H Radical Sensitivity-Assisted Automatic Chemical Kinetic Model Reduction for Laminar Flame Chemistry Retaining: A Case Study of Gasoline–DME Mixture under Engine Conditions

Flame propagation commonly exists in a wide range of engine operating modes. In chemical kinetic model reduction, the prediction of flame speeds must be targeted when engines involve flame propagation. However, the time-consuming nature of 1-D flame code running largely limits the feasibility of trial-and-error type reduction approaches. In this study, an improved automatic reduction scheme is proposed by adding a normalized H radical sensitivity with rate constants. By comparison with flame speed sensitivity, the combination of H radical and flame speed sensitivity can be more accurate to construct the species group relevant to laminar flame chemistry. Then, the newly proposed reduction methodology is applied for the spark ignition-controlled autoignition hybrid combustion (SCHC) with dimethyl ether (DME) as the pilot injection into gasoline, which proves promising in the engine performance by flexible controlling and stabilizing the combustion process. By constructing a new detailed chemical kinetic model for PRF–DME mixtures (348 species), a 143-species skeletal model is developed by considering both ignition and laminar flame. 3D-CFD simulations of experimental SCHC cases show that the detailed and skeletal models can capture the engine phenomena accurately. The results of a 103-species skeletal model reduced without flame speed target indicate that the flame propagation must be emphasized in the SCHC mode