Probabilistic multi-scale optimization of hybrid laminated composites

A. B.Ilyani Akmar, O. Kramer, Timon Rabczuk

Research output: Contribution to journalReview article

4 Citations (Scopus)

Abstract

This study presents a hierarchical multi-objective optimization over multiple scales of hybrid laminated composites. The fine-scale optimization problem is treated as a meso-level single-ply representative volume element (RVE) problem or lamina wherein the weave pattern is embedded in a matrix pocket. The weave pattern is the design variable of the first task considering the stochastic effects under uncertainties wherein four uncertain mesoscopic parameters are investigated: yarn spacing, yarn width, yarn height and misalignment in yarn angle. The fine-scale objective functions are to maximize the macroscopic elastic properties of single-ply RVE with periodic boundary conditions and optimize the pattern arrangement using evolutionary algorithm. The fine-scale optimization problem is done for a selected set of uncertainties by utilizing Latin Hypercube Sampling. The coarse-scale optimization problem is presented as the stacking sequence optimization of hybrid fiber-reinforced composite plate with two nonlinear objectives and two design constraints. The coarse-scale optimization goals are to minimize the cost and weight of the laminated plate with constraint on the first fundamental frequency and the buckling load factor. A multi-ply-ed, fiber reinforced and hybrid laminated composites are reconsidered with respect to the optimized macroscopic elastic properties of single-ply RVE in the fine-scale optimization problem. The investigated single-ply RVE is made of alumina oxide-aluminum (Al2O3-Al) and silicon carbide-aluminum (SiC-Al) plies to combine the toughness and economical attributes. Ant colony optimization (ACO) is utilized to formulate the Pareto-optimal solutions by optimizing a convex combination of the two nonlinear objectives, weight (W) and cost (C) based on a series of multiplier values (α). Simultaneously, the latter task could be simplified into a single-objective optimizer by employing the concept of weighted sum method. Conclusively, the best hybrid laminated composites based on the series of multiplier values are presented in the coarse-scale optimization problem.

Original languageEnglish
Pages (from-to)1111-1125
Number of pages15
JournalComposite Structures
Volume184
DOIs
Publication statusPublished - 2018 Jan 15
Externally publishedYes

Fingerprint

Laminated composites
Yarn
Aluminum Oxide
Aluminum
Fibers
Ant colony optimization
Multiobjective optimization
Silicon carbide
Evolutionary algorithms
Toughness
Buckling
Costs
Alumina
Boundary conditions
Sampling
Oxides
Composite materials

Keywords

  • Ant colony optimization
  • Evolutionary algorithm
  • Geometrical uncertainty parameters
  • Hybrid laminated composites
  • Multi-objective optimization

ASJC Scopus subject areas

  • Ceramics and Composites
  • Civil and Structural Engineering

Cite this

Probabilistic multi-scale optimization of hybrid laminated composites. / Akmar, A. B.Ilyani; Kramer, O.; Rabczuk, Timon.

In: Composite Structures, Vol. 184, 15.01.2018, p. 1111-1125.

Research output: Contribution to journalReview article

Akmar, A. B.Ilyani ; Kramer, O. ; Rabczuk, Timon. / Probabilistic multi-scale optimization of hybrid laminated composites. In: Composite Structures. 2018 ; Vol. 184. pp. 1111-1125.
@article{77b8ca27a4c248de9400e3593b61f5dc,
title = "Probabilistic multi-scale optimization of hybrid laminated composites",
abstract = "This study presents a hierarchical multi-objective optimization over multiple scales of hybrid laminated composites. The fine-scale optimization problem is treated as a meso-level single-ply representative volume element (RVE) problem or lamina wherein the weave pattern is embedded in a matrix pocket. The weave pattern is the design variable of the first task considering the stochastic effects under uncertainties wherein four uncertain mesoscopic parameters are investigated: yarn spacing, yarn width, yarn height and misalignment in yarn angle. The fine-scale objective functions are to maximize the macroscopic elastic properties of single-ply RVE with periodic boundary conditions and optimize the pattern arrangement using evolutionary algorithm. The fine-scale optimization problem is done for a selected set of uncertainties by utilizing Latin Hypercube Sampling. The coarse-scale optimization problem is presented as the stacking sequence optimization of hybrid fiber-reinforced composite plate with two nonlinear objectives and two design constraints. The coarse-scale optimization goals are to minimize the cost and weight of the laminated plate with constraint on the first fundamental frequency and the buckling load factor. A multi-ply-ed, fiber reinforced and hybrid laminated composites are reconsidered with respect to the optimized macroscopic elastic properties of single-ply RVE in the fine-scale optimization problem. The investigated single-ply RVE is made of alumina oxide-aluminum (Al2O3-Al) and silicon carbide-aluminum (SiC-Al) plies to combine the toughness and economical attributes. Ant colony optimization (ACO) is utilized to formulate the Pareto-optimal solutions by optimizing a convex combination of the two nonlinear objectives, weight (W) and cost (C) based on a series of multiplier values (α). Simultaneously, the latter task could be simplified into a single-objective optimizer by employing the concept of weighted sum method. Conclusively, the best hybrid laminated composites based on the series of multiplier values are presented in the coarse-scale optimization problem.",
keywords = "Ant colony optimization, Evolutionary algorithm, Geometrical uncertainty parameters, Hybrid laminated composites, Multi-objective optimization",
author = "Akmar, {A. B.Ilyani} and O. Kramer and Timon Rabczuk",
year = "2018",
month = "1",
day = "15",
doi = "10.1016/j.compstruct.2017.10.032",
language = "English",
volume = "184",
pages = "1111--1125",
journal = "Composite Structures",
issn = "0263-8223",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - Probabilistic multi-scale optimization of hybrid laminated composites

AU - Akmar, A. B.Ilyani

AU - Kramer, O.

AU - Rabczuk, Timon

PY - 2018/1/15

Y1 - 2018/1/15

N2 - This study presents a hierarchical multi-objective optimization over multiple scales of hybrid laminated composites. The fine-scale optimization problem is treated as a meso-level single-ply representative volume element (RVE) problem or lamina wherein the weave pattern is embedded in a matrix pocket. The weave pattern is the design variable of the first task considering the stochastic effects under uncertainties wherein four uncertain mesoscopic parameters are investigated: yarn spacing, yarn width, yarn height and misalignment in yarn angle. The fine-scale objective functions are to maximize the macroscopic elastic properties of single-ply RVE with periodic boundary conditions and optimize the pattern arrangement using evolutionary algorithm. The fine-scale optimization problem is done for a selected set of uncertainties by utilizing Latin Hypercube Sampling. The coarse-scale optimization problem is presented as the stacking sequence optimization of hybrid fiber-reinforced composite plate with two nonlinear objectives and two design constraints. The coarse-scale optimization goals are to minimize the cost and weight of the laminated plate with constraint on the first fundamental frequency and the buckling load factor. A multi-ply-ed, fiber reinforced and hybrid laminated composites are reconsidered with respect to the optimized macroscopic elastic properties of single-ply RVE in the fine-scale optimization problem. The investigated single-ply RVE is made of alumina oxide-aluminum (Al2O3-Al) and silicon carbide-aluminum (SiC-Al) plies to combine the toughness and economical attributes. Ant colony optimization (ACO) is utilized to formulate the Pareto-optimal solutions by optimizing a convex combination of the two nonlinear objectives, weight (W) and cost (C) based on a series of multiplier values (α). Simultaneously, the latter task could be simplified into a single-objective optimizer by employing the concept of weighted sum method. Conclusively, the best hybrid laminated composites based on the series of multiplier values are presented in the coarse-scale optimization problem.

AB - This study presents a hierarchical multi-objective optimization over multiple scales of hybrid laminated composites. The fine-scale optimization problem is treated as a meso-level single-ply representative volume element (RVE) problem or lamina wherein the weave pattern is embedded in a matrix pocket. The weave pattern is the design variable of the first task considering the stochastic effects under uncertainties wherein four uncertain mesoscopic parameters are investigated: yarn spacing, yarn width, yarn height and misalignment in yarn angle. The fine-scale objective functions are to maximize the macroscopic elastic properties of single-ply RVE with periodic boundary conditions and optimize the pattern arrangement using evolutionary algorithm. The fine-scale optimization problem is done for a selected set of uncertainties by utilizing Latin Hypercube Sampling. The coarse-scale optimization problem is presented as the stacking sequence optimization of hybrid fiber-reinforced composite plate with two nonlinear objectives and two design constraints. The coarse-scale optimization goals are to minimize the cost and weight of the laminated plate with constraint on the first fundamental frequency and the buckling load factor. A multi-ply-ed, fiber reinforced and hybrid laminated composites are reconsidered with respect to the optimized macroscopic elastic properties of single-ply RVE in the fine-scale optimization problem. The investigated single-ply RVE is made of alumina oxide-aluminum (Al2O3-Al) and silicon carbide-aluminum (SiC-Al) plies to combine the toughness and economical attributes. Ant colony optimization (ACO) is utilized to formulate the Pareto-optimal solutions by optimizing a convex combination of the two nonlinear objectives, weight (W) and cost (C) based on a series of multiplier values (α). Simultaneously, the latter task could be simplified into a single-objective optimizer by employing the concept of weighted sum method. Conclusively, the best hybrid laminated composites based on the series of multiplier values are presented in the coarse-scale optimization problem.

KW - Ant colony optimization

KW - Evolutionary algorithm

KW - Geometrical uncertainty parameters

KW - Hybrid laminated composites

KW - Multi-objective optimization

UR - http://www.scopus.com/inward/record.url?scp=85033466278&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85033466278&partnerID=8YFLogxK

U2 - 10.1016/j.compstruct.2017.10.032

DO - 10.1016/j.compstruct.2017.10.032

M3 - Review article

AN - SCOPUS:85033466278

VL - 184

SP - 1111

EP - 1125

JO - Composite Structures

JF - Composite Structures

SN - 0263-8223

ER -