Data-driven modeling of two-dimensional detonation wave fronts

Historical experimental testing of high-altitude nuclear explosions (HANEs) are known to cause severe and detrimental effects radio frequency signals communications infrastructure. In order study predict the impact HANEs, tractable computational approaches required model complex physical processes involved in detonation wave physics. Modern reduced-order models (ROMs) can enable long-time many-parameter simulations with minimal cost. However, translational scale invariances inherent this type propagation problem limit traditional ROM approaches. Specifically, dimensionality reduction methods typically ineffective producing low-rank when present data. work, an unsupervised machine learning method is used discover coordinate systems that make such amenable methods. The method, which has previously been demonstrated on one-dimensional translations, extended higher dimensions additional invariances. A surrogate HANE system, i.e. a HANE-ROM, one captured well at extremely low-rank. Two detonation-waves also considered various amounts interaction between waves, improvements for multiple quantities limited interaction.