The promise of digital twins has been their ability to test and validate a process or product before it even exists in the real world—and thus reduce engineering errors, inefficiencies, and costs. While digital twins are still in their infancy, they have already demonstrated benefits in applications ranging from fault prediction for machinery on the factory floor to performance optimization of wind turbines. The extension of digital twins to manufacturing control across multiple machines on the factory floor, or to life cycle management of a complex system, presents new challenges. The question this research addresses is how to enable a collaboration of digital twins so that they can communicate with each other and share relevant data in order to make even better decisions in extended applications.
A digital twin typically comprises the semantic description and modeling of a single asset—for example a machine or robot—in order to simulate its performance virtually. A collaboration of digital twins would require the sharing of information among a group of digital twins—for example a cell of turning machines and tending robots. Collaborative digital twins could also model cross-company decision-making that includes suppliers, vendors and customers. This research explores the creation of a cloud-based and learning-enabled collaborative ecosystem of digital twins to achieve even broader information exchange and benefits.
While there are many challenges in building a collaboration of digital twins, one of the most immediate is the creation of an appropriate data management framework—to include multiple data models and the communication protocols among them. Currently, real-time data acquisition and transmission technology is focused on interactions between a single physical asset and its digital representation. However, feedback control in a shared environment of multiple twins and real-time exchange and analysis remain a challenge. Other concerns include privacy, security, transparency, and ownership of these data models. Specifically, this research develops a collaborative data management framework for testing strategies of digital twin collaboration for manufacturing control.
PI. Noel Greis, Mechanical Engineering and Engineering Science, University of North Carolina at Charlotte