CoChat

ReactionMechanismSimulator.jl: A modern approach to chemical kinetic mechanism simulation and analysis

Open access
Johnson, Matthew S., Pang, Hao‐Wei, Payne, Allen Mark, Green, William H. · 2024 · International Journal of Chemical Kinetics

Published version of a preprint

The DOI matches its Crossref record; 45 of the 46 references it cites were found in a registry or an index; none of the 17 sources its claims cite is checked for support yet.

Abstract

We present ReactionMechanismSimulator.jl (RMS), a modern differentiable software for the simulation and analysis of chemical kinetic mechanisms, including multiphase systems. RMS has already been applied to problems in combustion, pyrolysis, polymers, pharmaceuticals, catalysis, and electrocatalysis. RMS is written in Julia, making it easy to develop and allowing it to take advantage of Julia's extensive numerical computing ecosystem. In addition to its extensive library of optimized analytic Jacobians, RMS can generate and use Jacobians computed using automatic differentiation and symbolically generated analytic Jacobians. RMS is demonstrated to be faster than Cantera and Chemkin in several benchmarks. RMS also implements an extensive set of features for analyzing chemical mechanisms, including a library of easy‐to‐call plotting functions, molecular structure resolved flux diagram generation, crash analysis, traditional sensitivity analysis, transitory sensitivity analysis, and an automatic mechanism analysis toolkit. RMS implements efficient adjoint and parallel forward sensitivity analyses. We also demonstrate the ease of adding new features to RMS.

Overview

CoChat Summary

TL;DR

ReactionMechanismSimulator.jl (RMS) is a Julia-based differentiable simulator and analysis toolkit for chemical kinetic mechanisms, including multiphase systems. It provides analytic, automatic-differentiation, and symbolic Jacobians plus sensitivity, flux, crash, and automatic mechanism analysis, and benchmarks faster than Cantera and Chemkin on selected cases.

Key findings

  • RMS was faster than Cantera and Chemkin in simulation benchmarks on the 443-species, 26,108-reaction methyl propyl ether mechanism: about 2.1× and 1.4× faster than Cantera, and about 4.0× and 1.5× faster than Chemkin for the ConstantTP and ConstantV cases, respectively.
  • In sensitivity benchmarks on RMG’s propane mechanism (82 species, 1,110 reactions), RMS Adjoint was about an order of magnitude faster than Cantera and about 6.8× faster than RMS parallel sensitivity on one thread; running on 6 threads reduced this to a factor of 4.
  • RMS implements automatic mechanism analysis that ran in under a minute on a 1,600-species dodecane-isooctane mechanism, where traditional sensitivity analysis was challenging.
  • RMS supports multiphase/liquid-phase simulation; it was demonstrated on thermal cracking of a 1,326-species, 116,045-reaction model at 350°C and 10 bar with gas-liquid mass transfer.

Methods

  • Software design: Julia package with species, reactions, phases, domains, interfaces, and reactor objects; uses a moles-based mass-action formulation from Schwer et al. 2002; supports .rms YAML, Chemkin .inp, and Cantera YAML input.
  • Numerical methods: Jacobians via an optimized analytic Jacobian library, ForwardDiff.jl automatic differentiation, and ModelingToolkit.jl symbolic generation; ODE solving through DifferentialEquations.jl with solver flexibility; adaptive preconditioning and sparsity-triggered GMRES when Jacobian sparsity is larger than 80%.
  • Benchmarks/case studies: simulation benchmarks on the methyl propyl ether mechanism; sensitivity benchmark on the propane mechanism; a liquid-phase two-phase pyrolysis case; and a qualitative comparison to Cantera and Chemkin.

Limitations

  • Stated: RMS still lacks steady-state flame simulation and non-ideal thermodynamics; the adjoint approach is not always efficient for very large mechanisms; finite differences are intentionally excluded; and embedded neural network training “has not been done yet.”
  • Stated: the Chemkin sensitivity benchmark uses an approximation and does not compute sensitivities to thermochemical parameters, so that comparison has caveats.
  • Evident from the text: benchmark evidence is limited to selected mechanisms and a qualitative comparison table; no broad quantitative validation across all domains, phases, and interfaces is reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1002/kin.21753

Cites

45 of 45 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 2 Comparison of the computational cost of CVODE_BDF with the default dense linear solver and the preconditioned GMRES solver. 8 chemical kinetic mechanisms with a range of sizes (7 to 1400 species) and Jacobian sparsity (30 - 97%) are used, with a range of dropping tolerances τ (1e-3 - 1e-5).
Figure 2 Comparison of the computational cost of CVODE_BDF with the default dense linear solver and the preconditioned GMRES solver. 8 chemical kinetic mechanisms with a range of sizes (7 to 1400 species) and Jacobian sparsity (30 - 97%) are used, with a range of dropping tolerances τ (1e-3 - 1e-5).CC-BY-NC-ND · From the PDF
Figure 3 Radical rate of production example using the methyl propyl ether mechanism from Johnson et al. 2021.
Figure 3 Radical rate of production example using the methyl propyl ether mechanism from Johnson et al. 2021.CC-BY-NC-ND · From the PDF
Figure 4 Example RMS flux diagram for methyl propyl ether combustion. Lighter-colored and thicker arrows indicate larger fluxes. Thicker circles around species indicate higher concentrations.
Figure 4 Example RMS flux diagram for methyl propyl ether combustion. Lighter-colored and thicker arrows indicate larger fluxes. Thicker circles around species indicate higher concentrations.CC-BY-NC-ND · From the PDF
Figure 5 Mole fraction profiles corresponding to the methyl propyl ether ConstantTPDomain benchmark.
Figure 5 Mole fraction profiles corresponding to the methyl propyl ether ConstantTPDomain benchmark.CC-BY-NC-ND · From the PDF
Figure 6 Temperature profile corresponding to the methyl propyl ether ConstantVDomain benchmark.
Figure 6 Temperature profile corresponding to the methyl propyl ether ConstantVDomain benchmark.CC-BY-NC-ND · From the PDF
Figure 7 Simulation benchmarks on the methyl propyl ether mechanism from Johnson et al. 2021 for ConstantTP and ConstantV test cases. 15
Figure 7 Simulation benchmarks on the methyl propyl ether mechanism from Johnson et al. 2021 for ConstantTP and ConstantV test cases. 15CC-BY-NC-ND · From the PDF
Figure 8 Sensitivity benchmarks for computing the sensitivity of propane to rate coefficients and Gibbs free energy for the propane test case.
Figure 8 Sensitivity benchmarks for computing the sensitivity of propane to rate coefficients and Gibbs free energy for the propane test case.CC-BY-NC-ND · From the PDF
Figure 9 Mole fraction profiles for vapor-liquid thermal cracking of a mixture of phenyldodecane and undecane to gasoline-range hydrocarbons.
Figure 9 Mole fraction profiles for vapor-liquid thermal cracking of a mixture of phenyldodecane and undecane to gasoline-range hydrocarbons.CC-BY-NC-ND · From the PDF
Figure 10 Flux diagram for vapor-liquid thermally cracking of heavy oil to gasoline at 1 year.
Figure 10 Flux diagram for vapor-liquid thermally cracking of heavy oil to gasoline at 1 year.CC-BY-NC-ND · From the PDF

Verification

Claims the paper makes

0 Valid · 12 Not valid · 13 Not checked · 53 Own results and interpretations

32 more claims

Create a free account to see all 78 claims and to verify the 1 reference still open.

See all claims — free account
  • It may state two separate claims as one. 15
  • Probably not a substantial claim. 32

Create a free account to see these claims.

Citations

Cites

Show all 45 references

Research works printed but not linked: 2

  1. KeeRJ ColtrinME GlarborgP ZhuH.Chemically Reacting Flow: Theory Modeling and Simulation Second Edition. In:Intergovernmental Panel on Climate Change ed.John Wiley & Sons Inc.;2017.
  2. Blum P, 2021, US National Combustion Meeting

Cited by

See all 20

Metadata

Authors
Johnson, M. S., Pang, H., Payne, A. M., & Green, W. H.
Affiliations
Massachusetts Institute of Technology (US)
Corresponding author
Johnson, Matthew S.
Venue
International Journal of Chemical Kinetics
Publisher
Wiley
Published
2024-07-05
Volume
56
Issue
12
Pages
732-747
ISSN
  • 0538-8066 (ISSN-L)
  • 1097-4601
OpenAlex
W4400366443
Cited by
20
References
46
Citations by year
2026: 4 · 2025: 5 · 2024: 10
FWCI
1.69
Type
journal article
Language
en
Open access
Locations
  • International Journal of Chemical Kinetics (publishedVersion, cc-by-nc-nd)
  • ChemRxiv (submittedVersion, cc-by-nc-nd)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc-nd)
  • OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) (submittedVersion)
Funders
Advanced Research Projects Agency - Energy
Topics
  • Physical Sciences › Chemical Engineering › Fluid Flow and Transfer Processes › Advanced Combustion Engine Technologies
  • Physical Sciences › Mathematics › Modeling and Simulation › Modeling and Simulation Systems
  • Physical Sciences › Engineering › Computational Mechanics › Combustion and flame dynamics
Keywords
chemical kinetic mechanisms, mechanism simulation, multiphase systems, automatic differentiation, sensitivity analysis, adjoint sensitivity analysis, accident analysis, Julia, combustion, catalysis, electrocatalysis, pyrolysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-19, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-19 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: ReactionMechanismSimulator.jl: A modern approach to chemical kinetic mechanism simulation and analysis. https://app.cochat.ai/papers/10.1002/kin.21753
The paper
Johnson, M. S., Pang, H., Payne, A. M., & Green, W. H. (2024). ReactionMechanismSimulator.jl: A modern approach to chemical kinetic mechanism simulation and analysis. International Journal of Chemical Kinetics. https://doi.org/10.1002/kin.21753

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.