The storage and production of natural gas from shale formations are significantly influenced by the presence of an adsorbed phase on the pore walls. The effects of adsorption on storage and production of natural gas are currently quantified by the establishment of an absolute adsorption isotherm, which is fitted to a Langmuir isotherm. However, the establishment of absolute adsorption isotherms requires the assumption of a value for adsorbed-phase density. The assumed adsorbed-phase density affects the estimated maximum adsorption capacity and the shape of the isotherm. Therefore, it is currently impossible to accurately establish an absolute adsorption isotherm. In order to understand the effects of adsorption on the storage and production of natural gas from shale formations we have measured the total methane storage capacity of core plugs from the Barnett and Eagle Ford shale formations. The methane storage capacity was obtained by measuring the NMR T2 spectra of the shale core plugs at pore pressures between 500 and 4,000 psi while maintaining a constant confining pressure of 5,000 psi and a temperature of 30°C. Our measurements show that the effects of adsorption on the storage and production of methane at pore pressures above 2,000 psi are negligible for the two samples studied. However, at pore pressures below 2,000 psi, the increase in total methane storage capacity due to adsorption can reach a factor of 2.5. Therefore, below 2,000 psi, the production of methane will be negatively impacted by the adsorption of methane molecules, while above 2,000 psi the adsorbed and free gas can be produced without distinction.
Many petrophysical properties of tight rocks, such as permeability and electrical resistivity, are very stress sensitive. However, most mercury-injection measurements are made using an instrument that does not apply a confining pressure to the samples. Here we further explore the implications of the use and analysis of data from mercury-injection porosimetry or mercury-injection capillary pressure measurements (MICP). Two particular aspects will be discussed. First, the effective stress acting on samples analyzed using standard MICP instruments is described. Second, results are presented from a new mercury-injection porosimeter that is capable of injecting mercury at up to 60,000 psi into 1- or 1.5-in. core plugs while keeping a constant net stress up to 15,000 psi. This new instrument allows monitoring of the electrical conductivity across the core during the test so that an accurate threshold pressure can be determined.
Although no external confining pressure is applied (unconfined) when using the standard MICP instrument, this doesn’t mean that the measurements can be considered as unstressed. Instead, the sample is under isostatic compression by the mercury until it enters the pore space of the sample. As an approximation, the stress that the mercury places on the sample is equal to its threshold pressure. Thus, the permeability calculated from standard MICP data is equivalent to that measured at its threshold pressure. Not all the samples have the same stress dependency, thus comparing measured permeabilities at a single stress with values calculated from standard MICP data, corresponding at different threshold pressures, can lead to erroneous correlations. Therefore, the estimation of permeabilities from standard MICP data can be flawed and uncertain unless the stress effect is included.
Results obtained from the new mercury-injection system porosimeter under net stress, are radically different from those obtained from standard MICP instruments, such as the Autopore IV. In particular, the measurements at reservoir conditions produce threshold pressures that are three times higher and pore-throat sizes that are one-third of those measured by the standard MICP instrument. The results clearly indicate that calculating capillary-height functions, sealing capacity etc. from the standard instrument can lead to large errors that can have significant impact on subsurface characterization.
Digital rock technology (DRT) has experienced tremendous progress in the last decade, with an increasing number of companies providing imaging hardware, modeling software and digital core analysis services. While prediction remains the most discussed application of DRT, this paper discusses its use to quality control water-displacing-oil relative permeability (kr) experimental measurements. The relative permeability data were collected from three wells over a span of seven years, and they showed a very large spread. To identify potential outliers, we performed micro-CT imaging on six samples that were selected based on similarity in rock properties but differences in measured relative permeability behavior. The three-phase segmentation process was guided by experimental values of porosity, permeability and clay. Consistency checks verified that we could reproduce permeability, drainage capillary pressure, and gas-oil relative permeability. Water-displacing- oil relative permeability was then calculated using pore-network models for water-wet and oil-wet conditions and used to establish a maximum range for each sample. This range was instrumental in identifying suspicious behavior, and reducing uncertainty in recovery predictions by confirming potential outliers and assisting in the upgrading of the experimental relative permeability data.
Digital rock technology (DRT)-based prediction of primary drainage and imbibition water-oil relative permeability can be in good agreement with experimental data if the pore structure, connectivity, and wettability of the porous media are captured accurately (Bakke and Øren, 1997; Øren et al., 1998; Blunt et al., 2002; Al-Kharusi and Blunt, 2008; Idowu et al., 2014; Golab et al., 2015; Massalmeh et al., 2015). However, accurate characterization of wettability inputs, such as contact angles and distribution of oil-wet surfaces, is a challenge (Bondino et al., 2013; Idowu et al., 2015). In this work, we use wettability measurements only as a guide, and focus on comparing experimental results with pore-network simulations of strongly water-wet and oil-wet relative permeability behavior. We expect these simulations to be reasonably accurate (Øren and Bakke, 2003; Valvatne and Blunt, 2004).
A densitometer is used for quantitative density determinations of fluids being produced from core samples during flooding experiments at reservoir conditions. The densitometer is situated in the flowline immediately after the core holder, and measures the density of all fluids being produced from the core sample at the actual pressure/ temperature (P/T) conditions of the flooding experiment. In addition, the densitometer provides timing information about dynamic events during the experiment, e.g. water breakthrough or gas breakthrough.
In the case of two-phase experiments, the densitometer may be used for determining the volumes of the two produced phases, if the density of each of the two fluid phases is known; this is the case in many flooding experiments using oil and water. In such cases, the densitometer may provide data for the produced volumes of oil and water that agree reasonably with fluid volumes determined by an acoustic separator. In complex and prolonged flooding experiments, the densitometer volume determinations may provide an independent confirmation of the volume determinations of an acoustic separator or possibly other devices.
During coreflooding experiments at reservoir conditions it is important to keep track of the fluids being produced from the core sample. For this purpose, a densitometer situated in the flowline immediately downstream to the core sample has proved useful. The densitometer (Paar DMA HPM) has been used at GEUS for obtaining precise density measurements of the fluids being produced from core samples at temperatures up to 115°C and fluid pressures up to 420 bar (Olsen, 2011). However, the rating of the device allows use up to 200°C and 1,400 bar.
An important aspect of preliminary core preparation for SCAL experiments is the restoration of the core sample to its original wettability. The prevalent method for restoration of core samples, to either strongly oil-wet or weakly oil-wet, is largely dictated by increasing or decreasing the time of aging at reservoir temperature. There is no consistent or reliable method ascribed for a specific sedimentary core sample to restore to its original state, which is obtained from preserved core samples. In this study, we have identified three important parameters—brine salinity, restoration temperature, and restoration time (age in number of days)—as contributing to wettability. The objective of this study was to determine the optimum level of these independent variables (brine salinity, temperature, and age) for restoring wettability. We extend the Box- Behnken model of surface response methodology to analyze wettability determined via three methods: contact angle, USBM, and a new method using scanning electron microscopy-mineral liberation analysis (SEM-MLA) over extended ranges of brine salinity (10,000, 100,000 and 200,000 ppm total dissolved salts), temperature (60º, 90º, and 120ºC), and age of conditioning (2, 4, and 8 weeks). The samples for this study included 15 Berea Sandstone samples aged in crude oil and brine of varying salinity. The wettability was experimentally validated using contact-angle measurements, USBM tests, and a novel SEM-MLA imaging (at low vacuum conditions). A seminal effort in applying SEM-MLA image analysis for wettability determination was also explored. Linear regression models were developed and the adequacy of predicting the output variables (wettability) to nearly all conditions were verified. The study showed a comprehensive influence of brine salinity, aging time, and temperature towards wettability restoration. Further 2D and 3D surface plots were generated to show the interaction between the three independent variables in establishing a wettability value.
Rydzy, Marisa B. (Shell International Exploration and Production) | Anger, Ben (Shell International Exploration and Production) | Hertel, Stefan (Shell International Exploration and Production) | Dietderich, Jesse (Shell International Exploration and Production) | Patino, Jorge (Shell International Exploration and Production) | Appel, Matthias (Shell International Exploration and Production)
In this study, digital rock analysis was combined with a variety of experimental core-analysis measurements to investigate the effect of salt saturation and distribution on the porosity and permeability of halite-cemented core samples. Medical and micro-X-ray CT scans of core sections and 2.54-cm (1-in.) diameter plugs indicated that the halite generally occurred in the form of distinct layers. High-resolution micro-X-ray computed tomography (MXCT) images acquired of 0.6-cm diameter plugs revealed that, on the pore scale, halite appeared to be pore filling. Pores were either completely filled with halite or did not contain any halite at all. It was also observed that halite preferentially occurred in the larger pores associated with larger grain sizes. The porosity and permeability results, both measured experimentally on the core plugs and calculated by segmentation of the MXCT images, demonstrated the obstructive effect of halite on storage and flow as well as the decline of both properties with increasing salt saturation. Comparison of calculated and measured values showed that the measured porosity could be up to 6 porosity units (p.u.) higher than the calculated one, while the measured permeability of core plugs after salt removal was an order of magnitude lower than the one obtained by lattice Boltzmann simulation. One possible reason for this discrepancy may be the stratified nature of the samples. While the fully salt-saturated plugs appeared homogeneous in MXCT images, post-flood MRI images revealed that the plug was composed of layers with different MRI intensities, i.e., different amounts of water-filled porosity. Consequently, the petrophysical parameters calculated for the miniplugs may only be representative for a section of the core plug. The results of the MRI-assisted corefloods emphasized the importance of considering different scales when interpreting and applying the results of digital rocks analysis.
King, Hubert (ExxonMobil Research & Engineering Company) | Sansone, Michael (ExxonMobil Research & Engineering Company) | Kortunov, Pavel (ExxonMobil Research & Engineering Company) | Xu, Ye (ExxonMobil Research & Engineering Company) | Callen, Nicole (ExxonMobil Research & Engineering Company) | Chhatre, Shreerang (ExxonMobil Upstream Research Company) | Sahoo, Hemant (ExxonMobil Upstream Research Company) | Buono, Antonio (ExxonMobil Upstream Research Company)
Recent studies on several core-plug-scale samples from tight-oil reservoirs have demonstrated that such rocks can exhibit a substantial, irreversible permeability decline with increase in net confining stress. Because this effect closely follows the expected stress change during drawdown in the field, the origins of this phenomena, as well as a method to predict the magnitude relative to different rock types, is valuable information for reservoir management. To better understand this effect, we have undertaken a series of laboratory studies under in-situ conditions that demonstrate how an external stress field translates to microscopic strain at the pore scale and couples to the fluid transport. These studies rely on the coordinated use of low-field nuclear magnetic resonance (NMR) and X-ray microtomography (XMT). Making use of labeled fluids to enhance contrast, we are able to directly resolve how local strains affect fluid transport throughout the core plug. In a similar manner, proton NMR resolves how stress couples to deformation of the various pore systems, affecting the fluid content and their dynamics. Together, these techniques indicate that internal, high-permeability zones play an important role in the stress dependence. Matrix permeability is much less affected. These higher-permeability zones are not ubiquitous in tight-oil rocks. Characterizing these zones and relating them to mineralogy and rock fabric is an attractive pathway to greater predictability for stress-dependent permeability for reservoir rock types.
Chhatre, Shreerang S. (ExxonMobil Upstream Research Company) | Sahoo, Hemant (ExxonMobil Upstream Research Company) | Leonardi, Sergio (ExxonMobil Upstream Research Company) | Vidal, Keili (ExxonMobil Upstream Research Company) | Rainey, Jennifer (ExxonMobil Upstream Research Company) | Braun, Edward M. (ExxonMobil Upstream Research Company) | Patel, Prateek (Consultant)
Estimation of reservoir rock properties using multiscale imaging of the pore structure, followed by mathematical modeling of the segmented images, i.e., digital rock physics (DRP), is a promising technique. However, DRP workflows are highly variable in terms of imaging tools, resolution of those tools, segmentation algorithms, handling of unresolved porosity, gridding of the resolved pore structure, and mathematical modeling of flow properties. As a result, users familiar with physical measurements of reservoir properties struggle to judge the quality of DRP data, and to incorporate DRP data in commercial workflows in a suitable manner.
In this work, we present a DRP study on tight rocks (kabs < 10 mD) conducted at the laboratories of four vendors of digital rock services, anchored to high-quality physical measurements conducted in our laboratory. We selected core plugs from a set of six outcrop rocks. We cleaned the plugs, measured porosity (φ) and absolute permeability (kabs), and then split the plugs into four quarter plugs that are each 1.5-in. long. Four commercial DRP labs conducted blind porosity and permeability predictions on those quarter plugs using (a) only micro-CT based tools, and (b) all the tools accessible to DRP service providers. We also compare primary drainage capillary pressure (Pc) calculated by four DRP vendors on quarter plugs with centrifuge-based gas-water measurements conducted in-house on companion plugs.
As a result of this blind study, we gained insights into workflows, strengths/weaknesses of DRP predictions carried out by four vendors. Various levels of physical measurements (e.g., laboratory-based kabs, and φ data, MICP, or none) are used by different vendors to anchor DRP data. DRP predictions for porosity ranged from 33 to 96% of the measured values, whereas permeability ranged within a factor of 0.3 to 4 from the experimental measurements. At low Pc values, predictions by the four DRP vendors generally agreed with each other, and with experimental measurements. However, the values diverged significantly at high Pc. Based on this study, we conclude that the dominant source of error in DRP data is highly specific to a given sample, technique, or operator. A lot more uncertainty quantification is necessary to allow DRP data to be used instead of physical measurements for business decisions on tight rocks. We outline learnings for hydrocarbon resource owners and DRP data providers so that commercial workflows could benefit from DRP-based data.
Formation resistivity of tight sandstone in the Keshen area of the Tarim oilfield is greatly affected by geostress. Fluid identification and the accuracy of water saturation calculations are drastically affected by significant increases in rock resistivity resulting from geostress. In this paper, the rock constitutive relation (i.e., the relationship between stress and strain) calculation method of geotechnical structural mechanics and electric field theorem and the finite element method (FEM) numerical simulation are applied to study the effective mechanism of geostress difference on rock resistivity, and a method of geostress correction for rock resistivity is proposed. The paper explains why resistivity increases significantly with the horizontal geostress difference in the water-bearing layer subject to the high geostress of the Keshen area, and also establishes the basis for calculation of water saturation in the high-stress formation.
The Keshen area of the Tarim oilfield is one of the main blocks of Tarim Basin undergoing natural gas development. The target zones are in the Cretaceous Bashijiqike Formation, which consists of tight sandstone reservoirs with very low porosity and permeability. The actual logging data indicate that the formation resistivity in the area is significantly increased by the difference between the maximum horizontal stress and the minimum horizontal stress (hereinafter referred to as the horizontal geostress difference), resulting in resistivity logs that to fail to directly reflect the influence of the formation fluid on resistivity measurements, and lead to the inaccurate estimates of water saturation using the Archie equation. The purpose of this study is to establish a geostress-correction method for rock resistivity suitable for the area, and thereby calculate the resistivity of rock in the absence of geostress difference (hereinafter referred to as the true resistivity).
The Gas Research Institute (GRI) conducted pioneering work on measuring shale petrophysical properties in the 1990s, however, despite growing interest in shales, there are still no set standards with respect to obtaining core petrophysical measurements due to the inherent complexity of shales. Core cleaning is one aspect of this problem.
The objective of this study is to shed some light on the shale core-cleaning conundrum. The study shows the cleaning impact of different solvents on samples from different maturity windows and having different in-situ fluids. It also compares the cleaning efficiency between plug and powdered samples. Different cleaning apparatus, such as the high-pressure extractor (HPE) and the Soxhlet extractor, are also compared.
Different measurements, such as source-rock analysis (S1 and S2 values); gas chromatography-mass spectrometry (GC-MS) extraction analysis; Brunauer-Emmett-Teller (BET) surface area and pore-size distribution help to understand the dynamics of core cleaning. This study was carried out on samples from the Wolfcamp and Eagle Ford formations.
Cleaning has a major impact on various petrophysical properties like porosity (increases up to 50%), S1 (decreases up to 90%) and surface area (increases by 450%). This study showed that cleaning time is a function of maturity and sample state. Samples in the oil-maturity window are much more difficult to clean compared to the samples in the gas-maturity window. Similarly, plug samples are more difficult to clean compared to the crushed samples. Our study shows that toluene, dichloromethane (DCM) and chloroform have similar cleaning efficiencies but n-heptane is less efficient.
Coring is an integral part of any exploration program. The planning for a coring program, coring fluids and corehandling procedures at the wellsite are all very important for preserving the core and getting accurate measurements in the laboratory.