Lithofacies Prediction Research Articles

Enhancing Reservoir Characterization With High-Resolution Lithofacies Prediction Using Advanced Feature Engineering and Ensemble Classifiers

Enhancing Reservoir Characterization With High-Resolution Lithofacies Prediction Using Advanced Feature Engineering and Ensemble Classifiers

Improving lithofacies prediction in lacustrine shale by combining deep learning and well log curve morphology in Sanzhao Sag, Songliao Basin, China

The accurate identification of shale lithofacies is crucial for characterizing the hydrocarbon potential of lacustrine shale oil reservoirs. Petrophysical logging, serving as an effective tool for acquiring subsurface lithofacies information, provides a convenient and reliable lithofacies identification solution. Deep learning technology, capable of adapting to the nonlinearity and non-stationarity inherent in geological statistics, exhibits unique advantages in conventional reservoir lithofacies prediction. However, the lithofacies of lacustrine shale formations undergo rapid spatial and temporal changes, rendering lithofacies prediction more complex compared to conventional reservoirs. In this study, the lower Qingshankou member in the Sanzhao Sag was selected as the research target, and the Deep Residual Shrinkage Network (DRSN), known for its ability to handle complex nonlinear relationships and mitigate the effects of noisy data through residual connections and shrinkage mechanisms, was employed as a deep learning framework for predicting lithofacies in lacustrine shale formations for the first time. Well logging data, including natural gamma ray (GR), acoustic (AC), deep investigate double lateral resistivity log (RD), shallow investigate double lateral resistivity log (RS), and corrected compensated neutron log (CNC), were used as input features for the model. The results indicate that the DRSN model achieves an accuracy of 76.3% in predicting lithofacies in lacustrine shale formations. However, the DRSN model still exhibits shortcomings in capturing lithofacies change information. To enhance the model's ability to identify lithofacies change interfaces, this study further explicitly introduces Well Logging Curve Morphological Features (WLCM) as additional features and establishes a recognition method combining DRSN with WLCM. The combined DRSN-WLCM model was validated using a separate test dataset, demonstrating an improved accuracy of 85.5%, using the five well logging attributes and the derivative of the AC as inputs. Furthermore, the study reveals the lithofacies spatial distribution characteristics of the lower Qingshankou member in the Sanzhao Sag. This method can be widely applied to lithofacies delineation in lacustrine shale formations and similar stratigraphic units.

Application of Deep Learning for Reservoir Porosity Prediction and Self Organizing Map for Lithofacies Prediction

While there is a connection between petrophysical logs and reservoir porosity, finding analytical solutions for this relationship is still difficult. This paper presents a novel approach for forecasting porosity and lithofacies by using a convolutional neural network (CNN) model in conjunction with a bi-directional long short-term memory (BLSTM) network. The BLSTM network uses a self-organizing map (SOM) technique to form connections between input and destination data. The SOM is used to organize depth intervals with similar facies into four separate clusters, each exhibiting internal consistency in petrophysical parameters. The CNN is responsible for extracting spatial characteristics, while the BLSTM network gathers comprehensive spatiotemporal components, guaranteeing that the model accurately represents the spatiotemporal aspects of log data. The accuracy of the model was verified by analyzing simulation logging data. The findings indicate that the BLSTM network model successfully recovers significant characteristics from logging data, resulting in improved estimate accuracy. In addition, Facies-01 has lower gamma ray levels in comparison to other facies. Facies-01 is also suggestive of pristine sandstone formations, which are greatly sought as reservoir rocks. The BLSTM network model is effective in predicting physical characteristics of reservoirs, offering a new method for precise reservoir characterization parameter prediction.

Data-Driven Classification and Logging Prediction of Mudrock Lithofacies Using Machine Learning: Shale Oil Reservoirs in the Eocene Shahejie Formation, Bonan Sag, Bohai Bay Basin, Eastern China

As the world’s energy demand continues to expand, shale oil has a substantial influence on the global energy reserves. The third submember of the Mbr 3 of the Shahejie Fm, characterized by complicated mudrock lithofacies, is one of the significant shale oil enrichment intervals of the Bohai Bay Basin. The classification and identification of lithofacies are key to shale oil exploration and development. However, the efficiency and reliability of lithofacies identification results can be compromised by qualitative classification resulting from an incomplete workflow. To address this issue, a comprehensive technical workflow for mudrock lithofacies classification and logging prediction was designed based on machine learning. Principal component analysis (PCA) and hierarchical cluster analysis (HCA) were conducted to realize the automatic classification of lithofacies, which can classify according to the internal relationship of the data without the disturbance of human factors and provide an accurate lithofacies result in a much shorter time. The PCA and HCA results showed that the third submember can be split into five lithofacies: massive argillaceous limestone lithofacies (MAL), laminated calcareous claystone lithofacies (LCC), intermittent lamellar argillaceous limestone lithofacies (ILAL), continuous lamellar argillaceous limestone lithofacies (CLAL), and laminated mixed shale lithofacies (LMS). Then, random forest (RF) was performed to establish the identification model for each of the lithofacies and the obtained model is optimized by grid search (GS) and K-fold cross validation (KCV), which could then be used to predict the lithofacies of the non-coring section, and the three validation methods showed that the accuracy of the GS–KCV–RF model were all above 93%. It is possible to further enhance the performance of the models by resampling, incorporating domain knowledge, and utilizing the mechanism of attention. Our method solves the problems of the subjective and time-consuming manual interpretation of lithofacies classification and the insufficient generalization ability of machine-learning methods in the previous works on lithofacies prediction research, and the accuracy of the model for mudrocks lithofacies prediction is also greatly improved. The lithofacies machine-learning workflow introduced in this study has the potential to be applied in the Bohai Bay Basin and comparable reservoirs to enhance exploration efficiency and reduce economic costs.

The Bowland Shale Formation in the Blacon Basin: syngenetic processes, stacking patterns and heat productivity

Abstract We conducted a high-resolution multi-disciplinary analysis of two core sections in the borehole Ellesmere Port-1, Cheshire, UK. Biostratigraphic analysis indicates that the core sections are Kinderscoutian and late Arnsbergian–Chokierian in age, respectively. Both cores are assigned to the Bowland Shale Formation (Holywell Shale). Coupled core scan and discrete geochemical analysis enables interpretation of syngenetic processes at a high stratigraphic resolution. Both cores exhibit the classic cyclicity of limestones, calcareous to non-calcareous mudstones and siltstones, interpreted to represent sediment deposition during fourth-order sea-level fluctuation. Machine learning of the well log data coupled to the core scan data enabled prediction of the key lithofacies through the entire Bowland Shale interval in Ellesmere Port-1. The machine predictions show that the Bowland Shale is interfingered with three turbiditic leaves of the Cefn-y-Fedw Sandstone Formation and contains at least 12 complete fourth-order cycles. The Bowland Shale exhibits high radiogenic heat productivity in comparison with other sedimentary rocks, due primarily to relative U enrichment under intermittently euxinic conditions. Thermal modelling, however, shows that the radiogenic heat productivity of the Bowland Shale contributes a negligible source of additional heat at the scale of hundreds of metres.

Application of Machine Learning in Prediction of Lithofacies from Well Logs

As humanity’s reliance on mineral resources continuously grows, new technologies need to be implemented in the mining industry to fulfill this demand. A method of using well logging data to predict rock types in the surrounding area using a multiclass classification neural network is discussed as a potential way to increase efficiency. The model achieved an accuracy rate much higher than would be possible through guessing, 70% as compared to 11%, demonstrating the effectiveness of such technology. Potential ways this model can be applied and improved were also discussed.

Prediction of igneous lithology and lithofacies based on ensemble learning with data optimization

Igneous rocks are widely developed in various Mesozoic and Cenozoic continental and marine basins. Igneous reservoirs are the key reservoirs for current oil and gas development. Accurate prediction of lithology and lithofacies is a prerequisite for the effective exploration of igneous reservoirs. Igneous lithology and lithofacies are complex and correlated. The existing single-label igneous rock identification methods only consider the prediction of individual properties, and less consideration is given to the correlation of reservoir properties. Therefore, lithology and lithofacies prediction based on conventional logging data is regarded as a typical class-imbalanced multilabel classification problem when considering both attribute correlation in algorithms and evaluation metrics. To solve this problem, an ensemble method of data optimization combined with multigrained cascade forest (CF) is used in this study to develop a new multilabel lithology and lithofacies prediction model based on data from nine conventional logs in the eastern depressional reservoirs of the Liaohe Basin, and satisfactory results are obtained. The imbalance problem of the conventional logging data sets is first solved by using K-means and synthetic minority oversampling technique methods; then, the model is trained by scenario transformation and stripping with multigrained CF; and next, a multilabel classification evaluation index with multiple perspectives is introduced. The differences between the model and typical intelligent algorithms such as CF, adaptive boosting, random forest, and support vector machine are compared in simulated wells, and the new model is found to have obvious advantages. The model is finally applied to an actual well, and the accurate prediction results illustrate that the new model designed and trained for the class imbalance multilabel classification problem in this paper has application value in the lithology and lithofacies multilabel prediction of igneous rocks and also provides a theoretical basis for more complex multilabel reservoir evaluation using machine learning in the future.

An ensemble-based machine learning solution for imbalanced multiclass dataset during lithology log generation

The lithology log, an integral component of the master log, graphically portrays the encountered lithological sequence during drilling operations. In addition to offering real-time cross-sectional insights, lithology logs greatly aid in correlating and evaluating multiple sections efficiently. This paper introduces a novel workflow reliant on an enhanced weighted average ensemble approach for producing high-resolution lithology logs. The research contends with a challenging multiclass imbalanced lithofacies distribution emerging from substantial heterogeneities within subsurface geological structures. Typically, methods to handle imbalanced data, e.g., cost-sensitive learning (CSL), are tailored for issues encountered in binary classification. Error correcting output code (ECOC) originates from decomposition strategies, effectively breaking down multiclass problems into numerous binary subproblems. The database comprises conventional well logs and lithology logs obtained from five proximate wells within a Middle Eastern oilfield. Utilizing well-known machine learning (ML) algorithms, such as support vector machine (SVM), random forest (RF), decision tree (DT), logistic regression (LR), and extreme gradient boosting (XGBoost), as baseline classifiers, this study aims to enhance the accurate prediction of underground lithofacies. Upon recognizing a blind well, the data from the remaining four wells are utilized to train the ML algorithms. After integrating ECOC and CSL techniques with the baseline classifiers, they undergo evaluation. In the initial assessment, both RF and SVM demonstrated superior performance, prompting the development of an enhanced weighted average ensemble based on them. The comprehensive numerical and visual analysis corroborates the outstanding performance of the developed ensemble. The average Kappa statistic of 84.50%, signifying almost-perfect agreement, and mean F-measures of 91.04% emphasize the robustness of the designed ensemble-based workflow during the evaluation of blind well data.

Lithofacies prediction from seismic data using deep learning: A case study from North West Shelf Australia

A fully convolutional neural network was applied to prestack 3D seismic data to predict lithofacies. A key challenge was to segregate volcanics and shale, which have similar elastic properties. We achieved reasonable lithofacies prediction by using pseudowells as labeled data, which could incorporate geologic information into the model (i.e., vertical distribution patterns). In particular, the addition of stratigraphic information as input to the neural network enabled effective learning. In addition, multiple networks were separately trained to evaluate prediction uncertainty. Even in such cases, the number of pseudowells required for the robust training was much smaller than the number of traces in the 3D seismic data.

Multiscale data fusion reservoir modeling: The case study of the E3h formation, A gas field, X sag (East China Sea)

The study of reservoir characterization is challenging and uncertain owing to the low exploration level of gas field A in the East China Sea shelf basin, in which four exploratory wells have been drilled, the absence of drilling information, and the poor quality of seismic data. In this study, multiscale information is synthesized and refined from several sources, types, and dimensions, including experimental data from the deposition physical simulation experiment, field outcrop, logging curves, and seismic inversion. This study conducts neural network lithofacies prediction research and braided river 3D training image construction and forms a method for geological modeling of deep, heterogeneous, and thick reservoirs in offshore gas fields with few-well conditions by fusing multi-scale information. The lithofacies model is established using the multiple point geostatistical SIMPAT algorithm. The physical model is built using facies control and collaborative constraints. These provide methodological reference and geological support to effectively reduce the exploration risk and uncertainty of reservoir characterization in offshore few-well oil and gas fields.

Lithofacies prediction from conventional well logs using geological information, wavelet transform, and decision tree approach in a carbonate reservoir in southeastern Brazil

This work proposes a workflow through numerical simulations to determine the link between lithofacies and electrofacies. The correspondence between them was studied using the gamma-ray, resistivity, neutron porosity, bulk density, and transit time geophysical well logs. The lithofacies information, derived from conventional cores analysis, was used as the target in the process. On the other hand, the classification of lithofacies using logs has been performed by several authors using a wide variety of approaches. Thus, records from boreholes A10 (reference) and A03 (blind test) of a carbonate reservoir in Campos Basin, Southeastern Brazil, were used to develop this study. The cross-plot between the bulk density and gamma-ray logs suggests five electrofacies, which coincide with the five lithofacies proposed in the literature. Throughout the work, Pearson's correlation coefficient (R) and the mean square error (MSE) were used as statistical parameters to assess the goodness of fit. The gamma-ray electrofacies were better adjusted to the lithofacies of each well (R = 0.74 and 0.65, MSE = 0.63 and 1.36), which is recognized by publications that show this log with the best performance in identifications of this type. After that, the wavelet transforms of all normalized logs showed only a good correlation between lithofacies and the transformed signal to the gamma-ray log (R = 0.83 and 0.64, MSE = 0.01 and 0.04). This result also indicates that the electrofacies derived from the normalized and wavelet-transformed gamma ray log exhibited better results than those resulting directly from the simple gamma ray log in both wells compared to the lithofacies. Both results also show that gamma-ray electrofacies are the best input to the simulation process. The decision tree approach modeling then performed the estimates only with the normalized and wavelet-transformed gamma ray log, revealing that this method worked reasonably well in both reference and blind test wells (R = 0.79 and 0 0.63, MSE = 0.02 and 0.06). Outcomes were better with all normalized and wavelet transformed logs for both wells (R = 0.93 and 0.95, MSE = 0.006 and 0.007), with the blind test well having better R but a slightly higher MSE. So, this approach allows to derive lithofacies from electrofacies when a series of steps are taken: a) select the logs that have the most relevant lithological information; b) normalize lithology and logs on the same basis; c) transform the logs by wavelet, and d) use these results as input to the inverse process with the decision tree algorithm.

A lithofacies prediction method based on stochastic simulation

AbstractLithofacies prediction can facilitate the search for hydrocarbon reservoirs and is, therefore, crucial for seismic exploration. We proposed a new approach for lithofacies prediction based on stochastic simulation of pseudo-wells. First, we performed a pseudo-well simulation using a continuous-time Markov chain (CTMC), which can accurately describe geological deposition and can be created using logging, core or geological data. We used the CTMC to randomly generate many pseudo-wells with different lithofacies columns. The corresponding reservoir and elasticity parameters for these lithofacies columns were determined using the Monte Carlo random sampling method. Second, we used seismic modeling and seismic matching to predict the lithofacies. We applied the elastic parameters for each pseudo-well to calculate the reflection coefficients and performed seismic modeling to generate synthetic seismic data. Thereafter, we used a weighted matching function based on prior geological constraints to match the synthetic seismic data with actual seismic data. We predicted the lithofacies based on numerous pseudo-wells with the best fit and tested the influencing factors of the method using model data. Finally, we applied this method to the field data, and the predicted lithofacies agreed well with the wells. Compared to conventional methods, the proposed method can use prior information in the log and the geology data, and detects thin layers better.

Optimized feature selection assists lithofacies machine learning with sparse well log data combined with calculated attributes in a gradational fluvial sequence

Machine learning (ML) to predict lithofacies from sparse suites of well-log data is difficult in laterally and vertically heterogeneous reservoir formations in oil and gas fields. Meandering, braided fluviatile depositional environments tend to form clastic sequences with laterally discontinuous layers due to the continuous shifting of relatively narrow sandstone channels. Three cored wellbores drilled through such a reservoir in a large oil field, with just four recorded well logs available, are used to classify four lithofacies using ML models. To augment the well-log data, six derivative and volatility attributes were calculated from the recorded gamma ray and density logs, providing sixteen log features for the ML models to select from. A novel, multiple-optimizer feature selection technique was developed to identify high-performing feature combinations with which seven ML models were used to predict lithofacies assisted by multi-k-fold cross validation. Feature combinations with just seven to nine selected log features achieved overall ML lithofacies accuracy of 0.87 for two wells used for training and validation. When the trained ML models were applied to a third well for testing, lithofacies ML prediction accuracy declined to 0.65 for the best performing extreme gradient boosting model with seven features. However, an accuracy of ∼0.76 was achieved by that model in predicting the presence of the pay bearing sandstone and siltstone lithofacies in the test well. A model using only the four recorded well logs was only able to predict the pay-bearing lithofacies with ∼0.6 accuracy. Annotated confusion matrices and feature importance analysis provide additional insight to ML model performance and identify the log attributes that are most influential in enhancing lithofacies prediction.

Lithofacies Characteristics and Sweet Spot Distribution of Lacustrine Shale Oil: A Case Study from the Dongying Depression, Bohai Bay Basin, China

Abstract Lacustrine shale is characterized by rapid lithofacies transformation and compositional heterogeneity, which present challenges in shale oil sweet spot evaluation and distribution prediction and should be systematically studied. Field emission-scanning electron microscopy (FE-SEM), low-pressure adsorption isotherm analysis, mercury intrusion porosimetry (MIP), and triaxial compression testing were employed to comprehensively analyze the oil-bearing capacity, reservoir properties, fluidity, and frackability of different lithofacies. Via analyses of mineral composition, total organic carbon (TOC) content, and sedimentary structure, seven lithofacies were identified: organic-rich calcareous shale (L1), organic-rich laminated calcareous mudstone (L2), organic-rich laminated carbonate-bearing mudstone (L3), intermediate-organic laminated calcareous mudstone (L4), organic-poor laminated calcareous mudstone (L5), organic-poor thin-bedded calcareous mudstone (L6), and organic-rich laminated silty mudstone (L7). Considered together, the oil-bearing capacity, reservoir properties, fluidity, and frackability suggested that the L1 and L7 lithofacies were high-quality sweet spots, with satisfactory oil-bearing capacity (TOC>3.5%; S1>10 mgHC/grock), well-developed pores and microfractures, notable fluidity (as indicated by a high oil saturation index value), and suitable brittleness. The sweet spot distribution was predicted according to multiresolution graph-based clustering analysis of well logs. The results indicate that comprehensive research of the key factors for shale oil and lithofacies prediction can promote sweet spot prediction and enhance shale oil exploration.

FaciesViT: Vision transformer for an improved core lithofacies prediction

Lithofacies classification is a fundamental step to perform depositional and reservoir characterizations in the subsurface. However, such a classification is often hindered by limited data availability and biased and time-consuming analysis. Recent work has demonstrated the potential of image-based supervised deep learning analysis, specifically convolutional neural networks (CNN), to optimize lithofacies classification and interpretation using core images. While most works have used transfer learning to overcome limited datasets and simultaneously yield a high-accuracy prediction. This method raises some serious concerns regarding how the CNN model learns and makes a prediction as the model was originally trained with entirely different datasets. Here, we proposed an alternative approach by adopting a vision transformer model, known asFaciesViT, to mitigate this issue and provide improved lithofacies prediction. We also experimented with various CNN architectures as the baseline models and two different datasets to compare and evaluate the performance of our proposed model. The experimental results show that the proposed models significantly outperform the established CNN architecture models for both datasets and in all cases, achieving an f1 score and weighted average in all tested metrics of 95%. For the first time, this study highlights the application of the Vision Transformer model to a geological dataset. Our findings show that the FaciesViT model has several advantages over conventional CNN models, including (i) no hyperparameter fine-tuning and exhaustive data augmentation required to match the accuracy of CNN models; (ii) it can work with limited datasets; and (iii) it can better generalize the classification to a new, unseen dataset. Our study shows that the application of the Vision transformer could further optimize image recognition and classification in the geosciences and mitigate some of the issues related to the generalizability and the explainability of deep learning models. Furthermore, the implementation of our proposedFaciesViTmodel has been shown to improve the overall performance and reproducibility of image-based core lithofacies classification which is significant for subsurface reservoir characterization in different basins worldwide.

Lithofacies prediction driven by logging‐based Bayesian‐optimized ensemble learning: A case study of lacustrine carbonate reservoirs

AbstractAlthough lithofacies routinely is featured by distinct logging responses from each other, many types of lithofacies in practical cases show similar measuring characteristics on logs, and then to achieve a desirable solution from logging‐based lithofacies prediction actually is challengeable. Since the mathematical essence of lithofacies prediction can be explained as an issue of pattern recognition, a light gradient boosting machine, a state‐of‐the‐art ensemble learning, specifically developed to address supervised classification, could be a potential solver. Nonetheless, due to an incompatibility of inherent exclusive feature bundling algorithm for logs and usage of a great deal of hyper‐parameters, a raw light gradient boosting machine might not be suitable or functional to predict lithofacies. Thus, continuous restricted Boltzmann machine and Bayesian optimization, respectively, employed to process original logging data and optimize hyper‐parameters, are adopted as technical assistants for the light gradient boosting machine, and accordingly a new ensemble learning‐based predictor called continuous restricted Boltzmann machine – Bayesian optimization – the light gradient boosting machine, is proposed for lithofacies. To validate classifying capability and robust nature of new predictor, three experiments are designed purposefully based on the application of a dataset collected from the wells located within pre‐salt lacustrine carbonate reservoirs of the Santos Basin. Simultaneously, to highlight validating effect, other three sophisticated classifiers are introduced as competitors in all experiments, including supper vector machine, random forest and extreme gradient boosting. According to the analysis and comparison of all experimental results, overall four points are verified: (1) the integration of continuous restricted Boltzmann machine and Bayesian optimization indeed is beneficial for the light gradient boosting machine in data preprocess and modelling stage; (2) the light gradient boosting machine cored predictor shows more capability of producing reliable predicted lithofacies information compared to other three competitors; (3) training more learning samples is a simple and effective approach to enhance computing capability of any validated predictor, and under this circumstance the light gradient boosting machine cored predictor still performs relatively better; (4) the light gradient boosting machine cored predictor is characterized with a better robustness as its predicting outcomes, even derived from a sparse‐condition dataset, are also proved more qualified. Consequently, the new proposed predictor is demonstrated as a high‐efficient and robust solver for the prediction of lithofacies and deserves a widespread application in the geological, geophysical and petrophysical research.

Supervised machine learning application of lithofacies classification for a hydrodynamically complex gas - condensate reservoir in Nam Con Son basin

Conventional integration of rock physics and seismic inversion can quantitatively evaluate and contrast reservoir properties. However, the available output attributes are occasionally not a perfect indicator for specific information such as lithology or fluid saturation due to technology constraints. Each attribute commonly exhibits a combination of geological characteristics that could lead to subjective interpretations and provides only qualitative results. Meanwhile, machine learning (ML) is emerging as an independent interpreter to synthesise all parameters simultaneously, mitigate the uncertainty of biased cut-off, and objectively classify lithofacies on the accuracy scale.
 In this paper, multiple classification algorithms including support vector machine (SVM), random forest (RF), decision tree (DT), K-nearest neighbours (KNN), logistic regression, Gaussian, Bernoulli, multinomial Naïve Bayes, and linear discriminant analysis were executed on the seismic attributes for lithofacies prediction. Initially, all data points of five seismic attributes of acoustic impedance, Lambda-Rho, Mu-Rho, density (ρ), and compressional wave to shear wave velocity (VpVs) within 25-metre radius and 25-metre interval offset top and base of reservoir were orbitally extracted on 4 wells to create the datasets. Cross-validation and grid search were also implemented on the best four algorithms to optimise the hyper-parameters for each algorithm and avoid overfitting during training. Finally, confusion matrix and accuracy scores were exploited to determine the ultimate model for discrete lithofacies prediction. The machine learning models were applied to predict lithofacies for a complex reservoir in an area of 163 km2.
 From the perspective of classification, the random forest method achieved the highest accuracy score of 0.907 compared to support vector machine (0.896), K-nearest neighbours (0.895), and decision tree (0.892). At well locations, the correlation factor was excellent with 0.88 for random forest results versus sand thickness. In terms of sand and shale distribution, the machine learning outputs demonstrated geologically reasonable results, even in undrilled regions and reservoir boundary areas.

Joint use of multiseismic information for lithofacies prediction via supervised convolutional neural networks

Lithology prediction from seismic data is of great significance for sweet-spot detection, reservoir structure delineation, and geologic model building, hence it is important in reducing the risk of exploration and development. Traditional lithofacies prediction methods often are limited by the seismic inversion accuracy and reliability of the rock-physics relationships, which are challenging to be applied in complex reservoirs (such as those containing coal-bearing strata or thin layers). Convolutional neural networks (CNNs) can represent the coupling relationship of seismic characteristics in the time domain through multilayer convolution and effectively manipulate multitype and multidimensional seismic data. Under the framework of a supervised CNN, we jointly integrate prestack seismic gathers (Pre-SGs), seismic inversion results (P-impedance and [Formula: see text] ratio), multiseismic attributes (amplitude-variation-with-offset [AVO] intercept, AVO gradient, instantaneous amplitude, instantaneous frequency, and instantaneous phase), and spectral decomposition attributes (SDA) to predict lithofacies in a complex clastic reservoir interbedded with thin-layer coal. We determine that the fusion model with multiseismic information containing different perspectives and complementary information of seismic data is capable of achieving better prediction performance than only using one type of input feature. In particular, using the proposed methodology, the angle-dependent Pre-SG is essential to decipher the rich information of lithologic details. The models using only poststack seismic data or inversion results cannot reliably describe lithologic details (especially the thin-coal layers). In addition, by including the SDA into model inputs, the model’s ability to recognize thin layers has been further improved but lead to the slight sacrifice of overall prediction accuracy.

Application of Machine Learning for Lithofacies Prediction and Cluster Analysis Approach to Identify Rock Type

Nowadays, there are significant issues in the classification of lithofacies and the identification of rock types in particular. Zamzama gas field demonstrates the complex nature of lithofacies due to the heterogeneous nature of the reservoir formation, while it is quite challenging to identify the lithofacies. Using our machine learning approach and cluster analysis, we can not only resolve these difficulties, but also minimize their time-consuming aspects and provide an accurate result even when the user is inexperienced. To constrain accurate reservoir models, rock type identification is a critical step in reservoir characterization. Many empirical and statistical methodologies have been established based on the effect of rock type on reservoir performance. Only well-logged data are provided, and no cores are sampled. Given these circumstances, and the fact that traditional methods such as regression are intractable, we have chosen to apply three strategies: (1) using a self-organizing map (SOM) to arrange depth intervals with similar facies into clusters; (2) clustering to split various facies into specific zones; and (3) the cluster analysis technique is used to identify rock type. In the Zamzama gas field, SOM and cluster analysis techniques discovered four group of facies, each of which was internally comparable in petrophysical properties but distinct from the others. Gamma Ray (GR), Effective Porosity(eff), Permeability (Perm) and Water Saturation (Sw) are used to generate these results. The findings and behavior of four facies shows that facies-01 and facies-02 have good characteristics for acting as gas-bearing sediments, whereas facies-03 and facies-04 are non-reservoir sediments. The outcomes of this study stated that facies-01 is an excellent rock-type zone in the reservoir of the Zamzama gas field.

Identification of lithofacies and prediction of mineral composition in shales – A case study of the Shahejie Formation in the Bozhong Sag

The identification of shale lithofacies is the basic work of shale gas exploration and development. Accurate quantitative characterization of different mineral components in fine-grained mixed shale is of great significance for the identification and classification of lithofacies types, the enrichment conditions of shale oil and gas (hydrocarbon generation, reservoir, occurrence and preservation), and the evaluation of shale oil potential (reservoir, oil bearing, fracturability and oil mobility). When organic-rich argillaceous laminae are moderately mixed with brittle laminae, the laminated shale is not only the favorable interval of shale oil and gas enrichment for hydrocarbon generation (organic-rich argillaceous laminae) and storage (brittle laminae), but also the excellent “sweet spots” for continental shale oil and gas optimization. Different shale lithofacies have distinct rock texture, fabric and composition, leading to different brittleness and rock physical properties. For the identification of shale lithofacies containing various components, the overlaps of different wire-line logging responses and the vague boundaries between various logging data cause the large deviation in the logging prediction of lithofacies by traditional methods. In this study, shale samples from the lower part of the third member of Shahejie Formation in Bozhong Sag of the Bohai Bay Basin are selected to carry out the reservoir characterization and then the data mining of logging information by a Back Propagation (BP) neural network coupled with Atomic Search Optimization (ASO) algorithm. The BP algorithm based on the identified shale lithofacies (expected output) and logging data index (input) is used to train the neural network. The complex and unrecognized nonlinear relationship between shale lithofacies and logging data is mapped onto the high-dimensional identifiable nonlinear quantitative relationship to establish the prediction model of the relative content of clay minerals, silicate minerals and carbonate minerals. This study reveals the main lithologic characteristics of lacustrine shale lithofacies from Shahejie Formation in Bozhong Sag, and the main controlling factors for shale lithofacies prediction based on logging data. Our results show that the main mineral composition of shale lithofacies associations can be effectively predicted through the whole rock X-ray diffraction data, wire-line log data and neural network analysis, which provides the basis for lithofacies identification shale interval in well locations lacking core and test data. • Based on the optimized algorithm, a neural network prediction model of mineral composition is established to mine the hidden information in conventional logging data to predict lithofacies, and the accuracy is improved compared with the conventional neural network model. • The prediction of lithofacies mineral composition data for a wider range of non-coring wells by neural network provides more geological basis for the selection and prediction of favorable lithofacies zone - dessert area in subsequent exploration.