Early Prediction οf Spray Charactéristic Based οn Curvature Distribution

Overview & Synopsis

Atomization and spray processes play a critical role in various industrial and scientificapplications, including fuel injection in combustion engines, agricultural spraying, pharmaceutical drug delivery, and additive manufacturing. The performance of these applications ishighly dependent on the spray characteristics.Research on atomization aims to describe, understand, and control spray formation,ultimately leading to the stabilization of spherical droplets. However, spray characteristicssuch as droplet size distribution are typically not well defined during atomization, as sphericaldroplets are absent for most of the process.To address this, a novel methodology based on surface curvature distributions has beendeveloped, which remains well-defined at all stages of atomization. A mathematical transformation converts this curvature distribution into a direct diameter distribution withoutinformation loss. This model can then be applied to estimate a possible equivalent dropletsize distribution, providing insight into the final characteristics of the spray. Surface curvatureanalysis can therefore enable a continuous description of the spray.Initially validated through numerical simulations, surface curvature analysis has alsobeen proven effective in image analysis. However, since most experimental images are limited to 2D, we investigate the bias introduced by this limitation in different cases, such assuperposition or overlap in image analysis. To generate realistic liquid structures, numericalsimulations of homogeneous isotropic turbulence were employed under various scenarios.These structures were then projected onto a two-dimensional grid to mimic potential experimental 2D images. The surface curvature analysis was applied to both 3D and 2D data toassess the impact of dimensionality.In addition, a typical injection system, which is favorable for 2D experimental imaging,was selected for further study. The flat fan spray configuration was first examined through aliterature review, focusing on break-up mechanisms and numerical constraints. A detailedattempt to reproduce this configuration was conducted using direct numerical simulations(DNS), which revealed key characteristics such as velocity distributions and thickness evolution. The initial parameters, including injection velocity and opening angles, were shown tosignificantly influence the final spray characteristics.Finally, surface curvature analysis was applied to the break-up process of the flat fan sprayto further explore its utility. A conference paper analyzing the results from the experiment offlat fan spray, based on the methodology developed in this thesis, has been included as anappendix at the end.