This paper presents a phenomenological quasi-dimensional model of the processes that lead to charge preparation in a Direct-Injection Spark-Ignition (DI-SI) engine, focusing on the physics of atomization and drop evaporation, spray development and the mutual interaction between these phenomena. Atomization and drop evaporation are addressed by means of constant-diameter drop parcels, which provide a discrete drop-size distribution. A discrete Probability Density Function (PDF) approach to fuel/air mixing is proposed, based on constant-mixture-fraction classes that interact with each other and with the drop parcels. The model has been developed in the LMS Imagine.Lab Amesim™ system simulation platform for multi-physical modeling and integrated in a generic SI combustion chamber submodel, CFM1D [15], of the IFP-Engine library.
The validation of the approach is performed on an experimental test case consisting of a high pressure isooctane injection in a constant volume vessel for which mie-scattering and high-speed schlieren visualizations for different thermodynamic conditions were performed at IFPEN within the framework of the French government MAGIE R&D project. Liquid and vapor penetration as well as spray angle data from experiments are then used to tune the RANS CFD simulations performed with the IFP-C3D code. CFD provides further data which is not directly available from the experiments such as drop size and charge distributions as well as spray properties outside the optical measurement field, which are then used to tune and validate the 0D model.
Good accordance is found between validation data and the results obtained with the proposed model showing the advantages of a detailed - though phenomenological - description of the main phenomena involved.