2019 Volume 18 Pages 91-100
In recent years, laminar flamelet model for compressible flows (compressible flamelet model) was proposed for investigating complicated combustion phenomena such as supersonic combustion and combustion oscillation, which require reproduction of pressure propagation. In this research, we reduce the computational cost of compressible flamelet model by making it possible to refer to thermochemical state quantities of multi-component gas and terms that include spatial gradients of chemical species’ mass fractions from the flamelet tables. Then, we also show that by using Artificial Neural Network (ANN) to develop these flamelet tables, calculation of gradient values of chemical species' mass fractions become easier. For evaluation of these methods in actual combustion field, numerical simulations of the hydrogen combustion in a scramjet test engine of the German Aerospace Center (DLR) is conducted.
