A digital twin approach for online reconfiguration of robotic manufacturing systems toward plug-and-produce

The digitalization of manufacturing is being developed with frequent changes in customer demands. This requires greater flexibility of manufacturing systems resulting in updates in types and quantities of field devices. However, traditional reconfiguration approaches ignore the perception of dynamic states in the actual manufacturing environment, and their deployment often happens offline unfulfilling the plug-and-produce (PnP) requirement. This paper focuses on the online reconfiguration approach of robotic manufacturing systems (RoMS). First, a digital twin-based modeling approach for reconfiguration of RoMS was proposed. Considering dynamic demands for the manufacturing capability of RoMS, a reconfiguration strategy for the digital twin model was specified. Then, a digital twin-based online reconfiguration optimization approach was proposed. An architecture that combines the reconfiguration optimization method and Open Platform Communications Unified Architecture (OPC UA) discovery service was designed. A multi-objective optimization model was developed for the reconfiguration problem of RoMS; an improved migrating birds optimization algorithm was designed to solve the problem. Finally, on-demand PnP functionality was realized through the OPC UA discovery service during the RoMS’s real-time operation. The feasibility of the proposed approach was verified by a practical use case of RoMS, showing superior performance than other comparative optimization methods in terms of defined fitness function.