Enabling Technologies for Unified Life-Cycle Engineering of Structural Components (1991)

From the needs and concerns and the enabling technologies discussed previously, a set of conclusions and recommendations has been formulated and for each recommendation, a specific action item has been proposed. Most recommendations respond to one of the four critical issues presented in Chapter 7. The first three are general in nature; recommendations 4 and 5 derive principally from the first critical issue; 6, 7, and 8 from the second; 9, 10, and 11 from the third; and 12, 13, 14, and 15 from the fourth.

The breadth of the technologies encompassed by unified life cycle engineering (ULCE) and the level of development needed will require that significant resources be expended to generate a functional ULCE environment. These resources will become available from both the private sector, and government activities (both new and redirected). The proportion of total resource expenditures that will be attributable to each potential source will be dictated largely by the rapidity with which ULCE is implemented. The shorter the time, the greater will be the proportion of the total resource allocation to be met with additional government resources. A review of the capabilities required for full ULCE implementation in a major system (e.g., the advanced tactical fighter) suggests that research, development, and engineering (RD&E) for a period of 10 to 15 years will be needed. Thus, for planning purposes, a development program costing $100 to $200 million and directly supporting about 50 technical specialists should be envisaged, with impact in the near, intermediate, and long term. The actions proposed in subsequent recommendations show the scope of program needed; these have been summarized with an estimated time phasing in Figure 8-1. Appropriate time phasing is important if all the diverse technologies required for ULCE are to be developed in concert to provide the technical basis for an integrated ULCE system. Given the breadth and depth of the ULCE program, efforts will have to be leveraged by taking full advantage of R&D projects that are not ULCE-funded.

• A detailed research and development (RD&E) plan and budget for a 10-to 15-year technology development program must be prepared. The plan and budget should be cognizant of commercial and other government resource allocations. Responsibility for preparation should be assigned to a ''program office'' and program manager (Recommendation 3 below). However, ULCE R&D can be initiated before a complete plan is approved based on actions proposed in subsequent recommendations in this chapter.

The breadth of ULCE mandates a series of parallel studies developing the technological underpinnings. Such a series will not, however, provide opportunity for evaluating the effectiveness with which the differing technologies interact or for fostering ongoing technology transfer among contractors, technical personnel, and program managers. Since the major benefit of ULCE is its integrated approach to design, manufacturing, and support, the ULCE development program must include an effort to evaluate the effectiveness with which subsystems are integrated. This effort should be a demonstration project. The process of implementing a demonstration project would provide an opportunity to "look over the designer's shoulder" to better determine whether ULCE has the appropriate tools and prompts to guide the designer to generate the best design. Further, it would in some limited sense provide an opportunity to review the robustness of the design or its ability to accommodate evolutionary changes as the design progresses from concept to final design. In addition, it would provide a mechanism to evaluate how well the ULCE environment accommodates engineering changes.

• A major subsystem or module of a current vehicle should be identified and then redesigned, re-engineered, or remanufactured employing ULCE methods.

ULCE requires the coordination and integration of research, development, and implementation in a variety of diverse technologies. Without overall control of progress and timing, these technologies will not reach fruition simultaneously, so that the systems benefits of ULCE can be achieved. Because the timing is so critical, it is imperative that some agency or individual have overall responsibility for ULCE. The goals of ULCE are relevant to many government agencies, particularly to the Department of Defense (DOD), since it purchases, operates, and maintains equipment designed to its specifications. Because the Air Force has already made a commitment to ULCE and generated momentum, it appears to be the appropriate agency to initiate the ULCE program. Other agencies, including the Army and Navy and non-DOD activities (e.g., National Aeronautics and Space Administration (NASA) and Coast Guard), should be encouraged to define the extent of their involvement.

• The Air Force should initiate an aggressive ULCE program and provide mechanisms both for collaborating with the Services and other government agencies in their ULCE programs and for reporting on its program to DOD.

Recommendations 4 and 5 derive principally from the first critical issue, ULCE-driven development of materials design processing and repair methodologies reguires integration of research and development across disciplines.

Materials research and development is usually focused on the characteristics of specific composites, ceramics, etc. Usually the results for one material cannot be used to predict quantitatively results in another material, even in the same class (metals, ceramics, polymers, etc.). Therefore, materials R&D has generated a vast yet incomplete data base. For ULCE, materials behavior must be sufficiently understood to permit reliable extrapolation from a limited set of knowledge. Ideally this extrapolation should permit a valid system prediction for the service conditions anticipated and for the manufacturing procedure used. This capability exists for structural applications of conventional alloys in a qualitative sense but not in a quantitative sense. For development of advanced materials, it is essential to provide at least a semiquantitative capability as soon as possible. For this, an understanding of basic physical mechanisms is required.

• Materials R&D should be focused on developing lifetime prediction capabilities for generic classes of applications and materials phenomena. To achieve this goal, it will be necessary to initiate programs to identify the state variables and general principles underlying the mechanical behavior of solids under a broad range of environmental conditions.

The very specific needs of ULCE for materials information will lead to changes in the way in which materials research should be undertaken and funded. It is essential that the research philosophy and needs of ULCE be widely known and endorsed by both the funding agencies and the materials R&D community.

• A consensus should be formulated on a materials R&D strategy, consistent with the needs of ULCE, that would be supported and followed in future ULCE-funded research projects. A means to accomplish this would be through a workshop with representatives from all funding agencies, major materials R&D facilities, and major contractors. The responsible agency for ULCE should initiate the organization of this workshop immediately.

• A continuing series of conferences focusing on ULCE-related materials developments should be arranged, possibly in conjunction with national technical societies or with the Henniker Engineering Conferences.

Recommendations 6, 7, and 8 derive principally from the second critical issue: Advanced analytical modeling and simulation methods to credict actual component manufacture, operation, and logistics do not exist to the degree reguired to preclude the need for physical prototypes and mock-ups.

Both funding agencies and contractors need to follow a consistent procedure to predict and assess life-cycle costs, since life-cycle costs will be the basis for decisions on competing design and system attributes. (Note that this does not involve decision-making itself. The intent here is to develop an appropriate common denominator to which engineering characteristics can be reduced, so that alternative concepts can be quickly compared.) A complete life-cycle cost methodology involves many elements, including agreement on standard terminology and procedures; a means of assessing individual component costs while recognizing system costs; quantification of difficult-to-quantify performance parameters; a system that permits initial estimates to grow naturally into detailed estimates as the design evolves, without redoing all calculations; and detailed estimates that are consistent with initial estimates but are not constrained by those initial estimates. Some of these issues must be factored into the conceptual system design, whereas others can be addressed as costs are developed and integrated into the framework of the conceptual design. The key issue, though, is that all parties develop a consensus on the appropriate approach.

• A standard dictionary of terminology should be developed; this may involve little more than endorsement of MIL Standard 1388.

• A basic cost structure framework should be delineated that includes and clearly identifies each major ULCE cost category and (a) is flexible enough to permit rapid initial cost

assessments and yet powerful enough for final detailed costing; (b) will permit subcontractor data to be merged with a prime contractor's estimates without recalculation; and (c) is modular and self-consistent so that familiarity with any part of the system permits understanding of the whole system. The development of a method by which design attributes can be evaluated on a self-consistent basis is key to the success of ULCE. Thus it is essential both that this action item be developed rapidly and that consensus be achieved on the framework adopted. Government agencies, contractors, and subcontractors must be directly and intimately involved in this effort.

• A cost methodology for difficult-to-quantify mission parameters needs to be developed for inclusion in the cost calculator. Examples might be loiter time, maximum speed, and payload.

Current CAD-CAM systems provide excellent geometric part definition and display capabilities. However, a complete product description includes not only the part geometry but also tolerances, materials, process descriptions, etc. Commercial CAD-CAM systems do not have the capability of incorporating descriptive, nonquantitative, and nongeometric data into the part data base. Nor do the data bases include information regarding part manufacture, such as process scheme, tooling, fixturing, and cutter selection. In addition, the incorporation of features into the CAD-CAM system is essential. These will enhance and accelerate the mechanical design process and, more importantly, because features convey a more global sense of the design, will provide a convenient basis for linking design and performance attributes. By designing through features and associating features with specific performance attributes such as tool access, failure mechanisms, and inspectability, characteristics such as manufacturability, reliability, and maintainability can be rapidly assessed. Thus improved designs can be created more quickly. (Note that, even though this field is under active commercial development, progress has been slow. Additional support and sponsorship is required to meet ULCE goals, and new approaches are needed.)

• The development of industry standards (e.g., PDES) that could be incorporated in software should be fostered. This would significantly leverage individual industry efforts.

• R&D should be supported on features-based CAD-CAM systems where the features include all aspects of ULCE.

• R&D should be sponsored to define the association between product performance—particularly for manufacturability, reliability, and serviceability—and specific design features.

At present, the capabilities of complex systems can be determined with certainty only after they have been operated throughout the desired performance envelope. Inevitably, deficiencies and potential improvements will be identified leading to redesign, component modifications, and further evaluation. Typically this process has been undertaken by reducing the designs to hardware, assembling prototype vehicles, and conducting flight testing. A far superior approach, which is becoming feasible with current computer capabilities, is to construct models whose response accurately simulates that of the structure. With this capability the design can be optimized while it is still in the conceptual stage, thereby eliminating the costs (in both time and money) of modifying and repackaging the physical components. For ULCE, more emphasis needs to be placed on improving both the quality and application of analytical methods—new approaches to dealing with stochastic systems so that the influences of, for example, material variation, tolerance stack-up, operating variability, on the design can be explicitly evaluated. Better coupling of analysis with experimentation is needed, particularly at the component and subassembly levels. The recommended ULCE demonstration project would allow validation of some of these analytical approaches for predicting product behavior.

• Development of enhanced user interfaces for analysis programs and enhanced integration of analysis data with other CAD-CAM tools should be promoted. These are not just "user-friendly" issues like common function key assignments, but new modeling methods that reduce the time, effort, and skills required to accomplish the analysis. Needs should be defined by a committee of users and commercial vendors to develop a requirements and standards document. Market forces will lead to commercial development.

• R&D methods warrant support for analysis of stochastic systems that can properly account for variability in materials properties, manufacturing processes, operating environment, etc. The goal of these efforts should be the development of algorithms that can be used in commercial software packages.

• Application of new, high-performance computer architectures to finite element and stochastic modeling systems should be studied. This research should address compilers, application source codes, and basic algorithms and should lead to orders-of-magnitude improvement in processing time. ULCE software applications should be usable for various hardware and operating-system platforms.

• Provide governmental encouragement to computer mainframe producers to develop supercomputer systems that can support the computational speed, memory, and storage required for CAD-CAM solid modeling and automated manufacture.

Recommendations 9, 10, and 11 derive principally from the third critical issue: The information system for an integrated team approach to ULCE is inadequate.

A major element of ULCE is the capability of providing the designer a wide range of information, engineering data, design data, and engineering analysis capabilities. Since it is impossible to have all the information potentially required by a designer resident in a single computer system, multiple data bases will exist, and access, data distribution, and other issues should be resolved to ensure the appropriate data system structure. Information structure models for computer-integrated manufacturing (CIM), similar to those for ULCE, are currently under development, and there are significant advantages to working for compatibility between the CIM and ULCE reference models. The reference model should be independent of computer supplier, computer hardware, and software.

• Existing conceptual information reference models should be expanded to include ULCE by combining or extending existing models (e.g., IDS, PDDI/GMAP, ICAM, CIM-OSA) and extending these models to include logistics, distribution support, and refinement. A major focus should be on information feedback from manufacturing and support to design.

• With the reference models as a basis, areas of ULCE opportunities and priorities should be identified and recommended for further study, prototyping, and demonstration.

• The reference data model should be employed both in the ULCE demonstration project (Recommendation 2) to evaluate its capabilities and in education and training activities (Recommendation 14).

Because of the distributed nature of the data required to perform the total design function within ULCE, much of the design effort will involve accessing and manipulating distributed data, incorporating the data in the design, and storing this enhanced design for later access. Because of the volume of data transfer, storage, and retrieval required, it is clear that even very small error rates could significantly affect the process. Thus it is important to employ standard representations that may include a degree of redundancy in the information transmitted.

• Commercial developments in representation of entities should be pursued to determine if they are appropriate to and sufficient for ULCE needs.

• Shortfalls in commercial development in representation of entities should be funded through the ULCE program with input from industry.

The design process consists of several stages, beginning with the concept development and followed by numerous steps of concept refinement. Decisions and trade-offs on both component and systems levels are made continually throughout these design phases, but some major strategic decisions are made in conceptual design. Evaluation tools must be in place to permit rational design assessment at all levels. Furthermore, the results of the early analyses must be retained and used as the basis for the increasingly refined evaluation tools employed as the design evolves. Some vehicle characteristics, particularly performance, can already be analyzed in this way, but producibility assessment is only partially implemented, and supportability is not implemented at all.

• An approach to design analysis capabilities should be developed that includes a features basis for maintainability and supportability and that is endorsed by contractors and DOD. (Because of the wide variety of CAD-CAM systems in use, implementation of this approach would be the responsibility of the contractor.) Standards for electronic exchange of product definitions need to be extended to include features-based designs.

• A specific, although reasonably simple, structural component for each system should be designed and evaluated according to the approach in the above action item to determine the success with which the approach has been defined and implemented. This should be coupled with the demonstration project (Recommendation 2).

Recommendations 12 through 15 derive principally from the fourth critical issue: The ULCE team will need to make key decisions while still operating with incomplete information.

Although programs are in place to identify failures and replacement requirements for the current vehicle fleet, this information has little impact on design. There are two reasons for this: first, the data are frequently not communicated to the designer; second, the data are not organized in such a way that a design engineer can generalize the information to develop associations between failures and design characteristics. One approach to addressing this is to restructure the data gathered previously to make it more relevant. Another is to revise the information-gathering procedure itself so that the data, as gathered, have greater relevance. It is clear that an improved information-gathering procedure will eventually negate the need to restructure previously-gathered data. In the short term, however, it is desirable to make use of the previously gathered data if it can be restructured at reasonable cost.

• Design, process, and manufacturing engineers should be assigned to operating bases for relatively short (6-month to 2-year) assignments. Their responsibilities would be to monitor failures to understand underlying defect mechanisms and to provide direct feedback to designers about repair difficulties along with suggestions to improve reliability and serviceability.

• Current data and lesson-learned repositories related to repair and serviceability should be reviewed to determine whether the data, or a subset of the data, can be restructured to provide designer feedback and, if so, whether this can be done rapidly enough and economically enough to pursue.

• Current data derived from peacetime operations should be reviewed to determine relevance to wartime requirements.

Service operations expose structures to damage. Although some of the damage can be anticipated based on past history, many of the failure modes, particularly with new materials, will be incompletely known. Thus they cannot be incorporated in any model. The design options in this case are to overdesign significantly in hopes of providing a sufficient margin of safety for almost any eventuality or to design to a much lower margin of safety and continually monitor vehicle integrity using sensors. Clearly, the second approach should offer a better performance trade-off and therefore should be pursued. Furthermore, since the sensor characteristics are known, the modeling parameters can be coordinated with knowledge of sensor placement, sensor sensitivity or other parameters affecting sensor operation. In addition, by using modeling techniques in analyzing the sensor results, a rational assessment of vehicle capabilities, consistent with the original design approach, can be achieved. Note also that in the short term this technology provides the bridge between current understanding of materials behavior and the future state-variable approach of Recommendation 4.

• Development of a prototype component incorporating embedded sensors should be funded and evaluated for its performance under laboratory conditions.

• Parallel development and evaluation of production components incorporating sensors should be undertaken for both laboratory and service conditions. This should be incorporated in the ULCE demonstration project (Recommendation 2).

• Damage models should be reviewed to determine whether they can incorporate sensor data directly to provide immediate output on remaining life. If not, more direct coupling between sensor data and models must be obtained.

ULCE will require a broader knowledge of product behavior and will require the design engineer to interpret and evaluate a wide range of information in a variety of contexts. Currently there is no formal educational program that embraces these skills, and there is no opportunity to acquire these skills in a structured manner. Therefore, new educational initiatives are required to develop a framework where new, bachelor-level employees are aware of ULCE concepts and methods. The issue must be addressed on two levels, teaching and research, to ensure that ULCE enjoys sufficient prominence to attract and retain the most capable faculty. University-industry interaction will be vital in ensuring that the curriculum, well supported by research, addresses the technical and management needs of ULCE.

• On a cost-shared basis between government and industry, the components of an industrial short course on ULCE for personnel from all institutions should be developed. The curriculum development may be aided by faculty involved in university activities and by analysis of case studies. It should be equivalent to a single-semester, 3-credit-hour course at the graduate level.

• Initiation of Master of Science in Life Cycle Engineering programs at several universities should be encouraged. The administration and organization of the programs should be patterned after the graduate programs in manufacturing systems engineering recently started at several universities.

• The initial funding for research centers on ULCE at major universities should be provided through federal agencies. These programs should be organized along the lines of National Science Foundation's (NSF) Engineering Research Centers to provide a coherent environment for developing and integrating ULCE technical and management tools.

Vehicle performance will always be a major driver in vehicle design. Thus there will always be incentives to employ higher-performance materials, new manufacturing techniques that promise higher-performance components, etc. Although some body of data will be available to support these new approaches, the volume of knowledge required to completely categorize them, in particular long-term behavior, means that the data available will be incomplete. Thus some approach that rationally evaluates these missing attributes is desirable. Also, the development of appropriate reliable methods for dealing with incomplete information will permit selection of new approaches on the basis of even less information and will support even more aggressive implementation of new technology. Note, however, that most of the artificial intelligence approaches to working with incomplete information are at a very early stage of development, and their implementation will require significant basic research.

• Expert systems should be developed where appropriate, recognizing their limited domains of applicability.

• The types of technologies appropriate to the generalization and specialization of inferred knowledge should be developed. Three technologies are appropriate: statistically based, machine learning, and neural networks.

• Other artificial intelligence approaches should be pursued on a long-range basis.