Deakin University
Browse

File(s) under permanent embargo

Theory-guided machine learning for damage characterization of composites

journal contribution
posted on 2020-08-15, 00:00 authored by N Zobeiry, Johannes ReinerJohannes Reiner, R Vaziri
© 2020 Elsevier Ltd A novel approach for damage characterization through machine learning is presented where theoretical knowledge of failure and strain-softening is linked to the macroscopic response of quasi-isotropic composite laminates in over-height compact tension tests. A highly efficient continuum damage finite element model enables the training of a system of interconnected Neural Networks (NNs) in series solely based on the macroscopic load-displacement data. Using experimental results, the trained NNs predict suitable damage parameters for progressive damage modeling of IM7/8552 composite laminates. The predicted damage properties are validated successfully using experimental measurements obtained through cumbersome non-destructive data analysis. The proposed strategy demonstrates the effectiveness of machine learning to reduce experimental efforts for damage characterization in composites.

History

Journal

Composite Structures

Volume

246

Article number

ARTN 112407

Pagination

1 - 8

Location

Amsterdam, The Netherlands

ISSN

0263-8223

eISSN

1879-1085

Language

English

Publication classification

C1 Refereed article in a scholarly journal

Publisher

ELSEVIER SCI LTD