A Shape-Similarity Latent Space for Fluid Interfaces: Invertible Reduced-Order Modelling of Droplet Morphology
My rating
This paper addresses the rapid breakup of droplets, where sparse recordings cannot capture the intermediate fluid-interface states. It develops an invertible, low-dimensional latent space based on shape similarity for reduced-order modeling of droplet morphology. The representation uses morphologies observed across a device’s actuation range to reconstruct states missing from individual recordings without repeatedly running costly simulations.
Tags: