Abstract

This work studies the application of Dynamic Mode Decomposition (DMD) to analyze and reduce data from CFD simulations of military projectiles with base-bleed technology (a system that reduces aerodynamic drag). From 870 axisymmetric two-dimensional simulations, DMD identified dominant dynamic modes of the axial velocity field. The approach resulted in a dimensionality reduction of approximately 99.31% and highlighted relevant modes. The next steps involve using these modes in machine learning models for national defense applications.

ERAD-NE 2025 aims to encourage the study and research in high-performance computing, computer architecture, distributed systems, parallel processing, and their applications in fields such as biology, engineering, physics, mathematics, computer science, medicine, finance, chemistry, among others.