A lightweight model for real-time safety management of nanoparticle concentrations in industrial spray processes

Industrial spraying processes incidentally generate high rates of hazardous nanoparticles (NPs), posing significant health risks to workers. While ventilation systems are critical for mitigation, optimizing their operation via Model Predictive Control (MPC) requires fast predictive tools, making traditional Computational Fluid Dynamics (CFD) unsuitable due to high computational costs. This study develops and validates a physics-based Lightweight Model (LWM) designed to predict real-time NPs concentrations in a thermal spraying booth. Based on a multizone airflow approach, the LWM incorporates source emissions, airflows induced by ventilation and openings, and diffusive transport using a well-mixed mass balance formulation. The model was validated using experimental data from an industrial facility and benchmarked against detailed CFD simulations. Results demonstrate that the LWM accurately predicts both the transient growth and steady-state stabilization of NP concentrations, achieving comparable accuracy to CFD but with drastically reduced computation time (minutes versus days). By offering accurate predictions with minimal computational demand, the proposed LWM is a promising prediction model for future MPC-based occupational safety and ventilation control. Its implementation and validation within a complete MPC framework remain future work.