(60a) Autonomous Systems for Experimental and Data-Driven Modeling of Combustion Kinetics

As energy demands continue to grow, energy security, energy economy, and climate change become increasingly pressing societal issues, yet combustion will likely remain a substantial fraction of the energy landscape for many years to come [1]. Therefore, future combustion technologies to address these pressing societal issues will need to be more fuel-efficient, produce fewer emissions, and operate on a wider variety of fuels, including bio-derived fuels. Furthermore, these technological improvements must be accomplished on the rapid timescales necessary to address these societal issues in an effective manner, necessitating predictive modeling capabilities.

Combustion of transportation-relevant fuels proceeds through complex chemical reaction mechanisms such that behavior of a given fuel inside an engine depends on the thermochemical properties of hundreds of intermediate species and rate properties of thousands of reactions [2] – many of which are unique to each fuel molecule. Models of these complex reaction mechanisms are "validated" using experimental data from several experiments. The current paradigm in validating reaction mechanisms involves manually simulating individual experiments and comparing the results to experimental data. To move beyond this requires a framework for automated testing against a wide variety of experimental data.

In this talk, we present extensions to an autonomous framework to automatically compare complex reaction mechanisms against experimental data.

First, we introduce the main elements:

• ChemKED [3] (Chemical Kinetics Experimental Data Format), a standardized, open-source, human- and machine-readable schema for describing chemical kinetics experiments and the data they produce.
• PyKED, a Python library for validating and interacting with data in the ChemKED data format.
• PyTeCK [4] (Python Testing of Chemical Kinetics), a Python library to automatically test reaction mechanisms using experimental data. PyTeCK uses Cantera [5] to perform simulations of each experiment described in a ChemKED file and compares the simulation results to the experimental data.

Until now, these tools have been limited to autoignition experiments performed in a shock tube or rapid compression machine. In this work we extend them to work with Jet Stirred Reactor (JSR) experiments. Specifically, we introduce a new ChemKED schema to represent composition profiles of species measured in JSR experiments, new PyKED classes to validate and interface with these data, and new PyTeCK methods to simulate this type of experiment and compare the results. We demonstrate this new autonomous framework on oxidation of n-heptane (C7H16) because of its importance as a primary reference fuel. Furthermore, we make it easier to extend to additional experiment types, such as laminar flame speeds, flame species profiles, and flow tube experiments.

Finally, we describe how this enabling work fits into a broader framework of autonomous science, in which a chemical kinetics model optimization method known as Multi Scale Informatics (MSI) [6,7] is coupled with automated JSR experiments, automated reaction mechanism generation [8], and automated quantum chemistry calculations [9], to build an automated system to design, perform, and interpret experiments and calculations that constrain parameter uncertainties to reduce uncertainties in quantities of interest.

References

1. Assessment of Fuel Economy Technologies for Light-Duty Vehicles, National Research Council, 2011, ISBN 978-0-309-15607-3
2. C.K. Westbrook, F.L. Dryer. Chemical Kinetic Modeling of Hydrocarbon Oxidation. Prog Energy Combust Sci 1984, 10, 1-57
3. Weber, B. W., & Niemeyer, K. E. (2018). ChemKED: A Human‐and Machine‐Readable Data Standard for Chemical Kinetics Experiments. International Journal of Chemical Kinetics, 50(3), 135-148 https://doi.org/10.1002/kin.21142
4. Niemeyer, K. E. PyTeCK version 0.2., Zenodo (2017). https://doi.org/10.5281/zenodo.596724
5. David G. Goodwin, Raymond L. Speth, Harry K. Moffat, and Bryan W. Weber. Cantera: An object-oriented software toolkit for chemical kinetics, thermodynamics, and transport processes. https://www.cantera.org, 2018. Version 2.4.0. https://doi.org/10.5281/zenodo.1174508
6. Burke, M. P. (2016). Harnessing the combined power of theoretical and experimental data through multiscale informatics. International Journal of Chemical Kinetics, 48(4), 212-235
7. Burke, M.P (2013). A quantitative explanation for the apparent anomalous temperature dependence of OH + HO2 = H2O +O2 through multi-scale modeling. Proceedings of the Combustion Institute, 34, 547-555
8. Gao, C. W., Allen, J. W., Green, W. H., & West, R. H. Reaction Mechanism Generator: Automatic Construction of Chemical Kinetic Mechanisms. Comput. Phys. Commun. 2016, 203, 212–225. https://doi.org/10.1016/j.cpc.2016.02.013

9. Bhoorasingh, P. L., Slakman, B. L., Khanshan, F. S., Cain, J. Y., & West, R. H. Automated Transition State Theory Calculations for High-Throughput Kinetics. J. of Phys. Chem. A. 2017, 121, 6896-6904. https://doi.org/10.1021/acs.jpca.7b07361