We experimentally verified that material-averaging techniques can be used for numerically modeling the response of the crosswell electromagnetic (crosswell EM) system in a fractured medium. We have designed a scaled model of the crosswell EM system, and verified that the laboratory data are in good agreement with the simulated response for different cases of fractures present in a host medium. Simulations of realistic scenarios in a hydrocarbon-filled reservoir indicate that the presence of fracture clusters can significantly affect the crosswell EM system response. The magnitude of the effect primarily depends on the number of fractures and their dip relative to the transmitter and receiver wells. In particular, small clusters of fractures filled with saline connate water produce large artifacts on the response, provided that the relative dip to the wells is large enough. The presence of a fracture cluster creates a characteristic signature in the full crosswell EM tomographical data set; therefore, crosswell EM data provides information related to the main fractures that are present between the wells. The inclusion of the fracture information as an additional constraint to the inversion of the crosswell EM data is expected to provide an improved interpretation of the reservoir fluids in the interwell space.

Rock surface wettability controls the fluid distribution in the pore space of the formation and thus, has an influence on its electrical properties. The variation of dielectric dispersion data with the variation of formation wettability is not well understood at present time. In this study, we investigated the influence of wettability alteration on dielectric dispersion data by laboratory experiments using different dielectric data interpretation models. We used a reflection open-ended coaxial probe to measure dielectric dispersion of a 1.5-in. core plug in a frequency range of 10 MHz to 1 GHz. The studied rock wettability was varied in a coreflooding apparatus during a drainage experiment. We used low-polarity oil for the initial drainage to preserve the rock in the water-wet state and added different concentrations of surface-active stearic acid to the oilflood during later drainage stages to change the sample wettability. The used dielectric data interpretation models, in addition to standard interpretation answers, solve for geometrical parameters of the fluids and pore space, which allow us to evaluate wettability variation. Our results showed a decrease in dielectric dispersion at the low-megahertz frequency range as we inject higher acid concentrations into the rock. This decrease in the dispersion also corresponds to an increase of the inverted MN-parameter, which characterizes water phase tortuosity. The results of the inversion, using the textural model, showed a consistent decrease of the depolarization factor of water inclusions as we increase the acid concentration in the injected oil. At the highest acid concentration used (1 wt%), this depolarization factor drops down to 0.36, which is very close to the depolarization factor of spherical inclusions. It follows from our findings that water films, which are characteristic for water-wet rocks, are transformed to more spherical droplets due to the change of wettability to oil-wet state. NMR T2 measurements on the same plug and contact angle measurements on reference samples of same lithology support our conclusions on wettability variation. This study illustrates how to interpret the rock wettability state from dielectric dispersion data based on geometrical models and how to use the inverted geometrical parameters as a wettability indicator.

Petrophysical models are used to estimate formation water saturation and salinity from dielectric dispersion data. These models are sensitive to other attributes of the formation such as wettability. Previous studies, analyzing different parameters of dielectric petrophysical models have shown how the dielectric dispersion is qualitatively influenced by rock wettability. Such analysis requires prior knowledge of the water saturation. To date, there is no direct method to quantify wettability from dielectric data. In this work, we discriminate the effects of water saturation and wettability on dielectric dispersion data. We conditioned a 1.5-inch-diameter dolomite plug with solutions of different concentrations of stearic acid. In each set of experiments, we imbibed water into the sample and measured the dielectric dispersion in a frequency range of 10 MHz to 1 GHz using a reflection coaxial probe. This method enables us to obtain several dielectric measurements on a rock at defined wettability states and at different saturation levels. Then, the inverted dielectric interpretation answers such as water-phase tortuosity (MN) are analyzed for each dataset to distinguish their sensitivity to saturation and wettability. After each set of experiments, we measured the wettability index of the sample using US Bureau of Mines (USBM) and Nuclear Magnetic Resonance (NMR) methods and correlated it with dielectric inverted parameters. Our findings show that the inverted interpretation parameters from dielectric data correlate well with the wettability index from USBM. These parameters include MN, the grain aspect ratio for the bimodal model, and the depolarization factors of water and oil-matrix of the textural model. We aged the sample into three wettability conditions: neutral-wet, strongly water-wet, and strongly oil-wet. For extreme wettability conditions, we observed a consistent trend of dielectric interpretation parameters with wettability during water flooding independent from saturation. For the neutral wettability case, the inverted dielectric parameters are constant up to a water saturation of about 50% and then starts to change gradually. For the strongly water-wet case, we observed a similar trend for depolarization factor as found in the literature, in which the nonwetting phase depolarization factor is close to the spherical geometry. Our study illustrates how the inverted dielectric dispersion petrophysical parameters can correlate with well-established laboratory measurement of the wettability index, such as USBM method, for a specific core plug. This correlation can aid in estimating the wettability index directly from dielectric dispersion data for cores with similar rock textures. Different rock textures may impose different correlations, and further work is needed to establish a dielectric correlation for each rock type in a reservoir for application to downhole formation evaluation.

Due to processes of geological diagenesis, pores in rocks can be isolated from the rest of the connected pore networks. The amount and spatial distribution of isolated pores can have direct effect on petrophysical properties and performance of the reservoir. This paper introduces a new methodology to quantify the isolated porosity of heterogeneous reservoirs from multi-frequency (dispersion) dielectric measurements. Based on numerical simulation studies, digital rock physics techniques are used to generate rock models with different isolated porosities. 3D dielectric dispersion modeling is then performed on the models to obtain the dispersion of rock’s dielectric constant. Dielectric dispersion behaves differently as the pore connectivity changes due to increase in isolated porosity. Dielectric constant is sensitive to frequency when pores are isolated, while insensitive to frequency when pores are connected. Variation of dielectric constant is strongly related to the amount of isolated pores. For rocks having the same total porosity, their dielectric constant increases as the isolated porosity increases. This enhancement of dielectric constant is attributed to the increase in pore network tortuosity, resulting in increased accumulations of electric charges at the interfaces between solid and pores. Analytical relationships are developed to correlate isolated porosity with the rate of permittivity change and/or the permittivity ratio, derived from the dispersion of dielectric constants. The validity and applicability of the established method are demonstrated by the agreement of predicted isolated porosity with the true values used in building the rock models. Potentially, this method can be used for enhancing reservoir characterization with modern multifrequency dielectric logs.

Archie equation is the foundation of modern petrophysics and parameters m, cementation exponent, and n, saturation exponent, are critical inputs of Archie equation. Traditionally, m and n are obtained from core analysis, which are available only in cored formations, and the process is costly and time consuming. It is always desirable if m and n can be derived continuously from downhole measurements, such as multi-frequency dielectric logs, the objective of this study. Using an open-ended coaxial laboratory test probe operating between 10 MHz and 1 GHz, multi-frequency dielectric data are measured on clean outcrop core samples of both carbonate and sandstone. The data are analyzed by using the traditional dielectric interpretation models assuming that m equals to n i.e., considering matrix and hydrocarbon as one phase and water as the second phase, the MN approach. To extract m and n individually so that they can be properly used in Archie equation, our approach considers matrix, hydrocarbon and water as three different phases in dielectric data processing, by applying effective medium theory on the formation rock in two steps: first between water and hydrocarbon resulting in an effective fluid permittivity, then second between effective fluid permittivity and matrix. Results are compared and validated with m and n from core analysis using resistivity measurements. The derived m and n from dielectric data using the new method show significant improvement over the MN approach. When compared with data from resistivity tests, the single-phase property m derived from the new method performs much better than that from MN approach. As for the two-phase property n from the primary drainage experiment, a significant improvement is also realized. The findings were consistent for different lithologies, sandstone and carbonate. When using m and n derived from the new approach to interpret resistivity data for water saturation of core plugs, the average uncertainty is improved from 13%, when MN is used, down to 4%, when m and n are used. Using the new method presented in this study to interpret multi-frequency dielectric data has the potential to be implemented in interpreting downhole dielectric logs for continuous Archie parameters m and n, resolving a long-standing challenge in formation evaluation.