Effect of injection model and turbulent dispersion models on prediction of full-cone nozzle spray in OpenFOAM*

Present work simulates a full-cone spray (Delavan CT-1.5–30B) using FireFoam, a solver in OpenFOAM. The solver uses the notion of 'parcels' in forming computational spray cloud. Injection model generates the spray from input parameters provided from experiments viz. particle size and velocity distributions of droplets. For the analysis, point-injection and detailed-profile injection models are considered.  Once injected, the droplets are tracked in a Lagrangian manner in an otherwise Eulerian framework for the background air flow field. To model the turbulent fluctuations experienced by particles from surrounding flow, turbulent-dispersion models have been imported from the Lagrangian solver, viz. Stochastic and Gradient-k dispersion models.  The predicted sprays with different aforementioned models for injection and dispersion are compared with experimental spray characteristics like Volume Flux, d_v50 and velocity distribution at near-field and downstream. Detailed-profile injection predicts the spray better compared to point-injection. Stochastic dispersion model gives similar profile compared to no-dispersion, unlike Gradient-k model which shows significant deviation.