Log Response Research Articles

NMR log response prediction from conventional petrophysical logs with XGBoost-PSO framework

A combined approach that exploits both the eXtreme Gradient Boosting (XGBoost) method and the Particle Swarm Optimization (PSO) method was used here to predict the nuclear magnetic resonance (NMR) log parameters such as porosity and permeability from field measurements. Data obtained by NMR are extremely critical for modeling the reservoir behavior and are among the most vital inputs for reservoir simulations. Therefore, developing an intelligent framework that can accurately predict NMR logging response by using conventional, low-cost logging data can be a valuable tool for field development. A number of conventional well logs, such as neutron, density, sonic, caliper, gamma-ray, and resistivity logs, recorded from sixteen wells in an offshore oilfield in Persian Gulf were used in the XGBoost-PSO framework to predict NMR log response. The outputs are the free fluid porosity (CMFF), bound fluid porosity (BFV), permeability (KTIM), and total porosity (TCMR). Based on the results, in terms of prediction accuracy, comparing correlation coefficients, the XGBoost-PSO model performs in the range of 88.5–91.4%. Moreover, the application of PSO in hyperparameter optimization of predicted parameters confirmed at least five percent improvement. Overall, we were able to demonstrate that ordinary conventional logging data can be utilized to obtain more advanced NMR data without performing expensive, time-consuming, and high-risk logging operations to decrease operational costs without compromising the benefits.

Transcriptional Responses as Biomarkers of General Toxicity: A Systematic Review and Meta-analysis on Metal-Exposed Bivalves.

Through a systematic review and a series of meta-analyses, we evaluated the general responsiveness of putative transcriptional biomarkers of general toxicity and chemical stress. We targeted metal exposures performed on bivalves under controlled laboratory conditions and selected six transcripts associated with general toxicity for evaluation: catalase, glutathione-S-transferase, heat shock proteins 70 and 90, metallothionein, and superoxide dismutase. Transcriptional responses (n = 396) were extracted from published scientific articles (k = 22) and converted to log response ratios (lnRRs). By estimating toxic units, we normalized different metal exposures to a common scale, as a proxy of concentration. Using Bayesian hierarchical random effect models, we then tested the effects of metal exposure on lnRR, both for metal exposure in general and in meta-regressions using toxic unit and exposure time as independent variables. Corresponding analyses were also repeated with transcript and tissue as additional moderators. Observed patterns were similar for general and for transcript- and tissue-specific responses. The expected overall response to arbitrary metal exposure was an lnRR of 0.50, corresponding to a 65% increase relative to a nonexposed control. However, when accounting for publication bias, the estimated "true" response showed no such effect. Furthermore, expected response magnitude increased slightly with exposure time, but there was little support for general monotonic concentration dependence with regard to toxic unit. Altogether, the present study reveals potential limitations that need consideration prior to applying the selected transcripts as biomarkers in environmental risk assessment. Environ Toxicol Chem 2023;42:628-641. © 2022 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

Lithology identification using graph neural network in continental shale oil reservoirs: A case study in Mahu Sag, Junggar Basin, Western China

The continental shale oil reservoir of Fengcheng Formation in the northern slope area of Mahu Sag, Junggar Basin, Western China is very heterogeneous in lithology. Thus, the complex response characteristics of conventional logging and limited core availability in the study area has led to major challenges in lithology identification. Therefore, to resolve lithology identification by well logs in continental shale oil reservoirs, a graph neural network (GNN) method named GraphSAGE is used to train the lithology identification model based on a constructed graph, which connects the samples with adjacent depth and similar log response features on operator intention. The identification process is divided into two parts: first, based on the formation depth sequence and affinity propagation clustering method, the vertical distribution of the stratum and nodes logging curve similarity information are integrated into the graph structure, which structurally represents the conventional logging curves as graph instead of well logs as input data; Second, the nodes of the constructed graph are classified by GraphSAGE, which naturally supports combination generalization and improves sample complexity accompanied by strong relational inductive bias. To examine the effectiveness of GraphSAGE for lithology identification, a conventional log dataset labelled by direct core observations from two separate wells in Muhu Sag are used. The identification results showed that the accuracy of GraphSAGE for the lithologies exceeds 90% of the testing data, especially for transitional lithology such as dolomitic mudstone, silty mudstone and tuffaceous fine sandstone. Compared with the commonly used machine learning methods such as SVM, RF and XGBoost, GraphSAGE was more accurate in lithology identification, matching core observations. Collectively, this reflects the superiority of graph neural network in conventional logging lithology identification and effective means provided for lithology identification of continental shale oil reservoir in the Mahu Sag.

Hydrocarbon Prospectivity of Surma Group in Sylhet Area in Context of Sequence Stratigraphy

This research focuses on studying the Surma Group sediments with an integrated approach of using wireline logs and 2D seismic data from Sylhet structure and outcrop analogy from the Zaflong-Tamabil-Jaintiapur area to explain the geological evolution in the context of sequence stratigraphic framework and its implications to identify the hidden potential of hydrocarbon traps that remains overlooked. The 'SYL-07' well logs divided the encountered sedimentary strata by sequence boundaries (SB) and flooding surfaces (FS) into an upper third-order Lowstand System Tract (3_LST), followed by the third-order Highstand System Tract (3_HST), and underlying third-order Transgressive System Tract (3_TST) which correspond to the Tipam, Boka Bil, and Bhuban formations, respectively. The 3_HST has further been subdivided into fourth-order sequences. Well-to-seismic ties helped in identifying the sequence boundaries in the regional context. Three new prospects in the nose and flank area of the Sylhet structure have been detected based on seismic attributes and log responses. Seismic sweetness attribute anomalies confirmed the existence of these prospects. Rigorous fieldworks provided nineteen facies and seven sets of sedimentary facies associations and helped to establish the sedimentological stacking patterns and corresponding depositional environments. About 2250m lithocolumn constructed from field data revealed four third-order sequence stratigraphic system tracts roughly corresponding to Renji (3_LST), Bhuban (3_TST), Boka Bil (3_HST), and Tipam (3_LST) formations in the bottom-up order, while the obtained facies associations aid in identifying four fourth-order sequences including fourth-order system tracts in the 3_HST. The lithologic column corresponds well with the seismic sections and validates the seismically interpreted third-order system tracts and fourth-order sequences. However, the high-frequency fourth-order system tracts can only be correlated with the log motif patterns due to sub-seismic resolution. The erosional remains of the fourth-order sequences in 3_HST, for instance, the coarsening upward shoreface and shelfal facies topped by transgressive shale, have yielded potential future drilling sites based on seismic attribute and log response analysis. The analogy of field data with well log and seismic data helped the overall evaluation of the hydrocarbon prospectivity of the area. The Dhaka University Journal of Earth and Environmental Sciences, Vol. 11(1), 2022: 95-110

Vertically transferred overpressures along faults in Mesozoic reservoirs in the central Junggar Basin, northwestern China: Implications for hydrocarbon accumulation and preservation

Abnormally high pressure (overpressure) is common in oil reservoirs of the Mesozoic formations in the central Junggar Basin. However, the generation mechanisms of overpressure are still uncertain, which not only affects safe drilling implementation, but also restricts the understanding of hydrocarbon migration and accumulation. Based on measured pressures and mud weights, the pore pressure distributions in the Mesozoic formations are characterized in detail. The different generation mechanisms of overpressure in mudstones and sandstones are investigated based on logging responses and basin modeling. Finally, the significance of reservoir overpressure to oil accumulation and leakage is discussed. The results indicate that overpressure in the Mesozoic reservoirs in the central Junggar Basin is mainly developed below 4500 m, the excess pressures are between 12 and 58 MPa, the pressure gradient ranges from 12 to 21 MPa/km, the pressure ratios range from 0.6 to 0.95, and the overpressure ratios are between 0.25 and 0.91. The high-overpressure in reservoirs is mostly present in isolated sand bodies that are hydraulically connected to Permian source rocks by faults. The overpressure data corresponding to a pressure gradient of 11–14 MPa/km plot on the loading curve, showing typical disequilibrium compaction overpressure. The points with pressure gradients greater than 14 MPa/km obviously deviate from the loading curve, indicating unloading overpressure mechanisms. The logging responses and numerical simulation results confirm that the overpressure in mudstone is mainly generated by disequilibrium compaction and chemical compaction, and hydrocarbon generation has a limited contribution. The unloading overpressure in the sandstone is mainly generated by the vertical overpressure transfer process from the deep Permian overpressured compartment. The vertical transfer of overpressure is often related to oil migration and accompanied by high oil saturations. However, high-overpressure presents challenges for the sealing capacity of caprocks. Hydrofracturing or the activation of preexisting faults due to high overpressure determines the risk of seal failure, which controls hydrocarbon leakage. The research is of great significance for understanding the mechanisms of hydrocarbon migration along faults and evaluating caprock integrity in similar basins.

Plastic responses of below‐ground foraging traits to soil phosphorus‐rich patches across 17 coexisting AM tree species in a subtropical forest

Abstract Below‐ground plastic responses to soil nutrient ‘hot‐spots’ form a key nutrient foraging strategy of plants coexisting in natural ecosystems. However, it is unclear how plant species differ in these below‐ground plastic responses and how they covary. Plastic responses to soil phosphorus (P)‐rich patches of absorptive root, mycorrhizal, and exudation traits of 17 coexisting arbuscular mycorrhizal (AM) tree species in a subtropical evergreen broad‐leaved forest were investigated using a root bag method. There was considerable variation and heterogeneity in species‐specific responses to P‐rich patches. Negative log response ratios usually occurred for high‐cost traits and positive log response ratios for low‐cost traits. There were trade‐offs in the plastic responses between root acid phosphatase activity and extraradical hyphal length, which were unaffected by phylogeny, and between root acid phosphatase activity and specific root length. Thicker‐rooted species responded to P‐rich patches more through root exudation plasticity than mycorrhizal plasticity. Thinner‐rooted species relied more on mycorrhizal plasticity. Synthesis. Our results revealed diverse P foraging strategies comprising different combinations of plastic adjustments in absorptive root, mycorrhizal, and exudation traits among coexisting AM tree species, which suggest the potential for complementary exploitation of different soil P sources.

Functional characteristics rather than co‐occurrences determine the outcome of interactions between neighbouring plants in sub‐Antarctic ponds: Consequences for macrophyte community biomass

Abstract Understanding of the mechanisms underlying species coexistence within plant communities is crucial to predict their fate given the current context of biodiversity loss. Freshwater ecosystems are among the most abiotically constrained habitats because they harbour clonal macrophyte communities. Strong habitat filtering in these ecosystems influences the functional composition and diversity of macrophyte communities, determining the outcome of interactions between neighbours and ultimately affecting local spatial arrangement between neighbours (LSA). This influence may be modulated by environmental conditions in abiotically constrained habitats, such as freshwater ecosystems. We focused on macrophyte communities living in ponds in the Iles Kerguelen, in the sub‐Antarctic. These freshwater ecosystems are especially abiotically constrained (cold climate), and their plant communities are remarkably species‐poor, simplifying the study of interactions between neighbours. We measured several abiotic variables in the ponds, species LSA and interspecific interactions (using the log response ratio metric), and the functional composition of the community using aerial, root and clonal traits. We also determined the biomass of the whole macrophyte community. Our results showed that LSA does not effectively assess interactions between neighbours at very small scales, neither at the community (one species vs. all neighbouring species) nor the species level (between pairs of species). Secondly, aerial (leaf and stem) and root traits related to resource acquisition played a more important role in interactions between neighbours than clonal traits (i.e. internode length and specific internode mass related to space acquisition and resource storage, respectively). Depending on the target species, (1) interactions responded positively or negatively to mean trait and functional diversity of the community; and (2) different traits of neighbours (aerial or root traits) triggered an interaction response. Lastly, abiotic variables, in particular water temperature and light intensity, influenced macrophyte community biomass and plant community structure (i.e. species richness, functional diversity and LSA), either directly or through the modulation of other abiotic variables. Our results lead to the following conclusions: (1) interactions between neighbours do not reflect their local spatial arrangements but are strongly associated with abiotic variables and neighbourhood functional traits (depending on species); (2) joint integration of biotic and abiotic variables in multivariate analyses enables better inference of biodiversity responses to environment, and the subsequent consequences for ecosystem function, especially in the context of climate change.

Identification of Hydrocarbon Zones and Chimneys Using Seismic Attributes Analysis at Woda Oil Field, Western Offshore Niger Delta

This research investigated hydrocarbon zones and chimneys using seismic attributes analysis in Woda Oil Field, Western offshore Niger Delta. Seismic attributes are powerful tools in monitoring and predicting hydrocarbon zones. Detailed application of seismic attributes on the Woda seismic data aided in the extraction of important information about the geologic features on seismic profiles. Materials used in this research include well logs, seismic data and 3D Post Stack Depth Migration (PSDM) seismic volume. Methods used were seismic attribute analysis and well log interpretation. The reservoirs of interest were identified and analyzed using well log data based on gamma ray, resistivity, and Vp sonic log response. Seismic attributes such as chaos, root mean square velocity, similarity and reflection intensity were quality controlled, enhanced, and generated which assisted in identifying chimneys and hydrocarbon zones. The majority of the chimneys in this research may have nucleated in the lower part of the faulting system below the partially permeable portion of the fault planes and across a regional partially permeable barrier as revealed by thin fault likelihood attribute. Gas expulsion along the lower portion of the complex faulting system perhaps, may have preconfigured the spatial organization of the chimneys. Pockmarks formation above the chimneys as revealed by the sweetness attribute may have inherited the same linear planiform geometry, and both structures recorded the timing of hydrocarbon zones and gas expulsion within the Woda Oil Field.

Evaluation of Logging Characteristics of the Yablunivske Oil and Gas Condensate Filed

A study of the reservoir characteristics of the Tournaisian formation in the Yablunivske oil and gas condensate filed shows that the pore structure of the reservoir in this area is complex and the pore permeability varies widely. Based on the reservoir logging response characteristics, core analysis data and logging curves GR, DT and normal resistivity curves, the reservoir mud content, porosity, permeability, bound water saturation and gas saturation parameters were calculated by statistical regression methods. The calculation accuracy of the established log interpretation models is high, which can meet the development needs of the oilfield and also lay the foundation for the subsequent study of the reservoir.

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

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

Investigation of volcanic rock quantitative fracture evaluation and reservoir classification based on the logging method - an example from Nanpu Sag

Multi-stage development results in complex lithology and variable pore structure in volcanic reservoirs. Therefore, it is difficult to calculate fracture porosity and determine the volcanic reservoir type. In this study, we use the deep volcanic rock of the Nanpu No. 5 structure as an example to use a matrix porosity logging evaluation model in different lithology reservoirs. Dual Laterolog has a large detection range and is sensitive to fractures. Therefore, the relationship between igneous fracture and logging response characteristics is analysed via numerical simulation. Then, the method used for calculating fracture parameters was verified by combining formation MicroScanner image (FMI) logging technology with Dual Laterolog. For accurate reservoir classification, a fracture porosity less than 0.0005 is ineffective since the formation fluid cannot be transported. The quantitative fracture evaluation of volcanic rock reservoir was completed based on calculation model. The agreement with the actual production proved the effectiveness of the method. [Received: September 9, 2022; Accepted: December 30, 2022]

Lithofacies identification of shale reservoirs using a tree augmented Bayesian network: A case study of the lower Silurian Longmaxi formation in the changning block, South Sichuan basin, China

As important units for evaluating shale gas reservoirs, lithofacies exert obvious control on the abundance of shale organic matter and the enrichment degree of gas reservoirs. Therefore, it is of great significance to accurately identify the lithofacies of shale reservoirs relative to shale gas exploration and development. This paper takes the Longmaxi Formation shale in the Changning block, Sichuan Basin, China, as an example and uses a tree augmented Bayesian network for the quantitative classification and identification of lithofacies for the first time. First, based on thin section and scanning electron microscopy observations and X-ray diffraction, we can determine the shale lithofacies types of the Longmaxi Formation. Second, the logging response characteristics of different lithofacies were obtained by comparing conventional logging data. Finally, the tree augmented Bayesian network (TAN) was built according to the selected logging parameters and identified unknown lithofacies. The results are as follows: (1) Longmaxi Formation shale is rich in siliceous minerals. By comparing analysis of shale mineral composition and diagenetic characteristics, we divide the lithofacies into five types: siliceous rock, siliceous shale, argillaceous/siliceous mixed shale, calcareous/siliceous mixed shale and calcareous shale. (2) Based on the logging data, 6 logging curves (GR, AC, DEN, CNL, PE and RT) can sensitively indicate diagenesis and changes in mineral content. (3) The TAN model trained by logging information was used to identify 68 shale samples through the 5-fold cross validation method, with an accuracy rate of 95.58%. The results showed that the tree augmented Bayesian network not only can overcome the limitation of interdependence among well-logging properties but also represents an efficient and accurate new method for identifying shale lithofacies. This algorithm can be applied to lithofacies identification of shale reservoirs with only conventional well-logging data, which can help promote the development of lithofacies identification of unconventional oil and gas geology and provides a new method for shale gas reservoir evaluation.

A robust and readily implementable method for the meta-analysis of response ratios with and without missing standard deviations.

The log response ratio, lnRR, is the most frequently used effect size statistic for meta-analysis in ecology. However, often missing standard deviations (SDs) prevent estimation of the sampling variance of lnRR. We propose new methods to deal with missing SDs via a weighted average coefficient of variation (CV) estimated from studies in the dataset that do report SDs. Across a suite of simulated conditions, we find that using the average CV to estimate sampling variances for all observations, regardless of missingness, performs with minimal bias. Surprisingly, even with missing SDs, this simple method outperforms the conventional approach (basing each effect size on its individual study-specific CV) with complete data. This is because the conventional method ultimately yields less precise estimates of the sampling variances than using the pooled CV from multiple studies. Our approach is broadly applicable and can be implemented in all meta-analyses of lnRR, regardless of 'missingness'.

Reservoir quality and facies modeling of the early Eocene carbonate stratigraphic unit of the Middle Indus Basin, Pakistan

The early Eocene carbonate reservoir, Sui Main Limestone (SML), is the largest gas reservoir in Pakistan. In the Qadirpur area, more than 30 wells have been drilled, some of which have been declared dry or abandoned due to poor reservoir characteristics or facies variation. The aim of this work is to re-evaluate the reservoir characteristics and facies identification of Sui Main Limestone by using petrophysical analysis and rock physics modeling in the Qadirpur field in the Middle Indus Basin of Pakistan. The reservoir characterization of carbonate rocks is difficult because of their complex pore networking. The well data on five exploratory wells drilled in this gas field are used. The log-derived porosity varies in a very large range of 2%–36%, with an average of 14%–34%. The average porosity of the clean carbonate intervals is 10%–14% and thus has the capacity to accumulate hydrocarbon. The high porosity value indicates the presence of micro-fractures in Sui Main Limestone. These micro-fractures and secondary pores are well interconnected and allow the pore fluids to communicate. The secondary porosity is mainly due to the presence of vuggs and fractures in Sui Main Limestone. The volume of shale varies from 11%–38% in the wells. The net pay zones have hydrocarbon saturation in the range of 40%–70%, which is mainly gas. Using wireline log response, the formation is divided into three facies: limestone, shale, and shaly limestone. This comprehensive work may help improve the prediction of the reservoir quality in heterogeneous carbonate reservoirs and optimize field development.

Physical Simulation of Ultrasonic Imaging Logging Response

Ultrasonic imaging logging can visually identify the location, shape, dip angle and orientation of fractures and holes. The method has not been effectively applied in the field; one of the prime reasons is that the results of physical simulation experiments are insufficient. The physical simulation of fracture and hole response in the laboratory can provide a reference for the identification and evaluation of the underground geological structure. In this work, ultrasonic scanning experiments are conducted on a grooved sandstone plate and a simulated borehole and the influence of different fractures and holes on ultrasonic pulse echo is studied. Experimental results show that the combination of ultrasonic echo amplitude imaging and arrival time imaging can be used to identify the fracture location, width, depth and orientation, along with accurately calculating the fracture dip angle. The evaluated fracture parameters are similar to those in the physical simulation model. The identification accuracy of the ultrasonic measurement is related to the diameter of the radiation beam of the ultrasonic transducer. A single fracture with width larger than or equal to the radiation beam diameter of the ultrasonic transducer and multiple fractures with spacing longer than or equal to the radiation beam diameter can be effectively identified.

The impact of funding models on the integration of registered nurses in primary health care teams: protocol for a multi-phase mixed-methods study in Canada

BackgroundFamily practice registered nurses co-managing patient care as healthcare professionals in interdisciplinary primary care teams have been shown to improve access, continuity of care, patient satisfaction, and clinical outcomes for patients with chronic diseases while being cost-effective. Currently, however, it is unclear how different funding models support or hinder the integration of family practice nurses into existing primary health care systems and interdisciplinary practices. This has resulted in the underutilisation of family practice nurses in contributing to high-quality patient care.MethodsThis mixed-methods project is comprised of three studies: (1) a funding model analysis; (2) case studies; and (3) an online survey with family practice nurses. The funding model analysis will employ policy scans to identify, describe, and compare the various funding models used in Canada to integrate family practice nurses in primary care. Case studies involving qualitative interviews with clinic teams (family practice nurses, physicians, and administrators) and family practice nurse activity logs will explore the variation of nursing professional practice, training, skill set, and team functioning in British Columbia, Nova Scotia, Ontario, and Quebec. Interview transcripts will be analysed thematically and comparisons will be made across funding models. Activity log responses will be analysed to represent nurses’ time spent on independent, dependent, interdependent, or non-nursing work in each funding model. Finally, a cross-sectional online survey of family practice nurses in Canada will examine the relationships between funding models, nursing professional practice, training, skill set, team functioning, and patient care co-management in primary care. We will employ bivariate tests and multivariable regression to examine these relationships in the survey results.DiscussionThis project aims to address a gap in the literature on funding models for family practice nurses. In particular, findings will support provincial and territorial governments in structuring funding models that optimise the roles of family practice nurses while establishing evidence about the benefits of interdisciplinary team-based care. Overall, the findings may contribute to the integration and optimisation of family practice nursing within primary health care, to the benefit of patients, primary healthcare providers, and health care systems nationally.

Prediction of Reservoir Fracture Parameters Based on the Multi-Layer Perceptron Machine-Learning Method: A Case Study of Ordovician and Cambrian Carbonate Rocks in Nanpu Sag, Bohai Bay Basin, China

Developing a model that can accurately predict internal fractured reservoirs in the context of the ultra-low physical properties of carbonate rocks by only employing conventional mathematical methods can be very challenging. This process is challenging because the relationship between basic fracture parameters and the logging response in carbonate reservoirs has not been studied, and the traditional method lacks adaptability due to the complex relationship between basic fracture parameters and the logging response. However, data-driven approaches supplemented by machine learning algorithms based on multi-layer perceptrons (MLP) provide a more reliable solution to this challenge. In this paper, a classical fracture parameter evaluation data set is established using fracture porosity, fracture density, fracture length, and fracture width data that can be identified by resistivity and acoustic imaging logging. Another data set can be composed of different types of logs, and it can be used to identify reservoirs. Two different data sets were validated by regression task evaluation indicators in machine learning, and the correlation coefficient R2 is greater than 0.82. This means that the model accuracy of the algorithm can reach 82%. Combined with the comparison results of eight conventional machine learning algorithms, the reliability and application validity of the MLP model are verified. This method’s accuracy is also verified by oil test data, which show that the MLP machine-learning algorithm can effectively simulate the relationship between lithology and fracture development. In addition, it can be used to predict key exploration horizons before drilling. The relationship between lithology and fracture development degree is well-simulated by the MLP machine learning algorithm, which shows that the degree of fracture development is mainly affected by fractures, indicating that the method can be used to predict key exploration horizons before drilling.

Effect of lithological variations on the performance of artificial intelligence techniques for estimating total organic carbon through well logs

By the expansion of production from source-related unconventional petroleum resources, accurate approximation of Total Organic Carbon (TOC) through well logs has become progressively important. Accordingly, recent studies have mainly focused on increasing the precision of TOC estimation by using different types of AI techniques and/or optimizing algorithms. Along with utilizing these approaches, this study emphasized on removing an unaddressed source of error inherited from lithological heterogeneity with the same goal. Like organic matter quantity, lithological variations within a source interval also induce well log responses, which may interfere with the training process of Artificial Intelligence (AI) techniques. In the present research, the effect of lithological variations on the performance of TOC estimators was evaluated by employing Adaptive Neuro Fuzzy Inference System (ANFIS) and Multilayer Perceptron network (MLP). Firstly, ANFIS and MLP models were constructed and trained using a database containing different lithologies (original models). Then, a new methodology was developed based on modeling the relationship between log data and TOC values for each type of lithology (litho-based method). The result showed that the litho-based method estimates more reliable and accurate TOC values, proving the adverse effect of lithological variations on the original models. Furthermore, the litho-based ANFIS models provide the most promising results. Since metaheuristic algorithms are usually employed to optimize AI techniques, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) were also implemented into the original models (hybrid models). Accuracy of TOC values estimated by the hybrid models is slightly higher than those derived from the original models. However, these hybrid approaches are not as efficient as the litho-based method. Applicability of the proposed approach was guaranteed by performing it over Pabdeh source rocks in a well of SW Iran. Based on its high efficiency, the newly developed litho-based method can be used as a powerful tool to reliably evaluate unconventional hydrocarbon resources, as well as the potential of the conventional petroleum sources. Moreover, it can be utilized, instead of/along with traditional optimization approaches, to approximate other geochemical factors as well as petrophysical parameters from log data.

Single-well lithofacies identification based on logging response and convolutional neural network

To address lithofacies identification of single-well, we proposed two solutions based on logging response and convolutional neural network. Firstly, the lithofacies types are obtained according to the coring well data, and then the logging response characteristics of different lithofacies are obtained by associating the lithofacies with some logging responses. Finally, the mapping relationship between logging responses and lithofacies types is established by convolutional neural network. The trained network can be used for single-well lithofacies identification. In the first scheme, we employed a deep residual network based on transfer learning. This network receives three-dimensional image as input and use two-dimensional convolution to extract pattern characteristics. Hence, it needs to plot the logging responses of each lithofacies into an image in advance. In the second scheme, we employed a parallel model composed of residual network and long short-term memory network, which can directly receive the sequence data of logging response without transforming them into images. The lithofacies identification results of 2 test single-wells show that the identification accuracy of the two schemes is more than 92%, which provides a new way for lithofacies identification in oil and gas exploration and development.

New direction for regional reservoir quality prediction using machine learning - Example from the Stø Formation, SW Barents Sea, Norway

Recently, the petroleum industry has focused on deeply buried reservoir discoveries and exploring potential CO2 storage sites close to existing infrastructure to increase the life span of already operating installations to save time and cost. It is therefore essential for the petroleum industry to find an innovative approach that exploits the existing core- and well log data to be successful in their endeavor of effectively characterizing and predicting reservoir quality. Continuous data sources (e.g. wireline logs) have a huge potential compared with expensive, time inefficient and sporadic data from cores in determining reservoir quality for use in a regional context. However, whereas core analysis offers in-depth knowledge about rock properties and diagenetic processes, continuous data sources can be difficult to interpret without a formation-specific framework. Here, we demonstrated how the pre-existing core data could be effectively used by integrating petrographic- and facies data with a pure predictive machine learning (ML) based porosity predictor. The inclusion of detailed core analysis is important for determining which reservoir parameter(s) that should be modeled and for the interpretation of model outputs. By applying this methodology, a framework for deducing lithological and diagenetic attributes can be established to aid reservoir quality delineation from wireline logs that can be used in frontier areas. With the ML porosity model, a Random Forest Regressor, the square of the correlation was 0.84 between predicted- and helium porosity test data over a large dataset consisting of 38 wells within the Stø Formation across the SW Barents Sea. By integrating the continuous ML porosity logs and core data, it was possible to differentiate three distinct bed types on wireline log responses within the Stø Formation. Particularly, the relationship between Gamma ray (GR) and porosity was effective in separating high porosity clean sand-, low porosity cemented clean sand and more clay and silt rich intervals. Additionally, in the P-wave velocity (VP) - density domain, separation of high porosity clean sand- and heavily cemented low porosity clean sand intervals were possible. The results also show that the ML derived porosity curves coincide with previously published and independent facies data from a selection of the wells included in the study. This demonstrates the applicability of the model in the region, because the Stø Formation has been described to exhibit similar lithological- and mineralogical properties over large parts of the Western Barents Sea area. Even though, continuous porosity data could be estimated from other sources like VP, neutron or density logs, this would generally require matrix and fluid information. This study demonstrated the effectiveness of the ML model in generating continuous porosity logs that are useful for characterizing and predicting reservoir properties in new wells. This methodology offers a workflow for exploiting already acquired core and well log data for frontier exploration that can be adapted to other formations and exploration scenarios worldwide.