Feature Embedding of Molecular Dynamics-Based Descriptors for Modeling Electrochemical Separation Processes

H. K.Gallage Dona, T. Olayiwola, L. A. Briceno-Mena, C. G. Arges, R. Kumar, J. A. Romagnoli

Research output: Chapter in Book/Report/Conference proceedingChapter

1 Scopus citations

Abstract

Machine learning is increasingly being used as a modeling technique for the analysis of complex systems even when limited knowledge is available. Here, we demonstrate the application of machine learning models to predict macroscopic properties of electrochemical polymers from molecular attributes. The computational framework proposes a novel approach to estimate molecular dynamic simulations (MD) attributes and experimental activity coefficients. Results showed that the data augmentation and embedding strategy effectively produce unique representations of each polymer. The findings from this study allow for the estimation of activity coefficients of novel polymers without the need for new time-consuming MD simulation runs. This data processing technique could then be used map to material properties that influence for electrochemical ionic separation units.

Original languageEnglish (US)
Title of host publicationComputer Aided Chemical Engineering
PublisherElsevier B.V.
Pages1451-1456
Number of pages6
DOIs
StatePublished - Jan 2023

Publication series

NameComputer Aided Chemical Engineering
Volume52
ISSN (Print)1570-7946

All Science Journal Classification (ASJC) codes

  • General Chemical Engineering
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'Feature Embedding of Molecular Dynamics-Based Descriptors for Modeling Electrochemical Separation Processes'. Together they form a unique fingerprint.

Cite this