An alternative formulation of concrete fatigue S–N curves considering dynamic strength

Abstract

This paper addresses a limitation in the conventional S–N curve approach for concrete fatigue, which assumes that failure occurs in the first cycle when the maximum cyclic stress equals the static strength (S_max= 1). Since the compressive strength of concrete is sensitive to the loading rate, an alternative approach is developed and validated that replaces the intercept value of 1 with the dynamic increase factor (DIF). This proposal, termed Stress Rate Dependent S–N curves (SRD S–N), is integrated into the fatigue models of Eurocode 2, Model Code 2020, and Zhang, evaluating its accuracy against several experimental datasets. The results demonstrate that considering dynamic strength significantly improves the statistical fit; while conventional regulatory models show poor coefficients of determination (R² < 0.50), the SRD variants increase the average R² to 0.79. Likewise, the root mean square error (RMSE) is practically reduced by half, dropping from 0.54 to 0.34 logarithmic units. The SRD models predict longer fatigue lives for low-cycle fatigue and are more conservative for high-cycle fatigue. Finally, new design expressions are proposed for the SRD models, providing a physically-based predictive framework that optimizes the design of structures subjected to dynamic loading.