Analytical and Numerical Study of the Notched Tensile Strength of Open-Hole Composite Laminates Using the Whitney–Nuismer Point and Average Stress Criteria
The growing replacement of metallic alloys by polymer matrix composites in the transport and mobility industries has intensified the need for reliable methods to predict the strength of laminates weakened by fastener holes. As a preliminary step toward the design of composite bolted joints, this work investigates the open-hole (notched) tensile strength of carbon fiber/epoxy laminates using a uniaxial fracture framework based on both the Point Stress Criterion (PSC) and the Average Stress Criterion (ASC) of Whitney and Nuismer. Six balanced carbon/epoxy fabric stacking sequences ranging from [0]5 to [45]5 were analyzed for four different hole diameters (3.18, 6.35, 9.52, and 12.70 mm), together with the unnotched reference, totaling 30 evaluated conditions (six different stacking sequences × four different hole diameters and one unnotched laminate). The Konish–Whitney analytical stress distribution was applied to both criteria. The characteristic dimensions d0 (PSC) and a0 (ASC) were obtained by fitting reference experimental data. The ASC integral was evaluated numerically over the same stress distribution used for the PSC. A complementary two-dimensional finite-element model with 360-element azimuthal mesh refinement was developed and validated against the classical isotropic infinite-plate solution (Kt ≈ 3), with an accuracy better than 0.1%. Across all 24 (laminate × diameter) combinations, the ASC reduced the root-mean-square prediction error from 15.1% (PSC) to 11.0%, and the mean absolute error from 11.7% to 7.8%. For balanced [0/45] configurations both criteria remained within ±10% of the experimental notched strength, with the ASC narrowing the band to ±3.6%; unidirectional laminates, non-conservative by up to 16% under the PSC, were brought below 7% by the ASC, and ±45°, dominated stacks remained conservative (15–23%), reflecting matrix-driven failure mechanisms outside the scope of either point-wise criterion. Condensed error maps are used to delineate the envelope of applicability of each criterion on the fabric laminate dataset of Soriano and Almeida. The comparative assessment supports the ASC as the more accurate of the two criteria for fiber-controlled configurations within the dataset, while highlighting that any transfer of the calibrated characteristic dimensions to other material systems, geometries, or off-axis-dominated layups would require dedicated validation and, for off-axis stacks, progressive damage modeling.