GALCIT Colloquium

The simulation of the aerothermal environment surrounding vehicles moving at hypersonic speeds is a complex problem due to its multi-physics and multi-scale nature. Accurate modeling of these systems has been limited by the lack of reliable physical models for the thermochemical and transport processes dominating the flow dynamics, as well as the predictive capabilities of existing models that often rely on simplistic comparisons to legacy experimental measurements with poorly characterized accuracy. Recent advancements in computational chemistry and increased computational resources have enabled the development of realistic models based on molecular-scale dynamics anchored in fundamental physics. This talk will showcase the use of Physics-Aware Coarse Graining and the Maximum Entropy principle as powerful tools for deriving new macroscopic conservation equations, energy exchange terms, and chemical production rates for atmospheric entry plasmas. The systematic hierarchical coarse-graining of the kinetic equations enables the novel concept of Adaptive Mesh and Model Refinement of the Phase Space, which allows for the automatic tailoring of model complexity as a function of local plasma conditions. Given its mathematical structure, this platform enables the implementation of acceleration strategies based on physics-informed learning. Key aspects of model development will be covered, including: (1) utilizing ab-initio quantum calculations to construct high-fidelity, physics-based models; (2) defining reduced-order models for simulating 2D and 3D flows, such as coarse-grain modeling; and (3) validating physical models and determining the uncertainty in their predictive capabilities using the latest developments in Uncertainty Quantification (UQ) algorithms.