Year: 2024
Type: conference-paper
Abstract: In multiscale simulations traditional data driven methods have failed to produce reliable results for highly nonlinear problems in comparison to classical numerical computational homogenization algorithms. Deep material network is a novel hybrid machine learning solver framework which gives reliable predictions in the nonlinear elastic domain with just training using linear elastic responses. The work discusses the implementation of this workflow and successful benchmarking of this technology for state-of-the-art case studies.
Author: BHAT KEELANJE SRINIVAS, PAVAN
Institution: Fraunhofer Institute for Industrial Mathematics
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949906065
FWCI:
Cites:
Cited by:
Related to: