Direct Numerical Simulation (DNS)-Informed Drag Modeling for Droplet-Shock Interaction

Liquid droplets impacted by a shockwave experience highly unsteady forces due to the extremely short timescales of the interaction and the droplet’s dynamic deformation in the post-shock region. Conventional rigid-sphere and steady state drag models fail to accurately capture droplet-shock interactions because they neglect droplet deformation and only estimate the quasi-steady force established after shock passage. In this study, high-fidelity direct numerical simulations (DNS) of individual droplets are performed to generate detailed, time-resolved data on droplet forces and deformation for Mach numbers ranging from 1.5 to 7.0 and initial droplet diameters between 250 and 1000 µm. These transient data are then used to train an Artificial Neural Network (ANN) capable of predicting the drag coefficient as a function of droplet deformation and flow conditions. Preliminary results indicate that the ANN provides strong agreement with experimental measurements and significantly outperforms traditional rigid-sphere and steady-state models, offering a robust framework for modeling droplet-shock interactions under highly dynamic conditions.