Verification of LNG Dispersion Simulation Using FDS Based on Burro3 Experimental Results

Prior validation against well-documented benchmark experiments is essential for establishing a reliable computational foundation for liquefied natural gas (LNG) vapor dispersion simulations in safety assessments and hazard prediction, particularly as new versions of simulation tools are released. This study performed a pre-validation assessment of the predictive performance of Fire Dynamics Simulator (FDS) version 6.10.1 for LNG vapor dispersion modeling. The simulations were validated against data from the Burro 3 large-scale field experiment, with a computational domain constructed to replicate the experimental conditions using carefully refined simulation settings. The results were compared with experimental measurements and prior simulations conducted with FDS 6.3.5. The findings suggest that FDS 6.10.1, when configured with refined computational parameters, achieves concentration predictions within a ±20% error envelope at downwind distances of 57 m, 140 m, and 400 m, showing substantial improvement over the earlier version in near- and mid-field regions. By contrast, deviations in far-field predictions (particularly at 800 m under short-term averaging) reveal limitations in capturing atmospheric variability, such as wind fluctuations and topographic effects. In conclusion, with adequately adjusted computational settings, FDS 6.10.1 can be a reliable tool for simulating LNG vapor dispersion to support the initial emergency response planning.