Development of an Algorithm for Structure Identification and Droplet Distribution Reconstruction in Liquid Sprays

The present study aims to deliver insights into droplet generation in sprays and develop a suitable methodology to evaluate the droplet distribution. The goal is to aid the development of a novel simulation framework that is based on the Fully Lagrangian Approach (FLA), which is capable of modelling polydisperse evaporating droplets (gFLA) as well. Unlike more commonly used methodologies which study the overall droplet size distribution in sprays, the development and validation of the new modelling approach essentially requires characterising droplet size and velocity distributions at all locations. A commercial flat fan nozzle was adopted in our study to inject water under steady flow conditions using a low-pressure pump. The atomisation takes place such that the liquid sheet and droplets that break out from the spray do not deviate significantly from a 2-D plane. The spray is visualised at 5000 frames per second via a high-speed camera with a long-distance microscopic lens, the optical setup includes a pulsating LED for backlight illumination. The experimental data was collected from 2000 consecutive images at steady state conditions and a 2D spray analysis was performed using an in-house image processing code. The code includes a functionality to distinguish between structures in the spray and to locate the region where the two-phase flow transits from a continuous liquid film to ligaments and then to individual droplets. In the down-stream atomised region, the image processing was applied to obtain the position and size of the droplets in consecutive frames and across all sequences of images. The data was collected across the spray domain to acquire the local droplet size distributions. Finally, we mapped the spray characterisation parameters such as Sauter mean diameter, level of polydispersity, and number density.