Advanced Predictive Diesel Combustion Simulation Using Turbulence Model and Stochastic Reactor Model

Tim Franken, Arnd Sommerhoff, Werner Willems, Andrea Matrisciano, Harry Lehtiniemi, Anders Borg, Corinna Netzer, Fabian Mauss (Chalmers University of Technology)

CHALLENGE - Numerical modeling of diesel combustion processes and emission formation is an important tool for evaluating engine designs and developing new combustion strategies. The difficulty is that the best design is not necessarily the same for all operating conditions since combustion stability, engine protection, and others set variable boundary conditions. The limits of the CFD method are its high computational costs making it difficult to use for extensive parameter variation and optimization. Physical 0D models balance the trade-off between simulation accuracy and computational costs. The scope of this work is the modeling of turbulence time-scales in diesel engines within a 0D model framework and its effect on combustion and emission formation for different operating conditions. 

SOLUTION - A one-equation turbulence model is incorporated, which accounts for effects of cylinder geometry, density change, dissipation, swirl and squish flow and direct fuel injection on the cylinder turbulent kinetic energy. The combustion is described using the pdf based DI-SRM (stochastic reactor model). DI-SRM is able to account for the cylinder to cylinder variation effects on engine performance and emission formation. Measurements are conducted on a modern Euro-6 1.5l diesel engine and further processed in a multi-cylinder heat release analysis tool. For optimization, the Fast MOSA algorithm is applied using software modeFRONTIER. The DI-SRM simulation results show an accurate match of the experimental data. Considerable differences for NOx, CO and HC emissions are found within areas of light load and low engine speed conditions. It is revealed that high cycle to cycle variations impact the mean emission concentrations significantly.












BENEFITS - As a conclusion, the DI-SRM is able to predict heat release rates, CO2 and NOx emissions for different operating conditions over the whole engine map. The investigated engine shows distinct cylinder to cylinder pressure deviations for all operating points. Example pressure profiles of a 2000rpm and 93Nm operating point are highlighted in figure 15 above.