Digital Twin and Open Automation for Process Manufacturing
As part of its partnership with MxD, the nation’s digital manufacturing institute, ADI embarked on a research and development project with Dow and the University of Michigan to demonstrate advanced manufacturing concepts in an industrial open automation framework. This project is now featured on MxD’s future factory floor in Chicago, IL. The project is one of MxD’s many projects and demos meant to showcase new technology, train the workforce on new systems and tools, and demonstrate the need for cybersecurity in manufacturing.
This project is a collaboration between MxD, Applied Dynamics International (ADI), the University of Michigan Barton Research Group, and The Dow Chemical Company.
The Process Manufacturing Digital Twin – Open Architecture Testbed Framework consists of a digital twin and open automation sandbox with a self-contained Integration Test Environment (ITE) platform. This system demonstrates and evaluates digital twin and open automation technologies with cross-vendor systems. It also allows for manufacturers to develop, test, and evaluate new technologies without interrupting production operations and without costly R&D investments.
Open Automation Framework
- Reduces sole sources lock-in by connecting systems from different suppliers over different networks and protocols
- Provides plug-and-play connectivity across the full array of legacy and future interfaces
- Enables rapid deployment and reconfiguration of industrial data and control applications
- Enables open automation and digital twin capabilities for legacy and future systems
- Provides standards-based access to data from all networks and ISA-95 layers
Off-the-Shelf Industrial Computing
Achieves the lowest cost solution through leveraging mass market computing and connectivity products from leading companies including Dell®, Cisco®, NVIDIA®, Eaton®, Intel®, and Texas Instruments®
Digital Engineering Agile Development
Reduces risk and deployment time for digital manufacturing technologies by evaluating the capability in a virtual environment
- Provides an up-to-date representation of the physical asset and process in operation
- Reflects and evaluates the condition of the physical asset and process
- Runs in parallel to the real assets and process, and immediately flags operational behavior that deviates from expected behavior
- Lowers maintenance costs by predicting maintenance issues before breakdowns occur
- Provides enhanced insight into the performance of the process
Open automation frameworks making use of standard Microsoft Windows® and Linux cybersecurity features allow you to capitalize on your existing industrial cybersecurity policies
The project started when MxD and its members proposed addressing the challenges of “plug and play” connectivity and interoperability between different vendors. According to Rene Reinbigler, an enterprise architect at Merck, at a BioPhorum workshop, “communication between equipment from different suppliers will no longer be a nice to have, it will become mandatory in the industry.” MxD was searching for a computing architecture that allows swapping of different technologies with minimal levels of reconfiguration required and relying heavily on the use of standards, the Open Process Automation Standard (O-PAS), developed by the Open Process Automation Forum, a member of the Open Group, was of most interest to the project.
The project involved the development and implementation of a framework for gathering and analyzing all the data from a process manufacturing line necessary to improve visibility and control. This framework is a key first step in being able to implement proof-of-concepts for ‘mobile worker’, cybersecurity, predictive maintenance, and other digital twin use cases. This well-defined vision has the factory connected to: on-premise computing at the edge; cloud computing; and, the enterprise. This project looked to demonstrate key enabling technologies to achieve that digital manufacturing vision.
ADI believed the project scope was a great fit for the ADEPT software platform and believed the development of a demonstrator industrial computing rig, for this project called an Integration Test Environment (ITE) rig, would allow the concepts to be rapidly developed and clearly demonstrated.
ADI’s partnership with the Barton Research Group in open automation connectivity and Digital Twin technology is part of how ADI became aware of the MxD project opportunity. The scope of this project made it a natural fit for the Barton Group to partner with ADI for the purposes of rapid prototyping and demonstrating their best-in-industry Digital Twin architecture and technology platform.
The Dow Chemical Company joined the team as the project’s “Pilot Manufacturer” and voice of the customer, and provided a key input to the project with the Dow Statement-of-Work (SOW) document. The Dow SOW document provides a detailed “customer discovery” report including solution objectives, key use cases, target demographic, solution architecture, ranked must-haves, and ranked nice-to-haves.
The project team designed, implemented, and delivered an industrial open automation framework, based on ADI’s ADEPT software platform, that included multiple digital twin and open automation apps, running in real-time, on the purpose-designed and built ITE rig. Figure 1 illustrates a logical and functional view of the Digital Twin and Open Automation Framework. ADI’s systems engineering team performed an agile design engagement with the Dow/MxD team, designed, and built the ITE as an industrial-packaged computing rig based on Commercial Off-the-Shelf (COTS) industrial computing equipment including Dell® edge server, NVIDIA® machine learning server, Cisco® industrial network gear, Eaton® smart power distribution equipment, and Texas Instruments® industrial communication prototyping equipment.
One of the unique features of the ADEPT software platform is its support for standards-based models. A model built using any one of dozens of commercial modeling tools, such as Matlab/Simulink and Siemens Flomaster, can be dragged into ADEPT, assigned to an app and an edge server, connected to factory data and other apps, and deployed to the computing framework with relative ease. This allowed the Barton Group team to design and prototype their Digital Twin apps using their preferred modeling tools and enabled fast iteration between the ADI modeling and integration team and the Barton Group algorithm development team.
Apps and Models
A key goal of this project was to demonstrate Digital Twin (DT) and Open Process Automation (OPA) applications, or apps, running in real-time within the ADEPT plug-and-play testbed framework. The project did not support connecting the testbed framework to a real factory, at least for this initial project scope, due to schedule and budget constraints. In order to rapidly develop and validate the DT and OPA apps, a portion of a process manufacturing line was simulated and hosted within an ADEPT plug-and-play computing framework to enable the simulated plant to be connected to, and virtually operated alongside, the DT and OPA apps. The table below summarizes the apps and simulations demonstrated within the testbed normal operation.
|Function Name||App or Sim||Function Description||Developed In|
|Pump Health Digital Twin||App||Monitors the condition of equipment in the system by detecting degradation onset, estimating and classifying health state||Matlab/Simulink|
|PID Controller Health Digital Twin||App||Estimates the health state of the PID flow control loop||Matlab/Simulink|
|Process Health Digital Twin||App||Estimates the health percent and detects anomalies in the process||Matlab/Simulink|
|ITE Health Continuous Built-in-Test||App||Estimates the health of the ITE computing rig, plug-and-play computing framework, and hosted apps||Matlab/Simulink|
|Startup Readiness Check||App||Performs a cross-function check of readiness to start the manufacturing process by checking the readiness of connected systems at various level of ISA-95||Python|
|Framework Auto-Recovery||App||Performs a periodic check on the health of each app running within the OPA plug-and-play framework and performs recovery reset actions if/when needed.||Python|
|Process Simulation||Sim||Provides high-fidelity multiphysics simulation of a flow control unit with accurately modeled, and dynamically-controlled, process and equipment degradation.||Matlab/Simulink SimScape and C/C++|
Industrial Plug-and-Play Connectivity Interfaces
The early project requirements discovery efforts included defining the industrial data connectivity needs required to support Dow’s modern and legacy process manufacturing lines. The table below summarizes the industrial plug-and-play connectivity demonstrated on the testbed as well as those not demonstrated but natively supported on the testbed.
|Interface Description||Demonstrated or Supported|
|High-Speed (FPGA) I/O||Supported|
MxD, the nation’s Digital Manufacturing Institute and the National Center for Cybersecurity in Manufacturing, is where innovative manufacturers go to forge their futures. In partnership with the Department of Defense, MxD equips U.S. factories with the digital tools, cybersecurity, and workforce expertise they need to begin building every part better than the last. As a result, our approximately 300 partners increase their productivity, win more business, and strengthen U.S. manufacturing. Learn more at mxdusa.org, and contact us about projects at email@example.com.