Machine Learning based reduced models for the aerothermodynamic and aerodynamic wall quantities in hypersonic rarefied conditions

Since their development at the end of 50s, panel methods were widely used for fast simulation aerospace objects reentry. Although improvements proposed continuum regime formulations, bridging functions usually employed in transitional did not go through major changes since then. With current interest designing Very Low Earth Orbit satellites and more efficient reentry vehicles, a greater level preciseness is now required computation aerodynamic aerothermodynamic wall quantities rarefied regime. In this context, paper presents new approach to build Machine Learning based surrogates going from choice design variables Design Experiments, models training evaluation. Hence, kriging Artificial Neural Networks are respectively trained predict pressure heat flux stagnation coefficients, pressure, friction coefficient distributions portion any shape’s