Effects of Defects on Compressive Kinking and Fracture of Fiber Reinforced Composite Materials and Structures.

Abstract

The successful implementation of composite materials in a design is predicated upon establishing its properties early in the design cycle. Composite material strength allowables are especially important because these set the limits on the stability and integrity of a structures. These strengths, however, are extremely sensitive to material defects introduced during manufacturing. The study of defects and its influence on material strength is referred to as defect mechanics.

In this dissertation, studies on two prevalent failure mechanisms in defect mechanics are presented. These failure mechanisms are; (a) Kinking, and (b) Fracture. Kinking failure refers to kink band formation in composites under compressive load. Fracture refers to initiation and progression of a crack from a pre-existing crack like defect within a composite.

Experimental, analytical and numerical analysis were conducted to study kinking failure in composites with defects, like waviness and holes. Analysis methods with varying degree of fidelity are presented to compute strength knock-downs due to these defects. A new analytical formulation for kink band formation was also developed. This formulation provides pre-peak, peak and post-peak response during kinking failure. The formulation highlights some of the deficiencies of previous analytical models and can potentially replace micro-mechanics based finite element models for strength prediction.

The topic of fracture is explored at a fundamental level, where a critical review of classical energy based fracture mechanics was conducted. Using experimental results and examples from literature it is shown that Linear Elastic Fracture Mechanics (LEFM) based energy release rate methods are of limited use for predicting certain types of crack growth. An alternative theory of crack initiation and growth that does not rely on the criticality of the energy release rate for crack growth is presented. This theory is shown to cover a wider range of crack initiation and growth problems. It also provides a phenomenological explanation for non-smooth crack growth and R-curve behavior observed in experiments with composite specimens.