From Data to Decisions: Integrating Process Mining and Twinning Systems in Manufacturing Systems
Giovanni Lugaresi
KU Leuven
While recent investments in automation have boosted productivity, they also intensified the need for high-fidelity, adaptive decision-support tools across production planning, control, and system supervision. Current digital twin deployments, however, remain limited by high development and maintenance costs, alongside the inability to keep models valid, synchronized, and aligned with the real systems over time. Recent advances in continual online validation and autonomous calibration have shown promising results in preventing model drift and sustaining simulation fidelity during operation. Likewise, progress in twinning systems, multi-paradigm modelling, and lifecycle-oriented architectures provides a conceptual foundation for robust digital twin evolution and long-term operability. Complementing these developments, new model generation and MES integrated architectures enable digital twins that evolve with the system, maintain process awareness, and capture real control logic in high-fidelity environments. This talk synthesizes insights from recent research and experimental platforms to outline a cohesive agenda for next-generation Digital Twins. Emphasis is placed on the convergence of process mining, online model maintenance, MES-driven simulation, and twinning systems theory as a pathway toward sustainable, trustworthy, and cost-effective digital twins in manufacturing.
Speaker bio: Giovanni Lugaresi obtained a PhD with the thesis titled “Automated Generation and Exploitation of Discrete Event Simulation Models for Decision Making in Manufacturing” (Politecnico di Milano, Italy). In his PhD, he developed methods based on process mining for the automated generation of discrete event simulation models of manufacturing systems. In 2018 he co-founded the “THE FACTORY” laboratory in Politecnico di Milano, an innovative learning factory based on lab-scale models of manufacturing systems, which serves as test bench for integrating digital models with physical systems, as well as for testing Industry 4.0 applications and innovative teaching experiences. He has eight years of experience in industrial projects with leading companies, and he participated in research projects at both national, and European level. Today, he is Assistant Professor at the Department of Mechanical Engineering at KU Leuven, where he is the advisor of four junior researchers. His research focuses on digital twins for production planning and control of smart manufacturing systems.