In direct-injection gasoline engines, the internal engine processes are characterized by cyclic fluctuations, which are caused by interaction of several transient processes. Thereby, the unsteady fuel injection plays a determining role, since it strongly influences the structure of the fuel injection jets and the subsequent evolution of the downstream processes. Heretofore, the complexity of the interrelated mechanisms and their effects on the resulting initial droplet size distribution and on the partial effect-chain “evaporation/mixture formation/ignition” have been rather scarcely investigated. It is therefore urgent if not compelling to integrate the fuel jet breakup process in numerical simulations by means of direct numerical simulation (DNS) in order to extract at least the correct initial droplet size distribution which is typically required in frequently used numerical approaches based on the Eulerian-Lagrangian (EL) method.
To date, the most of DNS works on fuel atomization processes have been especially focused on the development of numerical methods which may enable to identify the fuel jet breakup products. The interface capturing methods, namely the Volume of Fluid (VoF) and the Level Set (LS) along with their combinations, are well established. In contrast, the interface tracking methods still require particularly high level of effort.
The overall objective of this sub-project (TP 4) is to understand the mechanisms by which the distinctive cyclic fluctuations of engine fuel injection jets are determined. For this purpose, numerical investigations will be performed by means of DNS based on VoF method of the complete injection system including internal nozzle flow (under consideration of cavitation and phase transition processes) and the near nozzle exit mechanisms along with the fuel primary and secondary breakup and atomization processes. From the generated database, various studies will be carried out.
First, the structures of the injection jet in a pressure chamber under atmospheric conditions will be analysed in term of macroscopic features (jet penetration, cone angle, jet volume, etc.) and with respect to breakup and atomization outcome (droplet size, droplet velocity, etc.). How the intake flow influences the fuel injection process including the primary fuel jet breakup and the first phase of the secondary jet breakup will be then analysed in detail under cyclic engine conditions. To identify and quantify the importance of these processes on the partial effect-chain “evaporation/mixture formation/ignition”, the influence of crucial parameters on the structure of the fuel injection jet and the jet breakup as well as their effects on the in-cylinder flow will be investigated. The relevant parameters will be selected in agreement with partners TP 3 and TP 5. The analysis will be realized for different cavitation numbers, fuel properties (single or multi-component, classical or biofuel), fuel injection pressures, nozzle geometries, and in-cylinder flow and temperature fields. In particular, the contribution due to the interaction resulting from successive fuel jet injections will be also addressed. Boundary conditions under considerations of different cycles will be provided by TP 5.
Relying on the comprehensive DNS database, effort will be further put in the development of coupled simulation approaches that shall account for the different processes including internal nozzle phenomena, fuel atomization, spray formation and spray evolution while saving the computational costs. Different coupling methodologies of DNS and LES (statistical coupling, direct coupling) will be designed, appraised and compared.
Thereby, information about droplet properties (size, velocity, fluctuations, joint PDF, etc.) after the spray disintegration will be made available as inflow conditions for the commonly used Eulerian-Lagrangian approach as applied in TP 5. A methodology will be developed to extract from the VoF results the initial droplet size distribution. The partner TP 5 will provide the jet surrounding flow and temperature field conditions in the in-cylinder region. Analytical models will be derived from the comprehensive statistical information of the dense spray region for a better statistical description of the fuel jet breakup process. Moreover, the findings in terms of spray atomization models from TP 4 and of sensitivities with respect to the mentioned partial effect-chain will be integrated later in TP7 to assess the impact of cyclic fluctuations on knocking combustion analysis.
For validation, the achievements from TP4 will be first compared with ECN database for the Delphi multi-hole nozzle (www.sandia.gov/ecn). From TP 3 valuable experimental data of the fuel injection jet close to the nozzle, e.g. spray penetration length and angle, droplet size distributions at the end of primary and secondary breakup as well as correlations of droplet size, droplet velocity and gas velocity, all evaluated under engine operating conditions, will be supplementary used.