Log Interpretation Research Articles

Total organic carbon variability of lower Jurassic successions in the Mandawa Basin, SE Tanzania

Over many decades, an understanding of total organic carbon (TOC) variability in various sedimentary basins has been constrained by the availability of suitable samples for geochemical analysis. Many attempts rely on selected sidewall core or cuttings samples, which do not provide a continuous lithological record of the succession. As a result, TOC variability as a response to facies evolution is often underestimated. An advancement in TOC evaluation from well logging techniques often has permitted a continuous analysis of source rock intervals, while eliminating issues associated with obtaining discrete physical samples. Geochemical and wireline geophysical logs investigation here aimed at constraining TOC variability and depositional conditions prevailing during the early Jurassic in the Tanzanian Mandawa Basin (Mbuo, Nondwa and Mihambia Formations), which to date are poorly documented. TOC was deduced from both well cuttings samples and wire log data. Depositional conditions were defined via the use of spectral Gamma-Ray (Th/U, U) logging and lithofacies data. Sediment composition and TOC from cuttings were compared to illustrate how organic matter distribution is influenced by changes in sediment type. Complex mineral composition of heterolithic facies dominated log-response, complicated log interpretation, and thus limited the effectiveness of well logging techniques to predict TOC. Cuttings data reveal three organic-rich intervals, i) at the transition of Upper Mbuo to basal Nondwa Formation with the highest TOC of 3.5 (wt%), ii) in the Nondwa Formation with a TOC of 2 (wt%), and iii) in the Mihambia Formation with a TOC of 2.8 (wt%). An inverse linear relationship between Th/U and TOC signifies relatively shallow deposition in a well-mixed water column under oxic conditions. Intervals with higher TOC imply organic matter preservation under suboxic conditions upon two marine incursions in the Mihambia and Nondwa Formations, whereas a limnic environment prevailed at the Upper Mbuo to Nondwa transition. Inverse trends in TOC versus clastic components, a positive linear relationship between TOC and carbonates, and higher amount of clastics (>57 wt%) in the Mbuo and Mihambia Formations suggest a dominantly clastic sedimentary system with relatively stable carbonate and organic matter input in proximity to the source area. We, therefore, argue that the organic rich intervals observed in the Upper Mbuo, Nondwa and Mihambia Formations were deposited during periods of higher organic productivity driven by a rise in eustatic sea level, and coincide with a decline in terrigenous input from the hinterland.

SVR-based weighted processing method for electrical imaging logging in oil-based mud

Electrical imaging logging in high-resistivity oil-based mud (OBM) has been one of the hotspots in petroleum exploration and development recently. Compared with the traditional electrical imaging logging in water-based mud (WBM), the core issue restricting the application of electrical imaging logging in OBM is that the mud-layer with high impedance prevents the emitted current from flowing into the formation and seriously disturbs the measurement of formation properties. Here, a SVR (support vector machine regression)-based weighted processing method is proposed to resolve the intractable problem and three key parameters are obtained simultaneously. First, the equivalent circuit model of electrical imaging logging in OBM is established and the drawback especially occurring in low-resistivity formation is analyzed. Then, a weighted coefficient applied to the mud-layer impedance is introduced for the quantitative calculation of formation resistivity, and the change of relative formation permittivity and mud-layer thickness can also be depicted with weighted processing. Next, a SVR model is established to inverse the weighted coefficient and then figure out the inversed parameters including the formation resistivity, the relative formation permittivity, and the mud-layer thickness. A stochastic parameter model is adopted to verify the validity of SVR model. Finally, two deviated formation models further test the advantage and robustness of SVR model. The results demonstrate that the SVR-based weighted processing method possesses a terrific performance and all the three parameters can be calculated simultaneously regardless of the borehole wall condition. This study will provide a new approach to the data processing and interpretation of electrical imaging logging in OBM.

Identification of Milankovitch Cycles and Calculation of Net Primary Productivity of Paleo‐peatlands using Geophysical Logs of Coal Seams

AbstractIndividual coal seams formed in paleo‐peatlands represent sustained periods of terrestrial carbon accumulation and a key environmental indicator attributed to this record is the rate of carbon accumulation. Determining the rate of carbon accumulation requires a measure of time contained within the coal. This study aimed to determine this rate via the identification of Milankovitch orbital cycles in the coals. The geophysical log is an ideal paleoclimate proxy and has been widely used in the study of sedimentary records using spectral analysis. Spectral analyses of geophysical log from thick coal seams can be used to identify the Milankovitch cycles and to calculate the period of the coal deposition. By considering the carbon loss during coalification, the long‐term average carbon accumulation rate and net primary productivity (NPP) of paleo‐peatlands in coal seams can be obtained. This review paper presents the procedures of analysis, assessment of results and interpretation of geophysical logs in determining the NPP of paleo‐peatlands.

The Mishrif Reservoir Characteristics Utilizing Well Log Data Interpretation in the Fauqi Oilfield in Maysan, Southern Iraq

The Mishrif Formation was deemed the main oil reservoir in the Fauqi oilfield, southern Iraq. This paper aims to predict characteristic properties of the Mishrif Formation depending on the interpretation of well logs data. The data were collected from five wells of log images which had been digitized with Didger software version 5. 12. 1762 to be analyzed and interpreted using interactive petrophysics 3.5.1.2 software. The main lithology of the Mishrif Formation in this region was determined using standard (M-N xplot) based on the integration of density-neutron logs. The results confirmed that the Mishrif Formation is mainly composed of limestone units. A Gamma-ray log was used to estimate the shale volume in the Mishrif Formation depending on old rock function. As a result, shale volume relative to the bulk volume was about 19%, and an increasing volume was noticed in the upper reservoir, reaching the highest in the MA and MB11 units. The Mishrif Formation porosity was calculated using the interpretation of neutron, density, and sonic logs. The results showed that the Mishrif Formation is distinctive with low to medium porosity values (4.8 to 17%). The calculation of cores manifested the dominance of secondary porosity in the MB21 unit of the formation in comparison to the other stratigraphic units in the studied area. These results led to a the fact that the MB21 unit was more exposed to dissolution diagenetic processes in the south dome, which provided a good permeability integrated with the medium effective porosity, puttinh it as the main reservoir unit.

Assisted Cement Log Interpretation Using Machine Learning

Summary The Assisted Cement Log Interpretation Project has used machine learning (ML) to create a tool that interprets cement logs by predicting a predefined set of annular condition codes used in the cement log interpretation process. The development of a cement log interpretation tool speeds up the log interpretation process and enables expert knowledge to be efficiently shared when training new professionals. By using high-quality and consistent training data sets, the project has trained a model that will support unbiased and consistent interpretations over time. The tool consists of a training and a prediction tool integrated with cased-hole logging interpretation software. By containerizing the code using an “API First” design principle (API: application programming interface), the applicability of this add-on tool is broad. The ML model is trained using selected and engineered features from cement logs, and the tool predicts an annular condition code according to the cement classification system for each depth segment in the log. The interpreters can easily fetch a complete cement log interpretation prediction for the log and use that as a template for their final interpretation. The ML model can easily be retrained with new data sets to improve accuracy even further. To improve cement log interpretation consistency in the industry, the code will be made available as open source.

A systematic review of empirical studies using log data from open‐ended learning environments to measure science and engineering practices

AbstractTechnology‐based, open‐ended learning environments (OELEs) can capture detailed information of students' interactions as they work through a task or solve a problem embedded in the environment. This information, in the form of log data, has the potential to provide important insights about the practices adopted by students for scientific inquiry and problem solving. How to parse and analyse the log data to reveal evidence of multifaceted constructs like inquiry and problem solving holds the key to making interactive learning environments useful for assessing students' higher‐order competencies. In this paper, we present a systematic review of studies that used log data generated in OELEs to describe, model and assess scientific inquiry and problem solving. We identify and analyse 70 conference proceedings and journal papers published between 2012 and 2021. Our results reveal large variations in OELE and task characteristics, approaches used to extract features from log data and interpretation models used to link features to target constructs. While the educational data mining and learning analytics communities have made progress in leveraging log data to model inquiry and problem solving, multiple barriers still exist to hamper the production of representative, reproducible and generalizable results. Based on the trends identified, we lay out a set of recommendations pertaining to key aspects of the workflow that we believe will help the field develop more systematic approaches to designing and using OELEs for studying how students engage in inquiry and problem‐solving practices. Practitioner notesWhat is already known about this topic Research has shown that technology‐based, open‐ended learning environments (OELEs) that collect users' interaction data are potentially useful tools for engaging students in practice‐based STEM learning. More work is needed to identify generalizable principles of how to design OELE tasks to support student learning and how to analyse the log data to assess student performance. What this paper adds We identified multiple barriers to the production of sufficiently generalizable and robust results to inform practice, with respect to: (1) the design characteristics of the OELE‐based tasks, (2) the target competencies measured, (3) the approaches and techniques used to extract features from log files and (4) the models used to link features to the competencies. Based on this analysis, we can provide a series of specific recommendations to inform future research and facilitate the generalizability and interpretability of results: Making the data available in open‐access repositories, similar to the PISA tasks, for easy access and sharing. Defining target practices more precisely to better align task design with target practices and to facilitate between‐study comparisons. More systematic evaluation of OELE and task designs to improve the psychometric properties of OELE‐based measurement tasks and analysis processes. Focusing more on internal and external validation of both feature generation processes and statistical models, for example with data from different samples or by systematically varying the analysis methods. Implications for practice and/or policy Using the framework of evidence‐centered assessment design, we have identified relevant criteria for organizing and evaluating the diverse body of empirical studies on the topic and that policy makers and practitioners can use for their own further examinations. This paper identifies promising research and development areas on the measurement and assessment of higher‐order constructs with process data from OELE‐based tasks that government agencies and foundations can support. Researchers, technologists and assessment designers might find useful the insights and recommendations for how OELEs can enhance science assessment through thoughtful integration of learning theories, task design and data mining techniques.

Igneous rocks lithology identification with deep forest: Case study from eastern sag, Liaohe basin

At present, volcanic reservoir exploration has been carried out widely in the world. In the eastern sag of Liaohe Basin, different volcanic eruption periods lead to complex facies and lithology of igneous rocks, which brings difficulties to logging interpretation. Therefore, a method to accurately identify igneous rock lithology is urgently needed. In this paper, the scheme for identifying igneous rock lithology uses multiple deep forest classifiers. To select a practical training set, 944 conventional logs (AC, DEN, GR, RLLD, and CNL) with lithologic labels were randomly selected from the target area. Compared with the test sets selected from other formations, the accuracy of the scheme reached 99.96%. The scheme was applied to three wells in the same area, and excellent results were achieved when the predicted lithology was compared with the lithology profile and thin section data.

Characteristics and Seismic Identification Mode of Ultra-Deep Carbonate Fault-Controlled Reservoir in Northwest China

Ordovician carbonate reservoirs in the Shunbei area of the Tarim Basin are characterized by large burial depth, small vertical fault displacement of the strike-slip fault zone, small karst development scale, strong reservoir heterogeneity, complex and diverse seismic reflection types, which lead to unclear seismic response characteristics of fault zone and reservoir, and further restrict the fine description of ultra-deep (>7300 m) fault controlled reservoirs. On the basis of high-precision 3D seismic data analysis, combined with drilling, logging, and comprehensive geological interpretation data in the Shunbei area, seismic geological models of different types of reservoirs and fault zones are established. The changes in seismic response characteristics of fault zones and reservoirs are simulated and analyzed by changing reservoir size, scale, fault and fracture cavity combination modes, as well as other factors. On this basis, two seismic identification models of fault-controlled reservoirs in the Shunbei area are summarized and established. The seismic identification mode of seismic reflection marker wave of fault-controlled reservoir is mainly “fault seismic response characteristics & weak reflection” and red trough anomaly under T74. The internal seismic identification modes are mainly “fault seismic response characteristics & random reflection”, “fault seismic response characteristics & beading”, “fault seismic response characteristics & linear weak reflection”, and “fault seismic response characteristics & random reflection”. Among them, except for “fault seismic response characteristics & random reflection”, the coincidence rate of reservoir prediction of other three types of seismic response characteristics is more than 70%. The coincidence rate of the other three types of seismic response characteristics of a reservoir is more than 70%. The research results provide a reference for the prediction and description of ultra-deep carbonate fault-controlled reservoirs.

The lithostratigraphical context of the English Chalk Rock (Turonian)

Correlations of borehole geophysical logs in the middle and upper Turonian Chalk Group are used to consider recent proposals for a revision in understanding of a unit of hardgrounds (Chalk Rock) and associated stratigraphy developed across parts of southern England. Along the northern edge of the London and Wessex basins, geophysical logs reveal a laterally continuous framework of correlatable inflection patterns in the New Pit Chalk, with the package of sediment immediately below the Chalk Rock showing a trend of lateral thinning and increasingly condensed sedimentation westwards into areas where the oldest of the Chalk Rock Hardgrounds (Ogbourne Hardground) is present. However, apart from local absence of the Glynde Marls Complex near the top of the New Pit Chalk, there is no evidence for the presence of a major erosion event. This questions recent interpretations of microcrinoid data, reportedly showing that the Ogbourne Hardground lithifies a stratigraphical level in the lower part of the New Pit Chalk, with the middle and upper parts of this unit corresponding with a hiatus and related short-lived globally significant sea level fall. Macrofossil biostratigraphy supports the geophysical log interpretations, with evidence of both younger and older parts of the New Pit Chalk below the Ogbourne Hardground. The data are consistent with the Ogbourne Hardground in Wiltshire and Berkshire representing a highly condensed equivalent of thickened nodular chalk fabrics at the base of the Lewes Nodular Chalk in the eastern Chilterns.

Lithology identification using principal component analysis and particle swarm optimization fuzzy decision tree

Lithology identification using geophysical log information is vital for log interpretation and reservoir evaluation. As a result of the highly similar features for log curves that characterize complex lithology, there is significant information redundancy regarding the process of lithology identification. In addition, as a result of the highly nonlinear characteristics of log curves, the mapping relationship with lithology has certain ambiguities and uncertainties, which affect the lithology prediction results. Combining principal component analysis (PCA) and the fuzzy decision tree (FDT) model, we propose a new intelligent lithology identification method that is capable of effectively solving these problems well. However, because of the inaccuracy for empirically set parameters, an adaptive fuzzy decision tree algorithm based on particle swarm optimization (PSO-FDT) was proposed after analyzing the main features of the fuzzy decision tree and using an improved particle swarm optimization (PSO) algorithm to determine the relevant parameters. Compared with the FDT algorithm which determines parameter values empirically, the performance of the PSO-FDT has been significantly improved. Finally, the proposed PSO-FDT model was verified using test data. Experiments confirm that the proposed model is more effective than other lithology identification models. The identification accuracy for all lithologies was equal to or greater than that of the other methods. In addition, the overall accuracy was improved by at least 9.71%.

Factors controlling sonic velocity in dolostones

Although the impact of rock texture and crystal size on porosity and permeability of dolostones is well documented, the impact of such factors on the sonic velocity of these sedimentary rocks remains poorly understood. In this study, 100 samples from five formations outcropping along the Arabian Plate, were analyzed to investigate the main factors controlling the sonic velocity in dolostones. The study integrates results obtained using different techniques including: thin-section petrography, SEM, XRD, digital image analysis, porosity and permeability measurements, velocity measurements, in addition to rock physics modeling. Results indicate that porosity is a major factor in controlling the sonic velocity in dolostone where the porosity-velocity relationship shows a determination coefficient (R2) of 0.82. The scatter in the porosity-velocity relationship can be mainly attributed to variations in the dolostone texture (fabric preserving, FP, versus non-fabric preserving, NFP) and the associated pore types. FP dolostones (characterized by macro interparticle, moldic and shelter pores inherited from the precursor limestone) show higher velocities than NFP dolostones (dominated by intercrystalline pores) at a given porosity. The experimental porosity-velocity data, and the rock texture and pore types observations are consistent with rock-physics modeling incorporating variable equivalent pore aspect ratio (EPAR) values in the differential effective medium (DEM) theory. FP dolostones display higher EPAR values (>0.15) compared to NFP dolostones; a criterion that can be used to differentiate between the two dolostone textures. When classifying data based on texture (FP versus NFP), we observe strong porosity-impedance relationship (R2 > 0.95) for both the studied samples, and literature datasets (regional and global) following a consistent distinct trend for each texture. This implies a universality of the porosity-impedance trends observed here, which when combined with the presented rock physics analysis, can inform the modeling and quantitative interpretation of sonic logs and seismic data in dolostone strata.

Simulation of flow characteristics and development of permeability model in fractured-vuggy carbonate reservoir

Fractured-vuggy carbonate reservoirs have complex pore structure, significant heterogeneity, and irregular permeability-porosity relation, which make existing evaluation methods unable to accurately calculate the permeability. To investigate the flow characteristics of carbonate rocks with triple-porosity, 3D digital rock models were constructed from X-ray computed tomography (CT) scanning images of carbonate cores collected in the field, and the physical properties of rocks were obtained by combining digital rock analysis (DRA) and nuclear magnetic resonance (NMR) experiment. The effects of fractures and vugs on the permeability of carbonate rocks was investigated using the Stokes-Brinkman method, and then a new fractured-vuggy carbonate permeability model was developed by artificially adding vugs and fractures. The results indicate that the carbonate pores can be divided into matrix pores, vugs, and fractures according to their sizes by combining digital rock analysis and T2 spectrums. The transport resistance is effectively reduced by fractures and vugs, which leads to the change of the flow path. For tight carbonate rocks, the dominant flow path formed through fractures is a crucial factor affecting permeability. On the contrary, unconnected vugs have less effect on permeability. Finally, the new permeability model was applied to the logging interpretation of field data. The new model is in better agreement with the permeability analysis of field cores, which greatly improves the permeability calculation accuracy of the triple-porosity reservoir in the study area.

High-quality fracture network mapping using high frequency logging while drilling (LWD) data: MSEEL case study

The Marcellus Shale and Energy Environmental Laboratory (MSEEL) provides a comprehensive dataset and field tests that can be used to study the significance of preexisting natural fractures in different subsurface engineering problems such as the effectiveness of the stimulation of an unconventional reservoir, optimized geothermal fluids movement and integrity of the CO2 storage site. Conventionally natural fracture intensity is obtained using sonic and micro-resistivity imaging logs. However, these techniques significantly suffer from two major deficiencies: Human bias in log interpretation and extremely long interpretation time. These two deficiencies are well-recognized in the industry; however, no standard procedures exist to address them. In this study, a new automated machine learning workflow (AMLW) is introduced that uses the LWD high-resolution acceleration data along the horizontal laterals to predict the natural fracture intensities originally obtained using sonic and micro-resistivity imaging. The accuracy and robustness of the new workflow to predict the near wellbore fracture intensities are tested using both regression and classification approaches. Both the regression and classification approaches were able to predict the fracture intensities with high accuracy (average Mean Squared Error of 0.0085 for regression and average accuracy of 0.94 in the confusion matrix for classification). We have shown that only 10%–15% of the labeled resistivity image log is required for training and validation of the machine-learning model. The Automated workflow resulted in K-Neighbors Regressor and classifier algorithms as the best algorithms with a 52.74% and 139.3% improvement in comparison to the Gradient Boosting Regression algorithm (i.e. the fifth best algorithm).

Petrophysical Regression regarding Porosity, Permeability, and Water Saturation Driven by Logging-Based Ensemble and Transfer Learnings: A Case Study of Sandy-Mud Reservoirs

From a general review, most petrophysical models applied for the conventional logging interpretation imply that porosity, permeability, or water saturation mathematically have a linear or nonlinear relationship with well logs, and then arguing the prediction of these three parameters actually is accessible under a regression of logging sequences. Based on this knowledge, ensemble learning technique, partially developed for fitting problems, can be regarded as a solution. Light gradient boosting machine (LightGBM) is proved as one representative of the state-of-the-art ensemble learning, thus adopted as a potential solver to predict three target reservoir characters. To guarantee the predicting quality of LightGBM, continuous restricted Boltzmann machine (CRBM) and Bayesian optimization (Bayes) are introduced as assistants to enhance the significance of input logs and the setting of employed hyperparameters. Thereby, a new hybrid predictor, named CRBM-Bayes-LightGBM, is proposed for the prediction task. To validate the working performance of the proposed predictor, the basic data derived from the member of Chang 8, Jiyuan Oilfield, Ordos Basin, Northern China, is collected to launch the corresponding experiments. Additionally, to highlight the validating effect, three sophisticated predictors, including k-nearest neighbors (KNN), support vector regression (SVR), and random forest (RF), are introduced as competitors to implement a contrast. Since ensemble learning models universally will cause an underfitting issue when dealing with a small-volumetric dataset, transfer learning in this circumstance will be employed as an aided technique for the core predictor to achieve a satisfactory prediction. Then, three experiments are purposefully designed for four validated predictors, and given a comprehensive analysis of the gained experimented results, two critical points are concluded: (1) compared to three competitors, LightGBM-cored predictor has capability to produce more reliable predicted results, and the reliability can be further improved under a usage of more learning samples; (2) transfer learning is really functional in completing a satisfactory prediction for a small-volumetric dataset and furthermore has access to perform better when serving for the proposed predictor. Consequently, CRBM-Bayes-LightGBM combined with transfer learning is solidly demonstrated by a stronger capability and an expected robustness on the prediction of porosity, permeability, and water saturation, which then clarify that the proposed predictor can be viewed as a preferential selection when geologists, geophysicists, or petrophysicists need to finalize a characterization of sandy-mud reservoirs.

Correction Method for Logging Curves in Clay-Rich Tight Glutenite Reservoir: Upper Wuerhe Formation in Mahu Oilfield, China

Mahu Oilfield is the largest tight glutenite oilfield in the world, and the upper Wuerhe formation is an important succeeding exploration horizon. However, the upper Wuerhe formation in the Mahu 1 zone has a high clay content, which can lead to serious wellbore collapse. Meanwhile, the horizontal well logging is not corrected. These factors lead to the inconsistency between the logging interpretation results and the oil test results. The interpretation precision of the clay content, water saturation, and porosity, which are crucial to reservoir evaluation, is very low. In this paper, a workflow of logging curve correction using multiple methods is proposed. The multiple linear fitting is used to correct boreholes, and then histogram frequency distribution matching is used to standardize multi-well logging curves. Finally, the optimization method is used to build a volume model based on skeleton analysis, and the results are calibrated with core analysis results. Horizontal well density logs are corrected using adjacent vertical well logs. The interpretation results of clay content, water saturation, and porosity with high precision are obtained. The reservoir interpretation is more in line with the oil test results than the original interpretation. The clay content distribution is more reasonable.

An advanced workflow to compress the uncertainties of stochastic distribution of Bahariya reservoir properties using 3D static modeling: An example from Heba Oil Fields, Western Desert, Egypt

The main objective of this paper is to construct a static model that compress the uncertainties of the stochastic distribution of the reservoir properties of the Bahariya Formation in Heba field, at the northeastern portion of the Western Desert. This model has been constructed through the integration of the interpretations of the eighteen 2D seismic sections and the analysis of well logs data for four wells (HEBA 300X, E.BAH-E−1X, E.BAH-D-1X, and HEBA 10X) drilled in the study area. This set of data was implemented in a harmonic workflow. Structural framework was the first step created on the basis of the seismic and well log interpretations. Model zonation was mainly managed by the marine flooding events took place during the Cenomanian period. The trapping faults position uncertainty has been compressed through the tying of the seismic profiles with the identified fault cuts in the well data. Effective porosity spectrum was broke up into three reservoir qualities. The results showed heterogeneous facies qualities for oil production in specific five zones in the topmost part of the Bahariya Formation. The effective porosity model was generated stochastically considering the normal distribution for each reservoir quality. Water saturation was distributed by two methods; 1) Sequential Gaussian Simulation that was co-simulated by porosity model. 2) Log-based saturation height function for each reservoir quality. This methodology provided as accurate as possible estimates for the volume calculation by quantifying the sensitivity of the important parameters such as oil contact. Additionally, the model was prepared to be used as a front end for dynamic simulation.

Coal seam gas reservoir characterisation based on high-resolution image logs from vertical and horizontal wells: A case study

Drilling horizontal surface-to-inseam (SIS) wells along low permeability coal seams, e.g. <10 mD, is a crucial technology to increase the drainage area for this type of unconventional reservoir. For the first time in the Bowen Basin, a continuous 1900 m long formation resistivity-based high-resolution image log from a SIS well is used to investigate key coal reservoir characteristics including faults, fractures, coal plies, porosity and permeability. This study presents a workflow from image log interpretation to discrete fracture network (DFN) modelling and then to well performance analysis. The interpreted faults from the image logs were used to analyse the uncertainty of fault interpretation from 3D seismic data. The image log was also used in analysing the variation of fracture distribution vertically and horizontally and compared to similar data obtained from vertical wells. All these data were used to understand the variation in productions from three nearby chevron well groups. Results show that the fracture density decreases with increasing Gamma-ray response from the in-seam section, including the intra-seam partings. Regionally, the orientation of fractures and cleats is parallel, sub-parallel or even perpendicular to the regional maximum stress orientation; however, in the study area, the majority of fractures are generally parallel or near parallel to the maximum stress orientation. This study confirmed the trend that the orientation of fractures/cleats rotated from NE to NW from the study area to the north. Because the maximum stress orientation is NNE, this cleat orientation will yield a decrease in coal permeability from south to north. This can explain at least partially the decrease in well performance from south to north. Based on the modelled fractures and their apertures, the upscaled fracture porosity ranged from 0% to 0.01%, with fracture permeability in the x-direction (easting direction) ranging mainly from 1 mD to 10 mD, and with the ratio of permeability in the x-direction to y-direction (northing direction) ranging from 5 to 20.

Direct Multi‐Modal Inversion of Geophysical Logs Using Deep Learning

AbstractGeosteering of wells requires fast interpretation of geophysical logs which is a non‐unique inverse problem. Current work presents a proof‐of‐concept approach to multi‐modal probabilistic inversion of logs using a single evaluation of an artificial deep neural network (DNN). A mixture density DNN (MDN) is trained using the ”multiple‐trajectory‐prediction” loss functions, which avoids mode collapse typical for traditional MDNs, and allows multi‐modal prediction ahead of data. The proposed approach is verified on the real‐time stratigraphic inversion of gamma‐ray logs. The multi‐modal predictor outputs several likely inverse solutions/predictions, providing more accurate and realistic solutions compared to a deterministic regression using a DNN. For these likely stratigraphic curves, the model simultaneously predicts their probabilities, which are implicitly learned from the training geological data. The stratigraphy predictions and their probabilities obtained in milliseconds from the MDN can enable better real‐time decisions under geological uncertainties.

Lithology Logging Recognition Technology Based on GWO-SVM Algorithm

Accurate identification of lithology is the basis and key process of fine logging interpretation and evaluation. However, reservoirs formed by different sedimentary environments and tectonic movements generally have the characteristics of complex and diverse lithology and strong heterogeneity, which brings great difficulty to the identification of reservoir lithology. This paper proposes an automatic identification technology for lithology logging based on the GWO-SVM algorithm model. The technology is actually applied, and the results are compared with the results of the support vector machine cross-validation optimization model, PNN (probabilistic neural network) model, and ELM (extreme learning machine) model. The results show that the GWO-SVM lithology logging recognition model can efficiently solve the lithology recognition and classification problems in complex reservoir analysis and has strong adaptability and higher recognition accuracy.

Deep-Learning-Based Noniterative 2D-Inversion of Unfocused Lateral Logs

Abstract —The work deals with the development of methodological and algorithmic tools for the quantitative interpretation of oil well resistivity logs. We review the results of applying the neural-network-based approach to the inversion of resistivity logs measured at thinly bedded high-contrast environments. The capabilities of the proposed approach are demonstrated by the example of the algorithm for noninterative express inversion of unfocused lateral logs (BKZ). BKZ is the unfocused array logging method widely used in the Commonwealth of Independent States for oil well studies. BKZ logs are known for their complexity: The signals of unfocused gradient probes are highly affected by the medium properties below and above the measuring point. The developed algorithm is based on the inversion of full logs into the parameters of a 2D axisymmetric model of the medium, which allows naturally taking into account the influence of surrounding rocks and borehole conditions. Transition from the “layered” parametrization conventional for BKZ logs interpretation to a quasi-continuous change of properties along the well axis allows extracting meaningful information at every measurement point and constructing high-resolution geoelectric models of the sediments. The noniterative nature of the algorithm provides a high computing efficiency. This opens up the possibility of using the 2D inversion advantages to increase the reliability of the initial express interpretation results. Testing the algorithm on the practical data from West Siberian oilfields has revealed the field of its maximum efficiency, namely, study of impermeable and low-permeability sediments, such as the complex shaly caprocks and bituminous deposits of the Bazhenov Formation. With high-quality input data, the approach is also efficient for studying permeable terrigenous sediments.