Facies Prediction Research Articles

Sediment bypass in a marl-dominated margin of a turbidite system in a narrow basin setting

Submarine sediment density flows play a pivotal role in transporting clastic material to the deep sea. The volume of sediment they transport, which bypasses a specific point or geographical location, shapes the stratigraphic record of the entire turbidite systems. Hence, recognition of bypass-dominated zones is crucial in facies prediction and understanding the architecture of turbidite systems. This understanding is linked to the economic aspects of exploring hydrocarbon deposits, the occurrence of geohazards that impact submarine infrastructure, the distribution of pollutants, and carbon sequestration.In the western part of the Ropianka Fm (Skole Nappe, Polish Outer Carpathians), three bed types showing evidence of sediment bypass were identified in the channel-mouth setting of marl-dominated slope and base-of-slope successions. The varying proportions of these bed types in studied successions and relationship with adjacent facies associations led to the identification of two channel-mouth zones. This study provides insights into deposits with characteristics differing from the previously described channel-mouth setting, evidenced by a significantly lower sand-to-mud ratio and smaller scale of erosional and depositional structures. The reported channel-mouth zones are interpreted as the marginal parts of a channel mouth, formed by subcritical flows. This study broadens the understanding of the channel-mouth setting by introducing dynamic and mud-dominated zones that experience sediment bypass and weak erosion. The identified bed types and typical characteristics of Marginal channel-mouth zones 1 and 2 can serve as a reference for interpreting marginal areas of channel-mouth settings in mud-dominated successions with scattered thin-bedded and coarse-grained deposits in other deep-water basins.

Permeability Prediction and Facies Distribution for Yamama Reservoir in Faihaa Oil Field: Role of Machine Learning and Cluster Analysis Approach

Empirical and statistical methodologies have been established to acquire accurate permeability identification and reservoir characterization, based on the rock type and reservoir performance. The identification of rock facies is usually done by either using core analysis to visually interpret lithofacies or indirectly based on well-log data. The use of well-log data for traditional facies prediction is characterized by uncertainties and can be time-consuming, particularly when working with large datasets. Thus, Machine Learning can be used to predict patterns more efficiently when applied to large data. Taking into account the electrofacies distribution, this work was conducted to predict permeability for the four wells, FH1, FH2, FH3, and FH19 from the Yamama reservoir in the Faihaa Oil Field, southern Iraq. The framework includes: calculating permeability for uncored wells using the classical method and FZI method. Topological mapping of input space into clusters is achieved using the self-organizing map (SOM), as an unsupervised machine-learning technique. By leveraging data obtained from the four wells, the SOM is effectively employed to forecast the count of electrofacies present within the reservoir. According to the findings, the permeability calculated using the classical method that relies exclusively on porosity is not close enough to the actual values because of the heterogeneity of carbonate reservoirs. Using the FZI method, in contrast, displays more real values and offers the best correlation coefficient. Then, the SOM model and cluster analysis reveal the existence of five distinct groups.

Evaluation of sweet spots for a tight sandstone reservoir: A quantitative study of diagenesis in the fourth member of the Oligocene Huagang Formation, Xihu Depression, East China Sea Shelf Basin

Reservoir sweet spot evaluation is very important for tight sandstone gas exploration and development. Due to the characteristics of large burial depth, strong diagenesis heterogeneity, and prominent importance of diagenetic facies, sweet spot evaluation is the focus and difficulty of unconventional oil and gas exploration and development. By combining with core observation, conventional core analysis, clay x-ray diffraction analysis, bulk rock x-ray diffraction analysis, powder grain size analysis, thin section quantitative statistics, fluid inclusion homogenization temperature, scanning electron microscopy, scanning electron microscopy mineral quantitative evaluation, well logging data, and pre-stack seismic data, integrated simulation approach of seismic diagenetic facies and diagenetic numerical simulation was used to simulate porosity evolution and spatial distribution of tight sandstone of second submember (Smbr 42) of the fourth member (Mbr 4) of the Oligocene Huagang Formation in the Xihu Depression in the East China Sea Shelf Basin, and to evaluate reservoir sweet spot distribution. Based on mineral texture relationship and diagenetic sequence of Smbr 42 sandstone compaction, clay mineral transformation, calcite cementation, quartz cementation, and dissolution, five diagenetic facies were identified. Based on various seismic elastic parameters and core-derived diagenetic facies, the distribution of seismic diagenetic facies was interpreted by seismic diagenetic facies prediction method. Seismic diagenetic facies distribution can provide a “hard-constraint” three-dimensional diagenesis heterogeneity model for diagenetic history reconstruction, and seismic diagenetic facies distribution provides constraints on grain size and quartz grain content distribution in diagenetic numerical model. Based on the detailed consideration of input parameters of burial history, thermal history and diagenetic history, integrated diagenetic numerical simulation was used to reproduce porosity evolution and spatial distribution of Smbr 42 sandstone, and was applied to reservoir sweet spot evaluation of tight sandstone. Smbr 42-1 sublayer developed type III reservoir sweet spot and non-sweet spot. Smbr 42–2 to Smbr 42-6 sublayers mainly developed type II1, type II2, type III, and minor type I reservoir sweet spot. Simulation results were consistent with the locations of multiple sets of drilled gas layers and proven gas reservoirs.

Facies prediction in a mature oil field of Cretaceous age in the Calumbi Formation (Sergipe-Alagoas Basin, Brazil) by using an outcrop analogue approach

Outcrop gamma-ray logging is crucial for understanding subsurface reservoirs, but it is rarely used on outcrops, especially for correlating gamma-ray data from borehole profiles. This case study focuses on a mature onshore reservoir in the Sergipe-Alagoas Basin, Brazil, which is currently discarded by the oil industry. The study used an ideal outcrop (M-17) to gather stratigraphic, sedimentological, and gamma-ray data to better understand the lateral heterogeneity and distribution of reservoirs thought to have formed in offshore environments in the Sergipe-Alagoas Basin. Seismic and paleontological data were used to reduce uncertainty in outcrop data. The results showed a paleoenvironment compatible with marine sedimentary deposits on the continental shelf, supported by evidence from Thalassinoides and Ophiomorpha ichnogenera. The model suggests sand was continuously deposited following cycles of accommodation generation, resulting in a single succession of rocks organized into three system tracts: lowstand, transgressive, and highstand. The gamma-ray response identified and correlated the Cidade de Aracaju Oilfield using a pseudo-well and a well in the Cidade de Aracaju Oilfield. The methodological approach was seen as useful as a correlation tool, revealing previously unknown exploratory potential in the surrounding areas of the overexploited Cidade de Aracaju Oilfield. This approach is both unprecedented and pivotal, revealing previously unknown exploratory potential in the surrounding areas of the overexploited Cidade de Aracaju Oilfield.

Identification of carbonate sedimentary facies from well logs with machine learning

Sedimentary facies identification is critical for carbonate oil and gas reservoir development. The traditional method of sedimentary facies identification not only be affected by the engineer's experience but also takes a long time. Identifying carbonate sedimentary facies based on machine learning is the trend of future development and has the advantages of short time consuming and reliable results without engineers' subjective influence. Although many references reported the application of machine learning to identify lithofacies, but identifying sedimentary facies of carbonate reservoirs is much more challenging due to the complex sedimentary environment and tectonic movement. This paper compares the performance of the carbonate sedimentary facies identification using four different machine learning models, and the optimal machine learning with the highest prediction accuracy is recommended. First, the carbonate sedimentary facies are classified into the lagoon, shallow sea, shoal, fore-shoal, and inter-shoal five tags based on the well loggings. Then, five well log curves including spectral gamma ray (SGR), uranium-free gamma ray (CGR), photoelectric absorption cross-section index (PE), true formation resistivity (RT), shallow lateral resistivity (RS) are used as the input, and the manual identified carbonate sedimentary facies are used as the output of the machine learning model. The performance of four different machine learning algorithms, including support vector machine (SVM), deep neural network (DNN), long short-term memory (LSTM) network, and random forest (RF) are compared. The other two wells are used for model validation. The research results show that the RF method has the highest accuracy of sedimentary facies prediction, and the average prediction accuracy is 78.81%; the average accuracy of sedimentary facies prediction using SVM is 77.93%. The sedimentary facies predictions using DNN and LSTM are less satisfying compared with RF and SVM, and the average accuracy is 69.94% and 73.05%, respectively. The predicted carbonate sedimentary facies by LSTM are more continuous compared with other machine learning models. This study is helpful for identifying compelx sedimentary facies of carbonate reservoirs from well logs.

Elastic rock properties analysis of diagenetic facies identification of deep tight sandstone in the Xihu Sag, China

In order to investigate the physical relationship between elastic properties and reservoir parameters, and determine sensitive seismic elastic parameters of diagenetic facies of deep tight sandstone in the sparse well network area, 16 samples of typical diagenetic facies were selected for rock physics experiment. Diagenetic characteristics and diagenetic facies of tight sandstone were analyzed using petrographic data, and mineral associations and diagenetic pathways of diagenetic facies were investigated. Then, based on rock physics experiment data and petrographic data, sensitive seismic elastic parameters were determined, by comprehensively using rock physics analysis, correlation analysis, and relationship between diagenetic characteristics, mineral composition and seismic elastic parameters. Sensitive seismic elastic parameters can identify four diagenetic facies of tight sandstone. The results show that shear modulus, p-wave to s-wave velocity ratio, p-wave impedance and porosity data volume are sensitive to diagenetic facies, and the correlation between four seismic elastic parameters is weak. The determination of sensitive seismic elastic parameters in this study makes seismic diagenetic facies research targeted to distinguish four diagenetic facies. Understanding the basic physical significance of four elastic parameters is very important for their application to diagenetic facies prediction.

Quantitative Classification and Prediction of Diagenetic Facies in Tight Gas Sandstone Reservoirs via Unsupervised and Supervised Machine Learning Models: Ledong Area, Yinggehai Basin

Quantitative Classification and Prediction of Diagenetic Facies in Tight Gas Sandstone Reservoirs via Unsupervised and Supervised Machine Learning Models: Ledong Area, Yinggehai Basin

Fisher-Bayesian inversion for estimating shale gas facies

The classification of shale gas facies from seismic properties is critical for shale gas reservoir characterization. Shale gas facies are affected by many petrophysical properties. Therefore, the characterization of shale facies should be carried out by multiple parameters, which is more reasonable and accurate. However, multiparameter inversion often leads to unstable results, and coupled properties are generally a way of solving this problem. We develop a Fisher-Bayesian inversion method for estimating shale gas facies by combining the Fisher projection and Bayesian inversion method. The mathematical method adopted for the inversion is the Bayesian framework. The link between different facies and coupled properties is given by a joint prior distribution. We derive the analytical formulation of the Bayesian inversion under the Gaussian mixture assumption for coupled attributes and different shale gas facies. Our approach realizes the fusion of multidimensional petrophysical parameters and establishes a shale gas facies prediction method based on coupled properties. The application to real data sets delivers accurate and stable results, wherein shale gas facies and coupled attributes are accurately predicted and inversed.

Using deep-learning to predict Dunham textures and depositional facies of carbonate rocks from thin sections

Petrographic analysis of thin sections is one of the most important and routinely used methods in a wide range of energy, earth and environmental science applications. There has been a growing motivation for computer-aided automated analysis of thin sections due to an inherent subjectivity of interpretation by geologists and the ever-increasing need for analysis of new and legacy thin section petrography data. Particularly in carbonate reservoirs, Dunham texture classification and depositional facies prediction based on thin sections are crucial for accurate description and modeling of the reservoir, but these two tasks are labor and time-intensive and their accuracy is highly dependent on the experience and the ability of the interpreter. This study addresses these challenges by exploring deep learning methods in two case studies to predict Dunham textures and depositional facies using thin sections. The proposed methodologies were applied to 1038 thin sections collected from outcrop analogues of the Upper Jurassic Hanifa Formation. To accelerate the learning speed, transfer learning was applied based on various pre-trained models provided by PyTorch. We used a pre-trained DenseNet model to classify Dunham textures from RGB images of thin sections. For the prediction of depositional facies we applied an integrated methodology consisting of two VGG models and a U-Net model. First, one VGG model was used to classify thin sections into three groups, namely, grain-dominated, mud-dominated, and stromatoporoid facies. Then the U-Net model was used to identify oncoids to further classify oncoidal facies from grain-dominated facies. Finally, another VGG model was applied to classify the peloid-rich facies from the left grain-dominated facies. The Dunham classification model resulted in a prediction accuracy of 89%. The most frequent misclassification was mainly from the identification of packstone, particularly mud-supported packstone, which the model misclassified as wackestone. The depositional facies prediction model achieved final accuracy of 86% where the misclassification was often due to oncoids identification. Overall, the results of this study demonstrate the potential of our proposed workflow to obtain automated prediction of petrographic features from a large number of thin sections. Deep learning offers a promising way forward for rapid and accurate prediction of petrographic features for geoscience applications, while minimizing bias associated with manual interpretation.

Seismic Facies Recognition of Ultra-Deep Carbonate Rocks Based on Convolutional Neural Network

Seismic facies is a seismic reflection unit defined by specific seismic reflection characteristics, that is, the seismic responses of sedimentary facies or geological bodies, whose accuracy will directly affect the reliability of oil and gas exploration results. Currently, seismic facies is generally recognized depending upon the differences between certain single trace seismic attributes (waveform, frequency spectrum, and amplitude, etc.) and adjacent units to conduct cluster analysis. Such methods, however, have ambiguity in identifying special reflective structures with continuous waveforms (e.g. massive carbonate deposits). In order to solve this problem, this paper incorporates artificial intelligence (AI) technology into automatic recognition of seismic facies with special reflection structures. Firstly, a 2D seismic facies classification sample label set is designed and formed. Then, a seismic facies prediction model is designed and constructed using a multi-layer convolutional neural network (CNN). Finally, the trained model is used to automatically track the seismic facies in the study area. This method was applied to seismic facies recognition for the Sinian Dengying Formation in an area of the Sichuan Basin, and the seismic facies recognized were compared with artificially interpreted ones. It is confirmed that the proposed method provides a better effect than artificial interpretation, with greatly improved accuracy and efficiency.

Markov chain Monte Carlo for seismic facies classification

Seismic facies classification aims to predict a facies model, or a set of facies models, from measured seismic data. We focus on stochastic classification methods to estimate the probability distribution of facies conditioned on seismic data. Bayesian classification methods based on analytical solutions are generally applied to seismically inverted elastic attributes or petrophysical properties rather than measured seismic data, and they do not include spatial correlation models in the prior distribution. Instead, iterative stochastic methods, such as Monte Carlo acceptance-rejection sampling and approximate Bayesian computation, can be directly applied to seismic data and can be implemented with complex prior models; however, their convergence is generally slow because it requires the simulation of a large number of facies models from the prior distribution. We develop an efficient implementation of a Markov chain Monte Carlo (MCMC) approach that adopts complex prior models, such as multiple-point statistics simulations based on a training image, to generate geologically realistic facies realizations. The novelty of the approach is that the proposal distribution of the proposed MCMC method is based on a gradual deformation algorithm, in which the proposal distribution is expressed as a linear combination of the indicator variables of the previously accepted model and a random prior simulation, according to a uniform random weight. The synthetic examples indicate accurate facies predictions with a success rate of approximately 80% for well-log facies classification based on elastic attributes and approximately 60% for seismic facies classification based on seismic data. The inversion is compared to an MCMC method with prior models sampled from a first-order Markov chain and Bayesian facies classification.

Lateral and temporal variations of a multi-phase coarse-grained submarine slope channel system, Upper Cretaceous Cerro Toro Formation, southern Chile

ABSTRACTUnderstanding variations in the sedimentary processes and resulting stratigraphic architecture in submarine channel systems is essential for characterizing sediment bypass and sedimentary facies distribution on submarine slopes. In the Santonian to Campanian Cerro Toro Formation, southern Chile, a coarse-grained slope system, informally known as the Lago Sofia Member, developed in a structurally controlled environment, with complex and poorly established relationships with the surrounding mud-rich heterolithic deposits.A detailed architectural analysis of the most continuous and best-exposed channel system in the Lago Sofia Member, the Paine C channel system, provides insights on lateral facies transitions from channel axis to margin, stacked in a multi-phase sequence of events marked by abrupt changes in facies, facies associations, and architecture.The Paine C channel system is incised into siltstones and claystones interbedded with thin-bedded very fine sandstones, interpreted to be either channel-related overbank or unrelated background deposits. The coarse-grained deposits are divided into a lower conglomeratic unit and an upper sand-rich unit. The lower conglomeratic unit can be further subdivided into three phases: 1) highly depositional and/or aggradational, dominated by thick and laterally continuous beds of clast- to matrix-supported conglomerate, herein named transitional event deposits; 2) an intermediate phase, including deposits similar to those dominant in phase 1 but also containing abundant clast-supported conglomerates and lenticular sandstones; and 3) a bypass-dominated phase, which records an architectural change into a highly amalgamated ca. 45-m-thick package composed purely of lenticular clast-supported conglomerates with local lenticular sandstones. Between the conglomeratic phases, a meter-scale package composed of interbedded thin- to medium-bedded sandstone and mudstone deposits is interpreted to drape the entire channel, indicating periods of weaker gravity flows running down the channel, with no evidence of bedload transport.The upper sand-rich unit is divided into lower amalgamated and upper non-amalgamated phases, and represents a rapid architectural change interpreted to record an overall waning of the system. The sandstone unit laps out onto a mass-transport complex which is interpreted to have been triggered initially at the same time as major architectural change from conglomerates to sandstones.While mindful of the fact that each system is a complete analogue only for itself, we propose a new depositional model for coarse-grained submarine channel systems, in which particular characteristics can provide significant insights into architectural heterogeneity and facies transitions in channelized systems, allowing substantial improvement in subsurface facies prediction for fluid reservoirs.

Prediction of carbonate rock facies from core probe permeability measurements and well log data: a case study from carbonate reservoirs, Phu Khanh basin

Probe permeameter (also known as Mini-permeameter) has been widely used in many field and laboratory applications where in-situ measurements and spatial distributions of permeability are needed. Mini-permeameter measurements have become popular techniques for collecting localized permeability measurements in both laboratory and field applications. It is designed to obtain fast, cheap, intensive and non-destructive permeability measurements and to describe the spatial arrangement of permeability. Currently the probe permeability meter is designed and manufactured as a portable air permeability for field applications and to be used in outcrop and core samples. In this instrument, the permeability is measured by air that flows from the samples to be measured into an air chamber through the vacuum created by increasing the volume of the chamber. In carbonate reservoirs, permeability predicted from pure porosity-permeability empirical relationship is often difficult due to complex rock pore systems leading to poor porosity-permeability relations. Once the relationships between permeability and textural rock properties are clearly established in carbonates, they can provide better permeability predictions from porosity data. Rock texture is an important parameter for the understanding of the porosity and permeability characteristics of carbonate reservoirs. In addition to predicting carbonate rock facies from routine core plug porosity and permeability measurements, there is an approach to determine carbonate reservoir facies based on core-plug probe permeability. The results of the probe permeability measurements, in this paper, can be used in combination with the porosity values derived from the well logs to classify and predict rock facies in carbonate cored or uncored reservoirs in Phu Khanh basin.

A Supervised Workflow for Predicting Lithofacies in Complex and Heterogeneous Tight Sandstone Reservoirs: A Data-Driven Approach Using Clustering and Classification Models

This study introduces a novel supervised workflow for predicting lithofacies in complex, heterogeneous tight sandstone reservoirs with intercalated facies. Using a two-information criteria clustering method, six distinct facies are identified, providing an unbiased, data-driven alternative to manual approaches. Among classification models, Gaussian Process Classification (GPC) outperforms others, including Support Vector Machine (SVM) and Artificial Neural Network (ANN), with Random Forest (RF) performing less effectively. GPC accurately predicts lithofacies in testing data and is assessed for similarity accuracy. Predicted lithofacies are integrated into acoustic impedance versus velocity ratio cross plots, resulting in 2D probability density functions. These, combined with depth data, feed a neural network to forecast synthetic gamma-ray log responses. Results show strong agreement between measured and predicted gamma-ray logs (R2 = 0.978) and nearly identical log trends. Additionally, the predicted lithofacies are classified using inverted impedance and velocity ratio volumes, yielding a facies prediction volume that aligns well with well site lithofacies classification, even without core data

Imaging Lowstand prograding stratigraphic pinch-out traps of a Lower Cretaceous clastic system using inverted seismic-based porosity simulation of SE-Asian Basin: Implications for petroleum exploration

Prograding sandstone-filled incised valleys of the Lowstand System tract (LST) form stratigraphic plays. The band-limited seismic data interpretation techniques constrain the prediction of porous and gas-bearing reservoir facies. Therefore, the development of a cost-economic petroleum-bearing reservoir system is a challenge for reservoir geophysicists. This research is intended to predict the porosity and thickness of incised valleys by executing the continuous wavelet transforms (CWT)-based spectrum decomposition (SD), seismic static inverted simulation (SWM), and inverted seismic-based porosity simulation (IVPS) on a Lower Cretaceous LST system of Pakistan. The seismic, root-mean-square (RMS), and absolute average seismic attributes have provided a limited capability to image the gas-bearing sandstones. The 24-Hz tuning frequency has imaged the progradational sandstone-filled incised valley and retrogradational fluvial point bars. The 37-Hz has imaged the aggradational barrier bars. The amalgamation of multiple attributes has accurately demarcated the reservoir sandstone and sealing shale. The SWS has resolved a 32-m-thick prograding incised valley, 30-m-thick aggrading barrier bars, and 23-m-thick retrograding point bars within the complete fluvial hierarchy. The well correlation has provided evidence for the distribution of point bars, barrier bars, and shallow-marine incised valley reservoir facies. The 24-Hz CWT waveform-based conventional IVPS has predicted 94 and 42 m-thick valleys and "point bars" between 25 and 13% pseudo-porosity wells. The 24-Hz waveforms-based treatment of conventional IVPS has predicted enhanced thicknesses of 75 and 29 m for valley and point bar deposits with porosities between 25 and 13%. The 24-Hz IVPS has resolved two valleys compared to the conventional IVPS during the rapid fall of sea level and negligible rise, which have accumulated the coarse-grained point bar deposits of a fluvial meandering stream. The integration of IVPS and GR (API) has provided strong implications for the exploration of a pure stratigraphic trap within a shallow marine sedimentary system of Pakistan. A NE to SW oriented regional change was observed in the lithology, thickness, and depositional environments of the petroleum system. The dominant lithology in the eastern zones was point bar sandstones. The dominant lithology in the central zones was barrier bar sandstones. The dominant lithology in the western zone was incised valley sandstone. The thickness of the reservoir was 18 m thick in the west, which was reduced to 8 m in the eastern flanks of this gas field. The depositional environment in the northeast was fluvial, which changed to shallow-marine in the southwestern zones of the Indus basin. These lateral fluctuations have proved that a pure stratigraphic fairway was present within the LST of the Indus Basin of the sedimentary system of Pakistan. The porous incised valley stratigraphic reservoirs are embedded inside the non-porous shales, which validates the presence of a pure stratigraphic petroleum play inside the Indus Basin, Pakistan.

Prediction of Diagenetic Facies via Well Logs and Petrophysical Properties in Tight Sandstone from Zhu-III Sag: Pearl River Mouth Basin, South China Sea

Prediction of Diagenetic Facies via Well Logs and Petrophysical Properties in Tight Sandstone from Zhu-III Sag: Pearl River Mouth Basin, South China Sea

Characterization and Evaluation of Carbonate Reservoir Pore Structure Based on Machine Learning

The carboniferous carbonate reservoirs in the North Truva Oilfield have undergone complex sedimentation, diagenesis and tectonic transformation. Various reservoir spaces of pores, caves and fractures, with strong reservoir heterogeneity and diverse pore structures, have been developed. As a result, a quantitative description of the pore structure is difficult, and the accuracy of logging identification and prediction is low. These pose a lot of challenges to reservoir classification and evaluation as well as efficient development of the reservoirs. This study is based on the analysis of core, thin section, scanning electron microscope, high-pressure mercury injection and other data. Six types of petrophysical facies, PG1, PG2, PG3, PG4, PG5, and PG6, were divided according to the displacement pressure, mercury removal efficiency, and median pore-throat radius isobaric mercury parameters, combined with the shape of the capillary pressure curve. The petrophysical facies of the wells with mercury injection data were divided accordingly, and then the machine learning method was applied. The petrophysical facies division results of two mercury injection wells were used as training samples. The artificial neural network (ANN) method was applied to establish a training model of petrophysical facies recognition. Subsequently, the prediction for the petrophysical facies of each well in the oilfield was carried out, and the petrophysical facies division results of other mercury injection wells were applied to verify the prediction. The results show that the overall coincidence rate for identifying petrophysical facies is as high as 89.3%, which can be used for high-precision identification and prediction of petrophysical facies in non-coring wells.

Diagenetic facies characteristics and quantitative prediction via wireline logs based on machine learning: A case of Lianggaoshan tight sandstone, fuling area, Southeastern Sichuan Basin, Southwest China

Tight sandstone has low porosity and permeability, a complex pore structure, and strong heterogeneity due to strong diagenetic modifications. Limited intervals of Lianggaoshan Formation in the Fuling area are cored due to high costs, thus, a model for predicting diagenetic facies based on logging curves was established based on few core, thin section, X-ray diffraction (XRD), scanning electron microscopy (SEM), cathodoluminescence, routine core analysis, and mercury injection capillary pressure tests. The results show that tight sandstone in the Lianggaoshan Formation has primary and secondary intergranular pores, secondary intragranular pores, and intergranular micropores in the clay minerals. The compaction experienced by sandstone is medium to strong, and the main diagenetic minerals are carbonates (calcite, dolomite, and ferric dolomite) and clay minerals (chlorite, illite, and mixed illite/montmorillonite). Four types of diagenetic facies are recognized: carbonate cemented (CCF), tightly compacted (TCF), chlorite coating and clay mineral filling (CCCMFF), and dissolution facies (DF). Primary pores develop in the CCCMFF, and secondary pores develop in the DF; The porosities and permeabilities of CCCMFF and DF are better than that of CCF and TCF. The diagenetic facies were converted to logging data, and a diagenetic facies prediction model using four machine learning methods was established. The prediction results show that the random forest model has the highest prediction accuracy of 97.5%, followed by back propagation neural networks (BPNN), decision trees, and K-Nearest Neighbor (KNN). In addition, the random forest model had the smallest accuracy difference between the different diagenetic facies (2.86%). Compared with the other three machine learning models, the random forest model can balance unbalanced sample data and improve the prediction accuracy for the tight sandstone of the Lianggaoshan Formation in the Fuling area, which has a wide application range. It is worth noting that the BPNN may be more advantageous in diagenetic facies prediction when there are more sample data and diagenetic facies types.

Seismic reservoir characterization of the Gassum Formation in the Stenlille aquifer gas storage, Denmark — Part 1

Seismic reservoir characterization plays an important role in carbon capture and storage analysis. The Havnsø anticlinal structure in Denmark is a prospective CO2 storage site due to its proximity to two large emission sources—a coal-fired power station and a nearby refinery. Although legacy 2D seismic lines over the area outline the anticlinal structure, their quality is insufficient for quantitative interpretation. Earlier studies have shown that the natural gas stored in the Stenlille aquifer exhibits a seismic response similar to the modeled CO2 fluid in the Havnsø structure. Thus, seismic reservoir characterization carried out on the Stenlille aquifer gas storage in terms of identifying spatial distribution of gas and outlining faults would provide insight regarding value addition that seismic data can bring into the proposed CO2 storage at Havnsø. Using the available poststack seismic data, we apply an integrated reservoir characterization analysis. After performing the adequate data conditioning, the impedance of the target Stenlille Formation is estimated through generation of an accurate low-frequency model. Thereafter, multiattribute analysis was used to generate volumetric estimates of porosity, gamma ray, and water saturation within the target formation so that the spatial distribution of gas can be mapped. The resulting porosity and gamma-ray volumes indicate encouraging results and were used for Bayesian classification to predict the probability of the more important lithofacies, namely, sand, shale, moderate-porosity sand, and moderate-porosity shaly sand, which enabled the mapping of high-porosity/facies zones in the two aquifer storage levels. Independently, we make use of unsupervised machine learning applications for seismic facies prediction and compare them at the two storage levels, which will be presented in the part 2 of this paper.

Integrated Carbonate Rock Type Prediction Using Self-Organizing Maps in E11 Field, Central Luconia Province, Malaysia

Reducing uncertainty in 3D carbonate rock type distribution is a critical factor that profoundly impacts field development for hydrocarbon or carbon capture and storage (CCS) projects. Miocene carbonate reservoirs in the Central Luconia offshore region are economically important global gas reservoirs. The nature of these carbonate rocks can be visually distinct in the core and the multiscale reservoir heterogeneity might vary in scale from the 100-m scale to the sub-millimeter scale. This work presents a series of steps workflow to obtain spatial information about the organization scheme of carbonate rock types, and capture the most important petrophysical and sedimentary controls on rock property distribution in the E11 field, a carbonate buildup, located in Central Luconia Province, offshore Malaysia. The spatial data were generated from a supervised neural Kohonen algorithm. The rock types predicted with this workflow were propagated using IPSOM probabilized self-organizing maps SOM. This tool is used for classifying multivariate data samples according to “patterns” or multivariate responses. The workflow includes several steps: A Step 1—Core data description, B Step 2—Thin section description, C Step 3—Well log interpretation, and D Step 4—IPSOM probabilized self-organizing maps for facies prediction SOM. The depth plots of the predicted rock type showed close correspondence to the core-based rock types in terms of the stratigraphic organization of tight and reservoir layers, proportions, and juxtaposition. This result is sufficient to merit the application of the rock type logs into a future porosity model of the E11 field, and to understand the lateral and vertical distribution of tight and reservoir rock types of distribution. The results can be used to build a future realistic digital twin of the subsurface, and in digital geological modeling.