Reliable estimation of kerogen density is a requirement for dependable well-log-based petrophysical evaluation of organic-rich mudrocks. As kerogen matures, hydrocarbons are generated and the chemical structure of kerogen is transformed, which can lead to measurable variations in kerogen density. Uncertainty in estimates of kerogen density can significantly impact the reliability of well-log interpretation results.
The objectives of this research are (a) to experimentally quantify the density of kerogen isolated from a variety of organic-rich mudrocks with different origins, (b) to investigate the impact of thermal maturity on kerogen density, and (c) to investigate the impact of synthetic maturation on density of kerogen. We used organic-rich mudrock samples from four formations, to cover a wide range in kerogen thermal maturity. We isolated kerogen from these mudrock samples and estimated the density of the naturally and synthetically matured isolated kerogen samples.
The experimental results indicated that the density of kerogen varies significantly among organic-rich mudrocks with different origins. We recorded densities ranging from 1.19 to 1.77 g/cm3 in kerogen samples when the hydrogen index varied from 603 to 48 mg hydrocarbon/g organic carbon. We also observed that kerogen density increases as a function of thermal maturity. Sensitivity analysis confirmed a measurable impact of kerogen density on estimates of petrophysical properties, such as porosity and water saturation in organic-rich mudrocks. The documented experimental results and procedures can be used to enhance petrophysical evaluation of organic-rich mudrocks, by taking into account the impact of kerogen thermal maturity in the models used for interpretation of core or well-log measurements.
Kerogen disseminated in organic-rich mudrocks presents challenges when performing well-log-based petrophysical evaluation. One such challenge is the lack of reliable estimates of kerogen density in organic-rich mudrocks. Inaccuracies in estimates of kerogen density can negatively influence assessments of porosity, mineralogy, and water saturation in organic-rich mudrocks.
This paper narrows down the knowledge gap in interpretation of electrical resistivity measurements in oil-wet and mixed-wet formations by analytically deriving a new resistivity model that can reliably estimate hydrocarbon reserves at different levels of wettability. The objectives of this paper include (a) to quantify the influence of wettability on electrical resistivity measurement, (b) to develop a new analytical resistivity model that takes into account the impacts of wettability on electrical resistivity, and (c) to improve the assessment of hydrocarbon saturation by introducing a wettability-dependent parameter into a new resistivity method. The new resistivity model not only incorporates wettability of the rock, but also a directionally conducting fractional pore network to honor rock fabric.
The aforementioned features are quantitatively evaluated from the three-dimensional (3D) pore-scale images, taken from each rock type in the formation. We apply a semianalytical streamline numerical model to estimate pore-network connectivity in the 3D binary images. The resistivity and the calculated geometry-related parameters are used as inputs to the new model in order to estimate water saturation. To test the performance of the introduced method at different levels of wettability and water saturation, we synthetically saturate the porescale images with water and oil at different wettability configurations and water saturation, honoring the physics of intermolecular interactions between different fluid and solid components.
The results obtained from the new method are compared against the actual saturation. We successfully applied the introduced method to carbonate rock samples with wettability ranging from strongly oil-wet to strongly water-wet. The electrical resistivity results obtained from numerical simulations were in agreement with the resistivity estimates from the new method. The results also showed that wettability has a significant influence on electrical resistivity of the rocks at water saturation levels below 50%. Moreover, we demonstrated that the proposed model provides reliable results when applied to field data. The outcomes of this paper are promising for well-log-based applications of the new method in complex mixed-wet formations.
Dielectric-permittivity measurements are typically used to estimate water-filled porosity. The dielectric interpretation methods, such as, complex refractive index model (CRIM) (i.e., volumetric techniques) are extensively used to correlate dielectric permittivity of fluid-bearing rocks to petrophysical properties such as water-filled porosity. However, volumetric techniques usually oversimplify the rock structure and do not take into account the impact of spatial distribution of solid and fluid components on dielectric properties of the rock. The lack of reliable rock-physics models to interpret dielectric-permittivity measurements can lead to significant uncertainty in estimates of water-filled porosity.
We applied a pore-scale numerical simulation method to quantify the impact of pore and grain structures and heterogeneity on dielectric-permittivity measurements, and introduced a new dielectric model to improve assessment of water-filled porosity in formations with complex pore/grain structure. A directional-tortuosity factor was introduced and calculated using pore-scale rock volumes to quantify the geometry of pore and grain networks.The introduced techniques were applied on 3D computed-tomography (CT) scan images of sandstone and carbonate rock samples as well as synthetic organic-rich mudrocks. We showed that the new model is more reliable for assessment of water-filled porosity compared to the conventional CRIM, in the 12 sandstone and carbonate rock samples evaluated in this paper. In the case of synthetic organic-rich mudrocks, we observed that (a) despite the change of the overall dielectric permittivity, the accuracy in estimates of water-filled porosity is not affected by the presence of kerogen, if the influence of kerogen is correctly taken into account by CRIM, and (b) the presence of pyrite and its spatial distribution significantly affect the dielectric permittivity of organic-rich mudrocks. Failure to consider the influence of pyrite and its spatial distribution on dielectric permittivity may cause large uncertainty in estimates of water-filled porosity.
Petrophysical evaluation of complex formations using well logs is challenging due to the uncertainty in conventional petrophysical models in these reservoirs. Core measurements, on the other hand, are usually sparse and do not provide real-time and in-situ depth-by-depth petrophysical characterization. Several studies focus on improvements in workflows for core measurements and well-log interpretation methods. However, we approach this challenge by introducing a method to enhance the sensitivity of well logs to petrophysical properties of the formation and to the presence of natural fractures using nanoparticles as contrast agents that can travel into the rock.
We quantify the effect of petrophysical properties of porous media on the spatial distribution of magnetic nanoparticles injected as contrasting agents to the formation. We achieved this goal by developing a two-phase fluid-flow numerical simulator for transport of injected nanoparticles in fractured porous media and synthesizing magnetic nanoparticles and characterizing their properties. The developed simulator takes into account the deposition/adsorption of nanoparticles based on colloid filtration theory and Brownian diffusion. We documented the results of numerical simulations for homogenous and heterogeneous naturally fractured porous media. The effluent history from the coreflood of sandstone with the synthesized nanoparticle solution confirms the results from numerical simulations. We showed that the nanoparticles are mostly concentrated in the connected natural fractures when the permeability of matrix is significantly lower than that of fractures. The sensitivity analysis shows negligible impact of nanoparticle size on spatial distribution of nanoparticles because of the dominance of dispersion over diffusion effects. The results are promising for measureable impact of nanoparticles on specific borehole geophysical measurements to better characterize naturally fractured rocks.
Assessment of micro-fracture density in hydrocarbon-bearing reservoirs is of special interest for designing production plans and selecting zones for fracture treatment. NMR (nuclear magnetic resonance) T2 (spin-spin relaxation time) distribution has been traditionally considered insensitive to the presence of fractures. However, in a recent publication, we documented a measureable NMR sensitivity to the existence of micro-fractures and proposed a new concept of fracture-pore diffusional coupling. The quantification of micro-fracture density in multipleporosity systems is a challenging issue, and distinguishing fractures from pore space is not possible from NMR T2 measurement alone. However, the inclusion of additional borehole measurements, such as induction logs, enables evaluation of micro-fracture density. In this paper, we introduce a new method to evaluate the porosity associated to micro-fractures and intra-/inter-granular pores in complex formations using combined interpretation of NMR and EM (electromagnetic) measurements.
Performance of hydraulic fractures highly depends on formation elastic properties. Empirical rock-physics formulations, developed for the assessment of elastic properties, might not be reliable in organic-shale formations as the result of heterogeneity and anisotropy in these formations. Several models based on effective medium theories have been used to assess elastic properties of organic-shale formations using well logs. Their reliability, however, is still questionable in anisotropic shale formations. Furthermore, many of the existing methods require calibration against core measurements, which can introduce additional uncertainty in the assessment of elastic rock properties.
In this paper, we applied different empirical rock-physics equations and effective medium theories to estimate elastic properties in the Haynesville and the Cana-Woodford shale formations. The inputs to the models include: (a) acoustic measurements, (b) well-log-based estimates or laboratory measurements of petrophysical and compositional properties of the formation, (c) parameters quantifying grain/pore shapes, and (d) elastic properties of rock constituents. Next, we performed tri-axial laboratory experiments at different stress levels to measure elastic properties and acoustic-wave velocities in different rock types for both field examples. We then compared the estimates of elastic properties obtained from different methods against core measurements and investigated the reliability of different models.
We showed that the elastic properties obtained from acoustic-wave velocities using conventional empirical techniques are not always in agreement with core measurement. We found the estimates from inclusion-based effective medium theories the most reliable representatives for elastic properties of organic-shale formations. However, the uncertainty in the shape of inclusions still remains a challenge in these techniques. The outcomes of this paper can potentially be a guide for selection of a reliable model for well-log-based assessment of effective elastic moduli in organic-shale formations, which contributes significantly to detecting zones for successful fracture treatment and production.
Reliable evaluation of rock mechanical properties is important to predict response to fracture treatment in organic-shale formations. However, characterization of rock mechanical properties in organic-shale formations is challenging, as the result of heterogeneity, anisotropy, and complex lithology. Although organic-shale formations are anisotropic and highly laminated, their mechanical properties are often estimated using conventional techniques, with the assumption of isotropy, either in the laboratory or from interpretation of borehole acoustic measurements. Isotropic assumptions in conventional laboratory measurement are not appropriate for organic-shale formations, and, therefore, they can result in uncertainties in stress prediction.
The objective of this paper is to estimate elastic properties, earth stress coefficient, and minimum horizontal stress of highly-laminated organic-shale formations using laboratory experiments. We use hydrostatic tests, conventional triaxial tests, and uniaxial strain tests to measure rock compressibility, anisotropic elastic properties, and earth stress coefficient. We introduce the measured elastic properties and earth stress coefficient, Ko, in the stiffness matrix of the TIV (transversely isotropic media with a vertical axis of symmetry) model and compare the values of the minimum horizontal stress for anisotropic elastic properties and for the earth stress coefficient, Ko.
We assessed mechanical anisotropy in the Haynesville shale-gas formation through experimental techniques. This formation is both highly laminated and naturally anisotropic on both, micro and macro scales. The estimated compressional modulus of organic-shale samples using the TIV stress profile model are more than 30% less than those estimated from conventional techniques. The Young’s modulus and Poisson’s ratio in the direction parallel to the bedding plane were higher than the values obtained in the normal direction to the bedding plane. The earth stress coefficient, Ko, which is a measure of rock stiffness can be used to determine mechanical behavior in organic-shale formations.
Interpretation of resistivity logs in organic-rich mudrocks has been challenging for petrophysicists. Conventional resistivity-porosity-saturation models (e.g., Archie, dual-water, and Waxman-Smits) assume that the pore water is the only conductive part of the formation. However, this assumption is not reliable in the presence of the thermally mature organic matter, clay, and pyrite found in many source rocks. Previous experimental studies indicate that as the aromaticity of kerogen increases with increasing thermal maturity, the conductive behavior of kerogen is also affected. In this paper, we investigated the effect of conductive kerogen on the resistivity of mature source rocks.
We investigated the reliability of conventional resistivity-porosity-saturation models for estimating fluid saturation in organic-rich mudrocks, using well logs and core measurements from the Haynesville Shale. We then numerically simulated the electric field, electric currents, and effective resistivity of synthetic pore-scale rock images. Finally, we used numerical simulation to quantify the effects of the (a) volumetric concentration of kerogen, (b) kerogen conductivity, and (c) spatial connectivity of the kerogen-water network on the resistivity of organic-rich mudrocks.
Well-log interpretation of the Haynesville Shale showed that conventional resistivity-porosity-saturation models underestimate hydrocarbon saturation in zones with high concentrations of kerogen. In this paper, we show that errors in the estimation of of hydrocarbon saturation could be due to the effect of kerogen on resistivity measurements. Our pore-scale numerical simulations show that there is a 10 to 23% improvement in hydrocarbon-saturation estimates, when the impact of conductive kerogen is taken into account. Results of this work can be used as a first step towards improving conventional resistivity-porositysaturation models for the assessment of fluid saturations in organic-rich mudrocks.
Heidari, Zoya (The University of Texas at Austin) | Torres-Verdín, Carlos (The University of Texas at Austin) | Mendoza, Alberto (ExxonMobil Exploration Company) | Wang, Gong Li (Schlumberger Sugar Land Product Center)
Estimation of residual hydrocarbon saturation remains an outstanding challenge in formation evaluation and core analysis. Standard interpretation methods for nuclear-resistivity logs cannot distinguish between mobile and residual hydrocarbon saturation. In extreme cases, fluid pump out or production testing are the only options to ascertain whether the reservoir’s in-situ hydrocarbon is mobile.
We develop a new method to distinguish mobile from residual hydrocarbon and to quantify residual hydrocarbon saturation. The method combines modeling of resistivity and nuclear logs with the physics of mud-filtrate invasion to quantify the effect of residual hydrocarbon saturation on both nuclear and resistivity logs. This strategy explicitly takes into account the different volumes of investigation of resistivity and nuclear measurements and does not assume that the near-borehole region is flushed to the level of residual hydrocarbon saturation. The method begins with an initial multi-layer petro physical model which is constructed via standard procedures of well-log interpretation and core measurements. Thereafter, we simulate the physics of mud-filtrate invasion and the corresponding resistivity, density, and neutron logs. Initial estimates of residual hydrocarbon saturation and parametric relative permeability are refined until achieving a good agreement between simulated and measured neutron and density logs. Next, we refine initial estimates of water saturation, porosity, and permeability until securing a good match between numerically simulated and measured resistivity logs.
The method of interpretation considers two specific options for implementation: (1) quantification of the influence of residual hydrocarbon saturation on the radial distribution of fluid saturation due to invasion, and (2) appraisal of invasion effects on the vertical distribution of fluid saturation within a flow unit that exhibits both hydrocarbon and water saturation in capillary equilibrium.
Application examples are described for the cases of tight-gas sand reservoirs invaded with water-base mud (WBM) and oil-bearing reservoirs invaded with oil-base mud (OBM). In the case of tight-gas sands, our method explains the marginal productivity of deeply invaded beds that exhibit cross-over between density and neutron logs. For a 15-porosity unit formation, when the residual gas saturation increases by 10 saturation units, the cross-over between neutron and density logs increases by 2.4 porosity units. Interpretation results indicate measurable sensitivity of nuclear logs to residual hydrocarbon saturation in cases of deep WBM invasion due to immiscibility between invaded and in-situ fluids. However, the accuracy of the method decreases with increasing values of both hydrocarbon pore volume and hydrocarbon density. In the case of OBM invasion, reliable estimations of residual hydrocarbon saturation are possible with relative density differences above 15 per-cent between mud filtrate and in-situ hydrocarbon.