During the combustion of hydrogen-oxygen mixtures, molecules transform into a variety of ephemeral intermediates and generate an exponentially growing number of sequences of chemical species. Experiments continue to extract data from these elementary chemical processes at ever shorter timescales. However, there is a fundamental need to transform these data into chemical mechanisms, despite the constraints on the experimentally accessible information. Theoretically, the combinatorial explosion in the number of sequences makes it challenging to efficiently learn what subset of sequences are necessary to include and which can be neglected in macroscopic descriptions. In this talk, I will introduce our recent work applying theoretical tools from information theory and statistical mechanics that overcome these challenges. I will introduce a variational generalization of the typical set that respects experimental constraints and allows for the systematic discovery of chemical mechanisms with molecular dynamics data and techniques for its efficient computation.

Information and mechanistic patterns of hydrogen oxidation

256th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.