Geometric Design of Dog-Bone Specimens for Accurate Fatigue Life Characterization of High-Strength CFRP Laminates

Tension–tension fatigue testing of polymer matrix composites (PMCs) conducted per ASTM D3479/D3479M using rectangular specimens is widely plagued by premature crack initiation and propagation at the edges of reinforcing grip tabs, which leads to severe underestimation of the material’s actual fatigue life. While dog-bone specimen geometries have been universally adopted to mitigate this issue, and benchmark studies have validated their ability to completely eliminate grip-region failures in low-to-intermediate-strength PMCs, our preliminary work identified a critical unaddressed limitation: standardized dog-bone configurations produce highly unreliable fatigue characterization results for T800-grade and higher-strength carbon fiber-reinforced polymer (CFRP) laminates, with experimentally measured fatigue lives deviating significantly from predictions derived from classical laminate theory. To resolve this discrepancy and enable accurate fatigue performance quantification for high-strength CFRP laminates, the present work focuses specifically on the transition region geometry of dog-bone specimens, which we hypothesized to be the source of spurious premature failures in high-strength laminate testing. The study is bounded to tension–tension fatigue loading regimes relevant to high-performance structural applications of T800-grade and above CFRP laminates, with the core objective of developing an optimized geometry that eliminates premature non-gauge-section failures. First, statistical analysis of a large dataset of preliminary tests confirmed that transition region geometric parameters exert a non-negligible effect on the measured fatigue performance of advanced high-strength fiber-reinforced polymer laminates; stress concentrations induced by non-optimized geometries were identified as the root cause of premature non-gauge-section failures even in conventional dog bone specimens. We then systematically varied transition region geometric parameters, performed finite element stress modeling to quantify full-field stress distributions for each candidate geometry, and conducted parallel tension–tension fatigue tests on all designed configurations to cross-validate simulation outputs and experimental performance. Our results demonstrate that the optimized dog-bone configuration developed in this work completely eliminates all spurious non-gauge-section failure modes. Fatigue lives measured using the optimized geometry show excellent agreement with classical laminate theory predictions, enabling robust, repeatable quantification of the intrinsic fatigue life of high-strength CFRP laminates. The proposed configuration addresses the longstanding reliability gap associated with standardized dog-bone geometries for high-strength fiber-reinforced polymer fatigue characterization.