A Framework for Distributed and Virtual Discrete Part Manufacturing
Howard M. Bloom
Neil Christopher
National Institute of Standards and Technology
Abstract
In response to marketplace demands, U.S. manufacturing enterprises are forming strategic partnerships with diverse companies, geographically dispersed throughout the country, that are most effective at carrying out needed manufacturing operations. The manufacturing software applications used by the companies in these operations must be integrated in a manner that allows for efficient flow of information between applications. A national consensus-based software infrastructure framework for manufacturing operations will allow this integration to occur. This framework will specify how manufacturing information, communications protocols, and services are to be represented and used. The project described in this paper will prototype a framework of manufacturing software systems that supports many of the operations for discrete parts manufacturing. This prototype will be based on emerging technologies that are likely candidates for international standardization. It will be the experimental basis for further development of standards and specifications. Key standards such as STEP and CORBA will be utilized in the prototype. The project will also develop test methods that demonstrate how applications fit interoperably into the framework.
The project is supported within the National Institute of Standards and Technology (NIST)'s National Advanced Manufacturing Testbed (NAMT) where the next generation standards and measurements needed for information-based manufacturing will be explored. The project will leverage the work of NIST's industrial partners in the National Information Infrastructure Protocols (NIIIP), the Technologies Enabling Agile Manufacturing (TEAM), and the Sematech CIM Framework programs.
1. Introduction
For U.S. industry to be competitive in today's global manufacturing environment, it is critical that companies form strategic alliances with partners that can help them manufacture products in a timely and cost-effective manner. These partners, distributed throughout the nation, are using information-based enterprise applications such as design engineering, manufacturing engineering, production engineering, and fabrication software. To successfully use such alliances, software supporting the distributed applications must be available. However, it is not enough to be able to physically connect these systems. Companies need the applications to be interoperable so that as new technologies, or different vendor implementations of the same technologies, become useful, the new application software can work together seamlessly with the existing applications. This paper discusses the use of an enterprise framework, which is a software infrastructure that creates a common environment for integrating the applications and sharing information in a given manufacturing domain, as a means of achieving interoperability.
In section 2, the National Institute of Standards and Technology (NIST) role in supporting industry is discussed in terms of the unique functions of NIST that are being used in projects that involve the development and testing of information-based manufacturing systems. Section 3 discusses a new program at NIST, the National Advanced Manufacturing Testbed (NAMT), that is designed specifically to tackle the types of problems addressed in the Framework project. Section 4 discusses the Framework project itself in terms of the technology and products to be produced. Section 5 discusses the partnerships with the industry consortia and standards bodies. Section 6 describes the NAMT approach to developing specifications and test methods in support of such needs as interoperability for manufacturing systems. Section 7 summarizes the efforts on the project.
2. NIST Role in Standards and Testing
The National Institute of Standards and Technology (NIST) is an agency of the Department of Commerce. Its mission is to promote U.S. economic growth by working with industry to develop and apply technology, measurements, and standards. The Manufacturing Engineering Laboratory (MEL), one of seven technical operating units at NIST, serves the whole manufacturing enterprise in a systems view, including each of the manufacturing sectors, and the range of manufacturing technologies, with special focus on the mechanical manufacturing industry. A major technical thrust within MEL is in manufacturing systems integration where the effort is to facilitate the development of interface standards that enable a fully integrated set of manufacturing systems that underlie a complete manufacturing enterprise to interoperate.
MEL scientists and engineers track emerging technologies which have advanced from fixed automation of the early twentieth century to the flexible manufacturing in the 1980s to information-based manufacturing in the 21st century. The work of MEL is then to facilitate the development of standards for the applications that support the changing technologies. In the thrust area of manufacturing systems integration, the effort has grown from manufacturing protocols in the 1980s to manufacturing data exchange to integration frameworks for the next century.
MEL's laboratory role in working with industry usually involves providing a testbed that can be used with industry for solving technical problems, but with MEL's contribution primarily in the area of developing testing methods and testing services for key standards. Some of the important areas of software testing are human-machine interface, software diagnostics and performance, security, mathematical software, and knowledge bases. There are various types of software testing, such as: conformance, performance, assurance, and relevance. Software testing requires a reference (e.g., a primary standard) which may be a specification, data set, or reference implementation.
Two examples of testing services are provided to demonstrate how MEL supports U.S. industry in the area of manufacturing software. The Algorithm Testing and Evaluation Program for Coordinate Measuring Systems (ATEP-CMS) evaluates the mathematical performance of data analysis software used in CMSs. Tested software is treated as a filter that transforms point coordinate data into feature parameters according to a defined transfer function. MEL evaluates the accuracy of the filter under conditions typical of those found in industrial practice. MEL independently compares the output of the software under test to predetermined corresponding reference values.
MEL has teamed with the Industrial Technology Institute (ITI) to provide a means by which vendor implemented STEP (Standard for the Exchange of Product Model Data) products can be objectively measured against the standard. MEL is developing a set of value-added software tools for use by vendors during product development. These tools are part of a modular system, the STEP Conformance Testing System (CTS), that has two elements: a test system which integrates various test tools and administers the actual tests, and a set of tools for generating test suites for each application protocol (i.e., an information model describing the data to be exchanged with a given manufacturing application such as process planning) which are used in the testing process. This system is available over the Internet to allow vendors to test products before they are released, as well as to allow end-users to assess the ability of the products to interoperate in an industrial context.
For both these testing programs to be successful, it was critical for NIST staff to play key roles within the standards community for the development of the underlying methodology required for testing the standard. In the case of ATEP, NIST has had a lead role in the draft national standard, ASME B89.4.10 - Methods for Performance Evaluation of Coordinate Measuring System Software. In the case of the STEP CTS, NIST has played a lead role in the development of ISO 10303-31, Conformance Testing Framework and Concepts, as an International Standard, and ISO 10303-32, Requirements for Test Laboratories and Clients, as a Draft International Standard.
Over the past 15 years, MEL has had testbed programs to address a variety of manufacturing integration issues. The Automated Manufacturing Research Facility was begun (with support from the Navy's Manufacturing Technology Program) in 1982 with the idea of developing and testing the standards and measurements needed for a flexible manufacturing environment of machine tools, robots, coordinate measurement machines, and material handling systems. Information technology related issues that were examined included distributed heterogeneous manufacturing databases, manufacturing network protocols, architecture for hierarchical control systems, and manufacturing data representation. In 1989, the National PDES Testbed program was initiated (with support from DoD's Computer-Aided Acquisition and Logistics Support Program) to facilitate the development of STEP. Two very important products of this program have been the STEP Conformance Testing Service described above and the Application Protocol (AP) Development Environment which has greatly facilitated the efforts to implement APs. In 1996, the National Advanced Manufacturing Testbed (NAMT) program was initiated to help enable the development of standards and measurements for information-based manufacturing that is distributed and virtual.
3. NAMT Program
The information revolution is having a profound effect on how manufacturing is done. MEL is prepared to develop new tools for this new kind of information-based manufacturing, and to find new solutions to the standards and metrology issues raised by this new way of doing things. The NAMT program will be a showcase for the future of manufacturing, demonstrating how machines, software and people can be networked together to achieve interoperability at all levels of a manufacturing enterprise. The NAMT contains a facility in which scientists and engineers from industry, NIST, academia, and other government agencies work together to solve measurement and standards issues in information-based manufacturing and develop the needed tests and test methods for industry that are part of NIST's mission.
The projects within the NAMT are characterized by: (1) collaborative industrial partners, (2) leading edge technologies, (3) development or use of advanced measurement technologies, (4) development of standards for manufacturing applications, (5) use of information technology, and (6) tasks and processes at multiple sites on-line.
The results of the NAMT will be metrology techniques, interface standards, and other infrastructural technologies and standards. The NAMT will accelerate efforts to develop components of a common information infrastructure to manufacturing, extending the capabilities of advanced computing, communications, and control technologies to multiple manufacturing applications and domains. It will leverage pools of manufacturing resources, including physical facilities, equipment, expertise, and software.
There are presently four projects within the NAMT. The Nanomanufacturing of Atom-Based Dimensional Standards project is focused on the distributed design, fabrication, and use of nanometer-scale dimensional artifacts. The Development of Machine Tool Performance Models and Machine Data Repositories project is creating a virtual manufacturing and inspection model and data repository to reduce the need for prototyping. The Characterization, Remote Access, and Simulation of Hexapod Machines project is aimed at measuring and extending the capability limits of parallel machines, and developing & evaluating techniques for open architecture control interface standards. The Framework for Distributed and Virtual Discrete Part Manufacturing project, the subject of this paper, is described in the next section.
4. Framework Project
The NAMT Framework project is implementing a distributed manufacturing system infrastructure based on interoperability specifications emerging from industry. From a systems point of view, the implementation is a "framework" in the sense that it describes classes of manufacturing applications, defines the interfaces between those applications, and identifies the interoperability services that support transactions across that application interfaces. From an organizational point of view, the project provides a framework of specifications in that it describes a concept of operations, the relationships between the specifications, and a manufacturing context against which the specifications are applied.
The NAMT Framework project will further the specifications of industry consortia and standards organizations as well as those specifications developed as products of NIST's laboratory programs. The Framework implementation will be used to provide validation of the specifications with respect to a set of manufacturing operations. The goal of the project is to accelerate the development of these specifications, to work on the standardization of specification elements, and to use the validation testing as an industry point of reference in the consensus building process.
The NAMT Framework project is implementing as a test scenario the manufacturing operation automated inspection of mechanical parts. This operation entails entering an inspection job into the system, scheduling and task allocation by a factory controller to an inspection workcell controller, using an object oriented database to capture production information, and a product data manager to maintain product and process programs and documents. A graphical user interface will be used to monitor status and intercede in operations. This test system, consisting of a factory controller, inspection workcell controller, product data management system, object oriented database, and graphical user interface will be the experimental apparatus with which to analyze the specifications and protocols used for manufacturing systems integration.
The Framework implementation will be modified over time to enable researchers to analyze alternative and competing standards against various manufacturing operations. The first set of modifications will add machine control of a hexapod machining cell as well as control of an advanced turning cell to the existing inspection operations. Experiments using this enhanced capability will exercise scheduling, job task decomposition, machine messaging, data management, information models, and interoperability specifications. Later enhancements and additions to the NAMT Framework implementation are likely to include simulated machines and upstream software such as automated process planning and mechanical design applications. Specifications supporting framework to framework integration in support of manufacturing supply chains will be evaluated in subsequent implementations of the Framework project. Throughout the project expertise gained will be banked against the need to develop standards conformance tests, tools, and services in the future.
Initially, the Framework project is evaluating the Sematech Computer Integrated Manufacturing Applications Framework (CIMF) production information models for suitability in the discrete mechanical parts production domain. The CIMF specification is an object-oriented manufacturing systems integration framework for semiconductor fabrication. This evaluation includes the comparative analysis of these models with the NIST-developed models of manufacturing information for planning and production systems (a product of the Manufacturing Systems Integration project within the AMRF). Other elements of the CIMF under analysis against the discrete mechanical parts production domain include the application object interfaces for routing, scheduling, material management, specification management, and machine control. Additional machine control related specifications and standards under development and implementation include NIST's Enhanced Machine Controller (EMC), ISO Manufacturing Message Specification (MMS), and Object Management Group (OMG) Real-time Special Interest Group (RTSIG). The planning and production systems models will be enhanced with the specifications being developed by other industrial activities in an effort to reach consensus.
Access to production information and factory status is a key element of the NAMT Framework. The Standard Data Access Interface (SDAI), a part of ISO 10303 STEP (Standard for the Exchange of Product Model Data), is the specification under evaluation for access to production information in the NAMT Framework project. However, since the Common Object Request Broker Architecture (CORBA) object services from OMG are part of the NAMT Framework project, the SDAI/CORBA binding under development as one of the National Industrial Information Infrastructure Protocols (NIIIP) program protocols will also be included as part of this effort. In the NAMT Framework effort, product data exchange will be evaluated in the context of production processes. Product Data Management (PDM) systems manage product and production data alike. PDM interfaces are under intense scrutiny by national industrial consortia as well as standards organizations. The NAMT Framework project will explore PDM interface issues using the Sematech CIMF specification and will relate its findings to other programs so as to converge on a national solution. A new specification that the NAMT Framework project personnel and industry experts will initiate is a model of information relating to automated inspection processes for mechanical components. This new model is the STEP Application Protocol for Inspection Data (AP219), an effort registered with ISO but not active at the present time.
The NAMT Framework project is committed to helping industry achieve its manufacturing information technology goals. The NAMT Framework project will consider itself successful if its contributions to the SDAI/CORBA binding protocols leads to a standard. It will consider itself successful if the models of manufacturing information for planning and production systems become a standard and are used as a basis for exchange by software vendors. The framework project will measure its contributions to the standardization of PDM interfaces against the time to standardize those interfaces. It will measure its success in development of the STEP AP 219 by the level of industry involvement in the effort. In fact each phase of the NAMT Framework project will establish metrics in line with the criteria established in the NIST report "Setting Priorities and Measuring Results at the National Institute of Standards and Technology."
5. Framework Collaboration
The NAMT Framework project staff is participating in the development of specifications and protocols with several national industrial consortia and standards bodies. This initial set of consortia and standards bodies will change over time as work proceeds across the industry requirements for distributed manufacturing systems integration standards.
Key standards partners include the International Organization of Standardization (ISO), the Object Management Group (OMG), and their respective liaison organizations. Within ISO, there is the ISO 10303 - STEP (Standard for the Exchange of Product Model Data), ISO 14258 - Industrial Automation Systems Concepts and Rules for Enterprise Models, and ISO/TEC 9506 - Manufacturing Message Specification (MMS). Within the OMG, there is the Manufacturing Task Force (MfgTF) and the Real-time Special Interest Group (RTSIG).
Industry consortia partners include: Sematech, the National Industrial Information Infrastructure Protocols (NIIIP) program, and the Technologies Enabling Agile Manufacturing (TEAM) program. Sematech, a semiconductor industry consortia of eleven corporations, has created the CIMF specification. NIIIP is a consortia of eighteen organizations working together to produce the NIIIP Reference Architecture, a set of protocols required to establish and operate a virtual enterprise. TEAM is a joint industry government program led by the Department of Energy national laboratories. TEAM is modeling manufacturing processes and is creating applications to support those processes.
In addition, the Framework project is also benefiting from its interactions with such organizations as the Software Engineering Institute (SEI), PDES Inc., NCMS, and CAM-I.
6. Specifications and Testing
The key contribution that MEL can make to the development and implementation of a consensus framework is to provide a set of specifications that have been tested and shown to work in a given manufacturing domain. The project will utilize as much as possible existing standards and specifications and develop new ones only if necessary. The project will incorporate the work of its collaborators into the specification development. For this work to get accomplished effectively, a specification methodology has been developed. This methodology involves the use of a rigorous approach to the development of successful consensus-based specifications that includes a close partnership between NIST and the industrial community. The Initial Manufacturing Exchange Specification (IMES) is meant to be a pre-standard specification addressing interoperability deficiencies of manufacturing software systems. It can be an interface between user and software, two applications, or an application and a data repository. The IMES has been defined with a goal of being accepted by the manufacturing community as an authoritative exchange specification for a given manufacturing function.
There are seven phases defined for an IMES that allow for interactions with industry and the standards community at each step. These phases have been defined based on the manner in which MEL works with its customers - manufacturers, vendors, standards bodies, universities - to achieve its goals. The first phase involves identifying an actual industry need for a specification. This will be accomplished through workshops and reports. The second phase results in a requirements analysis where all existing specifications or standards are examined for relevance. The third phase is to develop a technical solution, if necessary, which results in a published draft specification. The specification would be circulated to a broad industrial-based audience for comments. The fourth phase involves implementing a set of prototype applications to demonstrate the viability of the specification. This phase requires involvement of both vendors and end-users. The specifications will be validated through test procedures. The next phase is the validation by industry. The specifications would be used in pilot applications in real industrial environments. The next phase involves transitioning the technology to industry. The final phase is to initiate the standardization process by submitting the specification as a working draft to the appropriate standards body. These phases are not necessarily sequential. However, each phase has a set of deliverables that must be satisfied to proceed formally to the next phase.
There are several key issues to be worked through. First, it is necessary to get industry consensus on the IMES concept. An example of a similar successful procedure was the original development of the Initial Graphics Exchange Specification (IGES) in which MEL created a large user group that developed the specification and then submitted the results to a formal ANSI committee for final balloting. Second, the IMES must be able to build upon national program specification products such as the Sematech CIM Framework. The framework project should make every attempt to identify and incorporate existing specifications rather than developing new ones. In addition, the process should be able to resolve differences between specifications developed in different programs that might not be consistent with each other. Third, the relationship between the IMES and formal standards development process must be established. NIST has a close working arrangement with ANSI and all the accredited standards bodies. It is important that the IMES is not perceived as a competitor to the formal process, but rather as a facilitator. In terms of STEP and PDES (the U.S. version of STEP), NIST plays a key role as the secretariat for the international program and the technical management of the national program. In terms of CORBA, NIST has an active role as member of Object Management Group and specifically on the manufacturing task group. Last, experience must be gained to effectively gauge the time required to proceed through the IMES process. The concept is that the IMES process should accelerate the development of standards, not slow them down.
7. Summary
A project has been described that will develop and test a software framework to support distributed and virtual discrete part manufacturing. Over the next four years, various levels of a consensus framework will be in place in the NAMT (and distributed with its partners) to allow for the testing of interoperable applications that fit within this framework The initial application area to be demonstrated is the inspection of manufactured parts. Systems to be included are shop control, inspection, product data management, and factory status information. Later, the area will be extended to machining and simulation, and finally extended to design and planning for manufactured parts.
Specifications that have been identified for development or enhancement over the next two years include STEP application protocols such as AP213 (process planning) and AP219 (inspection planning), STEP Standard Data Access Interface, OMG Interface Definition Language (IDL) for application integration (shop control, product data management, etc.), and others as appropriate.
The primary emphasis of the project will be to develop testing methodologies (and possibly testing services) for the emerging specifications. The testbed will be available for industry to collaborate with NIST in the area of distributed and virtual manufacturing.
In late June of 1996, an industrial review was held at NIST and attended by managers of key national programs as well as by many of the software application vendors. There was strong consensus that a framework project that would result in increased interoperability among applications was needed. In addition, the reviewers felt that the technical approach was sound and that NIST could provide a real value-added function through facilitating the implementation of existing specifications and the developing of new ones as needed. Finally there was general consensus that having a testbed where the specifications could be validated will be extremely useful.