LAMINATED COMPOSITES BUCKLING ANALYSIS USING LAMINATION PARAMETERS, NEURAL NETWORKS AND SUPPORT VECTOR REGRESSION

Authors

  • RUBEM MATIMOTO KOIDE UTFPR - UNIVERSIDADE TECNOLóGICA FEDERAL DO PARANá
  • ANA PAULA C. S. FERREIRA UTFPR - UNIVERSIDADE TECNOLóGICA FEDERAL DO PARANá
  • MARCO A. LUERSEN UTFPR - UNIVERSIDADE TECNOLóGICA FEDERAL DO PARANá

Keywords:

COMPOSITE LAMINATE, LAMINATION PARAMETERS, BUCKLING, SUPPORT VECTOR REGRESSION, NEURAL NETWORK

Abstract

THIS WORK PRESENTS A METAMODEL STRATEGY TO APPROXIMATE THE BUCKLING LOAD RESPONSE OF LAMINATED COMPOSITE MATERIALS. IN ORDER TO GET REPRESENTATIVE DATA FOR TRAINING THE METAMODEL, SOME LAMINATE ANGULAR ORIENTATIONS ARE GENERATED WITH LATIN HYPERCUBE DESIGN. THESE LAMINATES ANGULAR ORIENTATIONS ARE CONVERTED INTO LAMINATION PARAMETERS SO THAT THE NUMBER OF INPUTS TO THE METAMODEL BECOMES CONSTANT. IT WAS COMPUTED THE BUCKLING LOAD USING FINITE ELEMENT FOR EACH LAMINATE. IN THIS WAY THE INPUTS-OUTPUTS METAMODEL TRAINING PAIRS ARE THE LAMINATION PARAMETERS AND THE CORRESPONDING BUCKING LOAD. NEURAL NETWORK AND SUPPORT VECTOR REGRESSION METAMODELS ARE DEVELOPED TO APPROXIMATE THE BUCKLING. THE PERFORMANCE OF THE SURROGATE MODELS FOR BUCKLING LOAD RESPONSE WAS COMPARED.

Downloads

Published

2014-12-17

Issue

Section

Articles