Methanol fuelling of a spark-ignition engine: Experiments and 0D/1D predictive modelling for combustion, performance, and emissions

Methanol is a promising fuel for decarbonising hard-to-electrify propulsion sectors because of the sustainable production pathways, and its physical and combustion properties (energy density, fast burning velocity, high knock resistance). Among the simulation models used to support timely development of methanol propulsion systems, phenomenological models can be used for predicting performance, combustion, and emissions in an engine combustion chamber while offering a reduced computational effort. In this work, 0D quasi-dimensional models are applied within a 1D engine model in order to verify the prediction capabilities with regard to an experimental dataset from a spark-ignition direct-injection single-cylinder research engine fuelled with methanol. Spark-ignition combustion is simulated through a fractal combustion model within a quasi-dimensional combustion chamber formulation. Different operating conditions with varying engine load, speed, air–fuel ratio, and external gas recirculation fraction are used to validate the model. Unburned hydrocarbon (uHC) emissions are simulated with a phenomenological model accounting for contributions from piston top-land crevice and wall flame quenching. An additional model to account for the contribution of piston top-land oxidation to CO formation is also implemented. For carbon monoxide (CO) and nitrogen oxides (NO_x) emissions, multi-zone 0D simplified chemical kinetic approaches are applied in the burned zone and the temperature distribution is verified with 3D-computational fluid dynamics (CFD) simulations of the same engine. The combustion models are demonstrated to reproduce combustion duration trends over all the operating conditions investigated while emission prediction deviations on average are within 11.7 % for uHC, 17.7 % for NO_x, and 16.7 % for CO. The main novelty of the presented numerical methodology is the validated capability to correctly follow the emission and combustion trends of the considered methanol engine under strongly varying operating conditions without any case-by-case tuning. Finally, the predictivity of the proposed model is verified on additional operating conditions fuelled with gasoline instead of methanol. The results show the capability of the model to correctly predict both combustion and pollutant emissions, requiring minor adaptation of the tuning constants. This demonstrates the model's reliability and suitability for investigating of further operating conditions, identifying optimal calibration settings, or analysing gasoline–methanol blends as well as potentially other fuels.