Project
Experimental and numerical investigations of mixture formation in H2 engines using PFI and DI
Hydrogen internal combustion engines are a promising technology to decarbonize parts of the transportation sector. However, challenges such as achieving a stable combustion behaviour and reducing NOx emissions remain. The hydrogen injection and mixture formation play a crucial role in order to achieve a stable combustion and minimal emissions. The design and operation of such injection hardware requires suitable numerical codes (3D CFD). Existing methods, that are mostly based on conventional hydrocarbon fuels, are often not directly applicable to accurately model hydrogen as fuel. The scope of this project is therefore two-fold: Models should first be studied and validated, and then be employed to deepen the understanding of injection and mixing in hydrogen combustion engines to enhance the technology. As 3D CFD simulations are numerically highly expensive, the use of an HPC cluster is necessary.
Project Details
Project term
March 1, 2025–February 28, 2027
Affiliations
RWTH Aachen University
Institute
Chair of Thermodynamics of Mobile Energy Conversion Systems
Project Manager
Principal Investigator
Methods
As presented in the proposal, the project is well structured: Optical data on hydrogen injection and mixing from different test-benches is supplied by research partners. The test-benches increase in complexity, beginning with an optical study of a simple underexpanded hydrogen jet (sub-project 1) and ending with investigations of an optical accessible single-cylinder engine (sub-project 3). This allows for a structured approach to employ, improve, calibrate and validate the numerical methods. Numerical grids, models and computational settings are all studied. This validation is conducted for both RANS and LES. Once the model validation has been finished, the methods are employed to two thermodynamic single cylinder engines (sub-project 4). Experiments are conducted on these two engines. Employing the before validated numerical models will deepen the interpretation of the experimental results and enhance the understanding of the in-cylinder mixing and combustion processes. Again, both RANS and LES methods are employed. RANS methods for a broader spectrum of operating conditions, multiple LES cycles only for a selected operating point. The insights into the combustion behaviour via 3D CFD will allow to further improve the technology.
Results
The study of the underexpanded hydrogen jet (sub-project 1) was conducted mostly in Q2 of 2025 and finalized in Q3 of 2025. The study was a success: For different injector positions (protruding and recessed) and different injection pressures (5 – 25 bar) both RANS and LES models were studied and validated thoroughly and selected for the remaining project. Turbulence and other models, but also numerical options (e.g. the gradient calculation) were studied. Especially a RANS setup of the Standard k-epsilon turbulence model, combined with a least square gradients calculation, is capable of producing highly accurate results, while remaining numerical efficiency. This is an improvement on existing methods and knowledge in regard to modelling hydrogen injection. The hand-over of optical validation data for sub-projects 2 and 3 was delayed, thus also delaying the possibility to validate numerical models with this data. However, first results have been handed over in Q4 of 2025 and Q1 of 2026. The geometries have been cleaned up, numerical models have been setup up (accordingly to the findings from sub-project 1) and simulations have been conducted in Q1 of 2026. The work-packages are not fully finalized yet, but the first results indicate that the defined models also are suitable to model hydrogen injection and mixing in a physically more complex setup with a co-flow, valve interaction and charge motion. Furthermore, a novel post-processing method has been developed, employing ray tracing to produce Schlieren images from 3D CFD results. This is especially helpful, as the optical validation data is in large parts Schlieren imagery. The models for the two engines from sub-project 4 (light-duty and heavy-duty engine) have been setup and been run for a different operating conditions and geometries and have already helped to enhance the understanding of the thermodynamic results.
Discussion
The results so far are valuable to both research and industry. The detailed study and validation of different models and numerical setups enhance the knowledge significantly and gives concrete guidelines how the injection and turbulent mixing of hydrogen can be simulated. The sub-projects 2 and 3 are slightly delayed, as there were delays in the hand-over of the optical validation data. However, the time has been used to setup all geometries and models required for this project, including sub-project 4. Now, with more optical data available and more expected in the next weeks, the simulations for work packages 2 and 3 can be finalized efficiently, enhancing the understanding of hydrogen injection and mixing further. Already broad thermodynamic data for sub-project 4 exists, but still needs to be interpreted and explained deeper. First RANS simulations have already helped by identifying advantages and disadvantages of certain injector placements and timings, results that are directly valuable to industry. Further studies will build on these results, also employing LES to study cycle to cycle fluctuations. Overall, the project has already produced valuable results, and is on track, with some shifts in the time-schedule, to finish successfully.
Additional Project Information
DFG classification: 404-04 Hydraulic and Turbo Engines and Piston Engines
Software: CONVERGE CFD
Cluster: CLAIX
Publications
Ye Feng, Niklas Mirsch, Fabian Steeger, Michael Blomberg, Michael Grill, Andre Casal Kulzer, Marco Günther and Stefan Pischinger,
Development of a 0D/1D Model System for the Cycle-to-Cycle Variation of High Tumble Spark Ignition Engines,
https://dx.doi.org/10.4271/2024-01-2083, April 2024
Ye Feng, Niklas Mirsch, Michael Grill, Andre Casal Kulzer, Marco Günther and Stefan Pischinger,
Numerical Investigation on the Mechanism of Tumble Caused Cycle-to-Cycle Variation of High-Tumble Spark-Ignition Engine,
https://dx.doi.org/10.1007/978-3-658-42236-3_24, August 2023
Jana Hoffmann, Walter Vera-Tudela, Niklas Mirsch, Dario Wüthrich, Bruno Schneider,Marco Günther, Stefan Pischinger , Daniel A. Weiss and Kai Herrmann,
Investigation of Flow Fields Emanating from Two Parallel Inlet Valves Using LES, PIV, and POD,
https://dx.doi.org/10.3390/en16196917, September 2023
Jana Hoffmann, Niklas Mirsch, Walter Vera-Tudela, Dario Wüthrich, Jorim Rosenberg, Marco Günther, Stefan Pischinger, Daniel A. Weiss, Kai Herrmann,
Flow Field Investigation of a Single Engine Valve Using PIV, POD, and LES,
https://dx.doi.org/10.3390/en16052402, March 2023