About the talk

As applications of composite materials have broadened from aerospace to automotive and energy industries, their cost-effectiveness has become a subject of increased importance. Since in most cases, major cost lies in manufacturing, the effects of defects on performance has gained renewed attention. Earlier failure analyses have either not evaluated these effects explicitly or have taken account of these in artificial ways. Recent work, described as computational micromechanics, has attempted to remedy this situation. The work to be presented takes that approach and develops a statistical simulation scheme for characterization of manufacturing defects such as nonuniform fiber distributions by quantifying deviations from uniformity based on fiber mobility during manufacturing. Algorithms are implemented to generate representation of the “real” microstructure with quantified anomalies. The representative volume element realizations thus generated are subjected to boundary conditions under which the local stress fields are computed. Hierarchical energy based failure criteria are then used to determine occurrence of the sequence of failure events and their progression. By parametric studies, the effects of defects as well as constituent properties on failure behavior of composites under different loading conditions are illustrated.