LAUSR

Abstract Knocking is the main factor restricting the further improvement of the gasoline engine. Ethanol has high Octane Number and antiknock characteristics. Adding ethanol to gasoline makes the blend withstand higher compression ratio so that the engine can output more power without knocking. In this study, for the toluene primary reference fuel/ethanol/diisobutylene (TPRFE/DIB), the suitable detailed mechanism of ethanol was selected first by numerical simulation. Under knocking condition of low-temperature and high-pressure, the detailed mechanism of ethanol was amended and reduced by Rate of Production and Sensitivity analysis, and coupled with the reduced mechanism of the toluene primary reference fuel/diisobutylene (TPRF/DIB) amended under the same condition to construct the reduced mechanism of TPRFE/DIB. Then, the effect of ethanol on the Research Octane Number (RON) of different gasoline surrogates was revealed. On this basis, Response Surface Methodology and Back-Propagation Neural Network were used to construct two models for predicting the RON, and it is verified that both models have high prediction accuracy.