A Modular Framework for Simulating and Visualizing Digital Twin Factories

Abstract

Digital Twin technology is becoming increasingly important in Industry 4.0, but many existing platforms are difficult to access, configure, and adapt for teaching and early-stage research. This paper presents a modular Digital Twin Factory framework that simulates and monitors three representative manufacturing assets: a CNC milling machine, a conveyor belt, and an industrial robotic arm. The framework combines configurable machine simulation, real-time data streaming through FastAPI, interactive visualization through Streamlit and Plotly, Random Forest-based predictive maintenance, and rule-based feedback control. The reported experiments show a mean real-time data refresh of 150 ms for a single-machine change and an overall predictive-maintenance F1-score of 0.87 on the synthetic evaluation data. The framework also includes interactive monitoring and automatic response to simulated operating conditions. Rather than presenting the framework as a replacement for a physical industrial Digital Twin, the work positions it as an open and modular simulation environment for education, experimentation, and prototype development.