Parisa Shokouhi

Accurately forecasting injection‐induced seismicity remains a major challenge that is critically important for energy production and storage as controlling mechanisms are complex and remain opaque. We clarify key processes by following time‐histories of bespoke parameters anticipated to control strength reduction (drawdown) through overpressure during fluid injection. Injection‐driven radial overpressure fields motivate key parameter groups representing magnitudes () and rates () of strength‐drawdown. We emphasize these physics‐informed features as inputs embedded in a two‐stage machine learning (ML) architecture applied to high‐resolution microseismic and injection data from the Utah FORGE and EGS Collab stimulation experiments. Permuted models, selectively including then excluding these physically‐inspired terms, predict temporal seismic evolution and capture trends and peaks with testable and discernible fidelity. Improved predictions result from including these diagnostic parameters—confirming that the derived physical parameters mechanistically mediate seismicity. The magnitude of strength drawdown () dictates the size and localization of seismic events during phases of steep spatial pressurization—particularly early in the injection phase near the well. Conversely, governs the timing and rate of onset of seismic activity during rapid injection transients, accounting for elapsed‐time effects. Our approach demonstrates the broader benefits of embedding physics‐guided features into ML frameworks toward mitigation of seismic hazard.

A range of natural (teleseismic waves) and industrial sources (hydraulic stimulation) of stress perturbations in the subsurface can alter the poromechanical response of fractured rock which may reactivate faults or otherwise breach reservoir seals. We describe laboratory experiments that seek to illuminate the physical processes that dictate these effects. Our experiments focus on the roles of fracture interface stiffness and fracture infill (sediment, proppant or fault gouge) and how they modulate the hydraulic and elastodyamic properties of fractured rock. Experiments were conducted on Westerly granite using flat surfaces with uniform roughness subject to triaxial stresses while monitoring fluid permeability evolution using deionized water. Thin layers of quartz powder within fractures were used to study the role of infill and the poromechanical response of clogging and unclogging of pore throats. Pore pressure oscillations of amplitude 0.2–1 MPa and frequency of 1 Hz were used to simulate subsurface dynamic stresses. Before, during, and after the oscillations we monitored fracture elastic properties using piezoelectric transducers (PZTs). We measured stress‐induced changes in transmitted ultrasonic wave velocity and amplitude to estimate the nonlinear elasticity of the fracture concurrent with permeability evolution. We find that increasing gouge layer thickness up to fracture roughness results in enhanced permeability and a greater nonlinearity in the elastodynamic response. Thicker layers, comparable to or larger than the fracture roughness, result in reduced permeability enhancement and elastodynamic response. Our results offer insights for techniques to monitor the evolution of subsurface fractures during energy production and dynamic stressing via distant earthquakes.

Physics‐based handcrafted features, such as P‐wave amplitude derived from early parts of active source ultrasonic waveforms, successfully predict shear stress evolution. This study investigates whether feature extraction can be automated by leveraging full active source ultrasonic waveforms recorded during a laboratory friction experiment. We compare four models: handcrafted feature‐driven model (reference), convolutional neural network (CNN) model, explainable data‐driven (XDD) model, and a model with an explainable generalized physics‐based (XPG) feature. The reference model uses the spectral amplitude of the P‐wave packet calculated at center frequency, while the CNN relies on convolutional and pooling layers for feature extraction. The XDD model combines CNN feature extraction with long short‐term memory units and an attention layer to capture feature time history and significance. The novel XPG extracts a generalized spectral amplitude feature through a custom multilayer perceptron designed to learn linear combination of signals' spectral amplitudes without prescribing specific frequencies. Although CNN‐extracted features resemble handcrafted ones, both the CNN and XDD models underperform compared to the reference, while the XPG model outperforms all others. The attention mechanism in the XDD model highlights the importance of first P‐wave packet, whereas the XPG model also identifies later‐arriving shear wave packet as significant. The prediction improvement by inclusion of shear waves is physically intuitive; S‐waves have particle motion polarized in the direction of fault shear and are likely more sensitive to frictional state than P‐waves. This work demonstrates the value of incorporating domain knowledge into automatic feature extraction with potential applications in seismic data analysis.

Thermal Modulation (TM) is a nonlinear approach that has shown success in detecting damage-related variations in uniform materials such as steel and aluminum. Its underlying principle suggests promising use in ultrasonic (US) monitoring systems by exploiting natural temperature fluctuations to evaluate structural integrity rather than compensating for them. Despite its potential, the application of TM to concrete remains relatively unexplored and not yet fully understood given the intrinsically complexity of concrete. The technique’s response is sensitive to several parameters, including the range and rate of temperature change, humidity, temperature gradients, and the inherent properties of the material. To contribute to this developing field, the present study investigates the influence of one key material property — the water-to-cement (w/c) ratio — on the TM response of concrete. Expanding upon previous work that considered two ratios, this research examines four (0.45, 0.50, 0.55, and 0.60). Twelve concrete specimens were casted, three for each mix. One specimen per group was maintained as a reference, while the remaining samples were subjected to incremental uniaxial compression ranging from 5% to 95% of their estimated compressive strength, in 5% steps. Between each loading stage, all the specimens experienced at least two 24-hour thermal cycles (10°C–25°C) at 60% relative humidity and no applied load. Throughout these cycles, ultrasonic measurements were continuously collected and processed using Coda Wave Interferometry (CWI) to calculate relative velocity variations (dv/v). The resulting data were used to determine Thermal Modulation Coefficients (TMC). Initial observations reveal that TMC values tend to rise just before the samples reach their elastic limit (approximately 50–60% of ultimate strength), with more pronounced effects in mixes with higher w/c ratios. This pattern indicates that both material composition and the progression of damage play significant roles in shaping the TM response. In summary, this study provides preliminary evidence of the potential of combining TM and CWI techniques for assessing damage in concrete. The results shed light on how the w/c ratio influences TM behaviour and offer valuable insights for refining the approach. Continued research is essential to enhance understanding and confirm the viability of TM as a diagnostic tool for Structural Health Monitoring (SHM) in concrete infrastructure.

Understanding what controls permeability is important in modulating crustal processes and the recovery of fuels and energy from the subsurface. Permeability may be created by the reactivation of fractures creating connected flow pathways that result in microearthquakes (MEQs). We explore the influence of temperature on the linkage between permeability evolution and induced MEQs through a series of concurrent flow‐through reactivation experiments conducted on laboratory faults at elevated temperatures (20°–140°C). A distinct proportional relationship is observed between changes in permeability () and seismic moment () across the tested temperature range. These observations are consistent with a mechanistic model accommodating shear dilation on the reactivated fault. We show a direct, power‐law link between permeability change and seismic moment. Two prefactors that scale this relationship reduce with increasing temperature and are proportional to shear stress drop and shear dilation angle and inversely proportional to shear modulus. The sensitivity to temperature is mainly due to a decrease in shear dilation angle when temperature‐activated deformation mechanisms increase. In addition to these laboratory data, observations from four geothermal reservoirs at temperatures in the range 20°–250°C confirm this power‐law relationship and similarly indicate that the prefactor λ in the expression decreases with increasing temperature. This temperature‐dependent relationship can provide credible constraints on the progress of permeability increase and induced seismicity in field‐scale reservoir stimulations.

Nonlinear ultrasonic parameters are highly sensitive to microstructural features that affect macroscale material behavior, providing a nondestructive means to characterize their evolution. Although dislocations are known to be a strong source of acoustic nonlinearity, establishing quantitative links between the acoustic nonlinearity parameter (β), measured via Second Harmonic Generation, and dislocation morphology—such as dislocation length and density—remains an open challenge. This work advances the numerical modeling of dislocation–β relationships using 3D dislocation dynamics (DD) simulations in two approaches: a “static” method computing strain and stress fields from dislocation configurations in the absence of external loading, and a “quasi-static” method to estimate β from the curvature of dislocation lines under applied load. First, the static method is combined with finite element analysis to investigate a recent assertion that heterogeneous initial strain fields can induce higher harmonic generation in a linear elastic medium; the present results do not corroborate this outcome. Then, the quasi-static method is applied to multiple-dislocation scenarios through parametric studies, revealing behaviors not predicted by analytical models, such as the competing interactions of edge and screw dislocations and the significant influence of applied stress on β. Finally, the simulations are used to model SHG experimental results and validate the hypothesis that β can decrease during plastic deformation, despite increasing dislocation density. As the DD code used here is open-source, it provides a practical platform for future investigation into microstructure–β relationships important to the interpretation of SHG results.

This study provides the theoretical foundation, numerical simulation and experimental verification of multi-frequency nonlinear Rayleigh wave mixing aimed at imaging the damage gradient in highly scattering and dissipative media. We focus on the mixing of two counter-propagating finite-amplitude Rayleigh waves generating a secondary higher frequency longitudinal wave detected at its source, directly above the interaction zone. This greatly mitigates the influence of the medium’s heterogeneity on the measured acoustic nonlinearity compared to the more classical self interaction experiments, where the measured higher harmonics are greatly affected by the scattering and attenuation along the propagation path. The Rayleigh waves’ excitation frequency defines the interaction zone within the bulk. By increasing or decreasing the frequency, the classical elastic nonlinearity of shallower or deeper regions are probed. The method is first investigated numerically before being applied experimentally to estimate depth-varying thermal damage-induced changes in acoustic nonlinearity in concrete and mortar blocks. The findings indicate the utility of nonlinear Rayleigh wave mixing for imaging the depth-dependent nonlinear elastic properties of highly heterogeneous materials.

Tight cracks and microcracks often go undetected by linear ultrasonic testing, motivating nonlinear ultrasonic techniques, such as nonlinear coda wave interferometry (NCWI). NCWI combines high-frequency coda wave interferometry with high-amplitude low-frequency pumping, gradually increasing the pump amplitude, to probe changes in wave velocity due to crack shearing/opening. The question is whether dynamic in-service loading could be used as pump to increase NCWI's utility for in situ monitoring. Since high-amplitude pumping induces transient time-dependent changes in the material, the pumping duration and history influence the material's response. This study presents a systematic investigation of the pumping protocol's influence on NCWI results and discussions of the possible underlying mechanisms. We conduct NCWI with varying duration and pump amplitude orders on two concrete blocks—one intact and one containing a closed crack—under controlled temperature and humidity. Our findings show that longer pump durations cause larger relative velocity changes, likely due to progressive crack sliding/opening, while descending amplitudes yield smaller changes as the material may not fully recover from prior higher-amplitude pumps. Nevertheless, relative velocity changes measure consistently higher for the damaged block compared to the intact one. This work highlights the time-dependent material response and the importance of consistent pumping protocols when deploying NCWI.