[D_Timbota]

Hi Group,

It looks like the ignition does not take place in my simulation. I need your advice please. I have attached pictures of Heat release rate and spark setup.

[MFGT]

Originally Posted by D_Timbota

Hi Group,

It looks like the ignition does not take place in my simulation. I need your advice please. I have attached pictures of Heat release rate and spark setup.

Hi,

that is not your spark setup, it looks like the SAGE setup.

Do you have an energy source specified?
How is the mixture at IGN timing?
What reaction mechanism do you use?

[D_Timbota]

Originally Posted by MFGT

Hi,

that is not your spark setup, it looks like the SAGE setup.

Do you have an energy source specified?
How is the mixture at IGN timing?
What reaction mechanism do you use?

Both mech.dat and therm.dat were found online(https://github.com/OpenFOAM/OpenFOAM...Data/therm.dat). I am thinking to use the one from PFI_E85 Case setup in converge. I don't know how to setup ignition time, I went through Converge Manual but couldn't figure out.

The simulations I have run were at a coarse grid 0.1m (I was trying to save computational time while working on setting the right parameters) which makes a 2 cycles simulation quite fast( 2-3 hours) and none of them are not giving the right value for heat release, but the in-cylinder pressure is around 22 bar. I currently run a GDI_SAGE simulation at a grid of 0.01, it progressed only 881 CAD in 24 hours (is this normal?), The heat release is right (1265 J) and the Mean Temperature around 1800K, but the pressure is low, around 20 bar.
I am expecting a higher pressure and always thought is low due to ignition not taking place, but it seems that is still low even in GDI Case Setup.

I'm totally confused and under time pressure.

[MFGT]

Originally Posted by D_Timbota

Both mech.dat and therm.dat were found online(https://github.com/OpenFOAM/OpenFOAM...Data/therm.dat). I am thinking to use the one from PFI_E85 Case setup in converge. I don't know how to setup ignition time, I went through Converge Manual but couldn't figure out.

The simulations I have run were at a coarse grid 0.1m (I was trying to save computational time while working on setting the right parameters) which makes a 2 cycles simulation quite fast( 2-3 hours) and none of them are not giving the right value for heat release, but the in-cylinder pressure is around 22 bar. I currently run a GDI_SAGE simulation at a grid of 0.01, it progressed only 881 CAD in 24 hours (is this normal?), The heat release is right (1265 J) and the Mean Temperature around 1800K, but the pressure is low, around 20 bar.
I am expecting a higher pressure and always thought is low due to ignition not taking place, but it seems that is still low even in GDI Case Setup.

I'm totally confused and under time pressure.

Ah okay, if your base grid is that coarse (100mm ), it is guaranteed that IGN is not happening.

0.004 (4mm) is recommended for a reason. Not forget the embedding at spark timing, that goes down to 0.125mm cells. With this cell size setup, most of my simulations (720°CA) take 1.5 to 2.5 days on 8 cores. How many cores do you use? Do you even run CONVERGE in parallel?

To modify IGN timing check source.in

If you dont have the CPU ressources and many cores, you could change the combustion model (G-Eq. maybe) that is less computational expensive and less grid dependent?

[D_Timbota]

Originally Posted by MFGT

Ah okay, if your base grid is that coarse (100mm ), it is guaranteed that IGN is not happening.

0.004 (4mm) is recommended for a reason. Not forget the embedding at spark timing, that goes down to 0.125mm cells. With this cell size setup, most of my simulations (720°CA) take 1.5 to 2.5 days on 8 cores. How many cores do you use? Do you even run CONVERGE in parallel?

To modify IGN timing check source.in

If you dont have the CPU ressources and many cores, you could change the combustion model (G-Eq. maybe) that is less computational expensive and less grid dependent?

I run it on 4 cores with Converge in parallel. Is G-Eq. turning similar results as regular setup? Is there any other way to reduce computational time? I'm running out of time...

Any advice on why pressure is not rising to more than 22 bar? Even on the converge tutorials on YouTube, the PFI_Sage turns low pressure.

[D_Timbota]

Originally Posted by MFGT

Ah okay, if your base grid is that coarse (100mm ), it is guaranteed that IGN is not happening.

0.004 (4mm) is recommended for a reason. Not forget the embedding at spark timing, that goes down to 0.125mm cells. With this cell size setup, most of my simulations (720°CA) take 1.5 to 2.5 days on 8 cores. How many cores do you use? Do you even run CONVERGE in parallel?

To modify IGN timing check source.in

If you dont have the CPU ressources and many cores, you could change the combustion model (G-Eq. maybe) that is less computational expensive and less grid dependent?

Hi,

I have changed the grid to 0.01 for two simulation I run ( the recommended 0.004 will hardly progress). First simulation was The GDI_Sage case setup using Iso-Octane as a fuel, the second was GDI_Sage using Ethanol 80/Water 20 as a fuel(using therm.dat and mech.dat files from E85 PFI case setup from converge). The ignition clearly takes place in the gasoline simulation releasing over 1200 J every cycle while whereas this is not happening in the alcohol fuel based simulation ( around 15 J heat release). What I find weird though that max temp is similar for both simulations. Please advise me on where is the mistake. I paid close attention to all case setups from converge software.
Please find the attached pictures.

Regards,
Dumitru Timbota