SORS: Sprays modelling and experiments

Abstract

This aim of this work is to develop and validate modelling of polydisperse sprays using the generalised Fully Lagrangian Approach (gFLA), supported by detailed experimental measurements. A flat-fan water spray under quiescent conditions was characterised using highspeed videography (HSV), where a novel image-processing method was developed to extract droplet size and velocity from time-resolved images. Images from different sub-regions were used to reconstruct all events occurring in the spray, capturing local variations in droplet behaviour and identifying the primary breakup zone. The HSV technique was benchmarked against Phase Doppler anemometry (PDA) measurements. While both methods showed consistent trends, HSV provided a broader droplet size range, and PDA offered higher temporal resolution. However, PDA was limited in providing spatial gradients and reliable number density estimates. As a result, HSV data was used to generate the initial conditions for the gFLA. The model tracks droplet trajectories by solving systems of ordinary differential equations and predicts droplet number density, size, velocity, and concentration fields. Jacobian determinant associated with the transformation from initial to current particle positions was used to evaluate number density. The Jacobian determinant approaches zero at locations where droplet trajectories intersect, helping identify folds/caustics in the flow field, while leading to non-physical values for field variables. Kernel regression was applied to reconstruct continuous number density, average droplet size and velocity fields on a Eulerian grid. The gFLA successfully reproduced droplet clustering and general flow structure, although predictions slightly differed from experimental values due to kernel smoothing and simplified assumptions, such as quiescent gasphase flow. Overall, this study demonstrates the feasibility of the gFLA for simulating droplet-laden sprays and provides high-resolution experimental database that can support the development of more accurate models for both primary breakup and Lagrangian spray simulations.