Experimental and probabilistic reliability of fiber-reinforced concrete (FRC) under drop-weight impact

Concrete structures subjected to impact loading exhibit considerable variability arising from material heterogeneity, non-uniform fiber dispersion, and stochastic crack development, making deterministic performance assessment alone insufficient. This study presents an integrated experimental and probabilistic framework for evaluating fiber-reinforced concrete (FRC) incorporating six fiber types: steel, glass, polypropylene, polyvinyl alcohol, carbon, and basalt. Two concrete grades (M25 and M80) were investigated through compressive strength and repeated drop-weight impact tests. Nested analysis of variance (ANOVA) was used to quantify the influence of fiber type and dosage, while a hierarchical Bayesian framework was employed to characterize uncertainty and reliability associated with compressive strength and impact resistance. The results revealed pronounced fiber and grade-dependent behavior, with the optimum steel fiber mixtures (SF-1.50 for M25 and SF-1.25 for M80) increasing impact resistance by approximately 141% and 166%, respectively, relative to the control mixtures. Steel fibers consistently exhibited the highest impact resistance and reliability, while basalt and carbon fibers emerged as the most effective non-metallic alternatives. ANOVA demonstrated that fiber type exerted a significantly greater influence on impact performance than dosage. Reliability assessment and probabilistic multi-objective optimization consistently identified SF-1.50 and SF-1.25 as the optimum mixtures for M25 and M80 concretes, respectively. The proposed framework provides a systematic basis for reliability-informed evaluation and selection of fiber-reinforced concrete subjected to repeated impact loading.