Concurrent Design at Aerospace

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.

Concurrent Design at Aerospace

Patrick L. Smith, Andrew B. Dawdy, Thomas W. Trafton, Rhoda G. Novak, and Stephen P. Presley

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.

Imagine 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.

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. Real-time 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-to-face 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).

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 high-level 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 conceptual-study 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 space-systems 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 system-development 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.

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 personal-computer 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 space-system 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.

Design information using linked spreadsheets

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.

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 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-subsystem-design 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 spacecraft-subsystem 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 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.

A dedicated facility for design sessions

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.

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 system-level 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 concurrent-design 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 (see sidebar, CDC Teams). 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 top-level 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.
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 staff-hours 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.

Aerospace specialists use linked spreadsheets during a design session. Left to right: Christopher Taylor, Eric Hall, Mark Mueller, Douglas Daughaday.

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.

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.

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 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 (see sidebar, CDC Successes).

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 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.

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.

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.