Quantitative validation in indoor dispersion modeling: Comparing large-eddy simulation results with experimental measurements

Modeling indoor contaminant dispersion is crucial for exposure analysis in fields like occupant safety and infection prevention. Accurate predictions necessitate appropriate computational approaches because numerical solutions to turbulent indoor flow conditions are vulnerable to modeling errors. Large-eddy simulation (LES) is a turbulence-resolving approach with the potential to describe the relevant flow physics governing indoor contaminant dispersion. This work documents a quantitative validation study of the PALM LES model against experimental indoor dispersion measurements. The experiments were conducted in a controlled chamber with a mechanical ventilation system operated at two different ventilation flow rates (2 and 5 air changes per hour). The LES results were obtained using three different resolutions (1, 1.5, and 2 cm), labeled Fine, Medium, and Coarse. The evolution of particle concentration was monitored identically in the chamber and the LES model using a multipoint measurement network. The validation analysis assessed the performance of the PALM LES model in predicting aerosol dispersion using four validation metrics. The results indicated strong performance for the fine model under both ventilation rates. Validation performance declined with reduced resolution, and the coarse model demonstrated evidently lower accuracy due to deficiencies in capturing thermal stratification effects. Sensitivity analysis revealed that the validation results were largely unaffected by changes in thermal boundary conditions. This study highlights the importance of model resolution in predicting indoor contaminant dispersion and cautions against assuming predictive capacity in thermal modeling based on dispersion modeling results.