Day 2 Wed, January 25, 2023最新文献

{"title":"Carbon/Oxygen Spectral Data Processing, its Affiliation to Scintillation Detector Selectivity & their Impact on Reservoir Saturation Monitoring, Lessons Learnt and Recommended Workflow","authors":"Y. Eltaher, S. Ma","doi":"10.2118/212613-ms","DOIUrl":"https://doi.org/10.2118/212613-ms","url":null,"abstract":"\u0000 For decades, it has been affirmed that pulsed neutron (PN) spectral Carbon/Oxygen (C/O) logging is the industry's most robust salinity-independent means for reservoir saturation monitoring (RSM); yet C/O logging still comes with considerable uncertainty that has to be identified and handled with ultimate care. In this paper we investigate two main aspects of such uncertainties and showcase some recommendations to enhance the accuracy of the measurement for improved reservoir saturation monitoring.\u0000 Two fundamental factors affecting C/O measurement are the type of gamma ray (GR) scintillation detector crystals used and the method for C/O spectral data processing. Currently, there are mostly six types of crystals used as GR detectors in commercial PN logging tools for routine operations. Each detector type has its advantages and limitations. With respect to data processing, the most commonly adopted method is the Windows method, due to its simplicity and statistical robustness. Whereas the Yields method is much more complicated to develop and prone to statistical variation, though it tends to provide more accurate results. Similarly, each of these two methods has its own set of advantages and disadvantages.\u0000 A comprehensive study involved different logging instruments and datasets acquired under various logging environments showed that both the physical properties of the detector, as well as the characteristics of the data processing method, have to be fully considered for optimum results. The Windows method, for instance, can be adequate for detectors of statistical nature. Unlike the Yields method, which requires an optimized set of detector and tool specifications. Where for certain GR detectors, significant differences in C/O data and consequently the calculated fluid saturation were observed when processed by using the Windows and the Yields methods. C/O data processing method selection is commonly fit for purpose; yet with the continuous advancement in GR detection technology, standardization is recommended for accurate and precise log measurement.\u0000 Accuracy and precision are keys to C/O logging and consequently successful reservoir surveillance and oil field management. Accordingly, a new standard RSM workflow is recommended where all available elements are properly tailored, to enhance the quality of the answer product.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132045315","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Application of Simulation Opportunity Index and Assisted History Matching for Fishbone Completion Strategy in the Existing Single Horizontal Well in Tight Gas Carbonate Reservoir","authors":"B. Bernadi, H. A. A. Al Shehhi, Sara Saleh Abdulla Abdulla Al Ameri, Fatima Omar Alawadhi, Mariam Ahmed Al Hosani, Ahmed Mohamed Al Bairaq, Maksim Kuzevanov, Konstantin Shelepov","doi":"10.2118/212668-ms","DOIUrl":"https://doi.org/10.2118/212668-ms","url":null,"abstract":"\u0000 The trajectory placement of a horizontal well that crosses a heterogeneous tight gas carbonate reservoir is one of the important elements that contribute to the success of well productivity. Proper placement can yield good production and vice versa improper placement may give low well productivity. An effort was studied to improve the well productivity from an existing single horizontal well that was initially placed in sub optimal location by implementing the Fishbones completion technology (work-over).\u0000 In this paper, Simulation Opportunity Index (SOI) has been selected as a method to indicate the remaining gas sweet spot throughout the reservoir. SOI integrates 3 independent components extracted from static and dynamic parameters; reservoir permeability-thickness, movable gas, and reservoir pressure from a historically-matched dynamic model. By utilizing SOI, a map of the prospective gas sweet spots can be created; hence low performance existing wells are utilized to exploit the surrounding potential sweet spots using Fishbones completion, which consists of tiny short needles with a maximum effective length of 32 feet placed along the horizontal section.\u0000 The study reveals that the Fishbones completion application on existing low-productive horizontal well can multiply the well productivity in multi-layer reservoir environment in addition to the significant production gain. Assisted History Matching (AHM) is used to explore the best scenario of Fishbones features combination, such as the number of Subs required as the container room to hold the needles, the optimum needles length, and the optimum lateral drain section to place the Subs and needles. Many sensitivities with the abovementioned variables are run at once, and the analysis is conducted to identify the most impacting parameter to bring the highest well recovery. The use of SOI method to scrutinize the best location of well candidatures for Fishbone application and the required Subs and needles placement is not only able to rejuvenate the performance of the problematic well, but it can aid in generating CAPEX saving and efficient project schedule to manufacture Fishbones with a proper number of Subs and needles compared to blind Fishbone technology installation.\u0000 A combination of two techniques between reservoir simulation study to generate SOI with the new advanced completion technology called Fishbones applied on low-performance existing horizontal well is the real study integration with the final objective to increase the ultimate gas recovery in a tight gas carbonate reservoir. This is a breakthrough solution to optimize the development of tight carbonate gas fields in addition to the conventional development strategy with normal infill drillings.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114217656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep Learning Models for the Prediction of Mineral Dissolution and Precipitation During Geological Carbon Sequestration","authors":"Zeeshan Tariq, E. U. Yildirim, B. Yan, Shuyu Sun","doi":"10.2118/212597-ms","DOIUrl":"https://doi.org/10.2118/212597-ms","url":null,"abstract":"\u0000 In Geological Carbon Sequestration (GCS), mineralization is a secure carbon dioxide (CO2) trapping mechanism to prevent possible leakage at later stage of the GCS project. Modeling of the mineralization during GCS relies on numerical reservoir simulation, but the computational cost is prohibitively high due to the complex physical processes. Therefore, deep learning (DL) models can be used as a computationally cheaper and at the same time, reliable alternative to the conventional numerical simulators.\u0000 In this work, we have developed a DL approach to effectively predict the dissolution and precipitation of various important minerals, including Anorthite, Kaolinite, and Calcite during CO2 injection into deep saline aquifers. We established a reservoir model to simulate the process of geological CO2 storage. About 750 simulations were performed in order to generate a comprehensive dataset for training DL models. Fourier Neural Operator (FNO) models were trained on the simulated dataset, which take the reservoir and well properties along with time information as input and predict the precipitation and dissolution of minerals in space and time scales. During the training process, root-mean-squared-error (RMSE) was chosen as the loss function to avoid overfitting. To gauge prediction performance, we applied the trained model to predict the concentrations of different mineral on the test dataset, which is 10% of the entire dataset, and two metrics, including the average absolute percentage error (AAPE) and the coefficient of determination (R2) were adopted. The R2 value was found to be around 0.95 for calcite model, 0.94 for Kaolinite model, and 0.93 for Anorthite model. The R2 was calculated for all trainable points from the predictions and ground truth. On the other hand, the average AAPE for all the mappings was calculated around 1%, which demonstrates that the trained model can effectively predict the temporal and spatial evolution of the mineral concentrations. The prediction CPU time (0.2 seconds/case) by the model is much lower than that of the physics-based reservoir simulator (3600 seconds/case). Therefore, the proposed method offers predictions as accurate as our physics-based reservoir simulations, while provides a huge saving of computation time. To the authors' best knowledge, prediction of the precipitation and dissolution of minerals in a supervised learning approach using the simulation data has not been studied before in the literature. The DL models developed in this study can serve as a computationally faster alternative to conventional numerical simulators to assess mineralization trapping in GCS projects especially for the mineral trapping mechanism.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125396143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep Learning-Based and Kernel-Based Proxy Models for Nonlinearly Constrained Life-Cycle Production Optimization","authors":"A. Atadeger, M. Onur, Soham Sheth, R. Banerjee","doi":"10.2118/212690-ms","DOIUrl":"https://doi.org/10.2118/212690-ms","url":null,"abstract":"\u0000 In this study, we investigate the use of deep learning-based and kernel-based proxy models in nonlinearly constrained production optimization and compare their performances with directly using the high-fidelity simulators (HFS) for such optimization in terms of computational cost and optimal results obtained. One of the proxy models is embed to control and observe (E2CO), a deep learning-based model, and the other model is a kernel-based proxy, least-squares support-vector regression (LS-SVR). Both proxies have the capability of predicting well outputs. The sequential quadratic programming (SQP) method is used to perform nonlinearly constrained production optimization. The objective function considered here is the net present value (NPV), and the nonlinear state constraints are field liquid production rate (FLPR) and field water production rate (FWPR). NPV, FLPR, and FWPR are constructed by using two different types of proxy models. The gradient of the objective function as well as the Jacobian matrix of constraints are computed analytically for the LS-SVR, whereas the method of stochastic simplex approximated gradient (StoSAG) is used for optimization with E2CO and HFS. The reservoir model considered in this study is a two-phase, three-dimensional reservoir with heterogeneous permeability which is taken from the SPE10 benchmark case. Well controls are optimized to maximize the NPV in an oil-water waterflooding scenario. It is observed that all proxy models can find optimal NPV results like optimal NPV obtained by HFS with much less computational effort. Among proxy models, LS-SVR is found to be less computationally demanding in the training process. Overall, both proxy models are orders of magnitude faster than numerical models in the prediction. We provide new insights into the accuracy and prediction performances of these machine learning-based proxy models for 3D oil-water systems as well as their efficiency in nonlinearly constrained production optimization for waterflooding applications.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"77 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130791013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maximizing Value from Geosteering Efficiency by Integrating Real-Time Petrophysical Analysis","authors":"A. Kundu, T. Raza, Lichuan Deng, A. Soliman, Eslam Elabsy, Sarah Zemiti, Alyazia Alhammadi","doi":"10.2118/212684-ms","DOIUrl":"https://doi.org/10.2118/212684-ms","url":null,"abstract":"\u0000 Conventional geo-steering approach use raw logging measurements to define wellbore positioning within the reservoir while drilling. The geo-steering specialist usually compares real-time logs to modelled logs (GR/Density/Neutron/Resistivity) and the geological model is then adjusted to make real-time decisions to deliver the well objectives. This conventional method is applicable to most reservoir conditions. However, it may be insufficient or inappropriate in heterogeneous reservoirs or wells with complex geological settings, potentially resulting in wells being sub-optimally placed and reducing the value of reservoir sections in terms of productivity. This paper aims to showcase a Petrophysics-based Geo-steering approach to maximize the value of reservoir sections.\u0000 Geo-steering aims to place the well trajectory in the lithology with optimum storage capacity, flow capacity and hydrocarbon saturation. The method of log-to-log comparison is popular for its simplicity and speed of use in real-time but is not enough for certain scenarios. For example, the real-time log response can be very different from modelled log response in the presence of gas or very light oil, irrespective of petrophysical properties (porosity/permeability) being similar. Moreover, real-time Sw estimation would be required in addition to porosity to minimize the risk of drilling a producer into water bearing intervals. In fact, the comparison between petrophysical parameters is more appropriate to heterogeneous reservoirs or wells with complicated geology. This approach requires good co-ordination between geologist, petrophysicist and geo-steering specialist. Prior to drilling, the petrophysical model from offset wells should be defined and used to derive porosity, permeability and saturation. While drilling, the petrophysical properties are then interpreted in real-time and based on the comparison between modelled and real-time petrophysical properties, decisions are to be made with respect to the well objectives.\u0000 An example with strong gas effect in a carbonate reservoir from Abu Dhabi is presented to demonstrate this novel approach. Real-time density/neutron does not have good correlation with modelled density /neutron due to gas effect. Such poor correlation can be attributed to proximity to a Gas Oil Contact (GOC) and dynamic invasion, complicating the real-time geo-steering. However, real-time total porosity from log analysis correlates very well with modelled total porosity, providing confidence in wellbore positioning and allowing the geologist and the geo-steering specialist to make the correct real-time decision to place the well in the optimum stratigraphic position in order to meet the well objectives.\u0000 Only conventional logs are utilized in this case, but if real-time NMR and resistivity image interpretation are available, it will provide additional information in term of permeability, secondary porosity and irreducible water saturation to aid efficient geo-s","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123886664","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating Pulsar Technology with Reservoir Centric Fracturing Approach to Restore the Production in a Mature Tight Oil Field","authors":"K. Slimani, Tayeb Khetib, H. Monfared, Rabah Lamali, Nasrine Bendali Amor, Mohamed Semch-Eddine Skender, Farid Mezali, Hakim Chekou, L. Belaifa","doi":"10.2118/212585-ms","DOIUrl":"https://doi.org/10.2118/212585-ms","url":null,"abstract":"\u0000 Tight oil reservoirs are of paramount importance for an operator holding several fields with important oil potential in Illizi basin, southeastern Algeria. Formation tightness, the presence of nonconnected sand lenses, and the lack of a geological model made it very difficult to maintain the oil production in the studied Devonian reservoir. Consequently, the service company and operator adopted an integrated approach to devise solutions that could restore production in a mature tight oil field that had been closed in 2011.\u0000 Construction of a geological model for the studied reservoir was challenging because of the high uncertainty in water saturation interpretation caused by the formation water properties (fresh water). A multifunction pulsed neutron service was proposed to provide standalone cased-hole formation evaluation and reservoir saturation monitoring. A unique modeling approaches was applied to characterize the studied tight reservoir and evaluate the reserves based on advanced uncertainty analysis. A hydraulic fracturing design workflow in a reservoir centric stimulation to production software was developed using an integrated approach (geological and geomechanical models) to place the fracture in the optimum reservoir quality and connect the sand lens bodies.\u0000 Two existing wells were selected to run with the pulsed neutron service, resulting in acquisition of comprehensive reservoir rock and fluid content data. The interpreted logs served to reduce the uncertainty in water saturation modeling and to enable perfect history matching of the producing wells. The constructed geological model was the basis for improving the stimulation designs and maximizing production for future wells. With significant oil initially in place (STOIIP) estimated from the model, the field showed more promise than the previous recorded recovery factor of less than 1%. The field development plan (FDP) identified the location of 20 new infill drilling wells targeting the sweet spots and considering the optimal well spacing. In addition, the plan specified a systematic hydraulic fracturing stimulation job for each newly drilled well to connect and produce the sand body lenses. Recently, a successful campaign of hydraulic fracturing operations was executed on four wells, allowing the operator to resume production from the field. The fracturing performance minimized water cut despite the water-oil-contact (WOC) proximity, and it enhanced the oil recovery.\u0000 The developed integrated approach has already shown its effectiveness in returning a field to production and improving oil recovery. The approach can be replicated on subsequent wells in the field as well as on similar tight reservoirs all over the world.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129026087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Rock Physics Modelling and Stochastic Seismic Inversion to Predict Reservoir Properties and Quantify Uncertainties of a Complex Upper Jurassic Carbonate Reservoir From Offshore Abu Dhabi","authors":"M. Waqas, Lian Hou, J. Ahmed, Santan Kumar, S. Chatterjee, N. Vargas, Julio Tavares, L. Michou, Franciscus van Kleef, A. S. Alkaabi, O. Colnard, Bertrand Six","doi":"10.2118/212672-ms","DOIUrl":"https://doi.org/10.2118/212672-ms","url":null,"abstract":"\u0000 The need to understand field-scale reservoir heterogeneity using seismic data requires implementing advanced solutions such as stochastic seismic inversion to go beyond the resolution of seismic data. Conventional seismic inversion techniques provide relatively low-resolution reservoir properties but do not provide quantitative estimates of the subsurface uncertainties. The objective of this study was to carry out a facies dependent geostatistical seismic inversion to generate multi-realization reservoir properties to improve the geological understanding of the two adjacent offshore fields in Abu Dhabi.\u0000 An integrated approach of rock physics modelling and geostatistical inversion followed by porosity co-simulation was undertaken to characterize the spatially varying lithofacies and porosity of the complex carbonate reservoirs. Necessary checks to ensure highest quality data input included: 1) Rock physics modelling and shear sonic prediction 2) Invasion correction and production effect correction of elastic logs 3) Seismic feasibility analysis to define seismic facies and 4) Six angle stacks optimally defined to preserve AVO/AVA signature followed by AVO/AVA compliant post-stack processing. Subsequently, the joint facies driven geostatistical inversion was conducted to invert for multiple realizations high-resolution lithofacies and elastic rock properties. Finally, porosity was co-simulated and later ranked to map important geological variations.\u0000 Based on the rock physics analysis, a 4 facies classification scheme (Porous Calcite, Porous Dolomite, Tight Calcite-Dolomite and Anhydrite) was adopted and used as input in the joint facies-elastic inversion. Before the geostatistical inversion, a deterministic inversion was performed that helped in refining the horizon interpretation of the surfaces used as a framework for the inversion. In geostatistical inversion, results are guided by variograms, facies, prior probability density functions, wells, inversion grid and seismic data quality. At start of the joint inversion, the parameters for inversion are defined in an unconstrained fashion aiming to obtain unbiased parameters which are blind to well control. Finally, using elastic properties constrained at the well locations, the joint geostatistical inversion was run to obtain multiple realizations of P-impedance, S-impedance, density and lithofacies. The cross-correlation between seismic and inverted synthetics was high across the whole area for all the partial angle stacks, with the lowest cross-correlation observed in the far angle stack. Lithofacies and elastic properties were used to co-simulate for porosity. The porosity results were then ranked to provide the P10, P50 and P90 models to be used for reservoir property model building.\u0000 This study is an example of stochastically generating geologically consistent reservoir properties through high-resolution seismically constrained inversion results at 1ms vertical sampling. Lithofacies and el","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129029032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integration of Seismic Poisson Impedance Data with Real-Time Geomechanics and Real-Time 3D Ultra-Deep Resistivity Inversion Enabled New Opportunities in Developed Field","authors":"Anar Abdulkarim, Alexander Kharitonov, T. E. El Gezeery, Mohamed Al Haddad, Y. Halawah, S. A. Sabea","doi":"10.2118/212656-ms","DOIUrl":"https://doi.org/10.2118/212656-ms","url":null,"abstract":"\u0000 The Wara sandstone reservoir in the Minagish field of Kuwait Oil Company is a complex deposition of a typical pro-deltaic environment consisting of shaly-silty sandstone sequences W7-W1. Three sequences (W6, W5, and W3) were expected in the case study well. The objective was to set 9⅝-in. casing at the top of W6 and then drill through the Wara sequences to connect all of them and land and drill the lateral section within W3.\u0000 The W6 sequence is typically the primary target in the Wara formation, being thick and consistent throughout the field. The next logical step in developing the Wara reservoir was to study and investigate the minor W5 and W3 members. Due to poor correlation of W5 and W3 channels in offset wells, the geological target was selected based on seismic Poisson impedance. Historically, targeting the Wara formation occasionally resulted in multiple sidetracks due to drilling challenges. A real-time geomechanics service was utilized to overcome drilling challenges and real-time 3D ultra-deep resistivity inversion was implemented to optimize well placement.\u0000 An extensive pre-drilling study for geomechanical and ultra-deep resistivity inversion modelling helped to develop road map for an optimal and safe well-construction process. The study showed that utilization of real-time 3D ultra-deep resistivity (UDR) inversion would help to optimize well placement and maximize sweet-zone exposure. The original well design, mud properties, and drilling parameters were modified based on the geomechanical study. Additionally, real-time geomechanics services were utilized to monitor and control the drilling process to follow the road map, which helped to avoid drilling issues, geostop at the W6 channel, and finally to run the casing smoothly. Real-time 3D ultra-deep resistivity mapping in the lateral section helped the operator to drill through W6 and W5, land precisely, and drill the lateral in the W3 channel, which was well developed, as expected from seismic Poisson impedance analysis. Formation evaluation of lateral section showed an average porosity of 24 p.u., water saturation 11% and up to 3 D/cp mobility.\u0000 The application of real-time 3D ultra-deep resistivity inversion helped to triple the planned formation exposure and to discover a geometric extension of the above deposited channels (W6 and W5), which will help for future field development. The flow test showed the highest production rates from W3 of the field. The integrated approach described above was recommended to be utilized for all future Wara wells.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122026233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An Efficient Simulation Approach for Long-term Assessment of CO2 Storage in Complex Geological Formations","authors":"Ziliang Zhang, Yuhang Wang, C. Vuik, H. Hajibeygi","doi":"10.2118/212635-ms","DOIUrl":"https://doi.org/10.2118/212635-ms","url":null,"abstract":"\u0000 We present an efficient compositional framework for simulation of CO2 storage in saline aquifers with complex geological geometries during a lifelong injection and migration process. To improve the computation efficiency, the general framework considers the essential hydrodynamic physics, including hysteresis, dissolution and capillarity, by means of parameterized space. The parameterization method translates physical models into parameterized spaces during an offline stage before simulation starts. Among them, the hysteresis behavior of constitutive relations is captured by the surfaces created from bounding and scanning curves, on which relative permeability and capillarity pressure are determined directly with a pair of saturation and turning point values. The new development also allows for simulation of realistic reservoir models with complex geological features. The numerical framework is validated by comparing simulation results obtained from the Cartesian-box and the converted corner-point grids of the same geometry, and it is applied to a field-scale reservoir eventually. For the benchmark problem, the CO2 is injected into a layered formation. Key processes such as accumulation of CO2 under capillarity barriers, gas breakthrough and dissolution, are well captured and agree with the results reported in literature. The roles of various physical effects and their interactions in CO2 trapping are investigated in a realistic reservoir model using the corner-point grid. It is found that dissolution of CO2 in brine occurs when CO2 and brine are in contact. The effect of residual saturation and hysteresis behavior can be captured by the proposed scanning curve surface in a robust way. The existence of capillarity causes less sharp CO2-brine interfaces by enhancing the imbibition of the brine behind the CO2 plume, which also increases the residual trapping. Moreover, the time-dependent characteristics of the trapping amount reveals the different time scales on which various trapping mechanisms (dissolution and residual) operate and the interplay. The novelty of the development is that essential physics for CO2 trapping are considered by the means of parameterized space. As it is implemented on corner-point grid geometries, it casts a promising approach to predict the migration of CO2 plume, and to assess the amount of CO2 trapped by different trapping mechanisms in realistic field-scale reservoirs.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125769742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Comparative Study for Deep-Learning-Based Methods for Automated Reservoir Simulation","authors":"Alaa Maarouf, S. Tahir, Shi Su, Chakib Kada Kloucha, Hussein Mustapha","doi":"10.2118/212594-ms","DOIUrl":"https://doi.org/10.2118/212594-ms","url":null,"abstract":"\u0000 Reservoir simulation is essential for various reservoir engineering processes such as history matching and field development plan optimization but is typically an intensive and time-consuming process. The aim of this study is to compare various deep-learning algorithms for constructing a machine-learning (ML) proxy model, which reproduces the behavior of a reservoir simulator and results in significant speedup compared to running the numerical simulator.\u0000 Initially, we generate an ensemble of realizations via the reservoir simulator to train the different ML algorithms. The data set consists of a comprehensive set of uncertainty parameters and the corresponding simulation data across all wells. The system utilizes recent advances in deep learning based on deep neural networks, convolutional neural networks, and autoencoders to create machine-learning-based proxy models that predict production and injection profiles as well as the bottomhole pressure of all wells. Thus, the proposed workflows replace the time-consuming simulation process with fast and efficient proxy models.\u0000 In this work we provide a comparative study of various ML-based algorithms utilizing deep neural networks and convolutional neural networks for constructing a surrogate reservoir model. The trained models can simulate the behavior of the physics-based reservoir simulator by correlating uncertainty parameters to various history-matched reservoir properties. The algorithms were tested on a mature oilfield with a notable number of wells and several decades of production and injection data. We analyze the performance of each ML approach and provide recommendations on the optimal one.\u0000 The best performing workflow for building the ML proxy model consists of two steps. The first step uses stacked autoencoders to learn a low-dimensional latent space representation of the highly dimensional simulation data. This step allows to reduce the complexity of predicting the simulation data and enhances the prediction quality. The following step constructs an ML model to predict the latent space features from input uncertainty parameters and produces highly accurate results.\u0000 Reservoir simulation is of paramount importance for various reservoir engineering workflows. Traditional approaches require running physics-based simulators for multiple iterations, which results in time-consuming and labor-intensive processes. We implement and compare several deep-learning-based methods to construct ML proxy models that automate and remarkably reduce the runtime of the reservoir simulation process.","PeriodicalId":215106,"journal":{"name":"Day 2 Wed, January 25, 2023","volume":"176 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132070185","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}