Speaker
Description
Rocks, as porous and heterogeneous materials, exhibit fracture behaviors governed by intrinsic properties such as pore structure and connectivity, grain size, mineralogical composition, texture, anisotropy, and pre-existing microcracks. These factors influence the initiation, propagation, and coalescence of cracks, shaping the overall fracturing process. Despite advancements in experimental techniques enabling the detection of cracking levels and failure mechanisms, current methods often fall short of fully capturing the complex dynamics of rock fracture due to the interplay of mechanical and material properties along with challenges of interpreting large and intricate datasets. This study investigates the application of machine learning (ML) to analyze acoustic emission (AE) signals and ultrasonic monitoring waveforms to uncover new insights into fracture dynamics in rocks during laboratory rock mechanics tests. The primary focus is on determining whether ML can elucidate the details of microcracking and rupture evolution during rock fracturing.
Initial findings reveal a clear relationship between the effectiveness of active versus passive seismic monitoring when integrated with ML. For brittle rock specimens, the combination of passive AE monitoring and ML algorithms shows higher accuracy. Brittle failure produces high-magnitude AE events with distinct waveforms, enabling ML to classify and interpret cracking mechanisms reliably. Conversely, ductile rocks, characterized by aseismic cracking and diffuse energy release, benefit more from active ultrasonic monitoring. The controlled input of ultrasonic pulses enhances signal detection and allows ML to identify subtle changes associated with crack propagation and strain localization in ductile rocks.
The results highlight the importance of tailoring monitoring techniques to the specific mechanical behavior of the rock type under investigation. The synergy between ML and seismic monitoring—active and passive—provides a valuable framework for decoding the complexities of rock fracture processes. This approach holds potential for significant advancements in geotechnical engineering, energy resource extraction, and hazard mitigation in both underground and surface excavations.
| Country | USA |
|---|---|
| Acceptance of the Terms & Conditions | Click here to agree |
