OFELIE

In 2025, hybridized light duty vehicles will still largely operate with direct injection (DI) spark ignition (SI) engines. Although DI reduces SI CO2 emissions, it leads to large soot particle emissions, formed by a pool fire due to DI. A second lever to reduce the global SI engine emissions is the progressive replacement of fossil fuels by ethanol in the first place, than by various other bio resourced oxygenated fuels in the longer term. Although these fuels contribute to soot reduction, they lead to increased aldehyde emissions and their impact on NOx remains unclear. As these pollutants will be regulated by the future Euro7 norm, car manufacturers need to improve their engines and this implies more and more the usage of CFD codes (Computational Fluid Dynamics). At the same time, codes available today fail to predict these pollutants correctly: present models allow to describe the pressure evolution in the cylinder but the semi-empirical soot models employed are not predictive and NOx and aldehydes chemical paths induced by oxygenated fuels are not described in present mechanisms. The OFELIE project proposes a collaborative research integrating experimental and modelling tasks aiming at developing such a CFD tool. For this purpose, we propose to represent oxygenated fuels by an iso-octane/n-heptane/toluene surrogate representing standard gasoline, to which ethanol is added to represent E10 to E100 type fuels, or butanol to represent other bio-resourced fuels. This fuel will be used for all experiments and simulations of the project, thus allowing a step-by-step validation of models. First, laminar sooting flame experiments will be performed at PC2A laboratory to measure aldehydes, soot and their precursors, the PAH, thanks to advanced diagnostics (jet cooled LIF, LII, CRDS etc…). Soot will also be measured in low turbulent diffusion flames, closer to pool fire combustion. These measures will then be used at LRGP laboratory to develop a complete chemical mechanism for the surrogate oxygenated fuel and the target species. The automatic mechanism writing by computer-aided generation system EXGAS will be used along with quantum chemistry codes for the most uncertain reactions. CFD of a pool fire in engine conditions has never been investigated with care to our knowledge and requires many complex sub-models that are difficult to validate (description of multi-component evaporation, liquid jet, film formation and displacement, soot formation above the film). This is why an unprecedented experiment is proposed at IFPEN in a pressurized vessel to reproduce a gasoline jet impinging on a plate, flame propagation up to this plate and pool fire formation above it. Optical diagnostics will be developed to measure the film thickness time evolution (2D-UV absorption, RIM) and soot repartition above the film (2D-DBI). These experiments will then be used at IFPEN to validate and set the parameters of the two-phase flow CFD models mentioned previously and to validate the sectional soot model developed at IFPEN in the ANR ASMAPE project, using the kinetic mechanism developed in OFELIE. Simulations will be performed with the RANS (Reynolds Averaged Navier-Stokes) approach, the one used by car manufacturers today, but also in LES (Large Eddy Simulation), an approach allowing a more refined description of turbulent phenomena. All IFPEN developments and simulations will be performed with the commercial code CONVERGE from CSI. This code is also the one used by PSA and Renault for their engine developments, thus allowing a direct transition from the research of OFELIE to the industrial usage. Finally, all these models and mechanisms will be evaluated by PSA, Renault and IFPEN on their respective DI-SI engine databases thus providing a final evaluation of the progress brought by OFELIE in the prediction of soot, NOx and aldehydes in DI-SI engines.