Jami Shah - US grants

An artificial intelligence (AI) workstation is purchased for use in the newly established Engineering Expert Systems Lab. The acquisition of such equipment will considerably enhance the research efforts of a large number of faculty involved in the development of expert systems, primarily in the areas of mechanical design and manufacturing. Development of such systems involves the use of symbolic languages, such as LISP and PROLOG, but the architecture of general purpose computers is not suitable for such languages. Additional hardware is needed for fast symbolic and logical processing, as well as additional local memory, to increase efficiency. AI workstations fulfill these needs and thus reduce considerably the development time for expert systems.

The long term objective is to develop integrated software systems for automation of design, engineering, and manufacturing tasks--a technology that is needed for the computer integrated factory of the future. This requires the development of product databases that contain the totality of information related to all aspects of the product. The proposed Expert Solid Modeling Shell will provide a high level environment for developing and documenting such a database. The shell concept and object-centered programming technique are borrowed from expert system technology to provide a system that can be customized to user needs. This will allow designers to define properties of features they wish to work with, to instantiate these from a library and to manipulate them as desired. "Intelligence" will be imparted by three sets of rules: Cognition, Inheritance, and Interpretation. Cognition rules will be used for detecting invalid operations; Inheritance rules for deriving feature parameter values from parents: Interpretation rules for decoding information required by an interface while building its own multi-layered database incorporating feature relationships and descriptions. Feature taxonomy codes will be used by a post- processing program to convert design features to manufacturing features so that product data can be used to drive automated process planning, group technology classification and manufacturability evaluation programs, etc.

This research investigates an advanced heterogeneous database environment for engineering design, with emphasis on using design knowledge and relating design decisions to design versions. The environment uses a blackboard approach to coordinate, integrate, and monitor product development. An object-oriented database provides an integrated, logical view of complex design entities that are stored in a variety of databases. Designers therefore have a global view of how heterogeneous data repositories form the complete product design. The blackboard also coordinates access to expert systems and deductive databases to provide engineering design knowledge. Design histories are captured as a series of design development steps and the corresponding engineering knowledge is used in each step and resulting decision. Both navigational and knowledge-based approaches to querying design histories are used. The impact of this research on database technology will be the advancement of heterogeneous databases through the use of object-oriented concepts to achieve a tightly-coupled view of design data. Object-oriented concepts will also be extended with a process model for capturing and querying design histories. This research will also lead to the improvement of the engineering design productivity by providing an integrated view of design knowledge and data and by supporting the ability to query design histories allowing re-use of previous designs.

9411522 Henderson ABSTRACT Holistic Approach for Preparing Students to Learn and Lead in a New Manufacturing Paradigm Arizona State University has team with Motorola, Intel, Honeywell and others to infuse manufacturing education across the engineering curriculum at Arizona State -- from the freshman year through the master's level. This project will focus on leadership training, total quality principles, recruiting displaced defense workers and underrepresented groups, and enhancing the curriculum. The long term goal is to produce a new breed of engineer for the 21st century, one equipped with the hard and soft skills necessary to change manufacturing. ***

9522971 Shah This project investigates several computational techniques for transferring product geometry data from design to manufacturing without human intervention. The project is divided into two parts. The first component of the project is the development of machining algebra for geometric representation of tool-workpiece interaction for all machining processes. This provides the mathematical basis for determining the machining process that produces each machining feature on a given part. The second component of the project is a computational model for mapping dimensions and tolerances (D&T) while preserving the design intent. The key elements of the model are geometric building blocks, directed geometric constraints, and degree of freedom analysis for validation. Methods for redistributing the designers D&T between machining features will also be developed. This project addresses a critical area of industry need, viz., design and manufacturing integration. The lack of this integration results in longer development times because manufacturing planning today is a very labor-intensive activity. The machining algebra has the potential to capture the fundamental characteristics of common machining processes, replacing shallow heuristic rules used now, enhancing the degree to which manufacturing can be automated. The methods proposed are generic; they are independent of particular computer aided design (CAD) or computer aided process planning (CAPP) systems used. Each of the computational methods can be used independent of each other, making it attractive to incorporate them into existing CAD/CAPP systems. This will enhance the communication between design and manufacturing and enhance process planning productivity, thus reducing time to market.

This grant will provide support for developing theoretical foundations for describing and solving gometric and topological problems in mechanical product design and manufacture. A theory will be developed to encompass a taxonomy and morphology of generic classes of geometric and topological problems, their characterization, and their respective solvability conditions. The theory will provide an understanding of the scope and limitations of generic solution techniques. The unified characterization will contain representation of topological and geometric entities at multiple abstraction levels, a standard set of topological and geometric relations, and a minimum set of generic design tasks or issues. High level problem definitions will be encoded in the form of generic geometry exemplars that will contain not only geometry templates, but also mechanisms to specify the design task or goal. Methods for mapping high level problems to low level standard geometric tasks will also be investigated.
If successful, this research will lead to a uniform treatment of geometric problems in diverse areas and will enhance integration between different parts of the product design process. In the long run, the results of this project have the potential to extend the scope of CAD/CAM in automating design and manufacturing tasks both upstream (preliminary design) and downstream (manufacturing planning). This research will also benefit the development of standards for exchange of parametric geometry data between different computer aided design systems used throughout the product development process, thus reducing development cost and the time to market.

This project will formulate a set of guidelines for conducting empirical studies to evaluate formal methods for idea generation in collaborative environments. In order to conduct good experiments on idea generation methods, a clear understanding of experiment variables, method variables, and data collection variables is needed. Also, a framework to analyze the data collected from such experiments is required. The research is divided into four major tasks: survey and evaluation of creative cognition models; identification of operation variables, metrics, and experiment design; development of a framework for analysis of design sketches; analysis and critique of previous experiments on progressive idea generation (C-Sketch, Method 635, and Gallery Method). This project will investigate the issue of scaling up appropriate cognitive models to study typical design problems which are far more complex than those used in psychological experiments. This study will attempt to understand the correlation between four sets of factors relevant to collaborative design: the individuals, the situation/environment, the outcome, and the process. This research will also develop a methodology for analyzing design sketches for studying the cognitive processes, mental structures, and ideas generated by tracking the evolution of design concept at successive stages of development. Studies have shown that the early stages of conceptual design have a substantial impact on the efficacy of a product; development of effective design methods to support the generation of alternative design ideas is critical. Along with that, designers must be convinced that the use of these methods has potential benefits. Therefore, it is essential to collect and analyze data on the effectiveness of idea generation techniques. The availability of experiment methodology and metrics for evaluating and/or optimizing creative design methods is an important step in this direction. If successful, future research will employ these guidelines to produce verifiable database of performance of formal creativity methods which may eventually lead to wider use of effective idea generation methods to explore alternative designs.

The objective of this project is to develop a new bi-level math model for geometric tolerances, that is, the range of variation in size and shape that can be permitted for a part to be acceptable when in use. At the local level, the new model would represent each tolerance zone for a plane or an axis (line) as a multidimensional region or map, and it would include the computational techniques that relate interdependencies between these regions and subregions within them. The mathematical model to be developed will include formulations for variations in size, form, orientation, position, and runout, and combinations of these, which can be modeled using points, lines, and planes. The primary method for the research is to use traditional mathematics for representing planes, lines, and points in a tolerance zone. At the global level, the proposed model interrelates all frames of reference on a part or an assembly. It will include a symbolic algebra that relates the degrees of freedom and constraints for a feature, such as a hole, to reference frames used when specifying dimensions and tolerances. In order to validate the integrated bi-level mathematical model and to demonstrate its feasibility and usefulness, a set of prototype software modules will be designed and implemented.
If successful, the results of this research will impact the state-of-the-art for geometric tolerances and the productivity of designers who use the results. At the local level, the model will treat dimensional and geometric tolerances in three dimensions and in a manner consistent with the standard on tolerances that is universally accepted by practitioners, whereas existing models for tolerances are only partially satisfactory in being consistent with the standard. At the global level it is projected that the new model will provide a unique bridge between mathematical characterizations of features (planes, lines, and points) and existing parametric computer-aided design models and statistical tolerance software packages. It is projected that successful completion of the project will provide the designer with a tool for making better decisions about allocating geometric tolerances. The completed model also should permit designers to identify trade-offs and to optimize the allocation of tolerances. The impact should be a shorter design time, fewer iterations in the prototype phase of design, and hence lower cost.

In this joint project between Arizona State University and University of California at Berkeley, a set of metrics and methods will be developed to evaluate and optimize design for manufacturing through all phases of product development. The first phase of the project will involve investigation of Design-for-Manufacturing (DfM) metrics and methods using several criteria, such as theoretical soundness and computational complexity. A novel DfM approach based on Benefit/Cost analysis will be developed to consider trade-offs between the quality of a design and its manufacturing cost, while taking into account market factors and a Company's strategic goals. Key performance or utility factors of each design alternative or competing product will be used for computing a Design Benefit Rating, and similarly relevant manufacturing factors will be used for computing a Manufacturability Rating. Then techniques of value engineering will be used to explore design changes to improve the overall value of the design and/or to choose between alternatives. In the final year of the project, a DfM Shell will be implemented to serve as a domain independent testbed to verify, refine, and compare DN methodologies. Domain-specific design and manufacturing agents will be linked to the testbed. The particular focus will be on three processes: CNC machining, injection molding, and fuse deposition. Six industrial partners will provide the part designs and process knowledge for several manufacturability case studies.

The primary benefits from the project are: (1) tools for manufacturing companies to improve rapidly the quality of their products to meet changing market conditions; (2) sound theoretical foundations for manufacturability analysis; and (3) a framework to evaluate the reliability of current and future commercial software for DfM.

The objective of this research project is to produce the knowledge needed to evaluate the effectiveness of engineering design idea generation methods and to distinguish between their necessary and superfluous components. This research will combine the highly controlled lab experiments and atomic process models from cognitive science with design experiments from design research. Instead of conducting experiments by using an idea generation method in its entirety, each method will be decomposed into its key "ideation" components and its overall effectiveness predicted by testing the effectiveness of its components and their mutual interactions. Lab experiments will be directed at studying key ideation components individually, while the design experiments will examine interaction between components and their relation to outcome by mutating their constituent components. Explanations for why and how ideation components are effective will be researched from accepted models of atomic processes and structures already available in cognitive science. Effectiveness will be evaluated in terms of four measures: fluency, quality, variety, and novelty of design concepts generated. The results of the study will be consolidated into a model of design ideation by combining and modifying several elements from existing models. The proposed ideation model will be validated through final experiments in industry settings.

Two important results should emerge from this research: (1) data from the empirical studies that can be interpreted in terms of both theory and applied practices, and (2) a system for pursuing additional research questions about ideational aspects of conceptual design theory. An understanding of the relationship of ideation processes to design outcome will help companies determine which method to use under given conditions and how to constitute design teams. This work will also help educators in finding better ways of teaching design synthesis.

Tolerances specify the ranges of permissible imperfections so a manufactured part will be acceptable for assembly and use. Standards exist to define classes of tolerances and to ensure proper communication among engineers. The purpose of this research is to continue creating a mathematical model of geometric tolerances. The results from our earlier work show that the proposed math model is compatible with the standards, and that it provides three-dimensional relations for assemblies. At the local level, the method of the new model is to represent each tolerance-zone for a plane or a line as a hypothetical point-space, or map, and to combine these for assemblies of parts. At the global level, the model inter-relates all reference frames of an assembly using degrees of freedom. The project includes the implementation of a tolerance analysis system for comparison of the new model with existing software.

Existing methods for analysing tolerances today are based strongly on ad-hoc conventions from engineering practice and less on mathematical principles. Consequently, full three-dimensional analysis of tolerances in assemblies is not done today, and contemporary design software is only partially compatible with existing standards. The outcome of this project will provide the tools to complete full three-dimensional statistical analyses of tolerances, and, thereby, it will improve quality and lower cost. This project will result in new teaching materials, which should make the subject easier to include in undergraduate and graduate engineering curricula. Based on our model, we are proposing to the American Standards Committee a new set of professional development classes and certification exams that are directed at engineers for tolerance analysis in the digital age.

This award provides funding for the conduct of the annual National Science Foundation Design, Service and Manufacturing Research Conference. This conference brings together researchers in all program areas within the Division of Design, Manufacture and Industrial Innovation (DMII), and researchers from related programs in the Divisions of Engineering Education and Centers (EEC), Electrical and Communication Systems (ECS), Civil and Mechanical Structures (CMS), Information and Intelligent Systems (IIS), Materials Research (DMR), International Programs (INT), and Social, Behavioral, and Economic Research (SBE), specifically the Innovation and Organizational Change Program (IOC). Grantees will present their research in poster sessions and attend sessions covering topics that extend beyond their immediate area of research. Ongoing research in manufacturing related research activities are presented by the grantees from the Consejo Nacional de Ciencia y Tecnologia (CONACyT), Mexico, and grantees from the Natural Sciences and Engineering Research Council (NSERC) and the National Research Council (NRC), Canada. The conference seeks to ensure that the individual researchers supported by NSF are informed about the ongoing activities of their colleagues, eliminate duplication of efforts and foster a higher degree of cooperation among researchers. In addition, the conference allows all Principal Investigators to meet each other and to meet with their respective NSF Program Officers.

The goals of the conference are: (1) to broaden the outlook of all participants; (2) promote transfer of ideas and technology from one area of research to another; (3) to allow those working in a related area of research to get to know their peers so as to avoid duplication of research and to encourage cooperation, and (4) to allow personal contacts from the collective research community, the NSF, CONACyT, NSERC, and NRC program staff to discuss manufacturing research in detail, clarifying issues and problems in ongoing projects.

The last two decades have seen increased globalization of commerce and flight of engineering, manufacturing, and IT jobs from the U.S. It is more critical now than ever to innovate technologies, products and services that provide new fuel to the engines of our national economy. A workshop is planned to determine how to best meet industry and educational needs in engineering design through research in this changing environment.
This award provides support for a workshop to be held March 26-29, 2004 in North/Central Arizona (Phoenix/Scottsdale locale). The objectives of the workshop are a) to review the past 25 years activities in research in design theory and its impact on product design and engineering education to date, b) to forecast technological developments and changes in social, business, and environmental factors that may have profound effects on future product development, and c) to formulate a strategic plan for engineering design research based upon lessons learned and expectations of the globalization of product development. The 2-1/2 day workshop will be organized by a Steering Committee that represents the broad engineering design community focused on product realization. Potential participants will submit position papers and 40-45 invitees will be selected based upon their potential to contribute to one or more of the objectives of the workshop. This is appropriate to assure that the many different perspectives and threads of research are represented as the future directions for engineering design research and education are discussed.

The results of the workshop will be published on a website and on a CD for public dissemination. The previous workshop, held a decade ago, produced a substantive report that provided key recommendations to NSF and the community at that time. It is expected that similar outcomes will be realized from this workshop. This time, however, the community is being challenged to not only benchmark where we are now and immediate future needs, but to look out to the 2030 horizon with vision.

The research objective of this award is to gain a formal understanding of design skills that are critical in early phases of engineering design. The major components of this project are: (1) Enumeration of design skills and their association with specific conceptual design tasks; (2) Characterization of these skills in terms of measurable indicators that are theoretically and empirically verifiable; (3) Study of relationships between sub-factors of different indicators of each skill; (4) Investigation of skill measurement instruments; (5) Evaluation and refinement of measurement instruments using criterion and construct validity checks; (6) Characterization of designer profiles and team profiles in terms of skill sets. This research will connect design research to established cognitive theories of human problem solving and learning, visual and spatial reasoning, pattern recognition and scientific discovery. It seeks to gain insights into how design knowledge is used and what differentiates good designers from ones less skilled.

If successful, the results of this research will improve design education and collaborative team based design. Design exercises and courses could target specific skills or sub-sets of those skills, and the effectiveness of these exercises or courses could be evaluated in an objective manner by tracking changes in designer profiles. Skill based grading would be a radical departure from current grading practices, identifying strengths and weaknesses of individuals explicitly. For forming collaborative design teams in industry, complementary designer profiles could be matched to ensure that the team as a whole has all of the skills deemed essential for the of project that it is to undertake. Since the human designer is the most important element in product development, it is important to gain a fundamental understanding of skill sets possessed by good designers.

The research objective of this award is to develop mathematical models for manufacturing variations. In these models, dimensional and geometric variations on manufacturing features will be mapped to points in convex bounded regions in higher dimensional space. When features or tolerances stack-up in an assembly, error accumulation corresponds to Minkowski sums of all contributing regions (Tolerance Maps), which then represent the entire population of assemblies conforming to the given design specification. This research will create a method for using these maps for manufacturing operations based worst-case calculation in process planning by fitting maps corresponding to manufacturing specification inside maps corresponding to design. Statistical analysis of tolerance accumulation for any combination of frequency distribution corresponding to characteristics of manufacturing machines producing the corresponding features will also be investigated. Methods for direct fitting of inspection data from coordinate measuring machines into these and a methodology for using them in statistical process control maps will also be developed.

If successful, the results of this research will enable precise mathematical interpretation of the tolerance standards, consequently preventing errors in manufacturing planning and inspection. Maps use in statistical process control will provide an effective visualization and diagnostic tool, based not just on proximity to worst case limits, but also on full statistical analysis. There is potential for this new method to streamline the analysis of Coordinate Measuring Machine data to where it can be an integral part of the feedback for statistical process control. This work will also generate material suitable for teaching the theory of tolerancing and metrology. It is envisioned that these new math models will enable future research on correlations between parameters of a manufacturing process, such as tool wear, and the shape and size of the measured subset, together with its location within the map.

The research objective of this Early-Concept Grant for Exploratory Research(EAGER)award is to determine the feasibility of combining elements of intuitive ideation methods with logical/experiential methods to investigate the creation of a holistic framework that will contain ideation strategies from both types. Engineering creativity lies at the intersection of originality and functional quality. A large number of structured ideation methods and tools have surfaced for enhancing engineering design creativity. They can be classified into intuitive and logical/experiential approaches, both of which have deficiencies. Intuitive and logical/experiential approaches represent opposite extremes. Given the diversity of design problems and designers, it is unlikely that the exclusive use of a single method or tool, whether intuitive, logical or experiential, would always lead to creative solutions. In this EAGER award, the feasibility of combining elements of these two approaches - intuitive ideation with logical/experiential - in order to overcome limitations of each and to expand the potentials for ideation will be explored. The foundations will be laid for a prototype Testbed for design and for creativity research. Preliminary investigations will be performed to determine if selected parts extracted from them can be combined with intuitive strategies. The experiential strategies to be included during the exploratory stage are: decomposition, abstraction, analogical reasoning, example exposure and conflict resolution. The intuitive strategies included will be: suspended judgment, incubation, provocative stimuli, action verbs and combinatorial play.

If successful, this EAGER award will be a preliminary step towards Testbeds that will provide a uniform and structured way of collecting data on creative design. This will lead to a better understanding of creative processes and strategies through empirical studies. By dissecting ideation methods, rather than using them in their entirety for empirical studies, the investigators will research at a finer granularity level (mini-strategies) for greater insights into design creativity. Studying the combination of intuitive and logical/experiential ideation in terms of how problems are posed in each and the ideation stage, problem type, designer type is unique and challenging. Success of this effort will lead to more extensive research on this topic. The impact could be extremely significant, given the importance of ideation to the engineering design process.

The research objective of this CreativeIT award is to study the major blocks to creativity that arise from the manner in which design problems are formulated. It will investigate why some engineering designers generate creative solutions while others converge on mundane ones, specifically examining the critical role played by formulation and representation in early conceptual design. A central idea of this research is that one can usefully cast problem formulation as a process of answering questions that reveal key characteristics of the design task. These include questions about design goals and objectives, relative priority of technical issues, relations among design parameters, conflicts and gaps in the problem statement, and fictitious constraints. This investigation will result in a dynamic representation, the Problem Map, which aims to characterize a designer's understanding of a problem at a given time. An interactive computational aid that lets users construct a Problem Map for a given design problem will be developed. The project differs from earlier efforts in that its focus is on problem formulation (that is pre-ideation), not just idea generation, and its emphasis on an interactive system that keeps humans in the loop rather than replacing them with automated systems.

If successful, this research will advance understanding of creativity in engineering design, and should also generalize to other design tasks in science and engineering. Problem definition is perhaps the most critical part of design, since it determines the boundaries and topology of the search space and hence determines the originality and quality of designs that can be generated. Hence, this research is critical and timely, with potentially broad impacts. The interactive system should also serve as a useful educational tool that increases creativity in novices. The tool will demonstrate the power of Information Technology is aiding human creativity. The tool will be used in design courses at Arizona State, and will be distributed widely to other universities over the Internet. If successful, this system and the principles that underlie it will change the way that apprentice designers approach their tasks and, in the longer term, lead to more creative engineering practices. Thus, our work will also lead to innovative educational approaches in engineering design that reward creativity. The results will be disseminated through normal academic channels that involve publishing papers and presenting talks at conferences.

The objective of this research is to integrate the three sequential stages that typically are required to get complex mechanical assemblies to function and assemble properly. The first of the three stages is the setting of limits on the geometric variations by design engineers. The second is conversion of these limits by manufacturing engineers to process plans, the step-by-step procedures used in a factory to accomplish the manufacture of a part or an assembly. The third is the conversion of measured points on the surface of a part during quality control inspection to a form that may be used for direct comparison with the limits set by the designer. The method is to extend the concepts of a newly developed mathematical model for design both to complex geometries and to the domains of manufacturing and inspection of mechanical parts. The new model represents the design limits as boundaries in multidimensional geometric space. The boundaries separate smaller variations, which are acceptable, from larger ones that are not. The new math model is unique because no other model represents all classes of limits used in the practice of design and manufacturing.

The benefits from this award, if successfully carried out, will be a first mathematical model to integrate manufacturing variations in mechanical parts in a manner that will be consistent with the standards used in practice. Further, until now, all geometric manufacturing variations have been attributed to manufacturing processes, not to the geometry of the feature and the specified limits to the variations. The proposed use of the new mathematical model for design will enable a new way of thinking by separating variations that are purely geometric from those that come from characteristics of manufacturing processes and machines.

The research objective of this EArly concept Grant for Exploratory Research (EAGER) is to create a framework for a web based holistic ideation system that invites users at large to attempt to solve their design problems using the tool. The goal is to collect massive amounts of data from diverse set of users over an extended period of time to explore deeper issues and solution strategies employed by real designers. As part of an ongoing project the investigators have created a framework for holistic ideation by mixing and matching disparate ideation mechanisms matched to diverse ideation states commonly encountered in design. This is transformative in two ways: it builds a bridge between intuitive and logical approaches and it creates an entirely new mode of research for design creativity. But for this approach to bear fruit, massive amounts of data need to collected and analyzed to determine ideations paths, use and user scenarios and corresponding level of effectiveness. The Testbed will provide a uniform and structured way of collecting data on creative design which will lead to a better understanding of creative processes and the influence of thinking styles and problem complexity on these processes through empirical studies conducted at a fine granularity level (ideation mechanisms). A holistic approach will introduce students to both logical/experiential and intuitive strategies which has a greater probability of acceptance. Studying ideation paths in conceptual design has not been previously explored and researched in a formal manner. The main impediment is the lack of a proper foundation for characterizing such paths. The holistic ideation framework will facilitate several different ideation methods/strategies and track success rates based on ideation paths using heat maps employed today by companies such as Google.

If successful, this research has the potential for impacting design practice, design research and engineering education. This project explores the feasibility of new software to support people in being more creative and evaluates their performance through user studies. Such systems will become partners rather than tools. Through user interaction, they will sense the ideation state and suggest possible strategies at a given state. In this role the tool becomes a virtual facilitator with virtual group members that generate provocative stimuli and contribute technical insights. Particularly for novices, such as engineering students, it is important to provide access to experiential knowledge through a multitude of resources. Results will be broadly disseminated through conferences and journals, as well as through the developed website.