Equivalent thermophysical properties identification of the multilayer thermal protection structure based on the Bayesian network

In the paper, an equivalent thermophysical property identification method for the multilayer thermal protection structure (MTPS) is proposed based on the Bayesian network (BN), so as to address the uncertainty of multiple thermophysical properties and the interlayer thermal contact resistance (ITCR) of the MTPS. In the method, an equivalent thermophysical property is proposed to deal with the ITCR between different layers. In order to alleviate the ill-posedness of the thermophysical property identification, the key thermophysical properties of different materials in the MTPS are first determined by sensitivity analysis, and then based on the BN, the thermophysical properties are identified only in their sensitive intervals chronologically, which reduces the number of thermophysical properties to be identified simultaneously. Finally, based on the identified thermophysical properties and the equivalent treatment of the ITCR, the internal temperature response is further predicted under different boundary conditions. Numerical and experimental tests were conducted to validate the proposed method. The maximum error of the temperature response prediction for the numerical case with noise and the experiment was 7.20 % and 7.52 %, respectively.