Data Modeling and Workflow Analysis of Cyber-Manufacturing Systems

Cyber-Manufacturing System (CMS) is a vision for the factory of the future, where physical manufacturing resources and processes are purposely integrated with computational workflows to provide on-demand, adaptive, and scalable manufacturing services. In CMS, functional manufacturing components in a factory floor are digitized and encapsulated in the production services; and are accessible by users throughout the network. CMS utilizes data-centric technologies—such as Cyber-Physical Systems, Internet of Things, Big Data Analytics, and Machine Learning—to acquire real-time working conditions of machines, the status of materials and parts flow, the control of inventories and other manufacturing activities in factory floors. By leveraging such advanced technologies, CMS can provide robust solutions in achieving better manufacturing agility, flexibility, scalability, and sustainability than from current factories. CMS also promises significant improvement in manufacturing performance metrics, such as in productivity, lead times, and product quality.

The deep fusion of manufacturing processes with computational processes empowers artificial intelligence in solving complex manufacturing requests and addressing manufacturing uncertainties. Data is the main driver of the manufacturing activities in CMS, which imposes the needs for (i) a generic data model of explicit representation of the engaged entities and stakeholders in CMS and (ii) workflow definition and analysis for service-orientated functionalities and manufacturing intelligence of CMS. However, the existing data models focus on formulating supply-demand relationship only. Also, they include enterprise information or product specifications of CMS, not the comprehensive profile of CMS. The available workflow models cover partial data-driven manufacturing activities, such as the sensor data processing and on-line machine health diagnostics but have limitations in elaborating other intelligent functionalities of CMS.

To address these research gaps, this research aims at developing and presenting the data model and workflow analysis of CMS to support the full implementation of an executable CMS. In this paper, a brief introduction to CMS, the role of data in manufacturing, and data models of manufacturing systems and products are provided. Development of the data model using Entity-Relationship (E-R) diagram is presented to describe the abstraction of CMS, followed by the analysis and graph representations of workflows along with how intelligence and services are enabled in CMS. Next described is the design of data pipelines and Extract/Transform/Load (ETL) processes for integrating, processing, and analyzing industrial data to automate the entire data processing in CMS. An industrial case is used to demonstrate the deployment of the data model, the design of database and instantiation of the workflows and pipelines in a CMS.