At ILASS Europe 2025, the 33rd European Meeting on Liquid Atomization and Spray Systems, held from August 31 to September 4, 2025, in Lund, Sweden, we presented results from an experimental study on optical spray diagnostics.
The study investigates the analysis of light scattering from colloidal droplets illuminated by a Gaussian beam to determine suspension concentration. The experiment features a controlled droplet chain of milk droplets with varying concentrations, enabling a detailed investigation of concentration-dependent light-scattering behavior. To detect the milk concentration, both artificial intelligence–based methods and statistical approaches were applied.
Analysis of the light scattering of a colloid droplet on a Gaussian beam to determine the suspension concentration
Abstract
Characterizing individual static suspension droplets, such as those levitated acoustically or electromagnetically, is well-established and primarily involves analyzing light scattering and absorption effects. However, accurately determining the concentration and composition of dynamic suspension droplets, such as those in a flow or spray, remains a significant challenge. This study addresses this gap by analyzing the light scattering of dynamic suspension droplets passing through an elliptical Gaussian beam to determine the concentration of suspended particles within the droplets. Measurements were performed at a fixed scattering angle in the backscatter region, with high temporal resolution achieved using a photomultiplier and a high-speed signal analyzer. The experimental setup employed a modified version of the commercial device SpraySpy®, utilizing the time-shift technique [1]. Monodispersed suspension droplets were generated with a droplet generator to ensure consistency. A range of suspension concentrations was analyzed to collect sufficient data for developing an empirical model correlating the light scattering signal with suspension concentration. Two approaches were used to establish this correlation: one utilizing an artificial neural network [2] and the other leveraging the statistical properties of the light scattering signal. This work provides a robust methodology for characterizing dynamic suspension droplets, advancing the understanding of their concentration and composition.
rtjhi