Prediction of elastic properties of Al 2124?SiC particulate composites using FEM and artificial neural networks. (21st January 2009)

• Record Type: Journal Article
• Authors: Vajralingam, B.; Hasan, Z.; Sehgal, D.K.; Pandey, R.K.
• Is Part Of: International journal of microstructure and materials properties. Volume 3:Number 6(2008)
• Page Start: 734
• Page End: 748
• Publication Date: 2009-01-21
• Subjects: representative volume element -- RVE -- particulate composite -- effective properties -- volume fraction -- FEM -- finite element method -- artificial neural networks -- ANNs -- elastic behaviour -- metal matrix composites -- MMC -- particulate composites -- uniaxial tensile loading -- simulation; Microstructure -- Periodicals; Materials -- Properties -- Periodicals; 620.11299
• Languages: English
• ISSNs: 1741-8410

Abstract

The elastic behaviour of Al 2124?SiC metal matrix particulate composites under uniaxial tensile load was studied. The ABAQUS Finite Element (FE) software was used for the simulation of different Al-SiC composites employing two-dimensional (2D) and three-dimensional (3D) models with different volume fractions of the second-phase material. A set of displacement boundary conditions was applied to the Representative Volume Element (RVE) for predicting the effective properties, such as the elastic modulus and the shear modulus. Spherical and cubical particles were investigated for different volume fractions. The FE results were combined with the Eshelby analytical model. The FE results were used as input to the back-propagation algorithm of Artificial Neural Networks (ANNs) to develop a programme for determining the effective properties of the particulate composites for different volume fractions. A comparison of results between the FE simulations, ANN technique and Eshelby model was made and further compared with the limited experimental data.