Development of an Artificial Neural Network to Transfer Microstructural Information of Thermal Protection Systems (TPS) into Vehicle-Scale Simulations

The entry of space capsules into planetary atmospheres results in high-temperature flow around the capsule. Thermal protection systems (TPS) is required to prevent the overheating of the base vehicle structure. The interaction of hot gases with the TPS material results in a myriad of surface and volumetric processes on the material. These processes are influenced by the microstructure of the material and recent microscale simulations have been successful in capturing volumetric gas-surface thermochemistry and computing effective material properties. However, the microscale information has not been transferred into vehicle-scale simulations that still rely on empirical models developed for material response codes. The proposed effort aims to address this shortcoming by developing a robust neural network framework to transfer microscale information into material response codes and enable vehicle-scale simulations at higher fidelity. Using neural networks is a new paradigm for TPS design and could provide game-changing solutions to evaluate the performance of TPS materials.