Information age manufacturing begins with information age design. The design process for products and the processes by which they are produced involve compiling and understanding the products’ requirements, converting the requirements into engineering specifications, and producing plans that marshal the materials, equipment, and people needed to make and deliver the products.
Central to the 21st century design environment will be product and process modeling techniques that specifically and unambiguously define product and process attributes. Such modeling techniques would thus enable designers to determine production requirements and communicate them to other designers and to individuals upstream (e.g., in marketing) and downstream (e.g., in production).1
Also, such modeling techniques would eliminate or minimize the need for physical prototypes and construction aids. Production of physical prototypes and aids for use in functional testing, tool construction, and related activities introduces more lead time, cost, and potential inaccuracies into the development process and can cause unnecessary iterations when design changes cause models and aids to get out of synchronization with the product. In programmable electronic products, such modeling allows the supporting software to be tested on early versions of the hardware, rather than waiting until hardware development is complete.2
New design challenges are presented by (and are enabling) new product trends, such as increasing complexity, shrinking size, growing numbers of parts, and increasing technological diversity. As a result, the amount of knowledge and data a designer must synthesize, compare, trade off, and optimize is rapidly becoming more than an individual can comprehend. It is already beyond human comprehension for some complex products, such as VLSI chips.3
Running through all of design and production are major issues of scale-up and complexity. Any research that is done must take scale and complexity into account and not expect that they will be included at a later stage of the development of any proffered solution to any aspect of these problems.
To enable the designer to overcome these challenges, the information age integrated product and process design environment must help the designer to link product design and manufacturing to an unprecedented extent with minimal use of physical prototypes or construction aids. Specific capabilities hardly possible today include the ability to
Create a concept of a product and make extensive simulations of its behavior without specifying all of its details;
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs 2 Integrated Product and Process Design Information age manufacturing begins with information age design. The design process for products and the processes by which they are produced involve compiling and understanding the products’ requirements, converting the requirements into engineering specifications, and producing plans that marshal the materials, equipment, and people needed to make and deliver the products. Central to the 21st century design environment will be product and process modeling techniques that specifically and unambiguously define product and process attributes. Such modeling techniques would thus enable designers to determine production requirements and communicate them to other designers and to individuals upstream (e.g., in marketing) and downstream (e.g., in production).1 Also, such modeling techniques would eliminate or minimize the need for physical prototypes and construction aids. Production of physical prototypes and aids for use in functional testing, tool construction, and related activities introduces more lead time, cost, and potential inaccuracies into the development process and can cause unnecessary iterations when design changes cause models and aids to get out of synchronization with the product. In programmable electronic products, such modeling allows the supporting software to be tested on early versions of the hardware, rather than waiting until hardware development is complete.2 New design challenges are presented by (and are enabling) new product trends, such as increasing complexity, shrinking size, growing numbers of parts, and increasing technological diversity. As a result, the amount of knowledge and data a designer must synthesize, compare, trade off, and optimize is rapidly becoming more than an individual can comprehend. It is already beyond human comprehension for some complex products, such as VLSI chips.3 Running through all of design and production are major issues of scale-up and complexity. Any research that is done must take scale and complexity into account and not expect that they will be included at a later stage of the development of any proffered solution to any aspect of these problems. To enable the designer to overcome these challenges, the information age integrated product and process design environment must help the designer to link product design and manufacturing to an unprecedented extent with minimal use of physical prototypes or construction aids. Specific capabilities hardly possible today include the ability to Create a concept of a product and make extensive simulations of its behavior without specifying all of its details;
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs Explore, trace, and compare the manufacturing cost, quality, and performance implications of design decisions without specifying the design in full detail; Obtain a full understanding of the product’s behavior modes, including major off-nominal variant behaviors that may seriously affect performance, product quality, environmental quality, or user safety; Express the desired functions of the product in a translatable and analyzable format and thus permit function descriptions and mapping of functional requirements onto engineering entities within the product, including decomposing functions into subfunctions and system or engineering entities into subentities; Identify the right manufacturing, assembly, and test processes for creating and verifying the elements of the product and matching them to the functional behavior specifications, and predict the expected variability of the resulting manufactured entities during design; Obtain current information characterizing manufacturing and assembly processes, including variation expectations, not only for individual process machines but also for entire production systems; and Achieve the above product definition depth through affordable computer assistance that is flexible enough to accommodate design innovation. Creating a design environment with such functionality presents a challenge in itself, as well as a number of research opportunities. These opportunities are listed in Table 2.1 and are discussed in the remainder of this chapter.4 MANUFACTURING-SPECIFIC RESEARCH Design by Function Currently, a designer specifies products in terms of production-oriented parameters such as length of shafts or diameters of bearings. In the design environment envisioned for 2010, the designer will be able to specify products in terms of function and performance directly in function-relevant parameters, such as the load capacity and life of the bearing or the allowed vibration frequencies and minimum fatigue resistance of the shaft. Design by function is now possible only in a few domains, such as application-specific integrated circuits.5 Research is needed to address the following design-by-function needs:6 Support of functional descriptions in rough conceptual terms; The ability to simulate meaningful behavior at the conceptual level; The ability to predict performance or production problems with conceptual-level information; Methods for decomposing functions (i.e., a top-level description of what a product should do) and subfunctions (the next level down) and mapping those subfunctions onto system and subsystem elements (i.e., deciding what physical parts of the product will accomplish them);7 Methods for quickly converting functional descriptions into system space allocation and part geometry; and Methods for predicting production costs and possible problems.
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs TABLE 2.1 Product and Process Design Subject Area Example of Research Needed Design by function Relation of geometry to function Functional simulation Product-process data model Data descriptions for many physical processes and entities in a unified form Descriptions of design interactions, analyses, and process steps integrated with product geometry and function descriptions Capture of nominal and variant behavior of products and processes in one model A mathematics of variation for performance modeling Descriptions of product function and variants directly related to descriptions of geometric or material variations Design methods and tools for groups of parts and systems Decomposition methods to break product concepts into subsystems Subassembly performance models and interface descriptions for joining subassemblies to each other Assembly planning Process description languages and models Set of process primitives from which process models can be built Languages with syntax checking for formal correctness and logical completeness of process descriptions Novel design considerations Easy, error-free configuration control at the selling or servicing stage Economical build-to-order in lot sizes as low as one Manufacturing of a robust final product from parts obtained from different sources Decision aids Data visualization Database searching using geometric features, performance criteria, or process descriptions Intelligent advisors Geometric reasoning Visualization tools Knowledge and information management Systems that capture corporate memory and knowledge Systems that support corporate learning Techniques for handling data legacy issues Systems that record design history and rationale
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs Product-Process Data Model Current data models capture only geometry. The envisioned design environment will support a “product-process data model” that captures function and relates it to geometry, manufacturing, and assembly processes, interactions with the user, and ultimate disposal or recycling. Specifically, the product-process data model will have the following capabilities:8 Describe function, model variations, and families of the product; Describe information on which process models can work; Represent performance, geometry, and process requirements in ways that are readable by design tools and practitioners in the allied domains of process equipment design and shop floor planning and operations, as well as by designers in other companies or in other technical domains; Capture input as the product realization process proceeds so that the process can be improved; and Contain systematic ways of relating function to geometry; for example, it may be feature-based.9 Data models should be capable of representing the product or component from many different views at different stages of design, manufacture, and use. The product designer should be able to view its geometry, tolerances, assembly problems, and repair scenarios. The process designer should be able to study different fabrication methods and assembly sequences, comparing their cost, tolerances, and final quality. The production planner should be able to price different sources for material and place orders. The process engineer should be able to log process improvements or design changes so that the product designer can learn and do better. At present, no such data model exists, nor are there adequate methods for supporting these activities. Ideally, the data model’s function and geometry will be mutually convertible from one to the other. Although such convertibility could be cost-prohibitive in some cases, it would reduce storage requirements, because compact functional descriptions could be stored and voluminous geometric descriptions could be generated from the functional descriptions when needed. Capture of Nominal and Variant Behavior In addition to capturing function and relating it to geometry, product descriptions and the process descriptions of the envisioned product-process data model will capture in a unified way both the nominal behavior of the product and the main off-nominal behaviors, or variants. Nothing like this is currently possible. The need is illustrated by the example of tolerances. Currently, there is no single way of expressing product function specifications and limits that translate directly into the performance variations of the production equipment that will create the product. If the designer wants output power to be within ±x percent, he or she must express this as a requirement that some set of material or geometric properties lie within ±y percent of nominal values. The relationship between the power and the properties is usually difficult or impossible to establish conclusively, and whether a process can deliver the properties is often indeterminable as well. Consequently, it is difficult or impossible to determine whether the right set of processes and tolerances has been selected to deliver a specified range of performance variations.10
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs Design Methods and Tools for Groups of Parts and Systems Overlapping the need for models to capture nominal and variant behavior is the need for methods or tools to handle groups of parts such as assemblies, subsystems, product families, made-to-order configurations, and selected combinations of parts that create different product models by virtue of which parts are selected. Again, tolerance issues illustrate the need. Although the support for addressing tolerances for single parts is deficient, methods and tools to assist the designer in dealing with tolerance implications for assemblies are virtually nonexistent. Combinatorics and scale-up barriers defeat current approaches.11 The envisioned design environment will support design methods and computer tools adapted for the design and creation of groups of parts or systems, in addition to individual parts. Such methods and tools will enable the designer to divide a product into subassemblies, design optimal in-process test strategies during assembly, and identify assembly sequences that minimize cost, tolerance errors, and part damage during assembly. Also, the methods and tools will include standard design modules and methods that facilitate optimization of part or all of a product for cost and quality. Such optimization requires a deep understanding of specific components and features of the product. Because such understanding is often proprietary to component suppliers, research in this and similar areas of collaborative design must consider nontechnical issues such as intellectual property rights and intercompany data exchange. Process Description Critical to all manufacturing operations is the manufacturing process. The process performs the value-added operations to the product that is being produced. To design processes (even to consider product design as a process) and to enable process improvement, a precise way of describing processes is needed. The range of processes that needs to be described includes Business processes (e.g., converting a customer order into a list of the required parts, subassemblies, and documentation); Design processes (e.g., converting customer requirements into engineering specifications, converting specifications into subsystems, and allocating function to regions of space); and Production processes (e.g., moving materials, converting material properties or shapes, and verifying process results). Current process descriptions generally are made of text that captures only a portion of the required information and are thus ambiguous and incomplete. The envisioned design environment will support a process description that enables the designer to characterize the performance of a process (the efficiency or time of a design process, the quality of a production shop) so that he or she can compare sources, select the right shop to make a particular item, name the right team to do a certain design, and select the right algorithm to do a certain calculation. The process description should be dynamically modifiable so that adjustments can be made as the product realization process proceeds and tolerances can be reduced as processes are improved. Such functionality will enable the process designer to ensure that the required results really will be obtained and will facilitate systematic process improvement.
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs Process Description Language The envisioned design environment will support a language to express such process descriptions. This process description language must be translatable across technical domains and have a syntax that enables checking for formal correctness and logical completeness. Also, the language must be able to express not only nominal process behavior but also variant behavior. Process Description Models Process description models are critical for the local control and planning of manufacturing operations. Compact, accurate models of spatial and temporal process transformations are critical for both the planning and the control of manufacturing activities. Accordingly, the envisioned design environment will support process models that Contain the necessary detail for dynamic control of the individual operations as well as the information required to simulate the operation of the manufacturing system; Indicate the effects of perturbing the operational parameters, as well as the effect of complex interactions among processes; Indicate the effects of changing endogenous variables in a generic manner; and Are environment-independent and yet accommodate environmental specifics as model parameters. To produce such models, research is needed on process model representation schemata (both aggregate and detailed) and on-line data collection.12 Novel Design Considerations The envisioned design environment will support product design aspects that are not often associated with design at all, such as designing so that configuration control at the selling or servicing stage will be easy and error-free, build-to-order in lot sizes of one will be economical, or a robust final product can be made from parts obtained from different sources. Research is needed on tools and techniques to enable these new aspects.13 NON-MANUFACTURING-SPECIFIC RESEARCH Decision Aids A design decision has several impacts at once, some of which are beneficial and some of which are adverse. Therefore the envisioned design environment will support techniques to express comparisons and trade-offs vividly so that the designer can assess the impact of a wide range of design decisions on the product’s cost, time, or quality. To produce such techniques, research is needed on decision tools that draw upon the product-process data model and performance simulations of the product being designed, as well as process models and various data on costs and tolerances of different processes. These tools will sort data and models on the basis of criteria supplied by the designer to aid in making comparisons between alternate designs and processes.14
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs Geometric Reasoning A generic intellectual activity required by mechanical design is geometric reasoning. A major difference between VLSI design and mechanical design is the degree to which three-dimensional geometric reasoning is fundamental to even the simplest mechanical designs.15 Today, the design process relies on talented individuals to do this reasoning, and that may always happen. But the envisioned design environment will support improved visualization tools or other design aids that will help make this process faster and efficiently achievable by a broader range of people. Knowledge and Information Management Basic to design are many issues of data management and of data themselves. The envisioned design environment will include a number of data management methods and tools. First, at the simplest level, the design environment will include ways to capture corporate memory and knowledge so that successors of current designers can tell what knowledge was used, what competitive methods were used, what errors were made, and on what factors success was based. Second, the design environment will have data management techniques that enable corporate learning and consequent improvement of processes. Third, the design environment will have to handle data legacy issues, such as converting data from one computer-aided design (CAD) system to another and preserving old data for decades or more so that they can still be read, edited, and processed. Today, such data are either lost, kept on paper, or accessed in a limited way by old hardware kept on hand for the purpose.16 A fourth, deeper data management issue, one strongly affected by issues of scale-up, is that of recording design history and rationale. Every design is a historical web of decisions that grow out of each other and depend on each other. Revision, whether for correcting an error, absorbing a new outside circumstance, or improving manufacturability, requires unraveling the web to a certain degree. The design selected, the web of decisions leading to it, and “roads not taken” indicate the corporate state of belief at the time the decisions were made and thus form a historic context. The envisioned design environment will therefore provide for recording this context, possibly in the product data model, so that it can be compared with current states of belief as a result of accumulated analyses or new inputs. NOTES 1. Several cooperative attempts have been made to develop a standard product model. Of these, the product data exchange using STEP/standard for the exchange of product model data (PDES/STEP) program is intended to apply to the broadest assortment of products. Some view PDES/STEP as a design tool, others as a means of doing business that supports interoperability between enterprises. PDES/STEP has been criticized for not adequately supporting design, for not supporting assemblies of parts, and for ignoring modeling and communications standards. National efforts in the project remain important, however, to ensure a voice for U.S. industry in setting associated international standards. Furthermore, as a large, public, cooperative effort (employing 200 to 300 full-time-equivalent U.S. personnel), it presents opportunities for leverage. 2. The design of disk systems is illustrative. Disk systems consist of the physical disk medium (plates/heads) as well as a microcoded controller. The amount of microcode (software) is quite large—sometimes over 1 million lines. By far the greatest difficulty in bringing these products to market is getting the microcode done—there is so much of it, and it cannot be tested until the hardware is actually built. Better modeling and more efficient simulation systems or rapid prototyping of the hardware would allow earlier testing of the microcode and could cut by half the time to market.
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs 3. A single VLSI chip may have 3 million transistors with sub-micron feature sizes. A complex system may have hundreds of chips and 20 million transistors. Functional simulation of such a system may involve 7 billion test cases. Design data for such systems must be coordinated over many views and processes. Logic design, functional tests, fault tests, timing, placement, and wiring data run to hundreds of megabytes per chip. Coordinating these data requires sophisticated process and design control functions. For example, because some of these views are derived, a change to one view may invalidate other views. Thus an error detected during functional simulation, for instance, may cause a change to the logic design, which in turn invalidates derived fault, timing, and place and wire views of the logic. A change made to one view may also require re-execution of additional processes, such as design rule checking. 4. The committee has attempted to move beyond many general topics that have been suggested before and to focus on certain specific topics of great importance. These include understanding nominal and variant behavior and matching processes to design elements so that the specified maximum variation would be respected. Although they are among the major goals of “design for X,” they are not captured in the frequently suggested notion of building DFX into the design platform. 5. One such domain is application-specific integrated circuit (ASIC) design. In ASIC design, a logic designer specifies function in a high-level language, such as very high speed integrated circuit (VHSIC) hardware description language (VHDL). He or she also gives parameters such as maximum area, aspect ratio, allowable power dissipation, and timing constraints. Tools do logic design, physical design, test generation, and design rule checking. The designer also provides test cases for functional verification of the design. 6. See Riesenfeld, Richard F. 1992. “Integrated and Distributed Manufacturing,” presented at Information Technology and Manufacturing: A Workshop, National Science Foundation, May 5–6. This presentation recommended research on feature-based approaches to design and manufacture, and, in particular, development of a high-level, broadly applicable model that supports design, manufacture, and all associated processes for objects with both traditional and sculpted boundaries. 7. For example, a car door has two main functions, namely, to close out noise and weather and to keep the passenger in. It has many subfunctions, one of which is to resist a side impact from another car. So the outer skin of the door and the rubber seal (elements) keep out the weather and noise, while the inner and outer skin keep the passenger in (main function), and a bar inside the outer skin (subelement) resists side impacts (subfunction). On some cars, the door has an extra latch hook to fasten it more strongly to the car’s frame when it is closed, and this hook can take some of the load of a side impact. So in this case there are two subelements (hook and bar) that share the subfunction (resist side impact). Different decompositions of functions and subfunctions have different performance (better resistance to side impact for less weight) and different costs (extra hook means more cost in parts and assembly but perhaps a lighter bar, which is better for fuel economy). These choices are difficult to make, and few analytical methods and no computer tools exist to help the designers make them. Yet these crucial decisions are made at the beginning of every design, and they create the concepts from which everything else follows. If they are made poorly, the rest of the design will be too heavy, or too expensive, or not safe enough, or inadequate in some other way. 8. Manufacturing Studies Board, National Research Council. 1988. A Research Agenda for CIM. National Academy Press, Washington, D.C., pp. 17–19. It recommends research on product and process design, including data structures for describing products in terms of conceptual design, functional features, dimensions and tolerances, manufacturable features, and so forth, and methods that allow such structures to be interfaced with other computer-integrated manufacturing (CIM) components, such as knowledge-based systems. 9. “Feature-based” is a way of describing a design that contains more information than just geometry. In a computer model a circle may represent a hole, but it is not a hole. However, in the database there could be text data saying that it is a hole, giving the diameter and tolerances, plus numerical control instructions for how to drill it. The “hole” is then called a feature, which in fact is
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs a data object that contains instructions for how to draw itself on the computer screen and how to drive a machine tool to drill it. Features can describe machined areas, or they can describe areas where a measurement will be taken (“measurement features”) or where one part will be joined to another (“assembly features”). Features are thus able to hold a great deal more information than a purely geometric model can. In the future, one could expect features to contain design intent as well as details about how to make and use the feature. For example, a pocket feature to hold a precision ball bearing would require tighter tolerances and a finer surface finish than a hole through which oil is squirted. 10. Manufacturing Studies Board, National Research Council. 1991. Improving Engineering Design: Designing for Competitive Advantage. National Academy Press, Washington, D.C., pp. 56–57. It recommends research on tolerance analysis, tolerance representations, tolerance-performance relationships, and tolerance standards and measurement methods. 11. Monte Carlo (M-C) methods are currently used to study different tolerance assignments, but their utility is limited. First of all, M-C is computationally complex. Second, it can be difficult, though not impossible, to determine sensitivities using M-C, that is, to find out which dimensions contribute the most to the final total error. These two problems are called tolerance analysis. Third, the designer wants to find out not only what the error is but also how to distribute tolerance limits among several potential contributing sources during design so that the desired final error is achieved at the lowest cost. This last problem is called tolerance allocation. The ideal method would link a mathematical definition of the problem directly to solid computer models so that analytical and M-C methods could be used together during design to arrive at better solutions. Note, however, that too little is known at present about the relation between shape errors and product performance to permit rational tolerance assignments except in a few cases, such as optical systems or journal bearings. In most cases, tolerances are assigned on the basis of experience or available equipment. Research is needed to improve this situation. 12. Manufacturing Studies Board, National Research Council. 1988. A Research Agenda for CIM. National Academy Press, Washington, D.C., p. 20. It recommends research on methods for representing decision objects, e.g., states, actions, utilities, prior and posterior probabilities, samples, costs, and decisions; methods for mapping these to manufacturing objects; development of relevant heuristics and algorithms; and exploratory assessment of the merit of these techniques in specific domains. 13. Manufacturing Studies Board, National Research Council. 1991. Improving Engineering Design: Designing for Competitive Advantage. National Academy Press, Washington, D.C. pp. 59–60. It states that designing for manufacturing at the conceptual stages and designing for other objectives at almost any other stage are not supported by any well-developed techniques. It recommends research to relate the crucial features of a product’s early description to its ultimate life-cycle quality and cost in terms of each of the many design objectives. 14. Examples of pushing manufacturing models beyond geometry already exist in the area of electronic design automation. In gate-array or standard-cell-style designs, libraries of technology primitives are created that are, in effect, subassemblies of designs that can be used by designers and CAD tools. Each cell has not only a geometric view but also a functional view (e.g., a three-input NAND gate) for use in logic synthesis, a behavioral view for use in high-level simulation, a fault view for use in testing, and a timing view for use in timing and delay analyses. Product families are captured at several levels. Cell libraries may be designed to be consistent across technologies with different ground rules, allowing for design portability. Within libraries, there are families of same-function circuits. Cells in the same family perform the same logical function but differ in timing, power consumption, test characteristics, and geometry. The functional view showing timing also illustrates the concept of variation. Typically, the timing view includes the data needed for calculating nominal, best-case, and worst-case delay. 15. This explains why electronic computer-aided design/computer-aided engineering (CAD/CAE) appears further along than mechanical CAD/CAE: although it is nontrivial, two-dimensional analysis is easier than three-dimensional analysis.
OCR for page 11
Information Technology and Manufacturing: A Preliminary Report on Research Needs 16. Siewiorek, Daniel P. 1992. “Rapid Prototyping: The Design Process, Tools, and Fabrication,” presented at Information Technology and Manufacturing: A Workshop, National Science Foundation, May 5–6. It posed the following research questions that must be resolved before concurrent design and rapid prototyping become integrated into industrial practice: How can design and manufacturing information be reused in future products? and How can the compatibility of new incremental information with all the previously acquired information be ensured?