Search

Geophysical logging series are valuable geological data that record the physical and chemical information of borehole walls and in-situ formations, and are widely used by geologists for interpreting geological problems due to their continuity, high resolution, and ease of access. Recently, machine learning methods are gradually bringing data science and geoscience closer together, and Intelligent Recognition using Logging Data (IRLD) is increasingly becoming an important interpretation task. However, due to the specificity of geological information, relatively low data quality makes the direct application of machine learning models to IRLD often not optimal. And to the best of our knowledge, IRLDs are not highly generalizable and technical surveys are still lacking. Therefore, this paper presents a comprehensive review of IRLD. Specifically, after systematically reviewing geophysical well logging and machine learning techniques, the main applications and general processes for the cross-discipline task of IRLD are summarized. More importantly, the key challenges of IRLD in the four stages of data acquisition, feature engineering, model building, and practical application are discussed in this review. The potential risks of these challenges are visualized by using real logging data from a study area in the South China Sea and the example of a lithology identification task. For these challenges, we give the current state-of-the-art methods and feasible strategies in conjunction with published research. This comprehensive review is expected to provide insights for practitioners to construct more robust models and achieve more effective application results in IRLD. [ABSTRACT FROM AUTHOR]

Well logging can be classified under the general category of big data, as the datasets are intricate for conventional data processing application software to handle. This study aims at working on three entire wells drilled up to 1751–2690.4 m in upper Assam basin for lithofacies identification with each well comprising of huge datasets. Conventional methods for lithofacies classification are a challenging task for this enormous amount of data. It can be subject to biases, substantially laborious and time consuming. To deal with this problem, machine learning (ML) classification algorithms have come into picture, which systematically speed up accurate prediction of lithofacies. Five such algorithms (Support vector, Random forest, Multi-layer Perceptron, K Nearest Neighbors and Decision tree) are chosen and then evaluate each of the model's performance exhibiting varying hyperparameters and input features. We established a thorough approach that entails all classifiers operating concurrently within the same ML framework. The best algorithm was selected following the comparison of each model's Jaccard index and F1-score based on optimized hyperparameters. In addition, introduction of water saturation, a derived petrophysical property affects the model's performance. This feature selection process brags the importance of inclusion of useful features to improve the accuracy of any ML model's prediction capability. We observed that after introduction of water saturation profile, Jaccard index and F1-score of the best classification model, Random Forest has increased by on average 10.74% and 7.82%, respectively. Finally, we correlated identified lithofacies, including the reservoir facies from well to well in the study area. [ABSTRACT FROM AUTHOR]

Copyright of Natural Gas Geoscience is the property of Natural Gas Geoscience and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The spatial variability of properties in sedimentary deposits can be assessed using approximation methods. A small number of direct measurements or their extremely uneven distribution leads to significant model errors. This article explores the possibility of using multidimensional approximation and regression splines, both considering spatially referenced direct observation data and using well log curves statistically linked to the modeled variables. It is possible to significantly reduce the forecast error by utilizing indirect data. The results can be computed for individual wells as for inter-well space, allowing for the creation of geological cross-sections of predicted properties and 3D models of their distribution. In order to demonstrate the effectiveness of the proposed approach, computational experiments were conducted using data from the stratigraphic range of the Georgievskaya and Vasyuganskaya formations in the Kazan field in southeastern West Siberia. Comparing the obtained forecast with the real, unknown values of the modeled variable at the time of its implementation showed a high quality model with determination coefficients up to 0.9.

Recent experimental investigations have demonstrated that a substantial amount of H−/D− ions can be formed by thermal desorption processes. Based on these new findings, new mini axial and coaxial-type neutron tubes have been developed for the production of high or low-energy neutrons via the d-d, d-10B, d-7Li or p-7Li nuclear reactions. By operating these mini neutron tubes with a high frequency AC high-voltage supply, short pulses of high intensity neutron beams can be generated. Multiple applications, such as carbon and well logging, neutron imaging, cancer therapy, medical isotope production, fission reactor start-up, fusion reactor material evaluation, homeland security and space exploration can be performed with the subcompact neutron generator system. It is shown that the performance of these new mini neutron tubes can exceed those of the conventional plasma-based neutron sources. [ABSTRACT FROM AUTHOR]

Evaluating irreducible water saturation is crucial for estimating reservoir capacity and developing effective extraction strategies. Traditional methods for predicting irreducible water saturation are limited by their reliance on specific logging data, which affects accuracy and applicability. This study introduces a predictive method based on fractal theory and deep learning for assessing irreducible water saturation in complex carbonate reservoirs. Utilizing the Mishrif Formation of the Halfaya oilfield as a case study, a new evaluation model was developed using the nuclear magnetic resonance (NMR) fractal permeability model and validated with surface NMR and mercury injection capillary pressure (MICP) data. The relationship between the logarithm mean of the transverse relaxation time (T2lm) and physical properties was explored through fractal theory and the Thomeer Function. This relationship was integrated with conventional logging curves and an advanced deep learning algorithm to construct a T2lm prediction model, offering a robust data foundation for irreducible water saturation evaluation. The results show that the new method is applicable to wells with and without specialized NMR logging data. For the Mishrif Formation, the predicted irreducible water saturation achieved a coefficient of determination of 0.943 compared to core results, with a mean absolute error of 2.37% and a mean relative error of 8.46%. Despite introducing additional errors with inverted T2lm curves, it remains within acceptable limits. Compared to traditional methods, this approach provides enhanced predictive accuracy and broader applicability. [ABSTRACT FROM AUTHOR]

In the process of lithology discrimination from a conventional well logging dataset, the imbalance in sample distribution restricts the accuracy of log identification, especially in the fine-scale reservoir intervals. Enhanced sampling balances the distribution of well logging samples of multiple lithologies, which is of great significance to precise fine-scale reservoir characterization. This study employed data over-sampling and under-sampling algorithms represented by the synthetic minority over-sampling technique (SMOTE), adaptive synthetic sampling (ADASYN), and edited nearest neighbors (ENN) to process well logging dataset. To achieve automatic and precise lithology discrimination on enhanced sampled well logging dataset, support vector machine (SVM), random forest (RF), and gradient boosting decision tree (GBDT) models were trained using cross-validation and grid search methods. Aimed to objectively evaluate the performance of different models on different sampling results from multiple perspectives, the lithology discrimination results were evaluated and compared based on the Jaccard index and F1 score. By comparing the predictions of eighteen lithology discrimination workflows, a new discrimination process containing ADASYN, ENN, and RF has the most precise lithology discrimination result. This process improves the discrimination accuracy of fine-scale reservoir interval lithology, has great generalization ability, and is feasible in a variety of different geological environments. [ABSTRACT FROM AUTHOR]

This article reports on the successful downhole testing of the prototype permeability logging instrument, developed by the Li Ning Innovation Team of PetroChina. The test was carried out on 3 March 2024 in the Ren 91 standard well in Huabei (North China). In this article, we analysed the instrument characteristics of porosity logging measurement, saturation logging measurement, and permeability logging measurement in three stages of development, as well as the current research status of continuous depth-permeability measurement in the wellbore formations. We highlighted the basic principle of permeability logging measurement with Stoneley waves and pointed out the main challenges in the development of permeability logging measurement devices and on-site application research. [ABSTRACT FROM AUTHOR]

Well logging is a significant method of geological formation description and resource assessment in the exploration and development of oil, natural gas, minerals, groundwater, and sub-surface thermal energy, as well as geotechnical engineering and environmental research. However, the challenging problem of estimating well logging data always exists because well logs can only be measured through a drilling process involving costly and time-consuming field trials. This study provides a brief overview of Deep Learning (DL) models and examines the well-logging issues in estimating missing well-log data. In addition, it discusses a literature review focusing on the application of DL models for well-log estimation. The outcome of this exploratory work necessitates suitable requirements for the design and implementation. [ABSTRACT FROM AUTHOR]

Purpose. The paper concerns additional geological appraisal and identification of extra localization criteria of cupriferous mineralization of a stratiform type. The research is intended to define physical and geological factors influencing sandstone distribution within the north-end of Zhezkazgan synclinal in the Central Kazakhstan. Methods. A complex analysis of geological and mineralogical, geophysical, and geochemical data was carried out relying upon the information obtained from deep-hole wells. Modern methods to process and interpret the field data were applied. The analysis involved lithological dismemberment of the productive Taskuduk suite as well as determination of sulfide mineralization zone boundaries. Findings. The research has shown that the productive levels reach down to 1500-m depth. The data interpretation has helped specify the geological structure and develop a model of ore-promising bodies of cupriferous sandstones at great depth. The applied complex data analysis has increased potential to prospect new loci of productive cupriferous sandstone deposits of Zhezkazgan type. Originality. The research has contributed significantly to understanding of the physical and geological factors influencing the copper ore distribution in deep layers. New physical and geological criteria, determining deep copper mineralization localization, have been identified which were not considered by earlier similar studies. Practical implications. The obtained new data as well as the developed methods are important while deep drilling planning and implementing to identify promising copper mineralization sites. The findings may be used to optimize exploration within the north-end of Zhezkazgan synclinal, and other districts having analogous geological conditions. [ABSTRACT FROM AUTHOR]

Recent experimental investigations have demonstrated that a substantial amount of H−/D− ions can be formed by thermal desorption processes. Based on these new findings, new mini axial and coaxial-type neutron tubes have been developed for the production of high or low-energy neutrons via the d-d, d-10B, d-7Li or p-7Li nuclear reactions. By operating these mini neutron tubes with a high frequency AC high-voltage supply, short pulses of high intensity neutron beams can be generated. Multiple applications, such as carbon and well logging, neutron imaging, cancer therapy, medical isotope production, fission reactor start-up, fusion reactor material evaluation, homeland security and space exploration can be performed with the subcompact neutron generator system. It is shown that the performance of these new mini neutron tubes can exceed those of the conventional plasma-based neutron sources.

Abstract The Peshawar Basin is a part of the lower Himalayas that contains an enormous amount of groundwater storage. The evaluation of groundwater potential in the southern Peshawar district was done using well logging, lithostratigraphic properties, and combined hydrogeological and geophysical techniques. A total of 13 Vertical Electrical Sounding (VES) profiles were utilised to assess potential groundwater zones for surface resistivity studies. The aquifer system was delineated by comparing the data from five boreholes with the VES findings. An exploration of super-saturated groundwater potential was conducted, utilising parameters such as transmissivity (T), hydraulic conductivity (K), storativity, and the Dar Zarrouk analysis. The Dar Zarrouk analysis yielded average values of transverse resistance (TR), longitudinal conductance (S), and anisotropy (λ), which were determined to be 8069.12, 0.51, and 0.561, respectively. Similarly, average values of transmissivity (T), hydraulic conductivity (K), and storativity were obtained, resulting in 28.67, 0.24, and 0.000177, respectively. The saturated confined layer, characterized by highly saturated zones, was identified to begin at a depth of approximately 119 m and extend down to the lower boundary of the aquifer. The examined aquifer is composed of clay, sand, gravel, boulders, and loose layers of lacustrine mud that are interlayered to form an unconsolidated groundwater aquifer system. The aquifers in the region are highly developed and consisted of unconfined, semi-confined, and confined aquifer systems. As a result, it is possible to use the aquifer for groundwater development in the study area because of its low -to-medium discharge.

The displacement efficiency and residual oil saturation are the most important characteristics determining the recoverable volumes of oil and the waterflooding efficiency. The main data about these parameters are provided by laboratory core studies. Specialized field tests make it possible to evaluate the same characteristics in-situ, directly on a well, thus excluding the influence of scale effect and impact of core sampling and preparation.This paper presents the results of a field study to evaluate the displacement efficiency and residual oil saturation on a well draining terrigenous Devonian formations on one of the areas of the Romashkinskoye field. The results are compared with generalized dependencies based on core data. Significant influence of the scale effect (the effect of sweep multiplicity) on the estimated parameter values was noted, which indicates the need to supplement laboratory experiments with specialized field studies.

The basis for solving almost all geological problems are models, which are a reflection of the structure of deposits. This places high demands on the detail and accuracy of the models.One of the most important parameters in the calculation of reserves is the oil saturation factor, observed in most cases using Dakhnov-Archie models. Despite the widespread use of this technique, it has a number of significant limitations. An alternative way to assess oil saturation in this case is to use a capillary model obtained from the results of capillary measurements of the core.The capillary model is a continuous multidimensional function of the dependence of the water saturation coefficient on the free water level, porosity and permeability. A well-compiled capillary model is a reflection of the geological model of the reservoir – with information about the water saturation of the reservoirs by the altitude of the structure, the levels of fluid contacts, fluid densities, surface and reservoir properties.This paper discusses the capabilities of capillary models, which are widely used in three-dimensional modeling of the degree of saturation of the interwell space, the calculation of the volume of oil-saturated rocks, geological reserves of hydrocarbons, as well as dynamic characteristics in the construction of hydrodynamic models.

The proposed method of predicting spatial distribution of leaching zones in carbonate reservoirs includes integrated interpretation of well logging data, including electrical microimages (Formation MicroImager – FMI), together with analysis of lithofacies within depositional cycles. Based on the comparison of FMI data with results of lithological and petrophysical studies of the core, an atlas of diagenetic porosity images for the studied formation was developed. Choquette and Pray international classification was used for pore typing. Applying the developed atlas to FMI logs, karstification zones were identified and classified in reference wells. In wells, where only standard well loggings was performed, karstification zones were identified using a developed decision tree. Sedimentological and sequence stratigraphic analysis made it possible to identify sedimentation cycles and to correlate the main lithofacies within the target reservoir. It was established that microbial and shallow facies form the most productive part of the reservoir. These facies with initially high primary porosity were subjected to the most intensive karstification along sequence stratigraphic boundaries due to subaerial exposure. The established relationships between leaching zones, lithofacies, and sequence-stratigraphic boundaries made it possible to carry out inter-well correlation of karstified intervals. The correlation will allow construction of leaching zones maps, that will be further applied for geological modeling.

Nowadays well logging curves are interpreted by geologists who preprocess the data and normalize the curves for this purpose. The preparation process can take a long time, especially when hundreds and thousands of wells are involved. This paper explores the applicability of Machine Learning methods to geology tasks, in particular the problem of lithology interpretation using well-logs, and also reveals the issue of the quality of such predictions in comparison with the interpretation of specialists. The authors of the article deployed three groups of Machine Learning algorithms: Random Forests, Gradient Boosting and Neural Networks, and also developed its own metric that takes into account the geological features of the study area and statistical proximity of lithotypes based on log curves values.As a result, it was proved that Machine Learning algorithms are able to predict lithology from a standard set of well logs without calibration on reference layers, which significantly saves time spent on preliminary preparation of curves.

Empirical methods often fail to accurately depict in-situ gas hydrate saturation distributions, despite their relationships with petrophysical and elastic properties remaining partially unclear. We proposed a data-driven approach to estimate gas hydrate saturation employing several machine learning techniques, including radial basis function neural network (RBFNN), random forest (RF), extreme gradient boosting (XGBoost), Adaptive Boosting (AdaBoost), support vector machines (SVM), and k-nearest neighbors (kNN). This study involved pre-processing data from laterolog deep resistivity and p-wave velocity logs, defining their increments as differences from the lowest values in gas hydrate zones. We identified data-driven patterns between pairs of laterolog deep resistivity and p-wave velocity increments, as well as core information corroborated with the traditionally predicted gas hydrate saturations, by adopting machine learning (ML) approaches. The approach tested on four wells in the Krishna-Godavari (KG) offshore basin (India) is extremely feasible. During the training and test phases, the minimum correlation coefficient between the true and predicted responses exceeds 0.94 and 0.88, respectively. The model accuracy hierarchy was RBFNN > AdaBoost > RF > XGBoost > KNN > SVM during training, and AdaBoost > XGBoost > RF > RBFNN > KNN > SVM during testing. This approach allows interpreters to select the most accurate ML model based on training phase performance. The proposed ML-based method is efficient, synergising p-wave and resistivity data increment, significantly improving gas hydrate saturation predictions, and avoiding the complexities of traditional calculations. The study indicates that gas hydrate saturation in the Krishna-Godavari region ranges from 0.17 to 86.84%. [ABSTRACT FROM AUTHOR]

The lithofacies analysis of logging data is an essential step in reservoir evaluation. Multiresolution graph-based clustering (MRGC) is a commonly used methodology that provides information on the best number of clusters and cluster fitting results for geological understanding. However, the cluster fusion approach of MRGC often leads to an overemphasis of the boundary constraints among clusters. MRGC neglects the global cluster distribution relationship, which limits its practical application effectiveness. This paper proposes a new methodology, named kernel multiresolution graph-based clustering (KMRGC), to improve the merging part of clustering in MRGC, and it can give more weight to the spatial relationship characteristics among clusters. The clustering performance of K-means, Gaussian Mixture Model(GMM), fuzzy c-means(FCM), Density-Based Spatial Clustering of Applications with Noise(DBSCN), spectral clustering, MRGC and KMRGC algorithm was evaluated on a publicly available training set and noisy dataset, and the best results in terms of the adjusted Rand coefficients and normalized mutual information(NMI) coefficients on most of the datasets were obtained using KMRGC algorithm. Finally, KMRGC was used for logging data lithofacies clustering in cased wells, and the clustering effect of KMRGC algorithm was much better than that of the K-means, GMM, FCM, DBSCN, spectral clustering and MRGC algorithms, and the accuracy and stability were better. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Petroleum Research & Studies is the property of Republic of Iraq Ministry of Higher Education & Scientific Research (MOHESR) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In order to predict the reservoir productivity, solve the problem of weak log response characteristics caused by complex pore structure and low productivity prediction accuracy by conventional logging methods. The method of crossplots was used to investigate the relationship between the enveloped area of well log curves in the clastic rock reservoir of the Xihu Sag in the East China Sea Basin and productivity. The sensitivity envelope area of productivity was selected as the productivity prediction index. Based on the principle of wave superposition, the envelope area productivity prediction index was constructed, the entropy weight method was used to determine the weight of each prediction index, and then the productivity prediction model was established by reservoir type. The results show that the four envelope areas of compensation neutron permeability, resistivity density, resistivity apparent wave impedance and apparent wave impedance acoustic wave lag have strong sensitivity to productivity. The accuracy of the productivity prediction model established is high, and the average error is 19. 4%. It is concluded that the productivity prediction model based on envelope area can be used to predict the productivity of oil and gas reservoirs without dynamic test, which provides a new idea for the oil and gas reservoir productivity prediction method. [ABSTRACT FROM AUTHOR]

As the petroleum industry increasingly exploits unconventional reservoirs with low permeability and porosity, accurate predictions of post-fracture production are becoming critical for investment decisions, energy policy development, and environmental impact assessments. However, despite extensive research, accurately forecasting post-fracture production using well-log data continues to be a complex challenge. This study introduces a new method of data volume expansion, which is to subdivide the gas production of each well on the first day according to the depth of logging data, and to rely on the correlation model between petrophysical parameters and gas production to accurately combine the gas production data while matching the accuracy of the well-log data. Twelve pipelines were constructed utilizing a range of techniques to fit the regression relationship between logging parameters and post-fracture gas production These included data preprocessing methods (StandardScaler and RobustScaler), feature extraction approaches (PCA and PolynomialFeatures), and advanced machine learning models (XGBoost, Random Forest, and neural networks). Hyperparameter optimization was executed via GridSearchCV. To assess the efficacy of diverse models, metrics including the coefficient of determination (R2), standard deviation (SD), Pearson correlation coefficient (PCC), mean absolute error (MAE), mean squared error (MSE), and root-mean-square error (RMSE) were invoked. Among the several pipelines explored, the PFS-NN exhibited excellent predictive capability in specific reservoir contexts. In essence, integrating machine learning with logging parameters can be used to effectively assess reservoir productivity at multi-meter formation scales. This strategy not only mitigates uncertainties endemic to reservoir exploration but also equips petroleum engineers with the ability to monitor reservoir dynamics, thereby facilitating reservoir development. Additionally, this approach provides reservoir engineers with an efficient means of reservoir performance oversight. [ABSTRACT FROM AUTHOR]

The identification of lithology is a crucial step in determining the characteristics of petroleum reservoirs, and many studies have investigated the application of deep neural networks in lithology identification. However, incorrect lithology labels as a result of manual interpretation can seriously affect network training when deep learning is used to identify lithology from well logging data. To address this problem, a method of learning with noisy labels (probabilistic end-to-end noise correction in labels, PENCIL) is applied to the network training process. Experiments are conducted on two real well logging datasets, and two types of label noise, random and pattern, are added to the lithology labels of the training data to simulate the lithology label noise that may exist in the actual data. To demonstrate the effect of this method, this study trains four network models, namely residual network (ResNet), bidirectional gated recurrent unit (Bi-GRU), ResNet-PENCIL, and Bi-GRU-PENCIL. The results of the experiments show that pattern label noise has a more serious effect on network training than random label noise, and network models that use the PENCIL framework effectively mitigate the effect of incorrect lithology labels on lithology identification results. [ABSTRACT FROM AUTHOR]

The Peshawar Basin is a part of the lower Himalayas that contains an enormous amount of groundwater storage. The evaluation of groundwater potential in the southern Peshawar district was done using well logging, lithostratigraphic properties, and combined hydrogeological and geophysical techniques. A total of 13 Vertical Electrical Sounding (VES) profiles were utilised to assess potential groundwater zones for surface resistivity studies. The aquifer system was delineated by comparing the data from five boreholes with the VES findings. An exploration of super-saturated groundwater potential was conducted, utilising parameters such as transmissivity (T), hydraulic conductivity (K), storativity, and the Dar Zarrouk analysis. The Dar Zarrouk analysis yielded average values of transverse resistance (TR), longitudinal conductance (S), and anisotropy (λ), which were determined to be 8069.12, 0.51, and 0.561, respectively. Similarly, average values of transmissivity (T), hydraulic conductivity (K), and storativity were obtained, resulting in 28.67, 0.24, and 0.000177, respectively. The saturated confined layer, characterized by highly saturated zones, was identified to begin at a depth of approximately 119 m and extend down to the lower boundary of the aquifer. The examined aquifer is composed of clay, sand, gravel, boulders, and loose layers of lacustrine mud that are interlayered to form an unconsolidated groundwater aquifer system. The aquifers in the region are highly developed and consisted of unconfined, semi-confined, and confined aquifer systems. As a result, it is possible to use the aquifer for groundwater development in the study area because of its low -to-medium discharge. [ABSTRACT FROM AUTHOR]

The North Hebei Plain is a typical example of a deep-covered area, with a thickness of more than 200 m Quaternary loose sediments. Previous studies have tried various stratigraphic classification methods but lack an effective combination of geophysical methods. Based on the analysis of rock strata and dating tests, combined with geophysical well-logging data, shallow seismic reflection, and paleomagnetic dating, the Quaternary strata of the deep-covered area in the Hebei Plain was divided into the Lower Pleistocene series, the Middle Pleistocene series, the Upper Pleistocene series, and the Holocene series. By observing the cores of the Quaternary borehole QHJ02, this paper focused on dividing its loose sediments into eight lithological sections, according to the lithology, sediment colour, sedimentary assemblage, and sedimentary environment. The geomagnetic polarity belt is divided according to the paleomagnetic inclination of the paleomagnetic sample in borehole QHJ02. The results of the paleomagnetic test show that the interval of 0–101.3 m is Brunhes normal polarity chron, the interval from 101.3 to 219.5 m is Matuyama reversed polarity chron, and the interval from 219.5 to 350 m is Gauss normal polarity chron. The amplitude of the geophysical logging curve shows a good correspondence with the lithology of the QHJ02 borehole. The lithology changes can be identified through the abrupt abnormal positions of logging curves such as gamma ray, resistivity, and magnetization to analyze the sedimentary environment combination and determine the formation boundary. According to the combination of paleomagnetic, shallow seismic reflection, and geophysical well logging methods, we can determine the bottom boundaries of the Quaternary and middle Pleistocene are 219.5 m and 101.3 m, respectively. According to the combination of geophysical well-logging, lithostratigraphic assessment, dating experiment of Accelerator Mass Spectrometry 14C (AMS 14C), and Optically Stimulated Luminescence (OSL), we can determine the bottom of the upper Pleistocene and the lower limit of Holocene to be at 45.4 m and 21.4 m, respectively. Based on a comprehensive analysis of rock stratigraphic characteristics and well-logging curves, this study investigated the paleoenvironmental characteristics of borehole QHJ02 at different ages. The Upper Neogene was characterized by branch channels and river floodplains, with localized occurrences of alluvial fans, avulsion fans, and inter-channel lowlands. During the Early Pleistocene, branch channels and inter-channel lowlands dominated, with localized river floodplains. The Middle Pleistocene witnessed alternating branch channels and river floodplains as the main paleoenvironments. In the Late Pleistocene, branch channels and inter-channel lowlands were prevalent. The Holocene saw a combination of natural levees and river flood-plain environments. These findings enhance our understanding of the geological evolution and paleoenvironmental changes in the North Hebei Plain. [ABSTRACT FROM AUTHOR]

The lithology and pore structure of conglomerate reservoir in the Mahu sag, China, are complex, which makes it difficult to identify reservoir fluids. In this study, the body relaxation characteristics of crude oil at different temperatures were obtained from block crude oil samples, and two simulated oil samples with different body relaxation characteristics were used to conduct experiments. The saturated water, bound water, and saturated oil states of some conglomerate samples were quickly established by vacuum compression saturation, high-speed centrifugation, and vacuum saturation simulation of oil, and the fluid and different simulated oil-water saturation states were analyzed by nuclear magnetic resonance (NMR) experiments. The experimental results showed that the distribution range of the water-saturated core NMR T2 spectrum is mainly affected by surface relaxation. The NMR spectra of saturated oil with different viscosity values are different from those of saturated water because of the relaxation property of the oil phase. The samples of this experiment indicated that the difference is more obvious when the oil is thinner. The distribution pattern of the oil-bearing NMR results was affected by the relaxation of the oil phase, surface relaxation, pore structure, and wettability. With consideration of the relaxation characteristics of the oil samples, the NMR T2 spectra of two oil-saturated conglomerates were analyzed by multi-component Gaussian fitting, and the relaxation signals of the oil phase were evaluated quantitatively.

Copyright of Reviews of Geophysics & Planetary Physics is the property of Editorial Office of Reviews of Geophysics & Planetary Physics and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Well logging data shows the change of physical properties of rocks and fluids in lithology units with depth. Well logging is one of the main parts of natural resources exploration. But in some cases, due to the lack of geophysical equipment or due to high exploration costs, it is not possible to record some geophysical logs. In this paper, electromagnetic log predicted using electrical logs for the first time. In such cases, estimating the desired log using other geophysical logs is a suitable solution. For the estimation of geophysical logs, machine learning algorithms are used in most cases. In this research, a new strategy developed for processing and preparation of geophysical logs. This strategy consists of three parts: data smoothing, correlation intensifier, and MLR (Multiple Linear Regression) or ANN (Artificial Neural Network). The purpose of the data smoothing and correlation intensifier section is to remove outliers and identify the pattern of main changes in the log data, and as a result, the accuracy in estimating the target log increases. In this article, the determination of the electromagnetic log has been done using electric logs. The well logging data have been recorded in Southern California and the Central Valley. A total of six wells have been selected, four wells for MLR and ANN training and two wells for testing. By applying data smoothing and correlation intensifier to these data, the correlation between electrical and electromagnetic data increased significantly and caused the estimation accuracy of electromagnetic log to be above 95%. The use of this strategy is not limited to the estimation of electromagnetic log and can be used in all well logging data. [ABSTRACT FROM AUTHOR]

Well logging dapat digunakan untuk mengidentifikasi zona produktif, membedakan antara minyak, gas, atau air di dalam reservoar, dan memperkirakan cadangan hidrokarbon. Salah satu metode dalam pendekatan parameter karakterisasi reservoir yang cukup efektif adalah dengan menggunakan analisis petrofisika. Pada penelitian ini melakukan analisis petrofisika berupa analisis kualitatif serta analisis kuantitatif dengan tujuan untuk mengetahui nilai volume shale, porosity, dan saturasi air, kemudian menghitung ketebalan reservoir berdasarkan nilai net pay. Berdasarkan analisis kualitatif didapatkan 2 zona reservoir yaitu pada interval kedalaman 1112-1116m dan 1211-1219m. Berdasarkan analisis kuantitatif dihasilkan nilai volume shale pada zona reservoir A dan B yaitu 17,7 – 26,9% dan 8,5 – 16,6%, kemudian nilai porositas total pada zona reservoir A dan B yaitu 16,5 – 22,7% dan 23,1 – 25,9% sedangkan nilai porositas efektif pada zona reservoir A dan B yaitu 15 – 20,6% dan 22,6 – 25.7%. Kemudian, untuk nilai saturasi air total pada zona reservoir A dan B yaitu 34,4 - 42% dan 28,6 – 36,4%, sedangkan untuk nilai saturasi air efektif pada zona reservoir A dan B yaitu 23,1 – 25,9% dan 22,6 – 25,7%. Kemudian didapatkan ketebalan reservoir yang berisi hidrokarbon (netpay) yaitu sebesar 3,058 m yang berada pada Formasi Tualang. Maka zona reservoir yang memiliki prospek hidrokarbon terdapat pada zona reservoir A pada Formasi Tualang.

In the process of lithology discrimination from a conventional well logging dataset, the imbalance in sample distribution restricts the accuracy of log identification, especially in the fine-scale reservoir intervals. Enhanced sampling balances the distribution of well logging samples of multiple lithologies, which is of great significance to precise fine-scale reservoir characterization. This study employed data over-sampling and under-sampling algorithms represented by the synthetic minority over-sampling technique (SMOTE), adaptive synthetic sampling (ADASYN), and edited nearest neighbors (ENN) to process well logging dataset. To achieve automatic and precise lithology discrimination on enhanced sampled well logging dataset, support vector machine (SVM), random forest (RF), and gradient boosting decision tree (GBDT) models were trained using cross-validation and grid search methods. Aimed to objectively evaluate the performance of different models on different sampling results from multiple perspectives, the lithology discrimination results were evaluated and compared based on the Jaccard index and F1 score. By comparing the predictions of eighteen lithology discrimination workflows, a new discrimination process containing ADASYN, ENN, and RF has the most precise lithology discrimination result. This process improves the discrimination accuracy of fine-scale reservoir interval lithology, has great generalization ability, and is feasible in a variety of different geological environments.

Abstract Geomechanical (GM) parameters play a significant role in geomechanical studies. The calculation of GM parameters by analyzing finite rock samples is very limited. The GM parameters show a nonlinear trend; thus, applying empirical relationships is unreliable to predict their quantities. Machine learning (ML) methods are generally used to improve the estimation of such parameters. Recent researches show that ML methods can be useful for estimating GM parameters, but it still requires analyzing different datasets, especially complex geological datasets, to emphasize the correctness of these methods. Therefore, the aim of this study is to provide a robust recombinant model of the ML methods, including genetic algorithm (GA)–multilayer perceptron (MLP) and genetic algorithm (GA)–radial basis function (RBF), to estimate GM parameters from a complex dataset. To build ML models, 48,370 data points from six wells in the complicated Norwegian Volve oil field are used to train GA–MLP and GA–RBF methods. Moreover, 20,730 independent data points from another three wells are used to verify the GM parameters. GA–MLP predicts GM parameters with the root-mean-squared error (RMSE) of 0.0032–00079 and coefficient determination (R 2) of 0.996–0.999. It shows similar prediction accuracy when used to an unseen dataset. Comparing the results indicates that the GA–MLP model has better accuracy than the GA–RBF model. The results illustrate that both GA–MLP and GA–RBF methods perform better at estimating GM parameters compared to empirical relationships. Concerns about the integrity of the methods are indicated by assessing them on another three wells.

Fracture porosity is crucial for storage and production efficiency in fractured tight reservoirs. Geophysical image logs using resistivity measurements have traditionally been used for fracture characterization. This study aims to develop a novel, hybrid machine-learning method to predict fracture porosity using conventional well logs in the Ahnet field, Algeria. Initially, we explored an Artificial Neural Network (ANN) model for regression analysis. To overcome the limitations of ANN, we proposed a hybrid model combining Support Vector Machine (SVM) classification and ANN regression, resulting in improved fracture porosity predictions. The models were tested against logging data by combining the Machine Learning approach with advanced logging tools recorded in two wells. In this context, we used electrical image logs and the dipole acoustic tool, which allowed us to identify 404 open fractures and 231 closed fractures and, consequently, to assess the fracture porosity. The results were then fed into two machine-learning algorithms. Pure Artificial Neural Networks and hybrid models were used to obtain comprehensive results, which were subsequently tested to check the accuracy of the models. The outputs obtained from the two methods demonstrate that the hybridized model has a lower Root Mean Square Error (RMSE) than pure ANN. The results of our approach strongly suggest that incorporating hybridized machine learning algorithms into fracture porosity estimations can contribute to the development of more trustworthy static reservoir models in simulation programs. Finally, the combination of Machine Learning (ML) and well log analysis made it possible to reliably estimate fracture porosity in the Ahnet field in Algeria, where, in many places, advanced logging data are absent or expensive.

Geomechanical (GM) parameters play a significant role in geomechanical studies. The calculation of GM parameters by analyzing finite rock samples is very limited. The GM parameters show a nonlinear trend; thus, applying empirical relationships is unreliable to predict their quantities. Machine learning (ML) methods are generally used to improve the estimation of such parameters. Recent researches show that ML methods can be useful for estimating GM parameters, but it still requires analyzing different datasets, especially complex geological datasets, to emphasize the correctness of these methods. Therefore, the aim of this study is to provide a robust recombinant model of the ML methods, including genetic algorithm (GA)–multilayer perceptron (MLP) and genetic algorithm (GA)–radial basis function (RBF), to estimate GM parameters from a complex dataset. To build ML models, 48,370 data points from six wells in the complicated Norwegian Volve oil field are used to train GA–MLP and GA–RBF methods. Moreover, 20,730 independent data points from another three wells are used to verify the GM parameters. GA–MLP predicts GM parameters with the root-mean-squared error (RMSE) of 0.0032–00079 and coefficient determination (R2) of 0.996–0.999. It shows similar prediction accuracy when used to an unseen dataset. Comparing the results indicates that the GA–MLP model has better accuracy than the GA–RBF model. The results illustrate that both GA–MLP and GA–RBF methods perform better at estimating GM parameters compared to empirical relationships. Concerns about the integrity of the methods are indicated by assessing them on another three wells. [ABSTRACT FROM AUTHOR]

90% of extracted oil-gas condensate is made up of productive layer sediments. A more detailed study of these sediments is of great theoretical and important practical importance. For this reason, studying its characteristics, determining the composition, conditions of formation, and origin of the collected sediments is one of the main issues. All natural events (catastrophes) have a unique effect on the sediment accumulation process. Through facies analysis, it is possible to recreate the depositional conditions. At the same time, facies analysis allows determining the development process of rocks with different lithological composition, especially rock-collectors, as well as places rich in minerals. [ABSTRACT FROM AUTHOR]

Unconventional reservoirs, including gas shales and tight gas sands, have gained prominence in the energy sector due to technological advancements and escalating energy demands. The oil industry is eagerly refining techniques to decipher these reservoirs, aiming to reduce data collection costs and uncertainties in reserve estimations. Characteristically, tight reservoirs exhibit low matrix porosity and ultra-low permeability, necessitating artificial stimulation for enhanced production. The efficacy of the stimulation hinges on the organic material distribution, the rock's mechanical attributes, and the prevailing stress field. Comprehensive petrophysical analysis, integrating standard and specialized logs, core analyses, and dynamic data, is pivotal for a nuanced understanding of these reservoirs. This ensures a reduction in prediction uncertainties, with parameters like shale volume, porosity, and permeability being vital. This article delves into an intricate petrophysical evaluation of the Nene field, a West African unconventional reservoir. It underscores the geological intricacies of the field, the pivotal role of data acquisition, and introduces avant-garde methodologies for depth matching, rock typing, and the estimation of permeability. This research highlights the significance of unconventional reservoir exploration in today's energy milieu, offering a granular understanding of the Nene field's geological challenges and proffering a blueprint for analogous future endeavours in unconventional reservoirs. [ABSTRACT FROM AUTHOR]

The photoelectric factor (PEF) is an important well-logging tool to distinguish between different types of reservoir rocks because PEF measurement is sensitive to elements with high atomic numbers. Furthermore, the ratio of rock minerals could be determined by combining PEF log with other well logs. However, PEF logs could be missing in some cases such as in old well logs and wells drilled with barite-based mud. Therefore, developing models for estimating missing PEF logs is essential in those circumstances. In this work, we developed various machine learning models to predict PEF values using the following well logs as inputs: bulk density (RHOB), neutron porosity (NPHI), gamma ray (GR), compressional and shear velocity. The predictions of PEF values using adaptive-network-fuzzy inference system (ANFIS) and artificial neural network (ANN) models have errors of about 16% and 14% average absolute percentage error (AAPE) in the testing dataset, respectively. Thus, a different approach was proposed that is based on the concept of automated machine learning. It works by automatically searching for the optimal model type and optimizes its hyperparameters for the dataset under investigation. This approach selected a Gaussian process regression (GPR) model for the accurate estimation of PEF values. The developed GPR model decreases the AAPE of the predicted PEF values in the testing dataset to about 10% AAPE. This error could be further decreased to about 2% by modeling the potential noise in the measurements using the GPR model. [ABSTRACT FROM AUTHOR]

The development of cost-effective methods for estimating hydraulic conductivity profiles has been an ongoing effort in the field of engineering practice, which can be used to increase availability to clarify the hydrogeological complexity of fractured rock aquifers for the aid of solving groundwater-related problems. A new methodology is presented, which combines electrical well logs, fluid conductivity logs, double-packer hydraulic tests, Archie's law, and the Kozeny–Carman-Bear equation to investigate relations between formation factor (F) and hydraulic conductivity (K). Available geophysical and hydraulic test data measured from 88 boreholes in fractured rock formations in Taiwan were collected to perform the correlation studies. The correlation investigation outcomes indicate that the established F-K relations have the potential to serve as the transformation function for estimating hydraulic conductivity through the geological directly. To improve F-K relations in response to the effect of clay mineralogy, two proposed clustering techniques (the natural gamma ray threshold method and the modified Archie's law method) successfully play an important role in filtering clayed data. The prevalence of clay content in most of Taiwan's fractured rock formations has been found, which implies that careful consideration of clay-related issues in complex geologic formations is essential while applying Archie's law theory. Finally, the predictive models for estimating hydraulic conductivity have been developed for three types of lithology (sandstone, schist, and slate). [ABSTRACT FROM AUTHOR]

As the petroleum industry increasingly exploits unconventional reservoirs with low permeability and porosity, accurate predictions of post-fracture production are becoming critical for investment decisions, energy policy development, and environmental impact assessments. However, despite extensive research, accurately forecasting post-fracture production using well-log data continues to be a complex challenge. This study introduces a new method of data volume expansion, which is to subdivide the gas production of each well on the first day according to the depth of logging data, and to rely on the correlation model between petrophysical parameters and gas production to accurately combine the gas production data while matching the accuracy of the well-log data. Twelve pipelines were constructed utilizing a range of techniques to fit the regression relationship between logging parameters and post-fracture gas production These included data preprocessing methods (StandardScaler and RobustScaler), feature extraction approaches (PCA and PolynomialFeatures), and advanced machine learning models (XGBoost, Random Forest, and neural networks). Hyperparameter optimization was executed via GridSearchCV. To assess the efficacy of diverse models, metrics including the coefficient of determination (R2), standard deviation (SD), Pearson correlation coefficient (PCC), mean absolute error (MAE), mean squared error (MSE), and root-mean-square error (RMSE) were invoked. Among the several pipelines explored, the PFS-NN exhibited excellent predictive capability in specific reservoir contexts. In essence, integrating machine learning with logging parameters can be used to effectively assess reservoir productivity at multi-meter formation scales. This strategy not only mitigates uncertainties endemic to reservoir exploration but also equips petroleum engineers with the ability to monitor reservoir dynamics, thereby facilitating reservoir development. Additionally, this approach provides reservoir engineers with an efficient means of reservoir performance oversight.