Full system predictions (FSPs) will be conducted using our coupled full-physics aero-thermo-chemo-mechanics sim- ulation capability, which will comprise: our hypersonic aero-thermo-chemistry flow solver Exasim; the rarefied, non-equilibrium direct simulation Monte Carlo (DSMC) solver SPARTA developed by Sandia National Laboratories; and our thermo-chemo-mechanics solver ΣMIT. The individual components and coupled facility will be subject to a comprehensive Verification & Validation/Uncertainty Quantification plan.

Hypersonic flow simulation. Hypersonic flow simulations will be performed using Exasim, a Navier-Stokes (NS) solver developed by PIs Nguyen and Peraire, and SPARTA, which offers a state-of-the-art DSMC implementation with massive parallel capabilities. We plan to use DSMC simulations to (1) assess the validity of the continuum assumption in our NS solver, (2) improve and extend the modeling capabilities of our NS solver, and (3) in a standalone manner when the NS description is found to be insufficient. For localized rarefied flows, we will employ a hybrid approach that couples the SPARTA and Exasim solvers using the Modular Particle-Continuum (MPC) method. Additionally, we will develop mesh adaptivity capabilities to resolve steep gradients and shock structures in hypersonic flow simulations.

To deal with thermochemical nonequilibrium flows in hypersonic flight environments, we plan to develop and implement consistent nonequilibrium chemistry models in Exasim and SPARTA. In particular, we will implement both Park’s models and quantum kinetic (QK) models.

For treating the solid surface, we will implement gas-surface chemistry models to accurately describe the flow in the near-wall region and account for gas-surface interactions, chemistry, pyrolysis and ablation. Consistency be- tween the continuum and DSMC models will be one focal point of this work. Another will be the physical consis- tency/compatibility of these models with the solid-side treatment, while a third will be a mathematically consistent formulation for their coupling.

Solid mechanics. Our solid mechanics solver, ΣMIT, builds on a thermodynamically consistent framework for bulk and interphase multiphysics and fracture driven by thermo-chemical processes. The numerical approach is based on a high-order discontinuous Galerkin / Cohesive Zone Model (DG/CZM) formulation that enables robust and scalable simulation of material response and fracture. This DG approach overcomes key limitations of traditional CZMs and has been successfully applied and extended by other research teams, including at national laboratories.

We have extended our DG/CZM library to incorporate coupled multiphysics relevant to TPS response, including pyrolysis and oxidation. Our framework naturally captures general species transport in both the bulk and along in- terfaces such as cracks and grain boundaries, and couples chemical reactions, heat transfer, and fracture mechanics across scales.

To address the unique challenges posed by carbon pyrolysis and SiC oxidation/nitridation, we will integrate coarse- grained quantum mechanical calculations, molecular dynamics, and continuum modeling. We plan to develop general-

purpose machine-learned interatomic potentials across a broad range of conditions, enabling detailed atomistic de- scriptions. These will be upscaled via physics-informed surrogates using equivariant neural networks and manifold learning techniques to model thermo-chemo-mechanical behavior, including stress-strain response and fracture. Par- ticular focus will be placed on modeling high-temperature effects such as swelling and creep, which are critical to understanding intergranular cracking and related failure mechanisms. In the first year, we will be able to demonstrate coupled thermo-chemo-mechanics responses resulting from pyrolysis and ablation of C, and passive/active oxidation of SiC using our implementation of legacy models developed by our team and others. In subsequent years, our overar- ching capability will incorporate higher-fidelity models resulting from the constitutive/surrogate multiscale modeling efforts of these materials. This will include embedded chemistry dimensionality-reduction manifold-learning sur- rogate models, interfacial multiphysics models, and microstructural-level heterogeneity effects (intergrain fracture, fiber-matrix delamination, anisotropy, etc.).

Coupled fluid-structure interaction. We plan to develop a suite of algorithmic coupling strategies, including splitting and staggered schemes, boundary-matching and non-matching meshes with specialized boundary data interpolation and transfer schemes across solvers; as well as software integration strategies assisted by the CS tool suite. We will improve the GPU performance and scalability of the individual and interface codes by a tight integration with the compiler, programming language, and performance portability technologies. Special emphasis will be given to the implementation of fluid-solid interface boundary conditions including consistent thermo-chemo-mechanics mass, momentum and energy balance (e.g., re-radiation, blown gas re-entrainment). Furthermore, radiation will be taken into account in coupled fluid-solid simulations to ensure the energy balance on the vehicle surface by incorporating the radiated surface flux.

Atomistic simulations. One of the most important challenges associated with the simulation of the overarching application is the disparity between the atomistic scale, which determines the material constitutive behavior and is thus central to its behavior, and the material damage (oxidation, ablation) characteristic scale. The magnitude of this disparity necessitates the use of coarse-graining approaches that bridge the time- and length-scale gaps. The first coarse-graining approach involves using density functional theory (DFT) and ab initio molecular dynamics (AIMD) simulations to develop variable-charge machine-learned interatomic potentials (MLIPs) for use in classical MD sim- ulations. The second coarse-graining approach involves using ab initio and classical MD simulations to develop gas- surface chemistry models and the thermo-chemo-mechanics constitutive models. These models are then integrated into the coupled FSI solver to predict the key quantities of interest.