|
|
|
|||||||
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
Predictive Process Engineering |
|
|
|
Goal: |
|
By 2005, to support the industry need for first part correct manufacturing, establish an industry-accepted and widely adopted integration framework for sharing predictive knowledge about machining processes and resources with engineering and control systems using a standard semantic-based process representation and validated physics-based models for milling and turning. |
|
|
|
|
Needs Addressed: |
|
Annual U.S. expenditures on machining operations total more than $200 billion, or about 2 % of the Gross Domestic Product (GDP). Total expenditures on all mechanical manufacturing processes, including forming, and stamping, would be substantially larger. Due to the inherent complexity of manufacturing processes, process development is often ad-hoc and empirical. Process parameters, such as machining speeds, feed rates, and tool selection, are typically chosen by costly, trial-and-error prototyping, with the resulting solutions often sub-optimal. A survey by the Kennametal Corporation dramatically demonstrates this, finding that U.S. industry chooses the correct tool less than 50 % of the time, uses cutting tools at their rated cutting speed only 58 % of the time, and uses cutting tools to their full life capability only 38 % of the time. These sub-optimal practices are estimated to cost U.S. industry $10 billion per year. Pressure from international competitors is driving industry to seek more sophisticated and cost-effective means of choosing process parameters through modeling and simulation. Optimal manufacturing performance requires sufficient understanding of the impact of individual parameters on the various levels of the control hierarchy, from the shop floor to the overall enterprise. A principal barrier to reducing these inefficiencies is the lack of access to validated, physics-based models of the manufacturing processes when key engineering decisions are made. While there has been significant progress in the predictive simulation of low-strain-rate manufacturing processes such as forming, rolling, or drawing, there is a need for better predictive capabilities for high-strain-rate processes such as machining. The state-of-the-art in predictive modeling of machining operations is severely limited because measurement capability and materials characterization are severely lagging model development. In other words, current models give impressive qualitative results, but data to validate these results is nearly nonexistent. Commercial models of machining operations are currently proliferating. However, without focused metrology efforts to understand and characterize the processes fully, it is likely that industry will lose faith in such models due to difficulty obtaining input data and inability to validate results. A second barrier to reducing manufacturing process inefficiencies is the lack of simple mechanisms to enable exchange of process information among manufacturing systems. Market analysts at Gartner Group Inc. estimate the costs associated with the exchange of process information among manufacturing applications in U.S. industry at approximately $2 billion per year. Through the availability of standards and methodologies to provide a rigorous foundation for representation, exchange, and integration of process-related data, an estimated 15 % to 20 % cost reduction could be achieved, which translates into a savings of $300 million to $400 million per year. U.S. manufacturers and industry groups identified these problem areas through numerous public forums and publications. In particular, the 1998 Integrated Manufacturing Technology Roadmap (IMTR) identified a number of industry priorities and critical capabilities that provide direct support for this program. Specifically, the IMTR identifies science-based manufacturing, first part correct, intelligent process advisors, and robust process models as key industry needs.
|
|
Technical Approach: |
|
The vision of "first part correct" demands a different approach in many areas of manufacturing engineering. This program will build the foundation to enable a new paradigm for predictive process engineering and science-based manufacturing. This foundation will be built upon a science-based understanding of material removal manufacturing processes, advanced process metrology methods, valid analytical models to predict process performance and optimize manufacturing decisions, and rigorously-defined representations for manufacturing process information. The new paradigm will be a shift from classical feedback quality assurance and optimization to model-based feed-forward process design and quality control. Process metrology methods will provide new understanding and data for process characterization to develop and improve the predictive model, the process specification, and the manufacturing process itself. Product and process designers will have knowledge of and access to process specifications, manufacturing knowledge, and predictive process models to generate product and process designs seamlessly to produce the correct part the first time. To meet the needs, a number of areas must be addressed in parallel. Specific activities in each area will be adapted to take advantage of collaborations with industry, academia, and international organizations. These collaborations will inform and guide program efforts throughout the course of the work to ensure the usability and applicability of the program outputs. Key technical areas are: (1) Development of measurement methods to provide in-situ, real-time, or process intermittent information about milling and turning material removal processes. Program activities in the area of process metrology address such topics as machine dynamics, thermal measurements during material removal, dynamic material response, and experimental data to evaluate models of material removal processes. Related measurement methods and measured quantities are also within the program scope. The absence of materials characterization data and an appropriate materials model represents a major limiting factor. Therefore, one program emphasis will be development of appropriate methods to generate materials response data under the high strain rate conditions needed for simulation of machining operations. (2) Based on the process metrology methods and resulting experimental data, development and validation of physics-based models of milling and turning processes to support next-generation industry priorities in analysis, planning, optimization, and real-time control. With the increased understanding of the manufacturing processes, physics-based models of milling and turning processes will be developed and/or extended from existing capability. Process models have been published that describe the transitions to segmented chip formation in turning operations and the non-linearities encountered in milling. Next steps include evaluation and application of such techniques as receptance coupling to predict dynamic behavior of systems based on the performance of individual sub-systems. Process models for milling operations will be developed and extended to predict process behavior and establish optimum cutting parameters. Creation of science-based process models provides a number of benefits, including the ability to understand and predict the effect of parameter changes for situations that are outside the currently available empirical data. Validation of the process models using empirical data is a key element in this work to give confidence in the performance of the process model, both within and external to the regimes of "known" data. (3) Development of standard representations for manufacturing process information, standard interfaces among applications, and standard integration methods to enable seamless integration of process-related applications for extended enterprises. There is strong evidence for the industry need for a standard, rigorously defined data representation for manufacturing process information. The challenge of system interoperability is evident in the growing complexity of manufacturing information and the need to exchange this information among various software applications. Interoperability is hindered, however, because these applications often associate different meanings with the terms representing the information being exchanged. An exchange language is needed that allows for the representation of all relevant manufacturing process information independent of any specific application, and that also provides unambiguous and rigorous semantics of the concepts being represented. This need is addressed primarily through the Process Specification Language (PSL) project that has been recognized and accepted as the basis for the international standard ISO 18629 currently in development. The PSL consists of a modular, extensible data model that captures the concepts required for process specification. These data representations and integration methods are necessary to enable complete use of manufacturing knowledge and process-related data throughout the product lifecycle. (4) Development and demonstration of methods for effectively using predictive process models for analysis, planning, optimization, and real-time control. To emphasize and demonstrate integration of the various program components, specific efforts will be directed toward implementation of the standard process representations and predictive models with application to typical industry challenges. In general, the process metrology activities provide new understanding and data to improve the process modeling activities; the process representation and integration methods activities combine with the process modeling activities to enable the process application activities. Results from the process metrology, modeling, and representation efforts will be used to implement a prototype process integration framework to share process knowledge with engineering and control systems. These efforts provide a level of validation for the program outputs and ensure sufficient usability and functionality of results. Demonstrations of typical process applications based on the concepts of predictive process engineering will also generate the industry interest and support that will be necessary during standardization activities.
|
|
Patents/Licenses: |
|
Renewable Polishing Lap (1999) The renewable polishing lap virtually eliminates the substantial time and expense for "renewing" traditional laps. This new manufacturing process technology enables manufacturers to drastically reduce the costs for a wide variety of grinding, lapping, and polishing operations. The U.S. Patent Office has awarded U.S. Patent Number 5,897,424 "Renewable polishing lap" to Chris Evans, MEL scientist, and Bob Parks , MEL guest worker, for the development of the renewable polishing lap. The renewable polishing lap uses a textured substrate over which a thin film is placed. The substrate provides the geometry of the lap and a localized texture, which depends on the film thickness, properties, and means by which the film is deformed over and adhered to the substrate. When abrasives are distributed over the film, the substrate is protected during the lapping or polishing action. Since the film protects the substrate from the harsh abrasives, the substantial time and expense typically required for renewing the lap is eliminated. The renewable polishing lap technology has been licensed to Rodel Inc., who is developing the technology for use in photomask blank polishing. Device for Estimating Best Speeds in Milling (2000) Optimum speeds in milling depend on the dynamics of the entire system. Frequently it is more effective to increase tool speed – in the presence of chatter – than adopt the intuitive approach of reducing the “severity” of the cut. Matt Davies and colleagues have applied for a provisional U.S. patent for a device which uses once per revolution excitation of the cutter and feedback from a simple sensor to evaluate optimum speeds for a specific tool configuration. The device – which could be mounted in the work volume much like a tool setting station – offers the potential for much improved machine productivity through cutting at the highest stabe speed available for the specific tool/spindle combination. |
|
Send questions or comments to Webmaster. |