Search

In this work, the Engine Combustion Network Spray G injector was mounted in the Darmstadt optical-accessible engine to study phenomena typical of multi-hole, early direct-injection events in spark-ignition engines characterized by tumble flow charge motion. Dedicated experimental measurements of both in-cylinder spray morphology and flow velocities before and after the injection process were carried out to assess the adopted numerical setup under real engine conditions. A dynamic secondary breakup model from the literature was coupled with an atomization multi-motion regime model. The model was validated against state-of-the-art ECN Spray G experiments for a constant-volume chamber under low evaporating condition. Then, the simulation of the spray injection in the engine was carried out and the achieved results were compared against the experimental data. Overall, good agreement between experiments and simulations was observed for the spray morphology and velocity fields in both cases. With reference to engine calculations the intake flow was seen to induce spray asymmetry. A partial vortex generated during the intake phase on the tumble plane interacts with the spray, developing into a full vortex which induces an upward flow that stabilizes the spray. The upward flows below the intake valve increase the dilution of the plume outside the tumble plane, which therefore exhibits reduced penetration. Moreover, the intake valves protect from the energetic intake flow the recirculation vortex generated at the tip of the plumes that lie outside the tumble plane. The intake flow helps fuse the vapor fuel clouds of the individual plumes near the injector tip, obtaining a vapor fuel with a shape like that generated by a horseshoe multi-hole injector. Finally, a phenomenological model of the interaction between the multi-hole injector jets and the engine intake flow was introduced to describe the spray evolution in a typical DISI engine.

Internal combustion engines (ICEs) are an essential and common power source for personal mobility and the transport of goods. Despite their improvement over the last decades, they are still responsible for a relevant fraction of greenhouse gas and pollutant emissions. It is thus crucial to optimize them to enhance their efficiency and minimize emissions. This task is far from trivial because different physical phenomena, from a strongly unsteady turbulent multiphase flow to mixture ignition and combustion, interact in a cause-effect chain inside an ICE. Hence, it is challenging to study a single phenomenon individually, and engine studies must combine several research areas into a single multidisciplinary approach. In this context, numerical investigation methodologies can flank the experimental investigations as they make it possible to analyze many aspects of the engine that can be prohibitive to measure. This thesis develops a numerical method for engine simulation, striving toward accuracy, flexibility, stability and computational efficiency. All of these are necessary for proficiency in the multidisciplinary numerical study of the engine, in both academia and industry. This framework is used to investigate the phenomena preceding the combustion phase in the Darmstadt research engine, a Direct Injection Spark Ignition (DISI) optically accessible engine. In this context, new mesh motion methodologies are developed that are suitable for DISI engines, characterized by complex geometries. Such methodologies are designed to achieve maximum accuracy for RANS and LES simulations with minimum effort. Moreover, methodologies for multi-cycle simulations are developed, including pre-processing and post-processing methods. The strategy is fully validated by simulating a full cycle of the Darmstadt engine running in motored condition at different operating conditions, using a RANS approach. Furthermore, this work investigates spray dynamics, including the spray impingement a