Reduced Reaction Sets based on GRI-Mech 1.2

Andrei Kazakov and Michael Frenklach

         University of California at Berkeley
         Berkeley, CA 94720-1740, USA

         Telephone: (510) 643-1676

Two sets of elementary reactions reduced from GRI-Mech are reported here. They were developed by truncation of the original GRI-Mech with the objective of developing a smallest set of reactions (i.e., still a detailed mechanism but with the smallest numder of variables) to reproduce closely the main combustion characteristics predicted by the full mechanism, GRI-Mech.

The two sets are:

By clicking the corresponding icons above, (remember to select a load to disk option in your Mosaic application) you can download files with the reduced mechanisms. These are ASCII files in the CHEMKIN II format. For the thermodynamics data you must use thermo12.dat, the thermodynamics data of GRI-Mech. The thermo12.dat file can be downloaded from the GRI-Mech Home Page. If you experience difficulties with downloading any of these files, please contact the authors.
To reference this reduced mechanisms, please use:
A. Kazakov and M. Frenklach,
The rest of this document describes the reduction technique used to develop these reduced mechanisms and their performance tested by the authors.

We will be glad to learn about your experiences with these reduced mechanisms (and include your results in this document as well).

This work was supported by the Gas Research Institute, under contract No. 5092-260-2454 (Dr. Robert V. Serauskas, Project Manager, 773/399-8208,


LEGAL NOTICE These files, both the ones intended for use as computer input as well as those comprising documentation, were prepared by The Pennsylvania State University, Stanford University, SRI International and The University of Texas at Austin as a result of work sponsored by the Gas Research Institute (GRI). Neither GRI, members of GRI, nor any person acting on behalf of either:

  1. Makes any warranty or representation, express or implied, with respect to the accuracy, completeness, or usefulness of the information contained in these files, or that the use of any data, method, or process disclosed in these files may not infringe privately owned rights; or
  2. Assumes any liability with respect to the use of, or for damages resulting from the use of, any information, data, method or process disclosed in these files.

Reduction Technique

The reduction in size followed a technique of detailed reduction described in previous publications. The technique is based on flux analysis performed using simple, zero-dimension calculations with a "full" detailed chemical reaction mechanism (GRI-Mech in the present case). The following criteria are applied to identify the non-contributing reactions:

| R(i) | < e(r) | R(ref) | and | R(i) delta-H(i) | < e(q) Q

where R(i) is the rate of reaction i, R(ref) is the rate of a reference reaction (e.g., the maximum rate), delta-H(i) is the enthalpy change of reaction i, Q is the maximum value among all the terms |R(i) delta-H(i)|, and e(R) and e(Q) are the chosen parameters considerably smaller than unity. The first criterion tests the contribution a given reaction makes to the main chain branching while the second criterion tests the contribution to the heat release.

Reactions whose rates, both forward and reverse, satisfy the above inequalities at all grid points of the calculation are removed from the mechanism. This procedure eliminates a specific species when all reactions of such a species happen to be removed.

The present reduction was performed using e(r) = e(q) = 0.02, under the following test conditions: shock-tube ignition of methane-air mixtures at wide ranges of initial conditions (equivalence ratio PHI = 0.2 - 2.0, initial pressure Po = 0.1 - 50 atm, and initial temperature To = 1300-2500 K); and methane-air adiabatic flames at 1 and 20 atm. This procedure generated a 22-species, 104 reactions set, which is referred to as DRM22.

Additional analysis of reaction fluxes and sensitivities to removal of selected species was performed on DRM19. Those chemical species which were showed near-zero sensitivity were removed. This resulted in a smaller set, 19-species and 84 reactions, which is referred to as DRM19.

Numerical Performance of DRM19 and DRM22

The developed reduced mechanisms, DRM19 and DRM22, were tested against GRI-Mech using a series of ignition delay and laminar flame speed simulations.

Ignition Delays

Ignition delays were calculated for a wide range of methane-air mixtures: the initial pressure (Po) was varied from 0.1 to 10 atm, the fuel-air stoichiometric ratio (PHI) from 0.2 tp 2.0, and initial temperatures was varied from case to case in order to cover the entire range of ignition delays, 1 microsec - 10 ms. The results are presented in Figure 1. DRM22 works well for all the initial conditions tested: the deviations from GRI-Mech do not exceed 4 % and are below 2 % for most cases. DRM19 also performs well and the deviations are typically within 6-8 % with respect to GRI-Mech. The accuracy of DRM19 drops at lower temperatures and higher pressures. For example, at To = 1100 K, Po = 10 atm and PHI = 0.2, the ignition delay calculated with DRM19 has 37 % relative error. The deviations increase further at even higher initial pressures, e.g., 89 % for stoichiometric methane-air mixture at Po = 100 atm and To = 1050 K (for comparison, DRM22 has 3.5 % error at these conditions). However, no extensive tests were carried out for these conditions since GRI-Mech itself was not validated at these conditions either.

Figure 1. Relative deviations of ignition delays calculated with the reduced mechanisms in comparison to GRI-Mech for methane-air mixtures: solid lines - DRM22, dashed lines - DRM19.

Laminar Adiabatic Premixed Flames

The performance of DRM19 and DRM22 was also tested on laminar adiabatic premixed methane-air flames at 1 and 20 atm, the conditions against which GRI-Mech was optimized and validated. Four parameters were used for comparison: adiabatic flame velocity and maximum mole fractions of major radicals, H, OH and CH3. (The adiabatic flame temperature computed with the reduced mechanisms was found to be practically indistinguishable from that predicted by GRI-Mech.)

1 atm

The numerical results obtained for 1-atm adiabatic methane-air flames are shown in Figure 2. The adiabatic flame velocities are predicted within 8 % by both reduced schemes for PHI = 0.6-1.5 (3-4 % for the stoichiometric mixture). The peak mole fraction of H atoms is predicted within 20 % for the entire range of PHI (1.0-1.5 % for the stoichiometric mixture). However, the flame velocity predicted by DRM19 begins to deviate significantly as PHI increases (see the top panel of Figure 2). The maximum mole fraction of OH is predicted well by both reduced schemes for lean to stoichiometric mixtures (less than 1 % for PHI = 1.0). As PHI increases, the accuracy of DRM19 drops (84 % for PHI = 1.5) while that of DRM22 still remains reasonable (7.5 % for PHI = 1.5). The maximum CH3 mole fraction is predicted within 20 % by DRM19 (10 % for PHI = 1.0) and within 10 % by DRM22 (8 % for PHI = 1.0 ).

Figure 2. Relative deviations of adiabatic flame velocities and peak concentrations of selected species calculated with the reduced mechanisms in comparison to GRI-Mech for atmospheric, adiabatic, laminar, premixed methane-air flames: solid lines - DRM22, dashed lines - DRM19.

20 atm

The numerical results obtained for 20-atm adiabatic methane-air flames (To = 400 K ) are shown in Figure 3. The predictions for adiabatic flame velocities are slightly worse than those seen for the 1-atm flames, but still remain within 11 % (2-4 % for PHI = 1.0). The predictions for the peak mole fractions of H, OH and CH3 have the trends similar to those observed for 1 atm flames. We note that the performance of DRM19 is better for the lean to stoichiometric mixtures.

Figure 3. Relative deviations of adiabatic flame velocities and peak concentrations of selected species calculated with the reduced mechanisms in comparison to GRI-Mech for 20-atm adiabatic, laminar, premixed methane-air flames: solid lines - DRM22, dashed lines - DRM19.