Research On Probabilistic Characteristics Analysis Method Of Thermal Protection System Considering Heat Flow Uncertainty

Thermal protection system is one of the most important subsystems of hypersonic vehicle.Thermal response of thermal protection system is random due to uncertainties from manufacturing error,material dispersion and aerodynamic thermal loads,which brings great challenges to engineering design and reliability evaluation.Aerodynamic thermal uncertainty is a stochastic process in terms of time history.Therefore,it is difficult to build surrogate models based on random variables using the conventional methods.It is of great interest to find a new surrogate model,which is suitable for probabilistic analysis of aerodynamic thermal uncertainty in thermal protection system.Therefore,based on the series decomposition of stochastic process,a high-precision surrogate model for transient heat conduction analysis is constructed in this paper,and the probabilistic characteristics and reliability evaluation method of thermal protection system considering the uncertainty of heat flow are also investigated.In terms of series expansion of random process,the theories of Karhumen-Loeve(K-L)series expansion method and(OSE)method are introduced,and the error analysis of the two methods is carried out with numerical examples.The computational efficiency of the two methods is also studied.The results of numerical examples indicate that both methods can be used to describe the original random field,and the K-L method is more accurate than the OSE method,whereas the OSE method is superior in computational efficiency.To build the surrogate models based on random process,this paper introduces the strategy of building surrogate model based on random process,and demonstrates the feasibility of the method through numerical examples,based on which the error of the approximate model is analyzed.Numerical examples show that the method based on stochastic process has satisfactory approximate accuracy.The increase of the kept orders will lead to the decrease of the accuracy and efficiency of the surrogate model.Therefore,to ensure the accuracy of the series expansion method describing the random process,the surrogate model should be built with as few orders as possible.The probabilistic characteristics and reliability evaluation method for the thermal protection system under random thermal loads are also investigated in this paper.The presented surrogate model is verified to solve the transient heat conduction problem with high precision.By using this model,the probabilistic characteristics of the thermal responses are investigated.Results indicate that through the geometry optimization or control thermal load dispersion can improve the thermal reliability of thermal protection system.