Formulating of model-based surrogates of jet fuel and diesel fuel by an intelligent methodology with uncertainties analysis

Abstract

In this study, an intelligent model to formulate fuel surrogates was proposed which includes intelligent surrogate blend optimizer and optimization module of chemistry kinetic mechanism. The components of surrogates were selected from the library of candidates with given compositions by the intelligent surrogate blend optimizer based on the targeted fuel properties. Hence, new jet fuel surrogate (JFS) and diesel fuel surrogate (DFS) models were formulated consisting of five components (n-dodecane/iso-octane/isocetane/decalin/toluene), with specified compositions of each component, respectively. The new surrogates contain major components of the target practical fuels and inherently emulate the key properties including liquid density, viscosity, surface tension, cetane number, hydrogen-carbon ratio, molecular weight, lower heating value and threshold sooting index. Moreover, a robust skeletal oxidation mechanism of fuel surrogate composed of five components was updated and multi-objective optimization algorithm of non-dominated sorting genetic algorithm II (NSGA-II) was introduced to achieve self-adaptive tuning of Arrhenius pre-exponential factor in fuel-related sub-mechanism. Thus, the optimal rate constants could satisfy the predictions of ignition delay times and species concentrations simultaneously. Then Generalized Polynomial Chaos (gPC) was employed to minimize the uncertainty of stochastic input coefficients and its propagation towards ignition behaviours. Consequently, the new surrogate models were well validated against available tested data from practical fuels, and better performance of current JFS model was highlighted when compared with the prediction results by previous model. It is feasible to employ this intelligent model to formulating surrogates for more practical fuels.