Real-time monitoring of reaction mechanisms from spectroscopic data using hidden semi-Markov models for mode identification

Data Fusion-Based Approach for the Investigation of Reaction Networks in Hydrous Pyrolysis of Biomass

“…11 Importantly, comparison with analyses for the HTL of cellulose (represented by levoglucosan) and lignin (represented by 2-phenoxy-ethyl benzene) revealed that the network hypothesized for biomass breakdown was a combination of the networks for individual decompositions of its components as A1 and A2 of the biomass network corresponded to elements in the cellulose and lignin networks. Thus, this work presents a data-driven approach to infer reaction networks for complex reaction mixtures and relate them to the reaction networks for individual constituents of the feed and can potentially be used to develop reaction hypotheses, process designs, and process monitoring techniques 70 for biomass feedstocks of varying composition. The dominant reaction network that is well represented through the spectroscopic profiles is captured through the process.…”

Iterative multiscale and multi-physics computations for operando catalyst nanostructure elucidation and kinetic modeling

“…Puliyanda et al. 131 developed the hidden semi-Markov models for both offline and online monitoring of FTIR spectroscopic data, extracted the “latent” features (species identities) in the spectra using Bayesian networks, and interpreted the reaction mechanism and timescales for the identified modes, for low-temperature thermal cracking of Cold Lake bitumen. In another study for the same reaction system, a data injection/fusion model based on the joint non-negative matrix factorization and Bayesian networks was developed to extract the pseudo-component spectra and hypothesize the reaction mechanisms, respectively.…”

Semi-supervised machine learning approach for reaction stoichiometry and kinetic model identification using spectral data from flow reactors

“…[19][20][21] Recently, soft-modelling approaches have also been applied to spectral data for monitoring and discovery of reaction pathways using spectral data. 22,23 Solutions obtained from soft modelling are ambiguous in nature even for reaction systems with less than four species, and hence deter its use for complex systems involving multiple reactions. [24][25][26] Although it is shown that hard modelling methods can partly solve the rotational ambiguity problem usually encountered with soft modelling approaches, they require prior knowledge of the stoichiometry and kinetic model or pure component spectra of the reaction species.…”