Answer to the Question for the Semester - Designing Open Engineering System | ME 6102, Papers of Mechanical Engineering

Download Answer to the Question for the Semester - Designing Open Engineering System | ME 6102 and more Papers Mechanical Engineering in PDF only on Docsity! ME6102 – Designing Open Engineering Systems: Answer to the Question for the Semester Submitted by: Nathan Rolander Prepared for: Dr. Farrokh Mistree & Jitesh H. Panchal Georgia Institute of Technology Thursday, April 22, 2004 Nathan Rolander ME6102 - Question for the Semester Spring 2004 I. ABSTRACT In this document, I will answer the Question for the Semester posed by professor Farrokh Mistree as part of the final submissions for ME6102. I will parse and personalize the question, and give the context in which it is applied, namely the work of 2020. I will then conduct a survey of the base methods that I will build upon, and critically evaluate them, deciding what needs to be improved for distributed design and development of Open Engineering Systems in the world of 2020. I will then perform a gap analysis, determining the requirements for augmenting the base methods in order to achieve Open Engineering System realization in 2020. I will next address how I will meet these requirements, leveraging and adding value to the work of others as well as through my own original ideas. I will finally validate and verify my method, discuss its utility, and present my learning through answering the Question for the Semester. Page 2 Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 5 8.1.3. Global Physical Region .................................................................................. 56 8.1.4. Political........................................................................................................... 56 8.1.5. Ethics............................................................................................................... 56 8.1.6. Storage ............................................................................................................ 56 8.2. CORE REQUIREMENTS FOR OES DESIGN & REALIZATION...................................... 56 8.2.1. Integrated Design of Products & Design Process.......................................... 59 8.2.2. Knowledge Storage & Re-use ......................................................................... 59 8.2.3. Hierarchical Design Process Organization.................................................... 60 8.2.4. Digital Design Support System ....................................................................... 60 8.2.5. De-centralization ............................................................................................ 60 8.2.6. Decision Support............................................................................................. 60 8.2.7. Parametric & Tolerance Design..................................................................... 61 8.2.8. Product Platform & Family Development...................................................... 61 8.3. SUPPORTIVE REQUIREMENTS LIST .......................................................................... 61 9. FRAME OF REFERENCE........................................................................................ 64 9.1. INCREASING CUSTOMER SATISFACTION.................................................................. 64 9.2. THE NEED FOR SHORTER PRODUCT LIFE CYCLES................................................... 64 9.3. INCREASING PRODUCT QUALITY............................................................................. 65 9.4. ADVANTAGES OF MODULARITY OF DESIGN & MANUFACTURE .............................. 66 9.5. ADVANTAGES OF PRODUCT PLATFORMS ................................................................ 67 10. EXTERNAL AUGMENTATIONS ......................................................................... 68 10.1. FUTURE TECH OPERATION.................................................................................... 68 10.1.1. The Engineering Spectrum............................................................................ 68 10.1.2. The Role of the Systems Engineer................................................................. 69 10.1.3. The Role of the Strategic Engineer ............................................................... 69 10.1.4. Engineering vs. Science ................................................................................ 70 10.1.5. The Bungie Software Corporate Approach................................................... 71 10.2. TECHNOLOGIES AND APPLICATION ....................................................................... 72 10.2.1. Radio Frequency Identification (RFID)........................................................ 72 10.2.2. Rapid Manufacturing .................................................................................... 73 10.2.3. World Area Network ..................................................................................... 74 10.3. THE ROLE OF SIMULATION IN DESIGN .................................................................. 74 10.3.1. Simulation as a Tool ..................................................................................... 74 10.3.2. Applicability of Robust Design to Simulation............................................... 75 10.3.3. Applicability of Information Re-use to Simulation ....................................... 76 10.3.4. The Need For a Definition of Clear Goals ................................................... 76 10.3.5. The Role of the Corporation in Development Efforts ................................... 77 10.4. ETHICS.................................................................................................................. 78 10.4.1. The Role of Ethics in Decision Based Design............................................... 78 10.4.2. The ASME Cannons ...................................................................................... 78 10.5. COMPLEX SYSTEMS .............................................................................................. 78 10.5.1. Defining Complexity in the Context of Simulation & Modeling ................... 79 10.5.2. System Integration Complexity ..................................................................... 79 10.5.3. Dealing with Complexity in Modeling & Simulation.................................... 82 Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 6 11. DIGITAL DESIGN SUPPORT SYSTEM.............................................................. 83 11.1. DISTRIBUTED OPERATION..................................................................................... 83 11.2. THE HUMAN-COMPUTER INTERFACE.................................................................... 83 11.3. SOFTWARE INFORMATION EXCHANGE................................................................... 84 11.4. HARDWARE COMMUNICATION .............................................................................. 86 11.5. INFORMATION STORAGE ....................................................................................... 86 11.6. PROGRAMMING WRAPPERS................................................................................... 86 11.7. MULTI-PHYSISCS................................................................................................... 87 11.8. CLASSIFICATION OF KNOWLEDGE......................................................................... 88 11.8.1. Templates in context of knowledge representation....................................... 88 11.9. DESIGN PROCESS LAYOUT .................................................................................... 92 11.9.1. Central Project Base..................................................................................... 92 11.9.2. Automated Bookkeeping................................................................................ 92 11.9.3. Automatic Version Control ........................................................................... 92 11.9.4. Communication............................................................................................. 93 11.9.5. Digital Fingerprint........................................................................................ 93 12. CONCEPT OF MY PROPOSED METHOD......................................................... 94 12.1. LEVELS OF MODULARITY...................................................................................... 94 12.1.1. Original Pahl and Beitz ................................................................................ 94 12.1.2. Updated Pahl and Beitz ................................................................................ 95 12.1.3. My ME6101 Augmented Pahl and Beitz ....................................................... 95 12.2. MY OES REALIZATION METHOD.......................................................................... 96 13. METHOD FLOWCHART....................................................................................... 97 14. CORE AUGMENTATIONS.................................................................................... 99 14.1. INTEGRATED OF DESIGN OF PROCESS & PRODUCT................................................. 99 14.1.1. Modularity..................................................................................................... 99 14.1.2. Mutability.................................................................................................... 104 14.1.3. Robustness................................................................................................... 105 14.2. HIERARCHICAL DESIGN PROCESS ORGANIZATION.............................................. 106 14.2.1. Hierarchy from Heterarchy ........................................................................ 106 14.2.2. Small Problems Approach .......................................................................... 106 14.2.3. Matrix Branch & Bound Approach............................................................. 106 14.2.4. Constructal Theory approach ..................................................................... 108 14.2.5. Phase Designation ...................................................................................... 108 14.2.6. Product Utility ............................................................................................ 109 14.3. KNOWLEDGE STORAGE & RE-USE ...................................................................... 109 14.3.1. Information Over Re-use............................................................................. 110 15. AUGMENTATION MODULES & TOOLS ........................................................ 111 15.1. AUGMENTATION MODULES ................................................................................ 111 15.2. AUGMENTATION TOOLS...................................................................................... 111 16. VERIFICATION & VALIDATION ..................................................................... 113 16.1. THE VALIDATION SQUARE.................................................................................. 113 Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 7 16.1.2. Theoretical Structural Validity ................................................................... 114 16.1.3. Empirical Structural Validity...................................................................... 115 16.1.4. Empirical Performance Validity ................................................................. 115 16.1.5. Theoretical Performance Validity............................................................... 116 17. UTILITY & CRITICAL EVALUATION OF METHOD .................................. 117 17.1. UTILITY .............................................................................................................. 117 17.2. CRITICAL EVALUATION ...................................................................................... 119 18. LEARNING & VALUE.......................................................................................... 121 18.1. A0 GOALS .......................................................................................................... 121 18.2. ADDITIONAL LESSONS ........................................................................................ 123 18.2.1. Leveraging .................................................................................................. 123 18.2.2. Extrapolation & Forecasting...................................................................... 123 18.3. FUTURE............................................................................................................... 123 18.4. SELF GRADING.................................................................................................... 124 19. REFERENCES........................................................................................................ 125 II.(A) LIST OF FIGURES FIGURE 4.1 - VISION OF 2020 SPOKE & WHEEL DIAGRAM................................................ 30 FIGURE 5.1 - THE BASE PAHL AND BEITZ METHOD FLOWCHART........................................ 37 FIGURE 9.1 - PRODUCT LIFE CYCLE GRAPH OF PHILLIPS TECHNICAL PRODUCTS ................ 65 FIGURE 10.1 - DIAGRAM OF A WEAKLY INTEGRATED SYSTEM ........................................... 80 FIGURE 10.2 - DIAGRAM OF A HIGHLY INTEGRATED SYSTEM............................................. 81 FIGURE 11.1 - DESIGN DSP TEMPLATE EXAMPLE.............................................................. 89 FIGURE 11.2 - MANUFACTURE DSP TEMPLATE EXAMPLE ................................................. 90 FIGURE 11.3 - MAINTENANCE DSP TEMPLATE EXAMPLE .................................................. 91 FIGURE 12.1 - INTEGRAL SYSTEM ...................................................................................... 94 FIGURE 12.2 - MODULAR SLOT SYSTEM............................................................................. 95 FIGURE 12.3 - MODULAR BUS SYSTEM .............................................................................. 95 FIGURE 12.4 - MODULAR SECTIONAL SYSTEM ................................................................... 96 FIGURE 13.1 - FLOWCHART FOR PROPOSED OES REALIZATION METHOD........................... 97 FIGURE 14.1 - LABVIEW FUNCTION PLACEMENT ........................................................... 100 FIGURE 14.2 - LABVIEW TEMPLATE EDITING ................................................................. 100 FIGURE 14.3 - THE AUGMENTED DSP PALETTE ............................................................... 101 FIGURE 14.4 - LABVIEW BLOCK DIAGRAM INPUTS......................................................... 103 FIGURE 14.5 - MODULE HIERARCHY................................................................................ 103 FIGURE 14.6 - LABVIEW SUB FUNCTION EDITING........................................................... 104 FIGURE 14.7 - PARAMETER & TASK RELATIONSHIP MATRIX............................................ 107 FIGURE 14.8 - THREE INDEPENDENT GROUPS OF TASKS AND PARAMETERS...................... 107 FIGURE 14.9 - PARTITIONED PARAMETER & TASK MATRIX.............................................. 107 FIGURE 14.10 - COLOR CODING AND STRUCTURE OF LABVIEW ..................................... 109 FIGURE 16.1 - THE VALIDATION SQUARE......................................................................... 113 Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 10 1.3. PARTICULARIZING THE QUESTION FOR THE SEMESTER In order to personalize ME6102 in order to obtain the maximum learning value I have personalized the Question for the Semester. Through this process I have made my final submission better address my personal areas of interest, research work, and A0 goals. My reasoning for this area of focus is presented through a critical evaluation of the base methods that I have leveraged from, which is presented in section 7. The original Question for the Semester is stated below for comparison: We imagine a future in which geographically distributed engineers collaboratively develop, build, and test solutions to design-manufacture problems encountered in the product realization process. We recognize that solutions evolve over time. Accordingly, we expect you to build on what has been done before. In this context, we want you to provide a method to support the realization of mass customized industrial products for a global marketplace through distributed design and manufacture. The proposed solution to the Question for the Semester presented in ME6102 is the development of Open Engineering Systems. However I believe that in order to create open engineering systems, I need not only a method that can be used to design open engineering systems, but have a method that is in itself an open system. This is important because implementation of new technology into a product/process/service that is being developed cannot be implemented unless the design and realization method being used is capable of effectively implementing this new technology in the design and realization process to create a superior solution. Further relevant information on Open Engineering Systems is presented in section 2. In this manner I am personalizing the Question for the Semester as follows: We imagine a future in which geographically distributed engineers collaboratively develop, build, and test solutions to design-manufacture problems encountered in the product realization process. We recognize that solutions evolve over time. Accordingly, we expect you to build on what has been done before. In this context, we want you to provide a computer supported method, enabling knowledge storage and re-use, to support integrated realization of design processes and products for a global marketplace through distributed design and manufacture. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 11 The changes I have made to the original Q4S are in the same manner as I generalized my ME6101 Q4S. I have shifted its focus to address the integrated realization of design processes and products through the use of a computer supported method, allowing for knowledge storage and re use. I have specified the key areas that I intend to address through answering this question as follows: computer supported – In order to implement a distributed design methodology effectively a support system must also be implemented. This computer system itself must also be distributed, yet allow for efficient communication and collaboration between distributed team members. enabling knowledge storage and re-use – In the computer science domain, the general motto is to never write the same algorithm twice. This approach can also be applied to engineering design, where the storage of design methods as building blocks allowing for re-use would dramatically increase development speed. support integrated realization of design processes and products – The breakdown of the design process into individual modules would add great flexibility, allowing for the construction of any design process. This had even greater value when concurrently designing both the product and the process, using the information generated at each stage in both areas to proceed with the next stage of the design process. The primary use of this modularity and flexibility is the effective application of design knowledge to new domains. I have particularized the Question for the Semester in order to provide a sense of direction for my proposed method. This particularized Question for the semester serves to act as a two line summary of what I intend to accomplish with my answer that is embodied within this document. 1.4. APPROACH TO ANSWERING THE QUESTION FOR THE SEMESTER My approach to this problem is one of leveraging, extensions and integration. I feel that I designed a very strong answer to the Question for the Semester in ME6101 and hence am using it as one of my base methods upon which to build. The strength of my ME6101 augmented Pahl and Beitz method was its flexibility and generality. However, as I explain further in the critical analysis of the base method, this needs substantial changing to support the development on open engineering systems. To this end I am leveraging from many sources from within ME6102 as well as from my colleagues in the Systems Realization Laboratory. I feel that the integration of these methods in conjunction with my own original work will give me a strong answer to the Q4S. 1.5. STRUCTURE TO THE ANSWER TO THE QUESTION FOR THE SEMESTER The format of this answer to the Question for the Semester shares aspects with the Pahl & Beitz systematic design method. I first clarify the task by parsing the Question for the Semester. I then set the context in which my solution fits, defining my world of 2020. I Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 12 next determine the needs of my proposed method forming my requirements lists. I then explore the base methods to be build upon, and identify their shortcomings, and why these must be addressed in the context of my world of 2020. I finally address how I will attend to these needs though my proposed method, and determine its effectiveness. However, there is a difference between how I have structure my answer to the Question for the Semester and the structure suggested by professor Mistree. I have addressed the base methods and critically evaluated them before developing my requirements list. This is because I feel it is more logical to identify the crux of the problem (in this case the shortcomings of my base methods used), and then detail out the requirements to fill based on these key issues. Therefore, the structure to my answer to the Question for the Semester is as follows: 1. My vision of the distributed environment of 2020, both general and relevant to my augmented Pahl and Beitz method. 2. The base methods that I will integrate and build upon including critical evaluations of these methods. 3. The requirements for product realization in this distributed environment, in regards to my proposed method and external augmentations to support it. 4. My proposed method, its overall structure and individual module details. 5. Verification, validation, and critical evaluation of my method ensuring I have indeed answered my Question for the Semester, including a tie in to my work undertaken though the ME6102 project. 6. What I have learned from completing this answer to the Question for the Semester. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 15 Can be upgraded with new technology, through flexibility through modularity The key advantages to an Open Engineering System are: Ability to implement mass customization Ability to deal with uncertainty in demand or environmental conditions Rapid production rates Modularity enables part/process/information re-use Relatively low cost compared to fully custom products/processes/information services (because of production scale) Ability to produce large quantities of goods Easy upgradeability of system/manufacture/product/process The key disadvantages to an Open Engineering System are: Higher cost of product/process/service (compared to mass production) Lower rate of production (compared to mass production) High capital costs (even more so than mass production) Higher skilled design/labor required Not required for all products, needs knowledgeable designer to determine if it is appropriate 2.5. PERSONALIZED DEFINITION OF OPENNESS I have defined openness as “readily adaptable to change”. However in the case of an open system this means active change, not passive reactions. In the case of Open Engineering Systems, some of the environmental factors that the system must be open to include: Changes in physical operating environment (external) Changes in physical operating condition (internal) Changes in the market Changes in the customer needs/requirements Changes in technology Changes in resources Changes in the system environs Changes in the government/legislation 2.6. PERSONALIZED DEFINITION OF AN OPEN ENGINEERING SYSTEM Based upon my analysis of the original definition of an Open Engineering System, and in the context of my personalized Q4S I have personalized the definition of an OES as follows. The Simpson, Lautenschlager, Mistree definition of an Open Engineering System (OES) as given in ME6102 is as follows: Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 16 Open Engineering systems are systems of industrial products, services, and/or processes that are readily adaptable to changes in their environment which enable producers to remain competitive in a global marketplace through continuous improvement and indefinite growth of an existing technological base. [2] I am also going to generally define the words “open” and “system” as context to my definition of an Open Engineering System in the same manner as Kevin Klein. My personalized definition of an OES is: Open – readily adaptable to change System – a single entity or conglomeration of entities that act to perform a function. A system has defined inputs and outputs, and a defined boundary defines the internal system from its external environment. Systems usually consist of products, processes, or services. Open Engineering System – Open Engineering Systems are systems of products, processes and/or services that are readily adaptable to a changing environment - customer market, physical environment, resources, and technology - and can function under a range of uncertain internal and external operating conditions. This adaptability allows for the continuous improvement and indefinite growth of an existing technological base. I feel that this definition keeps the core strengths of the original definition, the aspects of flexibility and indefinite growth, while removing aspects regarding industrial producers and adding the need for robustness to differing operating conditions as well as specifying how the system must be adaptable to the customer market, hence encompassing the need for mass customization. With my definition of an OES completed, I will now address why OES development in necessary in order to answer the Q4S. 2.7. WHY OES DEVELOPMENT? The original Q4S focused upon the development of mass customized products. In the following section I discuss the characteristics of a mass customization approach as well as its relation to OES realization. The original Q4S is posed as follows: In this context, we want you to provide a method to support the realization of mass customized industrial products for a global marketplace through distributed design and manufacture. The basic OES premise is that a quality product should be brought to market as quickly as possible and then that product line is continuously developed as quickly as possible to Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 17 in an effort to remain competitive. Thus open engineering systems are designed to be as flexible as possible. Flexibility can be obtained through [18]: Modularity - the relationship between the functional and physical structures of products, so that there is a one-to-one correspondence between physical structures and a minimization of unintended interactions. Mutability - the capability of a system to be contorted or reshaped in response to changing requirements or environmental conditions. Robustness - the capability of a system to function properly despite small environmental changes or noise. The combination of these features will also enable mass customization. A product that is open is easily adaptable, and in this manner it is possible to customize each product to a customer to a certain degree, resulting in mass customization. In an OES customer demand is modeled as a noise system input, and the product must be robust and flexible to the changing levels of demand. In this way an OES satisfies the needs of mass customization, while simultaneously providing a higher degree of functionality through continuous improvement and infinite growth through an expanding technological base. In my proposed method, another concept of customization is applicable, the concept of customizing the process to best suit the product being developed. 2.7.1. Mass Production vs. Mass Customization The drivers for mass production and mass customization at the very base level are the same: to produce a product that will satisfy a customer base in order to be competitive in the consumer marketplace. However, the methods that each process takes are very different. This is as a result of the different drivers of each method, how they strive to achieve customer satisfaction and productions success. To goal of mass production processes is to produce as large a volume of identical product as possible as a method of both reducing production costs and being profitable through large-scale production. This pleases the consumer base though low product cost and high product availability. However, every product made is identical, there is no personalization of the goods. To quote Henry Ford regarding the Model-T, “Our cars are available in any color you want, as long as it’s black”. The goal of mass customization process is to produce as many products as possible also, however each product is custom tailored to each consumer. This means that the cost will almost always be higher, and the availability will be less compared to mass production, however the customer satisfaction will be higher because the product is personalized for their individual needs and desires. 2.7.2. The Need for Open Systems Observing the success of robust systems, (primarily products and processes) over the last two decades, it is evident that the increased system quality and customer satisfaction results in increased sales and profitability. There are also smaller markets of customized Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 20 shifters, cranks, derailers and gears. However, as personalized and new as this would make the bike, it would be ultimately better for me to buy a new bike. Why? Because every few years a development life-cycle ends, usually where a break-though technology emerges. An example of this is when full suspension mountain bikes became the norm over rigid hardtails. A possibility would be to carry over the componentry from the old bike to a new full suspension frame, however technology evolves so fast that usually every piece is obsolete within a year, furthering the argument for complete replacement. What does this have to do with OES’s? Thinking more generally, it is possible to see that this trend will also occur on the manufacturing and design level of any modular product. The important part is to ensure that the manufacturing process is modular and flexible enough to create a good portion of the new parts required. At some point, product termination is required. This will depend upon the product, for example the B-52 is still going strong at over 50 years old because of its architecture, whereas electronics will be obsolete in the span of about four years. The solution to this problem is through modeling and metrics. The old system does not need to be abandoned entirely. The people who designed many of the old planes and rockets were geniuses. They build better products than many that are produced today. However, this information should be saved and re-used in the design of new products, utilizing and integrating new technology at the conceptual and embodiment phase such that the final product is designed with the best of what exists and works to the best effect in harmony with the new technology, not a patch to keep the enterprise afloat for a few more years. This has need has been synthesized into the requirements for my method, given later in the requirements section. If the product can be designed for effective modularity and then its life cycle continually measured and predicted, the company can determine when its time to sit down and determine what can be used for the next generation. Product families are important, but so is moving to a new generation of products, that may share little or nothing with the old system. Again the willingness to look beyond the current years bottom line and the acceptance of some risk is required. To articulate the key points: Re-use of function and principle is vital Re-use of product must be closely evaluated Enabling the easy upgrade of a system through modularity is vital Determining the dynamic life cycle of a system though metrics is imperative Continual long term strategic planning and analysis is the only way to achieve true success Deciding when a product should die through dynamic life cycle analysis and a comprehensive redesign is necessary is as important as creating a product that is upgradeable Sometimes its necessary to “re-invent the wheel” Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 21 3. LEVERAGING SOURCES The second paragraph in the Q4S as posed by Dr. Mistree in ME6102 is: We recognize that solutions evolve over time. Accordingly, we expect you to build on what has been done before. This is a new addition to the Q4S this semester. It shows that we must build upon existing work, using it as a base upon which to add value through our own original work and integrating new ideas. To specifically address this part of the Q4S I am briefly describing my major leveraging material and sources from ME6101, ME6102 as well as any other significant sources. 3.1. MY OWN PAST WORK I wrote a <Superb, Comprehensive, Beautiful laid out, and easy to read1> answer to the Question for the Semester in ME6101. An integral part of the success of this document was because of the huge amount of leveraging I utilized. I did not simply splice ideas, but rather built upon them in the context of my personalized question for the semester, infusing them into the framework of Pahl and Beitz creating something that was more than the sum of its parts. I also leveraged the best chapter ideas and layouts that I saw. This resulted in the development of my comprehensive answer and well flowing organization. Because of the success of this approach in ME6101, I am going to continue to build from these sources, as well as gather more from ME6102. 3.2. LEVERAGING MATERIAL FROM ME6101 In this section I present my list of leveraged sources from ME6101 that I am using for my ME6102 Answer to the Question for the Semester. I have formatted this list such that I list the work being leveraged, followed by the material and/or formatting I am using. For the sake of brevity, I am only listing the major components I have leveraged in building my answer to the Question for the Semester. 3.2.1. Previous Answers to the Question for the Semester Spierling, Todd ME6101 – “Engineering Design, Question for the Semester”, Dr. Farrokh Mistree, Fall 2000. o Posing the Question for the Semester – the giving of the context, structure, and viewpoint for answering the question for the semester o The envisioned world of 2020 – the breaking up of the world in general into four categories formating as well as content, the separation of visions of the world in general from the design and manufacturing relevant vision 1 Dr. Mistree’s feedback on my ME6101 Answer to the Question for the Semester Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 22 o The proposal that the 7MP tools can be used for communication in a distributed environment o The concept of using Suh’s independence axiom to enable concurrent engineering processes o The importance of team & time management o The pros & cons of operating in a distributed environment Mocko, Gregory ME6101 – “Answer to the Question for the Semester”, Dr. Farrokh Mistree, Fall 2001. o The concept of removing bottlenecks in the human-computer interface o The concept of a “digital notebook” o The concept of data interoperability o The approach of concluding remarks and value Matthew Lebeck ME6101 – “Question for the Semester”, Dr. Farrokh Mistree, Fall 2002. o The format of a simple abstract description o The approach of explaining the major requirements list headings for better understanding of the methods needs o The concept of allowing modules to be placed anywhere within the Pahl and Beitz method, adding flexibility o The format of the diagrammatic representation of adding modules to the Pahl and Beitz flowchart Flis Brian ME6101 – “ME6101: The Question for the Semester”, Dr. Farrokh Mistree, Fall 2001. o The approach of rigorously addressing all requirements and augmentation made o The presentation of the base method and its critical evaluation o The approach of addressing learning o The approach of addressing utility of the proposed augmented method o The verification and validation approach o The approach of providing a basis for the particularization of the Question for the Semester Ragu, Annand ME6101 – “Designing the Answer to the Question for the Semester”, Dr. Farrokh Mistree, Fall 2001. o The concept of separating method tools from the method process (steps and phases) o The idea of formally adding ethics to the method o The approach of creating “primary augmentations” Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 25 o The format of the Verification and Validation section – The diagrammatic approach of showing how each section of the square maps into the square as a whole adds strength. I have added value to this by also using the validation provided through my project as a strong argument. o The approach of a Closure section – I already have a learning section at the end of the answer to the Q4S, however I feel that a small paragraph at the end provides a clean finish. Klein, Kevin Answering the Question for the Semester, Spring 2003. o The approach of parsing the definition of an OES – I have utilized the formatting, however I am taking a different approach to the Q4S and hence have a different OES parsing and definition. Chamberlain, Matthew & Williams, Christopher Answer to the Question for the Semester, Spring 2001 o The approach of a joint answer to the Q4S – Although this is a fun and innovative format, I have not used it, and have utilized the classic leveraging approach instead of direct collaboration. 3.3.2. In Class Sources This semester I have again looked at work by Bjoern Avak, in particular his Assignment 3 and the link between his project and Q4S. I have also conversed with Phil Newton regarding differences in approaches he has seen in industry versus what I have learned in ME6101. I feel Phil is a very valuable resource because I have very little industry design experience and find the comparison between the two to often highlight issues that I must address in order to make my proposed method attractive to the industrial domain. 3.4. LEVERAGING MATERIAL FROM SRL Besides the papers and books referenced for material and concepts used in answering the Q4S I used some work by my fellow SRL lab members in order to address parts of my ME6101 Q4S that could stand improvement. In particular Dr. Mistree felt that my verification and validation section had loose arguments that needed tightening and strengthening. To address this I was pointed to Carolyn Conner Seepersad’s Masters Thesis, specifically her approach to verification and validation. I feel that this approach combined with the better graphical presentation leveraged from Jeffery Hoobler will significantly strengthen this section of my answer to the Q4S. I have also heavily leveraged from Jitesh Panchal’s work regarding digital design support systems and intergraded product and process development. Panchal, J. H., “Towards a Design Support System for Distributed Product Realization”, MS Thesis, GW Woodruff School of Mechanical Engineering, 2003, Atlanta, GA Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 26 o X-DPR concept – I am leveraging the X-DPR concept in conjunction with my own idea of the central project server from ME6101. This combination adds functionality provided through my ideas from ME6101 as well as new material with the strong distributed design support system of X-DPR. Panchal, J. H “Towards Integrated Design of Products and Design Processes”, PhD proposal o Integrated Design of Products and Design Processes – In my ME6101 method I designed a modular system that fit around the Pahl and Beitz framework. Through the use of Jitesh’s concepts of integrated product and design process realization I can further augment my work, removing the Pahl and Beitz framework greatly improving the flexibility of the method. The depth and strength of the material available for me to leverage from both previous classes and from SRL is integral to the completeness and creativity of my proposed method to answering the Q4S. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 27 4. DEFINING THE WORLD OF 2020 In order to answer the Question for the Semester correctly I must first define my world of 2020 in which my proposed method will be implemented, as a form of presenting the context in which the method will be implemented. I will describe my world of 2020 in several parts. First I will define the world in general, then specifically the world of design and manufacture, and then the specific advances that pertain to my augmented Pahl & Beitz method, acting as enabling technologies to support the method. 4.1. THE WORLD OF 2020 IN GENERAL My vision of the world of the year 2020 is based on looking at the world of 1980 and comparing it today, and extrapolating trends out for another 20 years, using the Observe- Reflect-Articulate tool. Barring sudden drastic changes that cannot be predicted, I feel that this process will give me accurate insight into the world of the future. Before I can discuss the details of how my augmented method fits in with the future environment of 2020, I must define a general vision to build from. For this definition I will leverage from Todd Spierling’s ideas in his ME6101 answer to the Question for the Semester [7]. I performed a critical evaluation of his paper in my assignments for this class, and found I agreed with many of his ideas. I will leverage these ideas, and build upon them where I wish to elaborate. My other source of leverage is Ashley Ceci, my partner for many assignments and projects in this semester’s ME6102 course as well as ME6101. We developed many of these visions as a group as we share similar styles of thinking. I will use our joint vision of 2020 as a base and extrapolate where more detail for my specific world is required. I have divided up the world of 2020 into five distinct aspects: 1. Political & Economic 2. Social & Cultural 3. Technological 4. Environmental 5. Corporate My vision of the political and economic future is leveraged from Todd Spierling in his answer to the Question for the Semester [6]. I also am also breaking my general vision of the world into the same four categories as Todd, however my vision in these areas is very different, and I felt I also need to address the environmental future of 2020 as an additional category. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 30 Industrial facilities that have better environmental management are no longer allowed to sell their “environmental points” (the net amount of pollutants they are allowed to put into the air, water and soil) to those companies that are not up to par. Instead, companies that cannot consistently meet environmental regulations are shut down in developed countries. 4.1.5. Corporate The corporation is responsible for dealing with greater levels of uncertainty in the consumer market due to shorter product life cycles. This will result in a greater investment in the study and research of design methodologies and theory. Corporations will consolidate in their individual industries, resulting in single large companies. These companies will be multinational, and operate on a global scale involving manufacture, design, and marketing, this equates to managing a much larger resource pool. The interaction of these companies with the supernational economies and individual countries will be closely monitored and tied with governments of the countries involved. Smaller, more agile companies will exist in new and niche markets, or working in conjunction with the large conglomerates working on outsourced projects. Continued work with a single company usually results in conglomeration in order to remain competitive. 4.2. THE DESIGN & MANUFACTURING WORLD OF 2020 Now that I have defined the general world of 2020 I will address the specific changes that will affect the design and manufacturing processes in the future. These changes will be reflected in my augmented method, as they dictate the environment in which the method will be implemented. I have broken the world into the following six categories in the spoke and wheel diagram show below in Figure 4.1. Figure 4.1 - Vision of 2020 Spoke & Wheel Diagram Manufacturing People R&D Marketing Information Resources Geographically Dispersed Resources Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 31 4.2.2. Manufacturing With manufacturing, there will need to be a standardization and modularization of component parts, machine parts and production facilities. Every physical aspect of manufacturing should be location independent; and every process step should be standardized. This will allow processes and people to become interchangeable. Standardizing selection processes become important for this system during the design of the manufacturing process. The strengths of standardized and modular manufacturing components are dispelled if the process for selecting the appropriate units can cause non-uniformity plant layouts and operations. 4.2.3. Information Information must be readily available to anyone on the project, in real time. This requires the use of customized data pools, and specialized global networking. This will be must for keeping in constant contact with a dispersed work force. This will facilitate swifter and easier communication of ideas, thus allowing for disperse groups to base decisions on a global/company wide set of requirements. Versioning control will also become much more important, as you’ll have people in different time zones working on the same task at different times of the day. This will allow for cleaner task and project handoffs, as well as almost 24 hour “working” shift. 4.2.4. Marketing Marketing needs to be handled at a global scale. No longer will items have country, or nationality specific markets. This isn’t to say that individuality will be a thing of the past, but base product needs will probably become more global as far as food, transportation, basic clothing and the like. Marketing will also advance on a personal scale, through mass customized advertising. Global access to records of activities and purchases, similar to the history and cookies features on web browsers, will allow for individual advertisements directed to individuals based on their previous records. Decisions in marketing will become more important as product life cycles will shorten, and consumer fickleness and expectations will increase. This will facilitate the need for a standardized and quicker means for making product decisions. 4.2.5. People The rise of mass communication will continue to contribute to the homogenization of cultures around the world. As groups become less isolated, they become influenced by the ideas of other cultures. Many people will Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 32 become multi-lingual, speaking languages most encountered in their global workplace. For people to be effective members of a group, they must be part of as many aspects of the overall design process as possible. To that effect, meetings must also become “virtual”. This will limit the unintentional exclusion of certain internal, or external divisions due to distance, or time constraints. Also, the information itself should be understandable by any person in any nation that is required to work on the project (diagrams vs. text). Selection is an important part of this information that must be communicated between dispersed group members. 4.2.6. Resources No longer will company resources be departmentalized. Teams and groups will be required to make faster decisions with the understanding that the footprint of effect will have a larger scale, not only affecting neighboring departments but also international divisions. Thus the networking of the future will need to be done on a global scale, through true wireless (satellite based) transmission. This will be a must for keeping in constant contact with a dispersed work group. In addition, physical resources, those substances that make up the components used in production, must become interchangeable. This will allow processes to become location independent. 4.2.7. Research & Development The impacts of a standardized selection process on research and development will be similar to those outlined above, particularly marketing, as development life cycle will be shorter. Because research and development encompasses most of the above points, it is not necessary to readdress here. 4.3. ENABLING TECHNOLOGIES The following technological developments are utilized in my proposed method as external augmentations, and as such are required technological developments by the year 2020. These are described in detail in the External Augmentations section; however, a brief introduction is pertinent as part of my vision of 2020. Radio Frequency Identification (RFID) – A method by which products can be tracked from manufacture, through distribution, onto store shelves and finally into the households of individuals. This will be an invaluable tool that, when integrated with a global, wireless data network, will allow companies to track, real time, changes in demand, and, must importantly, purchasing trends. Rapid Manufacturing (RM) – This is something much akin to Rapid Prototyping, but the output is not a simple model, but a product ready component part necessary for full, production ready assembly Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 35 Another issue is standards set by the government, for example air pollution standards are much higher in countries such as Holland than in the United States, requiring different designs or components. Sometimes these differences are difficult to know unless the company has local presence, and they can react quickly to local market changes as well. 4. Utilize Existing Facilities & Employees – Because of the scale of the large businesses in my future of 2020, it does not make financial sense to relocate the entire business. The costs involved would often be too high to recover in a suitable time frame. This can be avoided by the take over of existing facilities, incorporating them into the larger company as a whole. 5. Access to Specialized Skills – The education of some specialized skills only occurs at a few limited locations around the world. In order to tap these specialized skills, the company must buy them as needed as replication of these skills internally would be too costly and involve too much time. 4.6.2. Reasons against globally distributed realization: 1. Communication – With increased distance there are increased communication difficulties. For example I am much more likely to ask another member of SRL for help with a problem than call or walk all the way home to ask a roommate. Distance also removes the possibility of face-to-face contact, which is important in establishing relationships. 2. Lack of Commonality – Corporations are held together by their common goal, and this goal is recognized by all of its employees, creating a bond. These bonds are important for effective teamwork, which is a requirement for effective product realization. If the company is distributed globally, each part may establish its own identity, making teamwork between groups less effective. Again, this problem stems from the communication barrier. 3. Economies of Scale – There is an overhead cost associated with maintaining multiple facilities that does not exist with a single location. These can involve increased communication and travel costs, taxes from different nations, and more buildings to maintain. Overall I believe that the advantages of a distributed environment heavily outweigh its disadvantages. Most of the problems involve communication. I believe that advances in communication technologies such as video conferencing and better remote computer access will help, but for true progress to be made a paradigm shift must occur. The very method of design must change to better address the distribution of team members. This is an issue I have considered in detail during my development of both my proposed method and the system supporting its implementation. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 36 5. BASE METHODS The method proposed in this answer to the Q4S is build upon the foundations of two design methods and two design tools. A brief description of the two methods is given in this section, and a description of the tools in the following section, as context to the critical evaluation of each. This process serves to take stock of what methods are available today that I can use to build upon and augment to answer my Q4S. 5.1. THE PAHL AND BEITZ SYSTEMATIC DESIGN METHOD In this section I will describe the Pahl and Beitz method and identify the key elements, leveraging my ME6101 Assignment 1.b and the Pahl & Beitz text [1]. This is important as it forms the concept of the phased process design approach. I quote the authors often because I feel their words are ideal for the description of their own method. The Pahl & Beitz method of design is organized into four phases: 1. Planning and Clarifying the Task – Specification of task and requirements 2. Conceptual Design – Specification of principle(s) 3. Embodiment Design – Specification of layout (& construction) 4. Detail Design – Specification of production The flow of the Pahl & Beitz process is best demonstrated using a flow chart as given in the text in Figure 3.3 and shown below in Figure 5.1. Decision making steps are required after each phase, to determine the direction of the project and if it should continue. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 37 Figure 5.1 - The base Pahl and Beitz method flowchart Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 40 5.2. NATHAN ROLANDER’S ME6101 ANSWER TO THE Q4S In this section I will describe my ME6101 Augmented Pahl and Beitz Method. This method is a modularized and augmented Pahl and Beitz approach that has many additional modules to bring the method up to date and address the linearity, rigidity, and lack of information storage in the Pahl and Beitz approach. Therefore the foundational method it was built from is the base Pahl and Beitz systematic design method. The core elements of Nathan Rolander’s augmented Pahl and Beitz method are outlined in this section. For the sake of brevity, only a high level description is given. For more detail please see Answer to the Question for the Semester, Fall 2003. Nathan Rolander. [19] 5.2.1. Concept of Method The concept of my augmented Pahl & Beitz method was to retain the base structure of the system, but add modularity, concurrent processing, and design process re-use. These are three distinct augmentation concepts, however concurrent processing and process re-use build off of the modularity of the method. In addition to these core augmentation I add ed design process modules. Each module adds functionality to the base Pahl & Beitz method to address the needs of a distributed design method for 2020. These modules do not have a defined position on the overall Pahl & Beitz process, this is to maintain the flexibility of the method for personalization. 5.2.2. Core Augmentations Modularity The four phases of the Pahl & Beitz method were utilized as a core framework, with the augmented design modules and design tools being inserted into each phase as applicable. Each individual phase of the Pahl & Beitz method phases is also a module. This allows for the design process to be re-organized to better suit design processes such as reverse engineering or variant design. This modularity of each of the phases of the Pahl & Beitz method is what facilitates the rest of my augmentations, allowing for the insertion of new modules where applicable, concurrent processes, and the easy personalization of my augmented method Concurrent Processing Process modularity is required for concurrent design processes to occur. This is because steps must be added where the design process is to divide and then re-integrates. It is also important that the sub-tasks be able to use any combination of the four phases. For example, a task may be handed off that requires clarification, then conceptual design, but then needs to be re-integrated into the project as a whole for the embodiment and detail design phases. The modularity of the four phases allows for any combination of them to be completed before re-integration into the main design. This splitting into concurrent Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 41 processing from the main process flow was accomplished through the use of splitting and re-connection modules, based upon the implementation of Suh’s independence axiom [6]. Information Re-use The ability to re-use existing design processes and parts of design processes was an integral part of my augmented Pahl and Beitz method. This was enabled by the modularity of my design method, allowing individual parts to be saved and stored for later use. This is possible because of the central project server, as the whole design process, all its functions, schedules, calculations, and part files are stored in a digital format. My concept for this is best thought of though an analogy of programming sub-routines. In a program a sub-routine has inputs (specifications and requirements), and outputs (data, information, graphics, etc.) This is true of each of my modules in my augmented design process. The entire process is a single program, but each modules is a sub-routine. In this manner my design process is akin to augmented function structures, but now the function is actually the design process. For example, the input function to conceptual design module would be the requirements list, and the output would be the principal solution. These digital 'design function' structures and their linked associated files and parts components will be stored on the central server and re-used when applicable. 5.2.3. Modules & Tools The modules and tools are specifically not directly linked to any specific phase of the Pahl & Beitz method. This is because my augmented Pahl & Beitz method was designed to be customized for each customer’s needs of Future Tech. These modules each represent a design method concept in ME6101 that I thought should be integrated into the Pahl and Beitz method in order to make it applicable for design in 2020. The tools used were taught both in ME6101 as well as pulled from industry practices. The difference between tools and modules is that the modules represent a process to be followed, while the tools are methods of implementing that process. Modules: Tools: Quality Design Decision Making Communication Ethics Embodiment Centrality Decision Making Team Contract Planning & Time Management Ideation Attention Direction Standardization Observe-Reflect-Articulate Communication Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 42 5.2.4. Flow Chart The flow chart for my personalized and augmented Pahl & Beitz method is on the proceeding page, shown in Figure 5.. The core modifications to the Pahl & Beitz flowchart are in bold-italics. The only permanent flow changes to the base Pahl and Beitz method are the establishment of the project central server at the start of the process, and the phase gate and check for concurrent processing connectivity at the start and end of each phase. The rest of the modules and tools are inserted as needed by the design engineer in order to personalize the process; this is represented by the line item to each working phase and decision phase labeled appropriately. The Pahl & Beitz color coding system is as follows: Blue - Clarification of Task Green - Conceptual Design Red - Embodiment Design Yellow - Detail Design The Module & Tool color coding system is as follows: Blue - Process Modules Red - Process Tools Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 45 6. BASE TOOLS In order to support the proposed method I am building in this answer to the Q4S, I will also require tools to execute it. I believe that the two most relevant supporting tools for the distributed design of Open Engineering Systems are the Decision Support Problem and the eXtensible Distributed Product Realization environment (X-DPR). This process serves to take stock of what tools are available today that I can use to build upon and augment to support and implement answer my Q4S. 6.1. THE DECISION SUPPORT TECHNIQUE Another primary source of leveraging is the Decision Support Problem Technique. This approach is fundamental to the decision based design approach, which I believe has excellent principles. The principal role of the DSP technique is to provide support for human judgment in synthesis of design, manufacturing, (distribution,) maintenance, (and recycling) considered as a single process. The technique is based on the following assertions: Synthesis involves a series of decisions, some of which may be made sequentially and others that must be made concurrently. Synthesis involves hierarchical decision making and the interaction between decision must be taken into account. Productivity can be increased by the use of analysis, visualization and synthesis in complementary roles. The technique must be process-based and discipline-independent. The technique must be suitable for solving open problems and must facilitate self- learning. Decisions associated with most real-life engineering systems are characterized by the following: The decision problems are multileveled, multidimensional and multidisciplinary. Most problems are loosely defined, open to the environment and also open to changes in the environment and other bounding conditions; thus, virtually none of them has a unique solution, but solutions must be negotiated for all of them. Their solutions are less than optimal and are called satisficing solutions. There are multiple measures of merit for judging the level of acceptability of a design; all of the measures may not be equally important. All the information required may not be available. Some of the information may be hard, that is, based on scientific principles, and some information may be soft, that is, based on a designer’s judgment and experience. All of the information necessary for modeling systems comprehensively and correctly will not be available for real-world, practical systems. Therefore, the solution to the problem, even if it is obtained using optimization techniques, cannot be an optimum with Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 46 respect to real-world conditions. Therefore, the function of the DSP technique is to provide support for human judgment in seeking superior solutions. 6.1.1. Types of Decisions The DSP Technique focuses upon the two kinds of decisions encountered in design, selection and compromise. [20] Selection – a primary decision – the process of making a choice between a number of possibilities taking into account a number of measures or merit or attributes Compromise – a primary decision – the determination of the “right” values )or combination) of design variables to describe the best satisficing system design with respect to multiple goals. Derived/coupled decisions – a combination of primary decisions, for example selection/selection, compromise/compromise and selection/compromise decisions. Further classification of decisions can be broken down as follows: Hierarchical – decisions in which both selection and compromise occur. Conditional – decisions in which the risk and uncertainty of the outcome are taken into account. The DSP Technique can be used to solve any of these types of problems, including coupled variants of the problems. The selection DSP formulation is as follows: Given A set of alternatives Identify The principal attributes influencing selection The relative importance of attributes The feasible alternatives Rate The alternatives with respect to their attributes Rank The feasible alternatives in order of preference based on the computed merit function values The compromise DSP is as follows: Given A feasible alternative Find The values of the independent system variables The values of the deviation variables (they indicate the extent to which the goals are achieved) Satisfy System Constraints: These must be satisfied for the solution to be feasible Goal Constraints: These need achieve a specific target value as far as possible Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 47 Bounds: Lower and upper limits on the system variables and deviation variables Minimize An objective that quantifies the deviation of the system performance from that implied by the set of goals and their associated priority levels. 6.2. THE X-DPR DESIGN SUPPORT SYSTEM In an environment where globally distributed designers work collaboratively on a product development task, engineers want to make best use of available resources distributed across the globe for reducing product development life cycle time. In order to make use of distributed resources along with systematic design methods, design support systems are required, that support engineers in at least two areas – provide distributed computing capabilities to utilize resources over the Internet and provide capabilities to manage and execute design processes in collaboration with other designers. In this manner, we present an eXtensible Distributed Product Realization (X-DPR) framework that provides support for designers in reducing product realization time by facilitating the use of systematic design methods to manage design processes and by facilitating the use of distributed computing resources (or agents). The Decision Support Problem Technique is implemented in the X-DPR framework for modeling design processes hierarchically and collaboratively. Moving further along the design timeline, from modeling to execution of a design process, the framework is used for integrating heterogeneous software tools developed on different platforms and in different computer languages. The tasks defined in a design process are carried out remotely as web services. The remote invocation and execution of web services is done using XML based SOAP (Simple Object Access xvii Protocol), which is a language and platform independent standard for achieving interoperability between heterogeneous software systems. The outcome of this work is a framework for distributed product realization activities that helps in making effective use of distributed resources; automating repeated and monotonous tasks and process management thereby reducing product development time and enhancing productivity. [5] The X-DPR framework is designed based on peer-to-peer communication between agents, where each agent is an independent entity communicating with other agents. X- DPR is an open system in which different modules can be easily integrated into the system for enhancing the functionality of the overall system. Engineers can integrate their own applications residing on their machines with X-DPR, this will help to create a global library of engineering tools over the Internet. This library can then be integrated with tools from other areas such as marketing, sales or other business services to realize a global enterprise. X-DPR framework supports meta-design using the Design Support Problem (DSP) Technique. The system is designed so that a designer can easily model his/her design process using visual tools. This capability for meta-design is unique in X- DPR. Engineers can then connect process models with services available in the global library using the Internet and execute complete design processes online. X-DPR provides flexibility at a design process level. It enables designers to design a process and replace Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 50 7.2. NATHAN ROLANDER’S ME6101 ANSWER TO THE Q4S My ME6101 answer to the Q4S was comprehensive, addressing the three major identified shortcomings of linearity, rigidity and information storage in the base Pahl and Beitz method effectively. However, my ME6102 Q4S asks for more than an augmented Pahl and Beitz method, it requires the integrated development of open engineering products and design processes, something which is not possible in any manner in my ME6101 Q4S. 7.2.1. Core Augmentations The augmentations to the base PB method in my ME6101 Q4S were as follows and were previously described in the Base Method section: Concurrent Design Processing Design Process Modularity Design Process Reusability The core augmentation goals are still valid for the integrated design of open design processes and products. However, the methods for achieving these goals are no longer valid. The requirements of an OES’s flexibility through modularity, mutability, and robustness are not addressed, only the modularity and that is limited as described next. The method also does not address the design task hierarchy and organization. Finally, the central project server concept proposed contains good ideas, however the centrality of the system is a weakness that needs to be addressed. In summary I feel that the following needs must be addressed in order to answer my Q4S: Integrated development of open design processes and products – In order for truly open products to be realized, the process used to design them must also be open. Flexibility of both product and process through o Modularity o Mutability o Robustness The benefits of a flexible process will yield the most efficient design in the shortest amount of time and investment, while simultaneously allowing for the development of any product. A flexible product enables mass customization and easy updating and upgrading for the product as the technological base expands. De-centralization of the design process for enhanced distributed operation – A centralized process is not robust as is the system is disrupted the entire team is affected. De-centralization means that noise is much better handled, as it only affects the local entities involved and the rest of the project can move forward. Hierarchical task structure and organization of the design processes developed – The original structure to my method was given by the Pahl and Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 51 Beitz system. However, without this backbone, a system for the arrangement of workflow and module connectivity must be developed. A strong digital design support system for distributed collaborative operation – The central project server must also be de-centralized, and also be updated to better support the distributed design team. 7.2.2. Modules & Tools The modules and tools integrated into the base PB method in my ME6101 Q4S were as follows and were previously described in the Base Method section: Modules: Tools: Quality Design Decision Making Communication Ethics Embodiment Centrality Decision Making Team Contract Planning & Time Management Ideation Attention Direction Standardization Observe-Reflect-Articulate Communication Many of these modules and tools are still applicable as building blocks in my proposed method for answering the Q4S. However, many require updating to bring them in line with the OES paradigm or that need to be removed. Also, there have been new material introduced in ME6102 that I feel is also required for effective design for OESs in 2020. The modules to be removed are: Communication – I do not believe that this module is necessary as communication is integrated into the digital design support system. Centrality – The point of my proposed method is de-centralization, therefore a centralization module is obsolete. The modules to be added are: Game Theory – The modeling of supply chains and competitive or cooperative relationships. Phase Gates – The highly important stages of a design process where important decisions are made involving all team members which will be affected by the outcome of the decision. Tolerance Design – The consideration of tolerances and range variables rather than single parameters from the earliest stages of design. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 52 The tools to be added are: DSP Techniques – To provide support for human made decisions throughout the design process. Decision Support through Modeling & Simulation – To provide computer aided support to the decision process, also the focus of my ME6102 project. I will develop these directions for augmentation further in my requirements list in the following section. 7.3. THE DSP TECHNIQUE The DSP Technique is not a method, and hence cannot be critically evaluated as one. Instead I believe it is valuable to describe how I intent to use the DSP Technique to implement my proposed method. I intend to use these aspects and derivatives of the DSP technique: Interfacing – The DSP Technique provides a strong organizational structure, identifying the key variables of the system and clearly stating what is required. This system can be used as a foundation for module interfaces. Communication – The DSP Technique can also be used for communication, both between distributed members of a team, as well as between the human and the computer through different levels of abstraction provided in the DSP pallet. Decision Support Tools – The compromise and selection DSP tools are to be utilized as tools in my proposed method, as well derivatives including them such as the RCEM and PPCEM methods. 7.4. X-DPR The X-DPR framework is much stronger than the central project server I proposed in ME6101. However, it still has its limitations that need to be addressed for digital design support in 2020. These limitations are: Search Service – X-DPR relies upon a search service to provide the location of the other systems on the network, in a similar manner to the Napster file sharing system. I believe that this needs to be updated so that there is no central file listing service, as employed in the Gneutella file sharing system, adding both flexibility and robustness to the system through redundancy and de-centralization. Central Project Sever – The X-DPR system needs to also be augmented to include project management tools, such as those described in the central project server in my ME6101 Q4S. This would include a pseudo central collection of all material that would really be distributed and backed up throughout the network. Program Wrapper Utility – X-DPR does not contain a easy to implement or efficient system for users to wrap the programs of processes that are executed by Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 55 Existing Production Factory limitations, maximum possible dimensions, preferred production methods Modularity, process flexibility, Global distribution of plants Existing Quality Control Possibilities of testing and measuring International quality standardization Existing Assembly Special regulations, installation, siting, foundation Disassembly, part standardization, re-use Existing Transport Limitations due to lifting gear, clearance, means of transport Import/Export regulation, existence of high speed conduits, storage node location Existing Operation Quietness, wear, special uses, marketing area, destination Local area operation regulations Existing Maintenance Servicing intervals (if any), inspection, exchange and repair Self maintenance & repair capability Existing Recycling Reuse, reprocessing, waste disposal, storage Product component re- use, design for disassembly, local vs global regulations Existing Costs Maximum permissible manufacturing costs Global profit & cost considerations Existing Schedule End date of development, project planning and control, delivery date Parallel scheduling, schedule divergences & re-integration, WAN use New Communication Communications protocols, user interaction, remote operation & control New Automation Automation protocol, level of AI, learning parameters, awareness New Global Physical Region Climate, terrain, resources, education level, population, average age, experience New Political Stability, public interests, restrictions, humanitarian views, sanctions, tariffs & taxes New Ethics Company responsibility, individual beliefs, individual responsibility New Storage Data, hardware, software, materials 8.1.1. Communication Many of the products designed in the future will have to communicate with both the user, other machines, and globally for remote operation. This protocol must be considered and integrated into the development. For example the idea of internet controllable washing machines developed a few years ago, or the emergence of building control systems such as BacNET. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 56 8.1.2. Automation Another technology development, many products have automated functions, and the control and development of these systems will become formalized in the future, deciding upon control systems will become a requirement. Artificial intelligence would fall under this category as well, the selection of learning parameters, awareness of the surroundings and development protocols. 8.1.3. Global Physical Region Because of the distributed environment of the future, considerations must be taken for the physical location of the design and design process. These parameters can affect the design from the physical geography or the human aspect, and both must be considered. 8.1.4. Political Along with the physical environment, the local political environment must be considered. The local policies can have an effect on both the product design, and more importantly the design process. Issues such as tariffs, taxes, and the local stability of the region must be considered before a full scale investment in the area. 8.1.5. Ethics Continuing from the environmental consideration are the notion of humanitarian awareness. This must be conducted on the global and local scales, and include the respects of individual beliefs of individuals, and acceptance of the responsibilities of both corporations and its individuals. 8.1.6. Storage Considerations for the storage and retrieval of both physical products and components must be considered as well as information. Accessibility to both of these resources is of utmost importance for OES design and realization as the flexible nature requires quick adaptation and hence access to new parts and information. 8.2. CORE REQUIREMENTS FOR OES DESIGN & REALIZATION In order to address the identified shortcomings of the base methods from my critical evaluation of each, the following requirements list has been constructed. The major headings identify the major tasks to be accomplished, with the individual requirements given under each heading shows the steps or tasks to be completed in order to achieve the major heading requirement. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 57 Nathan Rolander Core Requirements List for OES Design & Realization 4/15/2004 Problem Statement: Provide a computer supported method, enabling knowledge storage and re-use, to support integrated realization of design processes and products for a global marketplace through distributed design and manufacture. Schematic: Information P E I P E Pr od uc t D es ig n D es ig n Pr oc es s D es ig n Product Requirements Process Requirements Product Specific Information Process Specific Information Time Hie rar ch y Conceptual Design Embodiment Design Detail Design Clarification of Task Conceptual Design Embodiment Design Detail Design Clarification of Task Function Behavior Structure Robust Design Reliability Design Axiomatic Design T1 D2 D3 D4 T5 T1 D2 D3 D4 T5 Changes Need/ Want Requirements Responsibility Integrated Design of Products & Design Process N Enable design process modeling, representation, analysis, and synthesis to support the integrated realization of products and design processes N Enable design process modeling in a reusable manner through representation in a computer interpretable manner N Allow the design process to be synthesized from existing process elements W Facilitate the effective leveraging of design knowledge N Utilize the DSP Technique approach of meta-design and design N Utilize decision based approach with phase gate modules at major decision points closing design space Knowledge Storage & Re-use N Store knowledge & processes in modular template format N Enable storage of both abstract and quantitative knowledge W Allow for both top-up & bottom-up re-use of knowledge N Implement hierarchy of data types for organization N Implement object oriented variant type module approach for automatic mutation to available data and type Hierarchical Design Process Organization N Utilize decision based approach to building modular design process N Transform heterachical structure into hierarchical structure N Process structure is dependent upon information generated from product and design process development W Provide design process module structure with “intelligence”, allowing the process to be formed dynamically N Provide metrics for analysis of design process modules N Provide rules & algorithms for developing structure of modular design processes Nathan Rolander for all Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 60 8.2.3. Hierarchical Design Process Organization A new challenge that has arisen because of the flexibility of the modular design process is actually forming the modules from a heterarchy to hierarchy. This process will still need to follow the Pahl and Beitz approach of the four phases of design, moving from the abstract to the concrete as the design is finalized. However, because of the flexibility of the process, and its dynamic nature as it is built concurrently with the product, means that there must be a systematic system for the construction of the design process. This should involve consideration of information generated in both the product and process domains. This is only possible with the establishment of a design metric for determining the time efficacy of a process versus another. With the system complete, the design process can be as flexible as required yet retain a strong and efficient structure built upon a set of axioms for determining efficient process organization. 8.2.4. Digital Design Support System In order to realize a distributed design methodology, a support system must be utilized for its implementation. In order to support the method I am proposing, this system must be a globally distributed computer network that enables communication and the execution of the design process modules from individual team members machines. This support system must also have a pseudo centrality element. This means that access to all of the machines, or databases will seem central to the user, however the information, processing, and users will be distributed. This allows for redundancy as well as parallel processing for greater security and efficiency. Through this pseudo centrality all of the information regarding the design process will be stored for fast retrieval and searching. 8.2.5. De-centralization Working in conjunction with the digital design support system is the concept of de- centralization. A centralized process is not robust as is the system is disrupted the entire team is affected. De-centralization means that noise is much better handled, as it only affects the local entities involved and the rest of the project can move forward. This decentralization is true of both the process and the digital support system implementing the system. Both of these entities must be considered in order for this decentralization to occur. 8.2.6. Decision Support Decisions occur throughout the design process, each decision closing an area of the design space, and making the design more concrete as the process moves towards the final solution. The final outcome of the design is therefore heavily dependant upon the quality of decisions made during the design process. Approaching these decisions in an ad hoc manner is an effort in futility. The DSP Technique has been shown to be an effective tool for decision support, particularly in the digital form with further support from modeling and simulation. The modularity and computer supported method I propose is founded upon the decision based approach, and therefore must include Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 61 modules for the decision support tools of the compromise and selection DSP to provide support for human made decisions throughout the design process. A final consideration of the applicability of game theory towards the modeling of human-human interactions for systems such as supply chains and cooperative verses competitive relationships should be considered. 8.2.7. Parametric & Tolerance Design The consideration of tolerances and range variables rather than single parameters from the earliest stages of design are vital to the development of robust designs. This continues the Taguchi approach to robust design, minimizing the systems sensitivity to both internal and external noise in its various parameters. The design method must also make available the use of robust manufacturing techniques, maximizing the quality of the manufactured goods with a minimal number defects. 8.2.8. Product Platform & Family Development In the development of any product, the consideration of a portfolio of products should also be considered. This will easily allow for expansion in to other quadrants of the market segmentation grid. If these considerations are not made through component architecture or commonality this is very difficult and costly. If modularity is already considered then this expansion is easily facilitated. In the same way as the consideration of parameter and tolerance ranges in robust design, the concept of matrix solutions of product variants should be considered for product family development to enable concurrent development of a product family if desired. 8.3. SUPPORTIVE REQUIREMENTS LIST The supporting requirements list is to be created in the same way as my ME6101 supporting requirements list. The material is predominantly leveraged from existing answers to the Q4S and classmates’ ideas that I feel are valuable, mixed with my own ideas. These requirements do not need to be addressed in the proposed method, nor do they need to have their utility shown. Rather, these are ideas that are simply good ideas that should be implemented during the application of the proposed realization process, or applied by Future Tech or its clients. This list will be built and developed through learning essays as the semester progresses. Because of the magnitude and small scope of many of these requirements, only the key points involving the management and polices of Future Tech will be addressed. Other simple requirements, such as the layout of the facilities, are assumed to be included in the operations manual of Future Tech, and are not part of my answer to the Question for the Semester, but are points to consider in the effective implementation of the method. For this list I have leveraged many sources, and they are listed accordingly, if source is blank then the ideas are my own. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 62 Nathan Rolander Auxiliary Requirements List for Personalized Question for the Semester 11/23/03 Problem Statement: How should the Pahl & Beitz systematic design method be personalized and augmented to allow concurrent processing and design process re-use, to support the realization of technical products and processes accounting for the geographically disperse manufacturing, business and managerial environment of the year 2020. Changes Need/ Want Requirements Source Organization N Production hand-off schedules must devise for tasks that can be easily passed on to other groups with limited explanation of progress made, and future direction. This will allow for 24-hour productivity. Ceci W Shift schedules must take into account needs of distributed teams in different time zones, to allow for both sequential and parallel operations. N Means of quickly integrating part time, or temporary employees into a project life cycle. Ceci N Incorporate slack time & parallel team planning into scheduling Murray W Include broad range of employees for all disciplines and backgrounds in teams and particularly ideation meetings Company Policy N Profit margins must be looked at on a global scale Ceci N Capital Investment orientation for investment Hoobler N Customer consulting prior to engaging in project Jeram N Research & Development Focus on Technological Innovation Hoobler N Reward structures for workers must be altered. It should no longer be based on costs, but on contribution. Ceci W Encouragement of seeking of patents & patent protection N Transition from short term to long term thinking and planning Jeram W Collaborative Supplier Relationship development Hoobler N Internal department transfer, and placement mechanisms must be available to maximize employee effectiveness and morale W Promote individually, but maintain clear sense of goals and purpose. W Involvement of company in local organizations and community, such as humanitarian efforts, community BBQs etc. N Use strong standardized security protocols W Provide a Pleasant & Efficient Workplace Environment Jeram N Encourage long-term employee retention: employee morale must be kept high Murray Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 65 products sold, with respect to each product. The solid lines represent price of each product. Figure 9.1 - Product life cycle graph of Phillips technical products Analyzing the graph it is evident that product life cycles are dramatically shortening. This means that the development time must also be shorter. This is because once a new product is introduced, there is a very short time period before it is copied by another company and sold for less money, becoming a commodity item. Looking at the VHS line, there is a hiccup in the graph during 1979, and correspondingly there is a flat line is sales. However, due to the lengthy product life cycle, Phillips is able to correct this problem and sales continue to rise. During the development of the DVD and DVD-R systems, a problem like this would have put Phillips out of the market. Analysis of this example demonstrates the need to of a product development system to accomplish two major tasks. The first get this right the first time, with no updates, patches or corrections. The second is to develop products at very high rates in order to stay profitable during the short period of total market penetration. These needs must be addressed in my proposed method for OES realization. 9.3. INCREASING PRODUCT QUALITY The ability to get quality products out the first time requires two aspects of a design process. The first is a systematic design process to make sure that nothing is overlooked and no errors are made. The second is to ensure a quality product. In this sense I will use two definitions of quality. The first is the Taguchi definition of quality, in the creation of robust products. These products will be insensitive to changes in operating parameters or the environment, and therefore will operate as intended in different condition and continue to perform well as 0 Price Quantity Price Quantity 03938883 78 73 98 VHS DVD DVD-R Date Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 66 they age. The second is the concept of creating superior solutions. This can be achieved through synergy effects through collaboration as proposed by Dr. Masataka Yoshimura in his presentation to the SRL. Through the collaboration of ideas and information, superior solutions can be obtained than if only parts of the problem or even all aspects of the problem were analyzed by only one person. This is because of the difference in areas of expertise offered by different team members. If this depth and breadth of experience and knowledge can be tapped effectively, the resulting product performance gains while simultaneously reducing production costs would be substantial. 9.4. ADVANTAGES OF MODULARITY OF DESIGN & MANUFACTURE In order to achieve fast product development and manufacture concurrent processing must occur. Examples of this methods success are available today, and I believe concurrent processing will be a requirement for competitive practice in 2020. One of my roommates was watching the Discovery channel, and the show was about great engineering feats. This particular episode focused upon cruise ships, specifically a single cruise ship and its construction, “Voyager of the Seas”, the largest cruise ship constructed by a factor of nearly 50%. After watching this show and the construction of the ship, I reflected upon the requirements for modularity and product families, and how it is dealt with today in order to extrapolate requirements that must be met for modularity in the world of 2020. This gargantuan cruise ship was the focus of the show “Engineering the Impossible”, shown on the Discovery Channel. The areas of the show that caught my interest were regarding the modularity of the construction and the requirements that modularity imposed upon the construction methods. Usually a ship is constructed in one location, with smaller items being shipped in from off-site such as room furnishings. In order to complete the construction in 2 years Voyager of the Seas was constructed in 5 large pieces, and the thousands of individual rooms were built off site and assembled once finished at the launch site. This was a challenge because of the tolerances involved. Humans could not hold tolerances small enough to allow for the modular construction of the main hull, and as such CNC plasma cutters were used at increased cost. Extrapolated to 2020, faster construction will be required, and parallel construction of modular parts can cut production times dramatically. However, as demonstrated by the construction of “Voyager of the Seas”, advanced construction techniques are also required to enable modularity. This is an angle I had not though of before for large scale projects. Other examples of modularity used on board were the engine pods, completed as individual off site projects, there are essentially gigantic outboard motors, individual pods that can rotate, making a rudder unnecessary. It was this simultaneous construction that allowed the ship to be completed in such a short time period, thus maximizing profits by maximizing the time “Voyager of the Seas” will spend as the king of the cruise ships. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 67 The construction of Voyager of the Seas has shown that modular construction is possible, and has demonstrated its benefits. It has also shown the higher requirements in terms of tolerances and precision required for the re-integration of the assembled parts. In the world of 2020, even more advanced manufacturing technologies will exist, and the top- of-the-line techniques of today will be commonplace. This will make modular construction not only feasible but a requirement for competitive large-scale product manufacturing. This need must be addressed in my proposed method for OES realization. 9.5. ADVANTAGES OF PRODUCT PLATFORMS A good example of manufacturing efficiency through platform sharing and modularity is Volkswagen/Audi. VW owns Audi, and has conglomerated all of its product lines to be based on the same platforms as its own cars. This means that all standards are help across all the product lines. The body shells may differ, but the locations of engine mounts, transmissions, suspension components etc. is standardized to around 5 different models. This means that all parts are interchangeable, and one transmission can be used in 5 different cars. This also allows for mass customization. Although the body shells are different, a car can be made with a different engine, transmission, and suspension components, making it higher performance, or more affordable. This is done extensively, creating a product to better fit the consumer – giving the maximum number products from a minimum number of internal parts This efficiency through product families and modularity through parts interchangeability has given VW/Audi a competitive advantage over other auto conglomerates, such as GM. By having limited platform sharing, each car requires individual parts, tools, and specialists to perform maintenance and repairs. Other auto manufacturers are catching on to this concept, such as Nissan with its new FM platform, which supports all models of the New 350Z as well as the Infinity G35 coupe models, covering nearly 10 different cars with a limited number of components. I believe this trend will continue to spread into other markets, becoming an essential part of competitive production practices. Many companies have begun following VW/Audi’s lead in product platforms and modularity, and I believe this will continue in the future, and will be required for competitive manufacture of complex products in 2020. I feel that my augmented Pahl and Beitz method will be particularly applicable in these product family situations. Because sections of the design can be saved and re-used with only some parts modified, the design of new products in the family can be based off the an existing design process and calculation, with new modules inserted corresponding to the different assembly plant and materials etc. This need must be addressed in my proposed method for OES realization. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 70 upon the systems engineer designation of Future Tech. The strategic engineer will divide their tasks between determining the abstract high level, and working directly with their technical and systems engineers, they must not become disconnected from the team. The strategic engineers will essentially be the management of Future Tech, providing both leadership and direction. However, they must realize that the life blood of engineering is and always will be the technical engineer, and it will always be this way. The strategic engineers will be those who have excelled in critical thinking, critical analysis, and abstraction, but whom have also show their competency in the core areas of engineering. 10.1.4. Engineering vs. Science First to establish my perspective, and that of Future Tech DME. I believe there is a fundamental difference between scientists and engineers. There are instances where the line is blurred, but I feel that this difference is best described as follows. Scientists look at a problem, see a problem/reaction and respond by determining exactly the cause, as well as exactly what is required to remedy the situation or obtain a solution. The engineering approach is to not necessarily have the exact solution, but one that is adequate, combined with a margin of a safety factor. This is the “ok, it failed. Lets analyze it, find what broke, make those parts twice as strong, and test it again”. This is a bit of an extreme but the point I am trying to make is that we do not really need to know everything about the problem in order to solve it. Engineers care about the solution to the problem, scientists study the problem as a means in itself. There are benefits to a scientific approach, namely a higher degree of accuracy and a better chance of success than a simple iterative “try again and again” approach. However, there are diminishing returns with continued study of a problem. The often academic nature of engineering is fostered by our educational system. It is often that the leaders of an engineering department will have been in academia long enough that they are not accustomed to the rapid development requirements of the commercial world. It is my option and the stance of Future Tech that investigation and solution of a problem is to go marginally beyond that which is required to solve it is to completed. It is the role Scientists and research engineers to study these problems further, and obtain information that will be useful for further iterations or developments, but not critical at the time. This is a very important role, but there must be a balance. Future Tech will consist of a majority of the problem solving engineers, but also a smaller core of research engineers. It is the responsibility of these research engineers to develop the new technology and theories that will progress in OES development by the rest of the engineers. To re-itterate a point I made in my last learning essay, engineering is always ultimately “bending tin” it is what separates us from mathematicians, and “imaginers”. In the commercial world, Future Tech will be fundamentally supported by the development Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 71 “problem solving” engineers. The research engineers will develop the new technology that the strategic engineers will find the most profitable and interesting directions for its implementation and development. Although people within the company have different roles, all employees will still have equal treatment and have the ability to make their thoughts and ideas heard and considered, in the same way as Bungie Studios. 10.1.5. The Bungie Software Corporate Approach Bungie Software is a company that has always done things their own way, and been very successful at it. If you have not heard of them, they are the people behind the recent smash hit X-Box game HALO (voted game of the year by about every major video game publication under the sun, I am sure your children know about it). What most people don’t know is that they started as two guys working in their basement, creating games for the Macintosh, a computer system widely regarded as impossible to produce profitable games for. I have been a fan of Bungie, and their corporate style for the last 10 years, and I have listed the most pertinent points below. These excerpts are taken from a Bungie Soapbox Rant, by Matt Soell, the official "Community Guy / Keeper of the Bungie Way". [22] During Matt’s rant, he asserts 7 key points about the approach that Bungie has taken over the years that has led to their success and difference. These points are: Recognizing and embracing individuality Working only with those you like and respect Eschewing status for teamwork Following your muse regardless of where it takes you Believing relentlessly in your own abilities Busting ass all the time Doing The Right Thing He then states that although these 7 points aren't especially original or groundbreaking; none of them sound insane or heretical. Maybe Bungie just applies them more assiduously than most. I do not want to quote the entire rant here, rather these points on their own make considerable sense, however in light of the application of the engineering spectrum for Future Tech, I consider this section to be the most pertinent. The Flattened Pyramid Bungie never spent any time building an internal hierarchy. The two founders of the company were always in charge, and they always had a vision for where we should go next - but there was no ladder beneath them, just a pool of employees whose contributions were valued equally. There was no jockeying for position, no politics, and no grandstanding for the press. Everyone pulled toward a common goal and no one was excluded by rank from contributing to the games or the company. In fact, Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 72 everyone was expected to contribute in any capacity they could. There's a popular conception of game developers as rock stars too busy "having ideas," driving sportscars and counting money to mix with the common volk, even within their own company. Maybe that's true at other companies, I dunno. It was never true at Bungie. When I first started as Bungie's tech support guy, people used to ask if I'd ever met Jason Jones3. Some of them could barely believe it when I replied that Jason and I, along with the rest of Bungie, sat on the same carpet and ate the same greasy burritos for lunch every day. The basic concept underlying Bungie's corporate structure - strong leaders but no egos - extended to our dealings with the people who played our games. We didn't hire booth babes and talking heads to meet our fans at trade shows. Even if we had that kind of money to throw away, there was something to be said for getting out there and meeting the people who use your product. We all went to the trade shows. We all read the fan sites. We all took tech support calls. Again, this was partly pragmatic - there was always more work than people available to do it - but it had the side effect (one might even say benefit) of keeping customer interactions on a palpable, human level. We never saw our fans as wallets with legs and they never saw us as faceless drones. [22] If Future Tech follows this mantra half as well and has been half as successful as Bungie Software has been, I would consider it a overwhelming success. 10.2. TECHNOLOGIES AND APPLICATION The technologies presented in this section were co-developed by Ashley Ceci and myself. However, their application to my proposed OES relation method are my own and unique to my answer to the Q4S. 10.2.1. Radio Frequency Identification (RFID) One of the largest challenges in engineering products design is the transformation of consumer demands into product requirements. In 2020 companies will need to effectively monitor not only their products, but also the purchasing trends of their customer base in such a way that they can react almost instantaneously to changes in demand. Within the world of product tracking, and purchasing, an emerging technology could usurp the ubiquitous bar code's quarter-century of quiet domination. Radio frequency identification (RFID) tags, which consist of silicon chips and an antenna that can transmit data to a wireless receiver, could one day be used to track everything from soda cans to 3 Jason Jones is the co-founder and CEO of Bungie Software, also the lead creative mind behind all of their recent software releases. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 75 in the details of the process. There will be projects that have a goal of developing new simulations techniques or refining old ones, but in a more general task, a simulation should only be run until the required precision is obtained. This is easy to lose sight of. I know from experience that I can often get obsessed with creating the most efficient program possible, and lose sight of the fact that it is the results and functionality of the program that is important, not the code itself unless it starts affecting the efficiency of the simulation process. For some problems, such as those faced by the NASA engineers regarding hyper-sonic flight and spacecraft launches, the models are both incomplete and inaccurate. The problem is that simulation alone in these instances is a slow approach to developing any useful techniques or launch vehicles. In these situations I believe that experimental correlation needs to run in tandem with theoretical development. For example the thermal conductivity of materials with respect to temperature has been correlated in this way. Physical experiments run in parallel with physics theory have been conducted to support the results of each. Ultimately the empirical data verifies the theoretically derived equations and curve correlations. This model can then be used to predict the conductivity behavior of materials that have similar known characteristics, eliminating the need for more lengthy empirical correlations. This may seem like an obvious direction – combine the theoretical and modeling with the empirical and experimental. However, I believe that over the last few decades the engineering profession in general has moved away from this stance. Many years ago the focus was entirely upon experimentation. People like Thomas Edison would discover and develop new technologies with minimal theoretical work. There were also those such as Nicola Tesla who also utilized modeling, however the very high level of mathematics required limited the scope of what was possible to model. The advent of powerful computers and numerical simulations and models provided us with a new very powerful tool that we have gotten lost within. We have become too obsessed with modeling and often lost sight of the original goal, slowing the development process when it should be speeding it along. This is because simulation does not require nearly the same capital investment as a prototype or experiment, which may be required for the task. Ultimately, we can run as many simulations as we want, but pretty pictures alone won’t get us to Mars. Simulation is a tool and should be treated as one. We have many times the development capability we did 30 years ago, yet development in many industries such as aerospace and astronautics have slowed down. Some of this is due to funding limitation, however this alone does not cover the regression. I believe that this is because we have moved too far towards fine tuning simulations, where we should really be moving ahead towards the overall project goals. This is the position that Future Tech must take, and remain in constant vigil towards. 10.3.2. Applicability of Robust Design to Simulation Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 76 Robustness and simulation go very well together, (some may say as well as beer and pretzels). Simulation is inherently inaccurate, and thus contains uncertainty. Through the modeling and prediction of uncertainty through methods such as the u-sDSP and ucDSP as well as Taguchi’s approach (robustness type I) and robustness type II a superb solution is created. A robust solution is one that is tolerant of changes in internal and external (depending upon the type of robustness) conditions. Because a model can only predict a solution to within a certain range of uncertainty, it is currently the approach to over-engineer the system to a certain extent to account for that uncertainty. A more efficient approach would be to make the system robust enough such that the degree of noise tolerance more than makes up for the degree of uncertainty. This is a similar approach to that of a safety factor, but it is more mathematically rigorous and defined and encompasses all aspects of the system not just a single variable. I believe that the combination of these two concepts is very strong, and supports both the argument for simulation, and the use of simulation as a tool to generate useful results quickly. 10.3.3. Applicability of Information Re-use to Simulation Another integration of ideas is that of the combination of information re-use and simulation. A general rule given during coding is to never write the same function twice, it should be saved as a sub-function and called as so. This can be true of simulations also. When developing a simulation, although the goal of the results should be kept in priority, another goal should be to have a clean model or sub-function that can be archived. This will then allow for the re-use or simple modification of the simulation at a later time. This represents a huge time saving and helps push the idea of digital simulation based design as a superior approach to ad-hoc methods or iterative design. If these models and sub-functions are stored on the central project server, all members can access and use them, giving the entire team great collaborative capability. 10.3.4. The Need For a Definition of Clear Goals A different reason for getting stuck in the simulation phase is the lack of leadership and goals as well as the draw of perfecting a simulation. As example of this is NASA as it stands today. 30 years ago it had a goal – to reach the mood as quickly as possible. They accomplished goal this in a span of 7 years. Today we are developing another moon mission plan. However, despite the same level of funding, and the huge advances in technology and computer development, our projected timescale is 11 years. This is not because we are lazier, or getting there is suddenly harder, we simply do not have the goals or the motivation as we did then, and this makes a huge difference. When without a clear goal, the research is not done until the funding is pulled or expires. This represents the true end a project, the last report will be written and the final presentation delivered. Unless there is a clear goal that is accomplished, there will always be a pull to keep the funding to obtain a few more data points to fit the curve a little better, or to continue the validation process a little longer. If a clear goal is developed from the start, including what happens after that goal is accomplished, this funding drain or low yield work is removed. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 77 10.3.5. The Role of the Corporation in Development Efforts Poor management techniques are another strangling point for development. Examples such as excess personnel on projects, excess accuracy on unnecessary projects, and too much simulation development among along with countless others. The drive to keep people in work is a strong one, however this should not be dealt with through “busy work”. The solution to this is to “reach for the stars”. To keep people employed, some risk must be taken of undertaking new projects. This was shown during the great depression, with the construction of the Hoover Dam and the Interstates. Doing work just for the sake of it will never get us as a race, or a company in the competitive commercial market, anywhere. “Pretty pictures won’t get us to Mars!” This statement is true of simulation, but also fostered in the company structure employed. Because of quarterly status reports, and the continual demand to cut costs, the risks associated with real, effective development are often not taken. It is far easer to request a small investment, and then come back for more later after showing some initial results and presentations than getting a large capital investment that is really required for fast progress. With the risk of failure and the associated cost, it is far easer to show that a simulation failed, than watching your $200K prototype incinerate itself. However, you will learn multitudes more from the analysis of the experiment than the simulation. You can then use that data to enhance the simulation before trying again. I believe the day of the backyard inventor is coming to an end. 50 years ago there were many people calling themselves “inventors” working around the world, developing patents, products, and concepts. Today we have a few huge engineering companies such as GE or Lockheed Martin that dominate and new technological development that is not computer or software based. Part of this is the difficulty of obtaining funding, however it is also true that the problems faced today simply are not going to be solvable by a single lone tinkerer. I believe that the government needs to provide for and support small start up technology based engineering companies to challenge the existing giants, and spur competition and innovation. Currently the government awards too many projects to companies simply because they are a huge company, regardless if their proposal is better or worse than a new, smaller competitor. This is not a good practice in my option. Finally, not all corporations are ineffective with their goals or strategy. The VW Toureg SUV is a good example. VW waited two years before entering the SUV market. Normally this means the loss of market share and a very difficult time penetrating the market once it is formed. However, VW did its homework so to speak, and was designing and very effective, and completely superior SUV to anything else offered. Through efficient design, patient reflection of existing products, and high precision and production manufacturing processes, VW has products a product that is not only superior, but also costs less than almost all of its competitors. This example goes to show that if there is a strong goal, the infrastructure to fulfill it, and a complete, effective design Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 80 Figure 10.1 - Diagram of a weakly integrated system A weakly integrated system is not very complex. The types of interactions that make up a weakly integrated system are shown in Figure 11.1 above. In a weakly integrated system: Hierarchical relationships dominate over any lateral erects Cause and effect and relatively obvious and direct (i.e. each component is only linked to one or two others, and the link is apparent) Implications of design decisions are relatively predictable because of this interaction understanding Influences on, and the implications of decisions are dominated by the effects on the local components, as the impact on the rest of the system is known to be small Risks are dominated by local risks in achieving contributing components Complexity arises from the detail, either the complexity of the individual components or the sheer number of components in the system Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 81 Highly Integrated System Figure 10.2 - Diagram of a highly integrated system A highly integrated system can be very complex. The types of interactions that can make up a weakly integrated system are shown in Figure 11.2 above. In a highly integrated system: Lateral influences dominate hierarchical relationships Cause and effect are not obvious or direct Small causes can have large and far reaching effects Implications of design decisions are far less predictable Influences of decisions are difficult to bound as implications to the entire system are often unknown Risks are dominated by system level risks, with unforeseen properties and widespread consequences Complexity arises because of the system dynamics, the interactions of the system components Measures of Complexity Different measures of complexity and their combination can be used to classify the complexity of a system. I have identified the following characteristics for measuring system complexity: The number of elements in the system The individual convolution of the system component dynamics The number of interactions in the system The strength of the interactions in the system The directness and apparentness of the interactions in the system The time scales of operation of the components in the system The diversity/variability of components in the system Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 82 The environmental interactions with the system as a whole, and its individual components All of these considerations need to be taken into consideration when modeling and simulating a system. They can be used to determine the scope of the models created, as well as the approach to be taken (direct, approximate, empirical, etc.) 10.5.3. Dealing with Complexity in Modeling & Simulation The definition of the system boundaries is analogous to the framing of a problem. This can play an important role in the analysis of a complex system. Depending upon where the boundaries are set, the interactions between the system and its environment will de defined. If the boundary is set well, there will be minimal and weak interactions between a system and its environment, and only the dominant control factors will act as inputs and the resulting outputs a the system. In a similar manner, a system can be sub-divided for modeling, simulation, and analysis purposes. The same approach should be taken of dividing the model along its natural “divisions” of weak interaction boundaries. This will allow for the simple and weakly integrated section of the system to be partitioned from a highly complex component or highly coupled section. These difficult to analyze parts can then be treated separately, easing the analysis and allowing a “black box” approach if necessary as the inputs and outputs to the system are clearly known though the partitioning procedure. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 85 engineering classroom and program interoperability for the students. I believe this need is far more widespread, and useful for a distributed engineering design team, as I will describe below. One of the most important features to retain is the ability to modify the file. Although this is often simple for pure text editors, it gets complicated if there are hyperlinks, graphics, and tables included in the document. The same is true of parametric CAD models, such as SolidWorks or Pro/ENGINEER files. These models contain the geometry and parametric data that define the model, and allow for its flexibility to change features and geometry quickly without changing the rest of the model. Current attempts to allow for streamlined unified data exchange are universal file formats, such as Adobe PDF (Postscript Document Format) for documents or IGES (International Graphics Exchange Specification) for CAD files. However, both of these file formats remove the ability to modify the document. This makes them effective for demonstration or documentation, but no good for exchanging information between team members during design. There are also sharing that must be addressed that extends beyond files of the same program type. For example, parts made in a parametric modeler such as SolidWorks must be able to be imported into a Finite Element Analysis program such as ANSYS or dynamic simulation package like ADAMS without the loss of its parametric editing capabilities. This process also must be connected both up and down stream in the analysis process. This can be accomplished through the communication of programs, or the creation of modular software packages. Communication between programs is done today in the Microsoft Office package. If you create a chart in MS Excel and bring it into a text document created in MS Word, then go back and change a data entry in Excel, the chart in both the Excel and Word documents are updated automatically to reflect this change. Modular software products are also being developed. For example Pro/ENGINEER now has a module package that can be purchased to allow for FEA for parts. This allows for one central program with functionality added through different modules. The point of this two way exchange is best illustrated in an example. If an engineer had designed a connecting rod in SolidWorks, imported it into ANSYS for analysis, and then discovered they had to make changes, whether they made those changes in ANSYS or in SolidWorks, the model would be updated across the board, in any other documents or programs it was being used in as well. This interoperability of programs can then be extended to be integrated across the World Area Network (WAN). This will allow for changes made by one design team to be reflected in the files of their partner teams located around the globe. Currently interoperability is a large bottleneck in the embodiment design process. During the conceptual design phase there is a lot of design freedom, however in order to make the design more concrete during the embodiment phase analyses of the design must be performed. The slow process of converting file types from part creation to analysis wastes both time and money. In order to perform an evaluation and selection of design variants, this analysis must be performed multiple times on different variants. Therefore Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 86 this bottleneck is multiplied. In order for this time for the evaluation and analysis during the embodiment phase to be reduced, program interoperability must be accomplished. 11.4. HARDWARE COMMUNICATION In order to keep all distributed team members operating at maximum possible efficiency, all hardware must be able to communicate effectively. This again builds off of the World Area Network concept, allowing for any electronic device to be connected to any other in the world. This will require a common communication protocol, similar to Blue Tooth that was developed recently, to allow all PDAs, cellular phones, computers, and digital notebooks to exchange data. This will allow for an engineer on a project to always have access to whatever information they need. Although most work will be done using the digital notebooks or computers, if the engineer in out of the office, they can use their PDA and still retain almost all of their functionality. 11.5. INFORMATION STORAGE In my envisioned world of 2020 information mass accumulation and sharing will occur. Currently there are many emerging and successfully established sites that deal with the accumulation and sharing of information. This information does not have to be centrally based, rather it could be a complied list of off-site resources, such as the site Lexus- Nexus or other research sites. I also believe that information will be stored both on central servers, such as commercial data warehouses, selling access to their information, along with subscriber or free services that are linked lists of data, operating in a similar fashion to the Napster or Gnutella sharing systems. The impact of these data warehouses is enhanced by the Global Area Network, a satellite based global access system, allowing users to connect in any location. Therefore remote engineers will never be without any data they need, as they will have instant access at all times though their digital notebook, PDA, or mobile phone. 11.6. PROGRAMMING WRAPPERS One of the most difficult aspects of any digital design support system is the implantation of the modular interface with the computer models that run the computational analyses. Systems such as Model Center require a “wrapper” to be programmed that identifies the inputs and outputs of the files, and the generation of a batch file to execute. This is slow and inefficient; I propose the following system for X-DPR 2.0. Macro programming is available in many programs such as Microsoft Excel. What is lesser known is that operating systems (OS) also have scripting abilities that record and repeat the commands executed for any of the programs that are run on them. This means that universal macros can be recorded using the graphical user interface of a program, not writing command lines. This is augmented through the addition of a macro pallet. This pallet allows the user to specify which inputs are controls, noise, or response variables in Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 87 the system. This finished wrapper is then stored as a module to be run by X-DPR 2.0 in my proposed method. In order to edit these modules, the macro recording can still be edited though a standard command line interface, but also through re-recording set portions of the script and changing the input/output variables. Finally, for cross platform capability a translation program must be written to change the OS level macros between the three major operating systems to avoid having a module only work on one possible system. This style of wrapping programs is far more efficient that anything that is used today, and actually makes the use of simulation and modeling modules in X-DPR 2.0 a very feasible system. 11.7. MULTI-PHYSISCS One of the hardest types in interaction to deal with is when interactions between components spans several disciplines of physics. This means that a multi-disciplinary team is required for analysis, and many discussions must occur in order to correctly model these interactions, as knowledge must be passed between team members in order to understand all physical principles active in the interaction under consideration. Computers and information databases can be used to alleviate this challenge to a degree. If these interactions are modeled in a general but highly encompassing manner, they can be applied as a universal reference model. This means that each expert in their domain can apply their know part of the model, and the expert system will determine the interactions between the different disciplines of physics. This is augmented at Future Tech though the overseeing of the process by a systems engineer, who can determine if the interactions between disciplines is working correctly, although their depth of knowledge in each area my not be as complete as the individuals creating the specific models. Finally, the use of mathematical analysis, such as statistical analyses and design of experiments methods should also be considered during the analysis process. The inclusion of these disciplines will allow for more efficient analysis of the system, as well as the accurate modeling of unknown phenomena that cannot be modeled in any other way through statistical means. Programs such as FEMLAB, the model multiple physical phenomena are a step in the right direction, however there is still no reference model for physics interactions. This additional database and functionality would be a strong addition to this package, as well as statistical analysis and interactions tools, and would bring this software package closer to fulfilling my proposed requirements for 2020. X-DPR 2.0 will include a database of physical interaction modules. This means that it is possible to interface the inputs and outputs of FLUENT and ANSYS, programs that normally cannot communicate with each other. The heat transfer analysis from FLUENT will be referenced to the original CAD geometry to generate the resulting thermal stresses using the multi-physics module. This module will then feed into the ANSYS model, to analyze the resultant strains and displacements of the system. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 90 Figure 11.2 - Manufacture DSP template example Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 91 Figure 11.3 - Maintenance DSP template example Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 92 11.9. DESIGN PROCESS LAYOUT Looking at the Bruce Brown tutorial, and HTML pages in general, there is a "zooming" out to higher levels of information, or in to more detail on a section. This is possible with the digital storage of the proposed method on the X-PPR system. The user can look at the overall structure of the method, on the phase level, zoom into a single phase, either in the core process or a concurrent phase, zoom in again to find a single step of that phase, and finally go deeper another level to all the relevant files, scheduling, and team roster for that individual step or sub-step. I feel this digital format is a far more efficient and logical organization that a purely linear and sequential design process layout. 11.9.1. Central Project Base My concepts regarding an advanced integrated hardware communication system and World Area Network facilitate the use of a single remote access server. This allows for all engineers on a project to work off of a single database for storage of all files. This means that any engineer will always have access to any files that have been generated or placed on the server by any team member. This will streamline operations by avoiding waiting for files to be sent individually by individual members when needed. 11.9.2. Automated Bookkeeping One of the biggest difficulties, particularly with a large distributed design team, is keeping all the material and information regarding the project in order. My concept of automated bookkeeping is augmented by my concepts of the improved human-computer interface. This interface means that all conversations can be digitally transcribed to text for storage and retrieval. All messages will occur through the server also, logging and cataloguing all e-mails and organizing them automatically by content and its relation to specific parts of the design. This will mean that all information regarding a design is easily located, from CAD files and video conference recordings to engineers comments and concept sketches, including cross referencing to interconnecting parts or standards to be followed. Through the automation of this system, less time will be spent on organization and locating information, and all of the engineers’ efforts can be spent on their design tasks. This includes work such as progress reports, which will be completed automatically, which can normally take up a large quantity of time in an inefficient large corporate environment today. 11.9.3. Automatic Version Control Another issue that will be tackled by the central server is versioning control. This system is enabled by my concept of the interoperability of software programs and files. One problem often encountered today with distributed teams is the updating and changing of a design. For example, if it is agreed that a design needs to be updated, and two engineers take on different aspects of the part, re-integrating the design can be difficult as the upgrades my be incompatible. There is also the problem of keeping all design team members using the latest version of the model, documentation, etc. The versioning Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 95 12.1.2. Updated Pahl and Beitz The later renditions of Pahl and Beitz included “hooks” in which things such as manufacturing, design for assembly, and even modularity and scalability could be considered. This system is now somewhat modular, and is best thought of as a slot modular system. In this system everything has its place and cannot be interchanged, however certain components can be removed or included, allowing the system to be bettor tailored to its specific use. This system is shown below in Figure 12.2. Figure 12.2 - Modular slot system This system is an improvement over the original set Pahl and Beitz system, however it is still inadequate for the development of OES. 12.1.3. My ME6101 Augmented Pahl and Beitz In ME6101 I augmented the Pahl and Beitz method by breaking it into modular components. I then further augmented it by adding a large pool of modules that could be considered and added to any part of the process. This system was quite flexible, and could be tailored to the design of products, processes or services. It also could consider advanced relation techniques such as MEMS of biotechnology manufacturing through its flexible approach. However, through further analysis of this method with respect to the OES paradigm, I realize that this system uses a modular bus system. The bus in this method was the Pahl and Beitz system. The Pahl and Beitz system acted as a backbone, providing a structure into which the design process was fleshed out. This is represented in Figure 12.3 shown below. Figure 12.3 - Modular bus system Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 96 This system is quite advanced, and have a high degree of flexibility. The modules that I included in the method were also good, addressing different needs of the base Pahl and Beitz method. It is for this reason I have considered it as a base method for my proposed method. However, there is one obstacle to complete modularity and flexibility in this system, the bus itself. In order for the next step of modularity to occur, the bus must be removed. This is my goal for my proposed ME6102 answer to the Q4S. 12.2. MY OES REALIZATION METHOD The concept I propose for my OES realization method is one of a sectional modular system. In this system there is no longer any framework for the modules to be placed within. This allows for greater flexibility, but also presents another challenge. This section system is shown below in Figure 13.3. Figure 12.4 - Modular sectional system The challenge presented in this kind of modular system is which modules to include and what type of interface to use. In this example the interface is simply the cable connection, however the choice of modules is no longer straightforward. The user must decide if they want a monitor, a printer, a CPU, or if they need a modem or keyboard. This means that the user must have a method or system to selecting and connecting the modules. For small systems this is simple, but for larger systems with many modules this is very complex. Both the issues of interfaces as well as system configuration are dealt with in the core augmentations of my proposed OES realization method. Nathan Rolander ME6102 - Question for the Semester Spring 2004 Page 97 13. METHOD FLOWCHART Because of the lack of a defined process to follow, I feel that the augmented PEI diagram shown in Figure 13.1 is the best visual representation of the overall structure of my proposed method. Information P E I P E Pr od uc t D es ig n D es ig n Pr oc es s D es ig n Product Requirements Process Requirements Product Specific Information Process Specific Information Time Hie rar ch y Conceptual Design Embodiment Design Detail Design Clarification of Task Conceptual Design Embodiment Design Detail Design Clarification of Task Function Behavior Structure Robust Design Reliability Design Axiomatic Design T1 D2 D3 D4 T5 T1 D2 D3 D4 T5 Figure 13.1 - Flowchart for proposed OES realization method This representation is based upon the work presented in Jitesh’s PhD proposal [6], however I have modified it and augmented it to reelect my own strategies for organization and how I think the design process should be carried out as embodied in my answer to the Q4S. This PEI diagram is essentially a fusion of two PEI diagrams, one for the designing the design process (meta-design) and another for designing the product. The information segment of the PEI diagram is shared between both of these design processes, as this information regarding both is essential to determining the next step of the design process. The third dimension of the augmented PEI diagram represents the Concurrent Processes High Abstraction Level Product Requirements Process Requirements Low Abstraction Level