zhixchen
commented
1 year ago Hi,
Thank you for raising this issue. Indeed, we have not yet included the thermal diffusion (Soret) effect in the DeepFlame code.
It is very important that you mentioned the potential influence on H2 flames and we should definitely look into this effect. To make it easy for us to benchmark, it would be very helpful if you could provide more information about this typical example case where the flame speed is changed by more than 10%. Then we can set up this case and implement this effect in the code to quantify the influence.
Best regards,
The DeepFlame Team
wangdu817
97 cm/s discrepancy at P = 18 atm and Tu = 781 K.
ref: Combustion and Flame, 194 (2018) 175-179
more obvious effect at flame initial stage
ref: Proc. Inst. Combust.,34 (2013) 695-702
One of the flaws for the Cantera transport model is that the thermo diffusion (or Soret effect) only can be turned on for multi-component model, However, Soret effect can lead to more than 10% discrepancy for H2 and H2 addition flame speed prediction. The thermo diffusion coefficient is obtained from the the full diffusion calculation in Cantera, it is possible to apply light component limit (species molecular mass< 5) as CHEMKIN done, and included additional thermo diffusion ratio calculation function for these species.