Search

Machine learning (ML) methods are among the most promising technologies with wide-ranging research opportunities, particularly in the field of education, where they can be used to enhance student learning outcomes. This study explores the potential of machine learning algorithms to build and train models using log data from the "3D Modeling" e-course on the Moodle platform at TTK University of Applied Sciences, Tallinn, Estonia. By predicting trends, identifying patterns, and optimizing resource allocation, machine learning can improve the efficiency of e-learning and provide students with tailored recommendations for acquiring relevant knowledge and skills. The results of the study show that machine learning algorithms can be used to process the available e-course log data, using the clickstream of e-course resources and for their automated processing. The results suggest potential applications in personalized course recommendations, prediction and dropout prevention strategies, resulting in a more effective and personalized educational experience. Future research will focus on improving models of available registration data, exploring and using advanced machine learning techniques to improve the accuracy and usefulness of predictions, and providing faster recommendations to help students navigate their studies more effectively. [ABSTRACT FROM AUTHOR]

Abstract Machine learning (ML) methods are among the most promising technologies with wide-ranging research opportunities, particularly in the field of education, where they can be used to enhance student learning outcomes. This study explores the potential of machine learning algorithms to build and train models using log data from the "3D Modeling" e-course on the Moodle platform at TTK University of Applied Sciences, Tallinn, Estonia. By predicting trends, identifying patterns, and optimizing resource allocation, machine learning can improve the efficiency of e-learning and provide students with tailored recommendations for acquiring relevant knowledge and skills. The results of the study show that machine learning algorithms can be used to process the available e-course log data, using the clickstream of e-course resources and for their automated processing. The results suggest potential applications in personalized course recommendations, prediction and dropout prevention strategies, resulting in a more effective and personalized educational experience. Future research will focus on improving models of available registration data, exploring and using advanced machine learning techniques to improve the accuracy and usefulness of predictions, and providing faster recommendations to help students navigate their studies more effectively.

The application of the basin simulation method in reconstructing the pressure evolution process is frequently constrained by the limited availability of measured data. In order to establish a valid relation between logging data and paleo-pressure, this paper proposes a novel approach to constrain the pressure reconstructed by basin simulation method. The Eaton index N corresponding to the paleo-pressure of each period is calculated using inversion method, based on measured paleo-pressure by fluid inclusion PVTx method and logging data, in accordance with the Eaton formula. Subsequently, this value is utilized to calculate the paleo-pressure of other wells. The paleo-pressure results are compared with those reconstructed by basin simulation method to validate the reliability of the findings. The findings demonstrate a strong concurrence between the paleo-pressure calculated using the mean value of index N for each period and the reconstructed paleo-pressure obtained through basin simulation methodology, with an average error margin of less than ± 5%. Therefore, utilizing log data to constrain the paleo-pressure reconstructed by basin simulation method is a viable approach due to its ease of accessibility, strong continuity, and broad applicability and potential for application. This is particularly advantageous in regions where acquiring paleo-pressure data poses challenges. The paleo-dynamic conditions exert a significant influence on the distribution of gas and water in tight sandstone gas reservoirs, providing valuable insights for identifying favorable areas. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Xinjiang University (Natural Science Edition) is the property of Xinjiang University 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 degree and amount of sand production in hydrate reservoirs is related to the selection of stable production processes, but there is currently a lack of quantitative sand production prediction research using real logging data and formation samples from hydrate reservoirs. To reveal the dynamic change characteristics of in-situ reservoirs during hydrate decomposition, and explore quantitative prediction methods for guiding production practice, it is conducted a series of numerical simulations and quantitative prediction experiments. The numerical simulations are carried out using different sand-out prediction methods by using hydrate logging data during drilling, while quantitative prediction experiments of water production and sand-out are carried out based on in-situ reservoir samples. Our experiments indicate that hydrate mining is facing a serious risk of sand-out. The particle transport in the reservoir changes from "large-channel seepage" to "umbrella seepage" and then to "uniform fine flow" as the replacement flow rate decreases. A quantitative prediction model for water and sand production is also established. As a result, our study can provide support for the advancement of technology for long-term stable production and sand control of hydrates, laying the groundwork for developing a stable production plan for natural gas hydrates in offshore areas and determining the optimal depressurisation method. [ABSTRACT FROM AUTHOR]

Reservoir in-situ stress is crucial in oil and gas exploration and development. With the transitional shale as the research object, the ultrasonic transmission test and uniaxial compression test of rock samples from different sampling angles were performed. Further, the transversely isotropic in-situ stress logging calculation method for transitional shale reservoirs was established based on the transversely isotropic constitutive relationship. Finally, the in-situ stress profile of the transitional shale reservoir in the study area was analyzed, and the reliability of the results was verified. The results indicated that the mechanical characteristics of rock and its acoustic velocity varied significantly depending on the sample angle, with some regularity in the changes between vertical and horizontal mechanical properties. The logging profile revealed that the transverse Young modulus was greater than the vertical, and the vertical Poisson's ratio was typically greater than the transverse. The suggested transversely isotropic model performed better than the conventional in-situ stress model at predicting the in-situ stress of transitional shale reservoirs, with an error of less than 2%. This demonstrated the applicability of the model's in-situ stress calculation of the transitional shale reservoir. [ABSTRACT FROM AUTHOR]

In response to the unclear drilling resistance characteristics of rocks in the Ji'yang Depression, low drilling efficiency of PDC drill bits, and difficulties in drill bit selection, this study selected rock samples from different depths in the area for indoor drilling resistance analysis testing. Based on logging data, a prediction model was established for drilling resistance characteristics parameters of the strata in the area, and a graph of drilling resistance characteristic parameters of the rocks in the area was drawn. The study showed that the uniaxial compressive strength of the strata rocks was 50–110 MPa, with a hardness of 500–1300 MPa, a plasticity coefficient ranging from 1 to 2, a rock drillability grade of 8–20, and an abrasiveness index of 5–20. Combining the analysis of on-site drilling bit failures, PDC drill bits adapted to the strata in the area were selected, and the mechanical drilling speed of the selected bits reached 12.58 m/h, successfully drilling through the target layer. The above research results are of guiding significance for understanding the reasons for the difficulty of drilling into the Jiyang Depression strata and for improving mechanical drilling speed and drill bit selection in this 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.)

Abstract Fracture pressure is an important reference for wellbore stability analysis and hydraulic fracturing. Considering the low prediction accuracy, significant deviations, and limited applicability of traditional methods for predicting formation fracture pressure, this paper proposes an intelligent prediction method for fracture pressure using conventional well logging data based on the Stacking ensemble algorithm. The base learners of the model include RF, KNN, and LSTM algorithms with low correlation. The meta-learner adopts the XGBoost algorithm. The effectiveness of the model is validated using the fracture pressure data from Dagang Oilfield. The prediction results indicate that the stacking algorithm outperforms individual algorithms. After optimization with genetic algorithm, the R2 of the stacking model is 0.989, RMSE is 0.009%, and MAE is 0.32%. The global sensitivity analysis results show that AC and DEN in the well logging data have higher sensitivity to the fracture pressure. When using intelligent fracture pressure prediction methods, it is essential to ensure the accuracy of AC and DEN data. The work demonstrates the reliability and effectiveness of the method proposed for the intelligent prediction of fracturing pressure using conventional well logging data through Stacking ensemble algorithm to overcome the limitations of traditional methods.

The degree and amount of sand production in hydrate reservoirs is related to the selection of stable production processes, but there is currently a lack of quantitative sand production prediction research using real logging data and formation samples from hydrate reservoirs. To reveal the dynamic change characteristics of in-situ reservoirs during hydrate decomposition, and explore quantitative prediction methods for guiding production practice, it is conducted a series of numerical simulations and quantitative prediction experiments. The numerical simulations are carried out using different sand-out prediction methods by using hydrate logging data during drilling, while quantitative prediction experiments of water production and sand-out are carried out based on in-situ reservoir samples. Our experiments indicate that hydrate mining is facing a serious risk of sand-out. The particle transport in the reservoir changes from “large-channel seepage” to “umbrella seepage” and then to “uniform fine flow” as the replacement flow rate decreases. A quantitative prediction model for water and sand production is also established. As a result, our study can provide support for the advancement of technology for long-term stable production and sand control of hydrates, laying the groundwork for developing a stable production plan for natural gas hydrates in offshore areas and determining the optimal depressurisation method.

Abstract The uniaxial compressive strength (UCS) and tensile strength (T0) are crucial parameters in field development and excavation projects. Traditional lab-based methods for directly measuring these properties face practical challenges. Therefore, non-destructive techniques like machine learning have gained traction as innovative tools for predicting these parameters. This study leverages machine learning methods, specifically random forest (RF) and decision tree (DT), to forecast UCS and T0 using real well-logging data sourced from a Middle East reservoir. The dataset comprises 2600 data points for model development and over 600 points for validation. Sensitivity analysis identified gamma-ray, compressional time (DTC), and bulk density (ROHB) as key factors influencing the prediction. Model accuracy was assessed using the correlation coefficient (R) and the absolute average percentage error (AAPE) against actual parameter profiles. For UCS prediction, both RF and DT achieved R values of 0.97, with AAPE values at 0.65% for RF and 0.78% for DT. In T0 prediction, RF yielded R values of 0.99, outperforming DT's 0.93, while AAPE stood at 0.28% for RF and 1.4% for DT. These outcomes underscore the effectiveness of both models in predicting strength parameters from well-logging data, with RF demonstrating superior performance. These models offer the industry an economical and rapid tool for accurately and reliably estimating strength parameters from well-logging data.

Lithology holds significant importance in reservoir evaluation and geological modeling. However, the complex relationship between logging and lithology leads to strong multisolutions in logging responses, resulting in inaccurate identification of traditional logging lithology methods. Given the impressive performance of deep learning in data classification, we delved further into the technology and presented a deep residual network for lithological classification. The deep residual shrinkage network (DRSN) model incorporates an attention mechanism and a soft thresholding strategy based on the residual network. The residual shrinkage mechanism is the core characteristic of DRSN. It enhances model sparsity by shrinking the weights in the residual block (soft threshold strategy), resulting in a simpler model, and mitigating the overfitting problem. To test the model, we selected data from two wells in the Tarim Oilfield, China. In this article, we innovatively use the Gram angle field (GAF) to convert 1-D logging parameters into 2-D images. These images are then input into the DRSN model, using the idea of image processing to tackle the lithology classification problem. GAF effectively captures time-series information and converts 1-D time-series data into a 2-D matrix representation. Furthermore, GAF enhances the ability to capture nonlinear structures and patterns in time-series data using trigonometric transformations. Experimental results demonstrate that the DRSN-GAF network outperforms both DRSN and convolutional neural network (CNN) networks in terms of accuracy. The lithology prediction tasks for the two wells achieve accuracies of 96.00% and 96.50%, respectively. [ABSTRACT FROM AUTHOR]

The total organic carbon (TOC) is an important index for evaluating the hydrocarbon generation capacity of shale gas reservoirs. It is crucial for accurate prediction of the geological “sweet spot” of shale gas reservoirs. The existing methods for predicting the TOC content of shale gas reservoirs are highly subjective and have weak generalization ability. The offshore shale gas reservoir in southern Sichuan was taken as the research object. By sorting out the logging data of the study block and the results of laboratory core analysis, logging parameters such as natural gamma and density were selected as the feature vectors for model training. BP neural network and support vector machine prediction models of total organic carbon content were established to analyze the differences between different models. The influence factors such as model feature combination and network result were analyzed. Finally, the predicted TOC results were compared with the real values. The results show that the prediction model based on BP neural network without energy spectrum logging data can reflect the nonlinear relationship between logging data and reservoir more truly, and could provide a new idea for TOC prediction. [ABSTRACT FROM AUTHOR]

Macrolithotypes control the non-homogeneity of coal reservoirs, and it is important to grasp their distribution patterns for the exploration and development of coalbed methane (CBM). Traditional evaluation of macrolithotypes often relies on drilling cores and downhole workings, which are costly and difficult for obtaining comprehensive physical properties of coal seams. In recent years, geophysical logging technology has been commonly applied for evaluating physical properties of coal reservoirs because of its high efficiency and economic features. Taking the Junlian block in Sichuan province as a case study, based on many CBM evaluation wells and data analysis, a logging prediction method for the macrolithotypes of a high-rank coal reservoir is established. The problem of "boundary effect" is solved by the Haar wavelet analysis. Through a Pearson correlation coefficient matrix combined with backpropagation neural network analysis, density, natural gamma ray, acoustic time, compensated neutron, and deep lateral resistivity are extracted as critical parameters for constructing the method of logging interpretation of macrolithotypes. Logging identification models of different macrolithotypes are established by principal components analysis. Based on the results of logging interpretation, the macrolithotypes in the study area are mainly semi-dull coal and semi-bright coal. The genesis of macrolithotypes is analyzed by trace elements and lithologic characteristics. The bright coal in the study area is formed in a sedimentary environment of low paleotemperature, high anaerobic, saline water, and low sea level. In contrast, dull coal is formed in a sedimentary environment of relatively high sea level, brackish water, relatively rich oxygen, and relatively high paleotemperature. [ABSTRACT FROM AUTHOR]

Traditional machine learning models have mainly been used to study geological logging data of a single sample point, ignoring the fact that logging data has a strong spatial correlation. In this study, we use convolutional neural network to extract single-point features, structural features, and multidimensional features from logging data and compare the identification effects of lithology identification models based on the three features. The identification model based on the multidimensional feature extraction achieves 77.94% correctness in the test set, which is the best result among the identification models based on CNN and the three machine learning models. Based on this feature extraction model, the feature fusion modules in U-net and feature pyramid are added respectively to build two feature fusion models to combine the features extracted from different convolutional layers and improve the effectiveness of the model. The model also introduces attention mechanism to improve the role of useful features in the model training process. The identification accuracy of the two feature fusion models, U-CNN and P-CNN, reached 79.67% and 80.02% on the test set, respectively, which verified the effectiveness of the feature fusion models for lithology identification in the study area. [ABSTRACT FROM AUTHOR]

An interpretability model for intelligent lithology identification is proposed, which utilizes Ensemble Learning Stacking, Permutation Importance (PI), and Local Interpretable Model-agnostic Explanations (LIME) techniques. The aim of this method is to provide more accurate geological information and scientific support for oil and gas resource exploration. Two logging datasets from the public domain were used as experiments, and support vector machine (SVM), random forest (RF), and naive Bayes (NB) were employed as base learners, while SVM was utilized as the meta learner for lithology classification via stacking algorithm. The accuracy of the model was verified using evaluation metrics such as Area Under Curve (AUC), precision, recall, and F1-score. The PI and LIME techniques were employed to explain the lithology identification model. The results indicate that the stacking algorithm produced the best indexes and highest prediction accuracy. With respect to overall interpretation, PHIND, GR, and RT were found to have the most significant influence on lithology identification in a natural gas protection area in the United States, while DEN, CAL, and PEF were observed to be the most influential variables for lithology identification in the Daqing Oilfield in China. From the perspective of a single sample, the LIME algorithm can provide a quantitative prediction probability and degree of influence of the characteristic variables. [ABSTRACT FROM AUTHOR]

The shale oil reservoirs in the Jimsar Sag are strongly heterogeneous. Therefore, it is of practical significance to relate the dynamic elastic stiffness to the static elastic stiffness under the assumption of transverse isotropy. The acquisition of stiffness parameters depends on lab measurements or Sonic Scanner logging. Aiming at the problem that the lab measurements results are limited and discontinuous, the acquisition of special logging data is difficult in the well site, a method using only the conventional well-log data to predict the dynamic elastic stiffnesses of shale is proposed. Inspired by ultrasonic experimental test method, the study simplified the types of tested waves and complemented missing datas. According to the experimental proof, we concluded that: (1) Under the assumption of transverse isotropy, using simplified Thmosen model and velocity conversion relationship can accurately calculate the P-wave and S-wave velocities in the 45° angle direction. (2) Compared with the conventional isotropic algorithm, specify this algorithm can better reflect real mechanical conditions of the formation, and better characterize the variation of mechanical parameters with depth. (3) Effective conversion models between dynamic elastic modulus and static elastic modulus in vertical and horizontal directions were obtained in some areas of Jimsar.

Accurate identification of oil and water layers is the basis of qualitative evaluation of reservoir fluid properties or industrial value and selection of testing layers of the well. The traditional oil and water layer identification is mainly based on the extensive use of the well's logging and logging data, which is inefficient and easy to leak interpretation or misinterpretation for those reservoirs with complex geological conditions. In this paper, the random forest method of machine learning is used to select the lithology, porosity, permeability, movable fluid, oil saturation, S0, S1, S2, Tmax of rock as characteristics; smote oversampling is used to expand the sample, and the packet estimation is used to establish the oil and water layer identification model. This method is simple and easy to use, not prone to severe overfitting, and can find the potential rules in the data. The classification performance is excellent, and the accuracy rate can reach more than 89.9%, which solves the problem of low accuracy in oil-water layer identification in the past. [ABSTRACT FROM AUTHOR]

Copyright of World Nuclear Geoscience is the property of World Nuclear Geoscience Editorial Office 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 total organic carbon (TOC) content is an important parameter for the evaluation of abundance of organic matter in source rocks, and its predicting accuracy is of great significance to oil and gas exploration and development. At present, TOC prediction is mainly based on statistical analysis methods such as ΔlogR method and multiple regression analysis, problems such as weak generalization ability and strong subjectivity exist. The introduction of machine learning methods can effectively solve these problems of instability, nonlinearity, and high complexity. However, current research remains at the level of method comparison and selection with no indepth analysis of good models and their applicability. In this paper, a Support Vector Machine (SVM) model with better application effects was used to predict TOC contents of source rocks with different salinity degrees. As source rocks of freshwater lacustrine facies, the Paleogene Dongying Formation in the Bozhong Sag of Bohai Bay Basin and Paleogene source rocks in the Shizigou area of the western Qaidam Basin as saline lacustrine facies source rocks were selected to test and compare the effectiveness of the model. Through correlation analysis and XGBoost feature importance analysis, the logging sonic differential time (DT), volume density (DEN), spontaneous potential (SP), Gamma ray (GR) and depth were selected as the input layer, while the TOC was used as the output layer to establish a TOC prediction model based on SVM. Results show a strong generalization ability when applied to different sedimentary environments. It can adapt to the geological characteristics of different regions. The sensitivity of logging curves to the abundance of organic matter in source rocks varies in different sedimentary environments, which makes the model more relevant when applying to the fresh water lacustrine facies area in the Bohai Bay Basin.

Abstract Revealing subsurface structures is a fundamental task in geophysical and geological studies. Logging data are usually acquired through drilling projects, which constrain the subsurface structure, and together with the description of drill core samples, are used to distinguish geological units. Clustering is useful for interpreting logging data and making log unit classification and is usually performed by manual inspection of the data. However, the validity of clustering results with such subjective criteria may be questionable. This study proposed the application of a statistical clustering method, the hidden Markov model, to conduct unsupervised clustering of logging data. As logging data are aligned along the drilled hole, they and the geological structure hidden behind such sequential datasets can be regarded as observables and hidden states in the hidden Markov model. When log unit classification is manually conducted, depth dependency of logging data is usually focused. Therefore, we included depth information as observables to explicitly represent depth dependency of logging data. The model was applied to the following geological settings: the accretionary prism at the Nankai Trough, the onshore fault zone at the Kii Peninsula (southwest Japan), and the forearc basin at the Japan Trench. The optimum number of clusters were searched using a quantitative index. The clustering results using the hidden Markov model were consistent with previously reported classifications or lithological descriptions; however, our method allowed a more detailed division of logging data, which is useful to interpret geological structures, such as a fault or a fault zone. Therefore, the use of the hidden Markov model enabled us to clarify assumptions quantitatively and conduct clustering consistently for the entire depth range, even for different geological sites. The proposed method is expected to have wider applicability and extensibility for other types of data, including geochemical and structural geological data. Graphical Abstract

Abstract Formation pressure is the most fundamental data in oil and gas drilling and production; it has an important position in the entire cycle of oil and gas extraction. However, most current prediction methods are limited to parametric methods with fixed models; such that the accuracy does not meet requirements. This is especially true for deeper layers of marine sedimentary basins where the safety density window is extremely narrow. In this study, we propose a novel method to predict pore pressure using machine learning techniques. For the first time, the effective stress (direct output variable) was accurately predicted by a combination of four input variables (2900 sets of data, of which 90% is the training subset and 10% is the testing subset), including longitudinal velocity, porosity, mud content, and density. As such, an accurate prediction of the formation pressure was achieved based on the effective stress theorem. The performance of machine learning techniques was verified by comparing and analyzing the prediction results with traditional parametric single and multivariate models; whereby the best algorithm was chosen by structural optimization and comparative analysis of five algorithms (multilayer perceptron neural network, radial basis neural network, support vector machine, random forest, and gradient boosting machine). Compared with the methods based on parametric one‐dimensional and multivariate models, the machine learning‐based method was determined to possess high accuracy, adequate self‐adaptation, and high fault tolerance (D2 = 0.9981, RMSE = 0.00718 g/cm3). Moreover, the multilayer perceptual neural network algorithm outperformed other machine learning algorithms in terms of goodness of fit, generalization, and prediction accuracy, with D2 = 0.9981 and RMSE = 0.00709 g/cm3. The formation pressure prediction model developed in this study is not affected by the mechanical depositional environment and is applicable to sandy mudstone formations, such that it can be a useful and highly accurate alternative to the traditional formation pressure prediction methods with fixed parameter forms.

Reservoir parameters are crucial indicators for reservoir evaluation and development and provide insights into long-term reservoir behavior. The primary methods for evaluating these parameters include direct core observations, experimental testing, and indirect evaluation techniques. Since its introduction, geophysical logging has been used to evaluate and invert reservoir parameters owing to its wide coverage. With an increasing focus on unconventional natural gas reservoirs, more refined reservoir evaluations and multiparameter analyses are required for their development to address the complex and microscopic models differing from those of the conventional petroleum reservoirs. Geophysical logging is important in several unconventional fields. Machine learning (ML) was used in unconventional gas reservoirs as an effective method to establish relationships between parameters and logging features. However, the accuracy of evaluating storage layers using a single ML method is limited. Studies focusing only on algorithm updates and indicator values are problematic in terms of interpretability and production applications. A need to standardize the use of algorithms and introduce validation comparisons such as geological methods is evident. In this study, we reviewed ML algorithms and models commonly used for logging inversion applications. The current research status and issues were analyzed for different unconventional gas reservoir parameters. Our findings emphasize the importance of combining geological and other methods for logging inversion using ML. We also used the random forest algorithm to accurately predict the reservoir porosity, gas content, coal structure, and macrolithotypes. Combined with established permeability and vitrinite reflectance models, factor analysis was used to comprehensively analyze and evaluate the coalbed methane reservoirs in the study area. In our assessment of the challenges and future work on ML-based inversion, we observed a clear advantage for ML algorithms under geologically validated methods and experimental control. ML has great potential for optimizing the application of logging inversion for unconventional reservoir parameters. [ABSTRACT FROM AUTHOR]

Natural fractures directly affect the permeability and mechanical strength of reservoirs, and their development degree has an important impact on the design and implementation of engineering and development projects. Although there is some correlation between logging data and fracture development, studies using algorithms to optimize logging predictions are still scarce. Meanwhile, there is a scarcity of calculations and analyses concerning the distribution of geostress at the block scale, and the pivotal role that geostress plays as a tectonic factor in the development of fractures. In this study, machine learning methods are used to predict reservoir fracture development, and regional geostress distribution patterns derived from well test data and finite element methods are combined for verification. The support vector machine (SVM), random forest (RF), extreme gradient boosting (XGBoost) and back propagation neural network (BPNN) algorithms are used to predict the fracture development of No.15 coal reservoir in the block. The results showed that the accuracy of SVM is 83.3 %, RF is 91.6 %, XGBoost is 93.7 % and BPNN is 95.8 %. The BPNN can effectively predict the reservoir fracture development of the block. Combined with the regional finite element stress-strain analysis and geostress measurement, the prediction of No.15 coal geostress distribution and fracture development model is established. Under comprehensive verification, the established distribution of the degree of regional fracture development under the control of geostress is consistent with the results of the BPNN prediction of fracture development. These results show that the regional geostress calculated in association with finite element analysis (FEA) can reflect the development of fracture in coalbed methane (CBM) reservoirs, and the neural network has good performance in predicting regional fracture development. This work provides a new approach to the application of machine learning in the field of geological engineering, and the comprehensively validated model provides geologists and geological engineers with ideas in algorithmic practice. • Machine learning-based cross-layer prediction of the natural fracture development in coalbed methane reservoirs is proposed • Establishment of a methodology for evaluating regional tectonic and overall continuity modeling geostress • Interpretation and validation of natural fracture development on control features of burial depth and geostress distribution [ABSTRACT FROM AUTHOR]

Since the accurate prediction of porosity is one of the critical factors for estimating oil and gas reservoirs, a novel porosity prediction method based on Imaged Sequence Samples (ISS) and a Sequence to Sequence (Seq2Seq) model fused by Transcendental Learning (TL) is proposed using well-logging data. Firstly, to investigate the correlation between logging features and porosity, the original logging features are normalized and selected by computing their correlation with porosity to obtain the point samples. Secondly, to better represent the depositional relations with depths, an ISS set is established by slidingly grouping sample points across depth, and the selected logging features are in a row. Therefore, spatial relations among the features are established along the vertical and horizontal directions. Thirdly, since the Seq2Seq model can better extract the spatio-temporal information of the input data than the Bidirectional Gate Recurrent Unit (BGRU), the Seq2Seq model is introduced for the first time to address the logging data and predict porosity. The experimental results show that it can achieve superior prediction results than state-of-the-art. However, the cumulative bias is likely to appear when using the Seq2Seq model. Motivated by teacher forcing, the idea of TL is proposed to be incorporated into the decoding process of Seq2Seq, named the TL-Seq2Seq model. The self-well and inter-well experimental results show that the proposed approach can significantly improve the accuracy of porosity prediction. [ABSTRACT FROM AUTHOR]

Copyright of Geography & Geographic Information Science is the property of Geography & Geo-Information Science Editorial Office 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.)

Copyright of Petroleum Geology & Recovery Efficiency is the property of Petroleum Geology & Recovery Efficiency 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.)

Due to environmental interference and operational errors, problems such as incomplete and random missing logging data have occurred during the geophysical logging data collection process. Since it is difficult to establish a geophysical model based on logging data and geological information, the data complementation effect of conventional methods is not very satisfactory. In this paper, we propose an MC-GAN-BiLSTM model based on spatiotemporal sequence prediction. In the model, we adopt a generative adversarial network (GAN) as a network framework, and a long short-term memory (LSTM) neural network and a bi-directional long short-term memory (Bi-LSTM) as the basic modules. We use the LSTM instead of a fully-connected layer in the GAN to extract the potential information in the logging data depth domain. We complete the logging data missing values through an encoding-decoding structure that includes the Bi-LSTM. In addition, the generator module also uses multiscale convolution to fully extract the logging data features. We use logging data random missing values and consecutive missing values to simulate a field data acquisition environment and threshold control to simulate a laboratory processing environment for experiments. The experimental results show that the coefficient of determination (R2) of the GAN-LSTM model reaches 0.906 when 30% of random logging data are missing and 0.851 when 30% of consecutive logging data are missing. The effect of the model proposed in this paper is significantly higher than the commonly used random forest (RF), sequence to sequence (seq2seq) and generative adversarial interpolation network (GAIN) models.

Prediction of the failure parameters, cohesion (C) and friction angle φ , is essential for drilling operation, field development, and hydraulic fracturing design. Several case studies have been conducted to predict these parameters. In all cases, empirical correlations were recommended based on experimental data where the issue of measurements' uncertainties remained a common factor in all of them. Hence, in this study, the artificial neural network (ANN) is employed to predict these parameters from readily available logging data of two wells. A sensitivity analysis showed a high correlation between bulk density (ROHB), compressional time (DTC), and neutron porosity (NPHI) with the friction angle and cohesion analysis. Hence, ROHB, DTC, and NPHI were selected as inputs for the ANN model. Comparing the measured cored data of cohesion and friction angle with those predicted from the ANN model, the results indicated a model error less than 3% average absolute percentage error (AAPE) and a correlation coefficient greater than 0.95. Moreover, an empirical correlation was developed and validated with another set of data, with AAPE less than 3% and R greater than 0.95. Comparison with the literature shows the better accuracy of the ANN in predicting cohesion and friction angle. The model is several times faster than the existing methods in-field application with high reliability and accuracy. [ABSTRACT FROM AUTHOR]

Assessment of organic carbon content (TOC) by geophysical logs has been a challenging task in the formation evaluation of shale gas. This research is conducted to estimate the unconventional hydrocarbon (shale-gas) potential of Talhar Shale in the Southern Indus Basin (SIB), Pakistan. In this study, total organic carbon content (%) was estimated through well logs by different methods and then correlated with well cuttings/core data to determine the best method for estimation of TOC content especially when well cuttings/core data are not available. The Talhar Shale’s thermal maturity, as well as the organic content, were assessed from geochemical analyses. Talhar Shale of Dangi-01 well has good to very good source potential whereas Chak7A-01 well has fair-good. According to Ven Krevalen cross-plot, Talhar Shale of Dangi-01 well has type III kerogen; it can only produce gas while Chak7A-01 has type II/III kerogen which produces both oil and gas. The TOC was estimated using two methods i.e., Schmoker’s and Hester’s and Multivariate Fitting methods. The estimated TOC is then correlated with well cuttings data and concluded that the Multivariate Fitting method is selected as an optimized method for estimation of TOC because it shows strong correlation values of 0.93 and 0.91 in both wells respectively for Talhar Shale SIB, Pakistan.

Revealing subsurface structures is a fundamental task in geophysical and geological studies. Logging data are usually acquired through drilling projects, which constrain the subsurface structure, and together with the description of drill core samples, are used to distinguish geological units. Clustering is useful for interpreting logging data and making log unit classification and is usually performed by manual inspection of the data. However, the validity of clustering results with such subjective criteria may be questionable. This study proposed the application of a statistical clustering method, the hidden Markov model, to conduct unsupervised clustering of logging data. As logging data are aligned along the drilled hole, they and the geological structure hidden behind such sequential datasets can be regarded as observables and hidden states in the hidden Markov model. When log unit classification is manually conducted, depth dependency of logging data is usually focused. Therefore, we included depth information as observables to explicitly represent depth dependency of logging data. The model was applied to the following geological settings: the accretionary prism at the Nankai Trough, the onshore fault zone at the Kii Peninsula (southwest Japan), and the forearc basin at the Japan Trench. The optimum number of clusters were searched using a quantitative index. The clustering results using the hidden Markov model were consistent with previously reported classifications or lithological descriptions; however, our method allowed a more detailed division of logging data, which is useful to interpret geological structures, such as a fault or a fault zone. Therefore, the use of the hidden Markov model enabled us to clarify assumptions quantitatively and conduct clustering consistently for the entire depth range, even for different geological sites. The proposed method is expected to have wider applicability and extensibility for other types of data, including geochemical and structural geological data. [ABSTRACT FROM AUTHOR]

Traditional lithology identification left the problems of low accuracy, recognition efficiency and generalization ability. Facing the logging data with outliers, unbalance and high complexity, we propose a lithology identification method based on an improved neighborhood rough set and AdaBoost. On the basis of the classical neighborhood rough set, the selection of the neighborhood radius and the running time are optimized. The redundant information in logging data is then effectively eliminated. Thus more sensitive logging curves are selected without changing the physical meaning of logging attributes. Then the selected data are input into the AdaBoost model to construct a lithology identification model. About 54,000 samples from 5 boreholes are tested in the study area. The accuracy of classification on the test set is about 98.42%. Compared with BP neural network and random forest algorithm, the proposed method owns advantages in recognition accuracy and generalization ability. It can provide help for complex lithology recognition in the study area. [ABSTRACT FROM AUTHOR]

Formation pressure is the most fundamental data in oil and gas drilling and production; it has an important position in the entire cycle of oil and gas extraction. However, most current prediction methods are limited to parametric methods with fixed models; such that the accuracy does not meet requirements. This is especially true for deeper layers of marine sedimentary basins where the safety density window is extremely narrow. In this study, we propose a novel method to predict pore pressure using machine learning techniques. For the first time, the effective stress (direct output variable) was accurately predicted by a combination of four input variables (2900 sets of data, of which 90% is the training subset and 10% is the testing subset), including longitudinal velocity, porosity, mud content, and density. As such, an accurate prediction of the formation pressure was achieved based on the effective stress theorem. The performance of machine learning techniques was verified by comparing and analyzing the prediction results with traditional parametric single and multivariate models; whereby the best algorithm was chosen by structural optimization and comparative analysis of five algorithms (multilayer perceptron neural network, radial basis neural network, support vector machine, random forest, and gradient boosting machine). Compared with the methods based on parametric one‐dimensional and multivariate models, the machine learning‐based method was determined to possess high accuracy, adequate self‐adaptation, and high fault tolerance (D2 = 0.9981, RMSE = 0.00718 g/cm3). Moreover, the multilayer perceptual neural network algorithm outperformed other machine learning algorithms in terms of goodness of fit, generalization, and prediction accuracy, with D2 = 0.9981 and RMSE = 0.00709 g/cm3. The formation pressure prediction model developed in this study is not affected by the mechanical depositional environment and is applicable to sandy mudstone formations, such that it can be a useful and highly accurate alternative to the traditional formation pressure prediction methods with fixed parameter forms. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Engineering (17264073) 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.)

The availability of two types of solar panels that are common in the market namely monocrystalline and polycrystalline types cause confusion in the selection so that many solar panel users are questioning the differences of these two types of solar panel. This study produced a data logger system using Arduino Uno R3 to control voltage, current and temperature sensors for logging data that stores power measurement data from monocrystalline and polycrystalline solar panel in a micro SD. After it we can manage data to compare power produced between two types the solar panel. From the results of testing this data logger system it can be seen that monocrystalline solar panel are 9.18% better on power produced than polycrystalline when the maximum power conversion is generated.

Permeability is arguably the most important property in evaluating fluid flow in the reservoir. It is also one of the most difficult parameters to measure in field. One of the main techniques for determining permeability is the application of Nuclear Magnetic Resonance (NMR) logging across the borehole. However, available correlations in literature for estimating permeability from NMR data do not usually give acceptable accuracy in carbonate rocks. In this research, two new empirical models are introduced for quantifying NMR extracted permeability in carbonate formations. These models are validated for three carbonate formations, namely, Yamama, Gadvan, and Daryan in one of Iranian offshore reservoirs in the Persian Gulf. The first empirical model applies the pore-related NMR data such as free and bound fluid parameters. The second model, however, is a novel approach that uses the geometric features of the occurring humps in T2 distribution. For assessing the performance of the proposed models, statistical parameters as well as graphical tools are utilized. It is found that the for the examined case studies, geometric approach gives more accurate and reliable estimates compared to the available models in the literature including Timur-Coates and SDR methods.

Rock mechanical parameters are essential for drilling operation optimization, reservoir health monitoring, and hydrocarbon production. Consequently, achieving high economic revenue from the reservoir is dependent upon an accurate prediction of these parameters. Cohesion and internal friction angle are two important parameters that represent the shear strength of the rock. Several works have been done to predict these two parameters empirically from lab measurement or wireline logging data. In this study, two machine learning (ML) techniques, decision tree (DT) and random forest (RF), were used to build a predictive model using actual well logging data of two wells. The input data include compressional time (DTC), bulk density (ROHB), and neutron porosity (NPHI), as the sensitivity analysis showed that friction angle and cohesion are more sensitive to this input than the other logging parameters. The first well data set was used to train and test the model, while another set of 635 data points from different well was used for the model validation. The model accuracy was measured using the correlation coefficient (R), and the absolute average percentage error (AAPE), in addition to the failure parameter profiles. The result indicated the high accuracy of the two models in the prediction as the DT model recorded an R-value greater than 0.98 and 0.96 and an AAPE less than 2% and 3% for friction angle and cohesion, respectively, while the RF model showed its outperformance with R values 0.99 and 0.97 and AAPE values less than 1% and 2.5% for the two parameters respectively. The validation set confirmed the accuracy of the two models in generating a complete accurate profile of the friction angle and cohesion, as the DT model's R values were calculated at 0.96 and 0.95 and AAPE of 1.6% and 3.3% for friction angle and cohesion, respectively. On the other hand, the RF model showed its better accuracy with R values of 0.97 and 0.96, while the AAPE values were 1.2% and 2.5% for friction angle and cohesion respectively. The obtained results confirmed the high accuracy and the robustness of these techniques in predicting the friction angle and cohesion; in addition, these methods are several times faster for field applications. [ABSTRACT FROM AUTHOR]

Abstract The shallow accretionary prism of the Nankai Trough is a location where both large interplate earthquakes and slow earthquakes occur. Since the physical properties of sedimentary materials are important topics for understanding the structure of the prism, numerous ocean drilling expeditions have been conducted in that region to obtain logging data and core samples. Although the physical properties of the obtained samples are normally measured onboard immediately after coring, estimations of in situ physical properties are difficult because of differences in laboratory and in situ physical conditions. Herein, we propose a new method for estimating in situ porosity from downhole electrical resistivity log data that evaluates in situ porosity and thermal structure simultaneously using correlations between the porosity and resistivity, and between the porosity and thermal conductivity that were established based on laboratory measurements. When constructing physical property correlations, X-ray computed tomography data play an important role in estimating the porosity of samples from which resistivity or thermal conductivity were measured. To validate our method, we compared the estimation with density log data collected at Site C0002 and found that the estimated in situ porosity shows good agreement with the in situ porosity converted from density log data. A comparison with porosity measured onboard for core and cutting samples showed that they are consistent with each other. With this new method, continuous distributions of in situ porosity and thermal structure can be estimated simultaneously based on resistivity log data and heat flow, which are basic quantities acquired during ocean drilling science expeditions.

Abstract Sand production is a major problem that the oil and gas industry has been facing for years. It can lead to loss of production, equipment damage or complete well abandonment. Prediction of sand has been historically challenging due to the periodic nature of sand production, insufficient laboratory tests and lack of field tests validation. Analyses have been performed to identify weak zones for planned wells, and common technique is the application of shear modulus and mechanical properties log (MPL) criteria developed by Tixier et al. (J Pet Technol 27:283–293, 1975). However, the set criteria have been found to be generally inadequate to detect transition zone or predict weak formation in some fields. In this study, using the knowledge of rock behavior, geomechanical properties and well log data, we have established new simple criteria for identifying fragile sections within a transition zone. In situ logging data from a field X, located in Sabah, Malaysia, and Field Y, located in Shimokita, Japan, were used in this study. Using the threshold for shear modulus and MPL, the criteria for the geomechanical properties are set to differentiate formation strengths at different depths. The threshold for Poisson’s ratio is 0.34, Young’s modulus at 1.6 × 106 psi and the unconfined compressive strength at 2400 psi. The MPL and geomechanical models were generated to predict sanding incident. The results were subsequently validated with artificial neural network using MATLAB. Also, critical wellbore pressure is calculated and acts as a guide to operate outside the sand failure envelope. Thus, the prediction of the weak formation using geomechanical properties has been further established in this study.

[Display omitted] • A new blue noise sampling is proposed to preserve the spatial distribution of standard wells. • A stratigraphic correlation model is designed to estimate the similarity between wells. • A discrete choice model is used to select standard wells with stochastic utility maximization. • A visualizaiton framework is implemented enabling users to visually select and evaluate standard wells. It is a time-consuming and laborious task to conduct accurate geological interpretation based on large-scale logging data. A subset of drilling wells are always selected undergoing detailed stratigraphic correlation, aiming to enhance the efficiency of geological interpretation and retain the matching accuracy of stratigraphic structures. However, standard wells are selected by hand in the course of traditional standard well selection, which will easily bring many uncertainties for subsequent interpretation of geological structures. In this paper, we propose a visual analytics system to support supervised standard well selection via a discrete choice model. Firstly, an adaptive blue noise sampling model is applied to determine spatial distribution of standard wells, and a stratigraphic correlation model is designed to measure the similarity between drilling wells from different attribute perspectives. Then, we utilize a discrete choice model to select standard wells with the spatial distribution and multiple attributes taken into consideration. Furthermore, several meaningful visualizations are designed allowing users to get deeper insights into the selection of standard wells, and a rich set of interactions are also provided enabling users to further optimize the discrete choice model. At last, a visualization framework is implemented to integrate the models and visual designs, by means of which the geological interpreters are able to visually select, evaluate and optimize the standard wells. The effectiveness of our work and its application values for geological interpretation are further demonstrated through case studies based on real-world datasets and interviews with domain experts. [ABSTRACT FROM AUTHOR]

Hydraulic fracturing of tight sandstone reservoirs is an important method of developing this type of unconventional oil and gas resources. The evaluation of in-situ stress distribution and rock mechanical parameters of fractured wells is the prerequisite for fracturing design. Therefore, the development of a calculation method to evaluate the stress distribution and mechanical parameters of rocks, based on a single well in-situ stress data, is the key to establish the optimal fracturing design. In this paper, based on the correlation between the laboratory experimental data and the logging response, we have plotted the initial stress profile and then applied the proposed method of combining density curve and natural gamma curve to correct the calculation results. In the case of a limited number of wells, the obtained profile correction model demonstrated satisfactory matching with laboratory experimental data. [ABSTRACT FROM AUTHOR]

Water-bearing structures in front of coal mine shaft drivages require effective advanced prediction. Based on the characteristics of coal field logging curve carrying rich geological information, it is introduced into the advanced prediction of water-bearing bodies in mine shafts. At the same time, mine geophysical prospecting technology, such as the DC resistivity method, can accurately and effectively track and detect geological conditions during mine shaft drivage. The proposed advanced prediction method of the borehole-ground electrical method based on the detailed interpretation of logging curves was experimentally studied using physical simulations of the detailed interpretation of water-bearing rock strata in logging data and geophysical advanced detection simulations of water-bearing structures in mine shafts. By constructing a physical model of the mine shaft drivage, the change in the water content of the rock strata is simulated by adding water to sand and gravel, and the apparent resistivity logging in a borehole is simulated. The water-bearing characteristics of the model under different water content conditions are interpreted by using 3-D visualization technology, and it is concluded that the change law of the electrical characteristics of sand and gravel after adding water is a logarithmic function relationship. Then, based on the results of the detailed interpretation of the logging curves, a model experiment using the DC resistivity method of ground-borehole joint advanced detection is carried out, and the response characteristics of changes in the model water diffusion resistivity are accurately detected. In addition, a shallow geological borehole is constructed in the field. Water injection in the borehole causes anomalous changes in the electrical properties of the strata around the borehole and simulates water-bearing structures at different depths and layers, which verifies the effectiveness of the borehole-ground electrical method in detecting water-bearing structures in front of mine shaft drivages. The simulation results show that the advanced prediction method of borehole-ground electromagnetic technology based on the interpretation of logging curves enhances the effectiveness and detection efficiency of advanced detection and prediction of the water-bearing characteristics of rock strata. Based on the logging curve, the water-bearing characteristics of the strata in front of a coal mine shaft drivage in the Dongsheng coalfield in Inner Mongolia are interpreted in detail, and the provided anomalous target areas can provide reliable guidance for advanced detection using the borehole-ground electrical method in the future. [ABSTRACT FROM AUTHOR]

High-quality hydrocarbon source rocks are present in the upper Cretaceous layer in the western slope of the Songliao Basin. Oil and gas have accumulated in these rocks at the shallow edge of the basin, which has led to the formation of oil sand resources. This study uses the back-propagation (BP) neural network method to predict the distribution of oil sand reservoirs and is the first study of its kind in China. First, based on the basic data collected by core sample, well log and geochemical analyses, and the reasonable selection of samples, the cores are divided into mudstone, siltstone, fine sandstone, medium sandstone, and sand, according to lithology. Second, a three-layer BP neural network model is constructed with two hidden S-type layers and one linear output layer. Third, through a comparison of the effect of different numbers of training sessions of the sample data, we demonstrate that the accuracy of the model can be increased to 90% after training the network 100,000 times. Then, the log-derived data of rocks with unknown lithologies are input into the neural network to predict whether they contain oil sands. We show that the BP neural network method can predict the distribution of oil sand reservoirs in the target horizon of the study area, and the results are consistent with research results on the corresponding sand reservoirs and sedimentary facies. Thus, we conclude that it is feasible to use the BP neural network method to predict the distribution of oil sand reservoirs. [ABSTRACT FROM AUTHOR]

Copyright of Oil Drilling & Production Technology / Shiyou Zuancai Gongyi is the property of Shiyou Zuancai Gongyi Bianjibu 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 sedimentary characteristics of the southwestern part of the Qaidam Basin are complex. So far, the sedimentary facies types and sedimentary models in the region are still controversial. In this paper, we use the rich core data, logging data, mineral test data and regional sedimentary tectonic background in the region to study the sedimentary model of the area. The results show that the lower segment of the Xiaganchaigou Formation (fo1 ) is mainly the western source, followed by the northwest and southwest sources; the Shangganchaigou Formation (N1) is mainly the western source, followed by the northwest and southwest sources; The Xiayoushashan Formation (N21 ) is mainly the western source, followed by the northwest and southwest sources. The Xiayoushashan Formation (N2 1 ) is formed in the underwater environment under weak reduction-weak oxidation conditions; the Shangganchaigou Formation (N1) is formed in the weak hydrodynamic environment under weak reduction conditions; and the Lower Ganchaigou Formation (fo1 ) is formed in a strong hydrodynamic environment under oxidizing conditions. Finally, the sedimentary facies model of YJ No. 2 Oilfield was established. The YJ No. 2 oilfield developed a braided river and lake sedimentary system. The E3 1 reservoir is mainly a sedimentary system of the delta plain distributary channel, the E3 2 develops a lacustrine sedimentary system, the Ni reservoir develops a braided river sedimentary system, and the N2 1 reservoir mainly develops a braided river sedimentary system. [ABSTRACT FROM AUTHOR]