Jialiang Le, Assistant Professor, Civil, Environmental and Geo-Engineering
Reliability analysis of asphalt structures is of paramount importance for pavement design, which usually aims to ensure a low failure probability. To determine the design strength of pavement against such a low failure probability, one must rely on some well-established probability distribution functions of the strength of asphalt mixtures. Asphalt mixtures at normal temperatures behave in a ductile manner, and the corresponding strength distribution must follow the Gaussian distribution, which can be easily calibrated by histogram testing. However, at low temperatures (< 20C), which are often encountered in Minnesota, asphalt mixtures behave in a quasibrittle manner. Compared to ductile materials, the strength distribution of quasibrittle materials is much more complicated. Recent studies provided a theoretical framework to derive the probability distribution function of quasibrittle materials, and it was shown that the strength distribution varies with the structure size. Consequently, histogram testing on specimens of different sizes is needed to determine the strength distribution of asphalt mixtures, which greatly increases the experimental cost as well as the risk of experimental errors. This research is aiming to develop a novel method to accurately determine the strength distribution of asphalt mixtures in an indirect manner, where tedious histogram testing is no longer required. The method will be derived from a rigorous probabilistic mechanics theory and be fully validated by a set of comprehensive experiments on asphalt mixtures.