Fatigue failure prediction for unidirectional CFRP using SFF model with entropy damage criterion

This study extends the Simultaneous Fiber-Failure (SFF) model to predict longitudinal tensile fatigue failure in unidirectional carbon-fiber-reinforced polymers (UD-CFRP) by embedding an entropy-based damage criterion for the matrix resin. The framework assumes that neither the carbon fibers nor the fiber–matrix interface exhibit fatigue degradation under the considered loading and temperature conditions; thus, strength reduction is attributed to progressive matrix degradation, which promotes clustered fiber breaks and lowers the composite load-carrying capacity. Matrix fatigue is quantified through entropy generation associated with viscous dissipation and damage-related loss of stored elastic energy in a nonlinear viscoelastic constitutive law based on the generalized Maxwell model. The constitutive model was implemented in finite element analysis to identify the relationship between entropy and residual matrix strength, which was then introduced into the extended SFF formulation to enable fatigue-life and residual-strength prediction while explicitly accounting for fiber-failure clustering. Since no existing model is available for predicting the longitudinal tensile fatigue life of UD-CFRP while considering both fiber failure and matrix resin damage, the predictive capability of the proposed framework was assessed through comparison with experimental S–N data at multiple temperatures. The predicted S–N curves reproduced the overall experimental trends, demonstrating the feasibility and predictive potential of the proposed SFF extension. The proposed framework may also be applicable to variable or interrupted loading histories, provided that entropy can be evaluated.