Development of a mixed-mode fracture criterion for a fabric composite manufactured by RTM

This investigation considers the influence of manufacturing on a composite material's resistance to delamination. Mixed-mode toughness data was measured for a five harness satin (5HS) carbon/epoxy manufactured by resin transfer moulding (RTM). The effect of fibre volume fraction on both initiation and propagation fracture toughness was also determined. Fibre volume fractions of 57% and 66% were considered under Mode I (interlaminar tension), Mode II (interlaminar shear) and Mixed-Mode I-II loading of 25%, 50% and 75%. Flat rectangular plates, from which specimens were obtained, were manufactured by RTM. Double cantilever beam (DCB) specimens were used for Mode I and Mixed-Mode I-II, following ASTM standards D5528 and D6671. End-notched flexure (ENF) specimens were used for Mode II. Initiation toughness therefore increased as the contribution of Mode II towards Mixed-Mode I-II delamination growth increased. Important toughening mechanisms were observed in all cases, which resulted in propagation fracture toughness being at least 200% higher than initiation fracture toughness. The dominant mechanism responsible for fracture work was the energy dissipated in creating and translating the crack tip damage zone across a non-planar path. The increase in fibre volume fraction decreased initiation fracture toughness at all mixed-mode ratios, and increased propagation fracture toughness at high Mode I contributions towards Mixed-Mode I-II delamination growth. Increasing the fibre volume fraction had no effect on propagation fracture toughness at a Mixed-Mode ratio of 75%. Additionally, the Virtual Crack Closure Technique (VCCT) implemented in AbaqusRTM was evaluated as an analysis tool to model the initiation and growth of delaminations in composites. Two finite element models of DCB specimen were built (2D and 3D). The material's delamination behavior was governed by the mixed-mode fracture data measured experimentally. Both DCB models had a fibre volume fraction of 57% and were analyzed under pure Mode I delamination growth in displacement control loading. The loads at which delaminations initiated in both the 2D and 3D models matched the average experimental loads to within 6%, and the minimum experimental load to within 0.7%. The subsequent loads that predicted delamination growth however were underestimated.