Model reduction of aerothermodynamic for hypersonic aerothermoelasticity based on POD and Chebyshev method

Abstract

Accurate modeling of aerothermodynamics with low computational cost takes on a crucial role for the optimization and control of hypersonic vehicles. This study examines three reduced-order models (ROMs) to provide a reliable and efficient alternative approach for obtaining the aerothermodynamics of a hypersonic control surface. Coupled computational fluid dynamics (CFD) and computational thermostructural dynamics (CTSD) approaches are used to generate the snapshots for ROMs considering the interactions between aerothermodynamics, structural dynamics and heat transfer. One ROM adopts a surrogate approach named Kriging. The second ROM is constructed by the combination of Proper Orthogonal Decomposition (POD) and Kriging, namely, POD-Kriging. The accuracy of Kriging-based ROM is higher than that of POD-Kriging-based ROM, but the efficiency is lower. Therefore, to address the shortcomings of the above two approaches, a new ROM is developed that is composed of POD and modified Chebyshev polynomials, namely, POD-Chebyshev. The ROM based on POD-Chebyshev has the best precision and efficiency among the three ROMs and generally has less than 2% average maximum error for the studied problem.

Calibrating hypersonic turbulence flow models with the HIFiRE-1 experiment using data-driven machine-learned models

“…Surrogate models have also been used in compressible aerothermodynamics (hypersonic flows), often to approximate temperatures, pressures and heating on wings and control surfaces [39,40]. They have been used within the context of aerothermoelasticity studies [41,42,43,44,45] or to design thermal protective systems [46]. In these surrogate models, the spatial variation of the fields are captured using POD bases, though a few studies have used kriging [41,40,45].…”

“…They have been used within the context of aerothermoelasticity studies [41,42,43,44,45] or to design thermal protective systems [46]. In these surrogate models, the spatial variation of the fields are captured using POD bases, though a few studies have used kriging [41,40,45]. In case of POD, the coefficient of the POD bases are modeled as a function of the environment (e.g., Mach number, altitude etc.)…”

“…Comparisons between various surrogate modeling techniques for hypersonic flow fields can be found in Refs. [39,41,44,45]. In these studies the dataset for training the surrogate model generally contained O(10) − O( 102 ) examples.…”

Parameterized Reduced-Order Models for Probabilistic Analysis of Thermal Protection System Based on Proper Orthogonal Decomposition

“…Chen et al [29,30] used three methods, namely, the radial basis function (RBF), POD-RBF and Kriging method, to construct a reduced-order aerothermodynamic modeling framework for hypersonic vehicles. Yan et al [31] proposed the POD-Chebyshev aerodynamic thermal reduced model, which can be used to predict the heat balance state of a hypersonic vehicle, with higher computational efficiency and accuracy than Kriging method. Pérez et al [32] analyzed the linear instabilities present in a defined hypersonic flow over a blunt object in thermal equilibrium.…”