(73a) Assessing Discrepancies in Kinetic Parameters and Improving Combustion Models through Metaheuristic Optimization

Recent developments have enabled scientists to develop complex reaction mechanisms to describe the combustion of novel fuels, but with the proliferation of new models, discrepancies in the kinetic and thermodynamic parameters arise. In order to detect and resolve these discrepancies, three key challenges need to be addressed: (1) how to compare kinetic paraments between models; (2) how to assess the truthfulness of a reaction mechanism; and (3) how to identify correct kinetic parameters. To address these challenges, we present a framework that can automatically identify isomorphic elementary reactions in various models, assess the error of models, and minimize these errors using a genetic algorithm. To identify isomorphic reactions from kinetic models published in CHEMKIN format, we have developed a tool we call the Importer [1]. The Importer (built upon the framework of the open-source Reaction Mechanism Generator (RMG) [2]) generates reactions, compares them with the CHEMKIN reactions, proposes matching species, and uses human input to confirm these proposals. Identified species and reactions are then easily comparable across models that have been imported into our database. To assess the error of these imported models, we collected experimental ignition delay data from the literature, using the YAML-based ChemKED data format [3]. PyTeCK (Python tool for Testing Chemical Kinetics) [4], a validation tool built using Cantera [5], is utilized to assess the error of a reaction mechanism against the experimental data. From the above, we implement a genetic algorithm to identify parameters that minimize the average error of all models. In our genetic algorithm, we vary the unique kinetics for each reaction present in our importer to minimize the average error of models we test. We demonstrate the efficacy of this tool using three models that describe the combustion of butanol [6], heptane [7], and 3-methyl-heptane [8] to generate models that accurately describe the ignition of these fuels. Finally, we intend to apply this workflow to the remaining 69 models in our importer tool and for varying experimental data types.

[1] West, R.. Combustion Mechanism Importer and Kinetic Models. (2017). https://doi.org/10.6084/m9.figshare.4787893

[2] Gao C.W., J.W. Allen, W.H. Green, R.H. West, Reaction mechanism generator: automatic construction of chemical kinetic mechanisms, Comput. Phys. Commun., 203 (2016), 212-225, https://doi.org/10.1016/j.cpc.2016.02.013

[3] B.W. Weber, K.E. Niemeyer, ChemKED: a human- and machine-readable data standard for chemical kinetics experiments, Int. J. Chem. Kinetics (2017), https://doi.org/10.1002/kin.21142

[4] K.E. Niemeyer, PyTeCK: a python-based automatic testing package for chemical kinetic models, S. Benthall, S. Rostrup (Eds.), 15th Python in Science Conference (SciPy 2016) (2016), 82-89

[5] D.G. Goodwin, H.K. Moffat, R.L. Speth, Cantera: an object-oriented software toolkit for chemical kinetics, thermodynamics, and transport processes, 2017, (http://www.cantera.org). Version 2.3.0. https://doi.org/10.5281/zenodo.170284

[6] S.M. Sarathy, S. Vranckx, K. Yasunaga, M. Mehl, P. Oßwald, W.K. Metcalfe, C.K. Westbrook, W.J. Pitz, K. Kohse-Höinghaus, R.X. Fernandes, H.J. Curran, A comprehensive chemical kinetic combustion model for the four butanol isomers, Combust. Flame, 159 (6) (2012) 2028-2055, https://doi.org/10.1016/j.combustflame.2011.12.017

[7] Mehl M., W.J. Pitz, C.K. Westbrook, H.J. Curran, "Kinetic Modeling of Gasoline Surrogate Components and Mixtures Under Engine Conditions", Proc. Combust. Inst. 33:193-200 (2011). https://doi.org/10.1016/j.proci.2010.05.027

[8] W. Wang, Z. Li, M.A. Oehlschlaeger, D. Healy, H.J. Curran, S.M. Sarathy, M. Mehl, W.J. Pitz, C.K. Westbrook, An experimental and modeling study of the autoignition of 3-methylheptane, Proc. Combust. Inst., 34, (2013), 335-343, https://doi.org/10.1016/j.proci.2012.06.001