How to get a physical laminar flame speed at high temperatures, pressures and rich mixture

[Leonardo Giordo]

Hello everyone,

I am studying the laminar flame speed at high temperatures and pressures for Methanol in the rich region (Lambda 0.5 --> Phi 2).

In the images the behaviour of the LFS vs Lambda. It seems plausible at low pressures, but then it diverges at high pressures.

Is there a way to get a physical solution at low Lambda values?

RM: Mechanism Downloads - University of Galway

Attached: 

• images of LFS vs Lambda
• Script with used parameters
• Reaction mech .yaml file
• Flame profiles at P200bar, T900K, Lambda 0.5

BR,

Leonardo

[Ray Speth]

Hi Leonardo,

Your script specifies a minimum grid size of 1 μm, but in the example you shared at 200 bar, the flame width is only maybe 10 μm, so the solution is substantially under-resolved. I tried reducing the value of min_grid in your example script, and this shows it’s definitely affecting your result:

grid_min = 1e-6: LFS = 126.4550 cm/s | grid pts = 38 | domain = 0.0600 cm grid_min = 1e-7: LFS = 111.0595 cm/s | grid pts = 102 | domain = 0.0600 cm grid_min = 1e-8: LFS = 110.6983 cm/s | grid pts = 165 | domain = 0.0600 cm grid_min = 1e-9: LFS = 110.6983 cm/s | grid pts = 165 | domain = 0.0600 cm

The later grid refinements aren’t even that expensive, at least for these refinement criteria and mechanism size, so I don’t it’s worth trying to limit the minimum grid size. The solver seems well behaved at least for this case, though I know there are other high pressure cases with different flame configurations (see for example this thread) where it is not.

Regards,
Ray

​

[Leonardo Giordo]

Hello Ray,

thanks a lot for the heads up.

I noticed that in my script I wrongly left phi = 1. 

With phi = 2 I do see the lfs decreasing with the increased grid spacing, but it is still higher than the value at phi = 1, which is not physical.

May you help me understand why?

BR,

Leonardo