Flexural fatigue behavior of ultra-high-performance concrete: A Comparative study of two-parameter and three-parameter Weibull models
Youyou Zhang, Jiarui Zhao, Hucheng Feng, Haohui Xin
Abstract
Ultra-high-performance concrete (UHPC) has become a pivotal engineering material in modern construction due to its outstanding mechanical properties and fatigue resistance. However, probabilistic fatigue modeling of UHPC remains inadequately addressed, particularly concerning the applicability of advanced statistical distributions. This study systematically investigates the fatigue behavior of UHPC under cyclic bending loads through comprehensive experimental analysis and integration of literature data, employing both two-parameter and three-parameter Weibull distributions for probabilistic flexural fatigue strength. Digital image correlation (DIC) uncovers stress-dependent failure mechanisms: fiber-bridging effects induce crack energy dissipation to delay failure under low-stress conditions, whereas unstable main crack propagation governs brittle fracture at high stress levels. Crucially, the three-parameter Weibull distribution demonstrates superior predictive accuracy over its two-parameter counterpart through a location parameter that quantitatively defines the fatigue damage initiation threshold. By synergistically considering fiber-matrix interface strength distribution and fatigue crack growth resistance, this model provides enhanced precision in characterizing UHPC's endurance limit and multi-scale failure characteristics. Furthermore, it effectively resolves the two-parameter model's over-conservatism in high-stress regimes. This paper illustrates the three-parameter Weibull distribution as a powerful tool for probabilistic fatigue analysis, providing important insights into reliability design and durability optimization of UHPC structures.