The semi-circular bend (SCB) test is widely used to characterize asphalt mixture cracking resistance. However, the practical usefulness of the test depends on the reliability of the measured fracture parameters. This study investigates SCB testing from a reliability perspective, with the aim of identifying the specimen number required for dependable interpretation and the testing conditions that provide the most stable response. The analysis considered nominal maximum aggregate size, notch depth, binder type, aging condition, test temperature, and loading rate. Fracture energy, peak load, and flexibility index, together with their cumulative coefficients of variation, were tracked from n = 3 to n = 6, while six-specimen raw datasets were used for Weibull reliability analysis. The results show that notch depth had the clearest effect on response stabilization, with the 15 mm notch providing the most reliable configuration and reaching the adopted variability limits earlier than the other notch depths. The descriptive Weibull analysis further indicated that the SBS-modified mixture exhibited the highest fracture-energy consistency within the tested dataset, whereas long-term aging, testing at 0 °C, and loading at 50 mm/min were associated with the lowest fracture-energy consistency within the tested dataset. Overall, SCB interpretation should be guided by response reliability, not mean fracture parameters alone. On this basis, a reliability-based SCB framework is proposed to support more dependable mixture comparison and more rational specimen planning.
Abbas et al. (Wed,) studied this question.