[simone.altobelli]

Hi,

I have some questions about engine combustion modelling with ECFM in CONVERGE (v 2.3.15): I see during the 50-90 % MFB phase of the combustion a sudden slope change in the CHR curve at about 15/20 CA ATDCF; of course the lean lambda distribution (mean lambda = 1.3) and the reduction of turbulence have their influence, but I've noticed from 3D results that at a certain angle (15/20, that corresponds to te slope change) the variable FLAME_SPD_COR starts to go down from values near unity to 0.5/0.6.

The manual describes this passive as "Flamespeed correlation factor due to the flame-wall interaction", but from 3D results I can see values near 0.5/0.6 also far from boundary walls.

The questions are:
- Is it possible that the slope change is due to this correction on flamespeed or do you think it's something else?
- Is there a CONVERGE variable in the code that corresponds to this passive (I cannot find it)? Is it editable? I'm using UDF laminar flamespeed yet with my Metghalchi coefficients and EGR correlation, but I didn't see in the CONVERGE original UDF for ECFM laminar flamespeed this correction, maybe it's in the ECFM code, also because SLAM values from 2D or 3D results are not affected by a sudden 50% reduction (in this case, is it possible to have a UDF version of ECFM to modify it?)
-Why this correction should be less than unity far from wall?

I've attached some 3D and graphs to show the problem, thank you in advance.

Simone

[tmburton]

Simone,

Here are answers to your questions:

1. Yes, the flame speed correction factor is decreasing since the flame is approaching/contacting the wall.
2. You can use ecfm->heatloss[kk] to access it. It isn't editable by the user at this moment. However, if you have a different way to consider the flame-wall interaction, you can use user_combust_flame_speed.c to correct the laminar flame speed.
3. The correction factor is considering the heat loss from the premixed flame to the wall. In ECFM, the premixed flame is solved by transporting flame surface density which is relatively "thick". The correction factor is consistent with the flame surface density.

For more details about flame-wall interaction model in ECFM, you can refer to:
T.J. Poinsot, D.C. Haworth, G. Bruneaux, Direct simulation and modeling of flame-wall interaction for premixed turbulent combustion, In Combustion and Flame, Volume 95, Issues 1–2, 1993, Pages 118-132, ISSN 0010-2180, https://doi.org/10.1016/0010-2180(93)90056-9.

Best regards,

Tristan

[simone.altobelli]

Hello Tristan,

thank you very much for your reply. Now it is clear to me the role of the correction factor, but unfortunately in my case I wasn't using this heatloss correction in the laminar flamespeed UDF, so if I add it (multiplying the laminar flamespeed for heatloss[kk] in the calculation) the result is even worse (in term of MFB curve slope change), as you can see in the figures!

Maybe in my case the reason of this combustion trend isn't wall quenching (that I excluded without using heatloss, even if it's not the most correct choice, but I think that's a secondary problem now) and there's something else that I'm not considering: I'm thinking about an excessive response of my laminar flamespeed correlation (based on Cantera simulations) to the lambda variation or an excessive response of the ECFM model to the turbulence variation (turbulence is very low in the final phase of combustion).

Simone