TY - JOUR

T1 - Modeling of Transport Coefficients in Nonequilibrium High-Speed Flows Using Machine Learning

AU - Истомин, Владимир Андреевич

AU - Павлов, Семён Анатольевич

PY - 2025/8/6

Y1 - 2025/8/6

N2 - Abstract: The accurate computation of transport coefficients in the simulation of high-speed reacting gas flows is fundamentally important for analyzing heat transfer processes in various gas dynamics problems, such as the thermal protection calculations of reentry vehicles. The most physically correct models within the continuum approximation are multi-temperature and state-to-state approaches, which are constructed using the kinetic theory of transport and relaxation processes. However, these models possess significant computational complexity. For this reason, simplified one- and two-temperature approaches are commonly used, which are still computationally expensive for real-time modeling. One effective approach to applying such models is the regression of transport coefficients using machine learning. This work applies this approach to the modeling of viscosity and thermal conductivity coefficients for an 11-component ionized air mixture in the case of high-speed flow around Fire II reentry spacecraft.

AB - Abstract: The accurate computation of transport coefficients in the simulation of high-speed reacting gas flows is fundamentally important for analyzing heat transfer processes in various gas dynamics problems, such as the thermal protection calculations of reentry vehicles. The most physically correct models within the continuum approximation are multi-temperature and state-to-state approaches, which are constructed using the kinetic theory of transport and relaxation processes. However, these models possess significant computational complexity. For this reason, simplified one- and two-temperature approaches are commonly used, which are still computationally expensive for real-time modeling. One effective approach to applying such models is the regression of transport coefficients using machine learning. This work applies this approach to the modeling of viscosity and thermal conductivity coefficients for an 11-component ionized air mixture in the case of high-speed flow around Fire II reentry spacecraft.

KW - Fire II

KW - OpenFOAM

KW - neural network

KW - regression

KW - transport coefficients

UR - https://www.mendeley.com/catalogue/f536ac23-0100-3a06-996e-8fda0e7280f7/

U2 - 10.1134/s1063454125700426

DO - 10.1134/s1063454125700426

M3 - Article

VL - 58

SP - 428

EP - 437

JO - Vestnik St. Petersburg University: Mathematics

JF - Vestnik St. Petersburg University: Mathematics

SN - 1063-4541

IS - 3

ER -