Three-dimensional Reservoir Model Research Articles

Numerical Simulation of Fracturing Fluid Storage in Shale Reservoirs Based on Experimental Measurements of Stress Sensitivity of Hydraulic Fracture Network Conductivity

Hydraulic fracturing is used in shale reservoir production, with low flowback rates and a large amount of fracturing fluid retained inside the reservoir. In this study, a stress sensitivity analysis experiment on the fracture inflow capacity was implemented to investigate the relationship between the hydraulic fracture (HF) and natural fracture (NF) inflow capacities and effective stress. A three-dimensional shale reservoir model was also constructed to couple the experimentally obtained laws with the numerical model to investigate the effects of the connection and closure of the fracture network on the retention of the fracturing fluid. The results show that the stress sensitivity of natural fractures is two orders of magnitude higher than that of hydraulic fractures. The seepage-absorption effect of capillary forces is not the whole reason for the large amount of fracturing fluid retention. The closure of the fracture network formed by natural and hydraulic fractures during the production process led to the storage of a large amount of fracturing fluid, and this process maintained the stability of the water production rate during the steady water production period.

Permeability Data Treatment from Analogous Outcrop for Use in Three-Dimensional Reservoir Model

Over the past forty years, the study of analogous outcrops has become a common practice in the petroleum industry to access more comprehensive geological datasets and better parameterize various geological features. As most reservoirs do not outcrop and all the data obtained are restricted to those collected from wells located in areas of major interest or from indirect measurements, the study of analogous outcrops has brought important contributions to improve the production routines and still represents a helpful information resource. The use of data from analogous outcrops permits, for example, to estimate the sedimentary facies distribution in true reservoirs, and thus to improve on the reliability of fluid flow modeling during the oil field exploration. However, such data must be processed to become equivalent to the data collected under the reservoir conditions and the results obtained to be more realistic. This work presents the statistical method used to make permeability data obtained from an outcrop compatible with the subsurface conditions of poorly consolidated and tectonically deformed sandstone reservoirs. The studied outcrop is predominantly composed of Eocene fluvial sandstones (Resende Formation) of the Volta Redonda Basin, located in the Central Segment of the Continental Rift of Southeastern Brazil. These deposits are affected by deformation bands and faults. From statistical analysis, it was intended that the permeability data become more suitable to be used, in the next phase of the studies, to build a 3D flow model, in which it will be possible to test the potential of different tectonic structures acting as hydraulic barriers in poorly consolidated and deformed reservoirs. The methodological approach applied in this study made it possible to turn compatible the permeability data obtained in conditions of no confining pressures from the outcrop with reservoir conditions under confinement pressures of 2,000 psi. Such pressure conditions are frequently observed in important Maastrichtian reservoirs found in the Brazilian Southeastern Margin.

3D geological and petrophysical modeling of Alam El-Bueib Formation using well logs and seismic data in Matruh Field, northwestern Egypt

There are several productive petroleum fields in the North Western Desert (WD) of Egypt, which received extensive investigations regarding their petroleum potential. However, a few studies tackled the Matruh Oil Field, which contains the oil prolific Early Cretaceous Alam El-Bueib Formation (AEB Fm) reservoir. The reservoir intervals of the AEB Fm show substantial lithological variations across the basin. Therefore, it is necessary to analyze the vertical and lateral distributions in terms of their lithological and petrophysical properties. To achieve this objective, wireline logs of four wells and 20-2D seismic lines were used to construct a depth-structure contour map for the studied part of the field. This map was used to build the field’s structure model and to identify the fault patterns in the basin through several seismic lines. Analyses of well logs data and lithology were used to estimate the petrophysical properties of AEB sandstone units AEB-1, AEB-3A, AEB-3C, and AEB-6. Results show that the AEB-6 Unit is the most promising hydrocarbon-bearing unit. It has a net pay of 20–160 feet, a shale volume of 5–20%, an effective porosity of 14–20%, and a hydrocarbon saturation of 70–88%. The structure-depth maps indicate a number of normal faults with two principal NE-SW and NW–SE trends, which probably act as structural traps in the Matruh Oil Field. The constructed structure-depth maps and calculated petrophysical parameters were used to build a three-dimensional reservoir model. A blind well was used to validate the accuracy and reliability of the facies, porosity, and saturation models for the AEB Fm units, ensuring a good match between log-derived data and built models. The AEB Fm shows regional heterogeneous variations in its petrophysical characteristics. It exhibits unconventional reservoir characteristics in a N–S direction and conventional reservoir characteristics in an E–W direction. This observed heterogeneity shows the need to carry out further investigations to comprehensively assess the hydrocarbon potential of AEB Fm in different areas of the Matruh Basin.

3D Geomechanical Finite Element Analysis for a Deepwater Faulted Reservoir in the Eastern Mediterranean

AbstractHydrocarbon reservoir structures are subjected to tectonic forces along the geological time that cause rock deformation and break into faulted zones. Faulted reservoirs, enclose certain complexity in terms of the distributed effective stresses, rock plastic alteration, slipping and fault block displacement. In this study, we develop a three-dimensional (3D) geomechanical reservoir model with faulted and compartmentalized geometry, located in the offshore deepwater environment of the Levantine basin in the Eastern Mediterranean, based on non-linear finite element analysis (FEA). A regional structural and stress map was also constructed, integrating various data sources, to present the regional stress setting to enhance this work. The assessment of the geomechanical impacts on the reservoir provides important information in reservoir studies, that can analyze potential stability issues during the depletion to optimize the field production planning. Stress–strain evolution in the reservoir is primarily affected by the in situ stresses, the geometry of faults, and the degree of compartmentalization. The results demonstrate clearly the mechanism of stress transfer transmission and the impact between the fault block compartments in the reservoir. Fault contacts exhibit a higher tendency for rock displacements and deformations. Plastic yielding develops at a narrow extent along the faults. The risk of fault slipping depends on the depletion strategy, but it is low in all cases. No significant reduction in permeability was found at the end of reservoir depletion. Overall, geomechanics integration enriches and improves the dynamic reservoir models and applications.

Simultaneous inversion of permeability, skin and boundary from pressure transient test data in three-dimensional single well reservoir model

This study presents a novel approach for simultaneous inversion of the key reservoir parameters like horizontal permeability, vertical permeability, skin, and boundary distances for spatial distribution across the grid cells in a 3D single well reservoir model (SWRM). These parameters are first estimated from the standard pressure transient analysis of well test pressure and rate data, which also act as a priori for the inverse problem. A field-worthy layer cake geological model is prepared based on the prior information obtained from pressure transient analysis, followed by a sequential flow simulation of field well test operation. The simulation results provide the model pressure versus rate data as the synthetic data for this study. A cost function is defined incorporating the well test pressure data and model pressure data, which would determine the convergence. The inversion process is to optimize the spatial distribution of reservoir parameters to minimize the difference between the measured pressure transient data and the modelled one, which is obtained from the multiphase fluid flow simulator that solves the implicit black-oil fluid-flow diffusivity equations at every step. A Gauss-Newton (GN) inversion scheme is used for the inversion. The reliability of inversion results depends on the accuracy of priori reservoir parameters fed to the solver, which can be refined if required through uncertainty parameter optimization (UPO). This approach helps to obtain a faster and reliable update of reservoir parameters in a layer cake homogeneous geomodel, hereby introducing the required heterogeneity. This increases the confidence and reliability of a geomodel, which is further used for various production prediction strategies.

Feature engineering process on well log data for machine learning-based SAGD performance prediction

The conventional numerical simulation requires a reliable three-dimensional reservoir model to evaluate the steam-assisted gravity drainage (SAGD) performance, but it may be time-consuming and need a significant amount of static data. This study aims to develop a machine learning (ML) model that predicts the SAGD performance based on well log data from a near single well. Although the ML model is useful in capturing underlying non-linear relationships between input features and target features, it requires a large number of training data. However, the number of data is usually small in commercial fields, limiting the availability of all log responses as input features to consider reservoir heterogeneity. In case of using small dataset, the overfitting can occur with an excessive number of model parameters. Therefore, this study introduces a feature engineering process to extract and select key features from well log data to alleviate the overfitting problem.In this study, the feature engineering process was optimized by sensitivity analysis on feature extraction and feature selection methods. For data preparation, the 63 SAGD well pairs and their matched logged wells were screened from the six commercial oil sand fields in Alberta, Canada. The SAGD process mechanism and petrophysics were considered to build data screening criteria. Each depth of the log data was categorized into pay-zone or shale barrier. For feature extraction, sampling intervals were placed on log data. This calculates the net pay thickness, averaged gamma ray, deep resistivity, and neutron porosity logs in the pay-zone. In addition, the elevation of two shale barriers and the distance between the SAGD well and logged well are extracted. Sensitivity analysis of the sampling interval length was conducted to determine the optimal conditions considering heterogeneity and feature reduction. For the feature selection, various combinations of the extracted features were used to examine their inherent meaning for predicting the SAGD performance. The multi-layer perceptron model was trained based on various selected features to predict the SAGD performance indicators of a seven-year operation. The optimal sampling interval length was determined as 10m and the combination of gamma ray, deep resistivity, shale barrier elevation, and the distance information was selected as key input features. The model showed reasonable prediction results for SAGD productivity and economic indicator with R2 of 0.67 and 0.72, respectively.

Deep reinforcement learning and adaptive policy transfer for generalizable well control optimization

Well control optimization is a challenging task but plays a critical role in reservoir management. Traditional methods independently solve each task from scratch and the obtained scheme is only applicable to the environment where the optimization process is run. In stark contrast, human experts are adept at learning and building generalizable skills and using them to efficiently draw inferences and make decisions for similar scenarios. Inspired by the recently proposed generalizable field development optimization approach, this work presents an adaptive and robust deep learning-based Representation-Decision-Transfer (DLRDT) framework to deal with the generalization problem in well control optimization. Specifically, DLRDT uses a three-stage workflow to train an artificial agent. First, the agent develops its vision and understands its surroundings by learning a latent state representation with domain adaptation techniques. Second, the agent is tasked with using high-performance deep reinforcement learning algorithms to train the optimal control policy in the latent state space. Finally, the agent is transferred and evaluated in several environments that were not seen during the training. Compared with previous methods that optimize a solution for a specific scenario, our approach trains a policy that is not only robust to variations in their environments but can adapt to unseen (but similar) environments without additional training. For a demonstration, we validate the proposed framework on waterflooding well control optimization problems. Experimental evaluations on two three-dimensional reservoir models demonstrate the trained agent has excellent optimization efficiency and generalization performance. Our approach is particularly favorable when considering the deployment of schemes in the real world as it can handle unforeseen situations.

Performance of enhanced geothermal system with varying injection-production parameters and reservoir properties

The injection-production parameters and reservoir properties significantly affect the heat extraction performance of the enhanced geothermal system (EGS), while quantitative rankings of the effecting parameters have been yet unknown. This paper is devoted to investigating numerically the heat extraction performance of EGS by three-dimensional reservoir models under the influence of varying injection-production parameters and reservoir properties. A thermal-hydro-mechanical (THM) coupled model is used to quantify the complicated heat extraction process of EGS. Three-dimensional geothermal reservoir models containing random fracture networks are established based on the Monto Carlo algorithm to quantify the effect of fracture numbers in the process of heat extraction. The parameter sensitivity of the injection-production parameters (injection mass flow rate and injection temperature) and the reservoir properties (rock matrix porosity, fracture permeability, and rock matrix permeability) on the heat extraction performance are ranked and evaluated over the Qiabuqia enhanced geothermal system. It is found the heat extraction performance shows non-monotonic relationships with the fracture numbers in the three-dimensional reservoir model, which indicates the necessity of EGS fracture optimization. The injection mass flow rate turns out the first while the fracture permeability and injection temperature are the secondary dominating factors affecting the heat extraction performance among the evaluated parameters. The outcomes are beneficial to the high-efficiency construction, operation, and optimization of the EGS especially those with similar conditions to the Qiabuqia EGS.

Retracted: CO2 Sequestration and Enhanced Shale Gas Recovery by CO2 Injection: Numerical Simulation Method

In previous decades, shale gas production has been given special attention as a clean source of energy. Technological advancements in multi-well especially horizontal drilling and hydraulic fracturing, have raised more research on how to produce shale gas commercially while studying its challenges. CH4 gas production by using CO2 injection is a combination of several parameters, including both fracking and geologic parameters. This paper investigated the influence of aligned and staggered multiwell placement patterns to enhance CH4 gas recovery and CO2 sequestration. A three-dimensional reservoir model was built and simulated, and the influence of reservoir and fracking parameters was analyzed. The results revealed that the staggered well pattern is the best pattern for CH4 gas recovery during continuous injection, with a 37% increase of recovery with late CO2 breakthrough, compared with the aligned well pattern, which had 31% increase recovery, associated with early CO2 breakthrough. In terms of CO2 gas storage, the aligned well pattern had a significant amount of CO2 storage during continuous injection and huff and puff injection by storing 4% more carbon dioxide during continuous injection and 7% more during huff and puff injection than that in the staggered well pattern. Different injection timings of 1, 2, and 3 years, and injection cycles of one, three, and five cycles were analyzed during huff and puff injection. The results showed that five cycles of injection had the highest methane gas recovery, 23% and 21%, in aligned and staggered well patterns, respectively. Furthermore, the influence of injection start time also was analyzed, and results showed that injection after 10 years of production produced higher methane gas recovery and CO2 storage in both aligned and staggered well patterns, followed by injection after 5 years of production, after 3 years of production, and during the first year of production. Sensitivity analysis revealed that horizontal matrix permeability, vertical matrix permeability, initial reservoir pressure, reservoir temperature, injection pressure, injection rate, fracture porosity, and fracture spacing played a vital role in both methane gas recovery and carbon dioxide gas storage in both well patterns. In addition, results showed that methane recovered during continuous injection is higher than during huff and puff injection in both aligned and staggered well patterns. Results obtained from this paper can help to improve shale gas production in multiwell placements, and save time and cost. This paper should motivate more research into the mitigation of global warming through CO2 geosequestration.

Geostatic modeling of the clastic reservoir: a case study the Late Cenomanian Abu Roash G Member, Hamra Field, Abu Gharadig Basin, Western Desert, Egypt

Recent advances in computer sciences have resulted in a significant improvement in reservoir modeling, which is an important stage in studying and comprehending reservoir geometry and properties. It enables the collection of various types of activities such as seismic, geological, and geophysical aspects in a single container to facilitate the characterization of reservoir continuity and homogeneity. The main goal of this work is to build a three-dimensional reservoir model of the Abu Roash G reservoir in the Hamra oil field with enough detail to represent both vertical and lateral reservoir heterogeneity at the well, multi-well, and field scales. The Late Cenomanian Abu Roash G Member is the main reservoir in the Hamra oil field. It is composed mainly of shale, carbonate and some streaks of sandstone, these streaks are shaly in some parts. Conducting the 3D geostatic model begins with the interpretation of seismic data to detect reflectors and horizons, as well as fault picking to explain the structural framework and frequently delineate the container style with proposed limitations to construct the structural model. The lithology and physical properties of Abu Roash G reservoir rock, including total and effective porosity and fluid saturation, were determined using well log data from four wells in the Hamra field. The constructed 3D geological model of the Abu Roash G has showed that the petrophysical parameters are controlled by the facies distribution and structure elements, whereas properties are the central part to the northern side of the deltaic environment than the other sides of the same environment. The model will be useful in displaying the reservoir community and indicating prospective zones for enhancing the dynamic model to improve the behavior of the flow unit productivity, as well as, section of the best sites for the future drilling.

A deep learning-accelerated data assimilation and forecasting workflow for commercial-scale geologic carbon storage

Fast assimilation of monitoring data to update forecasts of pressure buildup and carbon dioxide (CO2) plume migration under geologic uncertainties is a challenging problem in geologic carbon storage. The high computational cost of data assimilation with a high-dimensional parameter space impedes fast decision-making for commercial-scale reservoir management. We propose to leverage physical understandings of porous medium flow behavior with deep learning techniques to develop a fast data assimilation-reservoir response forecasting workflow. Applying an Ensemble Smoother Multiple Data Assimilation (ES-MDA) framework, the workflow updates geologic properties and predicts reservoir performance with quantified uncertainty from pressure history and CO2 plumes interpreted through seismic inversion. As the most computationally expensive component in such a workflow is reservoir simulation, we developed surrogate models to predict dynamic pressure and CO2 plume extents under multi-well injection. The surrogate models employ deep convolutional neural networks, specifically, a wide residual network and a residual U-Net. The workflow is validated against a flat three-dimensional reservoir model representative of a clastic shelf depositional environment. Intelligent treatments are applied to bridge between quantities in a true-3D reservoir model and those in a single-layer reservoir model. The workflow can complete history matching and reservoir forecasting with uncertainty quantification in less than one hour on a mainstream personal workstation.

Three-Dimensional Reservoir Modeling Based on Geostatistical Analysis to Determine Volumetric Oil in Place of Alpha Zone

The research of three-dimensional (3D) reservoir modeling of Alpha Zone is located in Balikpapan Formation Kutei Basin, East Kalimantan. The lithological analysis shows that Alpha Zone consists of sandstones, mudstone/shale and coal, which vertically and laterally distributed based on 206 wells and 3D seismic. Based on the results of petrophysical analysis, well and controlled correlation with seismic horizon picking, produced a pattern of fault trending NE-SW relative to the pattern trending anticline structure relative North-South, where the distributions of Alpha Zone have an average gross thickness of 9-15 meter and average net sandstone is 4-11 meter. Lithofacies approach method also used, which divided into three lithofacies sandstone, shale, and coal, which have sandstone facies distributions NW-SE, where the percentage of the distributions is 20%. The 3D reservoir modeling framework conducted is based on the reservoir depth structure maps generated from picking seismic and controlled by the correlation between wells. Distribution of reservoir properties include facies, Vshale, porosity, and water saturation (Sw), in making the property distribution is based on the analysis of the distribution wells using geostatistical variogram, where the results are validated against production data, so we will get an integrated 3D reservoir modeling of Alpha Zone. Furthermore, based on the volumetric determination of Original Oil in Place (OOIP) in Alpha Zone acquired 2.35 MMbbl.

Steam conformance investigation of flow control devices deployed in SAGD injection and production horizontal wells

The nonuniformity of the steam chamber has a significant influence on oil recovery in the process of Steam-Assisted Gravity Drainage (SAGD). To improve steam conformance, Flow Control Devices (FCDs) are generally installed in both the SAGD injector and producer. In this paper, we investigated the steam conformance of tubing-deployed FCDs in a SAGD well pair. First, a three-phase, three-dimensional reservoir model was developed. Then, a wellbore model for saturated steam flow in the FCD equipped injector and oil/water flow in the FCD equipped producer was proposed. Next, the reservoir and wellbore models were coupled and solved fully implicitly. Finally, after model validations, the effects of the steam injection mode, FCD parameters, steam injection parameters and reservoir geological properties on the conformance of the steam chamber were studied. The results showed that tubing-deployed FCDs were obviously better at improving the steam conformance than the simultaneous dual-tubing string injection technique and injecting steam at only a single point. The steam chamber conformance increased with a decrease in steam injection pressure and reservoir permeability and an increase in steam quality and oil saturation. The steam chamber uniformity gradually declined with time. To maintain the steam conformance over the life of the SAGD and ultimately oil recovery, the FCD open flow area design in SAGD wells should be adjusted at different injection times. When the FCD open flow area was small in the part with high permeability and large in the part with low permeability, a high degree of steam conformance and greater amount of cumulative oil production could be obtained.

Petrophysical Characteristics and Reservoir Modeling of Mishirf Formation at Noor Oil Field, South of Iraq

Petrophysical properties of the Mishrif reservoir at Noor oil field have been done. Based on the interpretation of the open hole data from wells (No-1, 2, 3, 4, and 5) .Which have been calculated total porosity, effective and secondary porosity, water and hydrocarbon saturation (moveable and residual hydrocarbon) in invaded and uninvaded zones. Depends on the calculated of petrophysical properties, Mishrif Formation can be divided into eight reservoir units (RU-1 to 8), separated by eight caped rock units (barrier) (Bar-1 to Bar8). Three-dimensional reservoir model of oil saturation was constructed using the Petrel Software, (2009). Distribution of these petrophysical properties for each reservoir unit within the studied field has been done. The results showed that the best reservoir units are the second, fourth and first reservoir unit. It’s worth mentioned here that the heterogeneity of the thicknesses of these units and its individual direction. In addition, observed that the oil saturation increases towards the north of the field at the well (No-5) and the center of the field at the well (No-4).

An Application of an Embedded Model Estimator to a Synthetic Nonstationary Reservoir Model With Multiple Secondary Variables.

A method (Ember) for nonstationary spatial modeling with multiple secondary variables by combining Geostatistics with Random Forests is applied to a three-dimensional Reservoir Model. It extends the Random Forest method to an interpolation algorithm retaining similar consistency properties to both Geostatistical algorithms and Random Forests. It allows embedding of simpler interpolation algorithms into the process, combining them through the Random Forest training process. The algorithm estimates a conditional distribution at each target location. The family of such distributions is called the model envelope. An algorithm to produce stochastic simulations from the envelope is demonstrated. This algorithm allows the influence of the secondary variables, as well as the variability of the result to vary by location in the simulation.

On the optimum operating temperature for steam floods

The technical, environmental and economic performances of a steam flood are partly influenced by the operating temperature (pressure). However, the definition and procedure for determining the optimum operating temperature are still debatable. Employing a combination of analytic modelling and numerical simulations, this paper investigates the existence (or otherwise) of an optimum injection temperature Topt for saturated-steam floods. Considering the maximization of productivity and thermal efficiency as objective, an analytic procedure, which explores the effects of temperature on injectivity, total steam enthalpy, oil viscosity and relative permeabilities, shows that the operating temperature (pressure) of a steam flood should not exceed 515 K (3.5 MPa). A simple closed-form expression is proposed for Topt as a function of basic rock and fluid properties. For an example three-dimensional reservoir model comprising an 8-m oil shale unit sandwiched between two sandy units each 15 m thick, numerical simulations show sensitivity to temperature (and viscosity effect) in the range 350–450 K, but becomes increasingly insensitive in the band 500–650 K. It is established that ~500–550 K is the optimum band when the optimization objective is to maximize both discounted oil recovery and cumulative oil-steam ratio. These results agree with an optimum injection temperature of ~501 K estimated from the proposed analytical model in this case. Therefore, based on the results of the analytical model, thermal simulations and other considerations, it is concluded that the optimum steam-injection temperature is project and system specific. The insights gained should find relevance in the design and management of steam floods, as well as other steam-based recovery processes.

Application of horizontal wells in three-dimensional shale reservoir modeling: A case study of Longmaxi–Wufeng shale in Fuling gas field, Sichuan Basin: Reply

Three-dimensional reservoir modeling is an important aspect to determine the heterogeneity of organic-rich shale reservoirs, an area of study that continues to be explored and refined. A large proportion of data acquired from horizontal wells causes issues in the structural and property modeling for shale reservoirs. Since horizontal wells are designed to drill into a specific, narrow zone, their horizontal section tends to parallel or nearly parallel formation surfaces. As a result, formation surfaces have a much more complex spatial location relationship with horizontal wellbores than with vertical wellbores. The existing algorithms are not good at addressing this issue during structural modeling. The major problem of using horizontal well data in property modeling is the biased data set because their horizontal section tends to stay within a narrow zone. The property distribution feature estimated from this biased data set, as a significant, default input of geostatistical simulation algorithms, causes the constructed property models to deviate away from the real case in the subsurface. A method to infer more formation tops in pseudovertical wells according to a series of assumptions was developed to provide more constraint points for structural modeling within the areas of the horizontal well section. To use the biased database from horizontal wells, distribution function and trend model methods were developed for continuous property modeling, and percentage and probability trend models were developed for discrete property modeling. The Longmaxi–Wufeng shale in the Fuling gas field of Sichuan Basin was used as an example to express and verify these methods.

Application of horizontal wells in three-dimensional shale reservoir modeling: A case study of Longmaxi–Wufeng shale in Fuling gas field, Sichuan Basin: Discussion

Application of horizontal wells in three-dimensional shale reservoir modeling: A case study of Longmaxi–Wufeng shale in Fuling gas field, Sichuan Basin: Discussion

A Novel Workflow for Optimizing Well Placement Under Geomechanical Constraints: A Case Study

Abstract A novel workflow to optimize well placement using geomechanical constraints is introduced to maximize production performance, reduce excessive simulation runs, and minimize drilling constraints by considering the local stress field and the petrophysical properties in a given reservoir. A case study is presented for optimization of horizontal well placement in the Monterey Formation of Miocene Age in California. First, a three-dimensional reservoir model of formation pressure, in situ stresses, petrophysical and rock properties were built from available petrophysical and well log data. Second, numerical modeling using material point method (MPM) was applied to generate the differential stress field, taking into consideration a three-dimensional natural fracture network in the reservoir model. Third, an optimization algorithm which incorporates petrophysical properties, natural fracture distribution, differential stresses, and mechanical stability was used to identify the best candidate locations for well placement. Finally, flow simulations were conducted to segregate each candidate location where both natural and hydraulic fractures were considered. Statistical methods identify optimal well positions in areas with low differential stress, high porosity, and high permeability. Several candidate locations for well placement were selected and flow simulations were conducted. A comparison of the production performance between the best candidates and other randomly selected well configurations indicates that the workflow can effectively recognize scenarios of optimum well placement. The proposed workflow provides practical insight on well placement optimization by reducing the number of required reservoir simulation runs and maximizing the hydrocarbon recovery.

Development and surrogate-based calibration of a CO2 reservoir model

Simulation-optimization framework is a widely used approach for numerical model calibrations, though its primary difficulty is its high-demand of computational efforts. In this study, Bagging MARS (BMARS) adapted from Multivariate Adaptive Regression Splines (MARS) algorithm, is used to construct the surrogate of a three-dimensional CO2 reservoir model, which is developed to simulate CO2 injection and migration in a fault-compartmentalized underground reservoir. The BMARS surrogate model is then used in model calibration to estimate specified reservoir model input parameters efficiently. The results demonstrate that the BMARS model can improve fitting stability and predictive accuracy against the ordinary MARS model. Parameter sensitivity analysis, which is efficiently conducted using the BMARS model, suggest that permeability of Fault#10 and caprock dominate the pressure buildup in this fault-compartmentalized reservoir. Hence priority should be given to investment in estimating these two reservoir properties. Overpressure propagation and CO2 migration in the reservoir responding to three years of CO2 injection are also analyzed using the calibrated model. The calibrated water-CO2 flow model could be a useful tool to evaluate the future operation and risk assessment of the reservoir. The results comparison and sensitivity analysis demonstrated the proposed BMARS-based simulation–optimization framework is an efficient and accurate model calibration approach.