Modelling the end-use performance of alternative fuel properties in flex-fuel vehicles

Renewable fuels and fuel-optimized engines play a key role in the time- and cost-effective decarbonization of current mobility. The present work introduces a state-of-the-art mathematical model that allows for the first time an accurate estimation of fuel consumption in flex-fuel vehicle engines by considering the impact of the most significant fuel properties exclusively. These engines are optimized to a higher concentration of non-drop-in fuels such as E85. Based on the literature data, a matrix was built with fuel properties as multiple independent variables and fuel consumption as a response variable. A multilinear regression with quantitative analysis was applied to develop a fuel consumption model for FFVs. The most significant fuel properties turned out to be octane sensitivity, vapor pressure, lower heating value, and density. All properties in the final model have a unique and important impact on fuel consumption, secured by extremely low p-values (P ≪ 1 %). The model reached very high accuracy represented by R-Square of 0.994, which turned into 1.41 % of the average absolute error in internal validation and only 1.9 % in external validation. The present study shows that in all alternative fuel cases, flex-fuel vehicles performed with better fuel economy than standard spark-ignition light-duty vehicles. Moreover, high concentration alcohol blends reduce energy consumption as well as tank-to-wheel CO₂ emissions despite their higher fuel consumption. The developed model can be applied to fuel consumption estimations in FFVs from single chemical compounds to commercial fuel products including new fuel blends.