Comment on "Strength Predictions of Composite Laminates with Unloaded Fastener Holes"

Abstract

ARBO and Ogonowski1 have presented interesting analytical and experimental results relating to the fundamental but complex problem of predicting the strength of composite laminates with holes. However, the following points need to be considered: 1) Interlaminar stresses have been shown2*4 to be important in the presence of stress-free boundaries. Although these stresses are generally restricted to highly localized regions, their influence on static and fatigue strength can be pronounced. The point-stress characteristic dimension is small compared to the thickness of the laminate, so the boundary-layer effect may have a significant influence on the characteristic dimension. The initial attempt by Whitney and Nuismer5 to estimate the point-stress characteristic dimension was based on a simple, approximate stress distribution. The present authors have resorted to the more elaborate stressfunction approach and, therefore, the analysis can be extended further to include the free-edge effects. 2) Notched tensile specimens with widths ranging from 1.9 cm to 7.6 cm (0.75 in. to 3 in.) are reported in the paper. Prabhakaran6 has shown that there is an inherent size-effect, reflected in the variation of unnotched tensile strength with specimen width; this size-effect has been shown to be significant when the 0° and 90° plies predominate and has to be filtered out in order to get the true hole-size effect. In comparing the strength of notched specimens with the strength of unnotched specimens, have the authors maintained the widths constant so as to avoid the inherent sizeeffect? Further, the authors state in one place that only data from tensile loaded tests are reported in the paper, while at another place they mention that the strength values of Table 2 are derived from tests performed on unnotched unidirectional laminate sandwich beams (and rail-shear specimens). This needs to be clarified. 3) The difference between the present paper and the twoparameter models of Whitney and Nuismer5 is basically twofold: a closed-form analysis is used here to determine the stress field around the hole, whereas an approximate expression is used in Ref. 5; also, the Tsai-Hill failure criterion is used here whereas the maximum normal stress criterion was used in Ref. 5. In view of the more elaborate analysis performed here, it is difficult to understand why the characteristic dimension was calculated empirically for one layup and then used for all layups. It would have been more logical to have calculated a characteristic dimension value for each