Development and application of a reduced chemical kinetic mechanism for e-gasoline surrogate fuel in engine combustion modelling

Renewable synthetic fuels offer the potential to reduce global carbon dioxide emissions when used in internal combustion engines of current and future passenger car fleets. To enable the simulation of combustion and evaporation behaviour of renewable gasoline fuels, a surrogate fuel algorithm has been developed. This algorithm embeds a physical model of the fuel distillation process and accounts for key properties such as density, research and motor octane numbers, carbon-to-hydrogen ratio, and volumetric oxygen content. The surrogate palette includes isooctane, isopentane, n-heptane, toluene, pseudocumene, and 1-hexene. The fuels studied are standard RON95E10 and its synthetic counterpart, MtG-E10. A reduced chemical reaction mechanism, consisting of 548 species and 3350 reactions, is derived from the Lawrence Livermore National Laboratory mechanism using the Direct Relation Graph with Error Propagation and Sensitivity Analysis reduction method. Laminar flame speed neural networks are generated from this reduced mechanism to support the calibration of the Eddy Burn-Up combustion model within a one-dimensional computational fluid dynamics simulation framework representing a single-cylinder spark-ignited research engine. The model is calibrated and validated for the standard gasoline fuel, then applied to the synthetic fuel as well. Results show that the model accurately predicts combustion behaviour for new gasoline-like fuels without requiring recalibration of the combustion model turbulent flame speed parameters; only the laminar flame speed metamodel needs to be updated when switching fuels. For MtG–E10, the root mean square error remained below 4 bar for peak in-cylinder pressure and within 1 crank-angle degree for the 50 % mass-fraction-burned angle.