Abstract
A novel approach for the fracture characterization of soft adhesive materials using spiral cracking patterns is presented in this study. This research particularly focuses on hydrocarbon polymeric materials, such as asphalt binders. Ten different asphalt materials with distinct fracture characteristics were investigated. An innovative integrated experimental-computational framework coupling acoustic emissions (AE) approach in conjunction with a machine learning-based Digital Image Analysis (DIA) method was employed to precisely determine the crack geometry and characterize the material fracture behavior. Cylindrical-shaped samples (25 mm in diameter and 20 mm in height) bonded to a rigid substrate were employed as the testing specimens. A cooling rate of -1 °C/min was applied to produce the spiral cracks. Various image processing techniques and machine learning algorithms such as Convolutional Neural Networks (CNNs) and regression were utilized in the DIA to automatically analyze the spiral patterns. A new parameter, "Spiral Cracking Energy (E(Spiral))", was introduced to assess the fracture performance of soft adhesives. The compact tension (CT) test was conducted at -20 °C with a loading rate of 0.2 mm/min to determine the material's fracture energy (G(f)). The embrittlement temperature (T(EMB)) of the material was measured by performing an AE test. This study explored the relationship between the spiral tightness parameter ("b"), E(Spiral), G(f), and T(EMB) of the material. The findings of this study showed a strong positive correlation between the E(Spiral) and fracture energies of the asphalt materials. Furthermore, the results indicated that both the spiral tightness parameter ("b") and the embrittlement temperature (T(EMB)) were negatively correlated with the E(Spiral) and G(f) parameters.