Mixing Model of Jet Fuel Based on Explicit Equations and Neural Networks and Its Application

Quanhong Xu,Zhentao Liu,Chi Zhang,Yuzhen Lin
DOI: https://doi.org/10.11715/rskxjs.R201703025
2017-01-01
Abstract:Surrogate model of jet fuel is one of the most important premises to high-fidelity numerical simulations of spray combustion in aero-engine.For complex physico-chemical properties of jet fuel,a strategy was proposed based on a mixing model combining explicit equations and implicit neural networks.Hydrogen-carbon ratio, viscosity,and distillation curve were chosen as the target properties,and the method of jet fuel surrogate formula-tion was developed,which could well simulate the target properties of mixture.The mixing model was used to build the surrogate of coal-derived synthetic jet fuel,and the difference of physico-chemical properties between surrogate components and real fuel could meet the simulating demands.
What problem does this paper attempt to address?