Thickness effect on the compressive strength of T800/924C carbon fibre–epoxy laminates

Abstract

Abstract In this study, the effect of laminate thickness on the compressive behaviour of composite laminates was investigated through systematic experimental work using the stacking sequences, [0 4 ] ns , [45/0/−45/90] ns and [45 n /0 n /−45 n /90 n ] s ( n =2–8) with a 3 mm diameter open hole at the centre. Parameters such as fibre volume fraction, void content, fibre waviness and interlaminar stresses, influencing compressive strength with increasing laminate thickness were also studied experimentally and theoretically. Furthermore the stacking sequence effects on failure strength of multidirectional laminates were examined. For this, two different scaling techniques were used: (1) ply-level technique [45 n /0 n /−45 n /90 n ] s and (2) sublaminate level technique [45/0/−45/90] ns . An apparent thickness effect existed in the lay-up with blocked plies, i.e. unidirectional specimens ([0 4 ] ns ) and ply-level scaled multidirectional specimens ([45 n /0 n /−45 n /90 n ] s ). Fibre waviness and void content were found to be main parameters contributing to the thickness effect on the compressive failure strength. However, the compressive strength of the sublaminate level scaled specimens ([45/0/−45/90] ns ) was almost unaffected regardless of the specimen thickness (since ply thickness remains constant). From the investigation of the stacking sequence effect, the strength values obtained from the sublaminate level scaled specimens were slightly higher than those obtained from the ply level scaled specimens. This applied to all specimen thicknesses regardless of the presence of an open hole. The reason for this effect was explained by the fibre waviness, void content, free edge effect and stress redistribution in blocked 0° plies and unblocked 0° plies. Finally the average strengths of open hole specimens obtained from both stacking sequences increased with increasing specimen thickness. The measured failure strengths were compared with the predicted values.