A Data-Driven Approach For Real-Time Estimation of Material Uncertainty

Modern Thermal Protection Systems (TPS) used for planetary exploration missions often utilize lightweight porous materials as their outer isolating layer. Due to its stochastic nature, such material exhibits a high level of material variability in various properties. Therefore, it is of paramount importance to accurately estimate the uncertainty margin and to understand how material response is affected. In this study, a well-established data-driven algorithm is used to estimate the FiberForm conductivity by using real-time experimental data. By combining the real-time experimental data and a high-fidelity simulation model, the inherent material property is obtained via the proposed method-the retrospective adaptation algorithm. The results also indicate that this methodology is of great potential and can be applied to a wide range of engineering problems, including parameter evaluation, and experimental data processing.