Nonprobabilistic uncertain model updating and optimization design of thermal protection system

Reusable launch vehicles are subjected to intense aerodynamic heating during the hypersonic re-entry stage. Thus, thermal protection system (TPS) design methods that consider uncertainty have become increasingly important in recent years. In this study, a nonprobabilistic TPS optimization design that takes into account deviations in temperature-dependent thermophysical property parameters is carried out with corresponding experimental verification. An improved Latin hypercube design (ILHD) is first proposed to solve the sampling problem in the case where the distribution domains of correlated uncertainty parameters interfere with each other. Based on the ILHD, uncertainty and sensitivity analyses of the TPS heat transfer are performed, in which the importance and effect trends of uncertainty parameters to responses are clearly identified. In terms of both computational costs and accuracy, the ILHD method has a significant advantage because of the excellent abilities of random sampling to satisfy certain constraints, space-filling, and nonlinear response-fitting. The ILHD's superiority in uncertainty and sensitivity analyses is also proved compared with response bounds by sampling. During experimental verification, an uncertainty-based model updating method is proposed to modify the heat transfer numerical model of test pieces. Finally, a lighter design is obtained and the correctness and validity of applied methods are verified.