CFD Online Discussion Forums

CFD Online Discussion Forums (
-   -   Low load operation point with internal EGR (

MFGT June 23, 2020 04:00

Low load operation point with internal EGR
Hey guys,

I am working on a low load operation point (2000 rpm 2 bar BMEP) with high internal EGR (28%). EGR is storred in the intake port. GT-Power boundaries (pressure and temperature) are used. The engine is a gasoline DI 4-cylinder engine with integrated exhaust manifold, simulated is cylinder 1 only. The other exhaust ports are filled with INFLOW boundaries (EGR, massflow rate of corresponding valves).

The problem that I have is that my geometry ends at the cylinder head boundary, which means that during gas exchange the exhaust gas is pused out of the INFLOW boundary. To compensate this, i use a species (O2, N2, CO2, H2O) profile over CA to model the EGR backflow.

However, of course the CONVERGE backflow does not match the GT-Power backflow so the species do not match, I get more EGR in my clinder (>35%) and missfire as result. Combustion model is ECFM.

How can I solve this issue?

a) I need some species tracing for the INFLOW boundary. I tried setting up a coupeld simulation, but this seems to be a bit unstable and I would need to modify the integrated exhaust manifold.

b) Adding the manifold geometry to the intake port could be a solution, but then I also have to deal with the backflow of the other 3 cylinders.

c) A complete 4-cylinder setup? INFLOW and OUTFLOW boundaries are far away from the cylinder then.

Best Regards,

MFGT June 25, 2020 02:50

I went for b) now, and since I am using ECFM I had to define a lot of inputs.

The backflow boundaries of the remaining 3 cylinders into the intake manifold now have massflow rate, temperature, species and passive profiles attached to them.

In total, the result is good, I have no lost EGR and all EGR that is storred in the intake port is sucked back into the cylinder.

However, the combustion was not as good as I thought. Even with 25% EGR only its quite slow and by far not complete. I need to investigate further.

All times are GMT -4. The time now is 06:16.