DeepSense: A Physics-Guided Deep Learning Paradigm for Anomaly Detection in Soil Gas Data at Geologic CO2 Storage Sites.

Driven by the collection of enormous amounts of streaming data from sensors, and with the emergence of the internet of things, the need for developing robust detection techniques to identify data anomalies has increased recently. The algorithms for anomaly detection are required to be selected based on the type of data. In this study, we propose a predictive anomaly detection technique, DeepSense, which is applied to soil gas concentration data acquired from sensors being used for environmental characterization at a prospective CO2 storage site in Queensland, Australia. DeepSense takes advantage of deep-learning algorithms as its predictor module and uses a process-based soil gas method as the basis of its anomaly detector module. The proposed predictor framework leverages the power of convolutional neural network algorithms for feature extraction and simultaneously captures the long-term temporal dependency through long short-term memory algorithms. The proposed process-based anomaly detection method is a cost-effective alternative to the conventional concentration-based soil gas methodologies which rely on long-term baseline surveys for defining the threshold level. The results indicate that the proposed framework performs well in diagnosing anomalous data in soil gas concentration data streams. The robustness and efficacy of the DeepSense were verified against data sets acquired from different monitoring stations of the storage site.

Effects of salinity and shear stress on clay deformation: A molecular dynamics study.

The deformation of clay minerals is an important phenomenon that is relevant to many problems, particularly those that occur in subsurface geological formations. The salinity of the formations and external shear stress applied to them are two important factors that contribute to the deformation of such porous media. To gain a deeper understanding of such phenomena, we have carried out extensive molecular dynamics simulations using the Na-montmorillonite (Na-MMT) structure as the model of clay minerals and have studied the effect of salt concentration on its swelling. As the NaCl concentration increases, so also does the basal spacing. We demonstrate the effect of the coupling between the applied shear stress and NaCl salinity on the swelling behavior of Na-MMT, namely, deformation of the interlayer space that results in swelling. According to the results, the extent of Na-MMT deformation depends on both the brine salinity and the shear rate.

Nanoconfinement Effect on Surface Tension: Perspectives from Molecular Potential Theory.

Liquid-vapor surface tension (ST) in nanopores attracts great attention in many industries because of the prosperity of nanoscience and nanotechnology. Here, considering the important emerging new physical phenomena induced by nanoconfinement effects, including curvature-dependent and shift-critical temperature (Tc)-dependent effects, the anomalous variation of ST in nanopores is captured from the molecular potential perspective. Furthermore, a simple analytical model is proposed to determine the ST in nanopores by correlating these two effects with an easily accessible parameter, that is, normalized pore dimension, which is the ratio of the pore radius to Lennard-Jones size parameter. The model is validated to be reliable for determining the STs of different substances both in the bulk phase as well as nanopores through comparison with the experimental results and molecular simulations. Our results show that the reduction of ST induced by the nanoconfinement effects is visible when the pore diameter is within tens of nanometers, and the reduction is more sensitive as the pore size decreases. In detail, the curvature-dependent effect is remarkable in the pores with diameters ranging from a few nanometers to tens of nanometers. Moreover, a simply generalized formula is obtained to determine the curvature-dependent effect and the Tolman length for different substances. The shift-Tc-dependent effect is not only related to the pore dimension but also depends on the temperature. As the pore size decreases, the critical temperature of confined fluids diverges significantly from the bulk values. While at high temperatures, the range of pore size impacted by the shift-Tc-dependent effect is enlarged. Additionally, the nanoconfined STs of different substances are calculated and compared. Overall, the new model captures the underlying physics behind the variation of STs in nanopores and can determine the nanoconfined STs reasonably. Moreover, the simple formulation of the model is beneficial to the practical applications in many chemical engineering processes, such as chemical separation, nucleation phenomenon, and capillary condensation.

Effects of grain size and small-scale bedform architecture on CO2 saturation from buoyancy-driven flow.

Small-scale (mm-dm scale) heterogeneity has been shown to significantly impact CO2 migration and trapping. To investigate how and why different aspects of small-scale heterogeneity affect the amount of capillary trapping during buoyancy-driven upward migration of CO2, we conducted modified invasion percolation simulations on heterogeneous domains. Realistic simulation domains are constructed by varying two important aspects of small-scale geologic heterogeneity: sedimentary bedform architecture and grain size contrast between the matrix and the laminae facies. Buoyancy-driven flow simulation runs cover 59 bedform architecture and 40 grain size contrast cases. Simulation results show that the domain effective CO2 saturation is strongly affected by both grain size and bedform architecture. At high grain size contrasts, bedforms with continuous ripple lamination at the cm scale tend to retain higher CO2 saturation than bedforms with discontinuous or cross lamination. In addition, the "extremely well sorted" grain sorting cases tend to have lower CO2 saturation than expected for cross-laminated domains. Finally, both a denser CO2 phase and greater interfacial tension increase CO2 saturation. Again, variation in fluid properties seems to have a greater effect on CO2 saturation for cross-laminated domains. This result suggests that differences in bedform architecture can impact how CO2 saturation values respond to other variables such as grain sorting and fluid properties.

Assessing the potential of composite confining systems for secure and long-term CO2 retention in geosequestration.

A potential geologic target for CO2 storage should ensure secure containment of injected CO2. Traditionally, this objective has been achieved by targeting reservoirs with overlying seals-regionally extensive, low permeability units that have been proven capable of retaining buoyant fluid accumulations over geologic time. However, considering that the amount of CO2 is limited by a decadal injection period, vertical migration of CO2 can be effectively halted by a composite system of discontinuous shale/silt/mudstone barriers in bedded sedimentary rocks. Here, we studied the impact of depositional architectures in a composite confining system on CO2 migration and confinement at reservoir scale. We stochastically generated lithologically heterogeneous reservoir models containing discontinuous barriers consistent with statistical distributions of net-sand-to-gross-shale ratio (NTG) and horizontal correlation lengths derived from well log data and observations of producing hydrocarbon fields in Southern Louisiana. We then performed an extensive suite of reservoir simulations of CO2 injection and post-injection to evaluate the sensitivity of CO2 plume migration and pressure response of the composite system to a series of geologic and fluid parameters including the lateral continuity of barriers, NTG, permeability anisotropy within the sand body, and capillary pressure contrast between the sand and shale facies. The results indicate that lateral continuity of barriers and NTG are the dominant controls on CO2 plume geometry and pressure build-up in the reservoir, while the impact of NTG is particularly pronounced. The significance of intraformational barriers becomes apparent as they facilitate the local capillary trapping of CO2. Those barriers improve the pore space occupancy by promoting a more dispersed shape of the plume and ultimately retard the buoyancy-driven upward migration of the plume post injection.

Simultaneous decolorization and bioelectricity generation in a dual chamber microbial fuel cell using electropolymerized-enzymatic cathode.

Effect of cathodic enzymatic decolorization of reactive blue 221 (RB221) on the performance of a dual-chamber microbial fuel cell (MFC) was investigated. Immobilized laccase on the surface of a modified graphite electrode was used in the cathode compartment in order to decolorize the azo dye and enhance the oxygen reduction reaction. First, methylene blue which is an electroactive polymer was electropolymerized on the surface of a graphite bar to prepare the modified electrode. Utilization of the modified electrode with no enzyme in the MFC increased the power density up to 57% due to the reduction of internal resistance from 1000 to 750 Ω. Using the electropolymerized-enzymatic cathode resulted in 65% improvement of the power density and a decolorization efficiency of 74%. Laccase could act as a biocatalyst for oxygen reduction reaction along with catalyzing RB221 decolorization. Treatment of RB221 with immobilized laccase reduced its toxicity up to 5.2%. Degradation products of RB221 were identified using GC-MS, and the decomposition pathway was proposed. A discussion was also provided as to the mechanism of dye decolorization on the enhancement of the MFC performance.

Pore-scale characteristics of multiphase flow in heterogeneous porous media using the lattice Boltzmann method.

This study provides a pore-scale investigation of two-phase flow dynamics during primary drainage in a realistic heterogeneous rock sample. Using the lattice Boltzmann (LB) method, a series of three-dimensional (3D) immiscible displacement simulations are conducted and three typical flow patterns are identified and mapped on the capillary number (Ca)-viscosity ratio(M) phase diagram. We then investigate the effect of the viscosity ratio and capillary number on fluid saturation patterns and displacement stability in Tuscaloosa sandstone, which is taken from the Cranfield site. The dependence of the evolution of saturation, location of the displacement front, 3D displacement patterns and length of the center of mass of the invading fluid on the viscosity ratio and capillary number have been delineated. To gain a quantitative insight into the characteristics of the invasion morphology in 3D porous media, the fractal dimension Df of the non-wetting phase displacement patterns during drainage has been computed for various viscosity ratios and capillary numbers. The logarithmic dependence of Df on invading phase saturation appears to be the same for various capillary numbers and viscosity ratios and follows a universal relation.

Image-based modeling of gas adsorption and deformation in porous media.

Understanding adsorption of CO2 in porous formations is crucial to its sequestration in geological formations. We describe a model for adsorption of CO2 and the deformation that it induces in a sandstone formation over wide ranges of temperature and pressure. The model couples the thermodynamics of sorption with elastic deformation of the solid. Finite-element computations are then used in order to compute CO2 adsorption isotherms along with the induced strain in the formation. We also compute the Darcy permeability of the porous medium using the lattice-Boltzmann method. All the computations are carried out with a three-dimensional image of a core sample from Mt. Simon sandstone, the target porous formation for a pilot CO2 sequestration project that is currently being carried out by Illinois State Geological Survey. Thus, no assumptions are made regarding the shape and sizes of the pore throats and pore bodies. The computed CO2 sorption isotherm at 195 K is in excellent agreement with our experimental data. The computed permeability is also in good agreement with the measurement. As a further test we also compute the sorption isotherm of N2 in the same formation at 77.3 K, and show that it is also in good agreement with our experimental data. The model is capable of predicting adsorption of CO2 (or any other gas for that matter) in porous formations at high pressures and temperatures. Thus, it is used to study the effect of hydrostatic pressure on adsorption and deformation of the porous formation under various conditions. We find that the effect of the confining pressure is more prominent at higher temperatures. Also computed is the depth-dependence of the capacity of the formation for CO2 adsorption, along with the induced volumetric strain.

Computer simulation of the effect of deformation on the morphology and flow properties of porous media.

We report on the results of extensive computer simulation of the effect of deformation on the morphology of a porous medium and its fluid flow properties. The porous medium is represented by packings of spherical particles. Both random and regular as well as dense and nondense packings are used. A quasistatic model based on Hertz's contact theory is used to model the mechanical deformation of the packings, while the evolution of the permeability with the deformation is computed by the lattice-Boltzmann approach. The evolution of the pore-size and pore-length distributions, the porosity, the particles' contacts, the permeability, and the distribution of the stresses that the fluid exerts in the pore space are all studied in detail. The distribution of the pores' lengths, the porosity, and the particles' connectivity change strongly with the application of an external strain to the porous media, whereas the pore-size distribution is not affected as strongly. The permeability of the porous media strongly reduces even when the applied strain is small. When the permeabilities and porosities of the random packings are normalized with respect to their predeformation values, they all collapse onto a single curve, independent of the particle-size distribution. The porosity reduces as a power law with the external strain. The fluid stresses in the pore space follow roughly a log-normal distribution, both before and after deformation.

Bioelectricity generation enhancement in a dual chamber microbial fuel cell under cathodic enzyme catalyzed dye decolorization.

Enzymatic decolorization of reactive blue 221 (RB221) using laccase was investigated in a dual-chamber microbial fuel cell (MFC). Suspended laccase was used in the cathode chamber in the absence of any mediators in order to decolorize RB221 and also improve oxygen reduction reaction in the cathode. Molasses was utilized as low cost and high strength energy source in the anode chamber. The capability of MFC for simultaneous molasses and dye removal was investigated. A decolorization efficiency of 87% was achieved in the cathode chamber and 84% COD removal for molasses was observed in the anode chamber. Laccase could catalyze the removal of RB221 and had positive effect on MFC performance as well. Maximum power density increased about 30% when enzymatic decolorization was performed in the cathode chamber.