Abstract
This study develops a machine learning-optimized combustion model for the ether-based fuel PODE(5) based on the HyChem approach, which decouples fuel combustion steps into lumped pyrolysis and detailed oxidation. Due to its main-chain oxygen and high oxygen-to-carbon ratio, PODE fuels differ from conventional hydrocarbon and methyl ester fuels that HyChem has been applied to, requiring special consideration for oxygenated fuel decomposition intermediates such as CH(2)O, CH(3)OH, and HOCHO. During the model development, the differential evolution algorithm was used to adjust the reaction rate parameters of the HyChem model based on the ignition delay time. By defining optimization objectives as ignition delay times predicted by a reference model, the use of the differential evolution algorithm significantly accelerated the adjustment process of reaction rate parameters during model development. The final combustion model demonstrated results closely aligned with existing models in terms of ignition delay times, flame speeds, and oxidation products' speciation with satisfactory performance. It demonstrated the applicability of the HyChem method to ether-based fuels with main-chain oxygen.