~CILI~Y ~ '0 0 j

__ ..

c~~~- PAYLOAD SPI.'~ ;..... ~. ""- 1lI <j)COMPLEX 3. d ~ & '~TYI~~:'. ."- ~J(!) I, ~ ~~P8 Canaver~I-~'~"' ~~ '. 0 :> .!- \ .<> I _ Ir Force St . .. .... ___ ~. SKID •. ' atlon~CaMPlEX17::- ... srRlP~ ~ ATLANTIC OCEAN

BANANA RIVER

SOURCE: National Aeronautics and Space Administration,

33

Box 3-A.—Titan 34D and Titan IV Launch Systems

The Titan 34D VehicleThe Titan 34D is an expendable launch vehicle developed by Martin Marietta Corporation under con-

tract with the Air Force. It is capable of launching up to 4,000 pounds of payload into geosynchronousorbit, or 30,000 pounds of payload into low earth orbit (LEO), and is now available commercially. * It con-sists of a two-stage liquid-fuel rocket core vehicle, which carries the payload, and two strap-on solid-fuelrocket motors.2

The solid-fuel rocket motors,3 which are 122 inches in diameter and 90.4 feet long, develop about 1.4million lb. thrust each. They are similar in construction to the solid-fuel rocket boosters for the Shuttle(though they rely on a different design for the seals between segments), and are built up of 5 individualsegments.

The Titan IV VehicleThe Titan IV expendable launch vehicle is the latest in the Titan family of launch vehicles. Developed

by Martin Marietta Corporation under contract to the Air Force, the Titan IV uses the same fuels and sub-systems as its sister vehicle, and provides the lift and cargo capability of the Space Shuttle. It is capableof launching 10,000 pounds of payload into geosynchronous orbit, or 39,000 pounds into low earth orbit,using the Centaur upper stage.

The overall design is similar to the Titan 34D but its core vehicle has been lengthened approximately20 feet to carry more liquid propellant. The solid rocket motors have been increased to seven segments.Recently, Martin Marietta selected Hercules Aerospace Co. as an additional supplier of solid rocket boostersfor the Titan IV.4 These boosters will provide a 25 percent increase in payload capacity, The Air Forceand Martin Marietta have developed three major versions of the Titan IV: a No Upper Stage (NUS), anInertial Upper Stage (IUS), and a Centaur Upper Stage (Centaur).

Generic Titan SystemsThe valves, ignition system, guidance system, and all the other systems aboard both launchers are

controlled either by avionics located in the second stage of the core vehicle, or by telemetry from the ground.Guidance signals for the avionics complement are generated in the upper stages. The launch complexesand associated facilities are similar for both vehicles; they differ primarily in size.

Typically, a Titan 34D launch costs about $90 to $110 million including launch services. Launch serv-ices account for 12 to 20 percent of the total cost, depending on the extent and complexity of services re-quired. A Titan IV launch costs $120 million, including launch services.

I In 1987, Martin Marietta formed a subsidiary corporation, Commercial Titan, Incorporated to market Titan launch services. Martin Mariettahas negotiated with the Air Force to use launch complex 40 at Cape Canaveral (for a fee) for commercial launches. Andrew Wilson, “Titan GrowsStronger, ” SPACE, vol. 3, pp. 8-11, 1987.

2Budt by Chemical Systems Division, United Technologies Corp. It supplies similar, larger solid rocket boosters for the Titan 4.The propellant in the core vehicle is hyperbolic, meaning that the fuel (Aerozine-50) and the oxidizer (nitrogen tetroxide) ignite on contact and

therefore need no ignition system; at the right moment during the countdown, computers at the launch control center direct fuel valves to open, allowing

the two fluids to mix. Turbopumps feed the liquid fuel engines, which are hydraulically gimbaled for steering.‘The propellant is a mixture of powdered aluminum, ammonium perchlorate, synthetic rubber, and various additives. The recent explosion at one

of the Nation’s two ammonium perchlorate plants has caused a severe shortage of the chemical for boosters.4See “Hercules Wins Contract for Titan 4 SRM Work, ” Aviation Week and Space Technology, Oct. 26, 1987, p. 31.

34

Figure 3-3.-Titan 340 Lsuncher

162’

——

SOURCE: Martin Marietta

I Core vehicle stage II

Solid rocket motorsd

Core vehicle stage I

- Stage engine

The Air Force maintains two Titan launch com-plexes at Cape Canaveral Air Force Station,Florida, and two at Vandenberg Air Force Base,California. As the result of their geographicallocation at 28° north latitude, the East Coastlaunch complexes are used for launching payloadstoward the East into low inclination and geosyn-chronous orbits (fig. 3-s). The West Coast launchcomplexes are used for launching payloads southinto high inclination orbits, such as sun-synchro-nous, polar orbits (fig. 3-6).

Figure 3-4. -Titsn IV Lsuncher

T / +

2a

I 1 , - - - 56’I I

Centaur mod G’ : :I 18 I upper stage ‘ :I 1 8

I I8“ : :t - - Ius upper : ;# staget --- 4

i i

stage II

Stage II/ e n g i n e ~. . . -

, .L*

+

I

SOURCE: Martin Marietfa

Figure 3-5.-Geosynchronous Orbit

174’

,—

SOURCE: Office of Technology Aaseaament, 19SS

35

Figure 3-6. - High Inclination Orbit in the Vertical Integration Building and Solid Mo-

N

/Orbit path a

s

SOURCE: Office of Technology Assessment, 1988

Launch Operations—Cape Canaveral2

Contractors’ assemble and prepare the launchvehicle and payloads, monitored by Air Force per-sonnel (figs. 3-7 and 3-8). At Cape Canaveral, thelauncher is prepared and launched following amodified integrate-transfer-launch procedure(ITL),4 in which the launch vehicle is assembled

‘Much of the data for this section was supplied by the “Space-craft Users Guide, ” Martin Marietta, April 1982.

3Martin Marietta, which builds the Titan launch vehicle, is theprincipal contractor. Other major contractors involved in the launchprocess include the Chemical Systems Division of United Technol-ogies (solid rocket motors), McDonnell Douglas (payload fairing),Boeing (Inertial Upper Stage), and General Dynamics (Centaur UpperStage).

4The Cape Canaveral launch complex was originally designed tosupport up to 60 launches per year. However, it has never reachedthat productivity goal, in part because launch procedures have grownmore complicated and require a greater number of facilities thanare available. In addition, high demand for launches never materi-alized. See chs. 4 and 5 for additional details.

tor Assembly Building, and wheeled out to thelaunch complex on a transporter (fig. 3-9). Otherprocedures, such as stacking of the solid rocketmotors, can be carried out in parallel while thecore vehicle is assembled and tested. This tech-nique minimizes the amount of time the vehiclemust remain on the launch pad.

At Cape Canaveral, the ITL concept makes itpossible to sustain about four Titan launches peryear from each of two pads (pad 40 and 41). Theaddition of some new facilities and upgrades ofexisting facilities would enable six to eightlaunches from each pad. Currently, the Air Forceplans to launch approximately six vehicles peryear from the East Coast. Martin Marietta expectsto launch an additional three commercial vehi-cles per year from pad 40 at Cape Canaveral.

The average time from receipt of launcher com-ponents and payload at Cape Canaveral to launchis 5 to 9 months, depending on the complexityof the vehicle desired.5 The Air Force expects eachTitan IV to spend 8 to 9 1/2 weeks on the launchpad, and each commercial Titan 34D about 6weeks.

Launch operations begin with the arrival of theparts of a launch vehicle, by rail, truck, andaircraft.

Solid Rocket Motors

Solid-fuel rocket motor segments and othersolid rocket booster parts arrive at the launch cen-ter by rail. After inspecting the parts visually,technicians inspect the propellant-filled motor seg-ments for internal manufacturing flaws in theNon-destructive Test Facility.’ These items are

5Assembly and checkout of the Titan IV is likely to take severalweeks longer than the Titan 111 because Martin Marietta is doingmost of the initial assembly at Cape Canaveral rather than the fac-tory in Denver.

bAfter a Titan 34D exploded 8.5 seconds into flight on April 18,1986, the Air Force launched a detailed investigation into the causeof the failure. After discovering that the failure was likely the re-sult of debonding of the insulation within the Titan solid rocket mo-tor (see “Titan Solid Booster Failure Caused Vandenberg Accident, ”Aviation Week and Space Technology, May 5, 1986, pp. 24-5), theAir Force developed an extensive non-destructive inspection pro-gram to qualify the booster segments. It also included additionalsensors to monitor launcher systems. See “Titan Mission SuccessBased on Tighter Heavy Booster Standards, ” Aviation Week andSpace Technology, Nov. 2, 1987, pp. 25-6.

36

IT-

— — — . — — — J

37

Figure 3-8.–Typlcai Titan Launch Scheduie,Cape Canaverai

Vehicle integration

Phase 1 definitionphase

Phase II –implementationperiod

Anafysis (loads, fttmechanics, etc.)

Softwaredevelopment

Targeting cycleManufacturing– Material

procurement– Detail &

subassembly– Major weid/

final assemblyLV pack and shipVIB & SMAB

operationsLaunch complex

operationsPayload fainng

procurementPayload fainng

preparationInitial launch;apability

Payload integration

Phase I definitionphase

Phase II –implementation

periodAnalysis (loads, flt

mechanics, etc.)Fit controlsManufacturing

(LV T34D)– Material

procurement– Detail &

sub-assembly– Major weld/

final assemblyLV pack and shipLaunch complex

operationsInitial launch

capability (ILC)

SOURCE: Marhn Marietta

!hS3 0 3 3 3 (

Monms3 6 9 12 15 18 21 24 27 3033 3(

Figure 3-9.—Titan integrate/Transfer/Launch

Launch

SOURCE. Martin Marietta.

transported to the Solid Motor Assembly Build-ing (SMAB), where they are stored. In the SMAB,the contractor erects the solid rocket boosters, be-ginning with the the aft subassembly. Electricalcabling and other supports are installed, as wellas emergency-destruct explosive charges.

Core Vehicle

Concurrently, the two stages of the core vehi-cle are inspected and assembled atop the launchvehicle transporter in the Vertical IntegrationBuilding (VIB). The VIB contains two cells for as-sembly and checkout and two cells capable ofstoring Titan core vehicles. Technicians connectand test the electrical umbilicals and staging con-nectors, check out the inertial guidance system,and install connections for the payload.

Launch Vehicle Assembly

The core vehicle, mounted on its transporter,is towed to the Solid Motor Assembly Building,where the solid rocket motors and other launchercomponents are positioned and attached. In theSMAB, technicians thoroughly check all sub-systems of the core vehicle and the solid rocketboosters—electrical, propulsion, flight control,hydraulic, guidance, airborne instrumentation,tracking, and flight safety subsystems (fig. 3-10).

Payload Integration and Processing

Payload integration begins months before thepayload or launch vehicle arrives at the launchfacility, often while the payload is in production.

38

Figure 3-10. —Titan 34D Checkout and LaunchControl Equipment

Launch control

Launch control center

+

A

— Control room.

— Instrumentation room ‘“~.

Launch complex ,~

SOURCE: Martin Marietta.

Payload managers discuss flight plan and payloadrequirements for vehicle services with the launchvehicle managers and determine the appropriatelaunch vehicle configuration. Several differentpayload fairings and upper stages are available,depending on the size, weight, and configurationof the payload.

When the payload arrives at the launch facil-ity, it is assembled, checked out, and tested in fa-cilities off the pad; it is installed on the vehicleat the launch complex several weeks beforelaunch. Payload technicians make all electricaland other connections and test for mechanical andelectrical integrity with the launch vehicle.

Launch Complex

The launch complexes 40 or 41,7 several thou-sand feet from the VIB, comprise many subsys-tems (fig. 3-11). Before the vehicle reaches thelaunch complex, contractor technicians also in-spect, test, and repair each subsystem in theappropriate launch complex—the concrete deckof the pad and its exhaust duct, water deluge sys-tem (to cool the exhaust and lower the pressurefrom hot exhaust gases), and vehicle air-condition-ing system. They fill the fuel and oxidizer stor-age tanks and erect support stands for the flex-

7Launch Complex 40 will be used for Titan 34D commercial andgovernment launches. Launch Complex 41 has been modified to ac-commodate the Titan IV.

ible hoses necessary for filling, venting, anddraining the rocket.

After the launcher is completely assembled inthe Solid Motor Assembly Building, the trans-porter tows it to one of the two launch complexes.The Mobile Service Tower rolls into positionaround the vehicle, and access platforms are low-ered into place. Payload specialists then install andtest the payload, which has been processed in par-allel with the vehicle. Launch vehicle specialistsalso carry out tests of the vehicle.

The actual time the vehicle must spend on thelaunch pad depends directly on the type and com-plexity of the upper stages and payload. With astandard payload, such as a geosynchronous com-munications satellite, which is encapsulated off-pad, average total pad time is about 6 weeks. Ifsignificant checkout and servicing must occur justbefore launch, as in the case of a payload requir-ing extensive fueling or the Interim Upper Stage,the payload may remain on the pad for as longas 11 weeks.

In preparation for the final countdown, all theordnance is installed, and the liquid oxidizer andfuel are loaded into the core vehicle. After bat-teries of tests, simulations, and verifications, therocket is declared ready for launch. The rangecontractor monitors and evaluates all pre-launchenvironmental conditions at the launch site, in-cluding ground winds, gusts, high-altitude winds,ceiling, cloud cover, and visibility.

A typical terminal launch countdown lasts lessthan one shift, and includes preparations and roll-back of the mobile service tower. Command-destruct loops are checked with range safety; thetelemetry system is checked. The mobile servicetower is rolled back. The core vehicle propellanttanks are pressurized to flight pressure, the pres-sures inside the solid-fuel rocket motor casings areverified as stable, and the guidance systemchecked. Final checkout, countdown, and launchare monitored and controlled from the controlcenter, located in the VIB.

Post Launch Activities

Launch processing is not complete with thelaunch. After liftoff, the launch control center isshut down and reset. The launch area is tested

39

Figure 3.11.— Launch Complex 40

AGE Building Fuel Holding PondFual Holding Area Film Camera

Fuel Vant Stack TO Ribbon Frame camera/ Refrigeration BuildingUmbilical Tower I

\Protective Clothing Building

\ \

w

CCTV Tower\ ; “~ Mobile Service Tower

.

‘————————

I

CCTV Tower - 1Fan House -f--

CCTVTower

Traffic Control House

Film Camem PadSecurity Fence

Oxidizer Vent Stack‘ To Ribbon Frame

Camera I Launch Pad

Oxidizer Holding Area

SOURCE: Martin Marietta

for toxic vapors and inspected to determine howmuch damage was sustained from the high-temp-erature rocket exhaust and vibration, and howmuch refurbishment the transporter and pad willneed before the next launch.

Whether the launch is successful or not, launchdata, both from the ground facilities and from thelauncher g are recorded and analyzed for anoma-lies or other indications of future problems. If thelaunch is unsuccessful for some reason, these datamay be the only means of determining the causeof failure.

aThe launch vehicle carries hundreds of sensors to record vari-ous environmental factors during launch, including temperature,vehicle acceleration, and vibration. This information is telemeteredto the launch complex for later analysis.

THE SPACE SHUTTLE

Unlike the Titan, the Shuttle orbiter was de-signed to be reused, which means that launchprocessing includes recovering and refurbishingthe orbiter and the solid rocket boosters (fig. 3-12). The current Shuttle (fig. 3-13; box 3-B) can

Launch Operations—VandenbergAir Force Base

As noted, vehicle assembly and processing dif-fer significantly at Vandenberg Air Force Base,primarily because the vehicle is fully assembledon the pad. The launch complex is therefore con-siderably simpler, as there is no Vehicle Integra-tion Building, Launcher Transporter, or Solid Mo-tor Assembly Building. Consequently the launchrate is lower, as an average of 163 days are re-quired for assembly and launch of each Titan 34Dwith its payload. This allows only about two andone-quarter launches per pad per year from Van-denberg.

lift up to 48,000 pounds into low-Earth orbit(about 160 nautical miles above Earth’s surface).

Shuttle launch operations requires the use ofa wide variety of both advanced and routine tech-

40

Figure 3-12. -Orbiter Ground Turnaround Operations

, Landing

I

‘ r

Orbiter processingOperations (OPF)a

~ ’Launch

IIIIIIII

Shuttle IIntegratedOperations

I

ET checkout (VAB) b I

OPS IH/SRB I

JVAB fl/SRB operations’

1

a ~rbtier Pmessing FaciliWbvehlcle Aasemb~ Buildingc Vehicle Assembly Building External Tank/Solid Rocket Booster Operations

SOURCE: National Aeronautics and Space Administration

nologies and interrelated subsystems. At KennedySpace Center (KSC), Shuttle processing is carriedout by contractors monitored by NASA employ-ees. Lockheed Space Operations Company cur-rently holds the prime contract for this work.Lockheed subcontracts with a variety of othercompanies, including Rockwell International, theprime contractor for the Shuttle orbiter. Approx-imately 6,550 contractors and 640 NASA employ-ees directly support Shuttle launch operations atKSC. Johnson Space Center has responsibility foron-orbit mission operations and some launchoperations, which involve about 5,675 contrac-

Figure 3-13.—The Orbiter Challenger Rests onthe Mobile Launcher at Pad 39A With the RotatingService Structure Containing the Payload Canister

Moved Back From the Shuttle (January 1983)

SOURCE: National Aeronautics and Space Administration.

tors and 1,065 NASA employees. Marshall SpaceFlight Center contributes engineering expertise forShuttle modifications and supports the Shuttlewith approximately 1,000 NASA employees and11,000 prime contractor employees.

Payload Processing

Payloads for the Space Shuttle can be installedeither horizontally or vertically (figs. 3-14 and 3-15). Horizontal payloads, such as Spacelab, areinstalled in the orbiter when it is still in the Or-biter Processing Facility, prior to being mated withthe external tank and solid-fuel rocket boosters.Vertical payloads are installed in the orbiter’s pay-load bay after the fully assembled Shuttle arrivesat the launch pad.

The payload owner assembles the payload toits final launch configuration by, for example, in-

41

Box 3-B.—The Space Shuttle Launch System

The shuttle launch vehicle consists of three ma-jor subsystems:

● the orbiter, with the crew compartment andpayload bay, which also contains the threeSpace Shuttle main engines (SSMES). Aboutthe size of a DC-9, the orbiter weighs about215,000 Ibs., without its payload, and hasa 15 by 60 foot cylindrical payload bay. I

● the extermd tank, which holds liquid hydro-gen and oxygen to fuel the SSMES.

● two segmented solid-fuel rocket boosters(SRBS). These are each made up of 5 motorsegments.

At launch, the main engines are ignited, fol-lowed seven seconds later by the SRBS. Two andone-half minutes into the flight, explosive boltsseparate the orbiter from the SRBS, which para-chute into the Atlantic Ocean and are recovered.After about eight minutes of flight, the Shuttlemain engines shut down and the external tankseparates from the orbiter, breaks up as itreenters the atmosphere and falls into the IndianOcean. In space, the orbiter maneuvering sys-tem, fueled by hyperbolic propellants, propellsthe craft into the orbit desired for the mission.

After the Shuttle crew completes its mission,the orbiter returns to earth where it can land onany of several runways,2 and is carried back toKennedy Space Center, to be refurbished for thenext launch. Although the Air Force built a Shut-tle processing center and launch pad at Vanden-berg Air Force Base, CA, it is now in the proc-ess of being mothballed.

‘Following the loss of Challenger, the fleet now consists of threeorbiters—Columbia, Discovery, and Atlantis. NASA has contractedwith Rockwell International to build a fourth orbiter, which will re-place Challenger by 1992.

‘For safety reasons, especially after the loss of Challenger, the Shut-tle WI1l normally land at Edwards Air Force Base, However, in anemergency, the Shuttle can touch down at Cape Kennedy, or oneof the alternative emergency landing sites.

stalling solar panels, antennas or other items, andperforming minor repairs. Payload owners havefive options for processing payloads, ranging fromminimum KSC involvement—essentially “shipand shoot, ” where the payload is ready for launchand can be installed in the orbiter 2 to 50 daysbefore launch with no servicing, to maximumKSC involvement—where all flight experimentsand component hardware must be delivered up

to a year before launch for technicians to assemble,integrate and test.9

Pre-launch Processing

The basic processing flow is organized accord-ing to the integrate—transfer—launch (ITL) con-cept, which separates the major processing ele-ments and allows certain functions to proceed inparallel until the vehicle is assembled in the Ve-hicle Assembly Building and carried to the launchpad (fig. 3-16).

Orbiter Processing

Orbiter processing constitutes the critical limitto the achievable flight rate. Refurbishing the or-biter Columbia for the second Shuttle flight con-sumed nearly 200 work days (three shifts per day).By the ill-fated flight of the Challenger in Janu-ary 1986, this turnaround time had been reducedto 55 days (fig. 3-17).10 However, the numerousmodifications to the process of orbiter refurbish-ment, made as a result of the Challenger accident,have led to an estimated future orbiter turnaroundof about 160 days, to decrease over four yearst. 75 days. 11 FOr a four orbiter fleet, a 75 day Or-

biter turnaround time would allow 12 to 14 flightsper year.

In the Orbiter Processing Facility, NASA con-tractors check and refurbish every major systemin the orbiter after each flight, including theavionics, brakes, electrical systems, and windows.They inspect the Shuttle main engines, and com-pletely replace the bearings and turbines. Any ofthe 31,000 ceramic tiles of the thermal protectionsystem that are missing or damaged during theflight are also replaced; all tiles are re-water-proofed. Modifications to the orbiters are madeduring refurbishment.lz Finally, any horizontal

‘For a detailed breakdown of these options, see James M. Ragusa,“Historical Data and Analysis for the First Five Years of KSC STSPayload Processing,” NASA, September 1986, ch. 4.

IOchar]es R. Gunn, “space Shuttle Operations Experience, ” Pa-

per IAF-87-216, delivered at the 38th Congress of the InternationalAstronautical Federation, Oct. 10-17, 1987.

llHoWever, some Shutt]e operations experts have raised concernsthat new safety related requirements for orbiter processing may makeit extremely difficult to reach the goal of 75 days turnaround within4 years.

IZThese modifications can significantly lengthen the time neces-sary to process the orbiter.

84-755 - 88 - 4 : (2L 3

42

- — — - — — — - — — - — — - - - - -

. — — - — — - — — — — — - - — - - -

43

I

L I

a1-U)z

avm

(suca

Figure 3·16.-Kennedy Space Center Space Shuttle System Ground Flow

SAflNG PA YLOAD OPERATIONS

.---------------------~ ORBITER CYCLE

~. ~

PAYLOADS INSTALLED ON PAD

~~ ~'\s...-c-._- - . ...:;;..~

~~~-~:-SRORETAIEVAL

SOURCE: National Aeronautics and Space Administration.

VEHICLE ASSEMBLY

~-----------------------

t~~-~ •. ::s:Z!e.......-.. ~ EXTERNAL TANK DELIVERY

SAO CYCLE SRO REFUAB'SHMENT

45

220

200

180

160

140

120

100

80

60

40

20

0

Figure 3-17.—Orbiter Ground Turnaround Experience

Shuttle fits: 1

d

.

2 3 4 5

1-

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

n PAD workdays

c 1VAB workdays

dOPF workdays

SOURCE: National Aeronautics and Space Administration.

payloads, such as Spacelab, are installed in the When the segments arrive at KSC, they are moved. -.payload bay. into the Rotational Processing and Surge Facility

where they are inspected and stored until needed.Solid Rocket Boosters (SRBS)

Assembling and testing the SRBS requires aboutThe contractor (Morton Thiokol) ships new 33 days (three shifts). After the aft booster

solid-fuel rocket motor segments and associated assemblies are attached to support posts on thehardware, including the forward and aft closures, mobile launcher platform, the rocket motor seg-nozzle assemblies, and nozzle extensions by rail. ments—all filled with live propellant—are added,

46

inspected and tested. During the stacking proc-ess, which lasts several hours per segment, every-one but the stacking crew must evacuate the Ve-hicle Assembly Building (fig. 3-18).

External Tank

Meanwhile, the external tank, which is manu-factured by Martin Marietta, is transported toKSC by sea barge from the Michoud AssemblyFacility at New Orleans, Louisiana. In the Vehi-cle Assembly Building (VAB), contractors inspectthe external insulation and comections for groundsupport equipment, as well as the cryogenic tanksfor holding liquid oxygen and liquid hydrogen.

A crane hoists the tank to a vertical position andtransfers it to the mobile launcher platform, whereit is mated with the twin SRBS. The electrical sys-tems are checked and the many fluid valves tested.

Vehicle Assembly and Integration

After orbiter processing is complete, it is towedto the VAB High Bay, lifted to the vertical, andmated to the external tank and SRBS (fig. 3-19).After mating all the sections of the Shuttle andconnecting all umbilicals, technicians test eachconnection electrically and mechanically.

The computer-controlled launch processing sys-tem, which is operated from the firing rooms of

SOURCE: National Aeronautics and Space Administration

47

Figure 3-19.—The Orbiter Columbia, UndergoingLaunch Processing for Its Second Voyage Into Space,Is Gently Lowered Toward the Solid Rocket Boosters

and External Tank for Mating (August 1981)

SOURCE: National Aeronautics and Space Administration.

the launch control center, semi-automatically con-trols and checks out much of the Shuttle vehicle,both in the VAB and at Launch Complex 39. Ifany subsystem is found unsatisfactory, the com-puter will provide data that help isolate the fault.

Transfer and Launch

When the Shuttle is fully assembled on the Mo-bile Launch Platform, a crawler-transporter slipsunder and slowly (1 mph) moves the Shuttle toLaunch Complex 39A or B (fig. 3-20). Once atthe pad, workers gain access to the Shuttlethrough the fixed service structure, which alsoprovides liquid hydrogen and oxygen to the ex-ternal tank. The rotating service structure givesaccess to service fuel cells and the life-support sys-tem, to load and remove payloads, and to loadhyperbolic fuels for the orbital maneuvering sys-tem and the reaction control system. Those pay-loads to be installed vertically are transported tothe rotating service structure in a protective pay-load canister.

After the Shuttle arrives at the pad, most check-out operations are controlled from the launch con-

trol center. After these operations, power is ap-plied to the orbiter and supporting ground supportequipment, launch-readiness tests are performed,and the tanks are prepared to receive their fuels.The Shuttle is now ready for the cryogenic propel-lants to be loaded and the flight crew to board.

During the final six or seven hours of the count-down, the mission software is updated and theliquid hydrogen and liquid oxygen loaded into theexternal tank. Finally, the flight crew and opera-tions personnel complete all preparations and theShuttle lifts into space (fig. 3-21).

Mission Operations

Mission operations (table 3-1) comprise allactivities associated with planning and executinga mission and the payloads to be carried. The pri-mary focus is on gathering data, performing anal-yses, and developing the software required tomeet the mission’s objectives. Mission operationsbegin the first day a payload is conceived, con-tinue through the day of the launch, and end onlyafter the mission is satisfactorily completed andthe data analyzed. From early mission planningto completion, mission operations for a Shuttleflight may take two years or more.

Tracking stations scattered around the worldand the Tracking, Data, and Relay Satellite Sys-tem (TDRSS) give orbiter crews access to MissionControl for most of the orbit .13 The Johnson SpaceCenter (JSC) at Houston, TX is the central con-trol point for Shuttle missions; payloads and re-lated systems are controlled from the Jet Propul-sion Laboratory (Pasadena, CA), the GoddardSpace Flight Center (Greenbelt, MD), or JSC.

Post= launch Processing

Two minutes into the Shuttle’s flight, the twosolid-fuel rocket boosters are jettisoned and par-achute into the Atlantic Ocean downrange fromKSC. Two specially designed retrieval vessels re-cover the boosters and their various components.The smaller components are hauled on board theships, whereas the boosters are towed back to the

IJwhen TDRSS deployment is complete (three spacecraft), it willallow the crew to contact Mission Control nearly 100 percent ofthe time in orbit. The second of two TDRSS satellites is scheduledfor deployment on the first Shuttle flight after standdown.

48

Figure 3-20.—The Space Shuttle Columbia Begins Its Roll Up the Ramp to Pad 39A After Completing the3.5 Mile Journey From the Vehicle Assembly Building (September 1982)

SOURCE: National Aeronautics and Space Admlnlstration.

49

Figure 3-21 .–The Space Shuttle Challenger Lifts Offin the First Nighttime Launch of the Shuttle Era

(August 1983)

SOURCE: National Aeronautics and Space Administration

Table 3-1.—Shuttle Mission Operations Functions

●

●

●

●

●

b—

Design Shuttle trajectory and flight planDevelop flight and ground documentation to supportoperationsProvide mission support with flight planning, flightsystems, payload support, and trajectory teamsProvide maintenance, operations, missionreconfiguration, and sustaining engineering for supportfacilities (mission control center and simulation andtraining facilities)Develop plans and provide training to crews,customers, and flight controllersDevelop operating concepts and requirements

SOURCE: Johnson Space Center

KSC Solid Rocket Booster Disassembly Facility.At this facility, the boosters and other componentsare washed, disassembled, cleaned, and strippedbefore they are shipped by rail to the prime con-tractor (Morton Thiokol) for refurbishing andreloading.

When the Shuttle orbiter returns from its mis-sion in space, it normally lands at Edwards AirForce Base in California. In emergencies, it canalso land at KSC; White Sands, NM; Zarogosa,Spain; Casablanca, Morocco; Rota, Spain; orGuam. After landing, it must be drained of haz-ardous fuels and inspected for any exterior dam-age. Payload technicians remove any payloadsbrought back to Earth. If it lands anywhere butKSC, the orbiter must be lifted onto the back ofa specially equipped Boeing 747 and flown backto the Shuttle Landing Facility at KSC.

Chapter 4

Technologies andManagement Strategies

CONTENTS

PageIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Technologies for Ground and Mission Operations

Automated Data Management System . . . . . . . .Automated Test and Inspection . . . . . . . . . . . . . .Computer-Aided Software Development . . . . . .Expert Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Technologies for Launch Vehicles . . . . . . . . . . . . . .Vehicle Design Principles. . . . . . . . . . . . . . . . . . . .Vehicle Technologies . . . . . . . . . . . . . . . . . . . . . . .

Management Strategies . . . . . . . . . . . . . . . . . . . . . . .Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Operations Management . . . . . . . . . . . . . . . . . . . .

Assessing Technological Options . . . . . . . . . . . . . . \.Economic Criteria . . . . . . . . . . . . . . . . . . . . . . . . . .Non-economic Criteria . . . . . . . . . . . . . . . . . . . . . .Economic Benefits. . . . . . . . . . . . . . . . . . . . . . . . . .

BoxesBox

4-A. Shuttle Tile Automation System . . . . . . . . . . .4-B. Expert Systems for Launch Operations . . . . .4-C. Vehicle-Payload Interfaces on the Advanced4-D. Economic Criteria . . . . . . . . . . . . . . . . . . . . . . .

FiguresFigure4-I. Interim Problem Report: LES, the Lox Expert4-2. Universal Launch Complex . . . . . . . . . . . . . . . .4-3. Economic Benefits of Emerging Technologies

TablesTable

4-1.4-2.4-3.4-4.4-5.4-6.

4-7.

4-8.

Technologies for Operations . . . . . . . . . . . . . . .ALS Launch Operations Specifications . . . . . .Launch Vehicle/Payload Interface Issues . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

Launch. . . . . . .

System. . . . . . .. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . .

system”. . . . . . .

53535354545559596264646566676869

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

Page

55566167

Page

. . . . . . . . . . . . . . . . . . . . 57

. . . . . . . . . . . . . . . . . . . . 65

. . . . . . . . . . . . . . . . . . . . 71

. . . . . ..!!.. . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

Page

Vehicle Technologfis To Facilitate Ground Operations . . . . . . . . . . . . . . . . . . .NASA’s Advanced Operations Effectiveness Initiative . . . . . . . . . . . . . . . . . . . .Non-economic Criteria for Judging Space TransportationSystem Performance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Estimated Economic Benefits of Developing Technologies To FacilitateGround Operations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Technology Development Benefits: AComparison of Estimates by TwoSTAS Contractors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5362626266

68

69

70

Chapter 4

Technologies and Management Strategies

INTRODUCTION

Launch and mission operations could be mademore efficient and less expensive by employingemerging technologies in the three major compo-nents of the launch system—ground support fa-cilities, mission control facilities, and launch ve-hicles. These technologies must be put to workin an institutional structure and culture that fa-cilitates, rather than hinders, their use. Therefore,efficient management strategies must also receiveconsideration.

The first section of this chapter, Technologiesfor Ground and Mission Operations, introducesoperations technologies that could be used in anadvanced launch system specifically designed forlow cost. They are consonant with technologiesfor the Advanced Launch System (ALS) currentlyunder consideration by the Air Force and NASA.Many of them would also be appropriate either

for enhancing existing launch systems or for in-clusion in new launch systems built with existingtechnologies.

The next section, T’echnologies for Launch Ve-hicles, introduces launch design principles and ex-plores technologies that could be inserted into ve-hicles to reduce the costs of launch and missionoperations. The section on Management Str-ate-gies examines some methods of organizing andmanaging launch systems to achieve low costoperations. Finally, Assessing Technological Op-tions and Costs discusses the principal trade-offsto consider in designing new facilities and a newlaunch operations strategy, and explores howthese concepts and techniques may affect thedesign, costs, and processing of vehicles andpayloads.

TECHNOLOGIES FOR GROUND AND MISSION OPERATIONS

Because of considerable overlap in the technol-ogies that could be employed in launch and mis-sion operations, this section discusses them to-gether. Some of these technologies exist in oneform or another today, but would need to bemodified for specific space applications; others re-quire additional research and development. Ta-ble 4-1 lists some major categories of technologiesor applications. Those marked with an asteriskare described and discussed briefly in the text.

Automated Data Management System

Computer work stations linked through a net-work that provides a common database can as-sist the speed and accuracy of information trans-fer and make it possible to speed up sign-offprocedures. Such automated data managementsystems are in common use in manufacturing andservice industries.

“One of the highest cost items, if you look atthe Shuttle program today, is the operations cost

Table 4-1.—Technologies for Operations

Automated data management system*Automated test & inspection*Automated launch vehicle and payload handlingDatabase management systemsComputer-aided software development ●

Ex~ert systems*● Discussed in text.

SOURCE: Office of Technology Assessment, 19S8

associated with all the data processing systems in-volved, ” observed one OTA workshop partici-pant. On-board systems, flight-design-and-prepa-ration, training, launch processing, and missioncontrol systems have all evolved over the years.They are complex, written in different computerlanguages, and sometimes poorly documented.Each uses different, unlinked databases. Partici-pants further explained that individual programelements have their own autonomous missionplanning jobs and their own manner of sendinginformation among the subsystems; people use“bulky paper, communications, phone calls, ” and

53

54

group meetings, and use no integrated approachto transferring the data to all elements, eventhough they are all interdependent. As a result,although the flow of information within NASAduring the launch sequence is excellent, during themonths leading up to launch, information flowis very poor. The events before the failure of theChallenger provide an unfortunate example ofhow constrained the flow of information can bein the months prior to launch.1

An automated data management system shouldbe incorporated into any future launch systems.Workshop participants urged planners of futuresystems to: standardize the architecture of on-board and ground systems, standardize the codeused, and minimize custom hardware and soft-ware by using commercially available productswhere possible. One participant estimated that anintegrated paperless information-management sys-tem could reduce the time spent in launch opera-tions by one-half. The space station project plansan integrated, paperless information system to as-sist in managing space station operations. Manyof the lessons learned in that effort could be ap-plied to launch and mission operations.

Automated Test and Inspection

Automating certain test and inspection proce-dures could also reduce costs. However, beforeautomating current procedures, they should becarefully examined to see which ones are neces-sary, and whether some steps can be simplifiedor even eliminated. “It makes no sense to auto-mate nonsense, ” asserted one workshop partici-pant. Certain kinds of automation such as the as-sembly and test of electrical and electronicssystems may be technically straightforward, butdifficult to incorporate because workers under-stand current procedures well and are reluctantto change. Workers require incentives and addi-tional training to smooth the transition to newprocedures.

Automating the assembly and test of mechan-ical, pneumatic, and fluid systems is a major chal-

IReport of the Presidential Commission on the Space Shuttle&d)engerAcudent (Washington, DC: U.S. Government PrintingOffice, 1986), ch. 5.

lenge. Today, mechanical and fluid systems cre-ate the most operations delays and verificationproblems, whereas electrical and electronic sys-tems are already well-instrumented and tend tobe reliable. For example, on the Delta launch ve-hicle, more time is spent in mating the strap-onsolid- fuel rocket motors to the liquid rocket, fit-ting the cork insulation, and doing the leak checkon the pneumatic hydraulic systems than in check-ing out the entire electronic system. On the Atlas,part of the leak check test calls for looking forbubbles or listening for leaks—something very dif-ficult to automate.

Box 4-A presents an example of a semi-auto-mated system for inspecting the thermal protec-tion tiles on the Shuttle orbiter. The system, de-veloped by NASA, Stanford University, andLockheed Space Operations Company, promisesto make tile inspection more reliable and maylower its cost.

Computer-Aided SoftwareDevelopment

Traditional methods of developing softwareand writing the necessary computer code arehighly labor-intensive and require skilled pro-grammers. However, new techniques promise toimprove the speed and accuracy of software de-velopment.

Computer-aided software development optionsrange in power and complexity from commer-cially-available, so-called software engineeringenvironments, 2 which are program libraries, edi-tors, and program debuggers, to automatic pro-gramming. 3

The benefits of using computer-aided softwaredevelopment include reliability, economy, andresponsiveness, a key aspect of operational flexi-bility. Proponents of computer-aided software de-velopment suggest that it will be applicable toboth mission control and ground operations.

‘For example, SmallTalk: Adele Goldberg, SznaMalk-80 (Read-ing, MA: Addison-Wesley, 1983).

‘Skeptics contend that “automatic programming has always beena euphemism for pro ramming with a higher-level language than

fwas presently availab e to the programmer. ’’-D. L. Parnas, “Soft-ware Aspects of Strategic Defense Systems, ” American Scientist,November 1985.

55

Box 4-A.—Shuttle Tile Automation System

Inspecting the some 31,000 thermal protectionsystem (TPS) tiles on the Shuttle orbiters andrepairing damaged ones is highly labor intensive.Automating the inspection procedures could re-duce overall labor costs, and increase inspectionspeed and accuracy. In 1986 NASA began theSpace Systems Integration and Operations Re-search Applications (SIORA) Program as a co-operative applications research venture amongNASA-KSC, Stanford University, and LockheedSpace Operations Company. One of its initialtasks is to apply automation and robotics tech-nology to all aspects of the Shuttle tile process-ing and inspection system.

The team is developing an automated workauthorization document system (AWADS) thatwill enable technicians to document the condi-tion of each tile, determine any necessary repairsor replacement, and generate work instructions.With the automated system, the computer,which is programmed to recognize each techni-cian’s voice, prompts the technician to find thecorrect tile, enter its number, and report on itscondition in a systematic way. The TPS qualitycontrol technician first inspects the tiles after eachflight and enters the part number, location, andcondition of each tile into a computer base byvoice. The computer’s central database automat-ically generates a problem report in electronicformat, which a TPS engineer uses to identifyand recommend proper repair procedures for thetile. The problem report proceeds through anelectronic signature loop until final approval forthe repair. Finally, the TPS technician uses thevoice data entry method to indicate tile statusas repair procedures are completed.

The AWADS system and other automatedsystems developed in the SIORA program usethe Ada programing language,l the softwareenvironment that will be used in the space sta-tion, and other large NASA programs in the fu-ture. It offers the advantages of excellent porta-bility from one hardware system to another, arich set of programming functions and tools, anda uniform code documentation. The tile automa-tion system is expected to be operational by Jan-uary 1989.

‘Ada was originally developed for use by the armed services. Ithas become the DoD software standard,

When used in the appropriate application, “it willminimize programming time and effort . . . andimprove the probability of mission success. ”4

Expert Systems

Some systems attempt to capture experts’ prob-lem-solving knowledge in a computer program.So-called “expert systems” could provide consid-erable assistance in automating complex launchand mission operations procedures, such as fuelloading and gantry disconnect, where the experts’knowledge can be codified. Expert systems canalso be applied to maintenance checks and faultisolation procedures which are currently per-formed manually. In their most mature form, ex-pert systems are used as diagnostic assistants.Knowledge engineers and programmers have de-veloped expert systems for a diversity of dis-ciplines, including medicine, geology, chemistry,military science, electronics, education, agricul-ture, and law.

Expert systems solve problems arising in a par-ticular discipline using the same rules of thumbthat humans employ in decisionmaking. A typi-cal expert system has two parts:

●

●

A knowledge base: typically includingdescriptions of relationships among objectsor a set of rules describing actions. Theserules take the form, “if the power is turnedoff, then the system won’t work. ”An inference engine: typically including arule base (in this case, a set of rules of thumbto be used for problem-solving) and meta-rules (instructions that determine the orderin which to use the rules in the rule base whensolving a problem).

Each knowledge base is specific to a particulardomain of knowledge and must be appropriatefor the type of problem to be solved. The infer-ence engine, on the other hand, is generic; it isdeveloped by programmers trained in the meth-ods of artificial intelligence and, once developed,

4USAF Space Division, Launch Systems for the Strategic DefenseInitiative–Data Book (Los Angeles Air Force Station, CA: Head-quarters, U.S. Air Force Systems Command Space Division, De-cember 1986), p. 6-93.

56

can be used with appropriate knowledge bases tosolve a variety of problems.5 At present, knowl-edge engineers act as intermediaries in the proc-ess. The knowledge engineers and programmers

‘Cf. critiques by F.P. Brooks, Jr., “No silver bullet—essence andaccidents of software engineering, “ Computer, April 1987, pp. 10-19; and F. Flores and T. Winograd, Understanding Computers andCojmition: A New Foundation for Design (Norwood, NJ: Ablex,

are now aided, and may eventually be replaced,by computer programs that help translate the ex-perts’ rules of thumb into formats the inferenceengine can interpret. b Box 4-B discusses three ex-pert systems that could be used for launch oper-ations.

bW. B. Gevarter, “The nature and evaluation of commercial ex-19~6). pert system building tools, ” Computer, May 1987, pp. 24-41.

Box 4-B.—Expert Systems for Launch Operations

Expert systems that are potentially useful in space transportation systems include LES (LOX ExpertSystem), KATE (Knowledge-Based Automatic Test Equipment), and ISIS (Intelligent Scheduler and Infor-mation Systems).1

LES is a quasi-expert system built to demonstrate monitoring and troubleshooting of the portion ofthe Shuttle Launch Processing System that performs liquid oxygen (LOX) loading of the Shuttle at KSC.2

Sensors at numerous points in the LOX loading system report the temperature, pressure, and operatingstatus of various subsystems to the Shuttle Launch Processing System. LES was designed to:

1. identify abnormal sensor readings immediately;2. deduce whether an abnormal reading indicates a problem in the loading procedure or merely failed

instrumentation; and3. override reactions to apparent system failures, such as the safing operation, countdown hold, or

launch abort, if it identifies failed instrumentation as the cause.LES produces reports in the format of an Interim Problem Report, a paper form used at KSC for many

years (figure 4-I). LES can also display and print schematic diagrams of the wiring and plumbing it monitors.In developing prototype expert systems for use in launch operations, NASA engineers chose to apply

an expert system to the LOX loading system because a complete functional description of the Shuttle LaunchProcessing System was available. This improved LES’S performance but made LES a questionable modelfor other expert systems that must reason about domains about which they have only fragmented and some-times inconsistent descriptions. An upgraded version of LES (KATE-see below) subsequently demonstratedan ability to diagnose problems even when it had only limited information about the domain, by produc-ing a list of “suspect” faults.

LES’S developers also chose to use algorithmic reasoning rather than than applying “rules of thumb”gained by experience. In other words, LES follows programmed instructions to achieve full logical con-sistency of its diagnoses. In this respect LES is not a true expert system. It also cannot understand systemswith feedback or diagnose multiple failures.

Nevertheless, LES’S developers are enthusiastic about its potential for use on other KSC fluids sys-tems. They suggest that “the cost of . . . software would plummet while the reliability and safety of thecontrol software would rise dramatically. ”3

KATE is an expert system developed to demonstrate monitoring, diagnosis, and control of systemswith electrical, mechanical, hydraulic, and pneumatic components. 4 The present KATE system is being

‘For other examples, see NASA Advanced Technology Advisory Committee, Advancing Automation and Robotics Technology for the Space Sta-tion and for the U.S. Economy, NASA TM-87566, v. II, March 1985, and NASA Ames Research Center, “Systems Autonomy Technology-ProgramPlan,” briefing slides, 1987.

‘J. R. Jamieson, et al., “A knowledge-based expert system for propellant system monitoring at the Kennedy Space Center, ” Proceedings of the 22dSpace Congress, Cocoa Beach, Florida, 1985, pp. 1-9; E.A. Scarl, et al., “A fault detection and isolation method applied to liquid oxygen loading forthe Space Shuttle,” Proceedings of the Ninth International Joint Conference on Artificial Intelligence, 1985, pp. 414-416.

3Jamieson, et al., op. cit., pp. 1 - 9 .‘E.A. Scarl, et al., “Diagnosis and sensor validation through knowledge of structure and function, ” IEEE Transactions on Systems, Man, and Cyber-

netics, vol. SMC-17, No. 3, May/June 1987, pp. 360-368; M. Cornell, “Knowledge-Based Automatic Test Equipment, ” Proc. ROBEXS 86, NASA JSC,June 1986.

57

Figure 4=1

I N T E Ft I tl F’ R O E! L E M R E P O R T it*it**itit********t londay the twenty- f i f th of I larch, 1985; 3:38:23 pm

REPORTED FJY: LES the LOX Expert System

PROBLEM llESCi?IPTION:GLOX3043E the replenish valve open measurement ne. 2 i s n o t r e a d i n g c o r r e c t l y .It now reads: OFF , b u t s h o u l d r e a d : O N .

ANALYSIS & TROUBLESHOOTING STEPS:1) GLOP2045R the replenish valve s ignal pressure measurement detects the current

state of A3370&, because if A33706 was failed to ON, then EiLOP2045fi wouldhave to be reading between -0.5 and 0.5.

2) GLOP2045A the replenish valve signal pressure measurement detects the currentstate of A3370EI, because if A33708 was failed to ON, then GLOP2045A wouldhave to be reading between -0.5 and 0.5.

3) GLOP20450 the replenish valve signal pressure measurement detects the currentstate of A33709, becauseif h33709 was failed to ON, then GLOP2045fi would have to be readingbetween -0.5 and 0.5.however, GLOP204511 is reading 15, thus clearing:20-PSI-PR1, A33784, A33709, A33706, K105, and A33708.

Suspects now are: A 8 6 4 6 01 64019452, 6602111-F2, +22D180B 1 and GLOX3043E.41 GLOX2043E the replenish valve open measurement no. 1 is NOT reading

correctly thus clearing: GLOX3043E.Suspects now are A86460, 6401A452, 6602G1-F2, and +22D180B.

5) GLOH30449 the replenish valve position indicator no. 2 detects the currentstate of /!86460, because if A86460 was between 0.0 and 93, then GLOH3044Awould have to be reading between -5 .0 and 98.however, GLOH3044A is reading 99, thus clearing:A864LI0.

Suspects now are: 6401A452, 6602AI-F2, and +221)180B.61 GLOX2035E replenish valve secondary pressures okay measurements detects the

current state of +22D180)3! because if +22D180)3 was failed to OFF f thenFLOX2035E would have to be reading OFF.however , GLOX2035E is reading ON, thus clear ing:6602A1-F2 and +22D180B.

Suspects now are: 6401A452.Monday the twenty-f i f th of March, 1985:3:38:42 p mAt this point It appears that the most likelysingle point failure is 6401A452 thereplenish valve open limit switch. The restof the measurements will be searched forconflicting evidence.7) The balance of the RELATED MEASUREMENTShave been examined, and cannot add additionalinformation to the above analysis.CONCLUSION:

It is determined that the most likely single Pump

p o i n t f a i l u r e i s 6 4 0 1 ~ 4 5 2 t h e r e p l e n i s h v a l v eopen llmit switchThank VOU----LES r ’Monday’the twenty-fifth of March, 1985:3:38:44 pm aouRcE:wlonal&~ and~Mminhtration

58

demonstrated on a laboratory air purge system. Like LES, KATE is neither designed to understand systemswith feedback nor to diagnose multiple failures.

KATE was developed from LES by modifying the program so that it could not only diagnosis faultsin a monitored system but could also control the system. For example, it can turn valves and motors onand off in an attempt to keep sensor measurements within specified limits. LES diagnoses faults by: 1)hypothesizing faults that might cause sensors to report the undesirable measurements observed; and then2) deducing whether they would cause the undesirable measurements. KATE’s developers realized that thesame method could be used to hypothesize commands that might cause sensors to report acceptable meas-urements and then deducing whether they would cause all sensors to report acceptable measurements.

KATE’s developers added an ability to learn about its domain by experimentation. KATE’s LearningSystem enables KATE to construct a partial knowledge base, or a complete knowledge base of a simplesystem, by observing the performance of a system to which it is connected. KATE “issues combinationsof inputs, each time looking for measurement reactions and filing the results in a table. When all combina-tions have been tried, the table is evaluated to produce frames5 representing the tested system. ”b KATEhas produced complete knowledge bases about simple digital circuits by experimenting with them. KATE’sapproach may be inappropriate for learning about complicated real systems because experimenting withall combinations of inputs would take time and might even cause failures in some systems.

1S1S7 is a job scheduler. It is designed to solve “work flow” problems such as:

We want to produce Tethered Upper Stage Knobs (TUSKS)8 at the maximum rate possible without buyingnew tools. Each TUSK requires casting for five hours, milling for two hours, grinding for three hours, twodifferent half-hour inspections, and five different one-hour tests. We have two molds, one milling machine,two grinding machines, one inspector qualified to perform each inspection, and one test cell for each re-quired test. The casting must precede the milling, which must precede the grinding. Any tests and inspec-tions, which are the last stages of TUSK manufacture, can be done in either order, although it has proveneconomical to inspect before testing. The time required to transfer an unfinished TUSK from one work cellto another depends upon the origin and the destination; these times have been measured and tabulated. Bywhat path or paths should unfinished TUSKS be routed among the operations?

If only one TUSK were to be produced, this scheduling problem would be a “traveling salesman prob-lem” with additional constraints upon the routes the “salesman” [unfinished TUSK] can take through the“cities” [operations] he must visit. The additional constraints can be used to simplify the search for theshortest route, but the resulting simplified problem is still of the traveling-salesman type. The computa-tional effort required to solve such problems by the fastest published methodsg grows exponentially (inthe worst case) as the number of operations increases. The scheduling problem in the example above, al-though far simpler than an actual one, is even more complicated than a traveling salesman problem; itis analogous to the problem of coordinating the itineraries of a succession of traveling salesmen—one de-parting each day—so that the average trip duration is minimized, subject to the condition that no two canbe in the same city on the same day.

At KSC, Shuttle processing operations are now scheduled manually by individuals who maintain chartsshowing durations of individual operations as horizontal bars; these Gantt charts cover several walls. Por-tions of the charts are photographed, printed, and distributed daily and weekly. When schedule interrup-tions, delays, or speed-ups occur, schedulers modify the charts; they must determine a new schedule whichsatisfies all constraints, for example, on the order in which operations can be performed. Except fortui-tously, such a procedure will not result in the most efficient schedule for the workforce. Although schedulersalso try to minimize processing time, they find it impossible to determine and compare all possible sched-ules satisfying all constraints resulting in the most efficient schedule.l” How much vehicle processing timecould be reduced and costs saved by more efficient computer scheduling has not been explored.

5A “frame” is a list of statements about an object’s properties and relations (e. g., connections) to other objects.6M. Cornell, op. cit.7Mark S. Fox and Stephen F. Smith, “ISIS-a knowledge-based system for factory scheduling, ” Expert Systems, vol. 1, No. 1, July 1984, pp. 25-49.‘For illustration only. Any resemblance to acronyms in current or previous use is purely coincidental.‘S. Kirkpatrick, et al., “Optimization by simulated annealing, ” Science, vol. 220, No. 4598, May 13, 1983, pp. 671-680.‘°Critical path methods are used to monitor payload integration schedules at KSC but cannot be used to schedule processing operations; these

methods can identify the sequence of operations that will take the longest to perform (the “critical path”) but cannot rearrange the sequences to save time.

59

TECHNOLOGIES FOR LAUNCH VEHICLES

Vehicle Design Principles

As experience with the Space Shuttle illustrates,vehicle design significantly affects launch and mis-sion operations and plays a crucial part in the abil-ity to reduce costs. Many Shuttle subsystems areextremely difficult and time consuming to main-tain or repair because Shuttle designers focusedon attaining optimum performance and highsafety, often at the expense of ease of ground oper-ations, maintenance, or mission control.7 In or-der to determine which technologies might reducecosts most, launch system designers should con-sider certain design principles.

Include all segments of the launch operationsteam (including logistics personnel) in thedesign of any new launch system.

When plaming and designing a new launch sys-tem, it is essential to consider the entire systemas an interactive entity, including the operationsinfrastructure, and operations management. Thisenables system designers to anticipate potentialoperations and maintenance problems and pro-vide for them before the system is built.

Reduce number and complexity of tasksrequiring human intervention.

Complexity of documentation, maintenance,and interfaces among subsystems generally leadto higher system costs. ‘Therefore, reductions inthe number and variety of tasks necessary forlaunch preparation, especially those that requirehuman involvement, could assist in reducinglaunch costs.g However, vehicle subsystems them-selves can be complex, if they are designed to sim-plify each procedural step. For example, includ-ing self-testing electronics in an avionic subsystemmakes that subsystem complex, but reduces thenumber of tasks required of launch personnel.ESA achieved simplicity in the Ariane by using

‘George E. Mueller, Panel discussion, Space Systems Productivityand Manufacturing Conference IV (El Segundo, CA: AerospaceCorp., August 1987), pp. 233-35.

‘1-l. S. Congress, Office of Technology Assessment, Low-Cost,Low-Technology Space Transportation Options, staff paper, in prep-aration.

a high degree of commonality in the design ofdifferent vehicle stages, and evolutionary designfrom one vehicle to the next. In addition, Ariane-space has simplified the payload/launcher inter-faces that are required for Ariane.

Increase maintainability.

Launcher designers have paid relatively littleattention to providing the ease and simplicity ofassembling or maintaining vehicles. As one OTAworkshop participant observed, “One problemwith the Shuttle is that the systems on board arenot designed for changeout. You can pull a box,but you have to do copper path testing to get itback up there. And on the Shuttle, there area lotof boxes to fail. ” Even with ELVS, the amount oftesting that is done on the pad requires greaterattention to the principles of maintainability. Thefollowing would contribute to launch systemmaintainability:

Increase subsystem accessibility. It would behighly desirable to design subsystems that aremore accessible to repair and changeout. Oneway to assure more accessible subsystems isto include operations people in the designprocess.

Such involvement might avoid situationsin which some subsystems later turn out torequire a lot of detailed inspection and chang-ing, such as the Shuttle main engines, or eventhe air filters in the Shuttle crew cabins,which collect unanticipated amounts of hair,blue-suit lint, washers, and screws. Thesesubsystems are difficult to access and repair.Even where frequent maintenance has beenanticipated, such as in the avionics packages,pulling and replacing electronics boxes re-quires time-consuming retesting to ensure theintegrity of hundreds or even thousands ofelectrical connections.

Workshop participants noted that fluidand mechanical systems particularly needmore accessible design and an improved ca-pability for making internal tests. A futurereusable system might also take a lesson fromaircraft design: airliners are designed to havecertain parts and subsystems pulled and in-

.

60

●

●

●

spected after a given number of hours offlight. On the Shuttle, however, most sub-systems require disassembly, reassembly, andretest after each flight.Design for modularity. Workshop partici-pants also suggested that as much as possi-ble, components should be modular, stand-ardized and interchangeable. To achievedesign modularity means deciding whichfunctions must be handled separately fromothers and how they must be connected, andeven what standards (such as electrostaticdischarge protection) must be used. Havingstandardized interfaces would improve thechances for achieving modularity. This prac-tice is widely used in the design of both mil-itary and commercial aircraft. Thus a con-siderable base of experience already exists.

Arianespace has attempted to assure thatthe Titan 3 and the Ariane 4 use similar pay-load interfaces, because it is in the custom-ers’ best interests to have alternative launchvehicles to turn to. In the opinion of one par-ticipant, “an absolute mistake on the Shut-tle was marrying the payload to the vehicle, ”which results in major software changes foreach flight and complicates the task of re-manifesting payloads for other vehicles.Standardized interfaces do, however, exacta marginal cost in performance. Hence theyrequire that payloads be designed slightlylighter and smaller than the vehicle’s theo-retical maximum capacity. Moreover, it maynot be possible to design standardized inter-faces across all types of vehicles and missionsbecause the mission requirements for, say,launching into geosynchronous orbit are verydifferent than for low-Earth orbit.Include autonomous, high-reliability flightcontrol and guidance systems. This technol-ogy has yet to be fully developed for spacesystems, and will be very expensive.Build-in testing procedures, especally for me-chanical and fluid systems, as well as for elec-tronic systems. As noted in a later section,designers already know how to incorporatetest procedures in the electronic systems. Thebiggest hurdle is in the mechanical and fluidsystems, which are difficult to test.

Make payloads independent of launcher.

Payload integration constitutes a major frac-tion of the cost of launch operations. In the Shut-tle, payload integration has turned out to be a“long, complex, and arduous task, compared tointegrating a payload on an expendable launchvehicle. ”9 In large part this complexity results fromthe potential influence of multiple payloads oneach other as well as the interaction of the pay-loads’ weight distribution and electrical systemswith launcher subsystems. In the case of the Shut-tle, payload customers must also take part in plan-ning and training for deployment or operation oftheir payloads. Prior to the loss of Challenger,payload integration for the Shuttle typically re-quired about 24 months.l0

Workshop participants agreed that payloadsshould be designed to be as independent as pos-sible of the launch vehicle. From the standpointof the launch system managers, the payload-vehicle interfaces should be standard and incor-porate automated checkout procedures, off-lineprocessing, and testing prior to delivery at thelaunch site. However, such an approach finds fewadherents among payload designers, who gener-ally find themselves pushed by the payload per-formance requirements and weight limitations.

Several workshop participants warned thatnothing is gained in reducing overall costs bychanges in procedure or technology that merelysend problems elsewhere. Many of the ALS andSTAS concepts for improved launch operationstend to shift costs from operations to other stagesin the launch services process, such as payloadprocessing. For example, requiring payloads toprovide their own internal power, rather thanrelying on a source in the launcher, may reduceground operations costs, but may also increasethe cost of preparing payloads. On the otherhand, if launch costs per pound were sufficientlycheap, it might be possible to construct less costlypayloads (box 4-C).

9Charles R. Gunn, “Space Shuttle Operations Experience, ” pa-per presented at the 38th Congress of the International Astronauti-cal Federation, Oct. 10-17, 1987.

l~T’& too is likely to increase with the new safety requirementsfor the orbiter.

61

Box 4-C.—Vehicle-Payload Interfaces on the Advanced Launch System

One of the recommendations of the Space Transportation Architecture Study ] was to eliminate or sharplyreduce the burden of supplying special services from the launch vehicle to the payload. This strategy wouldassist in reducing vehicle turnaround on the pad, which in turn reduces the launch operations costs. ALSstudy managers have accepted that recommendation and adopted the philosophy that the ALS will provideminimum services to the payload.

Because the payload designers have been severely constrained by the total weight launchers could carryto orbit, their previous practice has been to consider the launch vehicle as almost an extension of the pay-load, and to expect it to provide a variety of special fittings, upper launch stages, and services such aspower, cooling, and fueling. Not only does such an approach lead to extra costs for the launch vehicleitself, it dramatically raises the costs of integrating and launching the payloads. ALS managers also main-tain that payloads could be much cheaper to build if the payload designers were less severely constrainedby weight capacity of the launch vehicle.

They have asked the payload community to consider the ALS as a transportation system capable oflaunching high mass payloads safely and on time, but which will provide only standard interfaces and limitedservices (table 4-2). As one Air Force manager put it, “This won’t sit well [with the payload designers]because it’s new and it won’t be the same way we’re doing business now. ” He went on to say, “We meanto force a revolution in the design of launch vehicles. An evolutionary approach won’t reach cost reduc-tions of a factor of ten. It just won’t do it. ”

Participants at two Air Force ALS workshops2 on the launcher/payload interface were asked to con-sider and analyze payload designs based on minimum services from the vehicle, and to identify any serv-ices they considered essential. In addition, they were asked to consider the effects on payload design ofdelivering payloads to a ballistic trajectory, just short of orbit. Such a plan would make the launch vehiclemuch simpler because it would avoid adding rockets to the core vehicle to send it into the ocean after deliv-ering its cargo. However, the payload designers would be required to provide their own boost to opera-tional orbit. In return, they could count on vastly reduced costs per pound to reach space.

Considerably more design work will be necessary to determine whether these stratagems could reducetotal launch costs dramatically, and consequently lead to cheaper payload designs. If successful, they couldease many of the current payload design restrictions. For example, with a ten-fold reduction in cost-per-pound to orbit, the weight of the payload and its upper stage could grow by a factor of two and still resultin reductions of a factor of five in cost per payload to orbit. In addition, if designers do not have to findinnovative but costly ways to shave weight, payloads could be much cheaper to design and build. How-ever, if these tactics lead primarily to shifting most launch costs to the payload accounts, the exercise willprove moot. In addition, if payload weights continue to grow to meet or exceed launch vehicle capacity,costs cannot be reduced. At present, the payload community, especially the designers of highly compli-cated national security spacecraft, have met the suggestions with profound skepticism (table 4-3).

‘U.S. Government, National Space Transportation and Support Study 1995-2010, Summary Report of the Joint Steering Group, Department ofDefense and National Aeronautics and Space Administration, May 1986.

‘Held at Aerospace Corp., Los Angeles, CA, November 1987, and January 1988.

SOURCE Of f]ce of Technology Assessment, 1988

Use less toxic propellants. spacecraft as well as for launch vehicles because

Storable high-performance propellants such as engines using such propellants can be started and

nitrogen tetroxide or monomethylhydrazine of- stopped easily.

fer significant advantages where the size of the Although such propellants will continue to havepropellant tanks is an issue, or where propellants an important role in space transportation, theymust be stored for long periods of time, especially are also toxic and corrosive, giving rise to humanon orbit. These propellants can be used in rela- health risks and maintenance problems. Launchtively simple engines and are frequently used for personnel must be protected by special suits from

62

Table 4-2.—ALS Launch Operations Specifications

● Separate launcher preparation from payload prepa-ration

● Place payloads in standard cannisters

● Provide no access to payloads when vehicle is on thepad

● Provide no flight power and communications interfacesSOURCE: Air Force Space Division, 1987.

Table 4=3.—Launch Vehicle/Payload Interface Issues

● How beneficial is standardization of launch vehicle/payload interface?

● How will needed payload sewices be provided ifminimize sewices provided by launch vehicle?

● Are total system costs lowered as launch vehicle costsare lowered and vehicle availability enhanced bykeeping launch vehicle-provided sewices at aminimum?

● Can generalized ALS mission analyses provide timelyloads and environment analysis to payload?

● Can payload requirements still be met if launch vehicledesign is insensitive to payload type?

SOURCE: Air Force Space Division, 1987.

exposure to carcinogenic or corrosive materials.When propellant technicians work with thesefluids, other launch personnel must evacuate thearea. Toxic propellants also tend to destroy sealsand metal containers and create internal leaks.Solving these problems could eliminate a signifi-cant amount of ground processing—especially forreusable systems, which require post-flight handl-ing. Developing better materials for storage ves-sels would help in this effort.

Cryogenic propellants offer lower productioncosts and higher energy density per pound ofpropellant. However, cryogenic rocket enginesand logistical support are generally more complexand expensive than storables.

Vehicle Technologies

The particular technologies used in a launch ve-hicle may enhance or hinder ground and missionoperations. Table 4-4 provides a list of several ad-vanced technologies that could lower launch oper-ations costs when incorporated in a vehicle.

Table 4-4.-Vehicle Technologies To Facilitate GroundOperations

Built-in test equipment (BITE)Thermal protection system (TPS)Fault-tolerant ComputersAutonomous and adaptive guidance, navigation, & control

(GNC) systemSOURCE: Office of Technology Assessment, 1988.

Built-in Test Equipment (BITE)

Future launch vehicles are likely to incorporatebuilt-in test equipment and software to detectfaults and reconfigure redundant systems; thiswould thereby reduce ground operations laborand cost. It could also increase vehicle reliabilityand autonomy during flight, easing mission con-trol requirements. Technology for built-in test ofavionics, especially computers, is most mature.

Technology for built-in test of mechanical andfluid systems, especially sensors, will require moredevelopment. More reliable sensors could reducefalse alarms, and software similar to the expertsystem KATE (see box 4-B) could diagnose sen-sor faults.

Thermal Protection System

Reusable vehicles, such as the Space Shuttle,require a thermal system to protect the vehicleupon reentry. As noted in chapter 3, the Shuttlethermal protection system, which was the first re-usable system to be developed, has proven ex-pensive to maintain. A more robust thermal pro-tection system would reduce the complexity ofinspection and repair and dramatically reduce thecosts of refurbishment. Advanced materials, suchas carbon-carbon composites and titanium-alumi-num alloy, which are being developed for the X-30 program,ll promise much greater tolerance tothe heating and buffeting experienced on atmos-pheric reentry than do the current Shuttle ther-mal protection tiles.

1lu. s. con~ss, General Accounting Office, National Aero-SpacePlane: A Technology Development and Demonstration Program toBuild the X-30, GAO/NSIAIM8-122 (Washington, DC: U.S. Gov-ernment Printing Office, 1988), pp. 37-38. For a detailed discussionof new structural materials and composites, see U.S. Congress, Of-fice of Technology Assessment, Advanced Materials By Design,OTA-E-351 (Washington, DC: U.S. Government Printing Office,June 1988).

63

Fault-Tolerant Computers

The on-board computers of future launch ve-hicles could consist of identical computer mod-ules “mass-produced” for economy and connectedby optical fibers for reduced susceptibility to elec-tromagnetic interference. These modules couldhold software that allows several of them to per-form each calculation, compare results, and voteto ignore modules that report “dissenting” results.This approach, now used on the Shuttle in anearly version employing less sophisticated com-puters,12 would enable the launch vehicle to toler-ate failures in one or more computer modules.Computers with a high degree of fault-tolerancewould allow the launch of a vehicle with a knownfault rather than holding the launch to replace afailed module and retest the system.

The full potential of fault-tolerant computersto reduce maintenance, turn around times, andcost may not be realized until space transporta-tion managers gain sufficient confidence to acceptthe small risks inherent in launching vehicles withcertain known faults. For example, the Shuttle hasa quintuple-redundant primary computer systemand dual-redundant software for these computers.Shuttle avionics also have triple- and quadruple-redundant sensors. Four of its five computerscould fail during flight, or before launch, with-out causing mission failure. Yet, NASA’s highlyconservative launch criteria now require all thesesystems to be operational before the Shuttle canbe launched. This increases safety and the prob-ability of mission success, possibly at the expenseof economy, resiliency, and access probability.In contrast, airlines will fly aircraft with faultyor failed equipment, as long as the equipment isnot vital to safety, or has sufficient redundancyfor safe operations.

Autonomous and Adaptive Guidance,Navigation, and Control

Current launch vehicle avionics that performnavigation, guidance, and control must be up-

12A Spmtor and D. Gifford, “The Shuttle primary Computer SYS-tern, ” Communications of the Association for Computing Machin-ery, vol. 27, No. 9, September 1984, pp. 874-901.

dated before launch with data such as payloadmasses and positions within the vehicle, launchtime, and predicted winds aloft. Launch delaysor changes in weather or payload configurationcan require additional updates.

Advanced avionics and software could makethe mission software less sensitive to payload con-figuration and weather, by monitoring a vehicle’sresponse to commands during the early flightstages. For example, an adaptive guidance andcontrol system could estimate payload mass dis-tribution and use the estimate to calculate guid-ance and control for the remainder of the flight.Similarly, wind could be measured by its effectson vehicle acceleration, trajectory, and structuralstrain, and control surfaces could be moved notonly to steer the vehicle, but also to alleviate struc-tural stresses. A vehicle with these capabilitieswould not only require less detailed programmingbefore launch, it could also be lighter because itwould “know” when to “give” under unantici-pated stress. Some current military and civilianaircraft use such systems.13

Computer-Aided Design and Computer-Integrated Manufacturing

Computer-aided design and computer-inte-grated manufacturing can make vehicle design andproduction faster and probably reduce life-cyclecost even at low launch rates. Computer-aided de-sign techniques can speed vehicle development byautomating the distribution, retrieval, and utiliza-tion of design information. Computer-integratedmanufacturing techniques can reduce productioncosts by automating selected manufacturing oper-ations. ’4 This will require specifically designingsystems to facilitate their manufacture by com-puter integrated methods.

‘3 E.g., Airbus Industries A320; tested on the B-52 and F-4.14 For examp]e, Hercules corporation has used automation to in-

crease the speed and safety of its manufacture of solid rocket mo-tors for the Titan IV.

64

MANAGEMENT STRATEGIES

Without appropriate management strategies,the ability of launch and mission operationsmanagers to utilize new technologies efficientlyis likely to be limited. In addition, managementstrategies that lead to improvements in the waysexisting technologies are used could result in costreductions. IS This section examines several man-agement strategies that experts have suggestedwould increase the efficiency of operations anddecrease costs.

Facilities

Use an integrate/transfer/launch (ITL)approach.

The ITL method of launch operations (figure2-9), in which each individual component in thelaunch process has its own dedicated set of facil-ities, is essential for achieving high launch rates.Separating the different launch operations func-tions in this way means that parallel processes,such as payload checkout and vehicle assembly,can proceed at the same time. However, ITL nec-essarily requires substantial investment in facil-ities. Further, ITL requires mobile platforms andother facilities for moving launch vehicles alongthe steps of the process.

Shuttle operations at Kennedy Space Center(KSC) and Titan operations at Cape Canaveralwere originally designed to use the ITL approach.By contrast, the launch complexes at VandenbergAir Force Base require assembly and integrationon the pad. Payloads must also be tested and in-tegrated with the vehicle on the pad. Such pro-cedures necessarily limit the rate at which Van-denberg can launch. However, even at KSC andCape Canaveral, the launch rate for existing ve-hicles is highly limited, in part because the avail-able facilities are too few and overscheduled toallow the ITL method to be fully realized.

Future launch complexes might be designed toaccept several different launch vehicles in the same

‘s’’Space Systems and Operations Cost Reduction and Cost Credi-bility Workshop,” Executive Summary (Washington, DC: NationalSecurity Industrial Association, January 1987), p. 2-1,2.

general size class (figure 4-2).16 As Arianespacehas demonstrated in a limited way, the samelaunch pad can be used for different launch sys-tems with a minimum of alterations to the launchcomplex.

Locate manufacturing facilities near launchcomplex.

Placing the launcher manufacturing facilitiesnear the launch complex would shorten and sim-plify the launch vehicle supply lines, and elimi-nate the need for most acceptance testing at thelaunch complex. However, unless the launch ratewas expected to be extremely high, such a strat-egy might not pay for itself, because it would re-quire substantial capital investment in facilities.In addition, it would require the manufacturingworkforce to relocate near the launch complexes.

Use off-shore launch pads.

New locations for launching space vehicles maybe needed if demand for launch services increasessignificantly. Because of restraints caused by lackof suitable real estate and cultural and environ-mental restrictions, the Air Force and its ALS con-tractors am studying several off-shore launch con-cepts, including offshore drilling rigs, smalloffshore islands, or even mid-Pacific islands. Inaddition to easing many of the restrictions ofcoastal launch pads, such options have potentialfor launching toward all azimuths.

The Air Force is taking a preliminary look atthe potential for using offshore drill rigs, becausethey seem to present greater opportunity for oper-ational flexibility than an island. It is exploringlaunch pad designs similar to floating oil drillingrigs that could be loaded with a rocket, towed toan offshore site, and, in a matter of hours, turnedinto a stable launch platform.17 For this and otheroffshore possibilities, technical feasibility (espe-cially safe handling of toxic and cryogenic fuels),

lbpeter L. Portanova and l-lardd S. Smith, “Strategic Defense Ini-tiative Launch Site Considerations,” Aerospace Corporation ReportNo. TOR-0084A(5460-04 )-1.

“’’USAF Studies Concept for Launching Heavy-Lift Rockets fromOffshore Rig,” Aviation Week and Space Technology, Feb. 1, 1988,p. 42.

65

Figure 4-2.—Universal Launch Complex

Modular Assemblv

r 7

v ; area* 1!

: ’as required]

1

Y- - - - - - -r

Mobile launch ~latform ~

/ . , / ● Propellant loading

k

‘ [h\> \

, ‘ /[● Countdown and launch

: / I‘ \ ’ \One universal launch stand can provide

<’...% , I high launch rate capability for various vehiclesI

:y -~ —

r

. ,.,. , .,’.“-:,:, / / > < > . . ? . ” >

\ ’

.’ ..* .●. . .

:“:.,,/ ,, ;/ / ‘. . / ,>; ‘ .: j>. - ~

Main principles ‘ \ \\ ? : ‘1

●

●

New operational philosophy ● Containerized cargo(enhanced ITL) c Rapid recovery capabilitySimplified launch stand sized for ● Integrated manufacturing/operat ionslargest vehicle processing on or near launch siteMinimum time on launch stand ● Adaptable launch mounts

(support beams adjustable for vehicle size)

SOURCE: Alr Force Space Dlvlslon

cost, logistics, and onshore facility and harbor re-quirements will all need considerable study.

Operations Management

Create incentives for achieving low cost,successful launches.

The current institutional structure tends topenalize launch failure, but is poorly structuredto lower launch costs or increase launch rate. Al-

though commercial launch service offerors nowhave the incentive of competition to encouragethem to drive down operations costs, similarincentives are not apparent for Governmentlaunches.

Centralize facilities, management, missioncontrol.

One way to lower the overall costs of launchand mission control is to centralize the facilities

66

and personnel. For example, because responsibil-ity for Shuttle launch operations and mission con-trol is divided among KSC, Johnson Space Cen-ter, and Marshall Space Flight Center, NASAmust duplicate some facilities and personnel, andprovide appropriate coordination among centers.

Develop and use computerized managementinformation systems.

Although computer systems play a major partin all parts of mission preparation, launch, andcontrol, they are seldom used for schedulinglaunch vehicle preparation and keeping track ofthe status of launch vehicle and payload systems.For the Shuttle at KSC, for example, Shuttle or-biter refurbishment, system status reports, andsubsystem alterations are all handled by paperdocuments. Not only is a paper system more cum-bersome and subject to error, it requires consider-ably more time.

The fundamental difficulty in changing over toa computerized system is that not only would itrequire the development and installation of a largecomputer system, it would lead to substantial al-terations of the ways in which managers interactwith each other. In other words, it would requirefundamental changes in the institutional cultureof NASA and the Air Force. The computer screenwould replace the “in-box. ”

Increase autonomous operations.

Many operations procedures now carried outby humans, especially routine ones, can be auto-mated to reduce the “standing army” of humanoperators. Automation could also reduce the timespent in preparing for launch and mission opera-tions, and increase system reliability. However,to be automated, operations procedures must alsobe standardized.

In fiscal year 1989, NASA’s Office of SpaceFlight will start an Advanced Operations Effec-tiveness Initiative to focus on automating portionsof launch and mission operations (table 4-s). Itsgoals are to:

●

●

●

●

improve the efficiency and productivity ofSTS operations;develop an integrated software strategy;develop an autonomous space flight opera-tions software test bed for:—determining enabling technologies for

“fully” autonomous operations;—performing hardware/software trade-offs

among operational systems;—characterizing flight operations proce-

dures/techniquesorganize existing autonomous capabilitiesin~o an integrated system.l a -

NASA’s program will serve as a test bed forinserting automated procedures into space trans-portation operations. It should also assist in mak-ing automation more acceptable to launch man-agers, if it is successful in demonstrating theapplicability and safety of autonomous oper-ations.

18NASA JSC Mission planning and Analysis Division, “Autono-mous Spaceflight Operations, ” briefing to OTA staff, Aug. 5, 1987.

Table 4-5.-NASA’S Advanced OperationsEffectiveness Initiative

Kennedy Space Center● Prelaunch processing/preparation● Launch operationsJohnson Space Center● Flight planning/preparation● Flight control● On-orbit operations● Postflight analysisSOURCE: National Aeronautics and Space Administration, 19S7.

ASSESSING TECHNOLOGICAL OPTIONS

Although launch system operations could be Do they contribute to other desired ends, such asimproved in a variety of ways, any proposed im- improving U.S. economic or political competitive-provement must meet the test of several measures ness? Do they help or hurt the morale of the workof merit. Do the intended changes improve effi- force? The primary criteria for judging spaceciency, reduce costs, and/or enhance reliability? transportation system performance and economy

67

include cost, reliability, access to space, and oper-ational flexibility.

19 Proposed changes in launchsystem operations can be evaluated on the basisof their effects on these criteria.

Economic Criteria

Space transportation analysts have used sev-eral economic criteria (box 4-D) to judge spacetransportation system economy. The decisionsabout which economic criteria to use to evaluatea particular technology for a new system, or forimprovements to an existing system, will affect

the choice of technology and even of launch sys-tem design. For example, selecting new launch sys-tems on the basis of the lowest non-recurring costgenerally favors existing technologies and maypenalize designs chosen for highest maintainabil-ity. The Nixon Administration apparently con-sidered low non-recurring cost as paramount inthe initial budgeting for the Space Shuttle,20 whichled in part to a vehicle design that is difficult andexpensive to prepare for launch.

On the other hand, selecting minimum recur-ring cost as the sole criterion may favor technol-

ZONationa] Aeronautics and Space Administration, “Shuttle19STAS Joint Task Team, Nationai Space Transportation and $UP- Ground Operations Efficiencies/Technolo@es Study, ” (Boeing Aero-

port Study, Annex C, May 1987; p, 12, Table 2-3: “Space Trans- space Operations Company), (Kennedy Space Flight Center, NASIO-portation Architecture Screening Criteria. ” 11344), vol. 1, p. 2.

Box 4-D.—Economic Criteria

nonrecurring costs include costs of vehicle design, development, test, and evaluation (DDT&E), and con-struction or improvement of facilities for manufacturing vehicles, processing and integrating vehiclesand payloads, and mission control. The costs of facilities and equipment existing at the beginning ofthe accounting period are considered to be “sunk” and are not included in most calculations of net bene-fit for new systems. Costs of developing technologies should be fully included unless they are being de-veloped for other purposes.

recurring costs include costs of flight hardware (expendable or reusable) and costs of operations and sup-port (e.g., wages). Some recurring costs (e.g., ELV purchases) increase roughly in proportion to the numberof launches; Recurring costs such as salaries are moderately insensitive to launch rate.

life cycle cost (LCC) is the sum of the nonrecurring costs and the recurring costs paid to operate a spacetransportation system for a specified period to transport specified payloads to their operational orbits.Unless otherwise indicated, LCC generally refers to the undiscounted life cycle cost, i.e., the total dol-lars spent regardless of the year in which they were spent.

present value (PV) LCC discounted at a specified rate to reflect the benefit of not investing for a deferredreturn. Depending on the goals they wish to achieve, experts differ concerning which discount rate shouldbe used. A 5 percent discount rate (often used in STAS) is considered a reasonable “real” discount ratefor use as an adjustment for risk and time preference for government investment but should be increasedto adjust for inflation. A 10 percent discount rate was sometimes used in STAS and could be consideredthe sum of a 5 percent “real” discount rate and a 5 percent inflation rate.

cost risk was defined in STAS as the percentage increase in the present value (of life-cycle cost discountedat 5 percent per annum) which would be exceeded with a subjectively estimated probability of only 0.3.

net benefit is the decrease in present value of life cycle cost that could be obtained by an improvementin a space transportation system, for example, by applying new technology to vehicles, building newsupport facilities, or changing management methods. The calculated net benefit of an improvement willdepend upon the discount rate assumed.

cost leverage: is the net benefit of an improvement divided by the present value of the cost of implementingit. The calculated cost leverage of an improvement will depend upon the discount rate assumed.

internal rate of return (IRR) is the discount rate at which the net benefit of an improvement would be zero.

‘Robert C. Lind, Discounting for Time and Risk in Energy Policy (Washington, DC: Resources for the Future, 1982).

68

ogy development with little regard for its cost,if system life expectancy is assumed to be long.The language of the act appropriating fiscal year1987 supplemental funding for the AdvancedLaunch System requires NASA and the Air Forceto obligate and expend funds “only for ALS var-iants which embody advanced technologies witha design goal of reducing the cost to launch pay-loads to low-Earth orbit by a factor of ten com-pared with current space boosters . . . “21 Becauselaunch operations account for a substantial per-centage of overall launch costs, the agencies wouldtherefore have to reduce these recurring costs sig-nificantly.

Present value of life-cycle cost is a flexible cri-terion that can be made to resemble either non-recurring cost (by using a high or variable dis-count rate) or recurring cost (by using a low dis-count rate). STAS analysts used present value oflife-cycle cost discounted at a rate of five percentper annum as the fundamental economic criterion;they also used discount rates of zero percent and10 percent.

The criteria of net benefit, cost leverage, andinternal rate of return (IRR) have been used toidentify technologies that could be applied benefi-cially in a space transportation system. All threecriteria have been used for evaluating options fortechnology development, system design, andmanagement. However, they are not equivalent;options may be ranked differently when evalu-ated by different criteria. An option may even bejudged an improvement according to one criterionand undesirable according to another; an optionwith a positive IRR would have negative net ben-efit and cost leverage at discount rates greater thanits IRR. For example, Boeing Aerospace Co. esti-mated that automating the handling and transferof components for a proposed cargo vehicle wouldyield an IRR of three or four percent and save $55million in an SDI deployment scenario, but wouldhave a negative net benefit ($17 million) at a dis-count rate of five percent, increasing the presentvalue of life-cycle cost by $17 million .22

ZIFiscal Year 1987 supp]em~tal Appropriations Act: Public Law100-71.

ZZUSAF Space Division, Launch Systems for the Strategic mfenseInitiative–Data Book (Los Angeles Air Force Station, CA: Head-quarters, U.S. Air Force Systems Command Space Division, De-cember 1986), pp. 7-22 and C-9.

Noneconomic Criteria

Several non-economic criteria have been usedto rate space transportation system performance(table 4-6 and app. B).23 In addition, other non-economic criteria, such as international techno-logical, political, or economic leadership, are oftenemployed in choosing among competing paths.

Because both ground operations and missionoperations are integral parts of the launch system,they play significant roles in meeting the non-economic, as well as the economic criteria. Forexample, the Titan 34D fleet had low operationalavailability for 1986 and most of 1987 as the re-sult of the failure of a liquid fuel pump and a solidrocket motor respectively, and of standdown pol-icy. During the standdown of the Titan, the AirForce developed non-destructive testing methodsto test the rocket motors prior to assembly. Thesemethods are now in use at Vandenberg and atCape Canaveral.

The example of non-destructive testing illus-trates the tradeoff among economic and non-economic criteria. Although developing and in-stalling these methods of testing were relativelyexpensive and will increase the costs of prepar-ing a Titan for launch, the perceived improvement

ZQther criteria u=d iri STAS are defined in Joint Task Team,National Space Transportation and Support Study 1995-2010, An-nex E (“DoD Functional/Operational Requirements”), May 1986,p. 5, and Annex C (“Space Transportation Architecture”), p. 12,Table 2-3 (“Space Transportation Architecture Screening Criteria”).

Table 4-6.—Non-economic Criteria for JudgingSpace Transportation System Performance

●

●

●

●

●

●

�

Capacity—maximum annual launch rate or payloadtonnage to given orbits.

F/exibi/ity—ability of a launch system to meetalterations of schedule, payload, and situation, and tosatisfy missions in more than one way.Re/iabi/ity—the probability with which a system willperform an intended function successfully.Resi/iency—the ability of a space transportationsystem to adhere to launch schedules despitefailures—to “spring back” after failure.

Operational Avai/abi/ity—the probability that a fleet, ora muHi-fleet system, will be operating (i.e., notstanding down).Access Probability—the probability that a launchsystem can launch a payload on schedule and that thepayload will reach its operational orbit intact.

SOURCE: Office of Technology Assessment, 1988.

69

in access probability was considered to outweighthe costs incurred.

Economic Benefits

Several contractors participating in the SpaceTransportation Architecture Study estimated theeconomic benefits24 to U.S. space transportationof developing or applying technologies to facili-tate ground operations. Table 4-7 illustrates oneset of estimated benefits. 25 The technologies as-sessed include some that would be used in launchvehicles (e.g., built-in test equipment) and othersthat would be used in ground support equipmentor facilities (e.g., automated data managementsystem). The technologies of table 4-7 are con-sidered “enhancing” technologies—they reducecosts in the assumed demand scenario but are notrequired to build or operate the chosen mix oflaunch vehicles and facilities.2b

The benefits of each technology are estimatedin terms of three criteria: internal rate of return,undiscounted net benefit, and net benefit dis-

24Thw ~timates app]y Ody tO U.S. Space transpofiation ‘x~nd-

itures. Benefits of technology transfer to other sectors is difficultto estimate even a posteriori; see NASA Advanced Technology Advi-sory Committee, Advancing Automation and Robotics Technologyfor the Space Station and for the U.S. Economy, NASA-TM87566,March 1985, vol. II, p. 104. It would be very difficult to forecastspin-off benefits from the technologies described above, and OTAknows of no such forecast.

ZSBWing Ae~SpaCe company; from space Transportation Ad?i-tecture Study, In-Progress Review Number 5 (Seattle, WA: BoeingAerospace Company, Apr. 7, 1987), p. 209.

ZbThe mix assumed in table 4-7 features a new piloted orbiter,a flyback booster, and a cargo vehicle core stage with a recover-able propulsion/avionics module,

counted at 5 percent per annum. The technologiesare listed here in order of their estimated internalrates of return. As the table illustrates, if they wereranked according to undiscounted or discountednet benefit, the order would differ. For example,“expert systems” rank highest in undiscounted netbenefit but tenth in internal rate of return. Thebenefits are sensitive to changes in mission model,discount rate, or mix of launch vehicles.

Similar estimates by other contractors differ indetail but generally predict that development ofthese technologies would be beneficial. Table 4-8 compares internal rates of return estimated byBoeing Aerospace C0.27 with estimates by theSpace Systems Division of General Dynamics.28

Comparison of the two sets of estimates is com-plicated because the two companies defined tech-nology categories differently. For example, Boe-ing defined a category it called built-in testequipment (BITE), while General Dynamics in-cluded BITE used for pre-flight testing in a cate-gory it called automated ground operations andBITE used for in-flight testing in a category itcalled flight management systems. The flight man-agement systems and automated ground opera-tions categories included equipment other thanBITE. For example, flight management systems

27Boeing Aerospace Company, op. cit.ZSGeneral Dynamics Space Systems Division, Space Transporta-

tion Architecture Study, Special Report—Interim Study Results, re-port GDSS-STAS-87-031 (San Diego, CA: General Dynamics SpaceSystems Division, May 12, 1987), vol. 2, book 3, pp. 7-90-7-91.

Table 4-7.—Estimated Economic Benefits of Developing Technologies to Facilitate Ground Operations(“Normal Growth” mission model)

Internal rate Net benefit Net benefitTechnology of return (undiscounted) (discounted 5°\0 pa.)Built-in test equipment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140.0?40 $2,617M $911 MAutomated data management system . . . . . . . . . . . . . . . . . .Automated test and inspection . . . . . . . . . . . . . . . . . . . . . . .Accelerated load calculations. . . . . . . . . . . . . . . . . . . . . , . . .Thermal protection system . . . . . . . . . . . . . . . . . . . . . . . . . . .Fault-tolerant computers . . . . . . . . . . . . . . . . . . . . . . . . . , . . .Automated launch vehicle and payload handling . . . . . . . .Database management system . . . . . . . . . . . . . . . . . . . . . . . .Computer-aided software development . . . . . . . . . . . . . . . . .Expert systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Autonomous and adaptive guidance, navigation, and

control system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

115.061.019.516.015.515.013.512.511.5

6.0

1,898M1,454M

247M218M106M830M

5,413M2,225M5,775M

716M

709M498M48M59M30M

227M1,472M

552M1,372M

43MNOTE: Not endorsed by OTA; sensitive to architecture and mission model.

SOURCE: Boeing Aerospace Co.

70

Table 4-8.—Technology Development Benefits: A Comparison of Estimates by Two STAS Contractors(“Normal Growth” mission model)

Boeing Aerospace Co. General Dynamicsproposed vehicles proposed vehicles

Technology IRR IRR Technology

Fault-tolerant computers 15.50/0Built-in test equipment 140.00/0Automated test and inspection 61 .t)”/oAutomated component handling 15.00!0

Automated data management systemb 115.OYODatabase management systemc 13.50/0

Expert systems 11 .50/0Accelerated load calculations 19.50/0

Computer-aided software development 12.59’oThermal protection system 16.0°\oAutonomous, adaptive guidance, navigation, 6.00/0

and control systemKerosene engine NAd

Actuators NAd

Hang gliders NAd

43.40/09.1 0/0

17.8°10

30.50/0

21.8°\o21 .7”!031.1 0/0

4.1 0/01.9?40

Flight management systemsAutomated ground operations

Advanced information processing

Expert systems

Automated software generation and validationReusable cryogen tankageAdaptive guidance, navigation, and control

Liquid oxygen/hydrocarbon enginePrecision recovery

%eneral Dynamics included built-in test equipment et al. in two categories: flight management systems and automated ground operations,bF or ground operations.cFor mission control.d Not assessed: Boeing assumed this technology to be enabling–ie. necessa~ for its recommended architecture–and assessed its net benefit but fIOt its IRR.e Not assessed: General Dynamics assumed this technology to be enabling—i,e, necessa~ for its recommended architecture—and did tlOt assess itS IRR.

SOURCES: Boeing Aerospace Co. and General Dynamics Space Systems Division

included fault-tolerant computers,defined as a separate category.

which Boeing

In table 4-8, OTA attempted to group overlap-ping technology categories. For example, the firstgroup of rows includes four categories defined byBoeing and two, covering the same applications,defined by General Dynamics. The second groupof rows includes the Boeing categories “automateddata management system” (for ground operations)and “database management systems” (for missioncontrol) and the General Dynamics category “ad-vanced information processing” (for both groundoperations and mission control). The third groupof rows includes the Boeing categories “expert sys-tems,” and “accelerated load calculations” and theGeneral Dynamics category “expert systems, ”which would include expert systems for load cal-culations. For those technologies that can be com-pared (e.g. “computer-aided software develop-ment” versus “automated software generation andvalidation”), the estimated internal rates of returndiffer because of differing vehicle concepts, tech-nology application concepts, estimation tech-niques, databases, and judgments.

The economic benefits of so-called “enabling”technologies, which are required in order to builda given mix of launch systems, can also be esti-mated. For example, figure 4-3 displays estimatesof the undiscounted net benefits of four technol-ogies that, if developed, would enable the Nationto develop a mixed fleet of advanced vehicles pro-posed by Boeing Aerospace C0.29 Boeing esti-mated the undiscounted life-cycle cost of the refer-ence launch system mix (featuring the Shuttle formanned flights and Titan IV for cargo) to be $248billion in the STAS “normal growth” scenario.Developing kerosene-burning rocket engines, re-usable liquid hydrogen tanks, actuators for thecontrol surfaces of reusable vehicles, and Rogallowings and control systems (“hang gliders”) to re-turn propulsion/avionics modules to the launcharea, would enable the United States to build newkinds of vehicles that could carry the assumedtraffic for $197 billion-about $50 billion (21 per-cent) less than the reference mix would cost.

ZqThis vehicle mix features a new piloted orbiter, flyback booster,and cargo vehicle core stage with recoverable propulsion/avionicsmodule.

71

Figure 4-3.— Economic Benefits of Emerging Technologies

● Built-in test● Auto data mgt.● Expert systems● etc.

.

Enhancenew vehicles i

I Enablenew vehicles

save $50B (21 O\O)

“ ~

● Kerosene engines. Reusable tanks● Actuators● Rogallo wing

SOURCE: Office of Technology Assessment. Based on estimates by Boeing Aerospace Co.

The figure also displays Boeing’s estimates ofthe undiscounted net benefits of applying ground-operations technologies and other technologies30

to enhance a launch system. Enhancing the refer-ence launch system would save $22 billion, or 9percent of $248 billion. Using the same technol-ogies to enhance operations of advanced vehiclesembodying the chosen enabling technologies

sOThe enhancing technologies include some (e.g., expendablealuminum-lithium tankage), which would not significantly affectground operations but which could reduce life-cycle costs in otherways.

would save $25 billion, or 13 percent of $197 bil-lion. The total savings afforded by enabling newvehicles to be built and then enhancing their oper-ations is estimated to be $76 billion or about 70percent of the undiscounted life-cycle cost of thereference Shuttle-Titan IV launch system. The en-hancing technologies would save $3 billion moreif applied to the new launch system (as Boeingrecommended) than if applied to the referencelaunch system, an example of synergism betweenthe enabling and enhancing technologies.

Chapter 5

Operations Policy

CONTENTS

Page

Improving Operations for Existing Systems. . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . 75Purchasing Launch Services Rather Than Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . 75Charging Payload Programs the Costs of Services. . . . . . . . . . . . . . . . . . . . . . . . . 76Congressional Oversight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Future Launch Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Technology Research and Development.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78Estimating Operations Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79The Role of the Private Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Purchasing Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Incentives for Reducing Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Launch Operations Test Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Institutional policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82New Unpiloted Launch Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82Existing Launch Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Foreign Competition in Launch Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

FigureFigure Page

5-1. ESA/Arianespace Relationship . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

Chapter 5

Operations Policy

As previous chapters emphasize, substantiallyincreasing the efficiency of space transportationoperations, and reducing costs, will require im-provements in both management strategy and theapplication of technology. Neither alone will besufficient. Reducing costs will also require greaterattention to launch and mission operations withinthe National Aeronautics and Space Administra-tion (NASA) and the Air Force, and oversight by

Congress. Although modest reductions in opera-tions costs are possible for existing launch sys-tems, new launch systems, especially designed forlow-cost operations, appear to offer the poten-tial for significant savings. The following discus-sion explores several policy options for puttingthe necessary technologies and management strat-egies to work.

IMPROVING OPERATIONS FOR EXISTING SYSTEMS

The current U.S. launch fleet consists of a rangeof vehicles and systems capable of placing from440 to 40,000 lbs, into low-Earth orbit. Even ifCongress funds development of a new launch sys-tem to be built in the mid to the late 1990s, theUnited States will continue to use most of the cur-rent fleet (with relatively minor modifications) un-til the beginning of the 21st century.1 Operatingthe existing systems will be expensive. For exam-ple, the Shuttle system will cost the U.S. taxpayerapproximately $2.5 to $3.5 billion per year (1988dollars) for the foreseeable future.2 Of this, some$1.5 to $2.0 billion per year will be devoted spe-cifically to ground and mission operations. An-nual operations costs for the ELV fleet could to-tal $120 to $150 million.3

Congress could assist efforts within the AirForce and NASA to reduce operations costs ofcurrent systems by using its legislative authorityto require both NASA and the Air Force to pur-chase launch services, rather than vehicles, fromprivate industry. Reducing costs for Govemment-operated launch systems is inherently difficult be-cause Government lacks the cost saving pressuresof a competitive market environment. Neverthe-

ILJ. S. Congress, Office of Technology Assessment, Launch Op-tions for the Future: A Buyer’s Guide, OTA-ISC-383 (Washington,DC: U.S. Government Printing Office, July 1988).

‘U.S. Congress, Congressional Budget Office, The 1988 Budgetand the Future of the NASA Program, Staff Working Paper, March1987.

3Based on an annual launch rate of six Titan IVS, 4 Delta 11s, and4 Atlas-Centaurs.

less, Congress could help by requiring NASA andthe Air Force to charge the costs of ELV launchesto the program that uses or “owns” the payload.Congress could also use its oversight authority toconduct site visits focused on launch and missionoperations, hold hearings on cost reductions, andmandate reports to Congress on the agencies’ ef-forts to reduce operations costs.

Purchasing Launch Services RatherThan Vehicles

The Administration’s latest space policy, whichwas released February 11, 1988, specifically directsNASA and other civilian agencies to purchaselaunch services rather than launch vehicles for ex-pendable launchers, unless they require the useof the Titan IV, which is under Air Force owner-ship and control.4 This policy is intended to as-sist the development of the private launch indus-try by putting the private sector in charge of theproduction and operation of most expendablelaunch systems. The policy should also have theeffect of lowering operations costs for these launchsystems, if launch services are procured competi-tively and if launch companies are given theauthority to determine the most cost-effective

“’Civil Government agencies will encourage, to the maximum ex-tent feasible, a domestic commercial launch industry by contract-ing for necessary ELV launch services directly from the private sec-tor or with DoD. ” White House, Office of the Press Secretary, “FactSheet, ” Feb. 11, 1988, p.9.

75

76

operations methods, consistent with safety andreliability y.

For national security space activities, the Presi-dent’s space policy allows the Department of De-fense (DoD) to procure either ELV launch serv-ices or launch vehicles.5 Congress could strengthenthese two policies by including them in legisla-tion. Congress could expand space transportationpolicy and encourage private sector competitionby directing the Air Force and other national secu-rity agencies to purchase launch services ratherthan vehicles for all expendable launches, includ-ing the Titan IV. Such a policy would requireturning over the responsibility for the operationof the Titan IV and any new launch systems toprivate industry.

Charging Payload Programsthe Costs of Services

Space transportation costs make up a signifi-cant fraction of the total cost of a payload mis-sion, yet within NASA and the Air Force thebudget for payload design, development, and con-struction is independent of the budget for spacetransportation services. Thus, payload programmanagers have little direct incentive to seek re-ductions in transportation hardware or operationscosts. One way to reduce operations costs overthe long term, especially for existing systems,would be to require NASA and the Air Force tocharge each payload program the recurring costsof transportation services provided. Such a pol-icy would enable each agency and the Congressto develop a more realistic picture of the overallcost of a scientific or applications program. If pay-load managers had to pay for launch services outof their payload budgets, they would have greaterincentive to encourage designers to design pay-load/vehicle interfaces that are cheaper to inte-grate and service.

This policy would be more effective for ELVSthan for the Shuttle, because the latter is still un-dergoing considerable modification and many ofthe costs for the Shuttle are development costs.In addition, a large portion of the investment inthe Shuttle is in the reusable orbiters, which func-

tion both as space platforms and launch vehicles.Apportioning costs of the Shuttle to the differentusers would also raise the question of competi-tion between using the Shuttle and using an ELV.

Congressional Oversight

Reducing the operations costs of existing launchsystems will require a series of evolutionary stepsinvolving new technology and incremental changesin management strategies. Inserting new technol-ogy into current space transportation operationsis costly and time-consuming. The complex pro-cedures and myriad rules of launch and missionoperations that have evolved over 30 years of thepublicly funded space program are extremely dif-ficult to change significantly, even for a newlaunch system, because they are so complex andinterrelated, and require such varied human skills.Altering one aspect of a launch system forcesother parts to change, often unwillingly. Becauseof these complex human and technological inter-actions, operations managers tend to be highlyconservative in adopting proposed changes inlaunch procedures. Reducing operations costs willrequire the willingness of NASA and Air Forcemanagement to focus continual attention on in-serting new cost-saving technology and innova-tive management strategies into operations proc-esses. It will also require congressional oversightto assure agencies’ attention to cost reductionstrategies and consistent funding by Congress.

Changes of management strategy alone canyield significant savings in operations costs (seealso ch. 2). For example, the Strategic Defense Ini-tiative experiment Delta 180, which was flown ona Delta launch vehicle, demonstrated that reduc-ing NASA and Air Force oversight and report-ing requirements can decrease launch operationscosts. b These requirements work against contrac-tor innovation because launch services companiesfind it easier to keep the payload customer happywith the same “tried and true methods” than tochange them. Current requirements reduce the po-tential for applying new technologies to launch

‘Ibid., p. 8.bDepartment of Defense Strategic Defense Initiative Office/Ki-

netic Energy Office, “Delta 180 Final Report, ” vol. 5, March 1987.

77

operations and raises the cost of launch servicesto the Government.7

Congress could assist the introduction of moreefficient management methods by directing NASAand the Air Force to reduce their direct involve-

7As one launch company official that OTA queried put it, “it costsin two ways. NASA hires more people than it needs to, and we [thelaunch company] have to hire extra people to respond to NASA’sconcerns. ”

FUTURE LAUNCH SYSTEMS

The most promising means of reducing theoperating costs of launch systems is to developa new system specifically designed with that goaluppermost. The Air Force and NASA are cur-rently working on an Advanced Launch System(ALS) program with the goal of reducing the costof launch per pound by a factor of 10 over cur-rent costs. Efficient, cost-effective ground opera-tions are essential elements of the ALS Program,but would require considerable initial investmentin new facilities and operations technology in or-der to realize the benefits of the ALS designs cur-rently proposed. ALS planners are developing ve-hicle designs that incorporate rapid, low costvehicle processing. Carrying out the goals of theALS program would also require implementinga management philosophy that stresses the impor-tance of low-cost operations.

If Congress wishes to lower the operationalcosts of future launch systems significantly, itmust be willing to appropriate funds for the nec-essary launch operations facilities as new systemsare developed. New launch pads and associatedfacilities would be needed. Many of the currentinefficiencies in U.S. launch systems are the re-sult of adding to or improving existing aging fa-cilities. For example, many of the facilities at John-son Space Center and Kennedy Space Center wereoriginally built in the 1960s for Saturn 5. TheShuttle launch complex SLC-6 at Vandenberg AirForce Base was modified from a facility originallybuilt for the Air Force’s Manned Orbiting Lab-oratory, a program that was canceled in the early1970s.

Some analysts have argued that the next ma-jor space transportation system, designed for low-

ment in the launch process, especially for expend-able launch vehicles. Congressional attention inthe form of site visits, hearings, and reports toCongress would play an important part in assur-ing that such direction is carried out effectively.In this, as in other areas of congressional over-sight, Congress should be cautious not to burdenthe agencies with reporting requirements that in-hibit the agencies’ ability to conduct efficient oper-ations.

cost operations, should be funded in a multiyearprocurement that would require Congress to com-mit the country’s resources over a period of sev-eral years, just as it does with certain weaponssystems. Such a multiyear procurement could al-low for larger production runs of launch vehiclesand major investments in new launch operationsinfrastructure. This strategy could reduce the costper unit vehicle, and provide additional incentivesto design and build modular facilities capable ofbeing modified to accommodate new technology.A recent study by the Congressional Budget Of-fice has shown that where the developmentand procurement of weapons systems has beenstretched out, whether because of Administrationor congressional action, costs of weapons haverisen. s The Space Shuttle program provides an ex-cellent example of this phenomenon in the civil-ian space program.

However, as the OTA Special Report, LaunchOptions for the Future: A Buyer’s Guide, pointsout,9 such a strategy would also require the Ad-ministration and Congress to agree on long-termgoals for the space program and a level of fund-ing to support such goals. IO

‘U.S. Congress, Congressional Budget Office, Effects of Weap-ons Procurement Stretch-Outs on Costs and Schedules (Washing-ton, DC: U.S. Government Printing Office, November 1987).

‘W.S. Congress, Office of Technology Assessment, Launch Op-tions for the Future: A Buyer’s Guide, OTA-KC-383 (Washington,DC: U.S. Government Printing Office, July 1988).

:Osee u.S. Congress, Congressional Budget Office, The NASAProgram in the 199(% andl?eyond (Washington, DC: U.S. Govern-ment Printing Office, 1988) for a comprehensive discussion of thebudgetary impacts of different development paths for the civilianspace program.

78

TECHNOLOGY RESEARCH AND DEVELOPMENT

Research and development efforts have a con-tinuing role in reducing costs of space transpor-tation. The Space Transportation ArchitectureStudy11 and OTA’s own examination of spacetransportation have provided a substantial list oftechnologies that could serve as a foundation forimproved launch efficiency and lower operationscosts (tables 4-1 and 4-4). Because many of thesetechnologies would be related to those requiredfor operating ELVS, the Shuttle or even a spacestation, the various governmental programs couldbe closely linked. For example, research devotedto the development of autonomous systems forhandling hazardous substances for the Shuttle andexpendable launchers could be applied to similarsystems for a space station.

In order to accomplish the goals of increasingefficiency and reducing costs, Congress could au-thorize, and appropriate funds for, a technologydevelopment and insertion plan specifically di-rected toward these goals. OTA workshop par-ticipants supported the development of a nationalstrategic technology plan designed to improvelaunch technologies for a wide range of launchproblems and activities. A national plan wouldalso provide interagency and interagency coordi-nation in order to reduce any duplication of re-search and development being done in NASA andthe Air Force.

Part of this national plan should be directed spe-cifically at launch operations. “We need long-termobjectives, ” said one participant, “which arehanded down from the highest levels to determinewhere we go with the technology. I think thattoday a great deal of money is spent and investedin shotgun development that really has a limitedyield. Budget constraints are real and are here tostay. With a well-developed strategic plan, youcan accomplish a lot more with the same moneythat is being spent in the agencies today. ” Sucha plan would enable NASA and the Air Force tocoordinate their efforts and to focus on a variety

IIU.S. Government, National Space Transportation and SupportStudy 1995-2010, Summary Report of the Joint Steering Group, De-partment of Defense and National Aeronautics and Space Admin-istration, May 1986, pp. 15-19.

of technologies having both near and far termpayoffs.

Since the early 1970s, when NASA decided tofocus its space transportation efforts on the spaceshuttle, the United States has invested very littlein technologies that would lead to improving theefficiency and reducing the costs of launch oper-ations procedures and payload processing. NASA’sCivil Space Technology Initiative, and the AirForce and NASA Focused Technology Programthat is part of the ALS program (see ch. 2, IssueB) could contribute to achieving these goals. How-ever, they do not spend enough effort on insert-ing technology into operations.

A thoughtfully constructed technology devel-opment plan would generate an ongoing programof incremental improvements to launch vehicles,facilities, and launch operations. One of the goalsof such a program should be to develop opera-tions technology and procedures designed to fosterroutine launching. It would assist the need for im-provements in current vehicle systems and sup-port research and development of operations foradvanced vehicles. A technology developmentplan should include work in all phases of tech-nology development:

●

●

●

broad technology exploration (basic research)—in areas of potentially high payoff such asautomation and robotics, built-in-test pro-cedures, and fault tolerant computers;focused research leading to a demonstration—of flight or ground operations systems suchas avionics packages, expert systems, auto-mated inspection systems; andimplementation to support specific applica-tions—in day-to-day operations. This phaseshould also include the development of meth-ods to insert such technology with minimumdisruption to existing procedures.

Even without a national technology plan, Con-gress could assist the integration of new technol-ogies into Shuttle launch operations by provid-ing modest additional funding specifically for aNASA technology insertion program. In addition,it could hold hearings to assess the progress NASAand the Air Force are making in coordinating ex-

79

isting research and development programs forlaunch and mission operations.

Congress could also enhance the developmentof new operations technologies by funding anoperations test center specifically designed to carryout tests of new technologies for incorporationinto existing and new launch systems (see TheRole of the Private Sector, below) .12 Such a cen-

12As ~art of its study of AIS for the Air Force, General DynamicsSpace Systems Division has specifically suggested turning LaunchComplex 13 at Cape Canaveral into an “ALS Operations Enhance-

ESTIMATING OPERATIONS COSTS

As noted in chapter 4, projected future savingsin operations costs will have to be examined care-fully to assure that the costs of making the pro-posed changes (up-front, fixed costs) are less thanthe savings realized in recurring costs over the lifeof the program. To accomplish this, design anddevelopment of improvements to operations willrequire adequate cost estimation models in orderfor the agencies and Congress to make informeddecisions about whether the proposed improve-ments meet cost reduction goals. As noted inchapter 3, existing cost estimation models haveproven grossly inadequate in estimating opera-tions costs. Workshop participants urged that newcost-estimating models be developed. Althoughthe current ALS Program includes funding to sup-port the development of accurate cost models,congressional oversight may be required to assurethat the agencies focus on this issue.

Congress could require that the Air Force andNASA report on their progress in developing

ter would consist of a mock launch complex andthe necessary supporting facilities for testing newconcepts and technologies outside the flow of nor-mal launch operations. It could enhance both theCSTI and the ALS Focused Technology Programand could also demonstrate the insertion of newmethods, techniques, and equipment into exist-ing launch systems.

ment Center, ” which would be available to the entire aerospaceindustry—General Dynamics briefing to OTA, Mar. 15, 1988.

more accurate cost estimating models. As pointedout in chapter 3, many of the data that could beused to verify the accuracy of new models havenot been gathered, particularly for launch oper-ations. In part this has been the result of congres-sional and Administration cost-cutting measures.However, such measures only inhibit future costestimation, because reliable models cannot be de-veloped without access to this important infor-mation. Collecting and maintaining such datacould be much cheaper in the long run than at-tempting to make decisions based on incompleteinformation. NASA and the Air Force should re-quire contractors to provide this information.

A new cost estimation model should be as freeas possible of potential bias. To avoid such bias,or a more direct conflict of interest, it may beappropriate to task an independent agency suchas the National Academy of Sciences, or the Gen-eral Accounting Office, to develop an independ-ent cost model.

THE ROLE OF THE PRIVATE SECTOR

One of the difficulties the Government faces inestablishing its own programs to reduce costs isthat such programs generally lack the sort of in-centives provided by the competitive environmentof the marketplace. Until recently, the develop-ment and operation of U.S. launch vehicles werethe sole responsibility of the Government. Now,however, three private U.S. firms offer commer-

cial launch services on ELVS originally developedwith Government funding—General Dynamics(Atlas Centaur), Martin Marietta (Titan), andMcDonnell Douglas (Delta). In addition, threestartup companies are also marketing spacelaunch services— Space Services, Inc. (Cones-toga), Orbital Sciences Corporation (Pegasus) andAmroc Corporation (Industrial Launch Vehicle).

80

The French firm Arianespace, the Soviet Union,and the Peoples Republic of China also offer awide range of ELV services.

Competition among U.S. firms, and with for-eign launch companies, which receive substantial

r !go ernment subsidy, may eventually spur U.S.fi ‘s to invest in additional facilities and tech-nologies for reducing the cost of launch opera-tions. However, the industry is not yet involvedenough in developing new operations technol-ogies. 13 Two Primaw factors in the existing in-

stitutional arrangements for launch operations im-pede privately funded innovations.

First, private launch firms only have incentiveto invest in new facilities and technologies for re-ducing costs if the up-front investment leads tosufficient future profits. Yet launch demand forcommercial payloads in the mid-1990s does notappear large enough to foster such private invest-ment today.

Second, all current ELV launch facilities, includ-ing the safety and range components, are ownedand operated by the Air Force. Private firms leasethem for commercial launches on a cost reimburs-able basis. Although industry can institute somecost savings measures in launch operations atthese Government-owned facilities, they are con-strained by the necessity to deviate as little as pos-sible from procedures and facilities used forlaunching government payloads. To do otherwisewould not in general be cost-effective. Unless thegovernment encourages such investment by re-moving unnecessary barriers of documentationand reporting and rewarding innovation, launchfirms are unlikely to assume such risks on theirown,

The Government could stimulate the innova-tive power of the launch industry by purchasingservices rather than systems; providing incentivesfor developing new, cost saving methods; and byproviding a Government-funded operations testbed.

‘3’’ Space Systems and Operations Cost Reduction and Cost Credi-bility Workshop,” Executive Summary (Washington, DC: NationalSecurity Industrial Association, January 1987), p. 2-22.

Purchasing Services

In 1985, Congress appropriated funds for theAir Force to procure an improved Titan launchvehicle (the Titan IV) to serve as a backup to theShuttle for critical DoD payloads. By committingto purchase several vehicles at once in a “blockbuy,” the Air Force saved money. Although theAir Force purchase of the Titan IV has stimulatedthe domestic ELV launch industry and resulted insavings on vehicle hardware, it has not reducedoperations costs very much. In a truly competi-tive environment, relatively high demand forGovernment payloads could lead to reduced oper-ations costs, especially if private firms had greatercontrol over launch operations. However, blockbuys are not in themselves likely to result in sav-ings on launch operations, because the Govern-ment still controls the manufacturing and launchprocesses.

More recently, the Air Force conducted a com-petition to purchase a lower capacity MediumLaunch Vehicle II (MLV-2). This purchase repre-sents a different strategy in which the Air Forcepurchases launch services rather than vehicles.Under this form of purchase, the Governmenttreats launch service providers much as it treatscompetitive commercial procurements from anyother service industry, and pays for the deliveryof a payload to a specified orbit. The launch serv-ices company provides the launch vehicles and allsupporting services, including launch operations.Government officials work with the launch firmto assure that the firm meets Government stand-ards of manufacture and service. However, theylimit their involvement in the details of the man-ufacturing and launch process. The launch firm,not the Government, accepts the financial risk ofa launch failure, and guarantees a reflight or othercompensation. However, because a launch fail-ure would mean losing an expensive payload aswell as a vehicle, Government officials havestrong incentives to maintain current levels oflaunch operations oversight despite attempts toreduce oversight. They are concerned that the riskof failure and a subsequent free reflight may notbe sufficient discipline for the launch firm.

81

General Dynamics Corporation, which won theAir Force MLV-2 competition, will provide 11Atlas-Centaur-2 launchers and associated launchservices for a firm fixed price of about $40 mil-lion each. The Government will rely on a deliv-ery schedule, performance, and reliability guaran-teed by General Dynamics. If the launcher failsfor reasons associated with manufacture or prep-aration, General Dynamics guarantees a reflight.This arrangement will require fewer Governmentoversight personnel and give General Dynamicsa financial incentive to improve the efficiency andreduce the costs of launch operations. Althoughprecise figures for the savings involved are im-possible to derive because this version of theAtlas-Centaur has not existed before, companyspokesmen estimate this method of procurementresulted in savings to the Government of 12 to20 percent. Savings of this magnitude are possi-ble both because the Government makes a “blockbuy, ” which reduces the cost of manufacturingeach vehicle, and because the Air Force will notbe overseeing the Atlas-Centaur production line.

Purchasing launch services rather than vehicleshas not resulted in immediate savings in launchoperations, in part because the Air Force stillmanages the launch pads. Additional savingsshould be possible as experience with this methodof providing launch services grows. For the pur-chase of launch services to be most effective, theGovernment will have to carry through with itsresolve to reduce oversight to a minimum and giveprivate firms greater control over launch opera-tions. Under the terms of a fixed price servicescontract, tasks outside the scope of the contract,such as increased documentation and reportingrequirements, will cost the Government more.

Government purchases of commercial launchservices offer the potential for synergism betweenGovernment and private sector attempts to reduceoperations costs. The large Government purchasesgive private industry an assured financial basefrom which to work in competition with foreignfirms. As the U.S. launch industry begins to dem-onstrate cost reductions in its commercial launch

operations, some of these gains may be transfer-able to Government launch operations.

Incentives for Reducing Costs

OTA workshop participants pointed out thatthe Government agencies had been less innova-tive than they might have been in providing con-tractors with direct incentives to lower the costsof launch operations. In part, this is the result oftheir historical focus on the performance andsafety of launch vehicles, and a desire to limit ini-tial investment, rather than on reducing long-termoperations and maintenance costs .*4

As the Delta 180 program demonstrated, cashincentives for meeting schedules can be an effec-tive means of increasing launch operations per-formance (see Issue A, ch. 2). The Governmentcould explore other possible incentives for reduc-ing costs. Existing types of incentives do not spe-cifically address the reduction of operations costs.

Launch Operations Test Center

As noted in the section above on TechnologyResearch and Development, a space transporta-tion operations test center could assist innovationin operations technology. NASA currently oper-ates several aeronautics test facilities, which theaircraft industry uses on a fee basis. For exam-ple, the NASA Wallops Island facility maintainsa runway and associated test facilities for assist-ing the private sector in improving the landingand flying characteristics of commercial aircraft.Such a facility could be operated as a Govern-ment-owned, contractor-operated establishment.

Alternatively, Congress might deem it appro-priate to establish a center that is funded in partby the private sector. Such a center could be oper-ated by a private consortium that brought to-gether experts from private companies, the Gov-ernment, and the university community.

I iNational Aeronautics and Space Adm inistrat ion, ‘‘ShuttleGround Operations Efficiencies/Technology Study, ” KSC ReportNAS1O-11344, Boeing Aerospace Operations Co, May 4, 1987, p. 2.

82

Congress could also direct the Air Force andNASA to fund research within private firms toexamine ways of reducing the weight and com-plexity of payloads. As noted in chapter 4, launchvehicles are only part of the equation for obtain-ing assured access to space. If launch operationscosts are to be reduced significantly, there mustbe a complementary emphasis on reducing pay-load costs and simplifying payload designs. The

INSTITUTIONAL POLICY

Attempts to reduce operating costs in planninga new launch system would be more effective ifthose responsible for managing launch and mis-sion operations and facilities were directly in-volved throughout the design and demonstrationprocess. However, they must not only be involvedin these processes, but have the institutional in-fluence, or “clout,” to make their views heard andacted upon. Giving operations experts broader in-fluence over launch system planning and designwill require substantial changes in the “institu-tional culture” of NASA and the Air Force.

New Unpiloted Launch Systems

As pointed out in chapter 4, to reap the poten-tial gains offered by revolutionary changes in tech-nology will also require revolutionary manage-ment changes. However, the United States is notlikely to achieve the desired result if the currentinstitutional structures of NASA and Air Forceare left intact. One way to effect change in theinstitutional culture of launch operations and giveoperations personnel more influence in launch sys-tem design would be to separate the responsibil-ity for system design and development from theoperations responsibility. For example, Congresscould decide to fund development of a new launchsystem under the management of the Air Forceand NASA with the understanding that the oper-ation of the new launch system would be con-ducted by the private sector. Under such an ar-rangement, the launch company would commenceoperation after the completion of developmentflights and would provide launch services to theGovernment on a contractual basis. In order toencourage attention to cost reduction, the com-

private sector could help with these. A wide va-riety of new ideas have surfaced with the DoDLightsat’ 5 and ALS programs. These and relatedideas should be examined for their potential ap-plicability to lowering launch operations costs.

lsThe LightSat program, funded by DARPA, is exploring waysto increase the cost+ffectiveness of spacecraft by reducing the weightand size of payloads.

pany would also be encouraged to market its serv-ices to other payload customers, either from theUnited States or abroad.

The European Space Agency has found such anarrangement effective. When planning for the de-velopment and operation of the European Arianelauncher, ” the European partners decided earlyto separate the functions of launcher developmentand operation. ESA, using the French spaceagency CNES as technical manager, has devotedits efforts to building an efficient, low-cost vehi-cle; Arianespace, S. A., a private French corpo-ration, has focused on developing cost-effectiveoperations (figure 5-1). Arianespace markets theAriane launcher and provides launch services.Neither institution can proceed without the helpand expertise of the other, but each contributesto the development of an efficient launch system.The result has been a relatively simple vehicle thatcan be prepared and launched quickly with a min-imum of personnel .18

The Ariane example presents an attractivemodel for launch operations because it created asubstantive division between the responsibility

l~see U.S. Congress, Office of Technology Assessment, Civi)ianSpace Policy and Applications (Springfield, VA: National Techni-cal Information Service, June 1982), for a description of the devel-opment of the Ariane and the role of ESA.

1l’Because it owns a substantial percentage of Arianespace stock,the French government has significant influence over decisions madeby Arianespace.

18The typical Ariane 3 launch requires about 35 full-time launchpersonnel, plus additional personnel who assist in preparing the ve-hicle. The Ariane 3 typically requires about 4 weeks to assemble,integrate, and test, and 2 weeks on the pad. The Atlas-Centaur,which has approximately the same payload capacity, takes abouttwice as long for the same procedures.

83

Figure 5-1. - ESA/Arianespsce Relationship

P

ESA memberstates

ESA as customer , ,m ‘

i s ~ Mt&?$paM$Other

customemESA as monitor

I CNES* Iv 4

Anane 1 Anane 2/3 Anane 4 Anane 1develop develop develop production

%antre Nationale D’ 6tudes Spatiales (the French Space Agency)

I

Launch Marketing

operations salesfinance

ProductionAriane 213/4 I

SOURCE: Arianespace

and authority for development and that for oper-ations. This institutional division gives each in-stitution a base of power from which to work inarguing its technical case for reducing operationscosts. Because Arianespace competes in the inter-national marketplace, it has strong commercialincentives to minimize these costs.

In order to make such an arrangement functionin reducing costs, the U.S. Government wouldhave to purchase services rather than vehicles,provide a strong economic incentive to reducecosts, and limit Government oversight, providedthe launch company proved its capacity to deliverpayloads to orbit within schedule a~[d budget. Thelaunch company would also have to assume a ma-jor portion of the economic risk of launch failure.

Such a division of responsibility would haveseveral advantages. First, launch costs would bemore visible and comprehensible than they arewithin the current institutional structure. Second,because the launch company would focus onlaunching vehicles and payloads for its clients,rather than on vehicle or payload development,

it would have a major stake in limiting the num-ber and extent of vehicle modifications that wouldnegatively affect its ability to launch on sched-ule. Third, because the launch operations com-pany would also be encouraged to compete forlaunch services in the international market, itwould have considerable incentive to lowerlaunch operations costs. Finally, if the ALS orother launch development program succeeds insubstantially reducing launch costs, the launchfirm would likely have many more private sec-tor and foreign customers for launches than canbe foreseen under existing demand projections.In general, the institutional tension that such adivision of authority and responsibility would cre-ate could enhance innovation and lower costs ofproduction.

This institutional arrangement would be suc-cessful only if the technology used in the ALSwere, and were perceived to be, well within thestate of the art. If significant components of theALS pushed the limits of technology, such an ar-rangement would likely to be considered too risky

84

by both Government officials and private in-dustry.

The arrangement would wrest much of the con-trol the Air Force and NASA now exert overlaunch operations from these institutions. How-ever, unlike earlier proposals for privatizing theShuttle system, this policy option would not in-volve piloted launchers, and therefore would notencounter the objection that a private corpora-tion should not have control over a symbol ofU.S. technological prowess. In other words, thismodel does not seem appropriate for operationof piloted reusable research and development ve-hicles.

Existing Launch Systems

In the absence of a massive reorganization ofthe launch institutions in NASA and the AirForce, it may still be possible to focus increasedattention on reducing the Government’s costs ofoperations.

In order to assist NASA and the Air Force inreducing operations costs for current launch sys-tems, Congress could direct these agencies toestablish an operations division independent oftheir launch development responsibilities. In bothagencies, these functions are mixed. Conse-quently, because budgets are also co-mingled, itis often difficult or impossible to determine whatoperations procedures really cost. Separating de-velopment activities from operations more clearlywould allow the agencies “and Congress to focusmore effectively on the true extent of operationscosts. Such an institutional change has the strongadvantage that it would lead to relatively few dis-ruptions of NASA’s and Air Force’s current or-

ganizational structure and procedures. However,it has the disadvantage that it would force onlya limited cultural change within the agenciestoward operating launchers on the basis of lo-wered costs. This option, if pursued by Congress,would require considerable congressional over-sight to assure that the agencies carried out thewill of Congress. It would only work if users wererequired to pay launch costs.

The proposed space station is another largeproject in which operations costs would consti-tute a significant proportion of life-cycle costs. Be-cause of its concern over the cost and manage-ment of space station operations, the NationalResearch Council recently urged a similar or-ganizational structure for the U.S. space stationprogram. It has suggested “an organizational en-tity, independent of the space station developmenthierarchy, with the ultimate responsibility foroperating the space station. ”19

In any event, Congress may wish to direct theAgencies to develop a plan with the goal of giv-ing launch operations and logistics managers astronger voice in the design of launch vehicles.The OTA workshop on launch operations af-firmed the importance of giving launch operationsand logistics managers an early and influential rolein the design of new launch vehicles. They shouldalso be given greater control over the budget foroperations. Most participants agreed that any de-sign changes to current vehicles should be madewith the principle of lower operations and main-tenance costs as a foremost criterion.

1gNatjona] Research Council, Report of the Committee on theSpace Station (Washington, DC: National Academy Press, 1987),p. 34.

FOREIGN COMPETITION IN LAUNCH VEHICLES

NASA, the Air Force, and commercial launch nology application, and we can go off and do itcompanies should examine launch operations in as well as or better than anyone else. Thus we areother countries. U.S. agencies and companies tend reluctant to look at other nations and learn fromto suffer from the “not invented here” syndrome. their approaches.” However, launch organizationsAs one OTA workshop participant put it, “we in in other countries may have something to offerthe U.S. believe that we are the leaders in tech- in reducing operational costs.

85

For example, Arianespace, the commercialoperator of the European Ariane launcher,20 hasfocused its attention on reducing the costs of con-structing and launching the Ariane, which unlikeU.S. expendable launchers was originally designedas a launch vehicle not a missile. In designing theAriane system, the European launch designerslearned a tremendous amount from previous U.S.experience and used it in their own designs.

ZOThe European Space Agency (ESA) developed the Arianelauncher in the late 1970s. After extensively testing the Ariane, in1984, ESA turned over management of launch operations to Ariane-space, a French corporation supported in part by the French gov-ernment.

Japan has made considerable progress in devel-oping its HI and HII launch systems. Because Ja-pan is likely to offer the latter commercially, itwill also give considerable attention to launchoperations costs, especially because Japan needsa second launch center located near the equatorin order to reach geosynchronous orbit more effi-ciently.

The United States can expect future foreignlaunch concepts like the U.K. HOTOL, the Ger-man Saenger, or the Japanese Spaceplane to bedirected in part at commercial sales. The HOTOLand the Saenger, especially, are being designedwith careful attention to improving launch oper-ations efficiencies.

Appendixes

Appendix A

Cost-Estimating Relationships

This appendix contains a brief summary of currentcost estimation methods to illustrate the uncertaintyand subjectivity involved. Methods used in the SpaceTransportation Architecture Study (STAS) are typi-cal and are used as examples for the general discus-sion. The second half of this appendix presents threeexamples from STAS. One illustrates the derivationof a cost-estimating relationship for vehicle structures; lthe other two illustrate estimated labor cost savingsachievable by developing and applying automationtechnology.

A Summary of Cost EstimatingRelationships Used in STAS

Ground Rules and Assumptions

A cost estimation effort such as undertaken in STASmust begin by making basic assumptions; those madeby the sponsor of the analysis are included in statedground rules, which also specify what the system mustdo. For example, STAS specified a matrix of possiblemission models and required: 1) design and cost esti-mates of a minimum of two independent vehicles, withno major subsystems in common, including launch,orbital transfer, and return of specific high-prioritypayloads “to provide assured access”; and 2) designof facilities and equipment allowing a surge factor of40 percent over the nominal launch rate.

Parametric Cost Estimation: When, as in STAS, sys-tems are to be designed for economy, designers do notknow optimal values of system parameters such as sizeand weight at the outset. The costs of vehicles or ve-hicle subsystems are therefore estimated parametri-cally; in other words, they are expressed as formulascalled cost-estimating relationships, or CERS, whichmay be used to calculate estimated cost in terms of pa-rameters such as weight.

A CER for a vehicle may be derived in a “bottom-up” manner by designing several launch vehicles thatare similar except in size, and, for each vehicle, add-ing up the estimated costs of the subsystems and la-bor required. The costs of the subsystems maybe esti-mated in a similar manner by designing them andadding up the estimated costs of the parts and laborrequired to build them, etc. This approach is labor-intensive, because it requires preliminary design of avehicle.

‘In STAS, costs of expendable hardware and spare reusable hardware areincluded in operations costs.

Alternatively, a CER may be derived by fitting acurve to a “scatter plot” of the weights and inflation-adjusted costs of similar vehicles that have actuallybeen built.z The most common procedure is a combi-nation of these approaches: designers develop a pre-liminary design for a vehicle in only enough detail toestimate the weights of its major subsystems in termsof the vehicle weight or its payload capacity. CERSfor the subsystems are then derived by extrapolationand interpolation from historic data, if they areavailable.

Individual CERS are derived for development costsand for the cost of producing the first unit. Incrementalcosts of additional units are assumed to be lower thanthe cost of producing the first unit by a factor that de-pends upon the number of units produced (learningeffect) and the production rate (rate effect). CERS forlabor costs of ground operations are derived in a sim-ilar manner.

Manifesting and Optimization

After CERS for vehicles, facilities, and operationshave been developed, manifesting and optimizationprograms are employed to determine the most eco-nomical types and sizes of vehicles for the missionmodel. First, a trial mix of vehicles is assumed. A sizeis assumed for each vehicle, and its cost is estimatedfrom the CERS. Then a manifesting program is run todetermine the least costly way of combining (co-mani-festing) payloads on vehicles so they reach their oper-ational orbits in the specified year, taking into accountany restrictions on co-manifesting for security andsafety. The launch rates in the resulting manifest de-termine the operations costs, and the maximum launchrate, inflated by 40 percent to provide a surge capa-bility, determines the number of facilities required forprocessing and launch. Costs of development, facil-ities, vehicle production, and operations are dis-counted and totalled to obtain a projected presentvalue of life-cycle cost. The process is repeated assum-ing different vehicle mixes, sizes, and technologies, anddifferent ground operations and mission control tech-nologies, to determine the most economic architectureand technology content.

‘For a comprehensive published description of this approach, see D.E.Koelle, “Cost Model for Space Transportation Systems Development, Fabri-cation and Operations” (Ottobrunn, FRG: Messerschmitt-Boelkow-BlohmGmbH, Bericht Nr. TN-RX1-328 B, 1983).

89

90

Estimation of Risk of Greater-Than-Expected Cost

STAS contractors estimated cost risk subjectivelyat the level of the vehicle systems mix, not in a bottom-up manner. The estimated risk, expressed as a cost,was given as percent weight in the overall score usedto screen system mixes, as specified by stud~ groundrules. Those who estimated the risk may halve beenunfamiliar with cost risks apparent to subsystem ex-perts. Moreover, cost risk was estimated assumingground rules were met; risk that ground rules will notbe met increases cost risk.

Cost Estimation Examples

Example 1: Parametric Estimation ofVehicle Structure Cost

To derive a CER for the cost of vehicle structures(e.g., inter-stage adapters), one contractor began byplotting the weights and inflation-adjusted costs of ve-hicle structures it had built previously.3 The resultingscatter plot is shown in figure A-1.

‘Such data are often proprietary, which makes valid comparisons withCER’S derived by other manufacturers very difficult.

Figure A-l. - Historical Costs and Weights ofV e h i c l e S t - n =

o

+ wg

1“ 0 -

●

/! ●●

● 0●

0.1 I I100 1,000 10,OOO

Weight (lb)Legend:● non-man-rated expendable+ mm-rated expendableO man-rated reusable

The contractor observed that the costs of the struc-tures rated for human use-especially the reusable one,were significantly higher in proportion to their weightsthan were the costs of the expendable structures notrated for humans. It therefore decided to derive a basicCER for unpiloted expendable structures and assumethat the greater costs of crew-rating and reusabilitycould be represented by “complexity factors” by whichthe basic cost should be multiplied. The basic CER isrepresented by a straight solid line in figure A-1, andcomplexity factors for crew-rating and reusability y arerepresented by arrows from this line up to the datapoints for the crew-rated systems. Figure A-2 was de-rived from three assumptions: 1) that design for reus-ability increases cost by a factor independent of struc-ture weight, or 2) rating for human use, and 3) thatthe complexity factor for rating for human use is in-dependent of structure weight.

Critique: The contractor could verify assumption(3) by comparison with a commercially available CER,presumably derived from different data and independ-ent assumptions; the contractor found good agree-ment, but this does not imply that the assumptionwould be correct in all cases. Assumptions 1 and 2were neither supported nor contradicted by the limiteddata available to the contractor; they are educatedguesses—the best the contractor could do under thecircumstances. Although they are not clearly incorrect,their accuracy is unknownj and the contractor pre-

Figure A-2.-Coat-Estimating Relationships (CERS)for Vehicie Structures (first-unit production costs)

0 . 1 ~100 1,000 10,OOO

Weight (lb)Legend:— non-man-rated expendable

SOURCE: Office of Techno@y Assesament, 19W SOURCE: Office of Technology Assessment, 1988

91

sented no estimates of uncertainty in the complexityfactors. If the contractor or a Government agency hadaccess to proprietary data of other contractors on ex-pendable structures rated for humans, there might beenough data points to test assumptions 1 and 2, al-though the variety of vehicles built to date maybe in-adequate to accept or reject the assumptions with highconfidence.

Example 2: Savings From AutomatingGround Operations

This example and example 3 summarize the proce-dures a different STAS contractor employed to esti-mate potential savings from automating ground oper-ations and mission control functions.

To estimate savings from automating ground oper-ations, the contractor first calculated the labor requiredto perform functions such as refurbishing avionicsusing current methods based on estimates of labor re-quired to perform similar functions on existing vehi-cles, and adjusted to reflect the fact that a new vehi-cle would have different needs. For example, a flybackbooster would require a more robust thermal protec-tion system than the Shuttle orbiter and would not re-quire refurbishment. A percentage reduction in laborfor each such function was then estimated for each newtechnology proposed (e.g., automated test & inspec-tion). This reduction was assumed to be achieved bydecreasing the crew size and the number of shifts byequal percentages; the reduction in processing time(number of shifts) allowed the required number of ve-hicles per year to be processed with fewer facilities,thus saving costs and lead time as well as direct laborcosts for new facilities. From these savings was sub-tracted the costs of developing the new technology re-quired and applying it; these costs “were estimatedbased on the costs associated with similar programs,including the costs of developing the STS Launch Proc-essing System. ”

Critique: The relevance of such costs as a basis forextrapolation could be questioned, because compara-ble automation was not developed for the STS LaunchProcessing System.

Example 3: Savings From AutomatingMission Control

To estimate savings from automating mission con-trol, the contractor first estimated the recurring andnon-recurring costs of performing five mission controlfunctions (flight planning, simulation and training,payload integration, data load preparation, and flightcontrol) using current technology. The costs of per-forming these functions in 1995, using 1990 technol-ogy, and in 2000, using 1995 technology, were alsoestimated; in general, the recurring costs were lowerwhile the non-recurring costs were higher. At the highlaunch rates assumed, the life-cycle costs were loweredby assuming use of the new technology, when avail-able. The fractions of net savings (cost reduction mi-nus cost of technology development) for each func-tion attributable to each new technology and toimproved management were then allocated accordingto a formula, e.g., 40 percent of net savings on flightplanning was attributed to use of expert systems. Tech-nology development funding requirements were listedby year, although the contractor’s reports do not makeclear the basis for their derivation.

Critique: The costs of developing “ordinary” soft-ware have proven difficult to estimate accurately, evena posteriori, when the size of the program (on whichsome estimates are based) is known.4 The accuracywith which existing methods can estimate costs of de-veloping software such as expert systems is not known.

‘Chris F. Kemerer, “An Empirical Validation of Software Cost EstimationModels,” Communications of the Association for Computing Machinery, vol.30, No. 5, May 1987, pp. 416-429.

Appendix B

Noneconomic Criteria

Capacity of a space transportation system can bedescribed in terms of maximum annual launch rate orpayload tonnage. However, capacity is actually mul-tifaceted and is better described by a set of numbers:the maximum launch rate of each fleetl to each of sev-eral reference orbits, along with trade-offs amongthese, if any (e.g., sharing of launch pads by differentfleets).z

Flexibility is the “ability of the space transportationsystem to . . . respond to schedule, payload, and sit-uation changes with . . . responsiveness . . . [and] ca-pacity, ”3 “ . . . and to satisfy missions in more thanone way.”4 For example, a fleet’s ability to share traf-fic (carry some or all payloads manifested for a differ-ent fleet) improves the flexibility of a multi-fleet sys-tem. Flexibility may be valued for its own sake orindirectly, if it contributes to resiliency, operationalavailability, or access probability.

Reliability is the probability with which a systemwill perform an intended function. A system designedto perform several distinct functions will have a relia-bility corresponding to each function. For example, afully reusable vehicle would be designed to transportpayloads to orbit safely anc3 return safely. The prob-ability that it will reach orbit and safely deploy a pay-load (its ascent reliability) is greater than its missionsuccess reliability—the probability that it will reachorbit, safely deploy a payload, and return. Missionsuccess reliability is a commonly used criterion, butreliabilities of non-critical subsystems are also of in-terest because they affect maintenance costs.

High reliability contributes to, but is not requiredfor, resiliency and operational availability; it is nec-essary for high probability of access and return. As-cent and return reliabilities determine the risks of lossof payloads and reusable vehicle components; theserisks include expected replacement costs as well as non-financial risks, e.g., to national security or politicalsupport of space programs. One of the difficulties inusing reliability as a criterion is the uncertainty in esti-mates of the reliabilities of operational vehicles and,especially, proposed vehicles.

1A fleet is that portion of a space transportation system which consists oflaunch vehicles of a single type, e.g., the Shuttle fleet, the Titan-IV fleet, etc.

‘Capacity was not used as a screening criterion in STAS because candi-date architectures were required to have sufficient capacity to fly all missionsin the mission model. Excess capacity contributed to architecture scores in-directly through effects on operational flexibility and resiliency, etc.

‘Joint Task Team, National Space Transportation and Support Study 1995-2010, Annex E (“DoD Functional/Operational Requirements”), May 1986,p. 5.

‘Ibid., Annex C, p. 12, tab. 2-3.

Because perfect reliability is unattainable, the mar-ginal cost of reliability must increase without boundas reliability approaches 1.0, and a lower reliabilitymust be optimal (see figure 2-4 in chapter 2). The op-timal reliability would be the reliability at which thevalue of reliability less the cost of achieving that relia-bility is greatest. The value of reliability has been esti-mated in some cases by calculating the expected re-placement costs of payloads and reusable vehiclecomponents; such estimates have been used to argue,e.g., that reusable systems with reliability below .985should not be considered as viable candidates. How-ever, such estimates are likely to undervalue reliabil-ity because they neglect intangible risks. The costs ofproviding various reliabilities have been estimated byconsidering different configurations of critical compo-nents (e. g., engines) with different degrees of redun-dancy, totalling component costs, and estimating relia-bility from estimates of component reliabilities.

Resiliency is the ability of a space transportation sys-tem to adhere to launch schedules despite failures—to “spring back” after failure. A fleet is considered tobe resilient if the probability of a failure while recov-ering from a failure is less than 0.35.5 This criterionis derived from several assumptions:

1. Payloads are launched at a nominal launch rateuntil a failure occurs.

2. After a failure, launch attempts cease for a dura-tion called the downtime, during which the causeof the failure would be investigated and cor-rected; however, the reliability of the launch ve-hicle is assumed to remain the same.

3. Reservations for some payloads scheduled to belaunched during the standdown are assumed tobe canceled; the rest, a fraction called the back-log factor, are added to the pre-failure backlog(if any) of payloads queued for launch.

4. When launches resume, payloads are launchedat the maximum launch rate (the surge rate) atwhich the system can operate in an effort to re-duce the backlog. Meanwhile, new payloads arereadied for launch at the nominal launch rate andare added to the backlog.

5. Launches continue at the surge launch rate untilthe backlog is eliminated, unless another failureoccurs first.

5The resiliency of each fleet of a multi-fleet space transportation systemmay be calculated as though it were the only fleet, if no fleet can carry pay-loads manifested for a different fleet. The resiliency of the multi-fleet systemis considered inadequate if the resiliency of any of its fleets is inadequate.If the fleets can share payloads, calculation of the probability of a failureduring a surge period can be very complicated.

92

93

If the probability of a failure during a surge periodexceeds 0.35, it is unlikely that the backlog can everbe eliminated after a first failure.’ However, resiliencymay be made as good as desired by changing any one(or any combination) of three of the four parametersthat determine fleet resiliency: reliability, downtime,and nominal launch rate. Increasing surge launch rateresults in only limited improvement in resiliency.7

Operational Availability is the probability that afleet, or a multi-fleet system, will be operating (i.e.,not standing down) at a randomly selected time.8

Operational availability is intimately related toresiliency; a fleet or system with high resiliency willhave a high operational availability. Reducing down-time can make operational availability as great asdesired; a fleet that never stands down is always avail-able, even if unreliable. A mixed fleet is not requiredfor high operational availability.

Access Probability is the probability that a spacetransportation system can launch a payload and that

“Harry Bernstein & A. Dwight Abbott, “Space Transportation Architec-ture Resiliency,” (El Segundo, CA: The Aerospace Corp., March 1987).

‘The probability of a failure during a surge period (PF) may be calculatedfrom the formula

T K(T#?n/(S–1)PF = 1– Ps d

where Ps is the launch vehicle reliability, I?n is the nominal launch rate, SRnis the surge launch rate, S is the surge factor (SRn/ Rn), Td is the downtime,and K(Td) is the backlog factor, which depends on the downtime. Down-time is defined as the interval from a failure until the next operational launchattempt. K(Td) = 1— Td/6 is assumed.

‘This does not mean that it must always have an unreserved vehicle ona pad ready to launch on a day’s notice.

the payload will reach its operational orbit intact. Twokinds of access probability can be distinguished: prob-ability of access on demand, and probability of accesson schedule. High probability of access on demand re-quires flexibility—so that a high-priority launch canbe scheduled quickly when required for national se-curity—and high probability of access on schedule, sothat the scheduled launch will most likely be success-ful. Probability of access on schedule is operationalavailability (the probability that a launch can be at-tempted when scheduled) times vehicle reliability (theprobability that the payload will reach its intended or-bit intact if launched).g 1°

A space transportation system is often said to pro-vide assured access if it provides means for placinghigh-priority payloads in their operational orbits ondemand with high probability. Access probability de-pends on the payload—and the variety of vehicles thatcan launch it—and is more important for some pay-loads than for others.

‘N .b.: traffic sharing is not required. Cf. STAS groundrule A-1: “Viablearchitecture will be based on a mixed fleet concept for operational flexibil-ity. As a minimum, two independent (different major subsystems) launch,upper stage, and return to Earth (~pecially for manned missions) systemsmust be employed to provide assured access for the specific, high prioritypayloads designated in the mission model. ”

10If downtime has not been minimized, traffic sharing may improve accessprobability, otherwise it can reduce it. If a reliable launch vehicle fails, usinga backup vehicle that is less reliable than the primary launch vehicle payloadwilI decrease the probability of getting the shared payloads to orbit safely.If the primary vehicle were indeed more reliable, it would be safer to useit while accident investigation proceeds; its unreliability, even if due to sys-tematic problems, would be lower. Moreover, it would be costly to main-tain a backup fleet.

Appendix C

Reliability

Launch vehicles can in principle be made very relia-ble by incorporating redundant components and sub-systems, and by detailed testing during manufactur-ing and launch operations. However, it is costly tomanufacture a highly reliable vehicle, and maintain-ing and verifying its reliability until launch imposesheavy burdens, high costs, and delays on ground oper-ations. Yet, operating a fleet of unreliable vehicles isalso costly: direct financial costs may include thereplacement costs of payloads and vehicles and thecosts of supporting the launch operations force dur-ing a standdown.

If only financial costs are considered, it is possibleto determine the most economical reliability for a ve-hicle, if payload replacement costs, etc., can be esti-mated before the vehicle is designed. This has beendone for some launch vehicles and orbital transfer ve-hicles.’ Intangible costs may also be included if ex-

‘Boeing Aerospace Corp. has developed a computer program called STROPto perform such an analysis; it was first used to determine the most economicalreliability for the Inertial Upper Stage.

pressed in monetary terms. However, this requires sub-jective judgement of the value of military satellites tonational security, 2 and other intangibles.

The reliabilities of currently operational launch ve-hicles are not known with certainty or precision. Onthe basis of actual launch experience, they can onlybe said to lie within certain confidence intervals withcorresponding statistical confidence. As more launchesare attempted, the confidence intervals will shrink andstatistical confidence will grow. The reliability of aproposed launch vehicle or variant can only be hy-pothesized on a semi-analytical, semi-subjective ba-sis. Confidence levels for the reliabilities of currentlyoperational launch vehicles are listed in table C-13

‘U.S. Congress, Office of Technology Assessment, Anti-Satellite Weap-ons, Countermeasures, and Arms Control, OTA-ISC-281 (Washington, DC:U.S. Government Printing Office, September 1985), p. 33.

3Confidence intervals are estimated from statistics in Harry Bernstein andA. Dwight Abbott, “Space Transportation Architecture Resiliency,” (ElSegundo, CA: The Aerospace Corp., March 1987), using the exact confidencebounds of Y. Fujino, Biometrika, vol. 67, 1980, pp. 677-681.

Table C-1 .-Launch Success Statistics (since 1976)

Successes Average Minimum reliabilityVehicle /attemDts success rate at eaual confidenceDelta . . . . . . . . . . . . . . . . . . . . . . . . 60/63 95.20/o 900/0Atias E . . . . . . . . . . . . . . . . . . . . . . 25/28 89.30/o 81 0/0

Atlas/Centaur. . . . . . . . . . . . . . . . . 26/29 89.70/o 81 0/0

scout . . . . . . . . . . . . . . . . . . . . . . . . 14/14 100.00/0 87°!oTitan . . . . . . . . . . . . . . . . . . . . . . . . 51/54 94.40/0 880/0Shuttle . . . . . . . . . . . . . . . . . . . . . . 24125 96.00/0 870/oAriane . . . . . . . . . . . . . . . . . . . . . . . 14118 77.8?40 690/oNOTE: Forecasts (for which OTA does not vouch) of the reliabilities of proposed vehicles differ, e.g.:

S1S.1 (post-Ch#lengerj: 0.98 [Aerospace Corp.], 0.997 [GD], 1.0 [NASA HQ].Tltm.lV: 0.98 [Aerospace Corp.], 0.98 [GD].MLV (Delta derivative): O.M [Aerospace Corp.], 0.978 [G D].

SOURCE: Office of Technology Assessment, 1988.

94