Manufacturing Services Network Models Project

Manufacturing re-shoring represents a huge opportunity for U.S. small and medium manufacturing enterprises (SMEs). To seize this opportunity, however, SMEs must be able to operate in advanced production networks. That means, following their discovery and qualification, the SMEs need to be able to receive and use digital manufacturing data, and effectively respond to the requests to make the parts. Standards for representing and communicating manufacturing service capabilities, including material processing and manufacturing information processing capabilities do not exist.  This project will develop measurement science necessary to build reference models, methods, and tools to support manufacturing service capabilities standard development.

Objective: To increase productivity of U.S. small manufacturers by enabling digital communication of manufacturing information for production and data processing through development of reference models and systematic, verifiable modeling methodology that enable new standards for manufacturing supply chain integration, by 2014.

What is the new technical idea? Manufacturing insourcing is on the rise^[1],[2]. The Boston Consulting Group estimates that by 2020 as much as 30% of the products the US imports from China ($55B/year) could be manufactured here. In addition, the U.S. small and medium enterprises’ capacities are underutilized by 15% to 36% on average^[3]. This represents a huge opportunity for the U.S. Small and Medium Manufacturing Enterprises (SMEs).

To prepare for this potential increase in insourcing, SMEs must be able to (1) receive and use digital data models, and (2) respond to requests associated with making parts using those models. Yet, there are no reliable and affordable means to assure correct use of such data. Methods and tools are lacking to describe precisely the required manufacturing service capabilities, including the material processing and manufacturing information processing capabilities.

To remove this barrier, this project will do the underlying measurement science research needed to develop expressive, shared models, most likely based on the Web Ontology Language (OWL)^[4], for describing manufacturing service capabilities. It will also develop a new methodology and tools so that SMEs can convert more easily to those models.

A key challenge is that models and terminologies of manufacturing service capabilities vary and change over time across manufacturing organizations. Traditional approaches using monolithic, static, text-based models are likely to fail. There is a gap in measurement science to model manufacturing service capabilities for today’s dynamic and distributed environments.

The technical idea is to advance a recently developed technique called representation patterns^[5],[6] to develop modeling approaches that enable precise, flexible, and open models of manufacturing service capabilities. NIST will develop manufacturing service capability modeling patterns that will have verifiably correct behavior.

What is the research plan? Our plan has two tasks^[7]: develop representation patterns for reference models of manufacturing service capabilities, and develop a process and tools to create manufacturing service capability data models based on these patterns. The first task analyzes requirements for and synthesizes a new approach to modeling manufacturing service capabilities. The second analyzes the new approach by developing tools and methods in support of the approach and experimentally analyzes their performance.

1. Develop representation patterns for reference models. The project will develop reusable patterns from the myriad schemes SMEs use for capturing and communicating concepts associated with their capabilities. Example concepts include length, tolerance, quality, material, software, and file formats. A pattern for part length capability, for example, might be the min and max interval in inches. New measurement science is needed to find the optimal way to represent these patterns.

   We will analyze requirements for manufacturing service capability information. Both the material processing (e.g., process, quality, and product properties) and manufacturing information-processing capabilities (e.g., engineering software functionalities) are included. We will analyze representative use cases and develop an OWL data model to capture manufacturing service capability information used by today's suppliers and information sharing portals. We will perform classification analysis of the information to identify similarity measures and common information types. We will use classification results to identify the necessary representation patterns.

   We will synthesize a methodology to derive representation patterns for manufacturing service capability information.  We will consider both semantic mediation (i.e., ability to reconcile differing manufacturing service models), and query requirements (i.e., ability to retrieve information using the manufacturing service models). We will review and advance state of the art in representation patterns as they apply to manufacturing services capability information. We will synthesize and experimentally verify behavior and expressivity of a pattern library for manufacturing service capability information, by demonstrating the target information retrieval behavior enabled by the patterns.

2. Develop a process and tools to create manufacturing service capability information. We will use a four-step process to develop tools that create and maintain manufacturing service capabilities models. First, we develop syntax transformation methods and tools to translate the differing syntaxes used in current SME practices (databases, HTML, XML, and text) into a common syntax, namely OWL. This will be focus of our work in FY13 for this task. Next, we will develop a method and tools to convert capability information that has more than one possible representation into a standard or “normalized” form that will allow application of patterns to the SME-specific proprietary models. This step will eliminate the differences in syntaxes and views in the SME-specific proprietary models, allowing objective semantic adjustments to be made next.

   In the third step, we will develop gap analysis method and tools to identify differences between the normalized (standard) model and the current state of an evolving manufacturing service capability model. The result of the analysis enables repeatable and traceable adjustments to the standard model. In the last step, an adjustment verification method and tool will assess the outcome of the adjustments for computational inconsistencies or errors. Results of the experimentally verified four-step process will be published as industry guidelines in support of manufacturing service capability modeling used as a basis for standardization.

[1] http://www.whitehouse.gov/blog/2012/01/11/everything-you-need-know-about-insourcing

[2] http://mba.tuck.dartmouth.edu/pages/faculty/matthew.slaughter/pdf/insourcing_study_final.pdf

[3] Federal Reserve report on “Industrial Production and Capacity Utilization.

[4] OWL (Web Ontology Language) is based on a logical formalism and, therefore, is much more powerful than HTML or XML.

[5] Blomqvist, E., Gangemi, A., and Presutti, V. Experiments on Pattern-Based Ontology Design. Accessible at http://dl.acm.org/citation.cfm?id=1597743.

[6] A representation pattern (also known as design pattern, knowledge pattern, ontology design pattern, etc) is used here to reference an archetypical solution to information modeling design problem given a certain context.

[7] In FY12, the project established a team of industry stakeholders that recommended investigating possibility of bridging the gap between product and process information. Throughout the FY13 thrust tasks, the project will explore possibility of including both product and process information within the patterns for modeling manufacturing service capabilities.

Recent Results:

Outputs:

• Organized, in April 2012, a workshop “Models of Manufacturing Services for Enhanced Manufacturing Sourcing” with participation from industry, academia, and government to document requirements, technical issues and potential approaches for developing shared representations of manufacturing capabilities.
• Prototype software tools developed and demonstrated at the workshop in collaboration with academic partner:
• • Ontology-based Extensible Dynamic Forms. The tool generates on-the-fly input forms to capture manufacturing service capabilities. The tool takes as input a part of a model of manufacturing capabilities (i.e., ontology) and creates a computer-processable definition of input forms.   The tool allows non-technical users to enter manufacturing capability data, which is captured in a computer-processable manner.
  • Ontology-matching System supports semantic similarity-based matching. The tool computes one-to-one correspondences between modeling concepts in two different formal models. The tool takes as input two different models that describe same domain and generates weighted relationship between classes to indicate their similarities. The tool supports analysis of different manufacturing capability models and identification of concepts to include in a shared model.
• Paper titled “Semantic Mediation Approaches to Enhance Manufacturing Service Models” submitted to the International Journal of Computer Integrated Manufacturing.
• Paper titled “An Analysis of Semantic Mediation of Manufacturing Models” submitted to the International Journal of Computer Integrated Manufacturing.

Standards and Codes:

NIST has initiated a new working group at the Open Applications Group (OAG), a US-based SDO. This working group will develop standard for manufacturing service capability information to enable digital communication of manufacturing information for production and data processing. Expected outcomes are draft standard requirement document for manufacturing service capability information, draft standard meta-model of manufacturing service capability information, draft standard for ontological representation patterns of manufacturing service capability information, and draft standard model for manufacturing service capability information.