Numerical study on the influence of gasoline properties and thermodynamic conditions on premixed laminar flame velocity at multiple conditions

In this paper, the effects of two gasoline key properties, research octane number (RON) and fuel sensitivity S, on the laminar flame velocity (LFV) were studied. Three environmental parameters, temperature, pressure and equivalent ratio 4), were also taken into consideration at multiple conditions. An effective hybrid computation method, combining the flame kinetics model and the machine learning (ML) algorithm, was proposed for the parameter study. The flame kinetics model coupled with a well-validation TRF reaction mechanism to model the flame propagation characteristics; four ML algorithms, Random forest (RF), Gradient boosting decision tree (GBDT), Support vector regression (SVR), and Adaboost were adopted to establish the data mapping between investigated parameters with fuel's LFV. The result shows all ML models perform well on the test set, especially for the GBDT algorithm, which with R-2 of 0.9984 and RMSE of 0.0084. Based on the power exponential formulas, the LFV correlations concerning temperature, pressure and equivalent ratio were concluded. It also found the influence mechanism of RON and S on LFV was controlled by the equivalence ratio. Besides, the tri-component diagrams of the LFV with the RON-S-phi under typical environments were analyzed, and the statistical information in different flame patterns was discussed in detail. (C) 2021 Elsevier Ltd. All rights reserved.