Difference between revisions of "Evaluation AC2-10"

Jump to navigation Jump to search

Internal combustion engine flows for motored operation

Application Challenge AC2-10   © copyright ERCOFTAC 2021

Evaluation

Comparison of CFD Results with Experimental Data

In-Cylinder Pressure

The in-cylinder pressure curves are shown in Figure 18. All groups match the experimental data well. Only for the peak pressure one can observe deviations due to its sensitivity in respect to the boundary conditions. The peak pressure was typically overestimated by the simulations. The influence of the piston top-land crevice is shown considering the data obtained using Ansys CFX, where simulations were carried out with and without the crevice volume. The specific treatment of the crevice volume considering the different computational codes can be found in the CFD Simulations dection. It was found, that the lower temperature inside the crevice is decreasing the overall in-cylinder temperature and hence the in-cylinder pressure. A detailed discussion about this deviation can be found in [2122].

 Figure 18: In-Cylinder pressure curve with enlargement at 0 CAD

Further, the trapped mass inside the cylinder after the intake valves are closed was calculated by all groups, see Table 6. Unfortunately, the trapped mass was not determined experimentally so that no comparison can be performed here.

Table 6: trapped in-cylinder mass after intake valves closing
Ansys CFX OpenFoam KIVA-4mpi
565mg 542mg 546mg

In-cylinder Flow Field

Particular attention needs to be paid to the intake phase, because the flow distribution over the intake valves determines the formation of the in-cylinder tumble flow. Therefore a comparison of the flow field in the valve middle plane (z=19mm) during intake at 270° bTDC is illustrated in Figure 19, where the piston is positioned at y = -51.98mm. The 3D flow field was used to calculate the phase-averaged velocity magnitude ${\displaystyle {|V|}}$. From the experimental side no PIV data are available for this plane. Therefore magnetic resonance velocimetry (MRV) data from a flow bench experiment are used for comparison instead. The flow bench is a 1:1 scale model of the TU Darmstadt engine, produced from polyamide for MRV measurements at a fixed valve position corresponding to 270° bTDC and uses water as fluid. A direct comparison of the PIV and appropriate scaled MRV measurements within the central tumble plane revealed a very good agreement between both measurements above y = −20mm. MRV measurements are disregarded below y = −20mm due to the absence of the piston. Hence, the data can be used for comparison of the motored engine flow as long as no compressibility effects need to be considered. Further information regarding the configuration and measuring technique can be found in Freudenhammer et al. [17].

 Figure 19: Flow field during intake at 270° bTDC at mid-valve position (z=19mm). Left Top: Ansys CFX, Right Top: OpenFoam, Left Bottom: KIVA-4mpi, Right Bottom: Experiment (MRV).

The general flow features are captured fairly well by all simulations. Some differences are seen in the vicinity of the intake valve where flow separation impacts the extent of the recirculation zone and flow redirection. This includes the region behind the valve stem, the corner on the left side of the intake valve and the region at the top along the cylinder head and exhaust valve. These small differences can have a strong impact on the motion of the tumble vortex later during compression.

Figure 20 and Figure 21 show the mean in-plane velocity components u and v as well as the corresponding velocity fluctuation urms and vrms in the central tumble plane (z = 0mm) at five different vertical positions for the intake stroke at 270° bTDC and compression at 90° bTDC, respectively. For this plane, PIV data are available from experimental side and are used for comparison. For the mean flow the largest deviations between the simulations are found for the horizontal velocity component u while the vertical component v which is strongly determined by the piston motion is captured well by all simulations. For the velocity fluctuations the agreement between the simulations and experiments is very good for 90° bTDC, while stronger deviations are observed for the intake stroke, where turbulence is generated within the shear layers of the high velocity intake flow resulting in high fluctuation levels.

 Figure 20: Flow field during intake phase at 270° bTDC at mid-plane position (z=0mm).

 Figure 21: Flow field during compression phase at 90° bTDC at mid-plane position (z=0mm).

The evolution of the tumble flow is shown by the path of the tumble vortex center in Figure 22 during late compression (100–30° bTDC). The tumble center detection is described in Stiehl et al. [41]. The differences in the tumble vortex path reflect the small differences of the intake flow as shown in Figure 19.

It should be noted that the results carried out using OpenFoam used a reduced cycle number of six for the averaging procedure and the results were obtained using a coarser mesh (1mm in the combustion chamber) than described in the mesh treatment section. For further information please see [21].

 Figure 22: Motion of the tumble center every 10 CAD in the range of 100° bTDC to 30° bTDC (OpenFoam results shows 100° bTDC to 40° bTDC).

Exhaust Pipe Pressure

To compare the pressure in the exhaust pipe with experimental data (see Figure 23), only the numerical data obtained using Ansys CFX can be utilized, since the data obtained by OpenFoam and KIVA-4mpi simulations used the experimental pressure data as boundary condition.

 Figure 23: Exhaust Pipe Pressure.

As can be seen, the time-resolved pressure boundary condition created by the software GT-Power (see the boundary condition section) reproduces the experimental data accurately.

Contributed by: Carl Philip Ding, Rene Honza, Elias Baum, Benjamin Böhm, Andreas Dreizler — Fachgebiet Reaktive Strömungen und Messtechnik (RSM),Technische Universität Darmstadt, Germany

Contributed by: Brian Peterson — School of Engineering, University of Edinburgh, Scotland UK

Contributed by: Chao He, Wibke Leudesdorff, Guido Kuenne, Amsini Sadiki, Johannes Janicka — Fachgebiet Energie und Kraftwerkstechnik (EKT), Technische Universität Darmstadt, Germany

Contributed by: Peter Janas, Andreas Kempf — Institut für Verbrennung und Gasdynamik (IVG), Lehrstuhl für Fluiddynamik, Universität Duisburg-Essen, Germany

Contributed by: Stefan Buhl, Christian Hasse — Fachgebiet Simulation reaktiver Thermo-Fluid Systeme (STFS), Technische Universität Darmstadt, Germany; former: Professur Numerische Thermofluiddynamik (NTFD), Technische Universität Bergakademie Freiberg, Germany

© copyright ERCOFTAC 2018