Measuring the notched compressive strength of composite laminates: Specimen size effects

Large fibre reinforced composite structures can give much lower strengths than small test specimens, so a proper understanding of scaling is vital for their safe and efficient use. Small size (scale) specimens are commonly tested to justify allowable stresses, but could be dangerous if results are extrapolated without accounting for scaling effects. On the other hand large factors are sometimes applied to compensate for uncertainties, resulting in overweight designs. The most important variables of scaling effects on the strength of composites with open holes have been identified from experimental tests as notch size, ply and laminate thickness. In this study, these have been scaled both independently and simultaneously over a large range of combinations. The specimens are fabricated from commercially available (Hexcel Composites Ltd.) carbon/epoxy pre-impregnated tapes 0.125 mm thick (IM7/8552). The material is laid up by hand in unidirectional [0(4)](ns) with n = 2, 3, 4, and 8 (i.e., 2, 3, 4 and 8 mm thick) and multidirectional laminates; two generic quasi-isotropic layups, one fabricated with blocked plies [45(n)/90(n)/-45(n)/0(n)](s) and the other with distributed layers [45/90/-45/0](ns) with n = 2, 4 and 8 are examined. It is shown that the critical failure mechanism in these laminates is in the form of fibre microbuckling or kinking. The unnotched compressive strength in unidirectional specimens thicker than 2 mm is found to be limited by the stress concentration developed at the end tabs and manufacturing induced defects in the form of ply waviness, fibre misalignment and voids rather than specimen size (scaling). In the open hole specimens, for both lay-ups, the strength reduction observed is due to hole size effect rather than specimen thickness or volume increase. The open hole (notched) compressive strength results obtained compare favourably to predictions by a linear softening cohesive zone fracture model developed in earlier work by the second author. (C) 2007 Elsevier Ltd. All rights reserved.