Reservoir Scale Research Articles

Geologic CO2 sequestration and permeability uncertainty in a highly heterogeneous reservoir

To understand the implications of permeability uncertainty in basalt-hosted CCS reservoirs, this study investigates the feasibility of industrial-scale CCS operations within the Columbia River Basalt Group (CRBG). It is generally accepted that plausible constraints on in situ fracture-controlled permeability distributions are unknowable at reservoir scale. In order to quantify the effects of this permeability uncertainty, stochastically generated and spatially correlated permeability distributions are used to create 50 synthetic reservoir domains to simulate constant pressure CO2 injections. Results from this research illustrate that permeability uncertainty at the reservoir-scale significantly impacts both the accumulation and distribution of CO2. After 20 years of injection the total volume of CO2 injected in each simulation ranges from 2.4 MMT to 40.0 MMT. Interestingly, e-type calculations show that the mean CO2 saturation over the ensemble of 50 simulations is concentric around the injection well, however, ensemble variance shows an ellipse of uncertainty that trends parallel to the long axis of CRBG permeability correlation (N40°E). These results indicate that a priori knowledge of permeability correlation structure is an important operational parameter for the design of monitoring, measuring, and verification strategies in highly heterogeneous CCS reservoirs.

Reducing Uncertainty in Seismic Interpretation using Crosswell Seismic

Crosswell seismic is an emerging technology which provides highly detailed images of the subsurface at reservoir scale. The technology has the potential to delineate complex structure, help verify existing interpretation and also can be used to monitor changes in the reservoir caused by enhanced recovery strategies. The technology employs tomographic surveying, whereby the transmitter and receiver are deployed in separate wells. Interwell structure and velocity can be obtained from reflection processing and direct wave processing respectively. In this paper an example of crosswell seismic is discussed where the technology produced high-resolution images of the geology between wells at 8 to 10 times the resolution of surface 3-D seismic data and revealed numerous faulted compartments in the reservoir previously unseen in the surface seismic. These results have lead to a new understanding of the field and improved the well design process.

Modeling the wave propagation in viscoacoustic media: An efficient spectral approach in time and space domain

We present an efficient and accurate modeling approach for wave propagation in anelastic media, based on a fractional spatial differential operator. The problem is solved with the Fourier pseudo-spectral method in the spatial domain and the REM (rapid expansion method) in the time domain, which, unlike the finite-difference and pseudo-spectral methods, offers spectral accuracy. To show the accuracy of the scheme, an analytical solution in a homogeneous anelastic medium is computed and compared with the numerical solution. We present an example of wave propagation at a reservoir scale and show the efficiency of the algorithm against the conventional finite-difference scheme. The new method, being spectral in the time and space simultaneously, offers a highly accurate and efficient solution for wave propagation in attenuating media.

Upscaling of CO2 injection in a fractured oil reservoir

The purpose of this study is to present the numerical evaluation of the upscaling of secondary and tertiary CO2 flooding at reservoir pressure and temperature. In all simulation cases, CO2 is injected from the top of the fracture in a fracture-matrix block system and displaces oil toward the bottom of the block. Our numerical models provide comprehensive sensitivity analyses that are proved to be important for the upscaling of CO2 injection from lab to reservoir scale. In addition, we address the scale dependency of the diffusion mechanism in a large matrix-fracture domain.In our earlier studies, we have conducted a significant number of reservoir-condition CO2 flood experiments by employing fractured core systems and North Sea reservoir oil. Our experimental and numerical studies are proved to be useful to investigate the complex physical phenomena affecting the efficiency of CO2 injection in oil recovery from a fracture reservoir. The height of the core plugs used in our experiments varies from 7 to 28 cm while the fracture spacing varies between ∼4 cm and 12 cm. In all cases, we find the important role of diffusion mechanism in increasing oil recovery in a fracture-matrix system at the lab scale. However, the field matrix block size is often 1-to-2 orders of magnitude larger than that in the lab scale. Recovery mechanisms are significantly affected by the matrix dimension. For instance, the impact of diffusion mechanism on oil recovery may alter with the size of the matrix block and the fracture spacing. This highlights the importance of upscaling from lab to the field scale that is addressed in this study.We employ a compositional reservoir simulation with a developed equation of state (EOS) to model CO2 injection in a fracture-matrix system at reservoir conditions. In all the numerical simulations, the single-porosity (SP) model consists of a single axial fracture with a typical fracture aperture of 1 mm. The SP model also contains a single matrix block that feeds the producer through fracture next to it. The dimension of the fracture-matrix block varies in the SP models to account for the scale dependency of the results. In our modeling framework, we assume a network of open fractures that are homogenously distributed in the reservoir and links the oil in the matrix to the production wells. We initialize the matrix with the North-Sea-Chalk-Field (NSCF) live oil and the fracture with the injected gas (CO2). The whole displacement process is operated at the constant reservoir conditions of 3750 psia and 110 °C, representing typical NSCF reservoir conditions. Prior to conducting simulations, we successfully tune EOS parameters against the extensive PVT data, sampled from a specific fractured NSCF reservoir. The tuned EOS model assures the proper estimation of the complex phase and volumetric properties for CO2 and oil mixtures at reservoir conditions. In addition, the multi-component diffusion coefficients that are tuned using experimental data are employed as input parameters for the compositional simulations.We study the effect of several key parameters on oil recovery; e.g. matrix block size, fracture spacing, CO2 injection rate, density difference, vaporization and the diffusion. The results show that the mass transport is mainly dominated by diffusion in the lab scale which is not the case for the large matrix block size. Finally, we test the accuracy of the dual-porosity (DP) model against the SP model at various displacement conditions. Results show that when diffusion and vaporization are significant the DP model fails to predict the results of the SP reservoir model.Our findings are an important step towards modeling the tertiary-CO2 flooding in an actual fracture-chalk system. We also provide some important inputs that are necessary for upscaling tertiary-CF from a lab-scale into a field-scale reservoir model.

Novel Observations of Salinity Transport in Low-Salinity Waterflooding

Summary Low-salinity waterflooding (LSWF) is a promising process that could lead to increased oil recovery. To date, the greatest attention has been paid to the complex oil/water/rock chemical reactions that might explain the mechanisms of LSWF, and it is generally accepted that these result in behavior equivalent to changing oil and water mobility. This behavior is modeled using an effective salinity range and weighting function to gradually switch from high- to low-salinity relative permeability curves. There has been limited attention on physical transport of fluids during LSWF, particularly at large scale. We focus on how the salinity profile interacts with water fronts through the effective salinity range and dispersion to alter the transport behavior and change the flow velocities, particularly for the salinity profile. We examined a numerical simulation of LSWF at the reservoir scale. Various representations of the effective salinity range and weighting function were also examined. The dispersion of salinity was compared with a theoretical form of numerical dispersion based on input parameters. We also compared salinity movement with the analytical solution of the conventional dispersion/advection equation. From simulations we observed that salinity is dispersed as analytically predicted, although the advection velocity might be changed. In advection-dominated flow, the salinity profile moves at the speed of the injected water. However, as dispersion increases, the mixing zone falls under the influence of the faster-moving formation water and, thus, speeds up. To predict the salinity profile theoretically, we have modified the advection term of the analytical solution as a function of the formation- and injected-water velocities, Péclet number, and effective salinity range. This important result enables prediction of the salinity transport by this newly derived modification of the analytical solution for 1D flow. We can understand the correction to the flow behavior and quantify it from the model input parameters. At the reservoir scale, we typically simulate flow on coarse grids, which introduces numerical dispersion or must include physical dispersion from underlying heterogeneity. Corrections to the equations can contribute to improving the precision of the coarse-scale models, and, more generally, the suggested form of the correction can also be used to calculate the movement of any solute that transports across an interface between two mobile fluids. We can also better understand the relative behaviors of passive tracers and those that are adsorbed.

Estimating smectite content from seismic data

Predicting the mineralogical composition of shales is crucial for drilling operations related to hydrocarbon exploration/production as well as for the assessment of their sealing capacity as hydrocarbon or CO2 barriers. For example, hydrocarbon exploration in the Northern Carnarvon Basin, North-West Shelf, Australia is hindered by the presence of a thick (up to 1 km) smectite-rich shale seal that spreads regionally. Complex structures of the channelised oil and gas fields in the area make it necessary to drill deviated wells through that seal. The maximum deviation angle at which successful drilling is possible depends strongly on the clay mineralogy and, in particular, on the smectite content in the shale. Here, we introduce a novel workflow combining seismic data, well logs and laboratory measurements to infer shale composition at the reservoir scale. It is applied to the Duyfken 3D seismic survey in the central part of the Northern Carnarvon Basin. Interpretation results are verified against the laboratory X-ray diffraction measurements from the test well that was not used for the interpretation. The results match the test data well within the determined uncertainty bounds.

Determining Conditions of CO2 Entry into Intraformational Baffles: A Prerequisite to Estimate Residually Trapped CO2 at Reservoir Scale

A lithologically heterogeneous reservoir contains low-porosity and -permeability flow barriers known as intraformational baffles, which impede the buoyancy-controlled flow of CO2 rising towards the surface. The contrast in petrophysical and flow properties, such as porosity, permeability and capillary entry pressure between baffles and reservoir rocks forces CO2 to accumulate underneath the baffle, which in turn increases the buoyancy pressure [1] and the CO2 saturation at the baffle-reservoir rock contact. Ultimately, the CO2 saturation at the baffle-reservoir contact builds up to a maximum critical value where the buoyancy pressure becomes equal to the capillary entry pressure and CO2 migrates into the baffle. In a lithologically heterogeneous reservoir, the range in the petrophysical and flow properties of the baffle and reservoir rock determine the degree of maximum CO2 saturation at the time of CO2¬ entry into the baffle. It is important to determine the range in maximum CO2 saturation values as it determines the volume of CO2 trapped underneath the baffle before entry takes place. In this study, lithological heterogeneity has been characterized at sub-meter scale in terms of five rock type classes: Coarse Sandstone, Fine Sandstone, Siltstone, Mudstone and Carbonated Sandstone based on field data obtained from the CO2CRC Otway site. 216 multi-phase transport model realizations have been built to simulate the entry of supercritical CO2 into the baffles using a range of realistic petrophysical and flow properties. The former two rock type classes are implemented as reservoir rocks while the latter three as baffles in the various model realizations. The results from simulations help determining relations between the range in maximum CO2 saturation at the baffle-reservoir rock contact at the time of entry and the range in petrophysical and flow properties for the different combinations of rock type classes. Using the degree of maximum CO2 saturation at the baffle-resevoir contact rather than using CO2 column height allows to account for the variation in CO2 saturation distribution within the column height. This study determines the threshold conditions of CO2 entry into an intraformational baffle and quantifies the maximum volume of residually trapped CO2 below the baffle by accounting for the heterogeneity in petrophysical and flow properties.

Constraining the Fluid History of a CO2‐H2S Reservoir: Insights From Stable Isotopes, REE, and Fluid Inclusion Microthermometry

AbstractReservoirs that host CO2‐H2S‐bearing gases provide a key insight into crustal redox reactions such as thermochemical sulfate reduction (TSR). Despite this, there remains a poor understanding of the extent, duration, and the factors limiting this process on a reservoir scale. Here we show how a combination of petrography, fluid inclusion, rare earth element (REE), and carbon (δ13C), oxygen (δ18O), and sulfur (δ34S) stable isotope data can disentangle the fluid history of the world's largest CO2 accumulation, the LaBarge Field in Wyoming, USA. The carbonate‐hosted LaBarge Field was charged with oil around 80 Ma ago, which together with nodular anhydrite represent the reactants for TSR. The nodules exhibit two distinct trends of evolution in δ13C with both δ34S and δ18O that may be coupled to two different processes. The first trend was interpreted to reflect the coupled dissolution of anhydrite and reduction to elemental sulfur and the oxidation of organic compounds and associated precipitation of calcite during TSR. In contrast, the second trend was interpreted to be the result of the hydrothermal CO2 influx after the cessation of TSR. In addition, mass balance calculations were performed to estimate an approximate TSR reaction duration of 80 ka and to identify the availability of organic compounds as the limiting factor of the TSR process. Such an approach provides a tool for the prediction of TSR occurrence elsewhere and advancing our understanding of crustal fluid interactions.

Characterization of Hysteretic Multiphase Flow from the MM to M Scale in Heterogeneous Rocks

Incorporating mm-m scale capillary pressure heterogeneity into upscaled numerical models is key to the successful prediction of low flow potential plume migration and trapping at the field scale. Under such conditions, the upscaled, equivalent relative permeability incorporating capillary pressure heterogeneity is far from that derived conventionally at the viscous limit, dependent on the heterogeneity structure and flow rate, i.e. dependent on the capillary number. Recent work at the SCA 2017 symposium (SCA2017-022) demonstrated how equivalent functions can be obtained on heterogeneous rock cores from the subsurface under drainage conditions; going beyond traditional SCAL. Experimental observations using medical CT scanning can be combined with numerical modelling so that heterogeneous subsurface rock cores can be directly characterized and used to populate field scale reservoir models. In this work, we extend this characterization approach by incorporating imbibition cycles into the methodology. We use a Bunter sandstone core with several experimental CO2 – Brine core flood datasets at different flow rates (2x drainage, 1x imbibition and 2x trapping) to demonstrate the characterization of hysteretic multiphase flow functions in water-wet rocks. We show that mm-m scale experimental saturations and equivalent, low flow potential relative permeabilities can be predicted during drainage and imbibition, along with trapping characteristics. Equivalent imbibition relative permeabilities appear as a function of capillary number, as in the drainage cases. We also find that the form of capillary pressure function during imbibition has a large impact on the trapping characteristics, with local heterogeneity trapping reduced (or removed), if the capillary pressure drops to zero, or below at the residual saturation.

Geomechanical responses during depressurization of hydrate-bearing sediment formation over a long methane gas production period

The geomechanical behaviour of hydrate-bearing sediments during methane gas production is complex due to the spatial and temporal changes in stress, pore pressure, temperature, and phase change. In order to evaluate the geomechanical risks during methane gas production, it is necessary to understand the thermo-hydro-mechanical (THM) responses in the production region that recovers methane gas from the hydrate-bearing sediments. In this study, a fully coupled THM numerical simulator code was used to examine the reservoir scale field behaviour observed during the gas production trial conducted at the Eastern Nankai Trough, Japan in March, 2013. Using the available field investigation data, history matching of the gas production test was conducted to evaluate the methane gas production process. The fully coupled model allowed examination of the mechanical response using the critical state based constitutive model proposed by Uchida et al. (2012). The model parameters were determined from the results of triaxial compression tests conducted on recovered core samples. Based on the reservoir scale simulation results, this paper investigated the mechanical responses of five selected elements at different locations in the hydrate gas production region. The mechanical responses of hydrate-bearing sediments at specific locations within the production region are related to their hydrate dissociation status, which typically can be divided into before, during and after stages of hydrate dissociation. The 260 days gas production simulation suggests an increase in effective stresses accompanied by shearing deformation, which makes the soil more plastic. Potential geomechanical risks (such as wellbore stability and formation compaction) associated with the observed changes in stress/strain were also identified.

Solute transport in heterogeneous reservoirs: Upscaling from the Darcy to the reservoir scale

In this work, we present a methodology for upscaling solute transport in heterogeneous reservoirs. The problem considers upscaling from the Darcy scale of a porous medium to the reservoir scale. We first define the mathematical problem that governs solute mass transport at the so-called Darcy scale, and then we apply the method of volume averaging. This method allows for upscaling the governing equations at the whole reservoir scale and, eventually, developing physical constraints, definitions of average variables and effective coefficients. One important characteristic of the volume averaging method is that the so-called closure problems for analytical or numerical computation of effective coefficients are identified. The closure problems consist of boundary-value problems, whose solution must be carried out in representative geometries (unit cells) of the reservoir. In this paper, we developed the theory for a two-facies reservoir, but the extension for more facies is straightforward. In this context, we present the two-equation model, and the numerical estimation of effective coefficients are presented and discussed for lenticular and layered arrangements of facies. In this manner, relevant coefficients are identified. The numerical solution of the upscaled model agrees well with direct numerical results when the governing equations are solved at the Darcy scale.

Insights into the fractional flow of low salinity water flooding in the light of solute dispersion and effective salinity interactions

The objective of Low Salinity Water Flooding (LSWF) is to improve oil recovery. While a significant number of laboratory tests have been carried out to investigate the impact of LSWF, field scale modelling is often reported in the form of sector models with relatively coarse cells. This paper assesses the impact of simulating flow at very fine scales and informs on the properties that should be captured at the coarse scale to avoid numerical errors. We have found that the weighting function that is used to control changes to fluid mobility combines with numerical and physical diffusion to induce a retardation/acceleration effect. This is a physical effect rather than part of a chemical reaction.In this study, numerous simulations of LSWF have been carried out at the reservoir scale to investigate flow behaviour for various salt concentration (salinity) weighting functions and dispersion coefficients. We have examined the effective salinity range over which the weighting function is applied as well as considering various shapes. Dispersion was varied to represent physical and numerical effects. These have been compared to analytical solutions from fractional flow theory.We also observed that the fractional flow of the oil bank will be same for both the secondary and tertiary flooding. We point out the relative importance of various parts of the relative permeability curves. An important finding of this work is that by spreading the salinity front through dispersion and setting a low value at which salinity impacts mobility, we saw the injected low salinity front advance more slowly while the high salinity front of formation water moves more quickly. This is an effective retardation effect. We related this to an effect equivalent to adsorption in the fractional flow theory and could measure it in a similar way. We were also able to develop a prediction of the effect using the analytical solution to the advection-diffusion equation.The outcome is that we can estimate a corrective term for the flow behaviour in situations where the dispersion is quite strong, particularly in numerical simulations. We consider that this enable corrections to be made for numerical dispersion effects in field scale models.

Foam Generation and Rheology in a Variety of Model Fractures

Gas is used in petroleum reservoirs to displace oil for enhanced oil recovery. The microscopic displacement efficiency of gas is very good, but at the reservoir scale the process suffers from poor sweep efficiency, especially in naturally fractured reservoirs. Foam can improve the sweep. There have been considerable scientific contributions toward understanding foam flow in nonfractured porous media, with relatively little work on foam flow in fractured porous media. We investigate foam-generation mechanisms in five fully characterized glass models of fractures with different apertures and correlation lengths of the aperture distribution. We also study the rheology of the in situ-generated foam by varying the superficial velocities of the gas and surfactant solution. We compare the measured pressure gradient against the fracture attributes, aperture, and the correlation length of the aperture. We also compare foam texture as a function of position within the fracture as the generated foam propagates through the fracture. Gas mobility was greatly reduced as a result of in situ foam generation in our model fractures. Foam was generated predominantly by capillary snap-off and lamella division. The measured mobility reduction depends on fracture attributes. Fracture-wall roughness, represented by both the hydraulic aperture and the correlation length of the aperture, plays an important role in foam generation and mobility. The average bubble size increases as the aperture increases, which results in a significant decrease in pressure gradient. Two model fractures show the same two foam-flow regimes central to the understanding of foam in nonfractured porous media: a low-quality regime where pressure gradient is independent of liquid velocity and a high-quality regime where pressure gradient is independent of gas velocity. The mechanisms thought to be behind these two regimes in nonfractured porous media do not apply to these experiments, however.

Hydrogeochemical Modeling to Identify Potential Risks of Underground Hydrogen Storage in Depleted Gas Fields

Underground hydrogen storage is a potential way to balance seasonal fluctuations in energy production from renewable energies. The risks of hydrogen storage in depleted gas fields include the conversion of hydrogen to CH4(g) and H2S(g) due to microbial activity, gas–water–rock interactions in the reservoir and cap rock, which are connected with porosity changes, and the loss of aqueous hydrogen by diffusion through the cap rock brine. These risks lead to loss of hydrogen and thus to a loss of energy. A hydrogeochemical modeling approach is developed to analyze these risks and to understand the basic hydrogeochemical mechanisms of hydrogen storage over storage times at the reservoir scale. The one-dimensional diffusive mass transport model is based on equilibrium reactions for gas–water–rock interactions and kinetic reactions for sulfate reduction and methanogenesis. The modeling code is PHREEQC (pH-REdox-EQuilibrium written in the C programming language). The parameters that influence the hydrogen loss are identified. Crucial parameters are the amount of available electron acceptors, the storage time, and the kinetic rate constants. Hydrogen storage causes a slight decrease in porosity of the reservoir rock. Loss of aqueous hydrogen by diffusion is minimal. A wide range of conditions for optimized hydrogen storage in depleted gas fields is identified.

An Integrated Analysis of the Eutrophication Process in the Enxoé Reservoir within the DPSIR Framework

The Enxoé reservoir in southern Portugal has been exhibiting the highest trophic state in the country since its early years of operation. The problem has attracted water managers’ and researchers’ attention as the reservoir is the water supply for two municipalities. Extensive research was thus conducted over the last few years, including field monitoring and modelling at the plot, catchment, and reservoir scales. This study now frames all partial findings within the Driver-Pressure-State-Impact-Response (DPSIR) framework to better understand the eutrophication process in the Enxoé reservoir. Agriculture and grazing were found to have a reduced role in the eutrophication of the reservoir, with annual sediment and nutrient loads being comparably smaller or similar to those reported for other Mediterranean catchments. Flash floods were the main mechanism for transporting particle elements to the reservoir, being in some cases able to carry up three times the average annual load. However, the main eutrophication mechanisms in the reservoir were P release from deposited sediment under anoxic conditions and the process of internal recycling of organic matter and nutrients. Reducing the P load from the catchment and deposited sediment could lead to a mesotrophic state level in the reservoir. However, this level would only be sustainable by limiting the P internal load ability to reach the photic zone.

Oceanic methane hydrate utilization system design and reservoir scale numerical modeling

In recent years, the exploration and utilization of methane hydrate from permafrost and off-coast regions has become a very hot topic. Aiming at its potential as a vast energy resources in future use, the current study is focused on the low emission extraction and utilization system for oceanic methane hydrate and its economic and technological feasibility. In this study, recent real extraction tests in both China and Japan in 2017 are firstly discussed and compared, where the challenges in production period and production rate are explained. In those tests, depressurization method is used to dissociate the methane hydrate for gas production form the unconsolidated reservoir layers below the seabed. Based on the conditions of Japan, a systematic utilization system is proposed for oceanic methane hydratewhich is consisted of extraction system, power generation system, Carbon Capture and Sequestration (CCS) system and hot fluid injection system. Vertical well systems are adopted in this design and the strategic warming up process is used before depressurization extraction, so as to obtain optimal production rate of methane gas. The inclusion of CCS makes it possible to utilize methane hydrate with very low carbon emission. It is estimated that the electricity price obtained by this system can be comparable with the current civil electric price in Japan if the production rate is around 1.0×105 m3/d. Also in this study, the real extraction scenarios and possible strategies of Nankai Trough in Japan and Shenhu Area in China are re-visited and discussed by numerical models. It is found that the only depressurization or thermal stimulation method have the problems of wellbore deformation, pressure problem and energy efficiency problems, while combined strategic modes show the possibility of long-term continuous production. For the basic dissociation parameter analysis, the conditions near the equilibrium curve show higher production rate, which is due to higher initial reservoir temperature. Such results indicate the importance of heat supply inside the reservoir layer as the methane hydrate dissociation is endothermic. It is shown that with the increase of reservoir pressure, such temperature effects become more critical as the optimal initial temperature increases according to equilibrium curve. This results support the system design with strategic warming-up process in a sister well during methane hydrate extraction. In addition, the effects of reservoir layer thermal management, hot reservoir fluid flow effects on the dissociation behaviors are analyzed in this study. For vertical well analysis, the production rates under different permeability parameters show the same trend with real production, which in turn help confirm the in-lab permeability tests. Horizontal well simulation show the importance of overburden and under burden. If the confining layers are less permeable, the production rate becomes higher due to the limited intrusive water flow from the confining layers. In sum, this study summarized the advantages and disadvantages of oceanic methane hydrate extraction/utilization methods and put forward a low-carbon emission system. It is hoped that this study will contribute to the future development of both production prediction and real extraction tests.

Stress-strain response and seismic signature analysis of phyllite reservoir rocks from Blue Mountain geothermal field

Geothermal energy is the heat contained in rock and fluid in the earth’s crust. In some areas the geothermal resource can be exploited by drilling wells to extract the heated fluids and piping the fluids to a power plant to generate renewable electricity. Geomechanical characterization of the rocks in geothermal reservoirs is necessary to (i) optimize a well stimulation program, (ii) reduce production decline, and (iii) build more accurate predictive models at reservoir scale. In this study, we present the results of a suite of experiments on mechanical characterization of phyllite rock specimens retrieved from DB-2 well (depth of 1260 m) at the Blue Mountain geothermal field in Nevada, United States. In order to investigate the effects of anisotropy on mechanical response of rock, one vertically-drilled and one horizontally-drilled sub-cores were extracted from the phyllite core. Multi-stage elastic, cyclic, creep, and multi-stage failure tests were performed to characterize the elastic, plastic, time-dependent, and failure properties as well as effects of anisotropy on the behavior of the phyllite specimens. The pressure-dependent static and dynamic moduli revealed that within the elastic region, the Young’s modulus increased with the increment of differential stress during the loading stage. While, 10% to 50% difference between the elastic moduli of the two specimens confirmed the anisotropic nature of the material. Negligible creep response was observed in both specimens, with a more pronounced change in P- and S-waves velocities during hydrostatic stage for vertically-drilled sub-core. The failure envelope, developed using the multi-stage failure test on the vertically-drilled specimen, indicated good agreement with both Linearized Mohr-Coulomb and Hoek-Brown failure criteria. On the other hand, significant changes were observed in the time-frequency maps of the transmitted seismic waves during multi-stage elastic test (i.e. change in confining level), however, slight changes in the time-frequency map were observed during creep test (i.e. application of differential stress for a relatively long period).

Permeability changes in coal seams: The role of anisotropy

The permeability of coal seams is pore pressure-dependent. A number of analytical models have been proposed to investigate this problem, but many disregard a crucial factor: the anisotropy of coal. This paper is devoted to investigating the role of anisotropy in modeling the change of horizontal permeability with pressure. Analysis is conducted using a fully anisotropic model that incorporates both the anisotropies of mechanical properties and of the permeability dependence on stresses. Analytical expressions of the pressure-permeability relationship are derived in oedometric and isochoric geomechanical conditions, and validations are conducted against both laboratory and field data. Then, the roles of the anisotropy of stiffness and of permeability dependence on stresses in the permeability-change model are explored. We demonstrate that the mechanical anisotropy can be simplified to an isotropic model without introducing significant errors in prediction of pressure-permeability relationship, while neglecting the anisotropy of permeability dependence on stresses leads to considerable errors. When both anisotropy sources are disregarded, the pressure-permeability curve can be exactly reproduced by a totally isotropic material in both oedometric and isotropic conditions. However, the material properties (e.g., bulk modulus) are skewed; moreover, such an equivalent material might lead to significant errors in other geomechanical conditions. Finally, the permeability change is investigated at the reservoir scale, and the reservoir simulation results confirm the conclusions obtained from the analytical analysis.

Quantitative evaluation of the impacts of drilling mud on the damage degree to the permeability of fractures at different scales in coal reservoirs

The current study aimed to investigate the degree of pollution from drilling mud on different scales of fractures in coal reservoirs. Drilling mud for field utilization was prepared in the laboratory and different scales of fractures were generated in coal and rock samples before and after being polluted by the drilling fluid and observed using scanning electron (SEM) and an optical microscope. Using Monte Carlo modelling, the networks of the millimeter (mm), micron- and nano-meter sized fractures, before and after pollution were separately determined. After programming, seepage models of mm, micron- and nano-meter sized fractures were separately constructed using Matlab software. The degrees of pollution of the drilling fluid on the different scales of fractures were calculated according to the seepage principle and compared with the measured permeability of the polluted boreholes. The micron- and nano-sized fractures were more sensitive to drilling fluids compared to the mm-sized and larger fractures. When the original permeability of the coal reservoir was low, different scales of fractures were all greatly polluted by drilling fluids, macroscopically showing high damage to permeability of the coal reservoir. However, when the original permeability of coal reservoir was large, the pollution degrees of micron, and nano-sized fractures failed to be macroscopically observed, having insignificant influences on the permeability of the coal reservoir. The results showed that the combination of the Monte Carlo method and the Matlab numerical simulations can more distinctly reflect the pollution degrees of the drilling fluids on different scales of fracture compared to traditional methods that have been used to test the permeability of coal reservoirs before and after pollution.

Stacked autoencoder with echo-state regression for tourism demand forecasting using search query data

Accurate tourism demand forecasting is fundamental in the tourism industry, while effective tourism demand forecasting using search query data (SQD) has become popular in the tourism management field. SQD is a type of statistical time series provided by search engines that can reflect netizens’ attention on certain events. Scholars attempt to establish a reasonable relationship between tourism demand and SQD for its timeliness and comprehensiveness. The current study proposes an effective deep learning technique called stacked autoencoder with echo-state regression (SAEN) to accurately forecast tourist flow based on search query data. In the proposed SAEN approach, stacked autoencoder is adopted to hierarchically learn high-level predictive indicators from substantial SQD and connected by an echo-state regression layer to model the nonlinear time series relationship between tourism flow and the learned indicators. Four realistic applications (i.e., one comparative case and three extended cases in the US and China with different SQD sources) are used to verify the forecasting performance of SAEN. Numerical results indicate that SAEN is better than the current literature findings, including time series approach, econometric model, common machine learning algorithms, and state-of-the-art deep learning techniques. The structure parameters of SAEN are further analyzed empirically and theoretically. Moreover, this study determined a different impact of network depth and echo-state reservoir scale on the performance of SAEN. The proposed SAEN can be an appropriate alternative for tourism demand forecasting in complex data situations.