From lab to field (1)

Rock salt formations are important candidates for radioactive waste disposal. The disposal concept of rock salt is based on the creation of galleries and boreholes to store the radioactive waste. These will subsequently be backfilled with crushed rock salt, which will converge and compact over time. To ensure safe, long-term storage, eventually this backfill should attain porosities and permeabilities comparable to that of dense rock salt, hence sealing in the waste. However, to accurately predict the timescale at which this sealing will occur physics-based descriptions are needed for modelling the long-term behavior of the backfill.
An empirical description of the compaction creep based on experiments on single grain sizes is already available. However, the backfill in a repository will have a distributed grain size, and the individual contribution of each grain size fraction to the overall compaction creep rate is poorly constrained. In addition, a full mechanistic-based model for compaction of backfill down to a few percent porosity is lacking. Hence, extrapolation to longer timescales is uncertain.  
Therefore, we have performed experiments on single and distributed grain size fractions and developed a mechanistic-based model that can predict the compaction creep rates of granular rock salt as a function of its grain size distribution. Depending on whether the material is assumed to be subjected to homogeneous stress or strain-rate (in a manner analogue to the Reuss and Voight bounds for elastic deformation), the compaction creep rate can vary up to a few orders of magnitude.
In addition to the effect of grain size, backfill compaction will also be influenced by humidity. To date, most experiments on crushed backfill are performed under either fully or partially brine saturated (i.e. (semi-)flooded) conditions or lab-air conditions. These experiments show that the presence of brine accelerates compaction creep. Therefore, many proposals consider adding brine to the backfill to accelerate creep. However, when no brine is added it is expected that the presence of humid air still accelerates creep with respect to truly dry conditions. The hygroscopic properties of rock salt make it very likely that in-situ backfill will adsorb water from the surroundings until an equilibrium (relative) humidity of around 75% is reached. At present, no experimental data is obtained under humid conditions. Therefore, we are currently performing creep experiments under conditions with a constant humidity to assess the effect of humidity on compaction creep rates. 

Natural and artificial phenomenon such as dike formation, vein growth and mineralization, hydrocarbon extraction through hydraulic fracturing, gas outbursts in underground mines can be simulated in the laboratory by performing fluid-driven tensile fracturing experiments. For safety of the engineered events of tensile fracturing, it is necessary to monitor the deformation from its offset and visualization of the localization patterns to demarcate sub surface damage zone.
In this study, anisotropic Vosges sandstone blocks retrieved from a quarry near the city of Hangviller was utilized to perform fluid-driven tensile fracturing experiments. Prismatic samples of volume 50×50×30 mm3, with a hole of diameter 10 mm and length 30 mm with its axis intersecting the centre of the 50×50 mm2 face was prepared. A white and black random speckled pattern was then applied on the 50×50 mm2 face of the specimen. The high pressure true triaxial apparatus (TTA) used for the deformation experiment with the unique feature of a transparent sapphire glass window in contact with the specimen’s speckled surface allowing to photograph the face of the specimen during deformation, so as to a posteriori reconstruct the surface strain field using Digital Image Correlation (DIC)[1]. DIC computation was performed in the SPAM[2] software by computing the correlation function between many subset (correlation windows) of corresponding images taken at two successive time steps.
Loading experiments were conducted in the specimens with two different orientations; one in which the applied axial load was perpendicular to the bedding plane of the sandstone and another in which the axial load was applied parallel to the bedding plane. In both the situations, isotropic loading was applied with a rate of 0.4 MPa/min to 8 MPa followed by the application of vertical load at a rate of 0.2 MPa/min to 10 MPa. Internal fluid pressurization in the circular cavity was achieved by injecting water at a rate of 0.2 MPa/min until visible fractures occurred.
Macroscopic stress-strain curves were correlated with volumetric and deviatoric component of strain tensor from onset of loading and fluid pressurization to post failure. Compaction and dilatancy regions around and away from the cavity are presented. Monitoring the advancement of strain concentration from early stages to failure, and its influence on the post failure behavior is reported. The study establishes a correlation between the mechanical response with the development and evolution of kinematic structures (full-field and global) during fracturing of porous sandstone with fluid pressure.

Porous reservoir rocks like sandstones have gained utmost important in the last decade as a potential sink for CO2. Most of the targeted reservoirs are depleted oil and gas fields, which has caprocks to ensure the containment of the injected CO2. Injecting CO2 into porous reservoirs increase the pore pressure, which therefore reduces the effective horizontal and vertical stresses. Depending on the pre-injection stress-condition and permeability of the reservoir, utmost care should be taken to define the upper limit of CO2 injection pressure, in order to prevent any permanent damage to the reservoir which can lead to leakage or induced seismicity. Lab-scale experiments provide key insights to the deformation behavior of reservoir rocks under different stress-conditions, which can be upscaled to understand reservoir scale processes. To simulate the stress perturbation caused by CO2 injection operations, we have subjected porous reservoir rocks (coreplugs) collected from different depths of offshore North Sea under cyclic axial loading and unloading with a confining pressure increment from 10-50 MPa between each cycle. The P and S wave velocities along the axial direction of the coreplugs were recorded in every 10 s to assess the change in wave properties during deformation. It was observed that during each loading cycle, wave velocities are highest at the elastic-plastic transition zone, which can be attributed to the compression of pores and closure of microcracks perpendicular to the loading direction. The wave velocities decrease sharply after the onset of plastic deformation, which can be attributed to the formation of microcracks in the coreplug due to increasing load. The static and dynamic Young’s modulus (E) of the coreplugs during each cycle of increasing confinement show linear increase. Plugs with lower porosity shows higher E with steeper increment at higher confining pressure. The correlation between the wave properties and mechanical response of the reservoir rocks under cyclic loading reveal that constant monitoring of wave velocities during CO2 injection can act as an efficient tool for monitoring stress-state of the reservoir, facilitating safer CO2 storage operations.

Rising demand for energy and green energy has led to increasing subsurface activities, such as geothermal energy sites. These increasing human activities in the subsurface have caused substantial induced earthquakes in more densely populated areas, increasing the risks of operating safely. Well-known examples of induced seismicity, due to geothermal sites, are the M5.4 earthquake in Pohang (South Korea) or the M3.4 earthquake in Basel (Switzerland).  
Monitoring and forecasting earthquakes have been a topic of interest for years. Predictions are often made by production scenarios, probabilistic models, or average earthquake size distribution (b-value). Only a few studies focus on predicting fluid-induced seismicity by using seismic monitoring methods. Pore fluid changes play an important role in the reactivation of the fault strength and stability. Variations in pore pressure can cause a drop in the stresses along the fault plane and cause fault instability and movement resulting in induced seismicity.  Monitoring and predicting these stress changes along the fault planes can therefore be essential in forecasting induced seismicity and mitigation, potentially reducing the risks of operating (in denser populated areas). However, monitoring the degree of these changes remains challenging. Most studies using seismic methods to monitor induced seismicity on a field scale or laboratory scale focus on either passive monitoring or active monitoring. This study combines the two methods and shows how they complement each other in monitoring and mitigation of fault reactivation in the laboratory. We have performed pore fluid injection experiments on faulted sandstones to reactivate the faults while monitoring both actively (active seismic) and passively (acoustic emission).
These results show that both acoustic monitoring techniques can be used to detect the different fault reactivation stages: linear strain build-up, early creep (pre-slip), stress drop (main slip), and continuous sliding phase. However, using active monitoring the early creep phase is detected slightly earlier than using passive monitoring. Combining the methods shows that the stress changes along the fault can be detected with more detail in more accuracy. As a result, the combination of passive and active techniques may be useful for monitoring faulted or critically stressed reservoirs that experience pore pressure changes.