Date of Award

2016-01-01

Degree Name

Master of Science

Department

Computational Science

Advisor(s)

Pavana Prabhakar

Abstract

Ultralight sandwich structures comprising of low-density core with stiff facings have attracted significant research interest for their considerable weight saving applications. The aircraft industries are focusing on decreasing the structural mass to lower the manufacturing and operating costs. Design analysis of the sandwich cores using finite element analysis has been developed as a promising concept to feature sandwich structures with maximum strength, stiffness, and reduced weight. To obtain multifunctional behavior of sandwich panels, a profound investigation of geometrical and mechanical properties in the transverse plane is required because it is very susceptible to any kind loadings. Structural optimization is one of the key factors for designing lightweight structures, where the main concern is not merely to ensure an intricate design, but also to identify the limiting factors and resolve the issues by generating optimum values of the main parameters.

This Thesis presents the design optimization of multifunctional sandwich panels in two chapters. The first chapter reports the shape optimization approach of four different core topologies considering three-dimensional isotropic patterns that are optimally designed for minimum weights. Additive manufacturing technology is a suitable and amenable method for the construction of sandwich structures because it ensures strong bonding between the facings and core to reduce the slipping. Fused deposition modeling method is employed to build the 3D printed structures. Short beam shear tests were carried out on the initially non-optimized structures to generate the structural response. Peak loads and deformations were recorded to compare the flexural properties. To obtain the new design of the sandwich cores with optimum stiffness and reduced weight shape optimization task is performed by ABAQUS. Stress and weight are the design variables to carry out the optimization method. Shape optimization process deals with the coordinates of surface nodes; eventually, it creates a new design of the cores that demonstrates versatile performance. Finally, based on the output of the optimization procedure new STL files are imported in the additive manufacturing machine to produce the optimized structure. Optimized panels are subjected to short beam shear test again to investigate their performance that has changed by employing shape optimization. Comparison using the mechanical properties are subsequently performed for the optimized and non-optimized panels to demonstrate the overall responses numerically. Results show that optimized structures are significantly lighter that perform decently from the strength standpoint with diverse characteristics such as ductility and brittleness.

Algorithms, like a genetic algorithm, mimics natural process can be employed in the structural optimization technique. In this paper, both finite element analysis and genetic algorithm are employed to obtain the optimum result of the cross- sectional area for truss structures. The area is the main variable for this optimization technique that can be expressed by the array of binary numbers to carry out genetic algorithm operation and subsequently stress analysis is performed using the material properties. Since minimization of the weight is the objective function, so decreasing the cross-sectional areas subjected to a higher stress of the truss members and allowable stress operates as a stopping criterion for this iterative process. Finally, stress analysis and genetic algorithm create a possible solution set for areas and weight of the unit cell for the truss structure is determined. FEA is conducted by combining FEA (using ABAQUS) and genetic algorithm that is implemented in MATLAB.

The findings shown in this Thesis have established appropriate weight saving technique for sandwich structures. The work provided a solid foundation for structural optimization that utilizes finite element package and a robust tool genetic algorithm which is not found in the commercial software packages.

Language

en

Provenance

Received from ProQuest

File Size

81 pages

File Format

application/pdf

Rights Holder

Mohammad Tauhiduzzaman

Share

COinS