Fire Safety Science Digital Archive

This paper reports a fundamental study of dynamic interactions between a buoyant reacting plume and evaporating droplets using large-eddy simulation (LES). An idealised prototype configuration is set up to numerically mimic a sprinkler nozzle placed right above the fire source, dispensing evaporating droplets of various properties. The gas phase of the reacting plume is described in the Eulerian frame while the discrete droplet phase is treated in the Lagrangian frame, with strong two-way coupling between the two phases through mass, momentum and energy exchange. Finite-rate chemistry is included and modelled by a scale similarity subgrid-scale model. The LES has six dynamically determined model constants, which minimizes constant tuning. A parametric study has been conducted by varying the initial Stokes number (St0) or non-dimensional droplet size, mass loading ratio (MLR0) and droplet velocity magnitude (|vd0|), independently. Droplets of three initial sizes (780?m, 390?m and 195?m) have distinctively different dynamic interactions with the buoyant reacting plume. Increasing MLR0 progressively from 0 to 9 increases the droplet effects, and in the case of the largest droplets used, the reaction is completely suppressed and the plume structure destroyed. Increasing |vd0| has mixed effects on the droplets’ fire suppression capacity. Detailed analysis of the budget equation for a non-dimensional gas temperature reveals roles played by the droplet-related terms in combustion suppression. The only “cooling” effect on gas temperature comes from the convective heat transfer between the phases, which drives droplet vaporisation, while there are three mechanisms (mechanical work done by droplet drag force, part of internal energy transfer into the gas phase due to evaporation and kinetic energy interactions between the phases) contributing to “warming” effects. On the whole, evaporating droplets in all cases studied result in significant reduction in reaction rate and gas temperature especially the peak values.