COUPLED 2023

Multi-model Investigation of the In-cylinder Direct-injection Engine Phenomena using Open-source Code

  • Pati, Andrea (TU Darmstadt - STFS)
  • Hasse, Christian (TU Darmstadt - STFS)

Please login to view abstract download link

Optimizing the combustion process inside Direct-Injection Spark-Ignition (DISI) engines is essential to reduce their environmental impact. Therefore, gaining full knowledge of all the involved phenomena is a priority, and numerical methodologies are an essential tool for this scope. In the context of engine simulation, the main challenge is the number of interconnected phenomena and, therefore, the number of models and phases involved in the simulations. However, despite the popularity of open-source codes, there is a lack of engine investigations covering different phenomena with open-source code. A library for the OpenFOAM code has been specifically developed to investigate the different phenomena occurring in DISI Engine. This code is applied to the Darmstadt Engine, an optically accessible research engine. First of all, advanced methodologies for mesh motions are developed. In addition, also a decomposition method specific to engine meshes is implemented. This new approach reduces the number of meshes required for a full-cycle simulation and improves parallelization. Then a fully-coupled lagrangian and wall film model is deployed to simulate the direct injection during the intake phase. It is proven that the setup is suitable for such low evaporating conditions. Strong in-cylinder flow and spray interactions are observed, mainly due to the intake flow and the cylinder geometry. In addition, it is confirmed the importance of the wall film model to investigate the distribution of rich fuel regions at the spark ignition and, therefore, the formation of soot in the subsequent combustion phase. Finally, the code's capability is used to perform motored multi-cycle LES. The resulting flow field matches experimental results in first-order and second-order quantities. In conclusion, the open-source code can fully support the numerical investigation of different engine phenomena.