A probabilistic prediction method for mode I fatigue life of concrete based on the statistical laws of material parameters

Accurately predicting the mode I fatigue life of concrete is critical for evaluating the safety of structures. However, due to the stochasticity of fatigue life, it remains a challenge to the deterministic methods in the existing literature. This study presents a probabilistic method for predicting the mode I fatigue life of concrete. The method considers four concrete material parameters in the initial fracture toughness-based crack propagation criterion and the tension-softening constitutive relationship as random variables following a two-parameter Weibull distribution. The detailed procedure for the implementation of the method is elaborated using the notched three-point bending (TPB) beam as an example, and the effectiveness of the method is verified by comparing the predicted fatigue life with the experimental results collected from the literature. Furthermore, the sensitivity analysis shows that the stochasticity of the initial fracture toughness is the most important factor leading to the variability of the fatigue life. Tensile strength stochasticity is the second most important factor. The effects of other material parameters on fatigue life are negligible. The proposed method allows reasonable prediction of fatigue life based solely on prior knowledge of the statistical laws governing the two factors. Meanwhile, the calculation results suggest that the large scatter in fatigue life is due to the high sensitivity to the stochasticity of concrete material parameters. The proposed method is expected to enhance the understanding of the stochastic nature of fatigue life and achieve accurate predictions at different failure probabilities.