Data-driven framework for input/output lookup tables reduction: Application to hypersonic flows in chemical nonequilibrium

In this paper, we present a novel model-agnostic machine learning technique to extract reduced thermochemical model for reacting hypersonic flows simulation. A first simulation gathers all relevant thermodynamic states and the corresponding gas properties via given model. The are embedded in low-dimensional space clustered identify regions with different levels of (non)-equilibrium. Then, surrogate surface from cluster-space output is generated using radial-basis-function networks. method validated benchmarked on flat-plate boundary layer finite-rate chemistry. reactive air mixture initially modeled open-source Mutation++ library. Substituting light-weight, machine-learned alternative improves performance solver by 50% while maintaining overall accuracy.