Flow dynamics in rock masses: 3D modelling for engineering problems management and resource safeguarding

This study investigates the role of complex networks of discontinuities (bedding, joints, fracture, and faults) for fluid flow dynamics within carbonate rock masses. Understanding and modelling how discontinuities interconnect and interact is crucial for resource management industries (e.g., geothermal energy and carbon capture), as well as for managing problem of water inflow during tunnelling. Our methodology integrates fieldwork studies with remote sensing techniques, including ground and drone-based photogrammetry, to build accurate models of discontinuity networks. We present how using machine and deep learning algorithms for semantic segmentation improves the identification of discontinuities in photogrammetric data and minimises biases induced by manual identification. This approach provides fast and reliable tools, essential in resource management and tunnelling operations. By merging theoretical insights with practical applications, we provide a comprehensive framework for understanding and modelling rock mass flow dynamics, vital for both management of engineering works and sustainable resource use.