Why Is It Difficult to Predict Mass Growth on Space Programs?
Richard Sugiyama and Louis Yang
Despite the correlation between system mass and mission cost, program managers lack tools and data for accurately predicting mass properties growth. Aerospace is working to redress this deficiency.
Predicting the mass of a space vehicle from conceptual design through final development is a complex and difficult task. Virtually every space vehicle experiences mass growth over time, yet engineers still lack a means to accurately predict, determine, document, and regulate this growth.
It would be difficult to exaggerate the importance of mass on a space vehicle program. Anticipating propulsion needs, figuring attitude determination, documenting control analysis—none of this can begin without knowledge of the vehicle mass. Further, an accurate prediction of space vehicle mass properties lays the foundation for effective space vehicle design, whereas an inaccurate prediction can lead to a vehicle that cannot achieve stability or meet its performance goals.
Aerospace has been working to produce better models to predict the evolution of space vehicle mass properties. A large part of this work entails the systematic study and documentation of mass data from past space programs. Additional efforts will focus on parsing the collected data to reveal relationships among factors that contribute to mass increase.
Mass Properties: Basic Parameters and Principles
A space vehicle can be mathematically characterized by a set of parameters that describe the distribution of mass in a rigid body. These parameters include mass, center of mass, moments of inertia, and products of inertia.
Mass is a measure of the amount of material in a body. Unlike weight, which is determined by gravity, mass doesn't change with a body's position, movement, or alteration of shape. The center of mass is a specific point at which the total system mass behaves as if it were concentrated. The center of mass is a function only of the positions and masses of the particles that comprise the system. In the case of a rigid body, the center of mass is rigidly fixed. In the context of a uniform gravitational field, the center of mass is sometimes called the center of gravity, because the net gravitational torque on a system is equal to the torque resulting in the system's weight applied at the center of mass. The center of mass of a system of particles is defined as the average of their positions weighted by their masses. If an object has uniform density, then its center of mass is the same as the centroid of its shape.
One property of mass is inertia, the resistance to being put in motion and the tendency to remain in motion once set in motion. Moment of inertia, also called mass moment of inertia, quantifies the rotational inertia of an object, i.e., its resistance to changes in rate of rotation. The moment of inertia of an object about a given axis describes how difficult it is to induce an angular rotation of the object about that axis.
Product of inertia is defined as the sum of the product of each element of mass by the perpendicular distances from two specified orthogonal axes. Nonzero product of inertia values can be a measure of a body's dynamic (or coupled) imbalance resulting in a precession when rotating about an axis other than the body's principal axis—that is, the body tends to wobble when rotating. Usually, products of inertia are defined with respect to the object's center of mass. The intrinsic inertia properties of a rigid body do not change with orientation; however, the values of the moment of inertia and product of inertia are dependent of the orientation of the body with respect to axes of the coordinate system used to describe the inertias.
Determination and control of these mass property parameters is vital to the launching and on-orbit stability of space vehicles.
Suspension of specifications and standards in the national security space arena may have contributed to reduced rigor in the oversight and control of mass for at least five major programs in the post-1990 period. |
The Knowledge Gap
Historical data indicate that half of the mass growth experienced on the average space vehicle program occurs between program start and the program preliminary design review. There is a direct correlation between system mass and mission cost, and a significant increase in mass during the design, fabrication, and build phases may be a good indicator of a significant increase in program cost. For example, in recent years, numerous space programs that experienced significant mass growth during the early phases of acquisition also experienced cost growth exceeding 15 percent, thereby triggering congressional review under the Nunn-McCurdy law. Subsequent penalties ranged from program cancellation to reduction in program scope. There is obviously a lot riding on the early, accurate prediction of space vehicle final mass. Why, then, is it so difficult today to predict final mass with a reasonable level of confidence?
Part of the answer is that most of the published guidelines and recommendations for managing mass growth are based on data from programs developed prior to 1990—and the industry, as a whole, has not increased its knowledge since that time. An Aerospace study of 21 space vehicle programs from that era showed a mass growth range of 9 to 53 percent, with an average of 28 percent. All but one of the space vehicles experienced a steep increase in mass during the initial part of the program. In 1992, the American National Standards Institute and the American Institute of Aeronautics and Astronautics published guidelines for minimum standard mass contingencies for spacecraft systems, somewhat consistent with the Aerospace findings. At the same time, Aerospace produced three separate papers examining the factors that tend to influence space system mass growth. Yet, little has been done since then to update the findings or act on the recommendations for future areas of study—largely because factors such as acquisition reform and industry consolidation limited any such efforts.
The other reason why growth in mass is difficult to predict is that the data are influenced by a number of programmatic factors (e.g., program maturity, technology maturity, system complexity, launch vehicle capability, mission category, etc.), and a more systematic approach is needed to understand their individual significance. For example, completely lacking is a quantitative correlation between mass growth and technology maturity and a simple listing of the typical causes of mass growth.
Renewed Efforts
Clearly, the industry needs to update its knowledge of satellite mass growth. To address that need, Aerospace initiated a mass growth study in 2002 to supplement those earlier findings with actual data from programs developed during the last 20 years. The goal was to compile, document, and maintain a database of space vehicle mass growth information that could be used to generate better predictions of mass requirements. The study was divided in two phases, a data collection and documentation phase and a data normalization and parametric study phase.
The collection effort focused on a wide range of government, commercial, and civil programs that were identified as programs of interest. Priority was placed on new or first-generation space vehicle designs and first vehicles in a significant new block purchase with increased complexity and capability. Specifically, researchers sought historical data pertaining to mass (dry nominal, dry with predicted growth, wet, launch), design maturity definition (percent estimated, calculated, actual mass), and functional subsystem mass breakdown to the unit level (mass fraction analysis). At a minimum, data would be gathered at several milestones in the procurement process: the authorization to proceed, the preliminary design review, the critical design review, and program completion (defined as final testing of mass properties, not the launch date). Additionally, other program variables (mission type, orbit, power requirements, battery capacity, pointing accuracy, launch vehicle, design heritage, etc.) were sought for future regression analyses to identify the most important variables or combinations of variables for predicting space vehicle mass growth.
The mass properties data collection and documentation phase is nearly complete. Aerospace has added 18 modern space vehicle programs to its database, bringing the total from 21 to 39.
Meanwhile, in a separate but related activity, Aerospace produced a technical report, TOR-2005(8583)-3970, "Mass Properties Control for Space Vehicles," as part of an Air Force effort to reintroduce specifications and standards into the acquisition process. Recognizing the importance of collecting the proper information to support future mass growth analyses, the new standard imposes specific requirements for including mass and mass-maturity information in the contractual mass properties status reports. The standard further defines a standardized format for reporting the reasons for mass change. That format lists the nine most typical reasons for mass change in a program and captures all the contractor in-scope changes that reflect the maturation of the design—including errors and omissions in the original estimate—as well as customer-directed out-of-scope changes (see sidebar, Nine Reasons for Mass Change).
Preliminary Analysis
Data collected from the 18 space vehicle programs completed after 1990 showed a mass growth range of 12 to 106 percent and an average of 39 percent. That's worse than before the era of acquisition reform. Six of the 18 programs experienced mass growth greater than 50 percent. In five of these six programs, the final mass predicted at the authorization to proceed was so grossly underestimated, one could only conclude that the proposal estimates were very poor. The reasons for the poor proposal estimates have not been analyzed in detail; however, the most probable causes were overly optimistic assessment of the hardware maturity (or lack of design maturity information), proposal estimates that were not scrubbed for completeness and accuracy, and requirements that were not fully defined or understood.
Whether these six programs are simply the products of a flawed acquisition strategy and are anomalies or outliers is still to be seen. In any case, SMC's systems engineering revitalization program is a first big step toward mitigating the risk of runaway mass growth on future programs. The lessons from the initial mass growth studies may help mitigate the risk of unplanned or unallocated mass growth on future programs.
The mass growth algorithms used to predict final space vehicle mass at program start, based on assessments of hardware design maturity, are not accurate enough. Data from 14 of the 18 modern era programs showed that average mass growth predicted by the algorithm was only 33 percent, and the average growth not predicted by the algorithm was 67 percent. The mass growth algorithms do not account for customer-directed out-of-scope changes, or poor estimates and hardware omissions, or errors in the proposal estimate. |
Data collected on 14 of the 18 programs completed after 1990 also showed that the mass growth algorithms applied industry-wide to predict final space vehicle mass at the start of a program (authorization to proceed) are not as accurate as would be desired. Part of the reason is that the algorithms used in the industry do not account for mass increases due to customer-directed out-of-scope changes and mass increases due to errors and omissions in the initial estimates. Aerospace's analysis from a very limited sample of programs estimated these changes (not accounted for by the algorithm) at 5–10 percent and 6 percent respectively. What this may imply is that additional mass contingency, in addition to that predicted by the mass growth algorithm, should be held by the contractor or customer in the early phases of the program.
Future Steps
Aerospace is embarking on the second phase of the 2002 mass growth study, which will focus on analysis of the data, such as mission, acquisition agency, or bus-payload mass fraction. Much of this work will involve analyses to determine whether parametric relationships exist between space vehicle mass growth and other program variables. Part of this effort will entail defining suitable indexes for maturity and complexity that can be assessed at program start. Another major task will be to collect and document the reasons why various programs experienced mass increases.
The results of these analyses could be used to standardize and fix the industry-applied mass growth algorithms that underpredict the final mass of the space vehicle, especially at program start. Mass contingencies, based on mass growth algorithms, applied industrywide to predict final space vehicle mass during the early program phase, are not accurate enough, and correcting this deficiency remains a primary objective. Using probability distributions to treat mass growth estimation as a parametric process may provide estimates that are more useful and meaningful; Aerospace is investigating that possibility. An ultimate goal is to provide guidance to the program offices and contractors on how much mass growth to plan for on future space programs.
The uncertainty of how much mass growth to anticipate remains the primary challenge. The initiatives prescribed in the mass growth study plan are the first steps in establishing a more disciplined systems engineering approach to mitigate the uncertainty in the mass growth risk on space vehicle programs. The long-term vision is to sustain this effort so that future programs can make more realistic mass estimates and, concomitantly, more accurate cost estimates.
