Construction of a skeletal multi-component diesel surrogate model by integrating chemical lumping and genetic algorithm

Towards high-fidelity combustion chemistry modeling of real diesel fuel, this study aims to construct a skeletal multi-component kinetic model with higher accuracy and heavy hydrocarbons (n-hexadecane, iso-cetane, n-octadecane, alpha-methylnaphthalene, decalin), which is feasible for engine computational fluid dynamics (CFD) applications. The developed skeletal model with 220 species and 1109 reactions is composed of lumped fuel-specific sub-mechanisms and a skeletal C-0-C-3 core mechanism derived from the well-validated detailed NUIG-Mech 1.1 mechanism. Chemical lumping integrated with skeletal reduction methods and reaction pathway analysis is used to develop compact fuel-molecule sub-models. Ignition delay and species concentration are then taken as targets simultaneously to optimize the original model automatically using genetic algorithm (GA) within given parameter ranges. Eventually, the optimized kinetic model is validated against various experimental results of pure components, their mixtures and real diesel fuels, including ignition delay times in both shock tubes and rapid compression machines, species concentration profiles in jet-stirred reactors, and laminar flame speeds in the counterflow configurations as well as actual engine data. In addition, it should be noted that due to the absence of test data on pure n-octadecane, the sub-model of n-octadecane is not directly validated in this work, but compared with measurements of real diesel. Overall, the simulation results are in good agreement with experiments under a wide range of engine-like conditions, suggesting that the present skeletal multi-component mechanism can be applied to mimic autoignition and oxidation of actual diesel. As such, this will provide a more accurate and robust combustion model for the development and optimization of advanced engines with high efficiency and ultra-low emissions.

Automated summary

This study aims to construct a high-fidelity simplified multi-component kinetic model for simulating the autoignition and oxidation processes of real diesel fuel. By using chemical lumping, skeletal reduction methods, and genetic algorithm optimization, a model that is in good agreement with experimental results can be obtained.