Computational modeling of failure in composite structures including uncertainties in material and geometrical properties

This paper is concerned with a progressive failure analysis methodology for fiber reinforced composite laminates combining various analytical models designed for investigating failure mechanisms at different length scales. The methodology here employs a fundamental mechanism based approach to predict failure or damage initiation with strong coupling between the multiple length scales. The discrete cohesive zone model elements are used to model the adhesion and delamination failure at macroscale while Schapery theory, a continuum damage theory based on thermodynamics, is used to model material degradation occurring at the lamina level. Furthermore, the present numerical framework is incorporated with a probabilistic analysis module, based on the NEESUS software, to consider material variability and manufacturing inconsistencies. The combined analysis modules are implemented in a non-linear finite element code for modeling the progressive failure of advanced composite structures. The proposed progressive failure analysis methodology is applied to several cases for validating its capability of predicting the evolution of the interactive failure mechanisms in composite structures.