Comparison between Conventional and Non-Conventional Computer Methods to Define Antiknock Properties of Fuel Mixtures

Abstract

Research Octane Number (RON) is one of the primary indicators for the determination of the resistance of gasoline fuels to autoignition. This parameter is usually determined with a test procedure involving a standardized engine that requires expensive hardware and time-consuming tests. In this work, a set of different methods with which to determine the RON of gasoline fuel surrogates is presented, considering only computer simulations, which allows to reduce both cost and time for the evaluation. A palette of 11 chemical species has been chosen as the basis for the surrogates’ database, which will be investigated in the work, allowing the representation of the complex chemical formulation of fuels in an easier way. A simplified zero-dimensional engine model of the standard variable compression ratio is used to provide pressure and temperature, then employed to calculate RON. This is done first by means of existing methods, and then by introducing new processes concerning a simplified chemical reactor built on kinetic schemes. Finally, these different methodologies are tested against a molar weighted sum of RONs of each chemical specie, allowing to have a criterion for comparison and evaluating their real prediction capabilities.

Keywords

• antiknock properties
• RON calculation
• machine learning
• engine thermodynamics
• chemical kinetics simulation