Al-Obaidi, Mohammed (The University of Texas at Austin) | Heidari, Zoya (The University of Texas at Austin) | Casey, Brian (University Lands) | Williams, Richard (University Lands) | Spath, Jeffrey (Texas A&M University)
Borehole image logs provide a continuous and high-resolution record of the rock sequence and can capture fabric-related features. However, automatic integration of image logs, conventional logs, and core measurements for rock classification can be challenging. In this paper, we propose a method for automatic characterization of rock fabric by joint interpretation of image logs and conventional well logs. This method also enables detection of optimum number of rock classes without a priori knowledge of rock types. The objectives of this paper include (a) developing an intelligent rock fabric recognition workflow that uses image logs as inputs, (b) integrating the identified rock fabric with conventional logs and core measurements for rock classification and for selection of optimum rock classes for completion, (c) using the rock classification outcome to improve formation evaluation, and (d) extending the introduced classification workflow to neighboring wells. First, we combine first-order statistics, cross-correlation, and Discrete Fourier Transform (DFT) methods to extract fabric-related features from image logs. We then integrate the detected fabric-related features with conventional logs and feed them into an unsupervised classification algorithm, called x-means, for optimizing the number of rock classes and rock type detection. After achieving an optimized rock classification, petrophysical models were improved on a class-by-class basis.
We successfully applied the proposed method to a data set from the Wolfcamp formation in the Southern Midland Basin and Ozona Arch. The data set consisted of eight wells with image logs, two of which contained detailed core data and description. We use the core measurements and description in these two wells to cross-validate the results from the proposed workflow. Comparison between the optimized rock classes and lithofacies revealed that the proposed method can automatically detect the lithofacies with an error of approximately 15%. The results showed that the estimates of petrophysical and compositional properties of the rocks improved relatively up to 20% compared to the cases where automatic rock fabric detection and image analysis was not implemented as part of the formation evaluation workflow. Analysis of the production data revealed that wells completed within the recommended zones outperform the next tier of wells by approximately 40%. The results showed that quantifying fabric-related features from image logs can be used to improve completion decisions.
Bennett, Nicholas (Schlumberger) | Donald, Adam (Schlumberger) | Ghadiry, Sherif (Schlumberger) | Nasser, Mohamed (Schlumberger) | Kumar, Rajeev (Schlumberger) | Biswas, Reetam (Schlumberger and University of Texas)
A new sonic imaging technique uses azimuthal receivers to determine individual reflector locations and attributes such as the dip and azimuth of formation layer boundaries, fractures, and faults. From the filtered waveform measurements, an automatic time pick and event localization procedure is used to collect possible reflected arrival events. An automatic ray tracing and 3D slowness time coherence (STC) procedure is used to determine the ray path type of the arrival event and the reflector azimuth. The angle of incidence of the reflected arrival is related to the relative dip, and the moveout in 3D across the individual sensors is related to the azimuthal orientation of the reflector. This information is then used to produce a 3D structural map of the reflector which can be readily used for further geomodeling.
This new technique addresses several shortcomings in the current state-of-the-art sonic imaging services within the industry. Similar to seismic processing, the current sonic imaging workflow consists of iteratively testing migration parameters to obtain a 2D image representing a plane in line with the desired receiver array. The image is then interpreted for features, which is often subjective in nature and does not directly provide quantitative results for the discrete reflections. The technique presented here, besides providing appropriate parameter values for the migration workflow, further complements the migration image by providing dip and azimuth for each event that can be used in further downstream boundary or discontinuity characterization.
A field example is presented from the Middle East in which a carbonate reservoir was examined using this technique and subsequently integrated with wellbore images to provide insight to the structural geological setting, which was lacking seismic data due to surface constraints. Structural dips were picked in the lower zone of the main hole and used to update the orientation of stratigraphic well tops along the well trajectory. 3D surfaces were then created and projected from the main hole to the sidetrack to check for structural conformity. One of the projected surfaces from the main hole matched the expected depth of the well top in the sidetrack but two were offset due to the possible presence of a fault. This was confirmed by parallel evaluation of the azimuthal sonic imaging data acquired in the main hole that showed an abrupt change in the relative dip of reflectors above and below the possible fault plane using the 3D STC and ray tracing. Dip patterns from both wells showed a drag effect around the offset well tops, further confirming the presence of a fault. A comparison of the acquired borehole images pinpointed the depth and orientation of the fault cutting both wells to explain the depth offset of the projected 3D well top surfaces.
Duarte, Sandra Buzini (SimWorx Engineering R&D) | De Jesus, Candida Menezes (SimWorx Engineering R&D) | Da Silva, Viviane Farroco (SimWorx Engineering R&D) | Arouca Sobreira, Matheus Cafaro (SimWorx Engineering R&D) | Cristofaro, Raphael Agostin Leite (SimWorx Engineering R&D) | de Lima, Larissa (Petrobras) | De Mello E Silva, Fernando Gomes (Petrobras) | Marques De Sa, Carlos Henrique (Petrobras) | De Oliveira Berto, Flavio Marcos (Petrobras) | Backheuser, Yeda (Petrobras) | Loureiro, Sebastiao De Andrade (Petrobras) | De Almeida Waldmann, Alex T. (Petrobras) | Fioriti, Lenita De Souza (Petrobras)
Severe fluid losses while drilling carbonate reservoirs have considerably increased well construction time and costs. Such extra expenses are mainly related to wasted time while struggling against such losses, material costs and its delivery/availability logistics. Besides the economic impact, severe circulation losses have HSE (Health, Safety & Environment) impacts since there is a risk of losing well control when losses can evolve to hydrocarbon inflow and simultaneous loss and gain issues. Extreme situations may lead to temporary or even definitive well abandonment.
Fluid loss predictions are usually performed by a specialist with knowledge of the geological model and the drilling history of the field. Such approach has proved to be a hard task, with limited success, especially in the pre-salt carbonate reservoirs due to their high structural and facies heterogeneity.
This study aims to improve prediction (as compared to the conventional expert-based approach) of the geological structures that might lead to severe fluid losses with impact on well construction costs by focusing on uncertainty reduction of critical resources allocation, such as Managed Pressure Drilling (MPD) and loss control materials (LCM).
Artificial Intelligence (AI) techniques have proved to be useful with high success rate in complex problem solving in many industrial segments. The focus of the present study was to search for AI algorithms to correlate seismic attributes, well logs, fluid loss occurrences and information from geologic and reservoir flow models. A pilot area comprising 38 wells drilled in Santos Basin (Brazil) was chosen for the present analysis. The first step was to use this data set to map the search space of the algorithms, i.e., to identify the critical intervals for severe losses. Information gain tests related to the fluid loss rate (dependent variable) were performed aiming to identify the most relevant independent variables for the case of severe losses prediction and to discard the ones with minor contribution.
Among the tested classifiers, an ensemble of Naive Bayes & Perceptron Neural Network had the best performance at predicting severe fluid losses for the pilot area. A global hit rate of 84% was achieved for metrics evaluated under a well-based standpoint. Blind test with 11 wells (from a different set) returned 82% of global hit rate. These results are considered superior to the ones obtained through the conventional approach. It is important to mention that due to the great uncertainty of the related variables, the output cannot be more accurate than the precision of the original data employed.
These results show a great potential for the use of AI techniques on severe fluid losses prediction in pre-salt carbonates. Therefore, the AI approach is being incorporated as a new tool to support the field experts to improve the performance of the predictions.
Shetty, Sushil (Schlumberger) | Liang, Lin (Schlumberger) | Zhan, Qiwei (Duke University) | Simoes, Vanessa (Schlumberger) | Canesin, Fábio (Schlumberger) | Boyd, Austin (Schlumberger) | Zeroug, Smaine (Schlumberger) | Sinha, Bikash (Schlumberger) | Habashy, Tarek (Schlumberger) | Singhal, Manu (Shell) | Guedes, Ana Beatriz (Shell) | Amorim, Claudia (Shell) | Abbots, Frances (Shell)
We present a new multiphysics inversion for petrophysical analysis of formations containing interspersed laminations of shale and sandstone or shale and carbonate. The inversion builds a petrophysical formation model consistent with triaxial induction data; sonic data such as shear, compressional, and Stoneley waves; and nuclear density data. All data were acquired in a vertical well drilled with oil-based mud. Lamination thickness is inch-scale or less, much thinner than the 1- to 2-ft vertical resolution of the sonic and triaxial induction tools.
The inversion is underpinned by new effective medium models relating petrophysical properties of the laminated formation to effective transverse isotropic (TI) elastic and electromagnetic properties. The effective medium models share a common microstructural geometry parameterized by shale and sand pore shapes. Consequently, shale TI parameters are estimated in the inversion and do not require manual calibration from logs or cores.
We design a robust inversion that considers sensitivity of data to different model parameters. The inversion outputs porosity and radial distribution of fluid saturations at each log depth. The inversion also outputs effective pore shapes for shale and sand laminations at each depth. The petrophysical properties and pore shapes provide corresponding radial distributions of TI elastic and electromagnetic properties (e.g., Thomsen parameters and resistivity anisotropy) for the laminations through the effective medium models.
We demonstrate the inversion on real data from a gas well drilled with oil-based mud in an interval with thin laminations of shale and sand identified from resistivity image logs and nuclear magnetic resonance data. The inversion shows four times more hydrocarbon saturation in the sand laminations relative to isotropic model that ignores the laminated structure. Furthermore, invasion of oil-phase filtrate into gas-bearing sand laminations is clearly elucidated and improves the porosity estimate. The sand and shale pore shapes are consistent with observations and provide TI properties of the laminations at each depth. The new inversion is therefore valuable for characterizing lamination-scale properties to guide formation evaluation decisions in complex environments.
Dernaika, Moustafa (Ingrain, Halliburton) | Al Mansoori, Maisoon (ADNOC Onshore) | Singh, Maniesh (ADNOC Onshore) | Al Dayyani, Taha (ADNOC Onshore) | Kalam, Zubair (ADNOC Onshore) | Bhakta, Ritesh (Formerly with Ingrain, Halliburton) | Koronfol, Safouh (Ingrain, Halliburton) | Uddin, Yasir Naseer (Ingrain, Halliburton)
Most carbonate reservoirs are characterized by multiple-porosity systems that impart petrophysical heterogeneity to the gross reservoir interval. This heterogeneity complicates the task of reservoir description and thus necessitates the establishment of accurate and detailed understanding of the geological heterogeneities and their impact on petrophysics and reservoir engineering.
One of the fundamental input parameters into reservoir models is permeability. The challenge would be to select appropriate samples that represent reservoir heterogeneity for accurate acquisition of vertical to horizontal permeability Kv/Kh data.
In an unpublished work, thousands of plug permeability measurements were performed to obtain Kv/Kh ratios across a large carbonate field in the Middle East. The results were largely influenced by reservoir heterogeneity and yielded large Kv/Kh ratios (greater than unity). Such data would need to be acquired on the same rock volumes for proper Kv/Kh ratios.
In this work, permeability measurements were investigated using digital and conventional techniques to determine the effect of heterogeneity. Detailed thin-section descriptions and mercury injection capillary pressure (MICP) tests were used to understand the different rock types. Advanced three-dimensional (3D) X-ray CT imaging was acquired at multiple scales for detailed digital rock characterization. Permeability was computed directly on the 3D images by the lattice Boltzmann methodology and upscaled to the plug and whole-core levels. Permeability varied largely among different scales/locations and was clearly linked to complex geological features in the rock samples. Integration of the CT images and thin-section photomicrographs provided geological variation in 3D and showed that permeability was influenced by macroscale heterogeneity that may only be examined through multiscale imaging. Larger volume samples were vital in capturing the reservoir heterogeneity, which gave proper Kv/Kh ratios less than unity. Our understanding of the comparisons among different scales will be crucial for upscaling laboratory-measured properties to grid-block scale in reservoir geological models.
Editor’s comment: This article is part of a series of short “tutorial-like” notes styled to mentor users of digital well logs in becoming confident practitioners of petrophysics.
At the close of Shaly-Sand Tutorial Part 2, I implied, or sort-of promised that this final part would provide some guidance or at least guidelines on the application of my favorite models (equations, transforms, or any of your personal preference names) used to calculate water saturation from a combination of electrical conductivity (i.e., inverse of resistivity), total porosity, formation water salinity, fitting parameters based on rock types (m*, n*), and of course, formation temperature. I need not remind you that all of these properties/parameters listed above can and do vary rapidly with depth (vertically) and transversely (horizontally) across a given reservoir.
I once requested a second core in a large reservoir which was turned down by the Division Engineering Manager using the following reason: “Easy, I am really doing you a favor because if the second core and its analyses are different than the first, you will spend a lot of time understanding the differences, and I know you will be back with a request for a third core to help in this investigation.” This fellow turned a deaf ear to all arguments that I fully expected that each core would be different and only through such sorts of studies could we ever improve our ability to predict recovery efficiency and sweep. “Anyway, that’s not the job assignment of a petrophysical engineer. Leave that job to the experts.” To me it was obvious that this manager did not know that he did not know the influence that geology and petrophysics played in selecting the correct earth model to simulate with mathematical models.
Seleznev, Nikita (Schlumberger-Doll Research Center) | Hou, Chang-Yu (Schlumberger-Doll Research Center) | Freed, Denise (Schlumberger-Doll Research Center) | Habashy, Tarek (Schlumberger-Doll Research Center) | Feng, Ling (Schlumberger-Doll Research Center) | Fellah, Kamilla (Schlumberger-Doll Research Center) | Xu, Guangping (Schlumberger-Doll Research Center and Sandia National Laboratories) | Nadeev, Alexander (Schlumberger Reservoir Laboratories)
Electromagnetic (EM) formation evaluation currently relies on low-frequency resistivity and high-frequency dielectric measurements that are typically not interpreted jointly. In consideration that formation EM responses in different frequency ranges are controlled by different physical phenomena, analysis of a wideband EM response can provide new and complementary sensitivities to formation petrophysical parameters.
We established a wideband rock model to describe the dielectric response of well-sorted clean sandstones in the spectral induced polarization (SIP) frequency range and the dielectric-dispersion frequency range. The model is based on a differential effective-medium approach that accounts for both the Maxwell-Wagner interfacial polarization related to the rock texture and the electric double-layer polarization due to the presence of charged grains. We aim to use a minimal number of parameters in our model to capture the essential dielectric properties in the frequency ranges of interest.
The SIP and dielectric-dispersion spectra were measured on a collection of quarried clean sandstones saturated with brines providing wideband core data. We analyzed these wideband data by applying the rock model simultaneously to the SIP and dielectric spectra. The joint wideband data inversion enabled the estimation of five formation parameters: water-filled porosity, water salinity, cation exchange capacity, dominant grain size, and cementation exponent. The ability to invert for this broad set of formation parameters provides a comprehensive characterization that is unattainable with currently practiced methods. Moreover, when the modeled and measured responses are compatible, the joint wideband inversion of SIP and dielectric-dispersion spectra potentially eliminates interpretation uncertainties if some parameters are independently provided as input.
Application of nanoparticles in the subsurface typically requires the use of surface coatings to maintain stability in dispersion and to provide particular functionality. However, the presence of surface coatings may hinder or mask properties of the bare nanoparticle cores, which may be a concern in nuclear magnetic resonance (NMR) applications. In this study, we used different amounts of 3-aminopropyltriethoxysilane (APTES) coating on Fe3O4 magnetic nanoparticles (A-MNPs). We measured the longitudinal relaxation time (T1) values of those A-MNPs suspensions, and computed and compared the surface relaxivities of A-MNPs with different amounts of APTES coating. Our results showed that when the mass percentage of APTES coating increased from 1.60 to 4.22 wt%, the A-MNPs’ surface relaxivity decreased by 26.1%. To determine the surface relaxation mechanism(s), we also used various volume fractions of D2O to dilute A-MNP dispersions to two concentrations: 0.01 and 0.002 g/L Fe. In the final mixtures, the volume fractions of D2O were fixed as 0-, 30-, 50-, and 70-vol%. The NMR measurements indicated that, at relatively high Fe concentration (0.01 g/L), electron-proton interaction dominates surface relaxation, and the hydrogen atoms in the APTES did not significantly alter the surface relaxation mechanism of the nanoparticles. At a lower Fe concentration (0.002 g/L), proton-proton relaxation, due to the APTES, also played a role in the overall relaxation mechanism on nanoparticle surfaces. A-MNPs with more APTES coating showed lower apparent surface relaxivities with higher D2O volume fractions in the mixture, indicating a greater amount of proton-proton relaxation on the nanoparticle surfaces.
With superior magnetic properties, nanoscale dimensions and nontoxic characteristics, iron oxide nanoparticles are of high interest in nanoscience and nanotechnology. As superparamagnetic nanoparticles (MNPs), Fe3O4 nanoparticles have been widely applied in biomedical areas, such as targeted drug delivery (Chertok et al., 2008), tissue repair (Jordan et al., 2001) and magnetic resonance imaging (MRI) techniques (Babes et al., 1999).
Craddock, Paul R. (Schlumberger-Doll Research Center) | Mossé, Laurent (Schlumberger) | Bernhardt, Carolina (YPF S.A.) | Ortiz, Alberto C. (YPF S.A.) | Tomassini, Federico Gonzalez (YPF S.A.) | Pirie, Iain C. (Schlumberger) | Saldungaray, Pablo (Schlumberger) | Pomerantz, Andrew E. (Schlumberger-Doll Research Center)
The determination of accurate density porosity requires an accurate matrix density. This determination is challenging in organic-rich shale using downhole logs because of the presence of insoluble sedimentary organic matter (“kerogen”) that is part of the solid matrix but has log-response characteristics more similar to those of pore fluids. Methods other than logs used to determine shale matrix density from drill core or cuttings, such as gas pycnometry, rely on remote laboratory services and may not be representative due to microcracks. This study describes a novel approach to quantify porosity in organic-rich shale by the integration of fast wellsite measurements of cuttings and a bulk-density log. The cuttings analysis uses diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), which provides an accurate estimate of matrix density explicitly including kerogen as part of the matrix. Matrix density is then used to estimate porosity per the well-known density-porosity relationship. The method is enabled because the DRIFTS analysis explicitly solves for both minerals and kerogen as components of the shale matrix, separate from that of pore fluids. It is demonstrated that kerogen density is a critical parameter in the determination of matrix density, and its determination is an integral part of the DRIFTS interpretation. Kerogen density is not measurable using traditional logging methods and is otherwise only obtainable using time-consuming laboratory procedures. The DRIFTS technique is advantageous because it requires minimal sample preparation and footprint, can be run at the wellsite, is rapid enough to keep pace with typical drilling rates, and can be performed in oil- and water-based muds. The method can be run in horizontal wells because cuttings are always available, and because the combined depth-resolution of the cuttings and the logs is typically much finer than the lateral interval over which formation properties vary. The integration of cuttings and logs provides a depth-by-depth estimate of shale porosity at the wellsite that is otherwise not obtainable from either basic logs or cuttings individually.
While permeability modeling follows a well-established approach in converting laboratory properties to subsurface conditions, ambiguity remains over the approach to be followed by laboratory-acquired capillary pressures (under ambient conditions, like most mercury injection capillary pressures (MICP) measurements). One approach (developed by the earlier work of Juhasz) recommends that capillary pressures be stress corrected (prior to modeling) according to a correlation. Another approach suggests the saturation-height model (SHM) be built with ambient measurements that when supplied with corrected properties (porosity and permeability) would generate in-situ saturations.
The effect of stress correction applied to porosity and permeability data (as part of routine core analysis (RCA) is not easily compared against the capillary pressure correction, potentially leading to inconsistencies.
The work presented here uses a recent methodology that aims at ensuring consistency between permeability and SHMs to provide guidance on the best approach to be followed in the process of building a SHM. The MICP or SHM carries an intrinsic permeability that can be compared to the permeability model. The results show that significant inconsistency can occur between the porosity-permeability data (a reliable, well-controlled and measurable property under stress) on one hand, and the MICP-/SHM-inferred permeability on the other.
The conclusion is that the most robust dataset for preparing the SHM is under the same conditions under which the MICP and capillary pressure (Pc) data have been acquired. When these data have been acquired under ambient conditions and the resulting model has stressed porosity and permeability as inputs, the SHM will predict the correct stressed entry pressures. The findings are validated against a dataset where the capillary pressures acquired under both ambient and stress conditions.
Saturation-height models (SHM) combined with fundamental rock properties (porosity and permeability) are the basis for a realistic reservoir model. In contrast to porosity and (single-phase) permeability that are rock properties, SHMs are the result of fluid-rock interaction.