Characterizing fatigue damage evolution in asphalt mixtures using acoustic emission and Gaussian mixture model analysis

Abstract

The identification and investigation of fatigue crack evolution and damage modes in asphalt mixtures are crucial for understanding the corresponding mechanisms and characteristics of fatigue failure. In this study, the four-point bending fatigue tests were performed on the asphalt mixture specimens with precast joints using the acoustic emission (AE) technique to analyze the damage evolution. The failure modes of the asphalt mixtures were identified using the Gaussian mixture model (GMM). The results revealed that the fatigue damage of the asphalt mixtures can be categorized into four stages: Stage I (void compaction), Stage II (microcrack initiation and stable propagation), Stage III (crack aggregation and unstable propagation), and Stage IV (complete fracture). This division was defined based on the inflection points of the axial displacement curves, the cumulative number of AE events, and the cumulative ringing counts. The evolution of the AE b value was found to be correlated with the crack propagation. The rapid decline in the b value in Stage III was considered a harbinger of the eventual complete fracture in the asphalt mixtures. The GMM clustering analysis indicated that the fatigue damage of the asphalt mixtures primarily involved tensile damage (90%), and shear damage made a small contribution of 10%. Existing AE studies have focused on the quantitative evaluation of AE parameters, while research on clustering identification algorithms for these parameters is scarce. The analytical methods and findings of this study enable an effective identification of fatigue damage in asphalt mixtures at the laboratory scale and offer insights into the fatigue damage mechanisms of asphalt mixtures.