Physics-Infused Reduced-Order Modeling for Analysis of Multi-Layered Non-Decomposing Ablating Hypersonic Thermal Protection Systems

This work presents a physics-infused reduced-order modeling (PIROM) framework towards design, analysis, and optimization of non-decomposing ablating hypersonic thermal protection systems (TPS), and is demonstrated via the modeling of transient thermo-ablative responses of multi-layered hypersonic TPS. The PIROM architecture integrates a reduced-physics model (RPM) as the building block, which is based on the lumped capacitance model (LCM) coupled to a surface recession model (SRM). The RPM provides a low-fidelity estimate of the thermo-ablative response of the TPS, based on average temperatures and one-dimensional surface displacements. This RPM is extended with data-driven hidden dynamics that are formulated via a systematic coarse-graining approach rooted in the Mori Zwanzig (MZ) formalism, and that are learned from high-fidelity simulation data. Therefore, while the LCM and SRM capture the dominant physics of the ablating TPS response, the correction terms compensate for residual dynamics arising from higher-order non linear interactions and temperature-advection effects due to surface recession. The trained PIROM consistently achieves errors of ≈0.5% for a wide range of extrapolative settings of design parameters involving time-and-space varying boundary conditions and SRM models, and improves by an order of magnitude by the LCM alone. Moreover, the PIROM delivers RPM-level computational costs, enabling evaluations that are two orders of magnitude faster than the high-fidelity full-order model (FOM). These results demonstrate that PIROM effectively reconciles the trade-offs between accuracy, generalizability, and efficiency, providing a promising framework for optimizing multi-physics dynamical systems, such as TPS, under diverse operating conditions.