CFD Modelling of Direct Gas Injection Using a Lagrangian Particle Tracking Approach

Abstract: CFD simulations of direct gas injection, especially in large dual-fuel engines, can be expensive both regarding time and computational power. The nozzle area needs to be resolved with a fine mesh to capture all phenomena and for a full engine model this results in a large amount of cells. A method using a Lagrangian Particle Tracking (LPT) approach was developed to handle gas injection by injecting gaseous parcels into the domain. The gaseous LPT method was implemented by modifying the LPT solver for liquid droplets in dieselFoam, which was already present in OpenFOAM 2.0.x to minimize development efforts. The method was evaluated by comparisons with RANS simulations of fully resolved subsonic jets in a simple chamber geometry, for different cases with varying inlet velocities and initial chamber conditions. It was found that despite that the gaseous LPT method under-predicts the spreading of the jet as compared with the fully resolved approach, resulting in a longer penetration length, the method provides overall reasonable trends regarding velocities and gas mass fraction. Therefore it was found in this thesis that, with a few modifications to the existing dieselFoam solver, it is possible to model direct gas injection of subsonic jet with reasonable results.