Interlaminar stresses in composite notched and unnotched laminates

The purpose of this paper is to determine the interlaminar stress distribut ion st curved and straight free edges for quasi-isotropic and zero-dominate d laminates. The interlaminar stresses are calculated using a three-dimensi onal finite-element code with error control. For straight free edges the re sults are compared to existing semi-analytical methods. Another objective w ith this investigation is to evaluate a quadratic failure criterion for pre diction of initial load for delamination and the most likely position for t he delamination. In this paper it is found, for quasi-isotropic stacking se quences, that it is not possible to tailor a stacking sequence so that sigm a(z) becomes compressive near hole boundaries or at straight free edges. On the contrary both tensile and compressive stresses appear in a periodic ma nner through the thickness. It can also be concluded that interlaminar stre sses exhibit a thickness effect for repeating bundles. The singularity of t he interlaminar components is concentrated to a narrow region, approximatel y 1-2 fibre diameters in thickness and radial direction. Since the interlam inar stresses are so localised, a continuum approach will not provide usefu l solutions in this area Consequently the material cannot be treated as hom ogeneous in this region, instead the heterogeneous nature within each ply s hould be modelled. The smallest element size in the present analysis for th e refined mesh is less than 1 mu m, i.e., far less than a fibre diameter. U tilising a quadratic failure criterion, for the studied quasi-isotropic lam inates, in conjunction with obtained FE-results, it is shown that global st rain to initiate delamination does not depend on the stacking sequence, but so does the number of possible initiation points. For quasi-isotropic layu ps with a circular cut-out, initiation of delaminations occurs at approxima tely the same global strain, 0.25%. For unnotched zero-dominated layups ini tiation and fracture strain are almost the same, i.e., similar to 1.2%. The investigation also shows that the number of +/-45 degrees/0 degrees or +/- 45 degrees/90 degrees alterations should be kept at a minimum to minimise t he interlaminar stresses and possible damage initiation points. It is likel y that the primary cause for delamination initiation for notched plates are the high interlaminar shear stresses and not the normal stresses. Finally it is demonstrated that available semi-analytical methods do not correctly describe the interlaminar stress distribution.