Focusing on precompetitive problems, such as enabling communication between digital twins, U-M and ASU are seeking industry partners for a potential NSF research center
Aiming to overcome barriers that prevent digital twins from delivering on their promise to improve manufacturing, the University of Michigan and Arizona State University are inviting industrial partners to participate in a new Center for Digital Twins in Manufacturing.
"Everyone's building digital twins, but we're trying to build the glue or connectivity that enables digital twins to work together-to be composable, reusable and maintainable," said Dawn Tilbury, the Ronald D. and Regina C. McNeil Department Chair of Robotics at U-M.
"Teams in this center will work on precompetitive issues that can bring broad benefits to everyone making and using digital twins, helping expand the use of these great technologies to improve manufacturing performance, quality and uptime."
With enough industry members, the team aims to formally launch the Industry-University Cooperative Research Center, funded by the National Science Foundation. For these types of centers, which are housed in current labs, the NSF contributes $1.5 million over 5 years to support administrative costs, while industry partners pay annual dues of $90,000 to fund the research projects.
An informational meeting for interested industry representatives will be held 1-4 p.m. July 23. Attendees can register online.
The meeting will include presentations about the center, its goals and operations, and the benefits to members, as well as breakout and discussion sessions to discover the topics of highest interest to potential members and identify specific industry needs.
"This meeting will be a great opportunity for potential industry partners to learn more about the vision of the center, understand the value of becoming a member, and help shape the first round of projects," said Wenlong Zhang, ASU associate professor of robotics and manufacturing systems.
A digital twin is a computer model of a device that is in communication with the real device, updating itself so that it matches the state of the device it is modeling. For instance, a digital twin of a milling machine might represent the progress of a part hewn from a block of metal, checking in every tenth of a second to make sure that its model is in lockstep with the real machine. The twin can make predictions about part quality, when maintenance will be needed and more. However, lack of standardization makes digital twins difficult to implement in manufacturing.
At present, Tilbury says most digital twins in manufacturing are for single devices. If a factory has a digital twin for an entire system or line, it is likely a bespoke piece of software that has to be rewritten if a part of the system changes-for instance, if a machine is upgraded. Meanwhile, many companies are providing digital twin software for individual machines. Linking them up with twins of the previous and next machines on the line would enable better coordination both up and down the line.
The prospective center plans to come at issues in digital twins for manufacturing from several directions. They are developing generalized digital twins for certain types of machines-for instance, a 3D printer that could be modified to represent a specific machine. These twins are to be reusable, extendable and maintainable, which means they can accommodate additions to the system, represent the natural degradation of machines, and show improvement following repair or part replacements.
Other aims of the center include:
- Quantifying and reducing uncertainty in digital twins
- Improving predictions and helping factory managers assess their accuracy
- Developing digital twins for human-robot collaboration
- Digital twins of human partners, tracking only the aspects of the human that are relevant to the task at hand, could help robots be better teammates while minimizing privacy concerns
- Using digital twin software for simulation and "what-if" analysis
- In simulation mode, rather than linked to a physical machine, digital twin software could enable factory managers to train personnel and experiment with potential machine upgrades or reconfigured factory lines
- Autotwin: Software that can generate and run digital twins
To test out ideas developed through the center and ensure that they operate as intended, U-M offers the SMART 4.0 testbed, featuring a "connected factory" of mobile robots, computer-controlled machines and 3D printers, connected through open process automation. ASU provides its own connected smart manufacturing system that includes robotics, programmable logic controllers, smart sensors and RFID tracking for milling and additive manufacturing
Tilbury is also the Herrick Professor of Engineering and a professor of mechanical engineering and electrical engineering and computer science.
