Concrete serves as a prevalent construction material in various infrastructure elements, including bridge decks, airfield and highway pavements, offshore structures, and machinery foundations. Many of these structures undergo fatigue loading, which is a process of gradually introducing permanent internal changes…
Concrete serves as a prevalent construction material in various infrastructure elements, including bridge decks, airfield and highway pavements, offshore structures, and machinery foundations. Many of these structures undergo fatigue loading, which is a process of gradually introducing permanent internal changes to the material, resulting in the reduction of remaining life of the structure. Despite its inherent heterogeneity, concrete is often treated as homogeneous, disregarding the influence of its variations. These disparities significantly influence the life concrete subjected to fatigue loading. Hence, it's prudent to adopt a probabilistic approach that accommodates these divergent effects when estimating the fatigue life of cementitious materials. Artificial Neural Networks (ANNs) have emerged as a promising computational tool for addressing this challenge. ANNs embrace a probabilistic framework to model the intricate relationships within cementitious composites. In this study, an ANN tool has been employed to predict the fatigue life of both plain concrete and reinforced concrete beams across varying sizes: small, medium, and large. The model is trained using experimental data corresponding to small and medium specimens and then validated using data from large specimens. By incorporating material and fracture mechanics properties associated with concrete's softening behavior as input, this model can forecast fatigue life in terms of crack length. Notably, this approach offers distinct advantages over alternative methods as it takes into account the stochastic nature of concrete characteristics under fatigue loading, consequently providing a reasonably accurate prediction of concrete's fatigue life.