ABSTRACT
A variety of heterogeneous multi-phase materials have found their utilization in commercial and industrial applications today. Among them are metals and alloy systems containing dispersed precipitates, defects and pores, or polymer, ceramic or metal matrix composite materials with fibers, whiskers or particulates. The microstructure in many of these materials exhibits strong non-uniformities in spatial dispersion, shape, size, and orientation of the heterogeneities. Material properties, especially those related to damage and failure, depend on these morphological characteristics of the microstructure. Various experimental and numerical studies (e.g., in [63, 59, 82, 281, 285, 366]) have established that the deformation and damage behavior of multi-phase materials can be highly sensitive to the local morphology, resulting from non-homogeneous deformation. Nucleation and subsequent growth of localized damage due to particle cracking, interfacial decohesion, or matrix localization and cracking are strongly affected by microstructural non-uniformity. Prior to their deformation and failure analysis, it is therefore necessary to accu-
rately represent and characterize their microstructure. Computational methods of predicting microstructure-property relations have been made in several ways. These include the use of simplified geometry unit-cell models, correlation functions of the material properties, and simulation of the microstructure by random sets. With continued advancements in modern computing capabilities, the ability to include local microstructural features in simulations of microstructure evolution or materials performance has become an active area of research. Recently, microstructural section images from optical microscopy, scanning electron microscopy, and orientation imaging microscopy are being used to generate high-fidelity, image-based models for micromechanical analysis [242, 240, 241, 48, 166, 167]. The efforts rely on the collection of serialsection data to reconstruct microstructures that are subsequently used in a finite element mesh and model. These techniques are attractive in their ability to translate microstructures into high-fidelity, computational models with minimal heuristic assumptions. The present chapter discusses a few methods of image-based microstructure reconstruction that are important in multi-scale response and failure analysis discussed later in this book.
