A Computational Framework for On-The-Fly Analysis of Droplet Genealogy
This work presents a real-time monitoring and analysis scheme for droplet breakup and lineage in multiphase simulations of spray formation in order to bridge the gap between high-fidelity outputs and meaningful statistical understanding. Droplet breakup and coalescence are central to systems like fuel injectors, agricultural sprays, and pharmaceutical manufacturing processes, yet it is difficult to numerically characterize because of their transient and unsteady nature. To address this, a liquid structure tracking tool is introduced and integrated into a state-of-the-art multiphase flow solver. The tool uses persistent IDs and data based on droplet events to capture the detailed genealogy of droplet break-up and coalescence dynamics directly in the simulation, while also allowing local physical properties to be accessed and avoiding snapshot-based post-processing limitations. This data is subsequently imported into a graph database, where droplets are nodes and their split/merge history are edges. The resultant graph supports complex querying and size distribution, generation number, and volume ratio analysis. The framework is compared to the breakup map of Farsoiya et al. (JFM 2023), and identifies three distinct breakup regimes characterized by trends in droplet generation rates, mean size, and breakup type. This method provides a scalable pathway toward enhanced modeling and understanding of atomization of turbulent multiphase flows.