Support Chemkin parameterization for multiple bath gas species

[ljxsbest]

The Chemkin format is not supported using the latest Cantera version. For example, in our mechanism, the format is:

H+O2(+M)=HO2(+M)    4.66E12 0.44 0.0E0 
    HE/0.57/
    AR/0.0/
    N2/1.0/
    O2/1.0/
    H2/2.0/
    CH4/2.0/
    CO2/3.25/
    H2O/17.6/
    CO/4.0/
    LOWMX  /4.0662E19 -1.4E0 -1.80537E2/
    TROEMX /5.0E-1 1.0E0 1.0E10 1.0E30/
    LOWSP  /N2 1.91E+20 -1.5568 253.86/
    TROESP /N2 5.0E-1 1.0E0 1.0E10 1.0E30/

However, the format is not supported in the latest Cantera version, and we must extract the N2 component and change the format into two separate reactions:

H+O2(+M)=HO2(+M)    4.66E12 0.44 0.0E0
    HE   /0.57/
    AR   /0.0 /
    N2   /0.0 /
    O2   /1.0 /
    H2   /2.0 /
    CH4  /2.0 /
    CO2  /3.25/
    H2O  /17.6/
    CO   /4.0 /
    LOW  / 4.0662E19 -1.4E0 -1.80537E2 /
    TROE / 5.0E-1 1.0E0 1.0E10 1.0E30  /

H+O2(+N2)=HO2(+N2)    4.66E12 0.44 0.0E0   
    LOW  / 1.91E+20 -1.5568 253.86    / 
    TROE / 5.0E-1 1.0E0 1.0E10 1.0E30 /

When we try to adjust the format to fit the Cantera, it’s not working by Chemkin.
Could you please try to fix this issue in your side to support the Chemkin format in the next version?

[speth]

I haven't seen these Chemkin keywords before: LOWMX, TROEMX, LOWSP, TROESP. I have several questions here:

• Is there any public documentation of what these keywords mean and how the resulting rate coefficient is supposed to be calculated?
• When were these keywords introduced?
• Is the second input form actually equivalent, or is the formula for calculating the rate coefficient more involved? I don't understand why additional syntax would have been added if the second form is equivalent.
• What do you mean that the second input form isn't working? That looks like syntax that would have been parseable all the way back to Chemkin-II.

I'm wondering if this ends up being a special form of Mike Burke's mixing rules (see #157), which @pjsingal is currently working on implementing in Cantera.

[rwest]

It's new to me too. I'm pretty sure fair-use doctrine allows me to post this excerpt of a Release 2021R2 technical manual that I just found online:

3.6.4. Usage of Multiple Bath-gas Species
The typical usage of the reaction types outlined in Unimolecular/Recombination Fall-off Reactions (p. 36) through General Pressure Dependence Using Logarithmic Interpolation (p. 40) is with a specific species present as the "bath" gas. Many times the reactions for different bath-gas species are all included in the same mechanism. Note that the reaction rate computed at the high or low pressure limit would, in theory, be identical regardless of which species acts as the bath gas. However, if multiple instances of the same reaction with different bath-gas species are present in a given mechanism, then the net rate reaction rate is computed by summing up the individual rates. If each rate expression gives the reaction rate at the limiting pressure as $R$ and suppose that there are $n$ such rate expressions (for $n$ different bath-gas species), then the total rate is $n*R$ (due to the summation). Note, however, that the expected value is $R$. To accommodate such cases, an enhanced way of specifying the reaction with multiple bath-gas species is provided. In particular, the rate expressions for different bath-gas species may be written together in the same reaction. For reactions written in this manner, the net rate is computed by summing the individual rates multiplied by the mole fraction of corresponding bath-gas species (or the mixture). Thus, the correct limiting rate is obtained.

Consider the following example:

H +O2 (+M) = HO2 +(M) 4.650e+012 0.440 0.0
LowMX/1.737e+019 -1.230 0.0/
TroeMX/ 6.700e-001 1.000e-030 1.000e+030 1.000e+030/ LowSP/ AR 6.810e+018 -1.200 0.0/
TroeSP/AR 7.000e-001 1.000e-030 1.000e+030 1.000e+030/ LowSP/ HE 9.192e+018 -1.200 0.0/
TroeSP/HE 5.900e-001 1.000e-030 1.000e+030 1.000e+030/ HE / 1.0 / AR / 1.0 / H2/ 1.30/ H2O/ 10.00 /

Note that the keywords LOW and TROE (see the Chemkin-Pro Input Manual) here are augmented by SP and MX; which indicate the rate expressions for a particular species (such as Argon (AR) and Helium (HE) in this case) and the rest of the mixture. The net rate for this reaction is then
$R = R_{AR}X_{AR} + R_{HE}X_{HE} + R_{MIX}X_{MIX}$ (3.38)
where $R$ and $X$ indicate rate and mole fraction, respectively, and
$X_{MIX} = 1 - (X_{AR} + X_{HE})$
is the mole fraction of the "rest" of the mixture. In the example above, each rate expression makes use of the third-body efficiencies. For the specific bath-gas species (that is 'AR' and 'HE' in this example) non-unity third-body efficiency cannot be specified.

This formulation is also available when (a) the reaction line input is for the low- pressure limit, that is, for individual species and for a mixture, the rates are indicated by the keyword HIGHMX and HIGHSP and (b) for general logarithmic pressure dependence use keywords PLOGMX and PLOGSP as shown in the example below:

H + O2 = HO2         1.0 0. 0.
   PLOGMX/1.      2.48E+17 -2.226  36. /
   PLOGMX/10.     1.90E+18 -2.194   6. /
   PLOGMX/100.    6.05E+19 -2.330 635. /
   PLOGSP/HE 1.   2.48E+17 -2.226  36. /
   PLOGSP/HE 10.  1.90E+18 -2.194   6. /
   PLOGSP/HE 100. 6.05E+19 -2.330 635. /

For PLOG input, third-body enhancements are not allowed. In the example above, the rates for the specific species (HE in this case) and for the rest of the mixture are first computed by appropriate interpolation for the specified pressure. The net rate is then computed by multiplying by the corresponding mole fractions as shown in Equation 3.38 (p. 42).