Automated Generation of Digital Models for Production Lines through State Reconstruction

Thanks to the rapid advances in information technologies, digital twins have been widely adopted in the manufacturing industry to support production planning and control. At the core of a digital twin is a digital model that mirrors the physical system in a virtual space. It is inefficient to develop digital twins by modeling the considered systems manually. Although significant research effort has been made to automate the generation of digital models, most approaches so far impose strong assumptions on the available data or cannot precisely capture the behavior of the physical system. Noticing the current gap, we propose in this paper a novel approach for automatically generating a graph representation of a production line from an event log through state reconstruction. The feasibility of the proposed approach has been demonstrated on three simulated instances