Title: Semi-inverse Monte Carlo reconstruction of two-phase heterogeneous material using two-point functions

Authors: H. Garmestani, M. Baniassadi, D.S. Li, M. Fathi, S. Ahzi

Addresses: Georgia Institute of Technology, School of Materials Science and Engineering, 771 Ferst Dr., Atlanta, GA 30332-0245, USA. ' University of Strasbourg, IMFS, 2 Rue Boussingault, 67000 Strasbourg, France. ' Georgia Institute of Technology, School of Materials Science and Engineering, 771 Ferst Dr., Atlanta, GA 30332-0245, USA. ' Faculty of Electrical and Computer Science, University of Siegen, Holderlinstr. 3, 57068 Siegen, Germany. ' University of Strasbourg, IMFS, 2 Rue Boussingault, 67000 Strasbourg, France

Abstract: A new Monte Carlo (MC) methodology using ant Colony and kinetic growth models is developed to reconstruct the microstructure of two-phase composites using correlation statistics. After the initial patterns are generated, the new patterns evolved by introducing optimisation parameters for rotation, shrinkage, transportation and distribution. The realisations are then optimised using kinetic growth rate applied to each representative cell. For each realisation, the final microstructures are then compared to the experimental results through minimisation of the error function. Finally an optimisation methodology was developed to introduce initial input parameters for MC simulation to minimise the error function.

Keywords: microstructure reconstruction; semi-inverse reconstruction; Monte Carlo simulation; two-point function; statistical continuum mechanics; ant colony optimisation; kinetic growth models.

DOI: 10.1504/IJTAMM.2009.029210

International Journal of Theoretical and Applied Multiscale Mechanics, 2009 Vol.1 No.2, pp.134 - 149

Published online: 10 Nov 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article