Digital Twin Methodology for Flexible Manufacturing System Design: Integration of VSM, Discrete-Event Simulation and Production Scheduling for Gear Wheel Production

This paper proposes a structured methodology for constructing a digital twin (DT) of a planned flexible manufacturing system (FMS) for gear wheel production, integrating value stream mapping (VSM), discrete-event simulation (DES), production scheduling, and CAD/CAM modeling into a hierarchical, iterative design framework. The methodology was validated on an industrial case study involving a Polish manufacturer of gear transmissions undergoing modernization from conventional machining to a fully automated system incorporating CNC machining centers, industrial robots, automated guided vehicles (AGVs), and an automated storage and retrieval system (ASRS). Production scheduling was performed using eight algorithms including a random-search method and an ant colony optimization (ACO) algorithm, applied to a representative of 1072 parts across 10 gear wheel types. Simulation models of both conventional and automated systems were developed in FlexSim 2024. The best random instance achieved a makespan of 72,317 s in the automated model, compared to 75,634 s for the list rule—a 4.4% improvement that corresponds to approximately 55 min of production time per batch. AGV fleet sizing experiments identified four vehicles as the optimal configuration, beyond which marginal gains fall below 2%. The proposed digital twin framework enables virtual commissioning, continuous production planning, and what-if analysis prior to and during physical system implementation.