Optimizing small-sample fatigue testing strategy: A unified S-N curve modeling for composite laminates

Abstract

Accurately predicting the fatigue life of laminated structures remains challenging, given the wide range of fiber and matrix types, lay-up sequences, and stress ratios. This paper proposes an optimal small-sample fatigue testing strategy based on a unified S–N curve model spanning LCF, HCF, and UHCF fatigue regions. This strategy eliminates the need for segmented fitting, reduces the number of parameters to be identified, and thereby minimizes the need for extensive fatigue testing. Based on the amount of available test data, four S-N curve models with two, three, or four parameters are developed, which each parameter reflects a clear physical meaning in relation to the fatigue strength. Experimental generalization capability validation on various laminate materials and stacking sequences as well as loading conditions demonstrates strong correlation (R² > 0.93). Moreover, the number of fatigue tests was reduced 40%, while maintaining a 95% confidence interval for a GFRP example. The proposed models support engineers in quickly and accurately obtaining material-level S-N curves, thereby finding practical application in material and layup comparison at the conceptual design stage and serving as input for detailed finite element simulations of FRP structures in aerospace and automotive industries.