|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Abstract Broadband relative dielectric dispersion measurements are considered interesting options for assessment of water-filled pore volume. Conventional models such as Complex Refractive Index Model (CRIM) and Maxwell Garnett (MG), often overlook or oversimplify the complexity of pore structure, geometrical distribution of the constituting fluids, and spatial distribution of minerals. This yields to significant errors in assessment of water saturation especially in rocks with complex pore structure. Therefore, it becomes important to quantify the impacts of pore structure and spatial distribution of minerals on broadband relative dielectric dispersion measurements to be able to make decisions about reliability of water saturation estimates from these measurements in a given formation. The objectives of this paper are (a) to quantify the impacts of pore structure and spatial distribution of minerals on relative dielectric permittivity measurements in a wide range of frequencies, (b) to propose a new simple and physically meaningful workflow, which honors pore geometry and spatial distribution of minerals to enhance fluid saturation assessment using relative dielectric permittivity measurements, (c) to verify the reliability of the introduced model in the pore-scale domain. First, we perform numerical simulations of relative dielectric dispersion measurements in the frequency range of 20 MHz to 1 GHz in the pore-scale domain. The input to the numerical simulator includes pore-scale images of actual complex carbonate rock samples. We use a physically meaningful model which honors spatial distribution of the rock constituents for the multi-frequency interpretation of relative dielectric response. To verify the reliability of the model in multiple frequencies, we apply the model to the results of relative dielectric simulations in the pore-scale domain on 3D computed tomography scan (CT-scan) images of carbonate rock samples, which are synthetically saturated to obtain a wide range of water saturation. We successfully verified the reliability of the introduced model in the pore-scale domain using carbonate rock samples with multi-modal pore-size distribution. Estimated water saturations from the results of simulations at 1 GHz resulted in an average relative error of less than 4%. We observed measurable improvements in fluid saturation estimates compared to the cases which CRIM or MG models are used. Results demonstrated that application of conventional models to estimate water saturation from relative dielectric response is not reliable in frequencies below 1 GHz.
This paper presents a continuation of our fundamental investigations into the 2D T1-T2 NMR response of saturating fluids in the organic-matter pores of pelletized kerogen isolates. We previously reported that T1-T2 data of heptane-saturated kerogen pellets showed two distinct peaks: (1) a slow-relaxing peak, interpreted as heptane in the kerogen intergranular pores created during pelletization, and (2) a fast-relaxing peak with large T1/T2 ratio, interpreted as heptane absorbed in kerogen granules, i.e. in intragranular pores.
In this study, we investigate the influence of bitumen extraction on the T1-T2 data of the heptane-saturated kerogen pellets, and we use supporting data, such as kerogen swelling effects, nitrogen adsorption BET, and UVVIS absorption spectroscopy to enhance the interpretation of the NMR data. We find that for the fast-relaxing peak, the T1 and T2 values remain roughly the same after bitumen extraction, however the porosity decreases, which strongly suggests that the fast-relaxing peak is associated with heptane absorbed in the intragranular pores of bitumen and kerogen. For the slow-relaxing (intergranular) peak, we find that the porosity remains roughly the same after bitumen extraction, however the T1 and T2 values increase due to a decrease in apparent surface relaxivity, which we attribute to (weak) diffusive-coupling effects between the inter and intragranular porosities.
Our findings provide key insight into the role of kerogen and bitumen on the NMR response in organic shale, which can be used to improve fluid typing and saturation estimates from 2D T1-T2 NMR data, both in the lab and from downhole logs.
In the past decade, as the oil and gas production from unconventional reservoirs increased dramatically, the investigations into organic shale have greatly stimulated both NMR log data interpretation and NMR core analysis (Jiang et al., 2013; Kausik et al., 2016; Reeder et al., 2016; Anand el al., 2017; Tandon et al., 2017; Washburn and Cheng, 2017). Among them, studies focused on kerogen have become more and more popular (Ertas et al., 2006; Chen et al., 2012; Singer et al., 2016, 2017; Zhang and Daigle, 2017). Kerogen, which is defined as solid, insoluble and immobile organic matter, constitutes most of the total organic content (TOC) of organic shale (Durand, 1980), which makes characterizing kerogen essential for formation evaluation.
Durand, Melanie (Shell Exploration and Production Company) | Nikitin, Anton (Shell International Exploration and Production) | McMullen, Adam (Shell Exploration and Production Company) | Blount, Aidan (Shell Exploration and Production Company) | Driskill, Brian (Shell Exploration and Production Company) | Hows, Amie (Shell International Exploration and Production)
ABSTRACT As activity increases in the Permian Basin and multiple billion-dollar acquisitions at upwards of &50,000/acre continue, there is a strong incentive for E&P operators to optimize the development in their existing acreage. Unfortunately, maximizing oil production typically results in significant amounts of produced water. Water cuts for individual Permian wells commonly range from 50 to 90% of total liquid production, thus the ability to predict water to oil ratio (WOR) of the produced fluids has a major importance for development planning (Scanlon et al., 2017). Petrophysicists are responsible for fluid saturation modeling, which provides the basis for predicting WOR. Core data acquisition and analysis are critical for developing a quantitative petrophysical model. However, accurately measuring saturations of cores taken from unconventional reservoirs continues to pose significant challenges originating from uncertainties in the acquired data, assumptions used to interpret these data and more broadly, due to increased relative uncertainty associated with tight, low-porosity formations. For example, the crushing of the core samples, which is required for efficient fluid extraction in tight rocks, causes systematic fluid losses which are not typically quantified. Instead, all as-received air-filled porosity is commonly assumed to represent hydrocarbons that have escaped during coring due to gas expansion. Additionally, fluid extraction from commercially available retorting systems can have widely variable fluid collection efficiency (<100%) resulting in significant inconsistencies between the weight of the collected fluids and sample weight loss during retorting experiments. The Dean-Stark technique removes not only fluids (water and oil) but an unknown volume of the extractable organic matter, and it only allows for direct quantification of the volume of extracted water. The reconciliation of fluid volume as well as fluid and sample weight data delivered by either of the two techniques (i.e., retorting or Dean-Stark) requires numerous assumptions about pore fluid properties which are typically not verified through direct measurements. We demonstrate that such assumptions can lead to up to 50% uncertainty in water saturation estimates. To address such critical uncertainties, a new core analysis workflow using improved core characterization and fluid extraction techniques was developed. To address fluid loss during crushing, this workflow employs advanced NMR measurements performed on both as-received and crushed samples to quantify fluid losses. Also, this approach uses retorting techniques with close to 100% fluid collection efficiency specially developed for core sample characterization. The workflow is further optimized to avoid fluid loss during sample handling and includes repeated grain density and geochemical measurements at different stages. As a result, the new workflow addresses uncertainties in acquired data and better informs the assumptions for interpreting the measured data into the desired petrophysical properties (e.g. total porosity, water saturation). The workflow is demonstrated for a set of Wolfcamp samples.
Seleznev, Nikita (Schlumberger) | Habashy, Tarek M. (Schlumberger) | Claverie, Michel (Schlumberger) | Wang, Hanming (Chevron U.S.A. Inc.) | Wang, Haijing (Chevron U.S.A. Inc.) | Hermes, Amir (Schlumberger) | Gendur, Jason (Schlumberger) | Feng, Ling (Schlumberger) | Loan, Mary Ellen (Schlumberger)
ABSTRACT Tight oil reservoirs present a unique opportunity for dielectric dispersion logging. Dielectric logging is sensitive to the water content and provides water-filled porosity without having to know Archie’s empirical parameters or water salinities, as is required with resistivity log interpretation. Moreover, because of the extremely low permeability of the shale reservoirs, there is effectively no invasion of the borehole fluids into the formation. Thus, in these reservoirs, dielectric dispersion logging directly provides the water-filled porosity of the undisturbed zone. In this paper, we investigate the interpretation of the dielectric dispersion measurements in tight oil formations. A representative core collection was obtained from two intervals in a field. The core material was characterized in terms of lithology and total organic carbon (TOC) content. The cores were cleaned and saturated with brines that match the formation water salinities. Next, the dielectric dispersion measurements on cores were obtained under controlled laboratory conditions of pressure, temperature, and brine salinity. On the basis of the analysis we conducted on these data, we have developed a new method for the interpretation of multifrequency dielectric logs in tight oil reservoirs. The new method has a significant advantage over the existing approaches because it does not require an input for either matrix or hydrocarbon permittivities, including kerogen permittivity, to derive water-filled porosity as is the case with the existing approaches. The new method enables the elimination of all associated uncertainties with formation mineral models in complex lithologies, unknown mineral permittivity endpoints, and, most importantly, the poorly defined permittivity of kerogen. The new method requires only the relatively well-known input of formation temperature. Thus, the new method provides a more robust, streamlined, and consistent interpretation of the dielectric dispersion logs in tight oil and reduces the uncertainty on the estimate of hydrocarbon in place. INTRODUCTION Currently the Permian Basin produces ∼4.8 million barrels of oil per day, constituting more than a third of total US production of ∼13 million barrels of oil per day (US EIA 2020). The Wolfcamp and Spraberry formations are the main producing intervals in the Permian Basin. Despite the economic significance of these reservoirs, challenges in their formation evaluation remain to be addressed.
Abstract This research proposed an alternative method for determining the saturation exponent (n) by finding the best correlations for the heterogeneity index using available core data and considering wettability changes. The log curves of the variable n were estimated, and the effect on the water saturation (Sw) calculations and the Stock Tank Oil Initially In Place (STOIIP) in the Tambaredjo (TAM) oil field was analyzed. Core data were employed to obtain the relationship between n and heterogeneity using cross-plots against several heterogeneity indices, reservoir properties, and pore throat size. After filtering the data, the clay volume (Vcl), shale volume, silt volume, basic petrophysical property index (BPPI), net reservoir index, pore grain volume ratio, and rock texture were defined as the best matches. Their modified/improved equations were applied to the log data and evaluated. The n related to Vcl was the best selection based on the criteria of depth variations and logical responses to the lithology. The Sw model in this field showed certain log readings (high resistivity [Rt] reading ≥ 500 ohm.m) that infer these intervals to be probable inverse-wet (oil-wet). The cross-plots (Rt vs. Vcl; Rt vs. density [RHOB]; Rt vs. total porosity [PHIT]) were used to discard the lithologies related to a high Rt (e.g., lignites and calcareous rocks) and to correct Sw when these resulted in values below the estimated irreducible water saturation (Swir). The Sw calculations using the Indonesian equation were updated to incorporate n as a variable (log curves), comparing it with Sw from the core data and previous calculations using a fixed average value (n = 1.82) from the core data. An integrated approach was used to determine n, which is related to the reservoir’s heterogeneity and wettability changes. The values of n for high Rt (n > 2) intervals ranged from 2.3 to 8.5, which is not close to the field average n value (1.82). Specific correlations were found by discriminating Swir (Swir < 15%), (Swir 15%–19%), and Swir (> 19%). The results showed that using n as a variable parameter improved Sw from 39.5% to 36.5% average in the T1 and T2 sands, showing a better fit than the core data average and increasing the STOIIP estimations by 6.81%. This represents now a primary oil recovery of 12.1%, closer to the expected value for these reservoirs. Although many studies have been done on n determination and its effect on Sw calculations, using average values over a whole field is still a common practice regardless of heterogeneity and wettability considerations. This study proposed a method to include the formation of heterogeneity and wettability changes in n determination, allowing a more reliable Sw determination as demonstrated in the TAM oil field in Suriname.