Uncertainty quantification and global sensitivity analysis for hypersonic aerothermoelastic analysis

A novel parameterized model for temperature distribution is proposed. A framework for uncertainty quantification and global sensitivity in hypersonic aerothermoelastic analysis is developed based on this model. The uncertainty quantification and global sensitivity analysis in hypersonic aerothermoelastic analysis for control surface due to hypersonic aerothermodynamics is investigated in this study. Firstly, based on Computational Fluid Dynamics (CFD) technology, Navier-Stokes equation is solved to acquire the temperature distribution of the control surface. A parameter method based on the temperature distribution is built. Then Monte Carlo Simulation (MSC) method and Spare Grid Numerical Integration (SGNI) method are used to generate temperature distribution samples. aerothermoelastic is analyzed under the temperature distribution for all samples. The process of aerothermoelastic analysis is as following: structural modal under the effect of structure thermal stress and material property which is based on the temperature distribution samples is analyzed, structural modes is interpolated to the aerodynamic grid, aeroelasticity stability-boundary of the control surface is analyze in state space based on CFD local piston theory. Finally, the uncertainty quantification and global sensitivity analysis of aerothermoelasticity is analyzed. The framework is applied in two hypersonic flow cases. The analysis results show that: the variation coefficients of structure natural frequency and flutter analysis are 5.83%, 8.84% under the two cases, And the global sensitivity of the two uncertainty parameter is about 50% under the two cases. And the coupling of two parameters is about 0. Comparing with MCS method, SGNI method improved the efficiency of uncertainty analysis significantly.