The Advanced Manufacturing Data Infrastructure and Analytics (AMDIA) program will lay the groundwork for advanced data infrastructure and corresponding analytics to improve the productivity, resiliency, security, and sustainability of manufacturing operations and enterprises across the supply chain. As the manufacturing industry produces increasing volumes of diverse data, stakeholders need robust data infrastructure and trusted analytics to prepare, model, understand, and utilize their data effectively and efficiently for improved control and better decision-making. Emerging technologies allow manufacturers to collect, structure, link, and analyze data in new ways. However, technologies are seldom one-size-fits-all solutions. Research into manufacturing data and the role of human expertise in the process is needed to effectively adopt and integrate solutions into existing operations.
The AMDIA program will develop test methods, standards, toolkits, models, datasets, industry pilots, and build communities of interest to advance manufacturing data infrastructure, with a focus on data collection, transformation, traceability, security, interoperability, and analysis. AMDIA outputs will lower the barrier to incorporate new technologies and analysis methods into existing and emerging operations. These outcomes will enable trusted, understandable, and reproducible analysis workflows across engineered products, manufacturing processes, production systems, enterprises, and supply chains to improve decision-making.
The Advanced Manufacturing Data Infrastructure and Analytics (AMDIA) Program will develop and deploy measurement science to advance data infrastructure and analytics that will improve the productivity, resiliency, security, and sustainability of manufacturing operations and supply chains thereby enhancing U.S. innovation and industrial competitiveness.
What is the Problem?
Manufacturers continually gain more capability to collect and monitor data throughout all levels of their operations and across their supply chains. As manufacturing operations become more advanced, so, too, does the amount, diversity, variability, and uncertainty of the corresponding data. Various Internet of Things (IoT) devices and even human observations also enable monitoring of legacy equipment. Manufacturing processes and constituent enabling technologies are rapidly evolving including advanced sensing capabilities, technical language processing tools, augmented reality systems, and cybersecurity measures. One commonality among all of these elements, and output from nearly every manufacturing process, is data. Manufacturers, and their industry partners are becoming greater generators and consumers of data output from their operations, particularly as automation increases. Turning this data into meaningful intelligence is non-trivial.
The manufacturing community confronts numerous challenges with respect to the lifecycle surrounding data generation and consumption. These challenges include determining what data is necessary for a sufficient understanding of an operation; what data is available for collection; the best approach to capture the necessary data (including if the collection capability exists or needs to be created); how best to organize and structure the data, including any pre-processing; what analysis methods should be applied to the data; how the resultant intelligence should be represented for specific consumers; and where should the intelligence be routed to augment decision-making throughout the factory and across the supply chain. These challenges are more daunting as additional data sources are being discovered; in addition to obtaining data from actual manufacturing operations, data can be generated from models, including digital twins and simulations. The answers to these questions must also explicitly consider security concerns. With data being passed through numerous processes, manipulated by various methods, and examined by a range of personnel, there are multiple ways data security could be compromised. For manufacturers to successfully leverage data to their competitive advantage, they must build a robust and flexible data infrastructure to aid in collecting and analyzing this data.
NIST is uniquely positioned to develop independent, unbiased measurement science to verify and validate emerging data infrastructure and analytical capabilities that are developed for implementation in the manufacturing industry. NIST researchers have built-up a vast network of collaborators, including data consumers from small to large manufacturers, technology integrators, technology developers, and forward-looking researchers. The sharing of data, resulting from these trusted relationships, enable researchers to objectively review and leverage external data sources to develop appropriate verification and validation capabilities for public release and benefit. Similarly, NIST has developed a suite of physical and virtual testbeds, along with a range of simulations and software tools allowing them to generate representative manufacturing data to accelerate and promote their research. Likewise, these in-house generated datasets can be used as reference datasets thereby allowing them to be exported into the public domain for community benefit.
What is the Technical Idea?
A robust data infrastructure creates a vital foundation for manufacturers to build upon with advanced and emerging technologies to improve their operations. While traditional infrastructures contain both hard and soft elements, improvements typically focus on hard infrastructure elements, such as physical networks (e.g., roads). Successful infrastructures explicitly consider both hard and soft elements; soft elements being the institutions that support the physical infrastructure (e.g., emergency services, education systems, and governing organizations). Similarly, a data infrastructure must directly consider both hard and soft elements for its success and sustainability.
Data infrastructures facilitate continued use and analysis of datasets among a variety of stakeholders. They allow researchers to preserve, find, access, and process data in trusted environments (https://eudat.eu/eudat-cdi). One example of a robust data infrastructure is the National Spatial Data Infrastructure (NSDI). The NSDI enables data from multiple sources to be available and integrated to enhance understanding. NSDI is defined as “the technology, policies, criteria, standards, and employees necessary to promote geospatial data sharing throughout the Federal Government, State, tribal, and local governments, and the private sector (including nonprofit organizations and institutions of higher education)” (https://www.fgdc.gov/gda/43-usc-ch-46-geospatial-data-geospatial-data-a…). Adapting this definition,the Advanced Manufacturing Data Infrastructure is defined as the technology, policies, criteria, standards, and employees, to promote manufacturing data sharing throughout the manufacturing supply chain, including individual enterprises.
The AMDIA program will focus on both internal and external data infrastructures. Internal infrastructures can support data sharing across an individual supply chain (e.g., internally shared maintenance data among suppliers). External infrastructures promote data sharing throughout the entire community, including items such as public datasets, open-source tools, and publicly available standard guidelines. A multi-disciplinary team will bring together data architecture and analysis expertise from manufacturing domain researchers, to ensure that the resulting data infrastructure advancements address both industry and research needs. This will improve manufacturing data infrastructure by understanding the nuances of the data and the specific needs of the manufacturing community for improved productivity, sustainability, resiliency, and security while also investigating new technologies that can enable improved analysis, decision making, and control.
What is the Research Plan?
The research plan consists of a portfolio of interrelated projects that focus on key areas of the standards, test methods, and measurement science needed to achieve successful development and implementation of advanced manufacturing data infrastructure and analytics. Collectively the projects provide a comprehensive approach that will lead to new industry standards and practices, which will result in improved productivity, resiliency, security, and sustainability of manufacturing operations and supply chains. The program will take a multi-faceted approach by not only focusing on different parts of the manufacturing supply chain, but also different stages of the data lifecycle and emerging advanced technologies and methods.
The AMDIA program will:
The AMDIA program consists of six projects:
The Product Digital Information Visualization and Exploration (ProDIVE) project will deliver standards, methods, and tools to improve interoperability of (i) data representations across the product lifecycle and (ii) product data standards with advanced visualization modalities, such as augmented reality (AR). Past and on-going work at NIST have made significant gains in harmonizing product lifecycle data standards to help realize the so-called digital thread, a concept purported to construct persistent links across lifecycle activities, such as engineering design, manufacturing planning, fabrication, and inspection. To support additional use cases, this project will focus on the upstream propagation of downstream observations, such as inspection results, towards design processes, such as part requirements.
The Model-Based Manufacturing Capability Definition project provides technical contributions to define, measure, and control these manufacturing capabilities. The challenge of this work is that manufacturing capabilities are inherently dynamic and vary based on the type of manufacturing system to control and the type of decision to be made. The research in this project focuses on the manufacturing capability of a flexible, on-demand, pull-production work cell composed of a machine tool, coordinate measurement machine (CMM), robot, buffer, and material conveyance. Success in this research would improve the agility, flexibility, and competitiveness of the US manufacturing base by allowing decisions to be made based on the measured and predicted capability of manufacturing systems as well as additional part and process information.
The Infrastructure for Agile Manufacturing Data Standards (IAMDS) project will contribute to standards and test methods that normalize models and methods for specifying and connecting shop floor information to operations decision making and execution. The core of the technical idea is to develop methods and tools that enable agile standard life-cycle management. Research will be conducted leading to new standards and methods that allow for more precise and cost-effective description, qualification, and quantification of a message standard functional characteristics and the quality of those characteristics. As the SOMS will exploit the economy of services, standards supporting it must also be agile. New models, computational methods, and tools will be developed to allow for more cost-effective management of the service descriptions along with their respective message standards.
The Monitoring, Diagnostics, and Prognostics for Manufacturing Operations (MDP4MO) project is developing and deploying measurement science to promote the implementation, verification, and validation of advanced monitoring, diagnostic, and prognostic technologies to minimize unplanned downtime and optimize planned downtime in manufacturing operations.
The Supply Chain Traceability for Agri-Food Manufacturing project will pursue its two research thrusts in parallel, incorporating results from one into the other where appropriate. The combined elements of research for the two research thrusts are as follows. Investigate traceability and cybersecurity requirements in the agri-food manufacturing domains working with industry partners and organizations such as AgGateway and identify key standards to extend to address those requirements. Develop a canonical model for Critical Tracking Events (CTEs) and associated Key Data Elements to support traceability of grain. Develop, test, and deploy enhancements to standards such as AgXML and OAGIS for message exchanges to support traceability and security configuration checklist compliance. Develop mappings from the canonical model for CTEs to related content in standards. Pilot use of new sensor and communication technologies to gather data for traceability and detect security vulnerabilities and misconfigurations of network endpoints. Develop models and tools to reduce the cost of security profile development and checklist content authoring.
Lastly, Knowledge Extraction and Application (KEA) project proposes a hybridized AI and expert-driven framework for quantifying human knowledge, in which Technical Language Processing (TLP) and graph-theoretic methods are introduced, to assist in “tagging” and analyzing unstructured information to enable decision making and continuous improvement. This project will contribute to domain-specific standards and associated open source toolkits for extracting knowledge to improve manufacturing decision making.