Experimental and Computational Exploration of the Fracture Energy in Fiber Composites When Subject to Crack Parallel Compression

Abstract

This paper explores the global Mode I fracture energy of a carbon fiber composite subject to a biaxial stress state at a crack tip. Both an experimental campaign, in which a novel test setup for composite materials is designed and developed, and a computational campaign are performed to quantify the effects of the biaxial stress state. Using photomicroscopy (PMG) the results from these efforts are explained via meso-scale damage mechanisms at the crack tip. This research represents a gradational step in the continued development of composite fracture mechanics and, as a salient result, indicates the criticality of modeling fiber composites with a finite width fracture process zone (FPZ) and a fully tensorial damage law when simulating quasi-static fracture. Experimentation is performed via a simple modification to the standard three-point-bend (3PB) test on a single edge notched bend (SENB) specimen. Polypropylene pads with a near perfectly plastic yield plateau are placed adjacent to the crack mouth and are used to apply a constant crack parallel compression prior to a bending moment which is induced by rigid rollers that are engaged only once the pads have yielded. Three different size test specimens, geometrically scaled in-plane and held at a constant thickness, are tested. A pre-impregnated, Toray T800H/3900-2 material system is used, and composite panels are manufactured via hand layup, vacuum bagging, and autoclave curing processes. All test specimens are symmetric cross-ply laminates (comprised only of 0° and 90° plies). Cracks are installed into the specimens manually with a mitre box that guides a thin kerf blade (0.508 mm thick) to the desired crack length. Testing is performed on an Instron 5585H load frame where force data is obtained via a transducer on the load frame and displacement is measured through 2D digital image correlation (DIC). The fracture energy is determined using Bažant’s Type II Size Effect Law (SEL). From which it is found that there is a monotonic decrease in the Mode I fracture energy as the crack parallel compressive stress increases. Compared to the nominal value of fracture energy, where no crack parallel compression is applied, the fracture energy is observed to decrease by up to 37% for a compressive stress equal to 44% of the compressive failure limit of the composite. This weakening effect is attributed to splitting cracks that are induced at the crack tip due to the crack parallel compression. This is a novel result that challenges the century old hypothesis of fracture energy being a constant material property. The experiments are also repeated computationally using finite element analysis (FEA) where two modeling approaches are employed. First, the Hashin damage criterion in conjunction with a crack band model and second, cohesive elements are used to define the crack band with a max stress damage criterion. When using the Hashin damage criterion in the crack band the weakening effect and decrease in fracture energy observed experimentally are accurately captured. Conversely, when cohesive elements and the maximum stress criterion model the crack band the simulations overpredict the fracture energy and are grossly off compared to the experimental results. This is explained by the complexity of the constitutive equations that govern damage initiation within the crack band. That is, the Hashin criteria is fully tensorial, including all experienced crack tip stresses in its failure calculation whereas the maximum stress criteria is a reduced tensorial damage law and the presence of a crack parallel compression is absent. This emphasizes the importance of using a crack band model coupled with a full tensorial damage law to accurately predict fracture in composites.