EAGE Conference on Energy Excellence: Digital Twins and Predictive Analytics
- Conference date: October 15-16, 2024
- Location: Kuala Lumpur, Malaysia
- Published: 15 October 2024
1 - 20 of 23 results
-
-
SqueezeNet-Based Fault Diagnosis and Predictive Maintenance for Sucker Rod Pumps using ECOC Classifiers for Asset Management
More LessAuthors M. Soni, S. Shukla, H. Sreenivasan and S. KrishnaSummaryThis study explores the application of SqueezeNet, a lightweight deep learning architecture, for fault diagnosis in sucker rod pumps within the oil and gas industry. Initially, SqueezeNet encountered challenges in accurately classifying fault conditions. To address this, the integration of Error-Correcting Output Codes (ECOC) classifiers, specifically Support Vector Machines (SVM) and k-Nearest Neighbors (k-NN), was implemented, resulting in a significant improvement in accuracy and a reduction in misclassifications. Additionally, data augmentation techniques were employed to enhance the diversity and robustness of the training data, further boosting the model’s generalization capabilities.
Our findings highlight the transformative potential of combining deep learning with advanced machine learning techniques for predictive maintenance in critical infrastructure. The improved fault diagnosis accuracy can optimize asset management, leading to increased operational efficiency and enhanced reliability. These advancements can guide decision-making processes, resulting in substantial cost savings and improved asset reliability for sucker rod pump operations and other applications. This study underscores the value of innovative approaches in the oil and gas sector, paving the way for more efficient and reliable operations through the adoption of cutting-edge technologies.
-
-
-
Multi-Component Analysis into Reservoir Connectivity in Complex Fluvio-Deltaic Hydrocarbon Fields, Gulf of Thailand
More LessAuthors J. Charoensuk and T. TonburinthipSummaryFields in the Gulf of Thailand are currently focused on optimizing infill well placement to maximize extraction from both shared and unshared hydrocarbon reservoirs. This strategy utilizes an advanced Gulf of Thailand model that integrates specialized concepts of infill well spacing. A key objective is to identify new reservoirs to unlock additional potential from these wells.
Traditionally, reservoir connectivity is identified using wireline logging, pressure tests, and geological data. Relationships between the width and thickness of sandstone bodies in specific depositional environments help predict reservoir connectivity. The methodology involves manual identification of reservoir connectivity and facies based on well data. The distance to connected reservoirs is measured and projected to determine true channel width trends, validating width-to-thickness ratios by facies and units. A Random Forest regression model assesses parameter influences to predict true channel width and sand connectivity among wells.
The width-to-thickness ratios differ significantly between the channel (1:20 to 1:700) and bar (1:200 to 1:450) sandstone facies, reflecting distinct reservoir geometries and connectivity patterns. The model’s accuracy is validated using Leave-One-Out Cross-Validation, achieving high performance (80%-90%) across geological units. This predictive tool supports new reservoir identification and optimization of reservoir productivity in the Gulf of Thailand.
-
-
-
Application of Advanced Artificial Intelligence and Machine Learning System for Subsurface
More LessAuthors M. SajidSummaryThe study proposes a new approach to interpret overburden faults in CCS storage studies using AIML. The approach aims to process large amounts of data with efficiency and precision than traditional methods. The study focuses on identifying potential risk zones for CO2 storage in depleted oil/gas reservoirs. The AIML system enhances the precision of fault location and identifies new faults, which could aid in future CCS storage program planning and risk mitigation studies.
-
-
-
Static Model Building Using Artificial Intelligence and Machine Learning Techniques
More LessAuthors A. Cheerappath Aravindakshan, B. Chennakrishnan and S. SenapatiSummaryThe study demonstrates the application of supervised machine learning algorithms to classify reservoir rock types using labeled data from well logs, cores, and other sources, improving accuracy and reducing manual interpretation time.
A 3D structural grid was constructed using Python, incorporating well tops, fault sticks, and depth horizons. Facies were predicted and populated throughout the reservoir zone using a Random Forest classifier algorithm.
Additional reservoir properties such as porosity, volume of shale, and saturation were populated across the entire grid using the XGBoost regressor algorithm, with the derived facies serving as an independent variable.
In-place resources were estimated using both deterministic and probabilistic methods, including sensitivity analysis, with estimates varying within 5% of those from a benchmark conventional model.
The study highlights that AI and ML methodologies can automate static modeling processes, facilitating data-centric reservoir description and swift updates to models. This is particularly beneficial for mature fields with extensive well data.
-
-
-
Application of Supervised Machine Learning in Porosity Prediction from Seismic Data and Uncertainty Quantification
More LessAuthors N. Rai, B. Chennakrishnan and A. RaySummarySeismic quantitative interpretation in oil and gas exploration benefits significantly from the integration of machine learning (ML) techniques, as highlighted in this study. By leveraging ML, the research focuses on optimizing the selection of elastic attributes derived from seismic data to predict porosity in reservoirs accurately and efficiently. Attributes such as P-Impedance, S-Impedance, Lamda-rho, Mau-rho, Poisson impedance_fluid, and Poisson impedance_Litho are systematically evaluated against porosity data from well logs to establish robust correlations using various ML algorithms.
The study emphasizes the automation of attribute selection, which traditionally is labor-intensive and prone to uncertainties. Through ML, the process becomes streamlined, enhancing the accuracy of porosity mapping and overall reservoir characterization. Moreover, the research addresses the challenge of uncertainty quantification by demonstrating high correlations between predicted and observed porosity values across training and validation wells. This validation underscores the reliability and generalizability of the ML models developed.
Ultimately, the application of ML techniques in seismic data interpretation offers substantial improvements in efficiency and reliability. By reducing uncertainties and providing rigorous uncertainty quantification, this approach enhances decision-making in the oil and gas industry, contributing to more informed reservoir modeling and exploration strategies.
-
-
-
Digital Field Development Planning
More LessAuthors B. Moradi, S. Völgyi, S. Behjat, J.O. Knutsen and S. BraeuningSummaryIdentifying and quantifying infill targets are vital for optimizing ongoing oil field production. This study a Norwegian North Sea field by a novel digital approach that combines physics-based principles with data-driven techniques. This digital hybrid method generates historical fluid distribution maps, predicts future fluid movements, and identifies potential locations for new wells to enhance recovery. The performance of identified targets was predicted by advanced machine learning techniques. A comparison between the hybrid workflow results and results predicted by numerical simulation showed a high degree of consistency. This hybrid approach enables rapid, accurate digital field development planning, addressing a wide range of uncertainties and providing valuable insights for the E&P industry. The findings demonstrate that integrating data-driven and physics-based methods significantly enhances the ability to locate and quantify remaining oil, optimizing field development strategies.
-
-
-
Innovative Machine Learning techniques for Improved Reservoir Modeling and Production Forecasting
More LessAuthors A. Kumar, A. Verma, R. Kumae and S. SenapatiSummaryThis paper investigates the use of advanced machine learning techniques, particularly the Random Forest algorithm, to improve reservoir modeling and production forecasting, highlighting significant improvements over traditional methods.
The study emphasizes the transformative potential of machine learning (ML) techniques in addressing the complexities of reservoir modeling and production forecasting by leveraging advanced algorithms to process data and model nonlinear relationship
The methodology involves using actual field production data, including parameters like flowing bottom-hole pressure (FBHP) and tubing head pressure (THP), to create a dynamic model. Porosity and permeability data from static models are integrated into the dynamic model, with the Random Forest algorithm selected for its adaptability to complex datasets
The model undergoes a rigorous history matching process and double-blind testing, demonstrating robust performance and maintaining consistent accuracy in forecasting production rates with an overall error margin of 15%.
Results indicate a notable improvement in forecasting accuracy compared to traditional methods, with significant implications for optimizing reservoir performance through strategic decisions such as infill drilling and workover opportunities
-
-
-
Optimizing Oil and Gas Assets Using Machine Learning Algorithms at Different Levels of Detail
More LessAuthors K. Pechko, S. Ivanov, A. Afanasev and M. SimonovSummaryThis paper presents a novel approach for optimizing oil and gas field development using multi-level and proxy models integrated with machine learning algorithms. The traditional method of calculating Integrated Asset Models (IAMs) sequentially is time-consuming and often impractical for timely decision-making. Our proposed multi-level optimization approach accelerates IAM computations, enabling the evaluation of a greater number of development scenarios within shorter time frames. By utilizing machine learning to create fast, less detailed models, this method prioritizes scenarios based on predicted outcomes, ensuring a balance between computational efficiency and result accuracy. The approach also incorporates data from reservoir models, well models, and surface gathering and transportation systems to enhance model fidelity. This method not only reduces computation time but also facilitates rapid and accurate strategic decision-making, essential for maintaining competitiveness in the dynamic oil and gas industry. The implementation of this optimization strategy promises to improve the efficiency of field management and development, leading to higher extraction rates and increased profitability.
-
-
-
Exploration Missed Opportunities Identification using Total Gas Augmented Machine Learning
More LessAuthors M. Salamat, I. M Fadhil, N.A.A. Azizul, G. Malo-Paul, H.F. Hasnan and T. Shu WenSummaryAccurately identifying net pay zones is crucial for maximizing hydrocarbon recovery and optimizing field development. Traditional methods often struggle with geological complexities. This study integrates total gas (TGAS) measurements from mud logs, gamma ray (GR), and resistivity (RT) from well logs to enhance net pay zone predictions. TGAS reflects the gas content of the subsurface formations which complements conventional logs.
We utilized the Random Forest Classification algorithm, which is adept at handling complex geological data. This involved preprocessing TGAS, GR, and RT data through cleansing, normalization, and division into training and validation sets. The model focused on maximizing recall to identify potential hydrocarbon zones, accepting a higher false positive rate for broader detection.
The model’s performance was evaluated across 12 wells. While the training data showed perfect scores the test data indicated overfitting with lower scores (F1 at 0.62, precision at 0.60, and recall at 0.63). Some wells, such as K1 (recall 0.34), performed better, though others had poor recall due to a lack of true positive identifications.
To minimise false negatives, future recommendations include adjusting classification thresholds, cost-sensitive learning, and ensemble methods. Integrating data science enhances efficiency, reduces costs, and maintains competitiveness in the energy sector.
-
-
-
Predictive Seismic Velocity from Rock Characteristics: Automating AVO Analysis for Hydrocarbon Exploration
More LessAuthors M. Salamat, I. M Fadhil, G. Malo-Paul and H.F. HasnanSummaryThis study focuses on automating the creation of Amplitude Variation with Offset (AVO) sensitivity matrices using machine learning (ML) to improve seismic exploration. Traditional methods are labor-intensive and prone to inaccuracies due to manual processes and expert interpretation. We developed a Random Forest model to estimate Bulk Density (RHOB), Compressional Velocity (VP), and Shear Velocity (VS) in the Balingian area, Offshore Sarawak.
Our model integrated well log data and employed sequential prediction for enhanced accuracy. Data preprocessing involved removing outliers and bad data points. Key features included gamma ray, bulk density, and seismic velocities. Hyperparameter tuning via Grid Search with cross-validation ensured optimal performance.
Results showed the model performed well under simpler fluid conditions (INSITU and BRINE) but struggled with complex conditions (OIL and GAS), as indicated by higher error rates and negative R-squared (R2) values. Metrics such as Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE) highlighted these challenges.
This study demonstrates ML’s potential to automate and improve seismic velocity predictions, aiding petrophysics in subsurface understanding and drilling target identification. Further refinement and additional data are necessary to enhance predictive accuracy for complex fluid scenarios, paving the way for more efficient and accurate seismic exploration processes.
-
-
-
Integrating Geostatistical Models and Machine Learning for Infill Well Placement and Production Forecasting in Mature Field
More LessAuthors A. Kumar, H. Golghanddashti, D. Selvaraj, R. Kumar, A. Verma and S. SenapatiSummaryOptimizing infill well placement in mature oil fields is complex due to reservoir dynamic complexity and extensive production history.
ombining geostatistical models with machine learning workflows yields quick and effective results. Geostatistical models capture geological and petrophysical properties, while machine learning analyzes and predicts time-dependent variables.
The production model involves history matching and forecasting using a workflow that includes initial production estimation, reservoir pressure prediction, generating missing data, predicting production rates, and model validation. The primary algorithm used is Random Forest.
The saturation model employs XGBoost for modeling lateral-vertical saturation movement and creating a 3D saturation model. This model predicts saturation distribution during history and forecasts future saturation distribution.
The infill location model identifies potential infill well locations and perforation intervals, and forecasts production potential. It uses well filtering, clustering analysis, and pressure estimation to optimize production efficiency..
The integrated approach successfully identifies multiple infill locations without simulation, providing better results than data-driven models alone. Validation of results after each model run and scrutiny of suggested locations are crucial for forecasting production profiles.
-
-
-
Automating Well Log Pattern Classification for Depofacies Identification using Deep Learning
More LessAuthors O. Ridhwan Sazali, T. Shu Wen, N.S. Mauzi, I. M Fadhil, G. Malo-Paul and H.F. HasnanSummaryConventional depofacies classification methods are often hindered by the massive data volumes, potential biases, and the labor-intensive nature of the process. Gamma ray logs are particularly effective for facies and sequence stratigraphy analysis due to their diverse shapes, higher resolution, and distinctive features. To enhance the accuracy and efficiency of depofacies prediction, this study employs a computer vision based deep learning approach to classify gamma ray logs into various categories of depofacies.
We applied convolutional neural networks (CNNs), known for their proficiency in handling graphical data. The volume of shale was derived from Gamma Ray (GR) by Steiber equation to identify the sand and shale interfaces, marking the top and bottom of these layers. Gamma ray (GR) logs were then preprocessed by converting them into images, cleansing the data, and dividing it into training and validation sets. The CNN model was trained to recognize coarsening upward, fining upward, bed, and serrated classes of gamma ray logs.
The model’s performance was assessed across a substantial number of images using a confusion matrix, achieving a training accuracy of 85%. This machine learning approach demonstrates significant potential in automating and improving the accuracy of well log classification which helps in identify depofacies.
-
-
-
Transforming OEE Optimization in Oil and Gas: Integrating AI and Sensitivity Analysis for Improved Performance Analysis
More LessAuthors Y.H. Hiew, P. Chandra Shegaran, C.A. Kang, N.S.H. Mohd Suhaimi and S. Mohamad TermidziSummaryThis paper presents an AI-driven approach aimed at improving Overall Equipment Effectiveness (OEE) in oil and gas production facilities. The study focuses on controllable factors affecting OEE, specifically internal unplanned shutdowns and slowdowns. For the analysis, four plants with varying capacities were selected. Data was then used to engineer monthly-derived features to analyze temporal dynamics. Feature selection was refined through Pearson’s correlation and input from Subject Matter Experts (SMEs) to ensure impactful features. Three models—Polynomial Chaos Expansion, Sparse Approximation, and Gaussian Process Regression—were developed and validated for each plant using Leave-One-Out Cross-Validation (LOO CV) and evaluated with Mean Absolute Percentage Error (MAPE). Sobol sensitivity analysis highlighted primary features influencing production loss variance. The model exhibiting the strongest feature-target relationship and lowest MAPE was selected. Scenario planning was employed to simulate the effects of hypothetical adjustments on future production loss, providing insights for strategic interventions. The findings revealed more predictable outcomes in worse performing plants and challenges in stable performance plants. The study highlights limitations such as monthly data reliance and suggests that daily data and early indicators could improve accuracy. The study underscores the necessity for customized optimization strategies tailored to the unique operational characteristics of each plant.
-
-
-
Applying Digital Transformation in Oil and Gas Reservoir for Real-time Management, Control and Optimization
More LessAuthors L. Pires and D. SchiozerSummaryAdvancements in computing, data storage, and communication have revolutionized the oil and gas industry through artificial intelligence (AI), machine learning (ML), and big data. These technologies, while enhancing analysis efficiency, are complemented by numerical simulators to manage and optimize reservoirs. The integration of AI and ML with traditional methods, such as simulation models, may be used to improve forecast accuracy by addressing the limitations of each approach. Digital transformation is central to this evolution, automating previously manual processes, optimizing equipment operations, and improving decision-making through advanced data analysis and visualization tools. This study presents the Digital Transformation application to short-term management, control and forecast. Key components include real-time monitoring, predictive analytics and control, and a three-stage optimization process that enhance short-term and life-cycle reservoir performance. These technologies are implemented in UNISIM-IV Benchmark to test and validate the proposed workflow to improve reservoir short-term forecast, management and optimization. This comprehensive approach leads to increased productivity, cost reduction, and more sustainable practices.
-
-
-
Prediction of Kinematic Viscosity for Lube Base Oils using Feedstock Properties and Reactor Operating Parameters
More LessSummaryIn the refinery industry, producing Group III Base Oils requires stringent control over viscosity index (VI) and kinematic viscosity (KV) as key measures of product quality. The traditional process of sampling and experimentally determining viscosity is highly time-consuming, often taking up to one full shift. This extended duration can lead to product waste and suboptimal yield, posing significant challenges for efficient production. To address these issues, a regression model using the LightGBM algorithm was developed to predict base oils’ VI and KV. By implementing Recursive Feature Elimination (RFE), crucial factors such as feedstock properties and compositions, flow rate, hydrogen flow rate, pressures, and temperatures were identified as significant contributors to the VI and KV of base oils. The model’s performance was impressive, demonstrating mean absolute errors (MAE) of 1.12 for VI, 0.88 for KV at 40°C, and 0.20 for KV at 100°C. On average, the three models achieved an R-squared value of 98% using LightGBM. The model developed in this study holds substantial potential to aid operators in estimating product rundown times accurately while maximizing product yields. By significantly reducing the time and labor associated with traditional sampling methods, this predictive model enhances operational efficiency and ensures consistent product quality.
-
-
-
Super Resolution Seismic Imaging through Deep Information Maximization: Optimizing Exploration and Development in the Chandon Field
More LessAuthors D. Nanda, M. Khanna, A. Yadav, S. Pattanaik and N. SeerviSummaryThis study introduces a novel generative deep learning-based methodology that enhances seismic resolution. This is crucial for optimizing field exploration and development as it enables more accurate geological interpretations. By integrating geological and geophysical knowledge with data-driven models, the methodology significantly improves subsurface imaging and inversion. Applied to the Chandon Field, offshore NW Australia, this approach has achieved substantial resolution enhancements, allowing for the detailed interpretation of previously unresolved stratigraphic and structural features. The improved seismic data and acoustic impedance models have enabled precise mapping of facies distribution and strati-structural plays, particularly within the Mungaroo and Brigadier Formations. The high-resolution seismic imaging has resulted in clearer visualization of structural and depositional features, facilitating confident fault interpretation and the identification of key geological structures, such as a strike-slip corridor. This work highlights the potential of deep learning combined with domain expertise to drive advancements in reservoir characterization and overall field management.
-
-
-
Digital Twin-Assisted Reservoir Management for Infill Drill Optimization
More LessSummaryThis study presents the development of an advanced digital twin technology designed to optimize infill drill locations within coal-bed methane reservoirs. By facilitating data-driven decision-making, the technology identifies the most productive drilling locations, enhancing overall reservoir management. The methodology encompasses designing well trajectories and the identification of coal seams, followed by the characterization of reservoir and geological properties. This is succeeded by the optimization of hydraulic fracturing parameters and the assessment of offset well interference at the planned well locations. These factors are used to train AI models capable of forecasting gas and water production during both transient and steady-state phases, with careful consideration of the varying physics and reservoir dynamics inherent to each phase. Testing and blind datasets validate these forecasts, demonstrating high accuracy. Additionally, the digital twin architecture performs Key Performance Indicator (KPI) analysis to identify the most influential input variables affecting production forecasts. By analyzing production forecasts, coal seam data, and reservoir characteristics, the technology effectively determines the optimal drilling locations. The modular design of the system supports continuous learning and customization with live datasets, aligning predictions with field observations. This leads to significant improvements in reservoir management, drastically reducing analysis time from weeks to hours.
-
-
-
Real-Time Decision Support in Optimizing Well Production and Hydraulic Fracturing Operations using Digital Twins
More LessSummaryThis study explores the development and application of digital twin technology to enhance field production and optimization through real-time decision support. It specifically focuses on optimizing production strategies and hydraulic fracturing job designs. The approach involves training an AI model with comprehensive field data, including reservoir properties, geological features, and operational parameters. By running numerous simulations using a multi-model dynamic approach and various AI algorithms, the model identifies optimal solutions, which are then validated through history matching, and identifying key performance indicators. Simulation outputs are integrated with financial modeling to reduce operational costs. The research highlights two key applications: optimizing production strategies and hydraulic fracturing job designs. For production strategy optimization, the digital twin model has led to a 6.3% increase in CBM field gas production over three months by fine-tuning parameters such as pump RPM and orifice bean size. In hydraulic fracturing, the model predicts optimal fracture length and permeability to enhance production efficiency, thereby identifying cost-effective fracking job design parameters. This AI-driven framework offers a powerful tool for real-time decision-making, significantly improving efficiency and profitability in field operations.
-
-
-
Transforming 2D Seismic into Pseudo3D
More LessAuthors D. Markus, R. Muammar, K. Rimaila and P.K. De GrootSummaryWe have developed a methodology known as Pseudo3D. This advanced technique is uses Machine Learning to enhance the value of 2D seismic data by transforming it into a more informative 3D format. Pseudo3D leverages input data from all available 2D migrated stacks and from various 3D vintages of seismic surveys, if available.
2D seismic data remains crucial for companies assessing new permits or exploring new basins, providing foundational insights that guide their decision-making processes. This type of data is particularly valuable in the initial stages of exploration, where comprehensive and detailed information about subsurface structures is essential.
While 2D seismic data is invaluable during the exploration phase, the lack of lateral continuity may make these datasets less suitable for appraisal and development stages. Additionally, the complexities associated with multi-vintage datasets necessitate careful and detailed analysis to ensure accurate and reliable results.
Our Pseudo3D is primarily a post-stack data-driven process, thus improving efficiency by not requiring complex and resource intensive methods that involve pre-stack processing such as de-migration and migration. We leverage machine learning models to approximate results of pre-stack quality while only using post-stack data.
-
-
-
Machine Learning in Real-Time Sampling Advisory to Predict OBM Contamination Levels
More LessAuthors S.Y. Lim, A. Abdollahzadeh and N.A. SulaimanSummaryReservoir fluid samples obtained from testing or production wells need to undergo PVT analysis to determine its properties under varying pressure, volume and temperature conditions. However, samples need to be tested for contamination, to ensure that the sample obtained is clean and representative of the downhole formation fluid, so that the PVT analysis gives reliable results.
During sampling, formation fluid is pumped through the wireline to the borehole. Initially, the fluid that flows will primarily consist of mud filtrate, and as pumping continues, the sample will gradually transition to a mixture of mud filtrate and formation fluid. Ideally, it is best for the sample to achieve the lowest possible contamination level, but it is impractical to pump indefinitely due to the cost of the rig time. The proposed solution, Live Integrated Sampling Advisory (LISA) leverages predictive modeling to perform real-time composition analysis of C1-C6+ hydrocarbon components to assess the contamination level of reservoir fluid. This could be implemented real-time, to aid reservoir engineers in determining the optimal time to sample reservoir fluid with contamination level within acceptance range.
-