NASA NTRS1993
The damage tolerance behavior of internally pressurized, axially slit, graphite/epoxy tape cylinders was investigated. Specifically, the effects of axial stress, structural anisotropy, and subcritical damage were considered. In addition, the limitations of a methodology which uses coupon fracture data to predict cylinder failure were explored. This predictive methodology was previously shown to be valid for quasi-isotropic fabric and tape cylinders but invalid for structurally anisotropic (+/-45/90)(sub s) and (+/-45/0)(sub s) cylinders. The effects of axial stress and structural anisotropy were assessed by testing tape cylinders with (90/0/+/-45)(sub s), (+/-45/90)(sub s), and (+/-45/0)(sub s) layups in a uniaxial test apparatus, specially designed and built for this work, and comparing the results to previous tests conducted in biaxial loading. Structural anisotropy effects were also investigated by testing cylinders with the quasi-isotropic (0/+/-45/90)(sub s) layup which is a stacking sequence variation of the previously tested (90/0/+/-45)(sub s) layup with higher D(sub 16) and D(sub 26) terms but comparable D(sub 16) and D(sub 26) to D(sub 11) ratios. All cylinders tested and used for comparison are made from AS4/3501-6 graphite/epoxy tape and have a diameter of 305 mm. Cylinder slit lengths range from 12.7 to 50.8 mm. Failure pressures are lower for the uniaxially loaded cylinders in all cases. The smallest percent failure pressure decreases are observed for the (+/-45/90)(sub s) cylinders, while the greatest such decreases are observed for the (+/-45/0)(sub s) cylinders. The relative effects of the axial stress on the cylinder failure pressures do not correlate with the degree of structural coupling. The predictive methodology is not applicable for uniaxially loaded (+/-45/90)(sub s) and (+/-45/0)(sub s) cylinders, may be applicable for uniaxially loaded (90/0/+/-45)(sub s) cylinders, and is applicable for the biaxially loaded (90/0/+/-45)(sub s) and (0/+/-45/90)(sub s) cylinders. This indicates that the ratios of D(sub 16) and D(sub 26) to D(sub 11), as opposed to the absolute magnitudes of D(sub 16) and D(sub 26), may be important in the failure of these cylinders and in the applicability of the methodology. Discontinuities observed in the slit tip hoop strains for all the cylinders tested indicate that subcritical damage can play an important role in the failure of tape cylinders. This role varies with layup and loading condition and is likely coupled to the effects of structural anisotropy. Biaxial failure pressures may exceed the uniaxial values because the axial stress contributes to the formation of 0 deg ply splitting (accompanied by delamination) or similar stress-mitigating subcritical damage. The failure behavior of similar cylinders can also vary as a result of differences in the role of subcritical damage as observed for the case of a biaxially loaded (90/0/+/-45)(sub s) cylinder with a 12.7 mm slit. For this case, the methodology is valid when the initial coupon and cylinder fracture modes agree. However, the methodology underpredicts the failure pressure of the cylinder when a circumferential fracture path, suggestive of a 0 deg ply split, occurs at one slit tip. Thus, the failure behavior of some tape cylinders may be highly sensitive to the initial subcritical damage mechanism. Finite element analyses are recommended to determine how structural anisotropy and axial stress modify the slit tip stress states in cylinders from those found in flat plates since similarity of these stress states is a fundamental assumption of the current predictive methodology.