3D-CFD RANS Methodology to Predict Engine-Out Emissions with Gasoline-Like Fuel and Methanol for a DISI Engine 2022-24-0038

Renewable fuels, such as bio- and e-fuels, are of great interest for the defossilization of the transport sector. Among these fuels, methanol represents a promising candidate for emission reduction and efficiency increase due to its very high knock resistance and its production pathway as e-fuel. In general, reliable simulation tools are mandatory for evaluating a specific fuel potential and optimizing combustion systems. In this work, a previously presented methodology (Esposito et al., Energies, 2020) has been refined and applied to a different engine and different fuels. Experimental data measured with a single cylinder engine (SCE) are used to validate RANS 3D-CFD simulations of gaseous engine-out emissions. The RANS 3D-CFD model has been used for operation with a toluene reference fuel (TRF) gasoline surrogate and methanol. Varying operating conditions with exhaust gas recirculation (EGR) and air dilution are considered for the two fuels. The laminar flame speed for the fuels has been tabulated by means of detailed chemistry 1D-flame calculations. The G-equation model is used to simulate flame propagation in combination with chemical kinetics to estimate the emission species correctly. The resulting G-equation calibration is compared between the two fuels. Emission results are analyzed, validated with experiments, and compared between the two fuels. The methodology shows a good overall predictivity regarding trends and absolute values. Simulated carbon dioxide (CO₂) shows to be mainly within a 4 % deviation from the measurements. Oxygen (O₂) deviations are within 3 % at lean operation and higher at stoichiometric conditions due to the low overall oxygen content. The maximum nitrogen oxides (NO_x) deviation for the TRF is 22 %, while higher deviations are observed for methanol up to 42.3 %. Total hydrocarbon (THC) emissions are mainly below 20 % deviation. Higher carbon monoxide (CO) deviations are observed due to high CO sensitivity to mixture formation prediction, even if the trends with EGR and air dilution are correctly reproduced. Overall, the methodology shows good potential for virtual pollutant evaluation, assessment of emission reduction strategies, and development of engines with methanol fuel.