Concurrent design Estimating probable system cost Launch vehicle reliability Small-satellite costs Designing a sensor system
Winter 2000/2001 Vol. 2 No. 1
Concurrent Design at Aerospace Clockwise from left: Thomas W. Trafton is Director of the Vehicle Concepts Department, which has been leading the development of the Concept Design Center. He holds a B.S. in mechanical engineering from California State Polytechnic University in San Luis Obispo and has been with Aerospace since 1966 ([email protected]
). Stephen P. Presley, Associate Director, Vehicle Concepts Department, oversees the work of the Concept Design Center. He holds a B.S. in engineering from the University of Washington and an M.A. in organizational leadership from Chapman University. He has been with Aerospace since 1990 ([email protected]
). Patrick L. Smith, Principal Director, Architecture and Design Subdivision, has many years of experience in Kalman filters and control. He supported efforts to improve the Air Force acquisition of space systems, particularly in the area of risk analysis. He holds a Ph.D. in control systems engineering from the University of California at Los Angeles and has been with Aerospace since 1968 ([email protected]
). Rhoda G. Novak, former Director of the Software Acquisition and Analysis Department, founded the Concept Design Center’s Ground Systems Team. She holds an M.S. in computer science from Loyola Marymount University and has been with Aerospace since 1978 (rhoda. [email protected]
). Andrew B. Dawdy, Vehicle Concepts Department, is one of the founding developers of the Concept Design Center. He manages a section responsible for systems engineering and conceptual system design. He holds an M.S. in aeronautics and astronautics from the University of Washington and has been with Aerospace since 1992 ([email protected]
2 Headlines 55 Links 56 Bookmarks 59
Aerospace Systems Architecting and Engineering Certificate Program
Estimating Probable System Cost Stephen A. Book, Distinguished Engineer, Systems Engineering Division, was a member of the 1998 Cost Assessment and Validation Task Force on the International Space Station and the 1998–99 National Research Council Committee on space shuttle upgrades. Book shared the 1982 Aerospace President’s Achievement Award for work showing that a particular nonuniform 18-satellite Navstar GPS constellation would satisfy the original GPS system requirements, with potentially large cost savings. He holds a Ph.D. in mathematics with concentration in probability and statistics from the University of Oregon and has been with Aerospace since 1979 (stephen.a.book @aero.org).
Cover: Progressive refinements of a conceptual design of a proposed boost-phase interceptor system. (See the article “Concurrent Design at Aerospace.”)
From the Editor
Small-Satellite Costs David A. Bearden, Civil and Commercial Division, is Systems Director of the JPL and NASA Langley Independent Program Assessment Office, supporting programs such as Mars Exploration, New Millennium, and Discovery. He led development of the Small Satellite Cost Model, as well as its application to NASA independent reviews and deployment of the Concurrent Engineering Methodology at JPL’s Project Design Center. Bearden is also editor of this issue of Crosslink. He holds a Ph.D. in aerospace engineering from the University of Southern California and has been with Aerospace since 1991 ([email protected]
Space Launch Vehicle Reliability I-Shih Chang, Distinguished Engineer, Vehicle Performance Subdivision, leads research on space launch vehicle failures and on solid- rocket-flow and thermal analyses. He has organized workshops on rocketry for the American Institute of Aeronautics and Astronautics and is a panel member of the White House Space Launch Broad Area Review Task Force. He has a Ph.D. in mechanical engineering from the University of Illinois and has been with Aerospace since 1977 ([email protected]
David A. Bearden mproving space systems is a challenge in an era of tightened budgets and reduced development schedules. Given fewer resources, how can designers and procurers of space systems assemble more capable systems? In their quest to comply with government and industry requirements to do more with less, designers have embraced modern design-to-cost practices and innovative design approaches. The Aerospace Corporation is continually working to develop advances that address this changing procurement environment and help designers and customers alike understand the performance and cost implications of early decisions. A widely accepted industry guideline is that 80 percent of design options are locked in during the first 20 percent of a project’s development time. It doesn’t pay to cut budgets and schedules during concept definition; early decisions are critical and often irreversible at later development stages. This issue of Crosslink showcases the experience, tools, and processes Aerospace uses to balance performance, cost, and schedule in designing space systems. Sequential design has given way to concurrent design. Through advanced design models, cost-estimating approaches, lessons-learned databases, and collaborative design teams, Aerospace has created a powerful environment for the concurrent engineering of space systems. In this setting, designers make difficult decisions about what objectives can be achieved, often relying on forty years of lessons learned from space system failures and their causes. The next time you see a launch or read about a satellite placed in orbit, think about the processes presented in this issue of Crosslink. We hope they’ll help you understand the many considerations associated with developing complex space systems.
Space-Based Systems for Missile Surveillance Terrence S. Lomheim, Distinguished Engineer, Sensor Systems Subdivision, leads focal-plane technology development and electro-optical payload design and optimization. He has published 29 articles on these topics. Lomheim received the Aerospace President’s Achievement Award in 1985 for his work with focal-plane technology. He holds a Ph.D. in physics from the University of Southern California and has been with Aerospace since 1978 ([email protected]
). David G. Lawrie, Director, Sensing and Exploitation Department, leads modeling and simulation efforts in support of space-based surveillance programs. He was the recipient of The Aerospace Corporation’s highest award, the Trustees’ Distinguished Achievement Award, in 1997 for his role in developing a new national early-missile-warning system. He holds a Ph.D. in astronomy from the University of California at Los Angeles and has been with Aerospace since 1986 ([email protected]
Headlines For more news about Aerospace, visit www.aero.org/news/ Precision Window for the Space Station
Photo courtesy of NASA
hen astronauts view Earth from the International Space Station, they will look through a glass porthole developed by Karen Scott of Aerospace. This 20-inchdiameter window provides a view of more than threequarters of Earth’s surface and is the highest quality window ever installed in a crewed spacecraft. Astronauts will be performing long-term global monitoring with remote-sensing instruments and Earth science photographic observations. As primary optical scientist for developing the window, Scott tested the viewing glass originally planned for the window and found that it would not support highresolution telescopes or precision remote-sensing experiments. Her recommended
Karen Scott of the Aerospace Houston office, flanked by Astronaut Mario Runco and Dean Eppler of Science Applications International Corporation (SAIC), looks through the space station’s optical research window.
upgrade was approved for the four-piece window, now consisting of a thin exterior “debris” pane, primary and secondary pressure panes, and an interior “scratch” pane. Scott led a 30-member team from Johnson and Kennedy Space Centers, Marshall Space Flight Center, and the University of Arizona Remote Sensing Group that conducted calibration tests on the upgraded window before it was installed in the Destiny module scheduled for launch in January 2001. The team determined that the window could support a wide variety of research, including the monitoring of coral reefs and Earth’s upper atmosphere. Scott’s efforts in completing the tests on a tight schedule brought her a Johnson Space Center group achievement award.
he Aerospace Corporation was recognized at a NASA press conference in June 2000 for its role in bringing down a 17-ton “giant.” The Compton Gamma Ray Observatory reentered Earth’s atmosphere and safely plunged into the Pacific Ocean June 4, 2000, after nine years in orbit studying gamma-ray emissions in space. It is one of the largest spacecraft ever launched by NASA. NASA began deliberating the probability of uncontrolled reentry after one of the observatory’s three attitude-control gyros failed in December 1999. Given the spacecraft’s size, scientists thought it likely that several large masses of the spacecraft could survive reentry. NASA enlisted the corporation’s assistance, and Aerospace technical experts helped to design and execute the successful splashdown. William Ailor and Kenneth Hagen participated in a NASA “red team” review to 2 • Crosslink Winter 2000/2001
Photo by space shuttle crew, Compton Science Support Center. Courtesy of NASA
Bringing Down a “Giant”
Compton Gamma Ray Observatory.
provide recommendations to NASA program management. They were assigned responsibility for reviewing the state of the remaining gyros and the likelihood of uncontrolled reentry. After NASA decided in
February to deorbit the observatory, Wayne Hallman and Benjamin Mains helped develop the deorbit plan with the operations team from Goddard Space Flight Center.
Preventing Power Failure on the Space Station
Popular Science Picks “Picosats”
aymond de Gaston, an Aerospace engineer at NASA Johnson Space Center Operations in Houston, received NASA’s prestigious Silver Snoopy award for his work in preventing possible electrical power failure on the International Space Station. The
(See Crosslink, Summer 2000.) Two more picosats, launched in July 2000, are scheduled for orbital release during the summer of 2001.
Battling Buffet Loads on the Titan IVB
n Aerospace investigation has revealed new data on why the Titan IVB vehicles have been experiencing much larger than expected buffet loads during transonic flight. To identify the origin of these loads, Aerospace engineer William Engblom conducted a computational fluid dynamics simulation of the transonic flow environment. Approximately 25,000 CPU hours were required to complete the calculations on a four-million-cell mesh over a real time of almost one second. Data-processing resources at the Wright Patterson Air Force Aeronautical Systems Center were key to executing this simulation accurately and quickly. The solution was obtained for the Titan IVB at Mach 0.8. Animations clearly illustrated a new, important fluid dynamic mechanism that is likely responsible for the anomalous pitch accelerations experienced on several Titan IVB vehicles. The mechanism involves strong pairs of vortices (see illustrations), which are alternately shed from the noses of the solid-rocket-motor upgrade boosters at a nearly constant frequency, causing substantial pressure loads on the vehicle surface.
Photo courtesy of NASA
he smallest operational satellites ever flown—built by Aerospace with Defense Advanced Research Projects Agency (DARPA) funding—were selected by Popular Science as one of the top 100 technologies for the year 2000. About the size of cellphones, these picosatellites, or “picosats,” were featured in the magazine’s December 2000 issue in the “Best of What’s New” section. Project director Ernest Robinson of the Aerospace Center for Microtechnology accepted an award for Aerospace at a Popular Science exhibition in New York in November 2000. A pair of these picosats flew a groundbreaking mission in February 2000 with the primary goal of demonstrating the use of miniature satellites in testing DARPA microelectromechanical systems (MEMS).
Raymond de Gaston, left, and Astronaut Frank Culbertson with award.
Map of unsteady pressure loads at Mach 0.8.
“Snapshots” of pressure contours. Pair of vortices (left); new pair of vortices (right) on opposite side of core vehicle (0.035 seconds later).
This newly discovered buffet behavior is common to all multibody vehicle configurations, according to Engblom, and should be considered in the design of future launch vehicles.
award, which was presented to de Gaston by Astronaut Frank Culbertson, recognizes outstanding performance contributing to flight safety or mission success. The award is given annually to less than one percent of the space program workforce and is always presented by an astronaut. De Gaston was recognized for finding a potentially disastrous shortcoming in component quality for the station’s direct-current-todirect-current converter units. These key components alter voltage generated by photoelectric arrays and provide all usable electrical power to the station. Corrections to Crosslink, Summer 2000 A news brief in Headlines incorrectly stated that the GPS industry is expected to grow to $16 million in the next three years. The number should have been $16 billion. The caption of a photo of a Titan IVA rocket at the beginning of the article “Aerospace Photos Capture Launch Clouds” incorrectly identified the rocket as a Titan IVB.
Crosslink Winter 2000/2001 • 3
Concurrent Design at Aerospace Engineers and customers work together to design new space systems in a setting that accelerates the development process. Real-time interaction between specialists is the key.
magine two engineers, each designing the thermal-control subsystem for a new satellite requested by a prospective commercial customer. Both are experienced and highly skilled; both have good tools at their disposal. They’re trying to accomplish the same objective, but they’re required to work in different environments, under vastly different circumstances. Consider for a moment the striking contrast in how they complete their tasks.
Technical specialists confer about a study in progress at a Concept Design Center facility. Study lead Ronald Bywater, standing, is discussing a design issue with cost specialist Vincent Canales, left, and communication specialist John O’Donnell, right.
Scenario #1: The first engineer puts the finishing touches on his design. A week later, he attends a program team meeting with representatives from all involved subsystem disciplines and learns that while his design is impressive, it’s too heavy. He goes back to the drawing board. In a couple of weeks, he’s got a new design that’s lighter. His colleagues give it a thumbs-up. But after a few days, the team learns that the customer has changed her mind about the payload performance requirements, so the entire mass budget has changed. It’s back to the drawing board again. Scenario #2: The other engineer puts the finishing touches on her design. She keys some values—power and thermal requirements, mass, and so on—into an electronic spreadsheet. She immediately receives feedback from the design lead, who tells her she’s a bit over the mass budget. After 15 minutes of reviewing, recalculating, and consulting with the customer, the engineer makes a small change to the design. It’s now within the mass budget. And the customer is pleased, to boot. It should be obvious that the second scenario has significant advantages. Realtime interaction between specialists enables an accurate dialogue that resolves issues right away. With concurrent designing, a study can be completed in hours instead of months. And getting everyone— including the customer—together in the same place not only speeds up the process but also affords participants the ability to clear up misunderstandings with face-toface communication. This scenario isn’t just a fantasy; it’s the way conceptual design studies are now being conducted at an innovative facility operating successfully at The Aerospace Corporation: the Concept Design Center (CDC).
Patrick L. Smith, Andrew B. Dawdy, Thomas W. Trafton, Rhoda G. Novak, and Stephen P. Presley CDC provides the opportunity for Aerospace customers, both government and commercial, to work directly with corporate engineering experts on the rapid development of conceptual designs for new space systems. Linked software models and a computer-aided design system for instant visualization of subsystems provide the concurrent design capability that characterizes CDC and makes it a potent facility for the development of new systems. The Intent of Conceptual Design A conceptual design study is a quick look at what is feasible to build and how much it could cost. The intent is to gain highlevel insight into a project’s scope, not determine the precise value of each design parameter. In a project at the conceptual stage, requirements are not yet well defined; detailed specifications are not locked in. Participants want to explore “What if…?” scenarios, changing a parameter here or there just to see what happens. Many, if not most, proposals for new missions never go beyond the conceptualstudy stage. Usually the mission cost turns out to be too high, or the study exposes a technical Achilles’ heel. Conceptual design studies are also useful for evaluating costs and benefits of new technologies (e.g., advanced solar cells, miniature sensors, inflatable structures) and for teaching the principles of spacesystems engineering. To get a feel for the questions that a study will answer, consider the example of a proposed mission to detect forest fires from space. What is the size of the smallest fire that the spacecraft must be able to detect? What types of sensors can be used? Who needs the data? How quickly must it be obtained? How many spacecraft are required? How much will the mission cost? Conceptual design studies are not new; they’ve always been part of the systemdevelopment process. To see how conceptual design has evolved into the sophisticated set of techniques now employed by CDC, consider how it was conducted in the early days of the space age.
Crosslink Winter 2000/2001 • 5
Conceptual Design at Aerospace: The Early Days In the 1960s and 1970s, conceptual design studies were performed by loosely organized teams of subsystem specialists. Such studies could take months. In the early days of the space age, technologies were new, and spacecraft-design methodologies and tools were still evolving. The personalcomputer era had not yet arrived, so designers had to develop computer programs that ran overnight on mainframes. And without computer-aided drawing software, they had to use manual drafting techniques to lay out spacecraft configurations. Like most companies in the space industry, Aerospace subdivided its engineering division into departments such as thermal, propulsion, structure, and cost. A spacesystem conceptual design study might draw upon expertise from a dozen or more of these specialty areas. The study leader recruited specialists directly through personal contacts or through department managers. Interpersonal relationships, a critical factor in the success of any team effort, were unpredictable; participants might or might not work together smoothly. A conceptual design study during this period was a sequential process, usually driven by the customer’s schedule. Team members would meet periodically as a group, perhaps weekly, to coordinate design details but otherwise would work alone and independently. Most studies were poorly documented—funding often ran out before reports could be prepared. A follow-on study would have little to build upon. The Space Planners Guide, published by Air Force Systems Command in 1965, was the first comprehensive reference source for the conceptual design of space systems. Engineers from Aerospace contributed much of the technical information published in the Guide, including pre-computer-age nomographs for orbit analysis and traditional (hard-copy) spreadsheets for cost estimation. The Guide was widely used for several years throughout the space industry, in both civilian and military space programs. Attempts in the 1980s and early 1990s to use computers to automate conceptual design studies largely failed. Researchers tried to capture each subsystem specialist’s knowledge in the form of rules of thumb and parametric sizing formulas, with the
6 • Crosslink Winter 2000/2001
Design information is passed among the team members using linked spreadsheet files such as the one shown here. The design process repeats until all designers are satisfied that their subsystems meet the requirements.
ambitious goal of optimizing design trades and costs. Several attempts to create a program for this purpose met with limited success. Subsystem specialists had said all along that automating conceptual design of spacecraft would be extremely difficult, if not impossible, because their knowledge and skills could not be fully captured in computer code. More successful efforts to automate conceptual design studies focused on computer-aided approaches that were less ambitious. One such effort was a program based on Mechanical Advantage, a commercial software product from Cognition Corporation that is basically an equation solver linked to a graphics program. While the Cognition application had some utility for certain aspects of conceptual spacecraft design, the program, which ran on powerful and (at the time) scarce workstations, was too limited for wide use. Users needed extensive training. Some spacecraft design models developed for the Cognition application, however, later became the basis for some of the subsystem spreadsheet models used in CDC today.
The Original Models With the proliferation of personal computers and the advent of powerful spreadsheet software in the early 1990s, more practical interactive approaches to computer-aided conceptual spacecraft design emerged. Spreadsheet sizing models were developed that linked mass, power, and other characteristics of various spacecraft subsystems, so that changing the design of one subsystem would have immediate impact on designs of the others. The original collection of spacecraft-subsystem-design spreadsheets developed by Aerospace proved very useful in conceptual design studies, but subsystem experts still needed to carefully check the spreadsheet outputs in order to ensure that a particular design did not exceed the limits of the spreadsheet models. Developing the spacecraft-subsystemdesign spreadsheets taught a valuable lesson to those working to build computer aids for conceptual design. System engineers were concerned that the subsystem specialists might be left out of the design process and that the models could be misapplied or give misleading results. Some
of the specialists were even reluctant to develop simplified models for the spreadsheets because they felt they could not guarantee the correctness of the models’ results in every context in which they might be used. In 1994, NASA’s Jet Propulsion Laboratory (JPL) asked Aerospace to adapt the spreadsheet models for the Advanced Projects Design Team, also known as Team X, in JPL’s Project Design Center. Team X’s job was to write proposals for “faster-better-cheaper” planetary exploration missions. The center was being designed as a facility where teams of JPL engineers could work together concurrently to rapidly design spacecraft for NASA’s planetary and other space-science missions. JPL needed to find a method for linking spacecraft-subsystem design models so that information on the different elements in a project (e.g., spacecraft, cost, operations) could be shared concurrently and archived for follow-on Team X studies. Two Aerospace engineers, Joseph Aguilar and Glenn Law, tried to adapt the spreadsheet models for use in the JPL design center. But they found the models were difficult to use in environments where team members worked on separate design elements at the same time. They then undertook the task of developing the computer network and interfaces that would allow the subsystem models to run on different workstations at the same time. Using the “distributed” version of the spreadsheet models, Team X eventually reduced its cost to produce a proposal from $250,000 to $80,000 and cut the time required from 26 to 2 weeks. Team X previously produced only about 10 proposals per year; it now produces 45. After this success, Aguilar, Law, and their colleague Andrew Dawdy proposed developing a similar concurrent design capability at Aerospace, geared to the conceptual design of military and commercial space missions. In the fall of 1996, management approved their independent research and development proposal for what was to become CDC. CDC Takes Shape The three Aerospace engineers spent a year linking new versions of spacecraftsubsystem spreadsheet models that were developed by subsystem experts. The development of a set of spreadsheet models to support fast-paced collaborative spacecraft design not only required experienced
CDC Teams Team
Space Segment • Payload and spacecraft subsystem design • Top-level ground segment and software sizing • Parametric cost and performance estimation
Astrodynamics, payload, command and data handling, communications, attitude determination and control, electrical power, propulsion, structure, thermal, ground, software, cost, configuration
• Constellation design and coverage analysis • Vehicle size and launch manifest determination • Concept of operations and ground segment architecture • Analysis of relative cost vs. requirements
Constellation, payload, spacecraft, availability, ground, cost, utility
• Payload and spacecraft subsystem design/analysis • Constellation and endgame performance estimation • Parametric system cost analysis
Configuration, sensor, ordnance, guidance, power, propulsion, structure, ground, carrier vehicle, cost
• Flexible, fast response, reusable launch vehicle (requires boost vehicle) • On-orbit payload support subsystem design and cost analysis
Command and data handling, communications, attitude determination and control, electrical power system, propulsion, thermal, ground, software, vehicle configuration, cost
• Estimation of facilities, personnel, processing, communications, and software • Determination of information, functional, and communication architecture trades
Information architecture, communications architecture, software, processing, staffing, facilities, spacecraft, cost
• Detailed communications payload subsystem design and trades • Top-level spacecraft estimation • Performance, hardware configuration, and cost estimation
Link analysis, radiofrequency units, modulation analysis, orbital analysis, digital, coding, laser, spacecraft
Kinetic Energy Weapons
Space Maneuver Vehicle
Crosslink Winter 2000/2001 • 7
CDC Successes CDC conducts studies for a variety of organizations, and the studies span a broad range of programs. Congressional staffers and senior Department of Defense officials suggested that the Space Based Infrared System Low mission could be combined with another program on the same bus. A quick study by the CDC System Architecture Team showed, however, that the total cost of the combined missions would exceed the costs of keeping them separate. A CDC study of the Communication/Navigation Outage Forecast System showed that proposed mission requirements would have to be reduced in order to meet cost and risk constraints. As a result, the Air Force Space Test and Experimentation Program Office rewrote the mission and technical requirements. The program recently received funding, and CDC team members participated in source selection. Subjects of CDC design studies for the Air Force Space and Missile Systems Center have included the Space Based Radar and the Global Multi-Mission Space Platform. The requirements for contractor feasibility studies for the space platform are based directly on results from the CDC study. The CDC Ground Systems Team designed several architectures for the Air Force Satellite Control Network. The network needed to compare its present acquisition plans to alternatives proposed by outside organizations. One alternative proposed using commercial service providers for most ground-to-space communications; another proposed moving most of the network’s communications to spacebased relays. CDC studies helped the customer understand key features and drawbacks of the alternatives. The Naval Postgraduate School in Monterey, California, has asked Aerospace to help develop a CDC-like design center as a learning tool for use by graduate students.
8 • Crosslink Winter 2000/2001
engineering judgment but also entailed very careful interface design so that the specialized subsystem spreadsheets could be appropriately linked. Using lessons learned at JPL, Aguilar, Law, and Dawdy carefully explained the concurrent design approach to other Aerospace engineers and to potential customers, whose acceptance was essential for the project to succeed. Recruiting the first design-team members, who would have to work together in a new type of environment, was initially a challenge. Fast-paced work on systemlevel concurrent design teams would be something new for Aerospace technical experts. Fortunately, the engineers recruited for the first CDC team not only possessed the required expertise but also were enthusiastic about trying this type of work. They seized the opportunity to apply their design skills in the new concurrentdesign environment. Working through some start-up problems, the engineers soon developed a strong team spirit that would prove essential in resolving technical and administrative problems. By the second year of the independent research and development effort, the original spacecraft design team had completed seven design studies and had received awards and recognition for its efforts. Word of CDC’s successes spread, and recruiting new team members became much easier. Today, more than 100 Aerospace engineers participate on CDC teams, working
in two dedicated facilities (unclassified and classified). A new Aerospace organization, the CDC Office, coordinates the center’s activities. Six teams currently make up CDC: • Space Segment Team, the original CDC team, focuses on the space vehicle (bus) segment. Each member designs a particular spacecraft subsystem and specifies the elements at the part level. Computer-aided-drawing layouts are used to visualize physical relationships among the subsystems. • Systems Architecture Team considers all of the space-system segments (space, ground, and launch). The level of detail does not extend below toplevel descriptions of each segment and their interactions—the minimum needed to understand the broad architecture trades. • Communications Payload Team focuses on communications subsystems at the part level. This team is in development. • Ground Systems Team examines elements of the ground segment of space systems, including facilities, staffing, software, communications, and processing equipment. • Kinetic Energy Weapons Team performs top-level design of space-based ballistic-missile interceptors. The team is similar to the Space Segment Team but uses a different set of performance metrics and technologies.
A dedicated facility for conducting design sessions. The configuration of workstations promotes face-to-face interaction between team members. The customer team sits at the center table. Overhead projectors can display any team member’s monitor. Video teleconferencing cameras are located at the front and back walls.
• Space Maneuver Vehicle Team is also similar to the Space Segment Team but focuses on the requirements of launch, orbital operations, reentry, and reuse. These teams follow the same basic guidelines and procedures that were established for the initial CDC spacecraft team—the use of well-defined processes, cross-department communication and teaming, ownership of models and technical data by engineering experts, and direct customer involvement during the design sessions. CDC Studies A typical CDC study takes about six weeks and requires about 300 to 500 staffhours of effort, depending on the amount of up-front preparation required and the scope of the study. Studies are conducted in three phases: presession preparation, design sessions, and postsession wrap-up. In the presession preparation phase, several meetings with the customer define the design trades to be performed. Team members often have to research new technologies and modify their models to handle unique features of the proposed concept. A formal proposal that lays out the objectives, schedule, and cost of a study is always provided to the customer before work starts.
The actual design sessions that are the heart of the CDC process take place in one of the dedicated facilities. Team members and customers work together in concurrent design sessions that last from two to four hours. During each session, they explore alternative approaches and gain insights into the design drivers. Working together in one room with the right tools and procedures vastly reduces the time required to complete a study and enables the design team to address customer questions and smoothly accept redirection from the customer if it becomes necessary. Two to five sessions spread over a week or two are usually needed to complete a study. The focus of the final phase, postsession wrap-up, is the creation of a report documenting the study. This report is published within three or four weeks after completion of the design sessions. The customer usually contributes a section describing the mission. Each participant in a CDC study has a specific role. In addition to the participants who bring their technical expertise to a project, some team members must exercise critical administrative skills to move the
study forward. Among the most important roles are the facilitator and the study lead. It is the facilitator’s responsibility to keep all hardware and software, including the computer network, up and running and to quickly resolve any interface problems that arise with the spreadsheet models. The number of people involved in a study and the rapid pace of the sessions make it essential that all supporting equipment and software perform reliably. The facilitator is also involved in training new team members to be effective participants in the CDC process. The study lead guides the customer and the technical experts through each step in the CDC process. It is the lead’s responsibility to ensure that customer expectations are realistic and are met. The customer must understand what he or she will get out of the CDC process, how it works, what the customer’s role in the process is, and what the team needs from the customer to do its work. The customer is the focus of everyone’s attention. Customers for CDC studies have included both military programs and commercial companies, but CDC also serves
Aerospace specialists use linked spreadsheets during a design session. Left to right: Christopher Taylor, Eric Hall, Mark Mueller, Douglas Daughaday.
Crosslink Winter 2000/2001 • 9
Aerospace Systems Architecting and Engineering Certificate Program Courses for Technical Teams The teamwork skills critical to CDC’s concurrent engineering process apply to all phases of system design, development, and deployment. Teams can bring to bear expertise from multiple mission partners and stakeholders to ensure clearer, more direct communication of varying points of view, resulting in the rapid development of consensus solutions. The Aerospace Institute has made teamwork skills an integral part of its offerings in both technical and business areas. In the Institute’s Technical Education and Development curriculum, the three-day course “Teaming for Systems Architects and Systems Engineers” serves as a knowledge-building component of the Aerospace Systems Architecting and Engineering Certificate Program (see page 59). This course reinforces team concepts through case studies and simulations. Topics include problem definition, communications, decision making, conflict management, leadership, and cross-functional team effectiveness. The concurrent engineering methodology critical to CDC is highlighted in the Institute’s Space Systems Design course. Part of the Space Systems Engineering series, this course describes the space systems design process, how design fits into the space-mission timeline, and how requirements flow between the vehicle payload, spacecraft, and ground system. The focus is on interactions and dependencies between subsystems and the relationships between subsystems and system elements. The student becomes familiar with methods and tools used in the space-vehicle conceptual design process that CDC employs.
10 • Crosslink Winter 2000/2001
This sequence of images illustrates how a conceptual design evolves during the course of a CDC session. A performance analysis indicated that five was the optimum number of ballistic-missile interceptors per spacecraft carrier. An initial layout of the carrier was developed that was compatible with the Delta II 3-meter launch-vehicle fairing. But further design iterations revealed the need for additional bus surface area to support larger solar panels. The resulting configuration proved well suited for manifesting four carrier vehicles on a larger Delta IV 4.5-meter fairing.
internal corporate customers, performing design studies for programs within Aerospace. Direct customer involvement in each step of a study is essential. It is the customer’s responsibility to define the trade space to be explored and to explain the big-picture context of the study to the design team. The Design Session A design session begins with team members arriving at one of the CDC facilities. The facilitator prepares by powering up the computers, video equipment, and audio systems. Participants take their places in front of their workstations and log on. Designers check over programs and data structures—software that could include, for example, a sensor database, a cost model, a computer-aided drafting program. The customer describes the objectives of the study, which may include development of a baseline design and cost estimate as well as the identification of cost drivers and areas of greatest technical risk. Then the study lead distributes a list of design options that had been developed in the presession planning meetings and other pertinent handouts—for example, a data sheet that describes the power profile for the mission, the payload operations requirements, and the technology freeze date
to be assumed in the study. The study lead moderates a brief discussion to ensure that everyone understands the objectives. The facilitator initializes the system parameters in each team member’s subsystem model, and team members begin working on their designs. The facilitator coordinates the flow of data among the models and periodically updates the master list of design options with the latest design parameters. As team members adjust their subsystem parameters, they exchange ideas about design issues with their teammates and the customer. They use parametric cost models (cost-estimating relationships, equations that predict cost as a function of one or more drivers) and many other parameters (mass, performance, etc.) to compare different designs. The biggest challenge for a CDC team is to come up with a first viable design; subsequent designs are usually easier, often just excursions from the baseline. Design issues surface as work proceeds. Discussions take place in side sessions where engineers try to resolve problems without full team involvement. Some team members might have to spend some time researching new design approaches or technologies. When it becomes necessary, the facilitator displays an individual’s monitor on a large screen for everyone to
Hundreds of interceptor “garages” like the one shown here would orbit Earth as part of a national defense strategy; upon detection of a hostile missile launch, interceptors would be fired to track down and destroy the missile. This proposed system is the product of a conceptual design study performed by the Concept Design Center.
see the subject of discussion. And at some points, the customer may be required to choose between several design options before the study can progress. Team members are given the opportunity to explain subsystem design issues so that the entire team understands how the design has evolved. The process continues, with continual redefinition and reevaluation of designs. As the design session winds up, the study lead discusses possible next steps with the customer and begins collecting data for the final report. Conclusion Thanks to CDC, the Aerospace role in front-end engineering and architecture studies has become more visible. CDC’s success has ensured that the company is widely recognized as a leader in up-front planning and technology development for new space systems. CDC has become an essential part of the systems engineering support that Aerospace provides. Six teams currently perform a total of about 12 to 18 conceptual studies per year. CDC has become largely
self-supporting, with most of its funding coming directly from customer studies. CDC teams and applications continue to proliferate. Planned future enhancements include increased contractor involvement, more powerful three-dimensional modeling and visualization capabilities, and geographically distributed design teams connected via the Internet. The basic principles that have guided CDC in its development have not changed since its origins—reliance on documented processes, cooperation between disciplines, and partnering with customers. These principles, which clearly are applicable to other corporate initiatives, such as mission assurance teams and information networking, have made CDC a resounding success thus far and will no doubt serve as an excellent foundation for its future development. Further Reading J. A. Aguilar and A. B. Dawdy, “Scope vs. Detail: The Teams of the Concept Design Center,” 2000 IEEE Aerospace Conference Proceedings (March 18–25, 2000).
J. A. Aguilar, A. B. Dawdy, and G. W. Law, “The Aerospace Corporation’s Concept Design Center,” 8th Annual International Symposium of the International Council on Systems Engineering (July 26–30, 1998). Capt. A. Bartolome, USAF, S. S. Gustafson, and S. P. Presley, “Concept Design Center Teams Explore Future Space-Based Tools,” Signal (July 2000). A. B. Dawdy, R. Oberto, and J. C. Heim, “An Application of Distributed Collaborative Engineering,” 13th International Conference on Systems Engineering (August 9–12, 1999). R. Novak, “Systems Architecture: The Concept Design Center’s Ground System Team—A Work in Progress,” 13th International Conference on Systems Engineering (August 9–12, 1999). S. L. Paige, “Solar Storm Sat: Predicting Space Weather,” 2000 IEEE Aerospace Conference Proceedings (March 18–25, 2000). S. P. Presley and J. M. Neff, “Implementing a Concurrent Design Process: The Human Element Is the Most Powerful Part of the System,” 2000 IEEE Aerospace Conference Proceedings (March 18–25, 2000).
Crosslink Winter 2000/2001 • 11
Estimating Probable System Cost Stephen A. Book
n estimating the cost of a proposed space system, cost analysts follow the adage “What’s past is prologue.” They use costs of existing systems to develop a cost-estimating relationship (CER), which can help predict the cost of a new system. The foundation of modern cost analysis, the CER is usually expressed as a linear or curvilinear statistical regression equation that predicts cost (the dependent variable) as a function of one or more cost drivers (independent variables). However, the CER tells only part of the story. At the beginning of a cost-estimating task, the analyst identifies the work-breakdown structure (WBS), a list of everything that has to be paid for to bring a system to its full operational capability. A space system’s WBS includes high-level categories such as research, development, and testing; operations, maintenance, and support; production; and launch. Lower levels of the structure include software modules, electronics boxes, and other components. In addition to CERs, the analyst bases estimates on costs of items already developed and produced, on vendor price quotes for offthe-shelf items, and on any other available information that can be used to assign a dollar value to items in the WBS. Until recently, sponsors and managers of space systems have expected cost analysts to provide best estimates of the costs of various options at each project milestone and decision point, from the initial trade-study stage to source selection, right on to project completion. Unfortunately, “best estimate” has never been precisely defined. For example, is it the most likely (most probable) cost, the 50-percent confidence level cost (the dollar figure that is equally likely to be underrun or overrun), or the average cost? It wasn’t clear what useful information about system cost this best estimate was conveying, so the figure proved inadequate for comparing competing options, as well as for planning system budgets. This unsatisfactory situation led the Department of Defense (DOD) to issue formal guidance on how system costs should be expressed and what the terminology should mean.
12 • Crosslink Winter 2000/2001
Basing a system cost estimate on past systems can be tricky. Analysts who do this find that the sum of the most likely costs of the elements of a space system in development does not equal the most likely cost of the entire system. Using probability distributions to treat cost estimation as a statistical process can provide estimates that are much more meaningful. The typical approach to obtaining a best estimate is to view each WBS item’s estimated cost as its most likely cost and then roll up (sum) those estimates to arrive at a total system cost. Because high-end risks outweigh low-end uncertainties, however, a roll-up estimate calculated this way tends to underestimate actual cost by a wide margin, leading to cost overruns attributable purely to the mathematics of the roll-up procedure. In fact, formal mathematical theory confirms that the sum of the most likely WBS-item costs is substantially less than the most likely total cost. Experience shows that assigning a confidence level of 30 percent or less to a roll-up estimate is not overly pessimistic. While generally thought to be a new phenomenon specific to DOD, the difficulties associated with cost-estimating uncertainty were recognized in France as early as 1952. The solution to the problem is to treat the cost-estimating process statistically, a technique known as cost-risk analysis. Even the use of the term “most likely” indicates a statistical situation, because it implies that other, less likely estimates exist. Probability distributions are established to model the cost of each WBS item. (A probability distribution contains the possible values a random variable could take on, as well as the probaThe three elements at the highest level of a bility of it taking on each one space-system work-breakdown structure (WBS) of those values.) Then, correlaare space-resident satellites, a launch system, tions among these distributions and a ground-based control system. In estimating a system’s probable cost, it is important to are estimated, and the distribuconsider how the costs of these elements (XS, tions are summed statistically, XL, XG) are correlated. typically by Monte Carlo sampling. The result is a probability distribution of total system cost, from which one can obtain meaningful estimates of the median (50-percent confidence
level), 70th percentile (70-percent confidence level), and other relevant quantities. The cost-risk-analysis approach yields a mathematically correct most likely cost (one that is precisely defined, along with its level of confidence), as well as costs for all percentiles. Estimates at the 50-percent and 70-percent confidence levels are much more valuable to decision makers in setting program budgets than an essentially meaningless “most likely” estimate. In the early 1990s The Aerospace Corporation developed many of the mathematical procedures now applied to estimate project costs statistically. Once the budget has been established and the system has been in development for a number of months, earned-value data become available. A measure of how much work has been accomplished on a project, earned-value data are typically collected by the contractor for use in comparing actual expenditures on scheduled tasks with the amounts budgeted for them. Over a specific time period, usually a month or three months, and cumulatively since the beginning of development, an earned-value management system tracks and compares three kinds of financial information associated with each WBS item: the budgeted cost of work scheduled to be done on the item during the period (the estimated cost), the budgeted cost of work actually completed on that item during that time (the earned value), and the actual cost incurred, per billing, for the contractor’s work done on the item during that time. Earned-value data can be used to forecast an estimate at completion at any point in the program. Measuring the Quality of a Cost-Estimating Relationship How does a cost analyst derive a CER for a WBS item? Let’s say the item under consideration is a satellite’s solar-array panels. Cost and technical data on solar arrays that have already been produced must be organized into a database. For each array, cost is tracked against appropriate technical characteristics such as weight, area, and storage Crosslink Winter 2000/2001 • 13
capacity. Algebraic relationships between cost and these potential cost drivers are compared to determine which relationship is the optimal predictor of solar-array cost. It is not a priori obvious which criteria are the best to use for assessing CER appropriateness, nor will all cost analysts agree on the best criteria. Three statistical criteria, however, lie at or near the top of almost everyone’s list: • percentage standard error of predictions made by the CER of values in the database: root-mean-square of all percentage errors made in estimating values of the database using the relationship (a one-sigma number that can be used to bound actual cost within an interval surrounding the estimate with some degree of confidence) • net percentage bias of predictions of the values in the database: algebraic sum, including positives and negatives, of all percentage errors made in estimating values of the database using the CER (a measure of how well percentage overestimates and underestimates of database actual costs are balanced) • correlation between estimates and actual costs (CER-based predictions and cost values in the database): If the relationship were a perfect predictor of the actual cost values of elements of the database, a plot of estimates against the actual costs would follow a 45-degree line quite closely (correlation, a statistical measure of the extent of linearity in a relationship between two quantities, would be high if estimates tracked actual values but low if they did not).
Structure and mechanical
Structure Interstage Adapter Thermal
Main propulsion Engines
Instrumentation Guidance and control Data handling Communications Electrical power Electrical wiring Software
Payload fairing Reentry Protection Landing system
A WBS is a hierarchical list of all items that must be paid for to bring a system to its full operational capability. A space system’s WBS includes high-level categories such as launch vehicle, as well as lowlevel items such as engines and software.
Standard error is better expressed in percentage terms than in dollars. Using percentage to express standard error in costestimating offers stability of meaning across a wide range of programs, time periods, and situations. An error of 40 percent, for example, retains its meaning whether the analyst is estimating a $10,000 component or a $10 billion program. Conversely, a $59,425 error is huge when reported in connection with a $10,000 component, but insignificant with respect to a $10 billion program. Even in less extreme cases, a standard error expressed in dollars often makes a CER virtually unusable at the low end of its data range, where relative magnitudes of the estimate and its standard error are inconsistent.
Similarly, in the case of bias, a dollarvalued expression is not as informative as an expression in terms of percentage of the estimate, because a particular dollar amount of bias would not have the same impact on all values in the database. Cost-Risk Analysis “Cost-risk analysis” is the term used by cost analysts for any estimating method that treats WBS-item costs and totalsystem costs as random variables rather than deterministically derived numbers. The term implies that, for any deterministic estimate, there is some degree of risk that the system will be unable to be delivered or meet its stated objectives at that particular funding level. Cost-risk analysis recognizes that a mathematical probability
Statistics at a Glance Probability distribution A function that describes the probabilities of possible outcomes in a “sample space,” which is a set that includes all possible outcomes. Random variable A variable that is itself a function of the result of a statistical experiment in which each outcome has a definite probability of occurrence. Determinism The theory that phenomena are causally determined by preceding events or natural laws; hence in this context, a "deterministically derived number" can be a system cost value derived from the cost of an earlier system. Standard deviation (sigma value) An index that characterizes the dispersion among the values in a population; it is the most commonly used measure of the spread of a series of values from their mean.
14 • Crosslink Winter 2000/2001
Bias The expected deviation of the expected value of a statistical estimate from the quantity it estimates. Correlation A measure of the joint impact of two variables upon each other that reflects the simultaneous variation of quantities. Percentile A value on a scale of 100 indicating the percent of a distribution that is equal to or below it. Monte Carlo sampling A modeling technique that employs random sampling to simulate a population being studied. Root-mean-square The square root of the arithmetic mean of the squares of a set of numbers; a single number that summarizes the overall error of an estimate.
of success is associated with each deterministic cost estimate. Are costs really random? The formal framework for working with a range of possible numeric values is the probability distribution, the mathematical signature of a random variable. Modeling costs as random variables does not imply they are random. It reflects how an item’s cost results from a large number of very small influences, whose individual contributions we cannot investigate in enough detail to precisely calculate the total cost. It is more efficient to recognize that virtually all contributors to cost are uncertain and to find a way to assign probabilities to various possible ranges. Consider coin tossing. In theory, if we knew all the physics involved, we could predict with certainty whether a coin would land heads or tails; however, the influences acting on the coin are too complex for us to understand in enough detail to calculate the parameters of the coin’s motion. So, instead, we bet that the uncertainties will average out in such a way that the coin will land heads half the time and tails the other half. It is more efficient to model the physical process of coin tossing—which is in fact deterministic—as if it were a random statistical process and to assign probabilities of 0.50 to each of the possible outcomes. System cost can be similarly represented as a random variable rather than a fixed number, because cost is composed of
many very small pieces, whose individual contributions to the whole cannot be specified in sufficient detail for precise estimation of the whole. Standard accounting technique considers total system cost to be the sum of the costs of all WBS items, so it first requires us to estimate the most likely cost (the mode, the most frequently occurring value) of each item and then to sum those costs. While this procedure seems reasonable, the roll-up is almost sure to be very different from the actual most likely value of the total cost. Statistical theory shows that the sum of modes does not generally equal the mode of the sum. Because of the preponderance of high-end risks (with more probability lying above the best estimate than below it), most cost probability distributions are not symmetric about their modes. As a result, the sum of the modes is usually considerably smaller than the mode of the sum. The practical impact of this is that estimates obtained by totaling most likely costs of WBS items tend to significantly underestimate actual cost. Examples are available to illustrate what kinds of errors might occur in typical cases if these mathematical peculiarities are ignored. They show that a roll-up estimate typically has a probability of 30 percent or less of being sufficient to fund the program. Here as elsewhere, shortcutting proper statistical procedures leads to erratic, unpredictable results.
Cost Correlation Between Items Correlation among WBS-item cost distributions contributes significantly to the uncertainty of the total system cost; this has been realized since the probabilistic approach to cost analysis became de rigueur over the past decade in response to DOD-issued requests for proposals. Understanding the degree of uncertainty in an estimate is a necessary aspect of cost analysis, one that is paralleled by representing system cost as a random variable for the purpose of appropriately modeling uncertainty. Because risks faced in working on different WBS items are often correlated, ignoring correlation in statistical computations makes the spread of the cost distribution narrower than it should be. Failing to account for correlation therefore deceives the analyst by making an estimate appear less uncertain than it really is. The most universal statistical descriptor of a random variable’s degree of uncertainty (spread) is its standard deviation (sigma value), and pairwise correlations between random variables are significant contributors to the magnitude of the sigma value of their sum. This makes correlation between program-item costs a critical factor in the estimation of total system cost uncertainty. Cost estimating would be much simpler if there were no interelement cost correlations; unfortunately, this is not the case. Consider the example of a space system. At the highest level of the WBS, there are three elements: space-resident satellites, a launch system, and a ground-based control and data-analysis system. The respective
Cost Estimating at a Glance Cost-estimating relationship (CER) A mathematical equation that predicts cost as a function of one or more drivers.
Net percentage bias Algebraic sum of all percentage errors made in estimating values of the database using the CER.
Work-breakdown structure (WBS) A hierarchical list of everything necessary to bring a system to its full operational capability.
Most likely cost The mode, the most probable value.
Cost-risk analysis An estimating method that treats WBS item costs and totalsystem costs as random variables rather than deterministically derived numbers. Earned-value management A set of procedures used to track program expenditures and their relationship to the amount of work that has been accomplished. Percentage standard error Root-mean-square of all percentage errors made in estimating values of the database using the CER.
Estimate at completion (EAC) An estimate of what the final cost of a program will actually be. Cost-performance index (CPI) A measure of the efficiency at which dollars are being spent on a project. Schedule-performance index (SPI) A measure of the rate at which work is being completed on a project. Schedule-cost index The product of the cost-performance index and the scheduleperformance index.
Crosslink Winter 2000/2001 • 15
30 60 10 Number of WBS items in the roll-up
0.4 0.6 Actual correlation
This graph illustrates the importance of working with the numeric correlations between WBS items. Assuming these correlations to be zero causes a detrimental effect on the estimation of total-cost uncertainty. Shown is the percentage by which the sigma value (standard deviation) of the total-cost distribution is underestimated, assuming WBS interelement correlations to be zero instead of the actual value (usually represented by ρ, the Greek letter rho). The horizontal axis tracks ρ, and the vertical axis tracks the percentage by which the total-cost sigma value is, for each nonzero correlation value, underestimated if the correlations are instead assumed to be zero. Each curve is keyed to a unique value of n, the number of elements in a roll-up. As the four curves show, the percent by which sigma is underestimated also depends on the number of WBS items for which the pairwise correlations are incorrectly assumed to be zero. For example, if n = 30 WBS items, and all correlations between WBS items (ρ) are 0.2, but the estimator assumes they are all zero, the total-cost sigma values would be underestimated by about 60%. (This is meant to be a generic illustration and therefore is only approximately true in any specific case. It has been assumed that the sigma values for the WBS items are the same throughout the entire structure.)
On the other hand, XS, XL, and XG may be negatively correlated for different reasons. Reducing the complexity of onboard satellite software and communications hardware may increase ground costs by complicating the ground software while, at the same time, decreasing launch costs as a result of reduction in size of on-orbit hardware. Further down into the WBS, costs are more highly correlated because they correspond to specific items that physically occupy adjacent locations within the
costs XS, XL, and XG may be positively correlated for several reasons. An increase in size, weight, and number of satellites to be placed in orbit results in an increase in launch costs, either through the number of launches required or the needed capability of the individual launch vehicles. An increase in number and data-gathering capability of the satellites forces an increase in ground-operations costs, either through the complexity of the tasking and control system software or the number and size of ground-station facilities.
L M Optimistic Most likely
Sample triangular probability distribution of WBS-item cost, based on estimation of optimistic, most likely, and worst-case costs. All statistical properties of the triangular distribution are determined by the lowest possible cost L, the most likely cost M, and highest possible cost H.
16 • Crosslink Winter 2000/2001
satellites or ground stations or to software packages that operate specific pieces of hardware. In any particular case, the actual interelement correlations have to be estimated along with the element costs themselves. While it may be difficult to justify use of a specific numeric value to represent the correlation between two WBS-item costs, it is important to avoid the temptation to omit the correlation altogether when a precise value for it cannot be established. Such an omission will set the correlation in question to the exact value of zero, whereas positive values of the correlation coefficient tend to widen the total-cost probability distribution and thus increase the gap between a specific cost percentile (e.g., 70 percent) and the best-estimate cost. Therefore, using reasonable nonzero values, such as 0.2 or 0.3, generally leads to a more realistic representation of totalcost uncertainty. Having been the first organization to vigorously advocate in Washington in the early 1990s that account be taken of the significant impact that correlation exerts on cost estimating, Aerospace has developed most of the practical methods currently in use for dealing with correlation-related issues. Obtaining Cost Percentiles Cost-risk analysis comprises a series of engineering assessments and mathematical techniques whose joint goal is to measure the degree of confidence attached to any particular estimate of system cost. A three-step procedure built upon results of a technical-risk study typically forms the cost-risk analysis. First, an engineering assessment of the technologies involved in each subsystem leads to probability distributions of subsystem costs. Second, these subsystem cost distributions are correlated and combined to generate a cumulative distribution of total system cost. Finally, once the cumulative distribution has been established, the 50th, 70th, 90th, and other cost percentiles can be read off the graph. Based on engineering assessments, the cost-estimation process is carried out by assigning low, best, and high cost estimates to each item in the WBS. These three estimates define a triangular probability distribution. All statistical properties of this distribution are uniquely determined by three parameters: the lowest possible cost L, the most likely cost M, and the highest possible cost H. The low estimate specifies subsystem cost under the
most optimistic assumptions concerning development and production capabilities. The most likely estimate is typically derived from the output of a CER-based cost model or other appropriate estimating procedure such as analogy or engineering buildup. The high-end cost encompasses the impacts of all technical risks faced in developing and producing the subsystem. Translating qualitative and quantitative assessments of technical risk into dollars to determine a realistic high-end cost typically requires extensive interactions of cost analysts with engineers knowledgeable in the technical state of the art. System-design concepts usually contain physical descriptions or lists of engineering requirements that may be translatable into dollars. Such translations may be derived from knowledge of technical precedents on which the concept is based. Alternatively, they may be derived from the fact that such precedents are lacking or have not been successfully pursued in the past, despite expenditures of known amounts of funds. Technical complexity of unproved methods for implementation are assessed using scales of increasing difficulty, complexity, or uncertainty analogous to related events in the historical record. These technicalrisk measurements may then be translated into cost risks, based on analogous cost experience or state-of-the-art relationships among cost, technical difficulty, and pace of development. Management and control of costs may then be implemented in accordance with a realistic understanding of the primary source of risk, namely technical difficulty. After probability distributions of individual WBS-item costs have been established, the next step is Monte Carlo random sampling from each subsystem’s distribution and combining these random numbers in a logical way to produce a representation of the cumulative distribution of total system cost. The ultimate objective of cost-risk analysis, the ability to read off percentiles of total system cost, is thus achieved. Trade Studies and Source Selections For programs in progress, probabilistic information allows budget planning to be based on the likelihood that any proposed dollar amount will be adequate to fund the program. Prior to formal program initiation, trade studies are typically undertaken to find out whether a certain type of system is feasible from the operational and cost
WBS-item triangular cost distributions Merge WBS-item cost distributions into total-cost normal distribution Most likely
Cost (dollars) Roll-up of most likely WBS-item costs Most likely
Most likely total cost
Once the cost analyst has established probability distributions of individual WBS-item costs, then Monte Carlo sampling is carried out from each subsystem’s distribution. The random numbers thus generated are combined in a logical way to produce a representation of the cumulative distribution of total-system cost.
points of view. Additionally, source selections are conducted to evaluate system approaches to a problem proposed by different contractors (under acquisition reform, the program for improving and accelerating government contracting and procurement while reducing costs, the approaches may possibly even meet different sets of requirements). Timeliness and simplicity are key requirements of analyses undertaken in support of trade studies and source selections
because not much technical detail is available about the system under study during either phase. In both cases, a decision has to be made by someone who has a very limited factual database. For trade studies and source selections, probabilistic information allows candidate systems to be compared on a level playing field; the goahead decision (in a trade study) or contract award (in a source selection) can be made on the basis of, say, the 50th percentile cost of each candidate.
2000 Mode 3000 Median
3674 Mean 50%
Cost (dollars) Sample lognormal probability distribution, a distribution in which the natural log of a random variable is normally distributed. Unique mathematical characteristics of the triangular and lognormal probability distributions make them both especially applicable to cost analysis at the trade-study and sourceselection stages. A random variable has a lognormal probability distribution if its natural logarithm has a normal probability distribution; a normal distribution is the familiar bell curve, a continuous distribution that is symmetric about its mean. The asymmetrical lognormal distribution, a good model for the statistical sum of a number of triangular distributions, is an excellent choice for representing total-cost distributions of systems that are to be compared on the basis of their relative cost. This is because the ratio of two independent lognormals is itself lognormally distributed.
Crosslink Winter 2000/2001 • 17
System test, launch, and operations
System verified by successful mission
System flight-qualified through test
System prototype demonstrated in space environment
System demonstrated in relevant environment (ground or space)
Component and/or breadboard validation in a relevant environment
Components validated in laboratory
Analytical and experimental critical function, characteristic proof-of-concept
Technology concept and application formulated
Basic principles observed and reported
Basic technology research
NASA technology readiness level (TRL) scale. This tool provides information for determining the worstcase high-end cost of a WBS item. For each spacecraft, aircraft, or payload subsystem under study, an engineer assigns one of the TRL indexes to that subsystem, then compares that index with the indexes of each item in the database that supports the cost estimate. The engineer can then derive the triangular distribution of cost for that subsystem from the relationship between the average TRL index of the database and the TRL index of the subsystem under study. Each time a new program’s cost is estimated, its TRL level will likely be higher than what it had been, if progress is being made. When a new program is at a lower level than the database, its cost will be more uncertain and its cost probability distribution will tend to range far from the estimate derived from the database. As work proceeds, though, the new program’s TRL level should eventually exceed the average level of the database. Then its cost will be less uncertain, so its cost probability will be concentrated somewhat closer to the database estimate.
But a nagging question remains: “What if System A, the lower-cost option at the 50-percent confidence level, faces risk issues that make its 70th-percentile cost higher than that of System B?” In other words, System B would be the lower-cost option if the cost comparison were made at the 70-percent confidence level, while System A would be the lower-cost option if the cost comparison were made at the 50-percent level. In this classic situation, the decision maker has to choose between a low-cost, high-risk option and a highcost, low-risk option. To take account of all possible risk scenarios, the decision maker can make use of all cost percentiles simultaneously, namely the entire cost probability distribution of each candidate system 18 • Crosslink Winter 2000/2001
(which reflects the candidate system’s entire risk profile), not simply the 50-percent or the 70-percent confidence cost. As it turns out, the expression of system cost in terms of a probability distribution makes it possible to estimate the probability that System A will be less costly than System B, and that probability can be part of the basis on which the decision is made. Learning Rates Standard cost-estimating practice involves the application of a cost-improvement factor, or learning rate, to account for management, engineering, and/or production improvements that save money as successive units are produced. Lack of credible analogous or applicable historical data, however, makes it difficult or even impossible to
determine in advance exactly what an accurate learning rate will be in any particular estimating context. Nevertheless, the estimator's choice of learning rate exerts a major impact on the estimate of the total spending profile of a large production program. Even if nonrecurring and firstunit production costs are estimated precisely, small variations in the learning rate will substantially outweigh all other contributions to uncertainty in the total system estimate. This is especially true in cases of largequantity procurements, such as aircraft, launch vehicles, or proposed constellations of numerous satellites. In the case of 50 units, for example, the average-unitcost estimate will be 46.6 percent lower at an 85-percent learning rate, compared with what it would be at a 95-percent learning rate. For 200 units, the estimate will be 57.3 percent lower at an 85-percent learning rate versus a 95-percent learning rate. In the case of 5000 units, the estimate will be 74.5 percent lower at an 85-percent versus a 95-percent learning rate. The learning rate should therefore be treated as another source of cost risk, with optimistic, most likely, and pessimistic learning rates factored into the total system cost probability distribution. Allocating Risk Dollars for Reserve Because users of common estimating methods often underestimate actual project cost, a management reserve fund is a smart idea. This fund is put in place to help overcome unanticipated contingencies that may deplete the budget prior to project completion. Percentiles of the cost probability distribution can serve as guidelines for the magnitude of the appropriate management reserve. Suppose that the number submitted as the best estimate falls at the 40th-percentile level of the cost probability distribution. Depending on the importance of the project, a prudent reserve might consist of the funding required to bring the total available dollars to at least the 50th, 70th, or even 90th percentile. Risk dollars in the management reserve pool may sometimes constitute a large percentage of the budgeted best-estimate funding base. Funding agencies are reluctant to set aside such large amounts of money for management reserve, believing that “slush funds” lead to waste and slack management. It is therefore advisable to justify such requests by providing a logical
Cost (millions of dollars)
400 300 200 100 0 70
75 80 85 90 Learning rate (percent)
Cost-estimating practice involves the application of a learning rate to account for improvements that save money as successive units are produced. This graph illustrates how cost can vary at different learning rates.
allocation of the requested risk dollars among the various project elements. A specific WBS-based cost-risk analysis can profile a probable need for additional monies beyond those included in the best estimate. Because a WBS item’s need for risk dollars arises out of uncertainty in the cost of that item, a quantitative description of that need must serve as the mathematical basis of the risk-dollar computation. In general, the more uncertain the cost, the more risk dollars will be needed to cover a reasonable probability (e.g., 0.70) of being able to complete that element of the system. Items whose cost distributions have relatively high probabilities of exceeding their own most likely estimates will need more risk dollars. Methods were developed at Aerospace to allocate management reserve properly, taking into account not only the skewness of each item’s probability distribution, but also correlations between the items, so that management reserve dollars will not be assigned to do double duty. Estimates at Completion Based on Earned Values Once a program is under way, program managers must monitor how work is being done and money is being spent. Earned-value management is a specific, well-defined set of procedures used in program control to track expenditures and their relationship to the amount of work that has been accomplished. Earned-valuemanagement documentation compares outflow of program funds with completion of various work packages against which the funds have been budgeted. This comparison, used properly, allows programcontrol personnel to quickly spot overruns and possible schedule discrepancies. In
addition, earned-value-management data are used to calculate estimates at completion at any stage in the program. Despite the wealth of data that earned-valuemanagement systems bring to bear on the estimating process, they have not been able to circumvent the statistical nature of cost that, for at least the past decade, has been the driving force behind the development and application of cost-risk analysis. The two main quantities formally tracked by earned-value-management systems are cost variance and schedule variance. Cost variance is the difference between the amount of money budgeted for work actually completed (or completed over some time period, such as a month) and the amount of money actually spent to do that work, regardless of how much work was supposed to get done during that period. Schedule variance is the difference between the budgeted cost of the work completed (or completed over some period) and the budgeted cost of the work planned for that time period, regardless of the amount actually expended. Mathematically related to the cost and schedule variances, although calculated slightly differently, are two other quantities: the costperformance index and the schedule-performance index. The cost-performance index is a measure of the efficiency at which dollars are being spent on the project; for example, a costperformance index of 0.90 means 90 cents worth of work is getting done for every dollar spent. The schedule-performance index measures the rate at which work is being completed; a schedule-performance index of 0.90 means 90 percent of the work is getting done that is supposed to be done during the time period in question. Focusing on the problem of using earned-value data to calculate estimates at completion, Professor D. S. Christensen of Southern Utah University tracked the historical performance of a number of common methods of making estimates-atcompletion calculations. His research led him to conclude that final program cost almost always falls between estimates at completion based on two earned-valuederived indexes, the cost-performance index and the schedule-cost index. (The latter is the product of the cost-performance index and the schedule-performance index.) Furthermore, he found that actual
Aerospace Systems Architecting and Engineering Certificate Program Courses in Cost Estimation The Aerospace Institute offers several courses in the field of cost analysis. Cost Estimation and Analysis The logical foundations, methods, and terminology of cost analysis. Historical cost data on space, launch, and ground systems are described, and work-breakdown structures are introduced. The course explains the adjustment of raw data to establish a consistent data set for application to cost-model development and costestimating tasks. Various costestimating techniques are discussed. The course concludes with a description of cost models in use at Aerospace and associated organizations. Cost-Risk Analysis An overview of cost-risk analysis—the techniques used to represent costs in the form of probability distributions. Primary topics include rolling up costs of work-breakdown-structure elements, accounting for correlation among element costs, and allocating risk dollars back to elements. Also covered are methods for using cost risk as a criterion or figure of merit in trade studies and source selections, the use of NASA technology readiness levels to estimate cost risk, and the impact of learning rate on a production cost estimate. The case is made for using cost-risk analysis to estimate probabilities of cost overruns of various magnitudes. Vital Issues in Earned-Value Management The capabilities and deficiencies of earned-value management. This course goes beyond formal earnedvalue-management theory, elucidating the conclusions that organizations can justifiably draw about program progress from earned-value-data compilations. It focuses on what program managers and their staffs must know about earned value to make appropriate, valid program-control decisions.
Crosslink Winter 2000/2001 • 19
System Development Undertake trade study Cost-Risk Analysis Assess subsystem technologies
Conduct source selection
Assign estimates to WBS items
Perform Monte Carlo sampling
Perform Monte Carlo sampling
Adopt specific hardware and software designs
Begin system development
Estimates at Completion Track cost variance and CPI
Monitor work done and money spent ?
Derive floor and ceiling Track schedule variance and SPI
Complete system implementation
program final cost is generally closer to the estimates at completion based on the schedule-cost index, which he refers to as a ceiling to the final cost. He calls the estimates at completion based on the costperformance index the floor to the actual final cost, based on his conclusions that program cost performance rarely improves as the program proceeds to its completion. The results of Christensen’s research, namely the clear historical precedents for estimating a floor and ceiling to program final cost, were recently applied by analysts at Aerospace to construct a statistical approach to computing estimates at completion. Such an approach allows us to associate levels of confidence with various estimates at completion, to distinguish the dollar value of a best estimate at completion from risk dollars or management reserve, and to identify WBS items that are most likely to require an infusion of risk dollars. While this kind of information is 20 • Crosslink Winter 2000/2001
usually inferred from a detailed technical risk analysis, it can also be derived from earned-value data. Summary At several stages in the system engineering process, it is necessary to conduct a cost analysis to assess the likely magnitude of program funding requirements. The cost-analysis process begins at the trade-study stage with an applicable set of CERs statistically derived from historical cost data. After specific hardware and software designs have been formally adopted, it may be possible to base estimates on components that were previously developed and produced, vendor price quotes for existing off-theshelf items, or other appropriate data that can enable dollar values to be assigned to all items to be paid for. These information sources typically allow the analyst to estimate the most likely cost of each item, but because many technical risk issues are of-
ten present, the sum of the items’ most likely costs is usually significantly below the total system’s most likely cost. This unfortunate situation led in the early 1990s to the development of the subfield of cost-risk analysis, a way of looking at cost through the lens of probability and statistics. Once one accepts the idea of evaluating costs probabilistically, one is locked into the need to estimate correlations between the cost impacts of various technical risks because correlation is a significant driver of total-cost uncertainty. However, while cost-risk analysis makes demands upon the cost estimator, it also provides benefits to program management, particularly when it comes to recommending a prudent management reserve. Having in hand a probability distribution of total program cost, rather than just a single best estimate, program management can propose, for example, that the basic cost estimate be budgeted at the 50-percent confidence level, but that sufficient management reserve be
Combine random numbers
Combine random numbers
Read off cost percentiles
Identify WBS items needing risk dollars
Associate confidence levels with EACs
Allocate risk dollars
Identify WBS items needing risk dollars
The basic steps in estimating probable system cost, with detailed expansion of the processes for performing cost-risk analyses and developing estimates at completion.
included to bring the success probability up to 70 percent. Finally, as the program progresses, earned-value data on expenditures and work accomplished provide program management with the ability not only to maintain current knowledge of cost overruns, but also to estimate cost at completion from inside the program itself rather than by statistical inference from historical information on other programs. Further Reading R. L. Abramson and S. A. Book, “A Quantification Structure for Assessing Risk-Impact Drivers,” The Aerospace Corporation, 24th Annual DOD Cost Analysis Symposium (Leesburg, VA, September 5–7, 1990). R. L. Abramson and P. H. Young, “FRISKEM— Formal Risk Evaluation Methodology,” The Journal of Cost Analysis, 29–38 (Spring 1997). Assistant Secretary of Defense (Program Analysis and Evaluation), Cost Analysis Guidance and Procedures (Department of Defense DOD 5000.4-M, December 1992), pp. 2–4 to 2–6.
S. A. Book, “Do Not Sum ‘Most Likely’ Cost Estimates,” The Aerospace Corporation, 1994 NASA Cost Estimating Symposium (Houston, TX, November 8–10, 1994). S. A. Book and E. L. Burgess, “The Learning Rate’s Overpowering Impact on Cost Estimates and How to Diminish It,” Journal of Parametrics, Vol. 16, No. 1, 33–57 (Fall 1996). S. A. Book, “Justifying ‘Management Reserve’ Requests by Allocating ‘Risk Dollars’ among Project Elements,” The Aerospace Corporation, Fall 1996 Meeting of the Institute for Operations Research and Management Science (INFORMS) (Atlanta, GA, November 3–6, 1996). S. A. Book and P. H. Young, “General-Error Regression for Deriving Cost-Estimating Relationships,” The Journal of Cost Analysis, 1–28 (Fall 1997). S. A. Book, “Why Correlation Matters in Cost Estimating,” The Aerospace Corporation, 32nd Annual DOD Cost Analysis Symposium (Williamsburg, VA, February 2–5, 1999). S. A. Book, “Do Not Sum Earned-Value-Based WBS Estimates-at-Completion,” The Aerospace Corporation, Society of Cost Estimating
and Analysis National Conference (Manhattan Beach, CA, June 13–16, 2000). E. L. Burgess and H. S. Gobreial, “Integrating Spacecraft Design and Cost-Risk Analysis Using NASA Technology Readiness Levels,” The Aerospace Corporation, 29th DOD Cost Analysis Symposium (Leesburg, VA, February 21–23, 1996). D. S. Christensen, “Using Performance Indices to Evaluate the Estimate at Completion,” The Journal of Cost Analysis, 17–24 (Spring 1994). H. L. Eskew and K. S. Lawler, “Correct and Incorrect Error Specifications in Statistical Cost Models,” The Journal of Cost Analysis, 105–123 (Spring 1994). P. R. Garvey, Probability Methods for Cost Uncertainty Analysis—A Systems Engineering Perspective (Marcel Dekker, New York, 2000). R. Giguet and G. Morlat, “The Causes of Systematic Error in the Cost Estimates of Public Works,” Annals of Bridges and Roads, No. 5 (September–October 1952, Paris, France). Translated from the French by W. W. Taylor, U.S. Air Force Project RAND, Santa Monica, CA, March 1958.
Crosslink Winter 2000/2001 • 21
Lockheed Martin Corporation
U.S. Air Force, 30th Space Wing, VAFB
Successful launches of five U.S. launch vehicles. Clockwise from far left: Delta II launched in 2000 from Vandenberg Air Force Base, California; Atlas vehicle carrying the DSCS III B13 spacecraft into orbit in October 1997; Scout lifting a radar calibration satellite into orbit in June 1993; an early morning liftoff of the space shuttle Discovery from Kennedy Space Center on November 27, 1989; Titan IV, the Air Force’s premier heavy lift vehicle, launched from Vandenberg Air Force Base on November 28, 1992.
Lockheed Martin Corporation
Space Launch Vehicle I-Shih Chang
The 1993 failure of a Titan IV K-11 launch vehicle prompted the U.S. Air Force to request The Aerospace Corporation’s participation in an analysis of space-mission failures. I-Shih Chang led the Aerospace study, which was later expanded to support the DOD Broad Area Review of U.S. launch failures between 1985 and 1999. This article is the second in the Crosslink series on the history of the corporation’s role in national space programs.
hen you get into your car and insert the key in the ignition, you expect the car to start. You expect to pull away from the parking spot and drive to your destination without a problem. These expectations are based on the assumption that your vehicle is reliable—you can depend on it to behave the same way time after time. This ideal applies to launch vehicles as well. If the vehicle has been designed, built, and tested well, we expect it to successfully leave its launchpad. But just as personal vehicles can fail in unforeseen ways (radiators leak, engines stall), so too can launch vehicles occasionally fail. Launch failures have been a fact of life for most space-faring nations since the space age began in 1957. Because a space mission involving a launch vehicle and a sophisticated satellite can easily cost hundreds of millions of dollars, investigation into why launches fail can provide valuable information for improving vehicle systems and cost savings. Any lessons learned from past experiences that could serve to mitigate launch failures in the future would make such an investigation extremely worthwhile. A space launch failure is an unsuccessful attempt to place a payload into its intended orbit. This definition includes all catastrophic launch mishaps involving launch vehicle destruction or explosion, significant reduction in payload service life, and extensive effort or substantial cost for mission recovery. It also includes the failure of the upper stage of a launch vehicle, up to and including spacecraft separation on orbit. However, this definition does not include the failure of an upper stage released from the U.S. space shuttle. The U.S. space shuttle is both a launch vehicle and a space vehicle. An upper stage released from the space shuttle in orbit is considered a transfer vehicle, not a launch vehicle. The space age began with the USSR launch of the first artificial satellite, a liquid-fueled Sputnik (SL-1), on October 4, 1957. At present, nine countries or consortia—the United States, the Common-
wealth of Independent States (CIS, formerly USSR), the European consortium, China, Japan, India, Israel, Brazil, and North Korea—possess space launch systems, demonstrate space launch capability, or conduct space launch operations. Many current major space launch systems are based on early ballistic-missile technology, which regarded launch costs and schedules a higher priority than launch quality and reliability. The design of these space launch systems left much room for improvement, as demonstrated by launch failures of the past. Financially, much is at stake in any kind of space launch. A small launch vehicle, such as the U.S. Pegasus, costs about $15 million, but a versatile, reusable launch vehicle, such as the U.S. space shuttle, costs well over $1 billion. A small experimental satellite can be purchased for a few million dollars, but an advanced spy satellite or scientific satellite may cost more than $1 billion. Furthermore, the possible monetary loss calculated for a launch failure does not include the expense, time, and effort spent during the recovery period or the cost of the damage to national prestige. Analysis of space launch failures is critical to a national space program’s future success. A systematic look at worldwide launch successes as well as failures, including scrutiny of various launch vehicle subsystems, can shed light on precise areas that might be at the root of many problems. This
Photo courtesy of NASA
Crosslink Winter 2000/2001 • 23
24 • Crosslink Winter 2000/2001
1999, the European consortium consistently conducted 10 or more launches per year. China and Japan have also invested heavily in space programs. China’s government publicized its first two successful space launches—in 1970 and 1971—as great national achievements in science and engineering. The Chinese CZ-2C vehicle holds a perfect record for Chinese launchers of 11 launch successes. Over the last few years, the Chinese government has made considerable investment in improving its launch-base infrastructure to gain a greater share of the commercial space market. In Japan, the National Space Development Agency (NASDA) and the Institute for Space and Astronautical Science
(ISAS) are responsible for space research and development. The NASDA H-II and ISAS Mu-5 vehicles use state-of-the-art technology and represent the Japanese government’s commitment to becoming a major player in space. Japan had an 18year streak (1977–1994) of successful consecutive space launches. The remaining nations whose launches have been tracked have, for the most part, conducted space programs for very brief time periods. They have experienced mixed success records. India, undaunted by a series of technical setbacks and launch failures in its fledgling space program, allocates a significant portion of its yearly budget to space technology development. Israel’s Space Agency launched its first satellite with the Shavit launch system on
2500 Space launches
2000 1500 1152
1000 500 164
N. Korea U.K. Brazil France Australia
CIS/ China India USSR The number of successful and failed launches for the space-faring nations of the world between 1957 and 1999. The CIS/USSR has carried out more launches than all other countries combined.
Space launch success rate (percent)
type of study can also help suggest what actions to take to address those problems. Worldwide Space Launches To understand space launch reliability, it is helpful to examine the history of launches worldwide since 1957. The progress made in space science and engineering during this period has been truly remarkable, as illustrated by the achievements of the United States and the CIS/USSR, the nations that have dominated the space launch arena. To get an idea of how great that progress is, consider that in 1957 the USSR’s Sputnik 1 weighed only 83.6 kilograms, and on July 16, 1969, the U.S.’s Saturn V, the largest and most powerful operational rocket ever built, lofted Apollo 11, with a mass of 43,811 kilograms, to lunar orbit during the moon-landing mission. Today, the U.S. Space Transportation System routinely launches the shuttle orbiter, weighing more than 110,000 kilograms, to low Earth orbit. The orbiter flies like a spacecraft and lands like a glider. The USSR was the first country to place a satellite carrying a person into Earth orbit. Its Soyuz vehicle has been statistically shown to be the most reliable expendable launch vehicle in the world. Since 1957, CIS/USSR has carried out more space launches than all other countries combined. Between 1957 and 1999, 4378 space launches were conducted worldwide, including 2770 CIS/USSR launches and 1316 U.S. launches. These figures include launches carried out individually by France and the United Kingdom (U.K.) over 25 years ago. France was the third country (after CIS/USSR and the United States) to attain space launch capability, with Diamant, a rocket that placed a satellite in orbit in 1965. The U.K. developed a small vehicle, Black Arrow, which was launched successfully in 1971. Currently, France and the U.K. participate through the European Space Agency (ESA) in the development of the Ariane launch systems. (The ESA is composed of 14 European nations.) The Ariane vehicles are launched from French Guiana in South America, which is only 5.2 degrees north of the equator and is therefore an excellent location for launching satellites into geosynchronous orbit. Recently, the European consortium has been catching up to CIS/USSR and the United States and capturing a large share of the commercial space launch service market formerly dominated by U.S. launches. From 1995 to
100.0 90.7 83.6
Europe CIS/ USSR
N. Korea U.K. France Australia
Each space-faring nation’s launch success rate as a percentage of its total launches.
September 19, 1988. Its third satellite, launched April 5, 1995, contains surveillance equipment designed to provide reconnaissance and military observation. Brazil’s satellite launch attempts using the Veiculo Launcadror de Satelites (VLS)—“satellite launch vehicle”—from the Alcantara launch site failed on November 2, 1997, and December 11, 1999. North Korea claimed to have successfully launched the small Kwangmyongsong-1 satellite into orbit by a Taepo Dong-1 vehicle on August 31, 1998. But other countries have received no signal from it, and the launch is considered a failure (third-stage malfunction). Australia launched a small Sparta (SPecial Anti-missile Research Tests, Australia) vehicle in 1967, which was a modified U.S.
Redstone rocket. Today Australia does not have its own launch system. The Woomera Rocket Range in Australia is currently dedicated to weapons and sounding rocket testing. It is also worth noting that since the end of the Cold War, national boundaries in the space launch business have become less distinct. Companies throughout the world have been marketing CIS/USSR launch vehicles for commercial launch service: Proton by Lockheed Martin, Zenit by Boeing, Kosmos by Cosmos U.S.A., Soyuz by a France-Russia consortium, and Rokot by a Germany-Russia consortium. Space Launch Failures Of the 4378 space launches conducted worldwide between 1957 and 1999, 390 launches failed (the success rate was 91.1
percent), with an associated loss or significant reduction of service life of 455 satellites (some launches included multiple payloads). A brief look at some of the most publicized, critical launch failures around the world will highlight the nature of system failures. In the United States, 164 space launches failed, with an associated loss or significantly reduced service life of 205 satellites. Most of the U.S. space launch failures (101 out of the 164) occurred during the first 10 years of space exploration (1957–1966). In that period, the United States was diligently attempting to catch up to the USSR, which had gained an early lead in space exploration. The first space launch failure involved a U.S. Vanguard vehicle, which exploded two seconds after liftoff on
U.S. Army White Sands Missile Range
A Brief History of Rocketry
All modern liquid rockets can be traced to the V-2, the first real ballistic missile. An unmanned missile guided by a gyroscopic system, it burned a fuel containing alcohol and liquid oxygen. The V-2 had a range of over 320 kilometers, and it could transport an explosive warhead that weighed a ton.
People have been building rockets in various forms for centuries. In 1232, the Chinese army shot “fire arrows”—solid rockets propelled by a gunpowder mixture of charcoal, sulfur, and potassium nitrate (saltpeter)—at invading Mongols in the Battle of Kai-Fung-Fu. Besides military weapons, other applications for rockets over the years have included fireworks, lifesaving devices (as early as 1882), whaling operations (as
early as 1865), and signal and illumination devices. The basics of rocketry have not changed. Ignited fuel burns in a combustion chamber, and the resulting gases are forcefully expelled through a nozzle, propelling the rocket in the opposite direction. Until the first quarter of the 20th century, all rockets were developed using solid propellant. Today’s space systems employ both solid-rocket motors and liquid-rocket engines to deliver payloads into orbit. Original rocket technology did not progress much until the early 19th century, when Colonel William Congreve of England developed effective bomb-carrying solid rockets. Long sticks trailed from his rockets to provide stability. Soldiers put Congreve’s rockets to use during the British bombardment of Baltimore in 1814, at Waterloo in the war with Napoleon in 1815, and in the Opium War with China in 1842. (The bombardment of Baltimore and “the rockets’ red glare” inspired Francis Scott Key’s famous poem “The Defense of Fort McHenry,” adopted later as the U.S. national anthem, “The Star-Spangled Banner.”) Congreve’s fellow Englishman, William Hale, developed a more accurate stickless, spin-stabilized rocket in 1844. In the early 20th century, Wilhelm Teodore Unge of Sweden improved the mechanical strength and workability of solid propellant and developed launcherrotated rockets that could travel eight kilometers with great accuracy. Y. P. G. Le Prieur of France invented stick-guided solid rockets for firing through steel tubes mounted on the wing strut of a biplane, which were used successfully in World
War I against German observation balloons. These constituted an early version of air-launched missiles. Konstantin Eduardovitch Tsiolkovsky of Russia first introduced the idea of using rockets for space exploration in 1903. Subsequently the idea was proposed by Robert Hutchings Goddard of the United States in 1919, Hermann Oberth of Germany in 1923, and Robert Esnault-Pelterie of France in 1928. In 1926, Goddard built a successful rocket using liquid propellant (gasoline as fuel and liquid oxygen as oxidizer). Germans used his 1939 design of a liquid rocket to build and test the first full-scale ballistic missile, Vergeltungswaffe-2 (V-2, “Weapon of Retaliation-2”), in 1942. All modern liquid rockets can be traced to the V-2. With a mixture of ethyl alcohol as fuel and liquid oxygen as oxidizer, it could travel 320 kilometers. The V-2 carried warheads from the European continent to England in the “Siege of London” during World War II. At the end of the war, Russia captured V-2 manufacturing plants, including many German rocket scientists and engineers, and set up its own ballistic missiles and space launch programs under the leadership of Sergei P. Korolev. Meanwhile, Wernher von Braun, the leader of the German V-2 program, and his coworkers continued their research on ballistic missiles and space launch vehicle development in the United States. During the 46 years of the Cold War between the United States and the USSR (September 2, 1945, through December 26, 1991), ballistic-missile buildup and the space race fostered the continued growth of rocket science.
Crosslink Winter 2000/2001 • 25
Launch Successes (s) and Failures (f), 1957–1999 Year 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Total 57–99 60–99 70–99 80–99 90–99
0 5 11 16 22 48 37 56 61 72 58 43 38 28 31 31 23 23 27 24 23 32 16 12 18 18 22 21 17 6 8 11 18 26 17 28 23 26 26 32 37 34 27
1 18 8 13 19 11 9 8 9 5 3 5 3 2 4 2 2 2 4 2 3 1 0 3 1 0 0 1 1 3 1 1 0 1 2 1 2 1 4 1 1 2 4
2 0 1 4 3 1 5 6 6 3 20 2 17 7 30 6 48 6 43 11 68 8 74 8 69 15 81 7 83 9 74 5 86 4 81 4 89 4 98 2 98 4 87 4 87 3 87 3 97 2 100 9 97 3 97 1 97 4 91 5 95 3 90 5 74 1 74 5 59 2 54 1 46 2 48 1 31 2 23 4 27 2 24 1 28 2
0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 2 4 3 2 2 7 7 5 8 7 7 6 11 10 12 11 10
1 1 1 1 0 0 0 0 0 0 0 0 1 0 1 0 0 1 1 0 0 0 1 0 0 0 2 0 1 0 0 0
1 1 0 0 0 3 2 0 1 0 0 1 1 1 2 1 2 2 4 0 5 0 3 1 5 2 2 6 6 4
0 0 0 1 2 0 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 1 0 0 1 2 0 0 0
0 0 0 0 1 2 1 0 1 2 1 2 3 2 2 3 1 3 3 2 2 3 2 2 3 2 1 1 2 1 1 2 1 0
1,152 164 2,589 181 117
87.5% 89.2% 92.9% 93.4% 93.6%
93.5% 93.6% 95.5% 95.8% 95.0%
90.7% 90.7% 92.1% 93.5% 95.6%
83.6% 83.6% 83.6% 88.9% 87.2%
December 6, 1957. The failure, which attracted tremendous public attention and criticism in the wake of two successful USSR Sputnik flights, was the result of a low fuel tank and low injector pressure that allowed the high-pressure chamber gas to enter the fuel system through the fuel-injector head. A fire started in the fuel injector, destroying the injector and causing a complete loss of thrust immediately following liftoff. The U.S. Saturn V had a single failure in the Apollo 6 mission on April 4, 1968, when the third-stage engine failed to restart because of fuel-injector burn26 • Crosslink Winter 2000/2001
2 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1 9
85.2% 85.2% 91.2% 92.5% 82.4%
0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 2 0 1 0 0 1
1 0 1 0 0 0 0 0 1 1 0 0 0 0 1 0 0 0 1 0 0
53.8% 53.8% 53.8% 58.3% 71.4%
1 0 1 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1 0
0 0 0
75.0% 75.0% 75.0% 75.0% 66.7%
1 0 1
0.0% 0.0% 0.0% 0.0% 0.0%
1 1 2 0 0 2 1 0 0 0 3
0 0 0 0 0 0 1 0 1 0 0
1 0 1 0.0% 0.0% 0.0% 0.0% 0.0%
through. The versatile Space Transportation System also suffered a single launch failure on January 28, 1986, when the Challenger, carrying a seven-member crew, exploded 73 seconds into flight. The launch management had waived the temperature-dependent launch commit criteria and launched the vehicle at a colder temperature than experience indicated was prudent. At such a low temperature, the rubber O-rings in the motor case joint lost their resiliency, and the combustion flame leaked through the O-rings and case joint, causing the vehicle to explode. The newly developed U.S. commercial launch
83.3% 83.3% 75.0%
50.0% 50.0% 50.0%
2 6 14 21 28 68 54 86 110 116 129 117 107 113 119 106 109 105 124 125 123 123 106 102 121 120 126 127 120 103 110 115 101 114 86 94 78 89 72 69 84 76 70
1 22 9 19 22 13 16 14 15 18 12 14 20 12 15 7 8 8 8 6 7 5 5 7 4 10 3 3 6 9 5 7 1 7 5 3 5 4 8 8 5 6 8
1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
91.1% 91.7% 94.1% 94.5% 93.4%
57–99 60–99 70–99 80–99 90–99
systems, including Delta III, Conestoga, Athena, and Pegasus, suffered launch failures during their early developmental flights, a repeat of Vanguard, Juno, Thor, and Atlas failures in the late 1950s and early 1960s. CIS/USSR experienced an impressive number of space launches and a strong launch success rate in the past. However, the number of space launches and the success rate in recent years have declined, mainly because of domestic financial problems. From 1996 to 1999, for example, the United States conducted more space launches than CIS/USSR for the first time in 30 years.
U.S. Launch Vehicle Successes (s) and Failures (f), 1957–1999 Year
2 8 10 9 10 8 6 8 7 8 7 8 8 7 6 5 3 5 5 3 7 1 0 3 2 4 4 2 3 1 5 4 4 5 2 3
1 2 1 1 0 0 1 0 2 1 1 1 1 0 1 0 0 0 0 0 0 1 1 0 1 0 1 0 0 1 0 0 0 0 1 2
1 0 1 4 11 9 15 14 31 14 6 5 2 4 6 4 2 3 4 6 14 4 6 5 3 6 4 5 3 2 2 1 3 3 4 5 7 12 7 8 6 5
0 1 4 7 5 1 3 5 2 0 2 0 1 2 0 0 0 2 0 1 0 0 2 1 0 0 1 0 0 1 0 0 0 1 1 1 0 0 0 0 0 0
1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
2 3 4 5 9 1 0 2 5 6 6 8 7 7 7 7 8 5 3
0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
57-99 99.0% 60-99 99.0% 70-99 99.0% 80-99 99.0% 90-99 100.0%
89.7% 89.7% 89.5% 89.2% 86.8%
2 3 9 7 4 7 8 12 7 9 7 4 7 5 6 12 9 8 10 3 3 5 7 8 4 0 1 2 1 8 11 5 11 7 3 2 10 10 12 10
1 0 0 0 1 1 0 0 1 2 0 1 0 1 1 0 0 2 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 1 1 1
257 44 259 16 85.4% 85.6% 91.3% 92.4% 95.2%
94.2% 94.2% 95.0% 96.0% 95.3%
Taurus Athena Pegasus Saturn s
3 3 1 1 3 4 1 2 2 4 0 1
1 0 0 0 2 1
0 0 0 0 0 0
0 0 1 1 2
1 0 0 0 1
1 0 0 2 2 1 4 5 6 3
100.0% 100.0% 100.0% 100.0% 100.0%
66.7% 66.7% 66.7% 66.7% 66.7%
0 1 0 0 1 1 1 0 0 0 4
85.7% 85.7% 85.7% 85.7% 85.7%
Thor** Conestoga Scout
0 0 0 0 1 0 0 0 0 0 0 0
0 7 12 13 25 17 25 24 14 16 12 11 9 8 4 1 2 1 1 1 1 1 0 0
3 3 6 7 4 5 2 1 2 0 1 0 0 1 0 0 0 0 1 0 0 0 1 0
1 205 37
96.2% 96.2% 100.0%
84.7% 86.5% 90.6% 0.0%
0.0% 0.0% 0.0% 0.0% 0.0%
0 1 3 4 7 5 8 6 5 1 3 5 5 1 6 2 2 1 1 3 0 1 0 1 1 2 1 1 4 0 1 1 2 1 1
1 3 2 3 1 0 0 2 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
86.0% 86.0% 97.9% 100.0% 100.0%
3 2 1 1
4 2 1 2
0 1 2
1 5 2
0 5 11 16 22 48 37 56 61 72 58 43 38 28 31 31 23 23 27 24 23 32 16 12 18 18 22 21 17 6 8 11 18 26 17 28 23 26 26 32 37 34 27
Year f 1 18 8 13 19 11 9 8 9 5 3 5 3 2 4 2 2 2 4 2 3 1 0 3 1 0 0 1 1 3 1 1 0 1 2 1 2 1 4 1 1 2 4
1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
87.5% 89.2% 92.9% 93.4% 93.6%
57-99 60-99 70-99 80-99 90-99
* Includes TAD (Thrust-Augmented Delta) ** Includes TAT (Thrust-Augmented Thor), LTTAT (Long Tank Thrust-Augmented Thor), and Thorad (Thrust-Augmented Thor Delta) *** Includes Juno-I and Juno-II
Space launch failure in a closed society like the USSR or the People's Republic of China was guarded as a state secret and not publicized in news media. Recently, though, because of “glasnost” in Russia, commercial competition, and requirements by the launch service insurance company, information flow on space activities has improved dramatically. Since the collapse of the USSR on December 26, 1991, CIS has released information on many USSR space launch vehicle failures that were not previously known to the West. Making this information accessible has provided a much more complete picture of worldwide space launches, although
the vast amount of information existing in CIS/USSR from both successful and failed launch operations is yet to be assimilated by space launch communities of the world. One CIS/USSR space launch failure involved an SL-12 Proton vehicle carrying a Mars-96 spacecraft on November 16, 1996. The second burn of the Proton’s fourth stage did not take place, and the spacecraft did not reach the interplanetary trajectory. It reentered Earth’s atmosphere over the South Pacific Ocean. For lack of funds, CIS launched this spacecraft without conducting a prelaunch systems checkout at the launch site. Some of the
mechanical integration of the spacecraft and launcher was carried out by the light of a kerosene lantern (electrical power had been cut off because of unpaid bills). Tight funding also made ground control difficult, even during the critical period immediately following launch. The spacecraft itself commanded the fourth-stage release, indicating that it had possibly sent incorrect commands. It took 10 years to complete the $300 million Mars-96 spacecraft carrying two dozen instruments supplied by 22 countries. This launch failure stalled plans for gathering valuable data about the planet Mars. Crosslink Winter 2000/2001 • 27
The failures of the European Launcher Development Organization (ELDO) Europa vehicle were reminiscent of the early launch failures in the U.S. space program. (ELDO was one of the predecessors of ESA.) After terminating the Europa program, Europe spent many years planning the Ariane launch systems, which have experienced eight failures since 1980. A recent failure involved a new Ariane5 vehicle, the most powerful in the Ariane family. During its maiden flight on June 4, 1996, it veered off its flight path and exploded at an altitude of 3700 meters only 40 seconds after liftoff. The failure was attributed to errors in the design and testing of the flight software. The flight software was programmed for Ariane-4 launch conditions, but it was never tested in conditions that simulated Ariane-5’s trajectory. The more powerful Ariane-5 travels at a much faster horizontal velocity than the Ariane-4. Significant horizontal drift caused an overflow error in the inertial reference system (IRS) software, halted the primary and backup IRS processors, and resulted in the total loss of accurate flight guidance information.
U.S. space launch vehicles. Notable among them is the Saturn V, the largest, most powerful operational rocket ever built. With an impressive 259 successes out of 275 launches, Delta is the United States’ most reliable expendable launch vehicle. (Numbers represent length in meters.)
From 1991 to 1996, the Chinese space launch record was marred by five failures. The most catastrophic failure occurred during the launch of a CZ-3B vehicle carrying a commercial satellite, Intelsat 708, on February 14, 1996. The 55-meter-tall CZ-3B is China’s most advanced vehicle. On its maiden flight, the CZ-3B began to veer off course two seconds after liftoff, before it even cleared the tower at the Xichang launch site. The vehicle and its payload hit the ground and exploded in an inhabited area near the launch site 22 seconds after liftoff. The explosion demolished a village and a nearby military base, and caused severe casualties and property damage. The cause of failure was traced to the CZ-3B’s guidance and control subsystem. A gold-aluminum solder joint in the output of one of the gyro servo loops failed, cutting electrical current output from the power module and causing the inertial reference platform of the vehicle’s guidance and control system to slope. This caused computers to send the vehicle veering off the planned trajectory shortly after liftoff. The failed module was the only one of six similar modules that lacked conductive adhesive to reinforce the solder joint. Japanese liquid-propellant rockets (H-II) suffered two launch failures during 1998 and 1999. Japan’s other seven launch failures (including four Lambda-4S failures during the period 1966–1969) involved solid-propellant rockets. One of the failures occurred on January 15, 1995, during the last flight of the Mu-3S-II vehicle. At 103 seconds after launch, the vector control thrusters, which partly control the rocket’s
pitch, began to oscillate. The rocket veered off course at 140 seconds after liftoff. The payload of the Mu-3S-II, a German satellite (Express 1), was put in the wrong orbit of 120 kilometers altitude, instead of the intended orbit of 210 kilometers altitude. The satellite, which fell into a jungle in Ghana after circling Earth two and a half times, failed because of improper modeling of the flight control dynamics relative to the weight of the payload. (Prior to the failure, the heaviest payload carried by the Mu-3SII had been 430 kilograms; Express 1 weighed 748 kilograms.) Extra propellant had been added to the three stages and to the kick motor of the vehicle to provide extra thrust for the flight of the Express 1 satellite. The flight was the eighth and final mission of the Mu-3S-II vehicle. Several satellites have plunged into Bengal Bay since India’s space program began in 1979. India’s Polar Satellite Launch Vehicle (PSLV) is designed to place payloads into a polar sun-synchronous Earth orbit. On its maiden flight on September 20, 1993, the PSLV experienced an unplanned change in pitch when the spent second stage separated from the vehicle at 260 seconds into flight. The third and fourth stages ignited normally, but the vehicle was unable to recover from the pitch change and did not reach sufficient altitude. The payload was placed in a 349kilometer orbit instead of the planned 814kilometer polar orbit. Shifting liquid fuel in the second stage of the vehicle may have caused the change in the vehicle’s pitch. Malfunction of the vehicle’s guidance system or failure of the control system to respond properly to the course deviation could have been the cause of the failure.
62.2 56.1 47.5 41.1 39 35.4 33.2
28 • Crosslink Winter 2000/2001
Causes of Failure Available launch-failure data reveal much about patterns in the possible causes of failure. Many failure causes fall into the category of human error: poor workmanship, judgment, or launch-management decisions. Some failures are the result of defective parts. Failure can have its root in any phase of launch vehicle development—difficulties have been noted in inadequate designs and component tests; in improper handling in manufacturing and repair processes; and in insufficient prelaunch checkouts. Many past failures could have been prevented if rigorous reliabilityenhancement measures had been taken. Launch vehicle failure is usually attributed to problems associated with a subsystem, such as propulsion, avionics, separation/staging, electrical, or structures. In some cases failure is ascribed to problems in another area altogether (e.g., launchpad, ground power umbilical, ground flight control, lightning strike), or to unknown causes (usually when subsystem failure information is not available). Launch vehicle failures between 1980 and 1999 have been investigated, and launch failure causes in the United States have been found to include fuel leaks (resulting from welding defects, tank and feedline damage, etc.), payload separation failures (from incorrect wiring, defective switches, etc.), engine failure (the result of insufficient brazing in the combustion chamber), and loss of vehicle control (because of lightning, damaged wires that caused shorts, and control-system design deficiencies). In Europe and China, launch failure causes during the same period included engine failure (from combustion
Launch Vehicle Subsystem Failures, 1980–1999 Country
U.S. CIS/USSR Europe China Japan India Israel Brazil N. Korea
15 33 7 3 2 1 1 2
4 3 1 1 1 1
instability, hydrogen injector valve leakage, clogged fuel lines, etc.), short circuits, engine thrust loss, software design errors that resulted in guidance system failure, wind shear, and residual propellants. Statistics show that among the causes of failure for space launch vehicles worldwide from 1980 to 1999, propulsion subsystem problems predominated. That particular subsystem appears to be the Achilles’ heel of launch vehicles. Fifteen of the 30 U.S. failures were failures of the propulsion subsystem. The unknown failures in the CIS/USSR could include many in the propulsion subsystem. The propulsion subsystem, the heaviest and largest subsystem of a launch vehicle, consists of components that produce, direct, or control changes in launch vehicle position or attitude. Its many elements include main propulsion components of rocket motors, liquid engines, and thrusters; combustion chamber; nozzle; propellant (both solid and liquid); propellant storage; thrust vector actuator and gimbal mechanism; fuel and propulsion control components; feed lines; control valves; turbopumps; igniters; motor and engine insulation. Similar components are
also used as separation mechanisms in the separation/staging subsystem. Propulsion subsystem failures can be divided into failures in solid-rocket motors and liquid-rocket engines. Solid-propellant launch systems include Taurus, Conestoga, Athena, Pegasus, and Scout. Liquidpropellant launch systems include Titan II, Titan IIIA, Titan IIIB, Atlas (except Atlas IIAS), and Delta DM19, A, B, and C. Hybrid launch systems, consisting of liquidpropellant and solid-propellant rockets, include STS, Titan IV, all other Titan III, Atlas IIAS, and all other Deltas. The success rate of the propulsion subsystem in the United States from 1980 to 1999 was 98.8 percent for solid-rocket motors and 97.5 percent for liquid-rocket engines. Addressing the Propulsion Problem Solid-rocket motors and liquid-rocket engines of the propulsion and separation/staging subsystems both require sets of precautionary measures to maximize reliability and safeguard against failure. First, consider solid-rocket motors. In the design phase, to reduce risk it is important to apply current analysis techniques to ensure fast, accurate, and low-cost modeling of precise configurations prior to hardware fabrication.
30 58 8 6 3 5 1 2 1
Crosslink Winter 2000/2001 • 29
Demonstrated Reliability Reliability for each space-faring nation’s launch vehicles between 1957 and 1999 is calculated from a cumulative statistical distribution function that is related to the percentage of
launch successes. Separate historical and operational (current) graphs represent the United States and the CIS/USSR, the two nations with the most extensive launch histories.
U.S. (Operational) 100
Demonstrated reliability (%)
Vanguard Pilot 56
China CZ-2C, -2D, -2C/SD
100 Mu-3S, 3H, 3C
N-I N-II Mu-3S-II
Europe 100 Demonstrated reliability (%)
Demonstrated reliability (%)
100 Demonstrated reliability (%)
80 84 Year
30 • Crosslink Winter 2000/2001
80 84 Year
In the construction of solid-rocket motors, a number of safeguards apply to the preparation of solid propellant: • Upon receipt from the supplier and prior to use, the propellant ingredient should be checked to ensure that it meets specifications, and the propellant mix procedure and burn rate should be checked for every mix before casting. • Motors should be cast from a single lot of material to minimize thrust imbalance of vehicles with multiple solidrocket motors. • Solid propellant should be cast in a vacuum, if possible, to reduce the number and size of internal voids. • Modern techniques (e.g., computer tomography) should be used to detect solid-propellant defects and propellantto-insulation bondline separation before motor assembly. Still other steps can be taken to help increase the likelihood of solid-rocket motor launch success. Chemical fingerprinting can be implemented for rare and sensitive chemicals such as propellant binder, motor case resins, flexseal elastomers, and adhesives. It should be possible to schedule into the development and qualification programs a motor-case cure study wherein a cured case is dissected to assess the adequacy of the process used in case-manufacturing. The liquid-rocket engines of propulsion and separation/staging subsystems should be designed and built with robustness and with high thermal and structural margins to allow for manufacturing deviations. Welded joints instead of seals ought to be used for fluid lines; high-pressure margins should be allowed in tanks, hydraulic lines, and plumbing; and 100 percent inspection, rather than spot-checking, must be applied to all welds. Other preventive measures for liquid engines include the application of redundancy in fluid valves and igniters; the utilization of effective liquid film cooling or ceramic coatings to increase thrustchamber durability; and the application of advanced high-strength (aluminummagnesium) welding and milling for the construction of thin-walled fuel and oxidizer tanks. Helium purging (for cryogenic propellants) or nitrogen purging (for storable propellants) of oxidizer/fuel pumps and pipelines needs to be done before engine start-up, and purging of the chamber cooling duct should be done at engine
shutdown, to provide a clean flow duct and to avoid the danger of fire or explosion. Testing liquid engines is also important. They should be qualified at above the maximum operating environment, conditions, and duration. And extensive tests on engine operation ought to be conducted under various conditions after transportation of the engine, since transporting an engine subjects it to a harsh environment that can alter its operation. Enhancing Launch Reliability Information gathered from failure studies of past launch vehicles indicates that following certain work practices could greatly enhance the reliability of launch vehicle systems. Of primary importance is the need to review and implement all lessons learned from past failure studies to avoid failure recurrences. It is necessary to incorporate preventive measures in all aspects of system development—design, building, testing, and operations. Launch vehicles should be designed for low cost in manufacturing, operations, materials, and processes rather than for maximum performance or minimum weight. Comprehensive design analyses should be conducted, with positive margins. In the manufacturing phase, only flightproven and defect-free components should be used. Advanced electronic beam welding, automation, and robotics should be applied for quality component manufacturing. Stringent control of raw materials, components, and semifinished products ought to be practiced. Multistring/redundant avionics, electrical, and ordnance components should be implemented for fault tolerance. Pyro-shock levels ought to be reduced whenever possible. Testing is a critical area for reliability enhancement. It is important for a design to be validated by testing components to the point of destruction or with a high enough margin to allow for manufacturing and operating environment variances, like the successful design margin testing performed on ballistic missiles. Electrical and pneumatic connection tests should be performed for each stage and between the stages before vehicle assembly. Components, software, and systemlevel electrical elements need to be tested under conditions that simulate an actual launch; system performance and flight simulation tests should be conducted; the results of testing during the development phase should be analyzed, and measures
Solid and Liquid Rockets Whether or not they know it as Newton’s Third Law, most people have probably heard the statement “For every action there is an equal and opposite reaction.” That expression is the principle behind rocketry. The action is the expulsion of gas through an opening; the reaction is the rocket’s liftoff. It’s not unlike what happens when you blow into a balloon, then release it. As air escapes (that’s the action), it propels the balloon (that’s the reaction), making the balloon zip through the room until its air is completely gone. In order to create a forceful expulsion of gas from a rocket, fuel in a combustion chamber is ignited. The fuel can be in the form of solid or liquid substances; some rockets (“hybrid launch systems”) may make use of both. These substances are the propellants that characterize rockets as either “solid-rocket motors” or “liquidrocket engines.” For the fuel to burn, oxygen or another oxidizing substance must be present. When the fuel burns, gases accumulate and pressure builds until the gases are expelled through an exhaust nozzle. In solid-rocket motors, the fuel and oxidizing chemicals are suspended in a solid binder. Solid motors are used as boosters for launch vehicles (such as the space shuttle and the Delta series). They’re very reliable, and they’re simpler than liquid engines, but they’re difficult to control. Once a solid rocket is ignited, all its fuel will be consumed, without any option for adjusting thrust (force). Liquid-rocket engines are more powerful than solid-rocket motors— they can generate more thrust—but the price of their power is complexity, in the form of many pipes, pumps, valves, gauges, and other parts. Liquid rockets require attention to storage issues and oftentimes the need to maintain very cold temperatures. Their complexity affects vehicle reliability, if only because the introduction of more components means the introduction of more opportunities for problems to occur.
Crosslink Winter 2000/2001 • 31
32 • Crosslink Winter 2000/2001
that any lessons learned from past failures are worth judiciously implementing if doing so can preclude future ones. Further Reading I-S. Chang, “Investigation of Space Launch Vehicle Catastrophic Failures,” AIAA Journal of Spacecraft and Rockets, Vol. 33, No. 2, 198–205 (March–April 1996). I-S. Chang, “Overview of World Space Launches,” AIAA Journal of Propulsion and Power, Vol. 16, No. 5 (September–October 2000). I-S. Chang, S. Toda, and S. Kibe, “Chinese Space Launch Failures,” ISTS paper 2000-g08, 22nd International Symposium on Space Technology and Science (Morioka, Japan, May 31, 2000).
The forces that drive the costs of today’s small satellites are very different from the forces that drive the costs of all other satellites. NASA and DOD needed a new model to gauge smallsatellite costs—and The Aerospace Corporation created one.
I-S. Chang, S. Toda, and S. Kibe, “European Space Launch Failures,” AIAA paper 20003574, 36th AIAA Joint Propulsion Conference (Huntsville, AL, July 17–19, 2000). R. Esnault-Pelterie, L’exploration par fusées de la trés haute atmosphére et la possibilité des voyages interplanétaires (Rocket Exploration of the Very High Atmosphere and the Possibility of Interplanetary Travel) (Société Astronomique de France, Paris, 1928). R. H. Goddard, “A Method of Reaching Extreme Altitudes,” Publication 2540, Smithsonian Miscellaneous Collections, Vol. 71, No. 2 (Smithsonian Institution, Washington, D.C., 1919). Also in The Papers of Robert H. Goddard, Vol. 1, Eds. Esther C. Goddard and G. Edward Pendray (McGraw-Hill, New York, 1970). S. J. Isakowitz, J. P. Hopkins, Jr., and J. B. Hopkins, International Reference Guide to Space Launch Systems, 3rd ed. (AIAA Publications, Washington, D.C., 1999). M. J. Neufeld, The Rocket and the Reich (Harvard University Press, Cambridge, MA, 1995). H. Oberth, Die Rakete zu den Planetenräumen (The Rocket into Interplanetary Space) (R. Oldenkourg, Münich, Germany, 1923; reprinted by Reproducktionsdruck von UniVerlag Feucht, Nürnberg, 1984). “Reliability Design and Verification for Launch-Vehicle Propulsion Systems,” AIAA Workshop on Reliability of Launch-Vehicle Propulsion Systems (Washington, D.C., July 25, 1989). K. E. Tsiolkovsky, A Rocket into Cosmic Space (published in 1903 and reprinted in Sobranie Sochinenie, Moscow: Izd. Akademii Nauk, USSR, 1951). Translated into English in Works on Rocket Technology (NASA, Washington, D.C., 1965).
Photo courtesy of NASA
taken to improve product reliability. The separation mechanism function should be confirmed with a full-size dummy booster. Hardware reworks should be minimized, and inspection testing should be tailored for specific reworks. When the system is operational, it is important to limit space launch operation to the design environment and flight experience. Prelaunch procedures and launch processes should be simplified to reduce contamination and damage in handling and processing. Launch-management training needs to be improved where possible. Finally, technical risks associated with schedule-driven launch dates should be reduced. Conclusion The technologies that have been developed for space applications and their spinoffs have dramatically improved human life, and they will no doubt continue to do so. Global high-speed telecommunication, videoconferencing, and Internet applications require many satellites, which means the need for launch services will continue to grow. In times of conflict as well as peacetime, space technology is of critical importance to the nation. Just as, more than half a century ago, the air advantage of the Allied Forces contributed significantly to the course of World War II, so in the future, space technology will have the ability to influence conflicts. The Falkland Islands War of 1982 and the Persian Gulf War of 1991 are two examples of how the intelligent utilization of space resources affects outcome. In coming years, needs for commercial and national defense space-related technologies are expected to multiply in many areas: propulsion, guidance and control, communications, navigation, tracking and data relay, weather forecast, remote sensing, surveillance, reconnaissance, early warning and missile defense, and interplanetary exploration. The demand for space launch services is ever increasing and may soon exceed the U.S. government’s defense budget. As more launches are conducted, more possibilities for failure will present themselves. The need for developing reliable launch systems will be ongoing. It is clear
Small-Satellite Costs David A. Bearden
ighly capable small satellites are commonplace today, but this wasn’t always the case. It wasn’t until the late 1980s that modern small satellites came on the scene. This new breed of low-profile, low-cost space system was built by maximizing the use of existing components and off-the-shelf technology and minimizing developmental efforts. At the time, many thought that because of their functional and operational characteristics and their low acquisition costs, these
small systems would become more prevalent than the larger systems built during the previous 30 years. But exactly which spacecraft fell into the new category? A precise description of small satellites, or “lightsats,” as they were also called, was lacking in the space literature of the day. The terms meant different things to different people. Some established a mass threshold (e.g., 500 kilograms) to indicate when a satellite was small; others used cost as a criterion; still
others used size. Even scarcer than good descriptions of small satellites, however, were guidelines for cost estimation of smallsatellite projects. Clearly, a more useful definition of small space systems was needed. By the 1990s, because of increased interest in small satellites for military, commercial, and academic research applications, the Air Force Space and Missile Systems Center (SMC) and the National Reconnaissance Office (NRO) asked The Aerospace Corporation for information about
Crosslink Winter 2000/2001 • 33
Spacecraft bus cost (FY97 dollars in millions)
DOD large satellites Modern DOD small satellites Traditional DOD small satellites
400 600 800 1000 1200 Spacecraft bus dry mass (kilograms)
Dollars-per-kilogram comparison of DOD large satellites (500 dollars per kilogram), modern small satellites (100 dollars per kilogram), and traditional small satellites (150 dollars per kilogram). Data points for these three categories cluster differently, and regression analysis shows that each set of points determines a different cost-estimating relationship. This information confirms the need for a new model using contemporary small satellites as its basis.
capabilities and costs of such systems. In response, Aerospace commissioned a study to compare cost and performance characteristics of small satellites with those of larger, traditional systems. Of specific interest was the ability to examine tradeoffs between cost and risk to allow assessment of how traditional risk-management philosophies might be affected by the adoption of smallsatellite designs. Estimating costs for small systems raised many questions. What parameters drove the cost of small satellites? Were traditional parameters known to drive the cost of large systems still applicable? How did small systems compare with large ones? Did small-satellite acquisition philosophies, which prompted reductions in levels of oversight, independent reviews, and paperwork, enable a reduction in cost-per-unit capability? What advantages might small
satellites offer for rapid incorporation of new technologies? Could they help reduce the long development cycle for military space programs? Were small satellites really economical for operational applications, such as navigation and communication? These questions led to a series of studies on technical and economic issues involved in designing, manufacturing, and operating small satellites. The studies found that existing spacecraft cost models, developed during the previous 30 years to support the National Aeronautics and Space Administration (NASA) and the Department of Defense (DOD), were of limited utility because of fundamental differences in technical characteristics and acquisition and development philosophies between small-satellite and traditional-satellite programs. This finding prompted NASA and DOD to seek cost-analysis methods and models
Total satellite cost (FY98 dollars in millions)
Traditional NASA missions NASA faster-better-cheaper missions 0
2000 3000 4000 5000 Satellite launch mass (kilograms)
This graph compares the dollars-per-kilogram ratio for traditional NASA missions (900 dollars per kilogram) with the ratio as noted in NASA’s faster-better-cheaper missions (120 dollars per kilogram). It’s clear that the different sets of data points determine markedly different cost-estimating regimes.
34 • Crosslink Winter 2000/2001
specifically tailored to small-satellite programs. To meet this need, Aerospace eventually developed the Small Satellite Cost Model, a small-satellite trade-study software tool that captures cost, performance, and risk information within a single framework. Before looking at the development of Aerospace’s trade-study tool, though, it will be valuable to backtrack to the late 1980s and review just exactly how smallspacecraft programs had been perceived. Streamlined Development Activities In the 1980s, the DOD Advanced Research Projects Agency and the United States Air Force Space Test Program served as the primary sources of funding for small satellites, which typically were used for technology experiments. The Space Test Program coordinated experimental payload flights for the Army, Navy, Air Force, and other government agencies. Reduced development complexity and required launch-vehicle size enabled affordable, frequent access to space for research applications. Relatively low acquisition costs and short development schedules also allowed university laboratories to participate, providing individual researchers access to space—a privilege previously reserved only for well-funded government organizations. Small satellites were procured under a specifically defined “low cost” philosophy. They were smaller in size and were built with maximum use of existing hardware. A smaller business base (i.e., a reduced number of participating contractors) was involved in the development process, and it was perceived that small satellites represented a niche market relative to the more prevalent large systems. Mission timelines from approval to launch were on the order of 24 to 48 months, with an on-orbit life of 6 to 18 months. Launch costs, either for an existing dedicated small launcher or for a secondary payload on a large launcher, remained high, but developments such as the Pegasus air-launched vehicle and new small launchers (such as Taurus and Athena) offered promise of lowering these costs. Additionally, small-satellite flight and ground systems typically used the most mature hardware and software available to minimize technology-development and flight-certification costs. Emerging advances in microelectronics, software, and lightweight components enabled system-level downsizing. Spacecraft often cost more than $200 thousand per
Cost model estimate/actual cost (percent)
STEP 1 Satellite
A cost-percentage comparison that makes use of an older model and the updated dollars-per-kilogram relationships shown in previous graphs to estimate modern small-satellite costs. Each bar’s height represents the percentage difference between a satellite’s estimated cost and its actual cost. Thus for Clementine, with a percentage of 109%, the older model’s estimate was twice
RADCAL SAMPEX Clementine
the actual cost, and for RADCAL, with a percentage of 801%, the older model’s estimated cost was nine times the actual cost. Because the estimates far outweighed the real cost in many cases, the chart illustrates the inadequacy of a traditional cost model for modern small satellites.
money, small-spacecraft programs came to be perceived as fast-paced, streamlined development activities. Dedicated project leaders with small teams were given full technical and budgetary responsibility with goals tailored around what could be done inexpensively on a short schedule. Fixedprice contracts became the norm, and requirement changes (and associated budgetary adjustments) were held to a minimum. The Next Decade With the advent of the 1990s came a movement toward realizing routine access to
space. The development of a broad array of expendable launch vehicles provided increased access to orbit for many different kinds of payloads. Satellite programs attempted to incorporate advanced technology and demonstrate that fast development cycles, low acquisition costs, and small flexible project teams could produce highly useful smaller spacecraft. This different paradigm opened up new classes of space applications, notably in Earth science, commercial mobile-communications, and remote-sensing arenas.
Cost vs. Pointing Accuracy
y = 1.67 + 12.98 x –0.5 16 data points Min = 0.25 Standard
Spacecraft cost (FY94 dollars in millions)
Spacecraft cost (FY94 dollars in millions)
kilogram and could reach $1 million per kilogram with delivery-to-space costs included. With miniaturization, every kilogram saved in the spacecraft bus or instruments represented a possible saving of up to five kilograms in launch, onboard propulsion, and attitude-control systems mass. Reduced power demands from microelectronics, instruments, and sensors could produce similar payoffs. For interplanetary missions, reduced mass had the capability to produce indirect cost savings through shorter transit times and mission duration. All this downsizing eliminated the need for large facilities and costly equipment such as high bays, clean-room areas, test facilities, and special handling equipment and containers. Engineering development units—prototypes built before the actual construction of flight hardware—were not built; instead a protoflight approach was favored, where a single unit served as both the engineering model and the actual flight item. Quality parts were used where possible, but strict adherence to rigid military specifications was avoided. Redundancy—the use of multiple components for backup in the event the primary component fails—was also avoided in favor of simpler designs. Designers relied on multifunctional subsystems and software to allow operational work-arounds or alternate performance modes that could provide functionality if something went wrong during a mission. As a result of these unorthodox approaches that sought ways to save time and
Cost vs. Satellite Dry Mass y = 0.7 + 0.023 x 1.26 20 data points Min = 20, Max = 210 Standard error = 3.3 $M
Spacecraft cost (FY94 dollars in millions)
Cost vs. End-of-life Power
y = 0.507 + 1.55 x 0.452 14 data points Min = 5, Max = 440 Standard error = 6.2 $M
EOL 5.00Power (W)
200.0 300.0 EOL power (W)
System-level cost-estimating relationships that were developed for early versions of the Small Satellite Cost Model. The first cost-estimating relationships related total spacecraft bus cost to individual parameters such as mass, power, or pointing accuracy. These were the early predecessors of today’s more sophisticated cost model that represents costs at the subsystem level utilizing a variety of cost drivers.
Crosslink Winter 2000/2001 • 35
1. Define problem
2. Examine existing cost models • Determine utility • Determine adaptability
3. Collect cost and technical data • Identify potential cost drivers • Normalize cost data
5. Develop cost model • Perform parametric weighting • Perform statistical analysis
4. Do parametric analysis $ • Perform regression • Identify correlation Mass • Consider multiple parameters
7. Compare with known costs
6. Validate cost model
9. Deliver model to user community
8. Apply model • Use in trade study • Use in cost analysis • Consider sensitivities
The cost modeling process. This is an ongoing iterative process that involves collecting data and performing regression analysis to arrive at cost-estimating relationships. The data are validated against actual program costs. The model is delivered to the users for trade analyses.
Small-spacecraft designers, in their quest to reduce costs through use of off-the-shelf technology, in many cases pioneered the use of microcircuitry and other miniaturized devices in space. Whereas small satellites had been unstabilized, battery-powered, single-frequency, store-and-forward spacecraft with limited applicability to operational space endeavors, the level of functionality achievable in small spacecraft took a dramatic leap forward in the early 1990s, mainly because of the availability of increased space-compatible computational power and memory. These advances led to the current rich suite of spacecraft bus capabilities and the large array of missions using small spacecraft. The trend toward cost reduction and streamlined management continued to gain momentum with increased interest in small spacecraft from NASA and DOD. A shift in philosophy, where a greater tolerance for risk was assumed, was evident in programs like the NASA-sponsored Small and Medium Explorer Programs, the Ballistic
36 • Crosslink Winter 2000/2001
Missile Defense Organization-sponsored Clementine, DOD-sponsored Space Test Experiment Platforms, and the Small Satellite Technology Initiative’s Lewis and Clark, among others. The end of the Cold War (in 1991) and the drive toward reduced development and launch costs created a political and budgetary imperative where small satellites were viewed as one of the few vehicles available for science and technology missions. In response to budget pressures and in the wake of several highly publicized lost or impaired billion-dollar missions, NASA’s administrator Dan Goldin in 1992 embraced small spacecraft and promoted the notion of a “faster-better-cheaper” approach for many of NASA’s missions. The programs implemented as a result of this tactic dictated faster and cheaper by specifying launch year and imposing a firm funding cap. These constraints laid the groundwork for what would become a decade of ongoing controversy about the definition and success of faster-better-cheaper.
The Need for a New Model It was against this backdrop that Aerospace began collecting a large body of information concerning technologies and programmanagement techniques that affected small-satellite cost-effectiveness. The programmatic aspects of traditional satellite programs (e.g., long schedules, large amounts of documentation, rigorous procedures, and risk abatement) were known to dramatically affect cost. In particular, two distinct but interrelated factors drove the cost of the system: what was built and how it was procured. In many cases, how the system was procured appeared to be as important as what was procured because cost and schedule were dependent on the acquisition environment. A study that compared spacecraft mass versus cost for traditional small spacecraft of the 1960s and 1970s, traditional large spacecraft of the 1970s and 1980s, and modern (post-1990) small spacecraft revealed two important messages. First, the modern small spacecraft differed dramatically from traditional large spacecraft as well as their similarly sized cousins of the past. It was postulated that the latter difference, as evidenced by cost reduction, was the result of a combination of new business approaches and advanced technology. Second, cost and spacecraft sizing models based on systems or technologies for traditional spacecraft were inappropriate for assessing modern small satellites. This was an understandable departure from traditional-spacecraft cost trends. New developments in technology are often based on empirical models that characterize historical trends, with the assumption that future missions will to some degree reflect these trends. However, in cases where major technology advancements are realized or where fundamental paradigms shift, assumptions based on traditional approaches may not apply. It became clear that estimating small-system costs was one such case. Early small-satellite studies showed that cost-reduction measures applied to smallsatellite programs resulted in system costs substantially lower than those estimated by traditional (primarily mass-based) parametric cost-estimating relationships (equations that predict cost as a function of one or more drivers). The studies analyzed the applicability of available cost models such as the Air Force Unmanned Spacecraft Cost Model and the Aerospace
Three-dimensional cost-estimating relationships (CER). Later versions of the Small Satellite Cost Model used multiparameter cost-estimating relationships derived at the subsystem level. Emphasis was placed on a combination of mass- and performancebased cost drivers.
16000 14000 12000 10000 8000 6000 4000 2000 0
Structures subsystem CER (composite)
100 440 780 1120 1460 1800 Beginning-of-Life power (watts)
Satellite Cost Model to predict costs of small satellites. These cost models—based on historical costs and technical parameters of traditional large satellites developed primarily for military space programs—were found inappropriate for cost analyses of smallsatellite programs. It became readily apparent in comparing actual costs against costs estimated by these models that a new model, dedicated to this new class of mission, was needed. Credible parametric cost estimates for small-satellite systems required new cost-estimating relationships derived from a cost and technical database of modern small satellites. The Making of a Model Developing a small-satellite cost model that related technical parameters and physical characteristics to cost soon became the primary objective of small-satellite studies. To accomplish this, a broad study of small satellites was performed, with emphasis on the following tasks: • definition of small satellite and identification of small-satellite programs • collection of small-satellite cost and technical data from the Air Force, NASA, and university, government laboratory, and industry sources • examination of cost-reduction techniques used by small-satellite contractors and sponsors • performance of parametric analysis to determine which factors should be used in the derivation of cost-estimating relationships by using best-fit regressions on data where cost correlation is evident
54 16000 42 14000 12000 10000 8000 6000 4000 2000 0
Structures cost (FY97 dollars in thousands)
TTC/CDH cost (FY97 dollars in thousands)
Telemetry, Tracking, Communications/Command and Data Handling subsystem CER (S-band)
180 120 60 0
90 Structures mass (kilograms)
• development and validation of a cost model with parametrics and statistics; evaluation of the cost model by performance of cost and cost-sensitivity analyses on small-satellite systems under development • creation of a corporate knowledge base of ongoing small-satellite activities and capabilities, technology-insertion opportunities, and project histories for lessons learned, systems studies, etc. • maintenance of a corporate presence in the small-satellite community to advise customers about relevant developments • development of a cadre of people with expertise and tools for continued studies of the applicability of small satellites to military, civil, and commercial missions The cost-modeling process entailed aggressive data acquisition through collaboration with organizations responsible for developing small satellites. One unanticipated challenge was actually gaining access to cost data. Small-satellite contractors, in their quest to reduce costs, would often not be contractually bound to deliver detailed cost data, so in many cases costs were not available. Despite this difficulty, Aerospace collected data over a period of two to three years for about 20 smallsatellite programs at the system level (i.e., total spacecraft or spacecraft bus costs only). From this initial database, analysts derived parametric costing relationships as
Solar-array area (square meters)
a function of performance measures and physical characteristics. The model estimated protoflight development costs and cost sensitivities to individual parameters at the system level. The model was of great value in instances where evaluations needed to be performed on varying proposals with differing degrees of detail or when limited information was available, as is often the case in an early concept-development phase. The Second-Generation Cost Model While initial system-level small-satellite studies were sponsored by DOD and internal Aerospace research and development, in 1995, the need to respond to increasingly frequent questions about NASAsponsored small-satellite architectures and a need for refined small-satellite system analysis at the subsystem level prompted NASA to seek better cost-analysis methods and models specifically tailored to smallsatellite programs. Consequently, NASA asked Aerospace to gather information regarding capabilities and costs of small satellites and to develop a set of subsystem-level small-satellite cost-estimating relationships. To allow assessment of a complete spacecraft bus cost, Aerospace collected more data in order to be able to derive costestimating relationships for each of the spacecraft bus subsystems: Crosslink Winter 2000/2001 • 37
Launch Mass Launch (kilograms) Date
NASA Small Planetary Satellites Clementine NEAR MGS Mars Pathfinder ACE Lunar Prospector DS1 MCO MPL Stardust
BMDO/NASA NASA NASA NASA NASA NASA NASA NASA NASA NASA
NRL JHU/APL Lockheed Martin JPL JHU/APL Lockheed Martin JPL/Spectrum Astro JPL/Lockheed Martin JPL/Lockheed Martin JPL/Lockheed Martin
494 805 651 890 785 296 486 629 576 385
Jan 94 Feb 96 Nov 96 Dec 96 Aug 97 Jan 98 Oct 98 Dec 98 Jan 99 Feb 99
Titan II Delta II Delta II Delta II Delta II Athena II Delta II Delta II Delta II Delta II
Lunar mapping Asteroid mapping Mars mapping Mars lander and rover Low energy particle Lunar science Technology demo Mars remote sensing Mars science Comet sample return
Jul 92 Apr 95 Oct 95 Jul 96 Aug 96 Nov 96 Nov 96 Aug 97 Feb 98 Apr 98 cancelled Jul 98 Dec 98 Mar 99 May 99 Jun 99 Jun 99 Dec 99 Mar 00
Scout Pegasus Conestoga Pegasus XL Pegasus XL Pegasus XL Pegasus XL Pegasus XL Pegasus XL Pegasus XL Athena I Athena I Pegasus XL Pegasus XL Pegasus XL Delta II Titan II Taurus Delta II
Science experiments Lightning experiment Microgravity experiments Ozone mapping Auroral measurements High energy experiments Science experiments Ocean color Space physics Solar coronal Science experiments Hyperspectral imaging Astronomy Astronomical telescope Space physics Space science Ocean wind measure Sun-Earth atmosphere Neutral atom/UV measure
Nov 85 Apr 90 Apr 90 Apr 90 Apr 90 May 90 Jun 91 Jul 91 Nov 91 Nov 92 Apr 93 Jun 93 Mar 94 Mar 94 May 94 May 94 Jun 94 Aug 94 Jul 95 Mar 96 May 96 Aug 97 Oct 97 Feb 98 Jul 98 Sep 98 Jun 00
Shuttle Pegasus Atlas I Atlas I Atlas I Scout Scout Delta II Pegasus Scout Pegasus Scout Taurus Taurus Scout Pegasus Pegasus XL Pegasus Pegasus XL Pegasus XL Pegasus Pegasus XL Pegasus XL Taurus STS/GAS Taurus Pegasus XL
Message relay Message relay Geomagnetic survey Communications Communications Communications Radiation Sensor experiments Communications Sensor experiments X-ray mapping Radar calibration tests Classified Autonomy experiments Sensor experiments Signal detect/modulation Atmospheric physics Power experiments Science/communications Radiation Hyperspectral imaging Science Science/communications Radar altimetry Science Tether experiment Remote sensing
NASA Earth-Orbiting Small Satellites SAMPEX MICROLAB METEOR TOMS-EP FAST HETE SAC-B Seastar SNOE TRACE Clark Lewis SWAS WIRE TERRRIERS FUSE QuikSCAT ACRIMSat IMAGE
NASA NASA/Orbital NASA NASA NASA NASA CONAE/NASA NASA NASA GSFC NASA NASA NASA NASA NASA, JPL NASA, BU NASA, APL NASA, NOAA NASA, JPL NASA GSFC
NASA GSFC Orbital CTA (Orbital) TRW NASA GSFC MIT/AeroAstro CONAE Orbital LASP NASA GSFC CTA (Orbital) TRW NASA GSFC NASA GSFC AeroAstro, LLC Orbital JPL/Ball Aerospace Orbital Lockheed Martin
161 75 364 295 191 128 191 372 132 250 266 385 288 255 288 1360 870 115 494
DSI (Orbital) DSI (Orbital) DSI (Orbital) DSI (Orbital) DSI (Orbital) DSI (Orbital) DSI (Orbital) Ball Aerospace DSI (Orbital) JPL/Spectrum Astro AeroAstro DSI (Orbital) Ball Aerospace TRW Spectrum Astro TRW TRW Orbital TRW DSI (Orbital) Spectrum Astro LANL/SNL TRW Ball Aerospace CTA (Orbital) Lockheed Martin Orbital
71 68 68 56 67 61 77 76 26 144 113 91 161 489 170 180 352 209 295 110 212 210 386 357 69 691 249
Other U.S.-Built Small Satellites GLOMR I PEGSAT POGS/SSR SCE TEX MACSAT REX LOSAT-X MICROSAT MSTI-1 ALEXIS RADCAL DARPASAT STEP 0 MSTI-2 STEP 2 STEP 1 APEX STEP 3 REX II MSTI-3 FORTE STEP 4 GFO MIGHTYSAT STEX TSX-5
DARPA DARPA STP, ONR ONR STP, ONR DARPA STP SDIO DARPA BMDO DOE STP, NRL DARPA/AF STP SDIO STP STP STP STP STP SDIO DOE STP U.S. Navy STP NRO BMDO/STP
38 • Crosslink Winter 2000/2001
• • • • • • •
attitude determination and control propulsion electrical power supply telemetry, tracking, and command command and data handling structure, adapter thermal control Emphasis was placed on obtaining data on spacecraft bus subsystem characteristics. In addition to technical data, costs in
ability to relate cost to those characteristics. Programs either already completed or awaiting launch in the next year were targeted. Because Aerospace operates a federally funded research and development center, it was in a unique position to receive proprietary data from private companies and enter it into a special-purpose database to support government space-system acquisition goals and provide value added to the industry as a whole. Proprietary informa-
SSCM98 DP Cost Input Data Code*
Cost Driver (Units)
Single contractor # Major Contractors Mission Orbit Earth Orbiting Apogee (km) Design life (months) Satellite Wet Mass (kg) Satellite Dry Mass (kg) 3-axis Control Stabilization Type Pointing Accuracy (degrees) Pointing Knowledge (degrees) Attitude Control Subsystem Mass (kg) Number of thrusters Cold Gas Fuel Type Propulsion Subsystem Dry Mass (kg)
1 1 600 48 805. 468.1 2 0.6 0.029 33.9 1 1 118.
Hydrazine propulsion type
2 1,880 105 2 103.2
Re He Tp,Ip R,L,Pp Ap,Ip,Se,Pp L A-spin A-3 axis Ap-3 axis Pe Pp
Ep,Tp Ie E Hp He Hp E S T Te,Le Ee Hp Sp Sp He
Beginning-of-Life Power (W) End-of-Life Power (W) Solar Array Cell type Power Subsystem Mass (kg) Solar Array mounting type Battery Capacity (A-hrs) Battery Type
Gallium Arsenide (GaA Deployable
1 9. 1 8.9 1 256. 5. 43.6 2 102.2 3.
Solar Array Area (m2) UHF/VHF Communications Band Downlink Data Rate (kbps) Transmitter Power (W) TT&C/C&DH Subsystem Mass (kg) Composite Structures material Structures Mass (kg) Thermal Subsystem Mass (kg)
Valid Range Low
400 6 75 26
4,200 60 489 410
1 0.016 2.4 0
5 1.5 59 8
64.6% 14.2% -40.0%
ERROR: Cold gas cannot be dual! 57 1,880 8 740
SMALL SATELLITE COST MODEL (SSCM) 7
Developed for NASA Headquarters 2.4 1 by 13
3,000 18 Bearden, 115
N.Y. Lao, T.J. Mosher, and J.J. Muhle
' 20 1998, The 183 Aerospace Corporation. All Rights Reserved. 0.1
This model may not be redistributed without the expressed written consent of The Aerospace Corporation Legend required input do not change! do not change!
Percentiles of Cost
Estimated Cost (FY97$K)
60% 65% 70%
39,945 40,780 41,679
These selected screen shots from the Small Satellite Cost Model demonstrate the parametric costestimating model. The model is easy to work with and provides useful outputs such as cost probability distributions.
the areas of spacecraft integration, assembly and test, program management and systems engineering, and launch and orbital operations were requested. To gather information on the state of the industry as a whole, as well as specific data, analysts surveyed and interviewed contractors who build small satellites or provide small-satellite facilities (e.g., components, launchers). A cost and technical survey sheet was distributed to virtually every organization and contractor in the small-satellite industry. It was important to obtain information about mass, power, performance, and other technical characteristics because the development of credible subsystem-level cost analyses of smallsatellite missions depends on the analyst’s
tion delivered to the corporation was treated in a restricted manner, used only for the purpose intended, and not released to organizations, agencies, or individuals not associated with the study team. The information was used exclusively for analysis purposes directly related to cost-model development. Only derived information depicted in a generalized manner was released, and the database itself has remained proprietary. In some cases, formal nondisclosure agreements between the companies and Aerospace were necessary to facilitate delivery of proprietary data. After properly categorizing cost data, adjusting it for inflation, and breaking it out on a subsystem basis, analysts developed cost-estimating relationships for each
of the subsystems, using a subset of the more than 70 technical parameters collected on each of the small satellites. The effort to develop a cost-estimating relationship for a small-satellite subsystem took full advantage of advanced developments in regression techniques. Choosing the proper drivers involved combining a knowledge of statistics, sound engineering judgment, and common sense. Graphics software tools assisted in the development of these cost-estimating relationships, enabling the analyst to view the shape of a function against its data points and to identify the function (whether linear, logarithmic, exponential, or some other form). The end product was a set of subsystemlevel bus-related cost-estimating relationships based entirely on actual cost, physical, and performance parameters of 15 modern small satellites. This was a major advancement over available tools for estimating small-satellite costs. Analysts also developed factors to use in estimating both recurring and nonrecurring costs of bus subsystems, to enable studies of multiple builds—such as the ones that are needed for constellations of small satellites. The cost-estimating relationships enabled the inclusion of cost as a variable in system design tools. They were also incorporated into a stand-alone, menu-driven computerized model that could be distributed to government organizations and private companies that contributed data. Cost Model Leaves Earth Orbit In 1996, NASA was moving to smaller platforms for planetary exploration. This movement afforded an important application for the Small Satellite Cost Model. Following well-publicized problems with the Galileo and Mars Observer spacecraft, there had emerged in the early 1990s a growing apprehension in the NASA planetary science community that opportunities for planetary science data return were dwindling. After Galileo was launched in 1989, the next planetary mission scheduled was Cassini, which would launch in October 1997 and begin returning data in 2003, a full six years after Galileo had stopped sending data. Since a steady stream of new data is important to maintaining a vigorous program of planetary and scientific investigation, the situation was naturally a cause for concern. Out of this concern emerged a new NASA small-spacecraft program called Discovery.
Crosslink Winter 2000/2001 • 39
40 • Crosslink Winter 2000/2001
Missions Evaluated by the Small Satellite Cost Model Launched in February 1999, Stardust is journeying to the comet Wild-2. It will arrive in 2004 and, during a slow flyby, will collect samples of dust and gas in a low-density material called aerogel. The samples will be returned to Earth for analysis in 2006. Stardust will also photograph the comet and do chemical analysis of particles and gases. The JPL/Lockheed Martin-built spacecraft is one of NASA’s Discovery Program missions. It is the first NASA mission dedicated to exploring a comet and the first U.S. mission launched to robotically obtain samples in deep space. NEAR was launched in February 1996. Its objective was to orbit the asteroid Eros for one year starting in January 1999, collecting scientific data. Developed in 29 months at JHU/APL, NEAR was part of the NASA Discovery Program. Its payload was composed of a multispectral imager, a laser rangefinder, an X-ray/gamma-ray spectrometer, and a magnetometer. A software-error-induced burn abort that occurred in December 1998 resulted in delaying the rendezvous and subsequent data acquisition until February 2000. The Lunar Prospector, a NASA-sponsored lunar polar orbiting probe developed by Lockheed Martin, was launched aboard Athena II in January 1998. Its primary mission was to map the moon’s chemical, gravitational, and magnetic properties. Data from instruments, including a gamma-ray spectrometer, a neutron spectrometer, an alpha particle spectrometer, a magnetometer, and an electron reflectometer, were used to construct a map of the surface composition of the moon.
Illustrations courtesy of NASA
The Discovery program’s primary goal was to conduct frequent, highly focused, cost-effective missions to answer critical questions in solar-system science. Formally started under NASA’s fiscal-year 1994 budget, the Discovery program featured small planetary exploration spacecraft—with focused science goals—that could be built in 36 months or less and would cost less than $150 million (fiscal year 1992), not including the cost of the launch vehicle. To apply its cost model to this new domain, Aerospace performed, in collaboration with Johns Hopkins University’s Applied Physics Laboratory (JHU/APL), a cost-risk assessment of the Near Earth Asteroid Rendezvous (NEAR) mission. This mission, one of NASA’s first two Discovery missions, was designed to leave Earth orbit on a trajectory to the near-Earth asteroid Eros. The study identified a number of limitations in applying the Small Satellite Cost Model to interplanetary missions. Out of this information came a concerted effort to gather data on small interplanetary missions to enhance the model. Analysts collected data on missions such as Mars Pathfinder, Lunar Prospector, Clementine, and Stardust, developing cost-estimating relationships appropriate to a Discovery-class mission. Less than a year later the model was again applied successfully to the Near Earth Asteroid Rendezvous spacecraft, demonstrating cost estimates within a few percent of the actual costs. Once Aerospace demonstrated the ability to assess small interplanetary mission costs, NASA’s Langley Research Center Office of Space Science asked the corporation to participate in the Discovery mission evaluation process. Aerospace evaluated 34 Discovery proposals submitted by government, industry, and university teams. These proposals included a wide variety of payloads (including rovers, probes, and penetrators)—more than 120 in all. The goals were to provide independent cost estimates for each proposal, identify cost-risk areas, determine cost-risk level (low, medium, or high) for each proposal, and evaluate proposals in an efficient and equitable manner. Five finalists were selected. In 1997, as a follow-on to the successful Discovery mission evaluation, the NASA Office of Space Science asked Aerospace to assist in the selection of Small Explorer missions. This was a series of small, lowcost interplanetary and Earth-orbiting
Mars Pathfinder, the second launch in the Discovery Program developed by JPL, consists of a cruise stage, entry vehicle, and lander. The mission of Mars Pathfinder was to test technologies in preparation for future Mars missions, as well as to collect data on the Martian atmosphere, meteorology, surface geology, and rock and soil composition. On July 4, 1997, Mars Pathfinder successfully landed on Mars and subsequently rolled out the Sojourner rover to analyze native rock composition.
science missions designed to provide frequent investigative opportunities to the research community. Aerospace served on the Technical, Management, and Cost review panel. Fifty-two Small Explorer mission concepts were evaluated, from which two final missions were chosen. NASA commended Aerospace for its work on Discovery and Small Explorer missions. Because of the work it had done on these programs, Aerospace was invited to
participate in a National Research Council workshop, from which a report titled “Reducing the Costs of Space Science Research Missions” was generated. Aerospace was also invited to join the editorial board of a new international peer-reviewed technical journal (Reducing Space Mission Cost, published by Kluwer Academic Publishers) and to become a member of the Low Cost Planetary Missions Subcommittee of the International Academy of Astronautics Committee on Small Satellite Missions.
The ACE (Advanced Composition Explorer) spacecraft carried six high-resolution sensors, mainly spectrometers, and three monitoring instruments. It collected samples of low-energy solar and high-energy galactic particles and measured conditions of solar wind flow and particle events. An Explorer mission sponsored by the NASA Office of Space Science and built by JHU/APL, ACE orbits the L1 libration point, a location 900,000 miles from Earth where the gravitational effects of Earth and the sun are balanced, to provide near-real-time solar wind information. SWAS (Submillimeter Wave Astronomy Satellite) was the third NASA Small Explorer mission. It was launched aboard a Pegasus XL rocket in December 1998. The overall goal of the mission was to understand star formation by using a passively cooled Cassegrain telescope to determine the composition of interstellar clouds and establish the means by which these clouds cool as they collapse to form stars and planets. SWAS observed water, molecular oxygen, isotopic carbon monoxide, and atomic carbon. Launched in October 1998, DS1 (Deep Space 1) was a NASA New Millennium Program mission. Its objective was to validate several technologies in space including solar electric propulsion and autonomous navigation. Instruments on board included a solar concentrator array and a miniature integrated camera and imaging spectrometer. The spacecraft, built by JPL and Spectrum Astro, was designed to monitor solar wind and measure the interaction with targets during flybys of an asteroid and a comet.
The primary mission objectives of the Clementine lunar orbiter, launched in January 1994 aboard Titan IIG, were to investigate long-term effects of the space environment on sensors and spacecraft components and to take multispectral images of the moon and the near-Earth asteroid Geographos. The Naval Research Laboratory-built spacecraft incorporated advanced technologies, including Lawrence Livermore National Laboratory lightweight sensors. After Clementine completed lunar mapping, its onboard computer malfunctioned on departure from lunar orbit and depletion of onboard fuel resulted.
Examining the Faster-BetterCheaper Experiment Successful NASA programs such as the Mars Pathfinder and the Near Earth Asteroid Rendezvous mission effectively debunked the myth that interplanetary missions could only be accomplished with billion-dollar budgets. They set a new standard against which all later missions were not only forced to measure up but go beyond. Designers were asked to meet unrelenting mission objectives within rigid cost
and schedule constraints in an environment characterized by rapid technological improvements, immense budgetary pressure, downsizing government, and distributed acquisition authority. As a result of these constraints, NASA had greatly increased its utilization of small spacecraft to conduct low-cost scientific investigations and technology demonstration missions. The original tenets of the small-satellite paradigm, including low cost, maximum use of existing components
and off-the-shelf technology, and reduced program-management oversight and developmental effort, had been applied to increasingly more ambitious endeavors with increasingly demanding requirements. This move had clearly benefited the scientific community by greatly diversifying the number and frequency of science opportunities. A number of failed small scientific spacecraft, however, such as Small Satellite Technology Initiative’s Lewis and Clark, and the Wide-field Infrared Experiment, fueled an ongoing debate on whether NASA’s experiment with faster-bettercheaper missions was working. The loss of the Mars Climate Orbiter and the Mars Polar Lander within a few months of each other sent waves of anxiety throughout government and industry that the recipe for successful faster-better-cheaper missions had been lost. Impaired missions or “near misses,” such as the Mars Global Surveyor, contributed to the debate as well, and many wondered whether programs currently on the books or late in development were too ambitious for the time and money they had been allotted. At the heart of the matter was allocation of cost and schedule. Priorities had changed. During the last few years the traditional approach to spacecraft design, driven by performance characteristics and high reliability to meet mission objectives, had completely given way to developments dominated by cost- and schedule-related concerns. While it was readily apparent that the faster-better-cheaper strategy resulted in lower costs per mission and shorter absolute development times, these benefits may have been achieved at the expense of reduced probability of success. Some questions lingered. When was a mission too fast and too cheap with the result that it was prone to failure? Given a fixed amount of time and money, what level of performance and technology requirements would cause a mission to stop short of failure due to unforeseen events? Risks often do not manifest ahead of time or in obvious ways. However, when examined after the fact, mission failure or impairment is often found to be the result of mismanagement or miscommunication in fatal combination with a series of lowprobability events. These missteps, which often occur when a program is operating near the budget ceiling or under tremendous schedule pressure, result in failures caused by lack of sufficient resources to Crosslink Winter 2000/2001 • 41
Complexity index 0–0.33
Complexity index 0.67–1
Small payload mass (~5–10 kilograms) One payload instrument Spin or gravity-gradient stabilized Body-fixed solar cells (silicon or gallium arsenide) Short design life (~6–12 months) Single-string design Aluminum structures Coarse pointing accuracy (~1–5 degrees) No propulsion or cold-gas system Low-frequency communications Simple helix or patch low-gain antennas Low data rate downlink (~1–10 kilobits per second) Low power requirements (~50–100 watts) No deployed or articulated mechanisms Little or no data storage No onboard processing (“bent-pipe”) Passive thermal control using coatings, insulation, etc.
thoroughly test, simulate, or review work and processes. Having maintained an extensive historical database of programmatic information on NASA faster-better-cheaper missions to support the Small Satellite Cost Model development, Aerospace was well positioned to examine the situation. With a decade of experience and more than 40 scientific and technology demonstration spacecraft flown, sufficient information existed for use in conducting an objective examination. To understand the relationship between risk, cost, and schedule, Aerospace analyzed data for missions launched between 1990 and 2000, using technical specifications, costs, development time, and operational status. The study examined the faster-bettercheaper strategy in terms of complexity measured against development time and cost for successful and failed missions. The failures were categorized as partial, where the mission was impaired in some way;
42 • Crosslink Winter 2000/2001
Large payload mass (~200–500 kilograms) Many (5–10) payload instruments Three-axis stabilized using reaction wheels Deployed sun-tracking solar panels (multijunction cells or concentrator) Long design life (~3–6 years) Partially or fully redundant Composite structures Fine pointing accuracy (~0.01–0.1 degrees) Monopropellant or bipropellant system with thrusters (4–12) High-frequency communications Deployed high-gain parabolic antennas High data rate downlink (thousands of kilobits per second) High power requirements (~500–2000 watts) Deployed and/or articulated mechanisms Solid-state data recorders (up to 5 gigabytes) Onboard processing (up to 30 million instructions per second) Active thermal control using heat pipes, radiators, etc.
catastrophic, where the mission was lost completely; or programmatic- or launchrelated, where the mission was never realized because of cancellation or failure during launch. A complexity index was derived from performance, mass, power, and technology choices, as a top-level representation of the system for purposes of comparison. Complexity drivers (a total of 29) included subsystem technical parameters (such as mass, power, performance, pointing accuracy, downlink data rate, technology choices) and a few general programmatic factors such as heritage (reuse of a part that flew on a previous mission) and redundancy policy. The process used to estimate spacecraft complexity included the following steps. • Identify parameters that drive or significantly influence spacecraft design. • Quantify the parameters so that they can be measured. • Combine the parameters into an average complexity index (expressed as a value between zero and one).
To determine whether the faster-bettercheaper experiment was successful, analysts plotted a comparison of complexity relative to development time and cost, noting failures. Some interesting trends emerged. Correlation between complexity, cost, and schedule was evident. A threshold, or “no-fly zone,” was apparent where project resources (time, funds) were possibly insufficient relative to the complexity of the undertaking. While it is unknown whether allocation of additional resources would have increased the probability of success of a given mission, this much is clear: When a mission fails or becomes impaired, it appears that it is too complex relative to its allocated resources. The observation of a correlation between cost and development time and complexity, based on actual program experience (i.e., actual costs incurred and development time required as opposed to numbers used during the planning phase), is encouraging because this model can be applied to future systems. The index may
200 Successful Failed Impaired
Total spacecraft cost (FY00 dollars in millions)
Development time (months)
72 48 24 0
Successful Failed Impaired
0.3 0.4 0.5 0.6 0.7 Complexity index
Cost and schedule plotted against a complexity index derived from performance, mass, power, and technology choices. The regression curves may be used to determine the level of complexity possible for a set budget or development time. Although the complexity index does not identify the
reveal a general risk of failure, but it won’t necessarily specify which subsystem might fail or how it will fail. Nevertheless, it does identify when a new mission under consideration is in a regime occupied by failed or successful missions of the recent past. This process should allow for more informed overall decisions to be made for new systems being conceived. Conclusion In summary, early small-satellite studies showed that older cost-estimation models based on historical costs and technical parameters of large satellite systems could not be successfully applied to small systems. It was necessary to develop a model that would be tailored specifically to this new category of spacecraft. To this day, there remains no formally agreed-upon definition of “small spacecraft,” although such spacecraft are typically considered to be Discovery-class in size or less (i.e., for interplanetary applications, they fit on a Delta II launch vehicle; for Earth-orbiting applications, they weigh less than 500 kilograms), and most are budgeted in the $50to $250-million range. Aerospace has been studying small satellites since 1991, and the main product of its ongoing work is the Small Satellite Cost Model. Based on actual physical and performance parameters of small Earthorbiting and interplanetary spacecraft flown during the last decade, this software tool was developed to estimate the cost of a small spacecraft. It has addressed many of the questions that were originally raised about cost estimation for small systems. The model is used in assessment of smallsatellite conceptual designs, technology
0.3 0.4 0.5 0.6 0.7 Complexity index
manner or subsystem in which a failure is likely to occur, it does identify a regime by which an index calculated for a mission under consideration may be compared with missions of the recent past.
needs, and capabilities, and it is continually updated to model state-of-the-art systems. The Small Satellite Cost Model has developed through several generations, with additions to the database and improvements to the cost-estimating relationships serving as the primary drivers from version to version. Currently, the small-satellite database has evolved to include more than 40 programs. While initial small-satellite studies were funded by DOD and Aerospace internal funds in the early 1990s, the development of Small Satellite Cost Model version 1.0 was funded by NASA in 1995. The database and model were updated in 1996 (version 2.0), and interplanetary capabilities were added in 1998 (version 3.0 or Small Satellite Cost Model 98). The fourth release is now in development. A version of the model is supplied to industry members who provide data that can improve it. There is also a publicly available version (see www.aero.org/software/sscm). The model is used extensively by DOD, JPL, and virtually all of the NASA field centers. It has become the standard for parametric evaluation of small missions worldwide and is used by the European Space Agency and the French Centre National d’Etudes Spatiales, among other organizations. A number of foreignbuilt small spacecraft are included in the database. The most recent application of the small-satellite study is the assessment of NASA’s approach, under constrained budgets and rigid schedules, to conduct faster-better-cheaper scientific investigations. Recent instances of failed or impaired spacecraft have brought into question the faster-better-cheaper strategy,
especially for interplanetary applications. While missions have been developed under this strategy with lower costs and shorter development times, these benefits have, in some cases, been achieved at the expense of increasing performance risk. Addressing the question of when a mission becomes “too fast and too cheap” with the result that it is prone to failure, studies have found that when a mission’s complexity is too great relative to its allocated resources, it fails or becomes impaired. Aerospace’s Small Satellite Cost Model has been highly successful, as evidenced by NASA’s recognition for the model’s application in the Discovery and Small Explorer programs. A critical component of many small-satellite evaluation activities and a major contribution to the space industry as a whole, the model is a stellar example of the value-added role Aerospace can play. There is every reason to expect that more success stories will be forthcoming. Further Reading R. L. Abramson and D. A. Bearden, “Cost Analysis Methodology for High-Performance Small Satellites,” SPIE International Symposium on Aerospace and Remote Sensing (Orlando, FL, April 1993). R. L. Abramson, D. A. Bearden, and D. Glackin, “Small Satellites: Cost Methodologies and Remote Sensing Issues” (Paper 2583-62), European Symposium on Satellite Remote Sensing (Paris, France, September 25–28, 1995). H. Apgar, D. A. Bearden, and R. Wong, “Cost Modeling” chapter in Space Mission Analysis and Design (SMAD), 3rd ed. (Microcosm Press, Torrance, CA, 1999). D. A. Bearden, “A Complexity-Based Risk Assessment of Low-Cost Planetary Missions: When Is a Mission Too Fast and Too Cheap?”
Crosslink Winter 2000/2001 • 43
The History Behind Small Satellites The Soviet Union’s October 1957 launch of Sputnik, the first satellite, stunned the world. It kicked off the “space race” between the Soviet Union and the United States, and in doing so, changed the course of history. Space would become an important setting in which nations could demonstrate political and scientific prowess. The United States responded to Sputnik in January 1958, launching Explorer I, a simple, inexpensive spacecraft built to answer basic questions about Earth and near space. Explorer and its immediate descendants were small satellites, but only because of launch-vehicle limitations. Size and complexity of later spacecraft grew to match launch capability. Not surprisingly, the early years of the space race saw U.S. projects expand on many levels. The Cold War spurred the buildup of a massive space-based defense and communications infrastructure in the United States. The government and its contractors, essentially unchecked by budgetary restrictions, developed large, sophisticated, and expensive platforms to meet increasingly demanding mission requirements. NASA followed the lead of DOD, building complex scientific and interplanetary spacecraft to maximize research capabilities. U.S. expertise in space science was escalating. Launch-vehicle capability continued to grow from the 1960s through the early 1980s, with large satellite platforms carrying more
Fourth IAA International Conference on LowCost Planetary Missions, JHU/APL (Laurel, MD, May 2–5, 2000). D. A. Bearden, R. Boudrough, and J. Wertz, “Cost Modeling” chapter in Reducing the Cost of Space Systems (Microcosm Press, Torrance, CA, 1998).
powerful payloads (and, often, multiple payloads). Engineers and scientists worked to perfect the technologies necessary for mission success and lengthier operations. Major research spacecraft took nearly a decade to develop, and they grew to cost more than $1 billion. As years passed, however, several factors pointed to a need to scale back. With the end of the Cold War, government spending in science and technology received increased public scrutiny. Funding for large, complex flagship missions would no longer be available. Budget constraints forced program managers to look seriously at smaller platforms in an attempt to get payloads onto less-costly launch vehicles. At the same time, the public voiced a growing concern over the potential for reduced research findings in the wake of several failures of large, highprofile, expensive NASA missions; for example, a crippling manufacturing defect was discovered on the Hubble Space Telescope. NASA came under fire for its perceived inability to deliver quality scientific research. The scientific community expressed frustration about the lack of flight opportunities because only a few flagship missions, with decade-long development times, were being undertaken. After the limited-capability launch of Galileo in 1989 and the loss of Mars Observer in 1993, the next planetary mission to be launched was the Cassini
D. A. Bearden, N.Y. Lao, T. B. Coughlin, A. G. Santo, J. T. Hemmings, and W. L. Ebert, “Comparison of NEAR Costs with a Small-Spacecraft Cost Model,” AIAA/Utah State University Conference on Small Satellites (Logan, UT, September, 1996).
D. A. Bearden and R. L. Abramson, “A Small Satellite Cost Study,” 6th Annual AIAA/Utah State University Conference on Small Satellites (Logan, UT, September 21–24, 1992).
E. L. Burgess, N. Y. Lao, and D. A. Bearden, “Small Satellite Cost Estimating Relationships,” AIAA/Utah State University Conference on Small Satellites (Logan, UT, September 18–21, 1995).
D. A. Bearden and R.L. Abramson, “Small Satellite Cost Study—Risk and Quality Assessment,” CNES/ESA 2nd International Symposium on Small Satellites Systems and Services (Biarritz, France, June 29,1994).
M. Kicza and R. Vorder Bruegge, “NASA’s Discovery Program,” Acta Astronautica: Proceedings of the IAA International Conference on Low-Cost Planetary Missions (April 12–15, 1994).
44 • Crosslink Winter 2000/2001
mission to Saturn in 1997, which wouldn’t start transmitting data to Earth until 2003, six years after Galileo stopped sending data from Jupiter. All these issues—budgetary changes brought about by the end of the Cold War, mission failures, predicted gaps in scientific data return— meant that future space-science research and planetary exploration would require a different approach. In the mid-1980s, with new developments in microelectronics and software, engineers could package more capability into smaller satellites. Funding from the DOD Advanced Research Projects Agency, the Air Force Space Test Program, and university laboratories allowed engineers to build lowprofile, low-cost satellites with maximum use of existing components and off-the-shelf technology and minimal nonrecurring developmental effort. Research organizations, private businesses, and academic institutions—all weary of waiting years for their instruments to be piggybacked on large satellites—began to develop small satellites that could be launched as secondary payloads on the shuttle or large expendable boosters. Small space systems were emerging that were affordable and easy to use, and thus attractive to a larger, more diverse customer base. A new trend had taken shape, and a whole new era in the history of space science was beginning.
T. J. Mosher, N. Y. Lao, E. T. Davalos, and D. A. Bearden, “A Comparison of NEAR Actual Spacecraft Costs with Three Parametric Cost Models,” IAA Low Cost Planetary Missions Conference (Pasadena, CA, April 1998). National Research Council report, Technology for Small Spacecraft (1994). L. Sarsfield, “Federal Investments in Small Spacecraft,” Report for Office of Science and Technology Policy, RAND Critical Technologies Institute, DRU-1494-OSTP (September 1996). L. Sarsfield, The Cosmos on a Shoestring: Small Spacecraft for Space and Earth Science (RAND Critical Technologies Institute, ISBN 0-8330-2528-7, MR-864-OSTP, 1998).
arly warning against missile attack is a key mission for military planners dealing with missile defense systems, the subject of considerable international debate. A space-based infrared surveillance system can provide such early warning. Enhancing its timeliness and usefulness to guarantee high system performance depends on accurate appraisal of the characteristics of the background against which the target will be detected and how these characteristics influence sensor design and performance. Optimum sensors must be deployed to ensure high system performance. The cost of optimum system performance, however, is being carefully scrutinized by decision-makers as part of the ongoing process of Department of Defense (DOD) acquisition reform that has distinguished the procurements of the past decade. Indeed, system cost has become a critical factor in decisions regarding DOD acquisitions. The inclusion of system costs in architecture studies represents a significant and important extension of the traditional role of The Aerospace Corporation in support of the Air Force Space and Missile Systems Center. The costs associated with the deployment of a surveillance satellite are closely related to the payload mass, both in regard to the payload itself and that of the launch vehicle required to lift it and its satellite platform into, for example, a geostationary orbit. The best performing sensor has the highest resolution and the best background clutter suppression, but is the heaviest and therefore most expensive. The study presented in this article deals with a constellation of spacebased infrared surveillance sensors as an example to show the link between system performance and system cost. Simulation tools are used to facilitate the trades required for optimizing sensor designs to meet mission requirements and to indicate the best approach for minimizing system costs. Sensor resolution (the ability to see detail) is varied over three fixedparameter designs in order to examine the impact of resolution on background clutter and, hence, system performance. The most valuable aspect of this example is the quantitative link between system-level performance and payload mass over conditions of variable clutter. Aerospace Simulation and Modeling Tools Aerospace regularly provides quick-response assessments of a variety of space-based-sensor concepts. These assessments, underpinned by reasonably detailed sensor design constructs, accurately determine system performance. The constructs are important for estimating sensor (and ultimately space-segment) mass, power, volume, and cost, and for evaluating associated technology risks for sensor subsystems and their components. A variety of analysis and simulation tools are used to assess sensor performance. An analytical approach is usually adequate, unless the background-scene structure interacts with the sensor to
Space-Based Systems for Missile Surveillance D. G. Lawrie and T. S. Lomheim
Aerospace uses sophisticated analysis and simulation tools to design systems, assess their performance, and link design and performance to system cost.
The constellation-level analysis tool TRADIX is used to assess the global missile-warning performance of space-based, infrared sensors. The satellites are propagated over an entire day (or epoch) with, typically, five-minute time steps. At each time step, target missiles are launched in up to 36 different azimuthal directions from almost 5000 launch sites uniformly distributed across Earth’s surface, resulting in more than 50 million booster launches per epoch. All relevant sensor parameters, target signatures and motions, atmospheric effects, and solar and Earth backgrounds are incorporated within the simulation. The effects of Earth background clutter on system performance, and hence sensor design, are a main topic of this article.
Crosslink Winter 2000/2001 • 45
create a significant component of the system noise (i.e., clutter). In these cases, the approximations required for an analytic approach are often violated; for example, cloud edges and land-and-sea interfaces frequently distort the normal background amplitude distribution. When this happens, detailed simulations of the spatial structure of the background scene, pixel by pixel in the focal plane, must be incorporated into the analysis for an accurate assessment of the sensor’s performance. If the emphasis is on the system performance of a constellation of sensors, such a level of detail has generally been viewed as too costly and time-consuming. Measuring the Impact of Background Spatial Structure Aerospace has developed methodologies for quantifying the effect of the background spatial structure on the performance of space-based infrared sensors. The results are coupled with sensor-design constraints and mission performance tools to allow high-level systems-engineering trades that provide insight into relationships between cost and performance. In effect, high-fidelity sensor and phenomenology (target and background) models generate constraints and databases for use within constellation-level simulations, enhancing their accuracy. This integrated simulation capability supports sensor and system trades for a number of space-based, infrared surveillance system studies, including those dealing with theater-missile warning. An early version of this capability was used in 1994 to support the Space-Based Infrared System (SBIRS) Phenomenology Impact Study, conducted by Aerospace in collaboration with the Massachusetts Institute of Technology Lincoln Laboratory. The study recommended collecting background characteristics, a strategy that ultimately involved the Miniaturized Space Technology Initiative 3 and Midcourse Space Experiment satellite experiments, as well as background observations from a high-altitude aircraft. The phenomenology database that the study generated was later made available to the SBIRS High and Low components. SBIRS High refers to a constellation in high orbit for full Earth surveillance; SBIRS Low is a constellation in low orbit for detection and precise tracking of postboost objects, such as reentry vehicles. The simulation capability was also used during the 1994 Surveillance Summer 46 • Crosslink Winter 2000/2001
FOVfocal plane FOVcross-scan Scan pattern of sensor for theater missile surveillance; the scanner field of regard typically covers most of the Eurasian landmass. The instantaneous field of view of the sensor is shown as FOVfocal plane, which depends on both the detector height and the array length perpendicular to the scan direction. The FOVfocal plane is scanned a length FOVin-scan over three successive adjacent “scan patches” to cover a crossscan extent, FOVcross-scan.
Study as a tool for developing the system requirements for the SBIRS program. Infrared Sensor Design An evaluation of a space-based infrared surveillance architecture for early missile warning in two potential theaters of operation will illustrate the models and analysis procedures for assessing sensor design and performance. System performance is derived for three generic infrared scanningsensor designs with varying degrees of spatial resolution of 1.8, 2.6, and 3.6 kilometers, with parameters specified for potential variations and uncertainties in the background structure. The sensor designs were analyzed in parallel to determine sensitivity, subsystem requirements, mass, and power. The results provide insight into the cost of the uncertainties in phenomenology, in terms of sensor mass. In order to relate system performance with system cost, the sensor performance must be linked to specific system architectures with well-defined sensor payloads
and associated spacecraft bus, communication, ground, and launch systems. The focus in this discussion is a key system element: the infrared-sensor payload configured for the detection and tracking of theater missiles. The sensor is assumed to be deployed in a geostationary orbit with a field of regard to cover the anticipated threat areas, for example, the Middle East and Northeast Asia. Typical mission requirements include the minimum detectable target, time of first report, size of the geographical area of interest, accuracy of the reported launch location, heading of the target, and accuracy of the predicted impact point. These are met with a set of properly sized infrared sensors that are configured with an optimized satellite-constellation architecture. The constellation is deployed in a geostationary orbit, which determines the number of satellites, data communication rates and other elements of the infrastructure, and space- and ground-based processing systems. Sensor sizing is driven by the required sensor sensitivity, spatial resolution at the target range, revisit time (time between looks) or target-update rate, and the selection of appropriate spectral bands for discriminating targets from backgrounds. Once the system performance parameters are defined and the constellation architecture selected, an infrared sensor-system design is synthesized and the sensor’s firstorder technical design parameters are obtained: • sensor type (scanner or starer) • system field of regard and scan pattern • telescope field of view, detector pixel instantaneous field of view, aperture and focal length • system scan rate/staring duration • focal-plane definition (single or dual color), sensitivity, topology, and frame rate • signal-processing data rates and functional definition • overall system digital output data rates The next level of synthesis fleshes out the infrared sensor subsystems in enough detail to allow reasonable estimates of the mass, power, and volume of the orbital components of the system. Meaningful technology risk assessments can now be formulated. A variety of linked analysis software tools and databases execute the payload design and sizing process. For example, defining the focal plane includes selecting
the focal-plane detector material and optical cut-off wavelength, spatial layout, detector or pixel dimensions, and readout rate(s). The sensor sensitivity requirement is translated into a focal-plane sensitivity constraint, which allows selection of the focal-plane temperature, using a thermalnoise model specifically tailored for the chosen detector material (e.g., mercury cadmium telluride). The focal-plane topology (total detector count) and maximum readout rate then allow determination of
point the line-of-sight mirror. The subsystem masses, power dissipations, and volumes are “rolled-up” into an overall payload mass, power, and configuration, which is then used to size the spacecraft bus and to select the launch system. The detailed subsystem parameters, along with the subsystem masses, power-consumption levels, and configurations, are then passed to an appropriate cost-estimating tool(s). In order to relate system-level performance with cost, the foregoing process is
Sensor Performance Simulation Tools The remainder of the discussion focuses on the sensor taken to be an infrared scanner; constellation system-level simulation tools are used to calculate end-to-end performance. An example is provided wherein the linear size of the sensor ground sample (spatial resolution or detector “footprint”) is systematically varied up to a factor of two to illustrate the impact of sensor susceptibility to background clutter. The level
Driving mission requirements
Driving sensor requirements
Sensor concept definition
Minimum detectable target Report time Area-of-interest size Tactical-parameter accuracy
Altitude Inclination Number of satellites Communications approach
Noise equivalent target Ground sample distance Coverage area(s) Revisit time Spectral band(s)
Example: Theater infrared scanner Scan pattern Line-of-sight (LOS) agility, stability defined
Derive field of view (FOV), scan rate, instantaneous FOV Determine aperture and focal-plane array (FPA) temperature
Spacecraft/launch vehicle synthesis
Spacecraft design/sizing Launch vehicle selection
Mass Power Size Performance
Size thermal radiator Select thermal heat transport hardware
Focal-plane size, topology Pixel size Pixel count Sensitivity Line rate Analog rates Digital rates Power dissipation Processing functions
Optical design LOS scan mechanism
Acceptable system and system performance ?
Mission performance analysis
Proceed with detailed system design
Dashed paths indicate potential sensor/payload/concept iterations Steps, decisions, and trade-offs in the process for deriving the specifications of an infrared space-based sensor system for detecting missile launches against highly structured backgrounds. In most cases, intermediate and overall iterations are required.
the electrical power, which, with the focalplane temperature, serve as inputs to a model for determining the technology, size, power, and volume of the cryogenic cooling system. A state-of-the-art optical design and tolerancing program uses the design parameters to select and refine a specific telescope optical design for the optical system. The optical program data are used to estimate the mass and volume of the optical subsystem and the power required to scan and
carried out as a function of sensor performance parameters by varying, for example, the sensor noise-equivalent target, revisit time, and ground-sample distance (resolution). For such parametric analyses, scaling relationships are often used to interpolate subsystem mass, power, and volume estimates between design points. This is appropriate once a detailed design is developed for a basis; excursions from this fiducial design then use the appropriate scaling relationships.
of this clutter is also varied over a wide range. Variable ground-sample size is used to derive corresponding sensor system designs from which payload masses are determined. To more clearly illustrate this sensitivity trade, cost is assumed to be related only to the infrared sensor mass, as a rough approximation. The specific costbenefit relationships developed by this example are illustrated in the next section. Simulating the performance of a spacebased surveillance system involves the use Crosslink Winter 2000/2001 • 47
Sensor Requirements • Noise-equivalent target • Ground sample distance • Coverage area • Revisit time • Spectral-band choice • Range to target
• Sensor type (scanner/starer) • Scan or step/stare pattern
• Select sensitivity model • Select FOV • Select detector size/focal length based on derived IFOV • Select line rate from scan geometry – Obtain integration time • Solve for optical aperture diameter/focal plane temperature • FPA temperature
Determine thermal radiator weight/power
• Number detectors • FPA temperature • Electrical dissipation
• FOV • Integration time
Determine FPA size, properties, data rate
FPA data rate
• Aperture • FOV • Focal length Determine signal processor mass/power
Sensor requirements for a hypothetical theater-missile surveillance system. The flowchart illustrates the process that links the sensor requirements to steps that determine the infrared sensor design and finally the sizes of the individual infrared sensor subsystems. Designing a sensor for a single mission
of an ensemble of software models and databases: • SSGM—Synthetic Scene Generation Model: encapsulates many phenomenology codes under one architecture; developed by Photon Research Associates Inc. for the DOD Ballistic Missile Defense Organization • CLDSIM—cloud-scene simulation model incorporated in SSGM • VISTAS—Visible and Infrared Sensor Trades, Analyses, and Simulations model: combines classical imageprocessing techniques with detailed sensor models to produce static and timedependent simulations of a variety of sensor systems, including imaging, tracking, and point-target-detection scanners and starers; designed and coded by Aerospace. • TRADIX—constellation-level analysis tool that combines electro-optical sensor models with target, background, and atmospheric models to evaluate system performance; developed by Aerospace. The simulation begins with the generation by SSGM of a set of shortwave infrared Earth-cloud background scenes. Scenes are selected from the database of weather satellite imagery. The images are pixelized, and the infrared properties of the scene elements are inserted into the image database. 48 • Crosslink Winter 2000/2001
• FOV • Integration time • IFOV • Spectral band
Determine optical subsystem mass/volume
Determine pointing mirror mechanism mass and power
and a limited set of engagement geometries is straightforward. However, surveillance systems must operate against a wide variety of viewing geometries, target signatures, and backgrounds. They must also be able to perform many types of missions, often simultaneously.
A key step in this process is the use of CLDSIM to simulate the solar scatter from cloud tops at various altitudes. Scenes with only terrain, sea surfaces, and low-altitude clouds usually do not generate much clutter in the chosen spectral band. Solar reflections from high-altitude clouds, on the other
hand, can cause a high degree of clutter, which can, in turn, stress the sensor’s ability to detect targets of interest. SSGM can generate selected atmospheric properties, a specified spatial resolution, and a matrix of scenes with a variety of viewing geometries in the desired spectral band.
Mean frequency of occurrence of clouds above 10 kilometers (University of Wisconsin HIRS-2 data, August 1989–1994). Earth’s surface is never completely covered with clouds on any given day. Highaltitude clouds are more likely to occur at lower latitudes. The analysis in this study generates the probability of missile warning when clouds of a given type and altitude are present at the locations of interest. A global cloud statistical model developed at Aerospace indicates that, based on six years of data, clouds at 10 kilometers or above occur over Northeast Asia about 20 to 30 percent of the time during the summer, with a maximum cloud coverage of 40 percent. In light of these statistics, sensor performance must be evaluated against clouds ranging in altitude up to 10 kilometers or possibly higher.
Global Cloud Database
Synthetic Scene Generation Model
Cloud cover, types, and altitudes Sensor Design and Optimization Optical design and straylight analysis Focal-plane layout and noise analysis Signal processing and communications Line of sight and thermal control
High-fidelity Constellation-level Clutter sensor model statistics performance tool Clutter statistics
Coverage, report times, and target-detection statistics
Other system-level simulation tools (e.g., tracking, resource scheduling) The various simulation tools developed and used by Aerospace and their interconnectivity for the evaluation of system performance. Such simulation tools must be built to accurately model the interaction of the sensors with relevant target and background characteristics over all possible viewing areas. One of the main goals of the tool development is to incorporate the effects of “real” clutter phenomena, such as cloud edges and sun glints, within system-level analyses. A scene-based methodology generates appropriate clutter statistics, which are included within the constellation-level simulations.
For the evaluation of sensor performance, VISTAS models the imaging chain of the electro-optical sensor, from the background scene input to the signalprocessor output. The sensor-system transfer function is applied to a high-resolution input scene, such as those produced by SSGM. The transfer function (output vs. input) includes the effects of the optical point spread function, that is, the blur, and for a scanner, the temporal aperture caused by the scan motion during the integration time. The blurred scenes are resampled at the system resolution, and clutter-rejection filters are applied. The output is calibrated to account for the sensor system’s response to the target intensity. The output scenes are then analyzed for figures of merit, such as the standard deviation of the sensor response, which are statistical representations of the clutter of the background scene and sensor noise for the sensor design under consideration. TRADIX models space sensors operating in both above- and below-the-horizon modes, from the visible to the long-wave infrared. It contains a model for the infrared signature of the missile body, plume intensity and trajectory profiles, clutter statistics generated by SSGM/VISTAS, background models including straylight from
nonrejected earthshine and sunshine, and the atmospheric path radiance and transmission. These models are integrated with the Aerospace orbit propagation library, ASTROLIB, to provide a dynamic simulation tool for studying the constellation-wide performance of electro-optical sensors. Critical inputs to the above tools include focal-plane pixel topology, sensor noise characteristics, details of the filters and signal processing, optical design and straylight rejection capability, platform drift and jitter, constellation orbits and phasing, and concepts of operation, that is, scan modes, revisit times, and others. Information on the frequency of occurrence of meteorological conditions is critical to the performance of a space-based infrared surveillance system against Earth backgrounds. Aerospace uses a global cloud statistical model that is based on the University of Wisconsin HIRS-2 (HighResolution Infrared Sounder) data from the National Oceanographic and Atmospheric Administration polar orbiters. The database incorporated in this model provides a basis for assessing system performance against clouds of various altitudes for a given time of year at a specified geographical location. Typical displays of such data would be shown on a world projection in
terms of the maximum, minimum, and mean distributions of the probability of clouds above a given altitude during a specified month. Scene-Based Clutter Analysis The performance of an Earth-viewing infrared sensor, designed to operate in the shortwave infrared atmospheric-absorption band to block signals from terrestrial sources, is dominated by the structure in the background scene. This structure is caused predominantly by sunlight reflected from clouds in the scene and depends on the sensor-cloud sun-viewing geometry, represented by the cloud “look-zenith angle” and the “solar-scatter angle.” The solar-scatter angle, which determines the intensity of solar scatter off the cloud tops, is dominant; very small values of the solarscatter angle result in very intense scattering. The other angle, the look-zenith angle, defines the projection of the SSGM cloud scene in the sensor line of sight, in reference to the range from the sensor to the cloud tops. The look-zenith angle is also directly related to the path length through the atmosphere to a target at a given altitude, and hence to the target’s apparent irradiance for a given time after launch. At high-zenith angles, low-altitude targets viewed within the Earth limb suffer from
Nadir angle LZA LZA = 0 deg To sun SCA
Solar-scatter geometry: As the sensor line of sight shifts from nadir to the limb, both the range to target and the path length through the atmosphere increase, while the minimum possible solar-scatter angle (SCA) decreases. When the sensor is viewing targets at the nadir (lookzenith angle [LZA] equals zero degrees), the minimum possible SCA is 90 degrees, whereas when viewing low-altitude targets at the limb (LZA equals 90 degrees), the sun can be directly in the sensor line of sight (i.e., SCA can be zero degrees). Unfortunately, most of the surface area covered by a given space sensor viewing Earth lies at the larger LZAs. Overlapping coverage provided by a constellation of many sensors mitigates this problem.
Crosslink Winter 2000/2001 • 49
the worst combination of range, atmospheric transmission loss, and solarinduced background clutter. Unfortunately, most of the Earth’s surface viewed by a sensor in space lies at the larger lookzenith angles. However, the overlapping coverage of multiple sensors can be used to mitigate this problem. The two SSGM scenes selected in this study represent a “nominal” case containing low- to mid-altitude clouds, and a “stressing” case with mid- to high-altitude clouds. The terms nominal and stressing refer to the level of the clutter generated when the scene is passed through a typical sensor simulation. Each scene covers an area of 512 × 1170 kilometers at nadir with a 200-meter spatial resolution, as seen through a midlatitude summer atmosphere at a look-zenith angle of 60 degrees and solar-scatter angle of 90 degrees; geometric projection effects at a look-zenith angle of 60 degrees shorten the apparent size of these scenes to 510 × 570 kilometers. The brighter clouds in these images are at higher altitudes where there is less attenuation of sunlight both before and after it scatters from the cloud tops. The most significant caveat concerning this analysis is the use of the CLDSIM model within SSGM. A number of uncertainties are inherent in this model, one being the model for the solar scattering from the clouds. Using detailed comparisons with actual space sensor data from the MSTI-3 (Miniature Sensor Technology Integration) sensor that collected information on Earth and Earth-limb clutter, the resultant uncertainty in the apparent cloud brightness has been estimated to be certainly less than a factor of three. The impact of this uncertainty was addressed by scaling the intensity of each scene, in the nominal case by one-third and one-half and in the stressing case by two and three, thus providing a total of six cases. In this way the effect on system performance of varying the cloud types and altitudes and the additional impact of the SSGM modeling uncertainties were quantified. In viewing the Earth-cloud background scenes, a scanning sensor typically responds only to changes in the signal, in effect performing a subtraction of the mean background radiance, thus providing an indication of the target intensity above the mean. The effectiveness of this suppression of clutter caused by radiance differences in the scene depends on the instantaneous 50 • Crosslink Winter 2000/2001
Nominal scene, left, used for clutter performance analysis: clouds at 2–8-kilometer altitude, 200-meter resolution, and midlatitude summer atmosphere. The database is referred to as nominal because it contains mostly low- to mid-altitude clouds and results in moderate clutter levels, depending on the sensor design and solar angle. Also, these clouds occur quite frequently over a wide range of latitudes. The brighter clouds are at higher altitudes where there is less attenuation of the sunlight. Stressing scene, right, used for clutter performance analysis: clouds at 4–10-kilometer altitude, 200-meter resolution, and midlatitude summer atmosphere. The database is referred to as stressing because it contains cirrus ice clouds at 10-kilometer altitude and can generate quite severe clutter levels. These clouds do occur less frequently than those contained in the nominal scene.
field of view of the detectors, the pixel footprint. In this study, scanning-sensor designs with footprints of 1.8, 2.6, and 3.6 kilometers were used at a nominal range of 40,000 kilometers (corresponding to a look-zenith angle of approximately 75 degrees for a geostationary satellite). The output scenes then exhibit a mean background level of zero, with positive and negative values apparent at the cloud edges. These simulation outputs were used to determine the probabilities of false exceedance (false indication of a target in the
field of view) versus intensity threshold for the background clutter. In order to set a threshold, an acceptable false exceedance rate for each sensor design must be determined. For architectures where the mission data are processed on the ground, this involves several factors: the number of detectors in the focal plane, the sampling rate, the number of bytes per sample, and the capacity of the downlink, as well as some knowledge of the target detection algorithm. For example, a scanner design with a 1.8-kilometer footprint
Infrared-scanner outputs in the simulation for the stressing background scene. For a scanner, each background input scene results in a single simulated static output scene. The intensity of a target relative to the surrounding background is the principal method for detection. Grey represents the mean level of zero, while white and black represent positive and negative exceedances (false indications of a target in the field of view). In order to provide a visual guide to the clutter levels in these two output scenes, four “target” markers with an intensity of 5 kilowatts per steradian are imbedded in each scene. All four are clearly visible in the 1.8-kilometer case, whereas only the central, isolated target is visible in the 3.6kilometer scene. These output scenes are used to provide a statistical representation of the clutter in terms of the number of exceedances vs. threshold level for the background scene and sensor design.
100 Stressing clutter: LZA = 60 deg SCA = 90 deg
10–1 Exceedance probability
could indicate a threshold of 6 kilowatts per steradian for a false exceedance rate per pixel of 10-4, whereas a design for a 3.6-kilometer footprint viewing the same scene would indicate a false exceedance rate per pixel of 10-1 for the same threshold intensity. Of course the smaller footprint design would require many more detectors and a larger telescope, hence a heavier payload and greater cost. The ground footprint of an infrared scanner is a key parameter in determining sensor performance against structured backgrounds. In a more complete analysis, one must account for the sensor noise arising from the natural fluctuations in the scene and from the electronics, usually expressed as a noise-equivalent target intensity. This accounting is expressed in a thresholdversus-exceedance distribution. If the clutter noise from the background is random in character, it can be expressed as a clutterequivalent target intensity; then a systemequivalent target intensity can be determined simply as the square root of the sum of the squares. However, clutter from clouds in natural background scenes is usually far from random, so that more complex methods are required for combining the sensor noise with the background clutter distribution. Target Response Target response is a scale factor that describes the attenuation of a target through the sensor system. The response of a scanning sensor to an unresolved target, that is,
10–2 3.6 kilometers
10–3 10–4 10–5
10 15 20 Threshold (kilowatts/steradian)
Probability of false exceedance plotted as a function of threshold intensity for infrared scanners viewing the stressing background scene. A threshold of approximately 6 kilowatts per steradian for the 1.8kilometer-footprint scanner design corresponds to a false exceedance of 10-4. When that same threshold is applied to the 3.6-kilometer-footprint scanner design, the false exceedance is approximately three orders of magnitude higher (10-1), whereas the reduction in the number of detector channels is only a factor of four.
a “point source,” depends on several factors. These are the blurring caused by the optics, the temporal aperture caused by the scan motion during the integration time, the sampling of the blurred target by the focal plane, the target phasing (i.e., the location of the target relative to the center of a pixel), and the electronic filtering. For a fast-scanning sensor system, the response does not depend on the temporal characteristics of the target so the calculation is fairly straightforward. The target response can be evaluated by constructing a
Infrared-scanner simulation: outputs thresholded at 6 kilowatts per steradian for the stressing background scene. If the 6-kilowatts-per-steradian threshold is applied to the output scenes, the impact of sensor footprint on clutter response is immediately apparent: clutter is reduced in both cases, but there is a more pronounced effect for the 1.8-kilometer footprint. Threshold levels are usually set by the limitations of the onboard processor or the ground communications link. The need to limit the number of false alarms reported by the system can also be a significant constraint. False-exceedance levels of approximately 10-4 to 10-3 are typical.
scene in which a grid of many point sources (usually 1 kilowatt per steradian) are spaced far enough apart to avoid interference, each offset randomly by a small amount to make the grid nonuniform, thus to ensure many different target phasings. This target grid is then passed through the same simulation process as the background scenes, namely blurring, downsampling, and filtering. The peak response from each target is determined, and the average is taken as the mean target response. All simulated clutter scenes are divided by this target response so they can be referenced to apparent intensities at the sensor aperture. Constellation Performance To evaluate the performance of a spacebased infrared surveillance architecture against a variety of target and background conditions, the sensor response to both must be combined in a constellation-level simulation. This is done by first choosing one of the cloud databases contained within the SSGM, then generating scenes spanning the entire range of viewing geometries and sun angles for the selected sensor constellation. The scenes are then processed through the detailed sensor simulation tool, VISTAS, to produce a set of false-exceedance-threshold clutter distributions. These are combined with the sensor noise-equivalent target intensity in TRADIX, and the threshold and minimum detectable target are calculated as functions of the look-zenith and solar-scatter angles,
Crosslink Winter 2000/2001 • 51
52 • Crosslink Winter 2000/2001
Shortwave infrared intensity (kilowatts/steradian)
for the required probability of detection and false exceedance. Within TRADIX, the targets and satellites are propagated in Earth-centralinertial coordinates with an appropriate sampling interval, typically 5 to 15 minutes. This is carried out over a period of a day at various times during the year to explore the effects of seasonal variations in the sun’s latitude. For a constellation in geostationary orbits and target launch sites in the northern hemisphere, the most stressing solar-scattering angles are close to the summer solstice, and the resultant background clutter is the dominant effect in the system performance. On the other hand, the effect of solar straylight can be more stressing close to the solar equinoxes. Each sensor-target line of sight at each time results in a look-zenith-angle/solarscatter-angle pair, which, with the clutter data, sets a minimum-detectable-target threshold for that sensor, and a time of first detection. For example, three out of four detections produces a “3of4” report from that sensor, and two such reports by separate sensors result in a stereo report for the constellation. Target-detection and reporttime statistics are thus generated for each sensor design against each structured background for the mission of interest. This procedure has been applied to missions of global- and theater-missile warning. For a typical global analysis, target launch sites are uniformly spread over the surface of Earth from −90- to +90-degree latitude. A target spatial pattern with a resolution of 3 × 3 degrees (Earth central angle) generates 4586 distinct target launch locations that represent equal-sized areas on the surface of Earth. Missiles are usually launched in 12- to 36-azimuthal directions in order to allow for aspect angle effects on the apparent booster signature. The point of such an analysis is to obtain a measure of system performance that is not scenario driven. As many as 50 million target launches may be run in order to determine global performance for a particular spacesurveillance constellation. For the theater-missile-warning mission, a representative short-range missile was assumed with an infrared intensity profile increasing as the missile rose through the atmosphere to a maximum, then decreasing as the afterburning of the exhaust diminished (as viewed from a broadside aspect at the two extreme look-zenith angles of 0 and 90 degrees, corresponding to the nadir
100 80 LZA = 0 deg 60 40 LZA = 90 deg 20 0 0
20 30 40 Time after launch (seconds)
The apparent shortwave-infrared intensity of a hypothetical theater missile vs. time after launch when viewed at nadir and the limb; includes attenuation by the atmosphere.
and the limb). The analysis was limited to the worst-case epoch (date or day of year) for clutter-limited detection and was focused on two theaters of operation, one in the Middle East, the other in Northeast Asia. A constellation of five satellites, with four of the satellites “pinched” to cover the Eurasian landmass, was selected to provide excellent overlapping coverage of both areas simultaneously.
A short-wave-infrared line scanner was selected as the sensor-design option. Although not the most effective choice for clutter suppression, a scanner reduces the program risks associated with extreme line-of-sight stability and focal-plane producibility associated with infrared-starer designs. On the other hand, for acceptable performance against highly structured backgrounds, it is necessary that a scanner
Instantaneous line-of-sight coverage of two theaters by five (four pinched) geostationary satellites. The 3- × 3-degree target grid results in approximately 70 launch locations per theater. A representative type of theater missile has a burnout time of approximately 60 seconds after launch. The ability to optimize satellite locations and thus trade between theater performance and global coverage is most easily achieved with geostationary constellations. The longitudes of the five geostationary satellites are shown here on a cylindrical projection of Earth, along with the boundaries of the two theaters and the constellation’s line-of-sight coverage. Essentially 100-percent triple coverage is provided over both theaters of operation.
Theater-missile-warning stereo performance of five (four pinched) satellites against highly stressing clutter. The stereo performance results across the entire Eurasian landmass are shown for all three sensor designs against the highly stressing clutter level. The performance for the Middle East is indeed very close to that for Northeast Asia. These designs would be even further stressed if the system were being asked to perform additional missions simultaneously.
have a relatively small footprint. In this study, three designs with footprints of 1.8, 2.6, and 3.6 kilometers were considered. The infrared-scanner optics consisted of a triplet refractor and a two-axis entry flat for scanning. For a low noise-equivalenttarget intensity, sensor aperture was traded against time delay and integration capability on the focal plane. The result was a 27centimeter aperture with 12 stages of time delay and integration. A 2-second revisit time taken as a mission requirement necessitated a scan rate of 4 degrees per second and resulted in a noise-equivalent-target intensity of about 1 kilowatt per steradian at 40,000 kilometers. A single-hit probability of detection of 95 percent was chosen; this led to a cumulative probability of detection of 99 percent for the 3of4-hits algorithm. The resultant probability of detection and false exceedance combination led to a minimum-detectable target as a function of viewing geometry for each sensor design, background scene, and clutter scale factor. The TRADIX constellation analysis tool was then used to evaluate the performance of the three sensor designs against the nominal and stressing backgrounds. Conclusion The costs associated with the deployment of a surveillance satellite are closely related to the payload mass, both in regard to the payload itself and that of the launch vehicle required to lift it and its satellite platform into a geostationary orbit. Accordingly, rather than attempting to carry out a detailed cost analysis, this study was focused on estimations of payload mass, which included the telescope, sensor housing, and scan mirror; the focal-plane assemblies and signal processors; power supplies; and other subsystem masses. The highest-performing sensor with the smallest footprint and best clutter suppression ends up also being the heaviest and most expensive. The overall costs of this highestperforming system would have to be weighed against the value of the mission to the national interest; such considerations are beyond the scope of this investigation. The study described here illustrates the application of analytical tools and databases designed and assembled at Aerospace to support the development of advanced space-based surveillance systems. The example shown represents a hypothetical system for the specific purpose of missile warning that covers two widely separated geographical areas of concern. Aerospace Crosslink Winter 2000/2001 • 53
100 Detection probability (percent)
Aerospace Systems Architecting and Engineering Certificate Program
Infrared Systems and Technology
Five geostationary satellites (four pinched) Northeast Asia theater
Stressing Stressing × 2 Clutter level
Stressing × 3
Theater-missile-warning stereo performance sensitivity to sensor footprint and clutter. Stereo tracking is necessary if stressing requirements associated with launch point determination and impact point prediction are to be met. The impact of the larger footprints on the stereo performance of the three sensor designs is severe. A 95-percent probability of stereo detection before burnout is only achievable for the 2.6-kilometer or better design against the stressing clutter level. The 3.6-kilometer design can only meet a stereo probability of detection of 95 percent at the nominal clutter level or below. Tenkilometer clouds can be expected over Northeast Asia approximately 20 to 30 percent of the time during the summer months, and six-kilometer clouds can be expected 40 to 50 percent of that time.
Footprint (kilometers) 2.6
80 70 60
Increasing clutter levels
Detection probability (percent)
Stressing ×2 Five geostationary satellites (four pinched) Northeast Asia theater
50 Stressing ×3 40 100
125 150 Payload mass (kilograms)
Performance sensitivity to sensor footprint and payload mass at the various shortwave-infrared clutter levels with line-of-sight stereo coverage. Essentially, the graph shows the cost of designing an infrared sensor with guaranteed performance against an uncertain level of background clutter. The larger footprint designs are less robust against increasing clutter levels. The data show that while there are significant cost savings to be made in lower payload mass (and power), the performance penalty in stereo track capability associated with a larger footprint design may be severe.
is currently applying comparable analyses in its support to the Air Force Space and Missile Systems Center and the Ballistic Missile Defense Organization for the development of both SBIR High and Low systems. Further Reading A. F. Pensa and J. R. Parsons, SBIR System Phenomenology Impact Study (SSPIS), (Sponsored
54 • Crosslink Winter 2000/2001
Infrared sensors detect reflected and emitted radiation from sources in a wavelength regime well beyond what can be seen by the human eye. Such sensors afford detection of hot missile plumes as well as cold space targets. The sensors also allow detection through haze, at night, and during conditions of atmospheric radiance and reflected solar radiation, among other applications. When placed on space-based platforms, these infrared sensor systems can detect, identify, and track missiles along most of their trajectory, from launch to impact. Infrared space sensor systems, such as the High- and Low-altitude segments of the Space Based Infrared System, provide global and theater-missile warning and have applications in missile defense. The Aerospace Institute offers a 30hour course, Infrared Systems and Technology, on infrared systems design, including the relationship between resolution, search rate, and elevation angle. The course also provides instruction on focal-plane arrays, space radiation effects, clutter, background phenomenology, passive cooling, infrared system simulation, system performance analysis, and alternative technologies and trades for single and multisensor systems. Infrared Systems and Technology is linked to the Space Systems Design course in the Space Systems Engineering series of the Aerospace Systems Architecting and Engineering Certificate Program. The Space Systems Design course describes the space systems design process, how design fits into the space-mission timeline, and how requirements flow between the vehicle payload, spacecraft, and ground system.
by the U.S. Air Force Space and Missile Systems Center, El Segundo, CA, December 1994). D. G. Lawrie et al., “Electro-Optical Sensor Simulation for Theater Missile Warning,” paper presented at the Fifth Symposium on Space Systems (Cannes, France, June 1996). T. S. Lomheim et al., “Performance/Sizing Relationships for a Short-Wave/Mid-Wave Infrared Scanning Point-Source Detection Space Sensor,” Proceedings of the 1999 IEEE Aerospace Conference (March 1999, Aspen, CO).
Conferences, Workshops, and Symposia Sponsored or Hosted by The Aerospace Corporation
Space Power Workshop Hosted by The Aerospace Corporation, the Air Force Research Laboratory, and the Air Force Space and Missile Systems Center. The workshop provides an informal, unclassified, international forum to exchange ideas and information on space power. The theme of this year’s workshop is “Commonality in Space and Terrestrial Power.” The sessions encompass the following areas of interest. • Power systems architecture • Power management and distribution (PMAD) • Energy generation • Energy storage • Program experience
Space Parts Working Group (SPWG) Sponsored by The Aerospace Corporation and the Air Force Space and Missile Systems Center. This joint government-industry working group provides an unclassified and international forum to disseminate information to the aerospace industry and resolve common problems with highreliability electronic piece parts needed for space applications. The meeting will include presentations from piece-part leaders in the commercial and military space industry, parts suppliers, and government agencies, including Air Force Space and Missile Systems Center, National Aeronautics and Space Administration, and Defense Supply Center Columbus.
11th Annual International Symposium of INCOSE The Aerospace Corporation is a patron of the symposium, which is hosted this year by the Systems Engineering Society of Australia. The theme of INCOSE (International Council on Systems Engineering) 2000, “Innovate, Integrate and Invigorate,” calls on delegates to discuss and debate the challenges presented by continued use of existing products and systems as well as the way forward in an era of rapid change and technological progress. The symposium will provide a forum for addressing the demands arising from the innovation of novel systems and systems-ofsystems, the integration of tomorrow's functionality with today’s systems, and the invigoration of systems thinking and practice as the globalization of technology and business accelerates in the coming decade. The program will be published after papers, panels, and tutorial submissions have been evaluated and final content has been selected.
April 2–5, 2001 In addition to the sessions, workshop groups will discuss timely space-power issues. An announcement of the preliminary schedule of presentations, registration form, and final hotel information will be available in February 2001. The workshop will be held at the Crowne Plaza, Redondo Beach, CA 90277. For hotel information call 800.368.9760 or 310.318.8888 or visit the hotel Web site at http://www.basshotels.com/crowneplaza?_franchisee=REDCP. For the latest updates on the workshop, please visit the Web site at http://www.aero.org/conferences/power. For general information, contact Jackie Amazaki by phone at 310.336.4073 or by email at [email protected]
May 1–2, 2001 The sessions will discuss the following subjects. • Piece parts and related materials and processes issues • Parts standardization • Enhancement of the part procurement process • Piece-part trends • Radiation hardness The meeting will be held at the Hilton Hotel, 21333 Hawthorne Blvd., Torrance, CA 90503. For hotel information call 310.540.0500 or visit the hotel Web site at http://www.hiltontorrance.com. For more information, contact Mel Cohen at 310.336.0470 or Larry Harzstark at 310.336.5883. Fax 310.336.6914.
July 1–5, 2001 The Academic Forum Program will address the following topics. • Involvement of academics in the INCOSE Plan • Academic contributions needed to support systems engineering • How academia can support stakeholders • Existing academic activity in support of systems engineering • What next? The symposium will be held at the Carlton Crest Hotel, 65 Queens Road, Melbourne, Australia. For hotel information visit the hotel’s Web site at http://asiatravel.com/australia/prepaidhotels/carton_ crest/melbourne/. For more information, visit the symposium’s Web site at http:// www.incose.org/symp2001/.
Crosslink Winter 2000/2001 • 55
Bookmarks Recent Publications and Patents by the Technical Staff Publications S. T. Amimoto, D. J. Chang, and A. Birkitt, “Stress Measurements in Silicon Microstructures,” Proceedings SPIE, Vol. 3933, 113–121 (2000). P. C. Anderson, D. L. McKenzie, et al., “Global Storm Time Auroral X-ray Morphology and Timing and Comparison With UV Measurements,” Journal of Geophysical Research, Vol. 105, No. A7, 15,757–15,777 (July 1, 2000). E. J. Beiting, “Measurements of Stratospheric Plume Dispersion by Imagery of Solid Rocket Motor Exhaust,” Journal of Geophysical Research, Vol. 105, No. D5, 6891–6901 (Mar. 16, 2000). K. Bell et al., “A Joint NASA and DOD Deployable Optics Space Experiment,” UV, Optical, and IR Space Telescopes and Instruments, Proceedings of the Conference (Munich, Germany, Mar. 29–31, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4013, 568–579. K. Bell et al., “Ultra-Lightweight Optics for Space Applications,” UV, Optical, and IR Space Telescopes and Instruments, Proceedings of the Conference (Munich, Germany, Mar. 29–31, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4013, 687–698. K. Bell, R. Moser, et al., “A Deployable Optical Telescope Ground Demonstration,” UV, Optical, and IR Space Telescopes and Instruments, Proceedings of the Conference (Munich, Germany, Mar. 29–31, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4013, 559–567. J. F. Binkley, J. B. Clark, and C. E. Spiekermann, “Improved Procedure for Combining Day-of-Launch Atmospheric Flight Loads,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 459–462 (Aug. 2000). J. B. Blake, J. Fennell, R. Selesnick, et al., “A Multi-Spacecraft Synthesis of Relativistic Electrons in the Inner Magnetosphere Using LANL, GOES, GPS, SAMPEX, HEO, and POLAR,” Advances in Space Research, Vol. 26, No. 1, 93–98 (July 2000). J. B. Blake et al., “Magnetospheric Relativistic Electron Response to Magnetic Cloud Events of 1997,” Advances in Space Research, Vol. 25, Nos. 7–8, 1387–1392 (Apr. 2000). J. B. Blake, M. Looper, et al., “SAMPEX Observations of Precipitation Bursts in the Outer Radiation Belt,” Journal of Geophysical Research, Vol. 105, No. A7, 15,875–15,885 (July 1, 2000).
56 • Crosslink Winter 2000/2001
W. F. Buell, R. W. Farley, et al., “Bayesian Spectrum Analysis for Laser Vibrometry Processing,” Laser Radar Technology and Applications V, Proceedings of the Conference (Orlando, FL, Apr. 26–28, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, No. 190, 444–455. D. J. Chang, P. R. Valenzuela, and T. V. Albright, “Failure Simulation of Composite Rocket Motor Cases,” Proceedings, 2000 National Space and Missile Materials Symposium (San Diego, CA, March 1, 2000), pp. 1–14. I-S. Chang, “Chinese Space Launch Failures,” Proceedings 22nd International Symposium on Space Technology and Science (Morioka, Japan, May 28–June 4, 2000), pp. 1–10. I-S. Chang, “Overview of World Space Launches,” Journal of Propulsion and Power, Vol. 16, No. 5, 853–866 (Oct. 2000). J. B. Clark, M. C. Kim, and A. M. Kabe, “Statistical Analysis of Atmospheric Flight Gust Loads Analysis Data,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 443–445 (Aug. 2000). J. E. Clark et al., “Overview of the GPS M Code Signal,” Navigating into the New Millennium; Proceedings of the Institute of Navigation National Technical Meeting (Anaheim, CA, Jan. 26–28, 2000), pp. 542–549. J. H. Clemmons et al., “Observations of Traveling Pc5 Waves and Their Relation to the Magnetic Cloud Event of January 1997,” Journal of Geophysical Research, Vol. 105, No. A3, 5441–5452 (Mar. 1, 2000). S. V. Didziulis, P. P. Frantz, and G. Radhakrishnan, “The Frictional Properties of Titanium Carbide, Titanium Nitride and Vanadium Carbide: Measurement of a Compositional Dependence with Atomic Force Microscopy,” Journal of Vacuum Science and Technology B, Vol. 18, No. 1, 69–75 (Jan/Feb 2000). J. F. Fennell et al., “Comprehensive Particle and Field Observations of Magnetic Storms at Different Local Times from the CRRES Spacecraft,” Journal of Geophysical Research, Vol. 105, No. A8, 18,729–18,740 (Aug. 1, 2000). J. F. Fennell et al., “Energetic Magnetosheath Ions Connected to the Earth’s Bow Shock—Possible Source of Cusp Energetic Ions,” Journal of Geophysical Research, Vol. 105, No. A3, 5471–5488 (Mar. 1, 2000). J. F. Fennell and J. L. Roeder, “Entry of Plasma Sheet Particles into the Inner Magnetosphere as Observed by Polar/ CAMMICE,” Proceedings of the 1998 Cam-
bridge Symposium/Workshop in Geoplasma Physics on “Multiscalle Phenomena in Space Plasmas II,” No. 15, Edited by T. Chang (1998), pp. 388–393. J. F. Fennell, J. L. Roeder, J. B. Blake, et al., “POLAR CEPPAD/IPS Energetic Neutral Atom (ENA) Images of a Substorm Injection,” Advances in Space Research, Vol. 25, No. 12, 2407–2416 (June 2000). A. Gillam and S. P. Presley, “A Paradigm Shift in Conceptual Design,” Proceedings, International CIRP Design Seminar (Haifa, Israel, May 16–18, 2000), pp. 41–46. E. Harvie, M. Phenneger, et al., “GOES OnOrbit Storage Mode Attitude Dynamics and Control,” Advances in the Astronautical Sciences, Vol. 103, pt. 2, No. 190, 1095– 1114 (2000). D. F. Hall et al., “Measurement of Long-Term Outgassing from the Materials Used on the MSX Spacecraft,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Bellingham, WA, Society of PhotoOptical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, 28–40. D. F. Hall et al., “Observations of the Particle Environment Surrounding the MSX Spacecraft,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, 21–27. D. F. Hall et al., “Outgassing of Optical Baffles and Primary Mirror During Cryogen Depletion of a Space-Based Infrared Instrument,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, 11–20. D. F. Hall et al., “Update of the Midcourse Space Experiment (MSX) Satellite Measurements of Contaminant Films Using QCMs,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Bellingham, WA, Society of PhotoOptical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, No. 190, 1–10. D. F. Hall, G. S. Arnold, T. R. Simpson, D. R. Suess, and P. A. Nystrom, “Progress on Spacecraft Contamination Model Development,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Belling-
ham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, No. 190, 138–156. C. Hay and J. Wang, “Enhancing GPS—Tropospheric Delay Prediction at the Master Control Station,” GPS World, Vol. 11, No. 1, 56–62 (Jan. 2000). J. H. Hecht, A. B. Christensen, et al., “Thermospheric Disturbance Recorded by Photometers Onboard the ARIA II Rocket,” Journal of Geophysical Research, Vol. 105, No. A2, 2461–2475 (Feb. 2000). K. C. Herr et al., “Spectral Anomalies in the 11 and 12 Micron Region from the Mariner Mars 7 Infrared Spectrometer,” Journal of Geophysical Research, Vol. 105, No. E9, 22507–22515 (Sept. 25, 2000). A. A. Jhemi et al., “Optimization of Rotorcraft Flight in Engine Failure,” Proceedings AHS International, Annual Forum, 56th (Virginia Beach, VA, May 2–4, 2000), Vol. 1, 523–536 (2000). A. M. Kabe, C. E. Spiekermann, M. C. Kim, S. S. Lee, “Refined Day-of-Launch Atmospheric Flight Loads Analysis Approach,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 453–458 (Aug. 2000). J. A. Kechichian, “Minimum-Time Constant Acceleration Orbit Transfer With First-Order Oblateness Effect,” Journal of Guidance, Control, and Dynamics, Vol. 23, No. 4, 595–603 (Aug. 2000). J. A. Kechichian, “Transfer Trajectories from Low Earth Orbit to a Large L1-Centered Class I Halo Orbit in the Sun-Earth Circular Problem,” The Richard H. Battin Astrodynamics Symposium (College Station, TX, March 20–21, 2000). M. C. Kim, A. M. Kabe, S. S. Lee, “Atmospheric Flight Gust Loads Analysis,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 446–452 (Aug. 2000). H. C. Koons, “Application of the Statistics of Extreme Valves to Space Science,” Proceedings of the American Geophysical Union Spring 2000 Meeting (Washington, D.C., May 30, 2000), pp. 1–22. H. C. Koons and J. F. Fennell, “Space Environment—Effects on Space System,” Proceedings of the Chapman Conference on Space Weather (Clearwater, FL, March 2000), pp. 1–13. J. C. Latta, “Production Cost Improvement,” Proceedings, ISPA 2000 (Noorwijk, The Netherlands, May 8–10, 2000), pp. 1–30. T. S. Lomheim, “Hands-On Science Instruction for Elementary Education Majors,” Optical Engineering Reports: SPIE Education Column, No. 195, 11 (March 2000). K. T. Luey, D. J. Coleman, and J. C. Uht, “Separation of Volatile and Nonvolatile Deposits on Real-Time SAW NVR Detectors
(Nonvolatile Residue Detectors for Spacecraft Contamination,” Optical Systems Contamination and Degradation II—Effects, Measurements, and Control, Proceedings of the Conference (San Diego, CA, Aug. 2–3, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4096, 109–118. D. C. Marvin et al., “Enabling Power Technologies for Small Satellites,” Proceedings of Space 2000: The 7th International Conference and Exposition on Engineering, Construction, Operations, and Business in Space (Albuquerque, NM, Feb. 27–Mar. 2, 2000), pp. 530–536. J. E. Mazur, “Interplanetary Magnetic Field Line Mixing Deduced from Impulsive Solar-Flare Particles,” Astrophysical Journal, Vol. 532, L79–L82 (March 20, 2000). J. E. Mazur et al., “Characteristics of Energetic (Not Less Than Approximately 30 keV/nucleon) Ions Observed by the Wind/STEP Instrument Upstream of the Earth’s Bow Shock,” Journal of Geophysical Research, Vol. 105, No. A1 61–78 (Jan. 2000). J. E. Mazur, J. B. Blake, M. D. Looper, et al., “Anomalous Cosmic Ray Argon and Other Rare Elements at 1-4 MeV/nucleon Trapped Within the Earth’s Magnetosphere,” Journal of Geophysical Research, Vol. 105, No. A9, 21,015–21,023 (Sept. 2000). D. L. McKenzie et al., “Global X-ray Emission During an Isolated Substorm—A Case Study,” Journal of Atmospheric and SolarTerrestrial Physics, Vol. 62, No. 10, 889–900 (July 2000). D. L. McKenzie et al., “Studies of X-ray Observations from PIXIE,” Journal of Atmospheric and Solar-Terrestrial Physics, Vol. 62, No. 10, 875–888 (July 2000). M. C. McNab et al., “Mappings of Auroral Xray Emissions to the Equatorial Magnetosphere—A Study of the 11 April 1997 Event,” Advances in Space Research, Vol. 25, Nos. 7–8, 1645–1650 (Apr. 2000). K. W. Meyer and C. C. Chao, “Atmospheric Reentry Disposal for Low-Altitude Spacecraft,” Journal of Spacecraft and Rockets, Vol. 37, No. 5, 670–674 (Oct. 2000). R. L. Moser, K. D. Bell, et al., “Experimental Control of Microdynamic Events Observed During the Testing of a Large Deployable Optical Structure,” UV, Optical, and IR Space Telescopes and Instruments Proceedings of the Conference (Munich, Germany, Mar. 29–31, 2000), Bellingham, WA, Society of Photo-Optical Instrumentation Engineers, SPIE Proceedings, Vol. 4013, 715–726. R. L. Moser, M. J. Stallard, et al., “Low Cost Microsatellites—Innovative Approaches to
Breaking the Cost Paradigm,” AIAA Space 2000 Conference and Exhibition (Long Beach, CA, Sept. 19–21, 2000). T. M. Nguyen and J. M. Charroux, “Baseband Carrier Phase Tracking Technique for Gaussian Minimum Shift-Keying,” AIAA Space 2000 Conference and Exhibition (Long Beach, CA, Sept. 19–21, 2000). T. M. Nguyen, C. C. Wang, and L. B. Jocic, “Wideband CDMA for NavCom Systems,” AIAA Space 2000 Conference and Exhibition (Long Beach, CA, Sept. 19–21, 2000). A. V. Rao, “Minimum-Variance Estimation of Reentry Debris Trajectories,” Journal of Spacecraft and Rockets, Vol. 37, No. 3, 366–373 (June 2000). A. F. Rivera, “The Impact of Power Architecture on Electromagnetic Interference Control,” Space Power Workshop-2000 (Torrance, CA, April 13, 2000), pp. 1–14. B. H. Sako, M. C. Kim, A. M. Kabe, and W. K. Yeung, “Derivation of Atmospheric GustForcing Functions for Launch-Vehicle Loads Analysis,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 434–442 (Aug. 2000). K. Siri, “Study of System Instability in SolarArray-Based Power Systems,” IEEE Transactions on Aerospace and Electronic Systems, Vol. 36, No. 3, 957–964 (July 2000). C. E. Spiekermann, B. H. Sako, and A. M. Kabe, “Identifying Slowly Varying and Turbulent Wind Features for Flight Loads Analyses,” Journal of Spacecraft and Rockets, Vol. 37, No. 4, 426–433 (Aug. 2000). P. Thomas et al., “MightySat II—On-Orbit Lab Bench for Air Force Research Laboratory,” 12th AIAA/USU Annual Conference on Small Satellites Proceedings (Utah State University, Logan (Aug. 21–24, 2000). W. M. VanLerbergher et al., “Mixing of a Sonic Transverse Jet Injected into a Supersonic flow,” AIAA Journal, Vol. 38, No. 3, 470–479 (Mar. 2000). R. L. Walterscheid, J. H. Hecht, et al., “Evidence of Reflection of a Long-Period Gravity Wave in Observations of the Nightglow Over Arecibo on May 8–9, 1989,” Journal of Geophysical Research, Vol. 105, No. D5, 6927–6934 (Mar. 16, 2000). H. T. Yura, L. Thrane, and P. E. Anderson, “Analysis of Optical Coherence Tomography Systems Based on the Extended Huygens-Fresnel Principle,” Journal of the Optical Society of America A, Vol. 17, No. 3, 484–490 (March 2000). A. H. Zimmerman, “Charge Management Issues for NiH2 Batteries,” Proceedings, 18th Annual Space Power Workshop (Torrance, CA, April 11, 2000), pp. 1–15.
Crosslink Winter 2000/2001 • 57
Bookmarks Continued Patents C. J. Clark, A. A. Moulthrop, M. S. Muha, C. P. Silva, “Frequency Translating Device Transmission Response System,” U.S. Patent No. 6,064,694, May 2000. A three-pair measurement method determines amplitude and phase transmission response of frequency translating devices, including a device under test and two test devices using a vector network analyzer and a controller where one of the devices has reciprocal frequency response characteristics. The measurement method provides a lowpass equivalent transmission response of the devices. J. T. Dickey, “Microelectronic Substrate Active Thermal Cooling Wick,” U.S. Patent No. 6,070,656, June 2000. This device relies on the latent heat of vaporization and on differences in the coefficient of thermal expansion of materials, which will cause a flex in the wick structure as it is heated or cooled. Current technology removes energy at a consistent rate from all areas of a device. Because high heat areas change with time in most devices, this cooling wick actively modifies the fluid flow paths in reaction to those time-dependent changes. This allows the heat-dissipating device to maintain more uniform, cooler temperatures. The purpose is to aid in thermal control of high heat flux components on the micro scale. L. K. Herman, C. M. Heatwole, G. M. Manke, I. M. McCain, B. T. Hamada, “Pseudo Gyro,” U.S. Patent No. 6,020,956, Feb. 2000. By means of software processes, a pseudo gyro emulates mechanical gyros. It processes space system appendage measurement data and reaction wheel tachometer data within reference and control systems of a satellite using principles of conservation of momentum to compute vehicular bus angular velocity rate data by accounting for the momentum transfer between satellite bus and appendages. This technique can be useful for extending hardware gyro lifetime and/or replacement of failed gyros on satellites. R. S. Jackson, G. M. Manke, “Control System for Counter-Oscillating Masses,” U.S. Patent No. 6,107,770, Aug. 2000. This control system stabilizes the flexible body bending modes of a space, airborne, or ground-based system, while providing angular position control of an oscillating mass connected to a counter-oscillating counterbalance. The actuating mechanism uses two drive motors to exert torques on the mass and counterbalance, under the control of a feedback controller. The controller’s two channels generate first and second torque
58 • Crosslink Winter 2000/2001
command signals for the two drives. The second channel filters out input frequencies in a predetermined bandwidth about the frequency of the first torque command signal. The system reduces sensitivity to flexible body system modes and allows maximum pointing accuracy for the payload. Applications include lidar or other similar systems. S. H. Raghavan, J. K. Holmes, “NRZ and Biphase-L Formatted Hexaphase Modulated GPS Transmission Method,” U.S. Patent No. 6,075,810, June, 2000. This transmission method provides for the simultaneous modulation of the C/A code, the P(Y) code, and a new code, all modulating a single carrier for Global Positioning System (GPS) use in one or both L1 and L2 bands. This allows addition of a new military signal in the same frequency band and C/A code deniability without affecting the new signal. C. P. Silva and A. M. Young “High Frequency Anharmonic Oscillator for the Generation of Broadband Deterministic Noise,” U.S. Patent No. 6,127,899, Oct. 2000. This invention provides a means to generate very-high-frequency, broadband, chaotic electrical signals that exhibit noise-like characteristics. The signals can be used to carry information in the same way a modulated sinusoidal carrier is used in a conventional AM or FM system. The oscillator is based on a forced second-order Duffing equation and is quite robust against the deleterious effects that normally arise in high-frequency operation. This design provides the fundamental enabling technology needed to develop and demonstrate an operational microwave chaos-based communications link and thus determine the unique application contexts and features of such a system. E. J. Simburger, “Power Sphere,” U.S. Patent No. 6,127,621, Oct. 2000. This invention addresses the problem of connecting solar cells mounted on a curved surface such as a sphere by having the cells connected in parallel to the array power bus through individual DC-DC converters. This allows each solar cell to deliver all of the power that it can produce based upon the amount of sunlight it is actually receiving. The DC-DC converters provide the mechanism of boosting the voltage generated by each individual cell to a level usable by the load connected to the solar array. This invention makes practicable solar array shapes other than flat panels or cylinders. R. P. Welle, “Ultrasonic Power Sensory System,” U.S. Patent No. 6,127,942, Oct. 2000. In a coupled transducer system, a power signal sent from an external controller energizes the first transducer, which converts the signal into an acoustic wave that is commu-
nicated through the coupling medium to the second transducer. The acoustic wave is transformed by the second transducer into an electrical signal that can be converted into useful power. The primary advantage is the transfer of power through a coupling medium without the use of electrical power wires. A. D. Yarbrough, “Micromachined Rotating Integrated Switch,” U.S. Patent No. 6,072,686, June 2000. This is a micromachined switch and filter combination suitable for use in high-performance, millimeter-wave receivers. It consists of an electrostatically driven switch mechanism with a bidirectionally rotating member having two positions for integrated circuit connection to the input feed for nearby filter structures. Control traces carry electrical signals generating electrical fields to provide electrostatic force upon the rotating member and thus turn the switch clockwise or counterclockwise. K. Siri, “Power Converters for Multiple Input Power Supplies,” U.S. Patent No. 6,088,250, July 2000. This improved power converter uses two conventional charging and discharging switches operating in a complementary mode and driven by a switch driver controlled by a power factor correction controller. Use of a blocking diode for blocking cross-coupled short-circuit paths and an absorption capacitor are key design features. Applications include switching power converters used for power factor correction in polyphase power systems and power sharing among distributed power sources to a common load. S. H. Raghavan, J. K. Holmes, “NRZ and Biphase-L Formatted Quadriphase Modulated GPS Transmission Method,” U.S. Patent No. 6,148,022, Nov. 2000. This transmission method provides for the simultaneous modulation of two codes such as the GPS C/A code and an arbitrary new code modulating a single carrier for GPS use in one or both L1 and L2 bands. This would allow addition of a new military signal in the same frequency band and C/A code deniability without affecting the new military signal. T. M. Nguyen, J. M. Charroux, “Precoded Gaussian Minimum Shift Keying Carrier Tracking Loop,” U.S. Patent No. 6,148,040, Nov. 2000. This invention provides carrier phase tracking using data precoded GMSK signals. It is an improved timing recover loop offering closed loop generation of a data timing signal at a baseband frequency. The technique improves noise rejection, and provides fast data acquisition by operating at baseband.
Aerospace Systems Architecting and Engineering Certificate Program
Aerospace Roles in Space Systems Architecting, Acquisition, and Engineering 30-hour “core” course
Knowledge and skills building
Teaming for Systems Architects and Systems Engineers 20-hour course
Aerospace Systems Architecting Program 120-hour curriculum
• System synthesis • Front-end planning: systems and system-of-systems
• System analysis and evaluation • Acquisition life cycle: for single system
Skills reinforcement Mentored one-year assignments
Bruce E. Gardner, Principal Director of Learning Systems for The Aerospace Institute, is responsible for directing the overall planning, development, and delivery of education/training programs and multimedia learning support resources for the corporation’s employees and customers.
he Aerospace Systems Architecting and Engineering Certificate Program has been the centerpiece of The Aerospace Corporation’s training curriculum since 1995. The program, which includes architecting and engineering tracks as well as internships and on-the-job training, has had a major positive impact on corporate culture and staff technical capabilities. The Aerospace Architect-Engineer Role Aerospace has always maintained a strong national reputation as an architect-engineer of space systems, and the corporation’s highly skilled technical staff accounts for much of its success. Engineers and scientists with advanced degrees and years of experience apply their expertise and crossprogram knowledge to complex tasks, from initial concept development to deployment and operation. The Aerospace architect-engineer role has evolved significantly over the last decade, affected by the ending of the Cold War, changing defense priorities, increasing cost-consciousness, the commercial-
Space Systems Engineering Program
Internship/on-the-job training programs
The Aerospace Systems Architecting and Engineering Certificate Program. Space systems engineering involves processes associated with the conceptualization, design, development, and fielding of a system, including system-interface requirements management, interdisciplinary effectiveness analysis, and independent verification and validation activities. Systems architecting is the aspect of systems engineering concerned with determining a system’s purpose, formulating the concept behind the system, structuring the system, and certifying its fitness for use.
ization of space, and numerous emerging technologies. One major impact of this evolution has been the increasing customer need for technical staff with heightened systems awareness and enhanced, multidisciplinary skills in the development of complex missions and systems. The Certificate Program To address this need, in 1994 the corporation created The Aerospace Institute to facilitate the development of staff members’ space-systems engineering competencies. In 1995, the Institute piloted the Aerospace Systems Engineering Certificate Program, which featured more than 100 hours of instructor-led classroom training. The program has since been expanded to incorporate a systems-architecting curriculum plus on-the-job training and mentoring. Today the Aerospace Systems Architecting and Engineering Certificate Program comprises more than 300 hours of classroom training, involving more than 50 instructors from a broad cross-section of Aerospace engineering and program office organizations.
The program’s purpose is to develop the next generation of Aerospace technical leaders and program managers by enhancing • awareness of Aerospace roles and responsibilities in national space systems architecting, acquisition, and engineering functions and processes, and how those roles and responsibilities relate to those of customers and their contractors • fundamental understanding of the Aerospace perspectives, concepts, methodologies, and tools associated with system-of-systems architecting, space-systems life-cycle engineering, space-systems engineering management, and technical-team leadership • basic skills related to the use of those perspectives, concepts, methodologies, and tools • application of those skills to real-world, multidisciplinary technical (and related nontechnical) issues Program graduates receive the Aerospace Systems Architect-Engineer certificate, a measure of successful career progression within the corporation. Crosslink Winter 2000/2001 • 59
To ensure the program’s relevance and quality, the Institute has formed a Systems Architecting and Engineering Mentor Team comprising 10 senior corporate managers and technical leaders. The team provides general oversight of the program’s curriculum structure, content, funding, and participation. Curriculum The curriculum consists of an initial awareness segment, a knowledge and skills building segment, and a skills-reinforcement segment. 1. Initial awareness This segment is a 30-hour core course, Aerospace Roles in Space Systems Architecting, Acquisition, and Engineering, targeted to Aerospace technical staff and customers. Content includes the corporation’s technical mission, philosophy, roles, capabilities, and approaches in space-systems architecting, acquisition, and engineering. The course orients participants to evolving DOD and NRO customer and acquisition environments, surveys major Aerospace systems architecting-engineering analysis and program-management support tools, and highlights lessons learned from cross-program experience. Senior technical experts and program managers conduct panel sessions and deliver interactive lectures. 2. Knowledge and skills building The next segment enables technical staff to develop strong foundational knowledge and competencies in systems architectingengineering. The focus is on gaining indepth understanding of state-of-the-art methodologies and tools and how to apply them in their jobs. All participants take Teaming for Systems Architecting and Systems Engineering. This three-day course develops influencing and negotiating skills for effective participation in multiorganizational technical teams and work groups for which Aerospace may not be the formal leader. Participants progressing beyond Teaming choose the Space Systems Engineering Program or the Aerospace Systems Architecting Program. Each curriculum track comprises approximately 120 classroom hours. Space Systems Engineering Program. The engineering curriculum consists of four courses that cover Aerospace analysis methodologies and lessons learned from the space-systems-acquisition life cycle:
60 • Crosslink Winter 2000/2001
• Concept Development: translation of the customer’s statement of mission need into the quantitative engineering requirements • Space Systems Design: translation of engineering requirements into preliminary designs of the physical elements of the space-system solution • Space Systems Development, Integration, and Test: evolution of product toward detailed design, fabrication, and integration; system test approaches and analysis methods leading to “go-ahead” decisions for deployment • Space Systems Operations: launch preparation and integration, satellite deployment, day-to-day operations, and mission effectiveness assessment Students apply the Aerospace Concept Design Center’s concurrent engineering methodology to a case study involving the development of a space-based theater missile defense system, and they tour a
contractor spacecraft-manufacturing facility as well as the Vandenberg Air Force Base launch and processing facility. Aerospace Systems Architecting Program. The architecting curriculum was developed in response to growing customer demand for support in initial planning and synthesis of architectural designs for largescale systems-of-systems in which the space segment may be only one component. Aerospace staff must work effectively with these customers to • determine the real purpose for which an architectural solution is sought • develop measures of effectiveness and utility for the architecture • develop technically feasible architectural solutions that are satisfactory to all stakeholders To teach participants the skills needed to accomplish these objectives, the architecting curriculum uses an interactive sequence of lectures and integrated case studies. The
Space Systems Design
Concept Development System concept design
Early warning satellite Uncued threat attack
Cued threat attack
Space Systems Development, Integration, and Test
Space Systems Operations Operational effectiveness
Parts, materials, assemblies Environmental test
Track, telemetry, control Requirements generation User services
Subsystem/software test Test and evaluation plan Test requirements SRR
Space Systems Engineering Program. The courses convey an ideal acquisition process from concept through on-orbit operations. The material is presented sequentially, but its application is cyclic and iterative: within any phase, movement is forward and backward as concepts, systems, and products are defined and refined.
architecting program also familiarizes participants with defense-community architectural description frameworks (such as C4ISR) and develops participants’ skills in identifying and using Aerospace corporate systems-architecting resources. 3. Skills reinforcement The final segment is an on-the-job-training and internship program where participants apply principles learned in the classroom. They complete a mentored assignment that consists of at least two major program systems architecting-engineering support tasks. The assignment lasts a minimum of one year. Afterward, participants provide a formal briefing on accomplishments and lessons learned to the mentor team and other corporate executives. Program Successes Since its inception, the certificate program has successfully strengthened the skills of the Aerospace technical staff in systems architecting and engineering. More than
30 percent of the technical staff have participated substantially in the program. Customer personnel have also attended the core course. The program has had a clear beneficial impact on Aerospace program-support capabilities and effectiveness. Graduates successfully apply course concepts, lessons learned, and course materials in working with customers. The design and development of this curriculum have led to the creation and enhancement of several major architectureengineering support tools and methodologies. Over 10 volumes of well-organized, high-quality course lecture notes and casestudy examples have been developed by nationally known senior Aerospace technical experts. Finally, graduates have enhanced their networking and career-development opportunities. As a result of interactions with instructors and fellow students, they have
Complex problem System (new or existing) or system-of-systems
Political and economic realities
Diverse needs Flexible, iterative method Aerospace Systems Architecting Methodology Existing processes
Raw needs, constraints
Problem structuring Use cases, domain specifics Harmonize
Purpose analysis Technology Architecture descriptions
Solution System structuring models
Satisfactory solution Multistatic air vehicle surveillance • Global coverage • Deploy UAVs as required
Defeats Stealth, LO
Stealth Stealth Cruise missile
Aerospace Systems Architecting Program. The curriculum uses lectures and case studies to teach participants the Aerospace Systems Architecting Methodology in dealing with complex, unstructured scenarios.
Mark Maier of the Aerospace Engineering and Technology Group developed the Aerospace Systems Architecting Methodology.
made significant improvements in their ability to network with colleagues in other organizations and to gain access to relevant corporate expertise. Some certificate recipients have reported that participation has led to a new position or expanded job function. What’s Next? A third knowledge and skills building curriculum, the Space Systems Acquisition Management Program, will be piloted during fiscal year 2001. Attendees will include both Aerospace technical staff and Space and Missile Systems Center personnel. A new multimedia learning support tool, SEEK (Systems Engineering Educational Knowledge), will provide on-line access by the entire Aerospace technical staff to the full range of certificate-program course materials, supporting technical reports, and customer acquisition process information. This tool, which features a user-friendly database-search capability to obtain critical information on specific systems architectingengineering topics of interest, will be deployed in the near future. The certificate program’s success has helped create and promote a highly favorable climate for the development of improved systems architecting-engineering awareness and competencies throughout Aerospace. Strong corporationwide interest and participation in the program are expected to continue.
Crosslink Winter 2000/2001 • 61
Crosslink Winter 2000/2001 Vol. 2 No. 1 Editor in Chief Donna J. Born
Board of Trustees
Bradford W. Parkinson, Chair
E. C. Aldridge, Jr. Chief Executive Officer
Howell M. Estes III, Vice Chair
E. C. Aldridge, Jr.
David A. Bearden
William F. Ballhaus, Jr.
Richard E. Balzhiser
Guion S. Bluford, Jr.
Daniel E. Hastings
Thomas C. Hamilton
Jimmie D. Hill
John A. McLuckey
John A. Hoyem
Thomas S. Moorman, Jr.
Ruth L. Novak
Ann C. Petersen
Robert R. Shannon
Donald W. Shepperd
William C. Krenz, Chairman
Jeffrey H. Smith
David A. Bearden
K. Anne Street
Harlan F. Bittner
John H. Tilelli, Jr.
Donna J. Born
Robert S. Walker
William F. Ballhaus, Jr. President Michael J. Daugherty Executive Vice President Jon H. Bryson Stephen E. Burrin Marlene M. Dennis Rodney C. Gibson Lawrence T. Greenberg Gordon J. Louttit John R. Parsons Joe M. Straus Dale E. Wallis John F. Willacker
Linda F. Brill David J. Evans Isaac Ghozeil David J. Gorney Linda F. Halle Michael R. Hilton John P. Hurrell Mark W. Maier John W. Murdock Mabel R. Oshiro Frederic M. Pollack
The Aerospace Corporation P.O. Box 92957 Los Angeles, CA 90009-2957
Copyright 2001 The Aerospace Corporation. All rights reserved. Permission to copy or reprint is not required, but appropriate credit must be given to The Aerospace Corporation. Crosslink (ISSN 1527-5264) is published by The Aerospace Corporation, an independent, nonprofit corporation dedicated to providing objective technical analyses and assessments for military, civil, and commercial space programs. Founded in 1960, the corporation operates a federally funded research and development center specializing in space systems architecture, engineering, planning, analysis, and research, predominantly for programs managed by the Air Force Space and Missile Systems Center and the National Reconnaissance Office. For more information about Aerospace, visit www.aero.org or write to Corporate Communications, P.O. Box 92957, M1-447, Los Angeles, CA 90009-2957. For questions about Crosslink, send email to [email protected]
or write to The Aerospace Press, P.O. Box 92957, Los Angeles, CA 90009-2957. Visit Crosslink online at www.aero.org/publications/.
NONPROFIT ORG U.S. Postage
PAID El Segundo, CA Permit No. 125
Change Service Requested