Maximizing structure performances of a sandwich panel with hybrid composite skins using particle swarm optimization algorithm

Abstract

A sandwich panel, composed of hybrid laminate skins of AL (aluminum)-CFRP-GFRP and aluminum honeycomb core, was optimized for maximizing the structural performance. Stacking sequence of the three different materials comprising the hybrid laminate skins and individual ply angles are taken as design variables in the present optimization problem. Synergizing a particle swarm optimization (PSO) algorithm method with a specially developed FEM program enables one to optimally decide the design variables and thereby significantly improve the sandwich performance. The present technique applying PSO to a hybrid sandwich in conjunction with FEA has extended the application area of optimization with a complex honeycomb sandwich that is not possible by the conventional method.

An A-STAR algorithm for thick composite laminate design

“…3 Evolutionary techniques constitute another approach capable of providing high quality designs, as shown in references. [7][8][9][10] These evolutionary alternatives do not require any gradient computation and, in most cases, they are not sensitive to stacking at local optima. Genetic algorithms (GA) have proven effective in high-level optimum search performing and have been widely used to determine the optimum laminate stacking sequence for a laminate subjected to external forces and manufacturing constraints, see references.…”

Engineering Applications of Satellite Structural Optimization Utilizing ANSYS/CATIA

“…It is quite important to create a FEA model with both reliable accuracy and practical feasibility. Generally, the research hot points in and abroad are centering around three aspects in order to avoid employing a whole satellite FEA model rashly with large degrees of freedom (DOFs) and fuzzy mechanical characteristics, or depicting design problem without quantificational formulation: FEA element formation of basic components [6] , idealization of real structures [7][8][9] , mathematical description of optimization model [10] . The large amount of design variables and design constraints, as well as the strong non-linear traits of state variables, have justified the utilization of an optimization method with efficiency and accuracy to avoid finding only local extremum design points or jumping into much computation cost [11][12] .…”