The vast majority of developments in fracture mechanics have been instigated by the need to avoid fracture of metals. Consequently, these developments are based on the assumption that the material is homogeneous—an assumption that is far more palatable in the case of most metals than it is for fiber-reinforced composites. Attempts to make fracture mechanics concepts more applicable to composite materials have resulted in the emergence of anisotropic fracture mechanics approaches. But, while this simple extension of linear elastic fracture mechanics does remove the assumption of material isotropy, it still assumes the material to be homogeneous. Experimental evidence suggests that the application of anisotropic fracture mechanics to fiber-reinforced composites still leads to inconsistencies in the prediction of fracture.

This paper presents an improvement in the application of a concept that allows for direct consideration of heterogeneous material behavior in the near crack tip region. The concept is similar to the singular perturbation and matched asymptotic expansion techniques used in fluid mechanics. The problem of a fiber-reinforced composite containing a flaw is divided into a local heterogeneous region (LHR) and a global anisotropic homogeneous continuum region. In the LHR, local failure events can be individually modeled, and then related to a global fracture parameter. Specifically, a quasi-three-dimensional finite-element LHR model is used to investigate micromechanical failure events at the tip of a crack in a fiber-reinforced composite laminate. These events are then related to a global fracture event, whereupon the fracture toughness of the composite can be estimated. Computational results, although in seemingly reasonable agreement with existing experimental data, are heuristic in nature. The work represents only a preliminary effort toward possible future development of a more sophisticated predictive model.