TY - JOUR
T1 - Knowledge database creation for design of polymer matrix composite
AU - Huang, Hannah
AU - Mojumder, Satyajit
AU - Suarez, Derick
AU - Amin, Abdullah Al
AU - Fleming, Mark
AU - Liu, Wing K
N1 - Funding Information:
H.H. would like to acknowledge the NSF REU program under the Grant No. MOMS/CMMI-1762035. S.M. and W. K. L. thankfully acknowledge the support provided by AFOSR (FA9550-18-1-0381). D.S. A.A.A. and W.K.L. acknowledge the support of the United States National Science Foundation (NSF) under Grant No. MOMS/CMMI-1762035. D.S. also gratefully acknowledges the Walter P. Murphy fellowship provided to first-year graduate students at Northwestern University.
Funding Information:
H.H. would like to acknowledge the NSF REU program under the Grant No. MOMS/CMMI-1762035. S.M. and W. K. L. thankfully acknowledge the support provided by AFOSR (FA9550-18-1-0381). D.S., A.A.A., and W.K.L. acknowledge the support of the United States National Science Foundation (NSF) under Grant No. MOMS/CMMI-1762035. D.S. also gratefully acknowledges the Walter P. Murphy fellowship provided to first-year graduate students at Northwestern University.
Publisher Copyright:
© 2022 Elsevier B.V.
PY - 2022/11
Y1 - 2022/11
N2 - We present a mechanistic data science (MDS) framework capable of building a composite knowledge database for composite materials design. The MDS framework systematically leverages data science to extract mechanistic knowledge from composite materials system. The composite response database is first generated for three matrix and four fiber combinations using a physics-based mechanistic reduced-order model. Next, the mechanistic features of the composites are identified by mechanistically analyzing the composites stress–strain responses. A relationship between the composite properties and the constituents’ material features are established through a mechanics constrained data science-based learning process after representing materials in latent space, following a dimension reduction technique. We demonstrate the capability of predicting a composite materials system for target properties (material elastic properties, yield strength, resilience, toughness, and density) from the MDS created knowledge database. The MDS model is predictive with reasonable accuracy, and capable of identifying the materials system along with the tuning required to achieve desired composite properties. Development of such MDS framework can be exploited for fast materials system design, creating new opportunity for performance guided materials design.
AB - We present a mechanistic data science (MDS) framework capable of building a composite knowledge database for composite materials design. The MDS framework systematically leverages data science to extract mechanistic knowledge from composite materials system. The composite response database is first generated for three matrix and four fiber combinations using a physics-based mechanistic reduced-order model. Next, the mechanistic features of the composites are identified by mechanistically analyzing the composites stress–strain responses. A relationship between the composite properties and the constituents’ material features are established through a mechanics constrained data science-based learning process after representing materials in latent space, following a dimension reduction technique. We demonstrate the capability of predicting a composite materials system for target properties (material elastic properties, yield strength, resilience, toughness, and density) from the MDS created knowledge database. The MDS model is predictive with reasonable accuracy, and capable of identifying the materials system along with the tuning required to achieve desired composite properties. Development of such MDS framework can be exploited for fast materials system design, creating new opportunity for performance guided materials design.
KW - Dimension reduction
KW - Materials design
KW - Mechanistic data science
KW - Mechanistic features
KW - Polymer composite
KW - Unidirectional fiber
UR - http://www.scopus.com/inward/record.url?scp=85135942734&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85135942734&partnerID=8YFLogxK
U2 - 10.1016/j.commatsci.2022.111703
DO - 10.1016/j.commatsci.2022.111703
M3 - Article
AN - SCOPUS:85135942734
SN - 0927-0256
VL - 214
JO - Computational Materials Science
JF - Computational Materials Science
M1 - 111703
ER -