Among the existing concepts to help improve the efficiency of spark ignition engines on low load operating points, Controlled Auto-Ignition™ (CAI™) is an efficient way to lower both fuel consumption and pollutant emissions at part load without major modifications of the engine design. The CAI™ concept is founded on the auto-ignition of a highly diluted gasoline-based mixture in order to reach high indicated efficiency and low pollutant emissions through a low temperature combustion.
The high dilution rates needed to successfully obtain CAI™ combustion imply the use of unconventional valve lift strategies. To correctly control this combustion mode, the rate of dilution has to be precisely known. From a numerical point of view, this induces the need for a CFD tool suited for air path computations, the most relevant one being a 1D simulation approach. On the other hand, the predictive simulation of the auto-ignition and combustion processes is at present only possible with 3D CFD or complex chemistry codes.
The suggested approach to build the most adapted simulation tool consists in combining 1D and 3D approaches: 1D simulation is used to simulate the gas exchange process and to provide the necessary inputs for 3D computations (e.g. volumetric efficiency, thermodynamic state, mixture composition, mass flow rate). The latter is limited to the combustion chamber and ensures an accurate description of auto-ignition, combustion and pollutant formation processes. This strategy allows to explore a wide range of operating conditions via the use of a predictive up-to-date model for each engine part while limiting the CPU time.
With this tool, the benefits provided by a fully variable valve actuation device and an EGR circuit are evaluated. It is shown that this allows an extension of the CAI™ operating range and an optimisation of the combustion timing in order to improve the fuel consumption.