Research
The Investigation of How Rocks Fracture, Flow, and Evolve Across Scales
From microscopic cracks to fault‑scale seismic processes, we study the fundamental physics that govern deformation and damage, revealing how these processes shape the mechanical and seismic behavior of Earth materials.
Our work integrates advanced laboratory experiments with multiphysics measurements, high‑resolution imaging, and modern data‑science tools. By combining these approaches, we capture the full lifecycle of rock damage across all stages to explore microstructural changes and the signals they generate.
Through this multiscale perspective, we aim to improve forecasting of fracture evolution, enhance subsurface energy and engineering systems, and deepen understanding of both natural and human‑induced seismicity.
Research areas
-
Graphic analysis of energy release
-
Seismic magnitude clustering
-
EGS permeability/flow
-
Finding seismicity within the noise of fiber optics DAS
-
Linking geophysical microcrack observations with rock damage evaluation
-
Non-local triggering in rock fracture
-
Microcrack damage observations via acoustic emission
-
Linking physical measurements with imaging
-
Influence of damage permeability
Graphic Analyses of Energy Releases
Applying graph neural networks (GNNs) to energy-release data allows us to make predictions about individual properties of events and the larger-scale behavior of the material. By exploiting the natural structure of earthquake (or microseismic) data, GNNs far surpass the capabilities of traditional models. Whether being applied at the node level (e.g. predicting the magnitude of events) or at the whole graph scale (e.g. predicting fracture coalescence), these models imitate the deeper relationships involved in stress transfer to generalize to different scales and environments. With larger-scale applications in energy extraction and hazard mitigation, these network-based models could improve forecasting and operational strategies.
Seismic Magnitude Clustering is Prevalent in Field and Laboratory Catalogs
Clustering of earthquake magnitudes is still actively debated, compared to well-established spatial and temporal clustering. Magnitude clustering is not currently implemented in earthquake forecasting but would be important if larger magnitude events are more likely to be followed by similar sized events. Here we show statistically significant magnitude clustering present in many different field and laboratory catalogs at a wide range of spatial scales (mm to 1000 km). It is universal in field catalogs across fault types and tectonic/induced settings, while laboratory results are unaffected by loading protocol or rock types and show temporal stability. The absence of clustering can be imposed by a global tensile stress, although clustering still occurs when isolating to triggered event pairs or spatial patches where shear stress dominates. Magnitude clustering is most prominent at short time and distance scales and modeling indicates >20% repeating magnitudes in some cases, implying it can help to narrow physical mechanisms for seismogenesis.
EGS Permeability/Flow Imaging and Optimization
Permeability/Flow imaging is needed as a first step to mitigate the impacts of ubiquitous flow heterogeneity/channeling that exists within geothermal systems. Initial focus is on imaging flow and permeability fields using large specimen sizes (> 30 cm) and realistic yet variable true-triaxial stresses (external stresses up to 30 MPa) and temperatures (up to 200°C). These experiments will also focus on thermal performance prediction and dynamic optimization of circulation using graph theory and operational surface controls.
Finding Seismicity Within the Noise of Fiber Optic DAS
Earthquake (EQ) signal detection and phase picking are challenging problems in seismology. Detection refers to identifying EQ signals within a dataset that contains noise and non-EQ signals, and phase picking refers to identifying distinct seismic phases, such as P- and S-waves, within an EQ signal. Deep Learning (DL) can be used for detection/picking because it learns from high-level representations of available data streams. In large datasets such as DAS, current computational constraints require high-level representations of the data to extract usable information on seismicity, with specific regions of the seismic signals receiving greater focus. This can be accomplished by replicating how humans interpret their visual field. Specifically, attention mechanisms can be used in DL to replicate the efficient way humans identify changes in space and time.
Linking Geophysical Microcrack Observations with Rock Damage Evolution
The goal of the proposed project is to identify and link critical microcrack parameters to macroscale physical property behavior, providing a method for interpreting and predicting the evolution of rock/concrete behavior under various loading conditions (i.e., stress and stress path). Laboratory investigations to separate the macro-scale contributions of microcrack parameters will be used to analyze the dependency of damage on stress and stress path. The deconvolution of microcrack parameter-controlling mechanisms into macro-scale behavior will help put site-specific observations of geologic/civil structure damage into perspective, thereby providing a context for predicting rock/concrete behavior due to microcracking.
Non-local Triggering in Rock Fracture
Combining multiphysics observations in the laboratory, we present novel phenomena for the analogue between rock fracture and seismicity. We show, for the first time in a laboratory setting, how a large‐scale flaw in rock can facilitate “non‐local triggering”—remotely triggered damage in rock mass. Results prove analogues of rock fracture evolution to natural seismicity including several lab‐analogues on seismicity model predictions beyond power law. We observe under specimen‐scale criticality a relatively small AE perturbation occurring at a seemingly random location can trigger cascading events whose spatial span can traverse the entire rock fracture system. The inclination angle of the non‐local triggering pairs shows dependence on the large‐scale flaw, and the subsequent generations of triggered AEs by these non‐local triggers decline significantly slower than that of the general triggers. The rock fracture system evolves to a state where the resistance to non‐local triggering continuously reduces, bridging the gap between an idealized lab setting and a natural fault. The non‐locally triggered events can also be reactivated as the rock fracture system evolves towards increased complexity. These lab‐observations as well as the understanding on the analogue relationship between rock fracture and seismicity may shed new light on the insight of the periodical, dynamic, and unpredictable nature of the multiscale rock fracture process.
Microcrack Damage Observations via Acoustic Emission
Hydraulic fracturing is a common practice in several industries and environments, including energy production using enhanced geothermal systems, hydrocarbon extraction from unconventional oil and/or gas reservoirs, and mining and civil engineering excavation methods. Understanding damage associated with coalesced fractures induced by hydraulic fracturing and the surrounding material is fundamental for predicting material and system behavior. Inducing fracture networks in rock can create extensive microcracking in surrounding regions not connected to the wellbore. The degree of microcracking can vary depending on the fluid, rock type, stress/temperature boundary conditions, and inherent material properties and heterogeneities. Regions of rock containing microcracks near coalesced macro-scale fractures can behave differently than the original matrix material due to the permanent structural changes. Understanding how coalesced fractures interact with the surrounding rock containing microcracks requires characterizing damage in terms of physical property evolution.
Linking Physical Measurements with Imaging
The use and integration of many physical measurements is crucial for understanding the impact of changing boundary conditions. For example, acoustic emissions observed throughout a laboratory fracture test can provide an immense amount of information regarding the microcracking processes and damage, but the nature of the coalesced fracture, its roughness, conductivity, or opening width is not readily captured with the AE method. Other geophysical or imaging techniques must be integrated to provide a full characterization of the deformation. Conversely, using CT images alone will not describe the progressive nature of the fracturing process. This research examines the integration of multiple physical measurements and imaging techniques to enhance understanding of fracture development and its interactions with existing discontinuities and the matrix material.
Influence of Damage on Permeability
Understanding the role of rock damage in the form of microfractures, compaction and dilation bands, and alteration of the geochemical make-up of rocks is fundamental for subsurface engineering and safe energy extraction and sustainable reservoir development. Currently, knowledge is lacking in the characterization of damage across multiple length scales and its influence on material behavior. In subsurface engineering of oil/gas extraction, rock damage is purposely induced to achieve specific goals, but the role of micro-scale damage on the performance of these systems is not well understood. Consequences of poorly characterized rock damage and the subsequent rock mass physical property changes can lead to poor reservoir development practices, lower production, and lost time and investment.
