A Fuzzy Logic-Based Digital Twin Framework for Risk-Oriented Safety Assessment in Waste Sorting Facilities

The increasing automation of waste sorting processes brings significant efficiency gains but also introduces new layers of operational uncertainty that further challenge safety management. This paper proposes a conceptual Digital Twin (DT) framework integrating fuzzy logic to support dynamic, risk-oriented safety assessment in automated waste sorting facilities. The novelty of the approach lies in combining Digital Twin-based system representation with fuzzy reasoning, enabling a continuous and interpretable evaluation of safety states under uncertain or incomplete information. The proposed model defines linguistic safety scales (" safe," "warning," "hazard") and a hierarchical fuzzy inference structure, in which sensor-derived indicators, such as temperature, vibration, or current deviations, are aggregated into intermediate risk sub-indices and a global Safety Index. The Digital Twin concept serves as a virtual representation of a cyber-physical sorting line, connecting the physical and cyber layers to provide context-aware insights into operational safety. A simplified proof-of-concept scenario based on a representative cyber-physical sorting system and synthetic input signals demonstrates the model's capability to detect early warning conditions and dynamically adjust the Safety Index according to the evolving system state. The results indicate that integrating fuzzy reasoning into a DT environment can improve the resilience, reliability, and transparency of safety monitoring, particularly in contexts where data availability is limited or operational variability is high.

Keywords: Digital Twin, fuzzy logic, risk monitoring, waste sorting systems.