Day 2 Tue, October 27, 2020最新文献

{"title":"3D Reservoir Model History Matching Based on Machine Learning Technology","authors":"E. Illarionov, Pavel Temirchev, D. Voloskov, A. Gubanova, D. Koroteev, M. Simonov, A. Akhmetov, A. Margarit","doi":"10.2118/201924-ms","DOIUrl":"https://doi.org/10.2118/201924-ms","url":null,"abstract":"\u0000 In adaptation of reservoir models a direct gradient backpropagation through the forward model is often intractable or requires enormous computational costs. Thus one have to construct separate models that simulate them implicitly, e.g. via stochastic sampling or solving of adjoint systems. We demonstrate that if the forward model is a neural network, gradient backpropagation becomes naturally involved both in model training and adaptation. In our research we compare 3 adaptation strategies: variation of reservoir model variables, neural network adaptation and latent space adaptation and discuss to what extent they preserve the geological content. We exploit a real-world reservoir model to investigate the problem in practical case. The numerical experiments demonstrate that the latent space adaptation provides the most stable and accurate results.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127143516","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":"Evaluation of a Field-Wide Post-Steam In-Situ Combustion Performance in a Heavy Oil Reservoir in China","authors":"Fang Zhao, Changfeng Xi, Xialin Zhang, Xiaorong Shi, Fengxiang Yang, Hetaer Mu, Wenlong Guan, Youwei Jiang, Hongzhuang Wang, T. Babadagli, H. Li","doi":"10.2118/201815-ms","DOIUrl":"https://doi.org/10.2118/201815-ms","url":null,"abstract":"\u0000 We evaluated the performance of a field-wide post-steam in-situ combustion (ISC) project conducted in a complex heavy oil reservoir in China using laboratory one-dimensional combustion experiments, reservoir simulation outputs, and data collected from the field application. The commercial ISC project showed vastly different production performances in different regions of the field and two types of representative well groups were identified. Type I group has a low oil viscosity (<8000 mPa.s) and a high steam-flooded recovery factor (>30%); after ISC treatment, these producers show a high initial water cut, while some experience channeling issues and hence produce a large quantity of flue gas. Type II group has a high oil viscosity (>20000 mPa.s) and a low cyclic steam stimulation (CSS) recovery factor (15-20%); these producers have a high air injection pressure exceeding the fracture pressure. Corresponding remedial methods were designed and applied to these two well groups. Presently, the evaluation methods described in this paper are being applied in the field, and initial results have been acquired.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122662666","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":"Automated Geosteering While Drilling Using Machine Learning. Case Studies","authors":"I. Denisenko, I. Kuvaev, I. Uvarov, Oleg Evgenievich Kushmantzev, Artem Igorevich Toporov","doi":"10.2118/202046-ms","DOIUrl":"https://doi.org/10.2118/202046-ms","url":null,"abstract":"\u0000 Today's oil & gas industry faces a number of different challenges. Drilling activities are ramping up due to an increase in hydrocarbon demand combined with a reduction of easy-to-recover reserves. Horizontal drilling is growing and has become an integral part of field development. The geology is becoming more and more complex requiring drilling through dense layers targeting thin-layered reservoirs with lateral changes and anisotropy. In recent years, companies have been looking at the ways of optimizing drilling costs by increasing efficiency and process automation. This has been a driver for many companies to stay profitable and efficient in the market.\u0000 One of the areas of interest for process automation has been a geosteering. Geosteering is the real-time adjustment well trajectory while drilling to maximize effective footage in the target zone. In this paper, innovative new approaches to automation of the geosteering process will be discussed. This approach has been successfully tested and deployed in several leading O&G companies.\u0000 The main objective of automated geosteering is to optimize horizontal well placement while freeing up time operational geologists had spent doing routine work in order to focus on complex and more intense tasks as well as the reduction of operational errors related to human factors. This paper will provide details on several automated geosteering algorithms. They have been tested successfully on large numbers of wells. The results of automated geosteering were as close as 90% to the manual interpretations done by geologists. When the results diverged, the geologists often \"agreed\" with the interpretation proposed by the algorithm.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121932950","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 New Data Analytics Based Method to Characterize Waterflood Strategy in Geologically Challenging Mature Oil Field","authors":"A. Yadav, D. Davudov, Y. Danişman, A. Malkov, E. Omara, A. Venkatraman, A. El-Hawari","doi":"10.2118/201929-ms","DOIUrl":"https://doi.org/10.2118/201929-ms","url":null,"abstract":"\u0000 The uncertainties associated with oil and gas field reduces with time. When oil fields mature, there is a potential to better understand the field due to the availability of historic production and injection data. In this research, a novel approach is presented which uses data analytics techniques to optimize waterflooding in a Gulf of Suez field. A combination of qualitative and quantitative techniques has been applied to develop a new workflow for analyzing and optimizing waterflood.\u0000 The presented technique involves combining qualitative analysis (random forest) and quantitative analysis (capacitance resistance model, CRM) to obtain a waterflood strategy for the producing field. The Random forest algorithm (machine learning technique) is used to compare two time series signals – production data and injection data from producer/injector wells. The data from each injector and surrounding producers are used for random forest analysis to identify the most effective and ineffective injector-producer pairs. Next, the qualitative analysis using the capacitance resistance model (CRM) is used to determine gain values between each injector-producer pair and to also obtain new injection rates for increasing oil recovery. Results obtained from the random forest model helps reduce the number of unknowns and further validate results in CRM.\u0000 The production and injection data reveal the most effective and ineffective injector-producer pairs that are the result of changes occurring in the reservoir during waterflood. Accordingly, the use of data analytics technique of random forest analysis and CRM on production injection data helps improve reservoir characterization. This combined analysis for the presented field uniquely helps identify effective and ineffective injector-producer pairs to determine the efficiency of waterflooding. The results from this novel analytical technique are presented for the Gulf of Suez field. These results compare well with the streamline approach presented for the same Gulf of Suez field.\u0000 In summary, a new method for reservoir surveillance using data analytics technique of random forest in combination with the capacitance resistance model is presented. The novel combination of the qualitative and quantitative methods presented also helps adapt the specific characteristic of this field – the presence of water drive (pseudo injector). The modeling of water drive as an additional injector (pseudo injector) improves the gain coefficient obtained from the CRM. The comparison with streamlines helps benchmark the model results especially in cases where such secondary data is not available. The model presented can be adapted to similar mature fields under waterfloods. This new approach can be used to optimize water injection more frequently using operations data being gathered for implementing digitization strategies for oil and gas companies.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134565109","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":"Wait or Get the Oil: How SAGD Technology Implementation Options will Vary Future Production","authors":"A. A. Terentiyev, Pavel Valeryevich Roschin, A. V. Nikitin, V. N. Kozhin, K. Pchela, I. I. Kireyev, Sergei Valerevich Demin, A. T. Litvin, I. Struchkov","doi":"10.2118/201819-ms","DOIUrl":"https://doi.org/10.2118/201819-ms","url":null,"abstract":"\u0000 \u0000 \u0000 One of the complications at the stage with steam injection start into the SAGD injection well in the Terrigenous reservoir with extra-heavy oil (EHO) is its injectivity rate. Traditionally, preheating the well bottom-hole (BH) zone with steam and subsequent recovery of hot water or steam through the annulus is used to get adequate injectivity. As an alternative to steam preheating it is proposed to inject an aromatic solvent/reagent to ensure sufficient well injectivity. Calculations were performed with the real reservoir model. The mutual influence of wells in SAGD blocks under the conditions of solvent/reagent injection was studied for this.\u0000","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130183186","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 Petrophysical Log Data with Computational Intelligence for the Development of a Lithology Predictor","authors":"Syed M Amir, Mohammad Rasheed Khan, Ekarit Panacharoensawad, Serhii Kryvenko","doi":"10.2118/202047-ms","DOIUrl":"https://doi.org/10.2118/202047-ms","url":null,"abstract":"\u0000 Wrong manual interpretation from the log data about the formation type and other important information can be catastrophic for the company-operator. With Machine-Learning (ML) (a branch of Artificial Intelligence) algorithms, the interpretation of formation type from the log data has been addressed. As a result, we have successfully developed a program able to accurately predict the type of formation.\u0000 Using the conventional Machine Learning technique of splitting the data into training, validation and test sets, we tried six different ML algorithms to fit with the training part of the data and then verify their prediction accuracy with cross-validation scores and cross-validation predictions which tests the performance of the classifiers (ML algorithms) on the validation set. The three best performing classifiers were selected and further improved by a search of classifier's best hyperparameters. These improved classifiers are further tested on unseen data to produce a comparative analysis.\u0000 Our prediction accuracy with Receiver Operating Characteristic (ROC) scores and ROC-Area Under-the-Curve (ROC-AUC) for each type of formation from the log data lies in the range of 95-99%, except for formations such as shaly sandstone and shale (50% and 84% respectively). The reason for this seemed to be under-fitting i.e., during the training, the classifiers did not see enough instances of these types of formation to know exactly what characteristics of the data make the type of formation to be shaly sandstone or shale. The issue of under-fitting was verified by skimming through the data. To resolve this problem, we suggest training classifiers with a larger data with more targets (types of formation). Furthermore, during the data cleaning (prior to classifier training) and data analysis phases we have discovered important relationships between well logs and defined relative importance of each well log for different formations. This observation can be investigated further to help eliminate the use of multiple well logs while dealing with some formations (based on prior geological knowledge) and reduce the cost of the well logging operations. Using our program with a larger well log data consisting of more formation type instances, we can train the classifiers to accurately predict the formation type irrespectively of differences in formation type.\u0000 Our program is dynamic in the sense that with different targets, i.e., type of formation fluid instead of type of formation or both together, it can successfully predict either or both targets. Increasing the numbers of data instances resulted in a better training and thus, more accurate predictions. Utilization of the program will make the formation-evaluation process easier, faster, automated and more-precise.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128563615","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":"Broaden Limits: Managed Pressure Drilling – A New Step for Achimov Horizontal Wells","authors":"Maxim Serggevich Zlobin, Sergey Vasilyevich Pilnyk, Y. S. Kolesnikov, Daniil Yurievich Kartinen, D. Krivolapov, P. Dobrokhleb, I. Masalida, A. Magda, Artem Andriyanovich Gerasimov, T. Soroka, I. Moiseenko, D. Andreev","doi":"10.2118/201866-ms","DOIUrl":"https://doi.org/10.2118/201866-ms","url":null,"abstract":"\u0000 The rapid development of the Achimov drilling program on the Yamal Peninsula requires an increasing of the drilling efficiency and complicating the wells construction. These factors create new challenges. The pursuit of increasing the productivity of the production wells leads to an increase in the length of horizontal reservoir sections and the number of hydraulic fracturing stages. Since the beginning of the complicated horizontal wells’ drilling at the Urengoyskoye oil and gas field, the average length of the horizontal wellbores has increased from 1000 to 1500 m, and the number of hydraulic fracturing – from 3 to 7 stages.\u0000 Further extension of the horizontal well interval is associated with an increase in the bottomhole pressure and the risk of hydraulic fracturing. A pilot project for the horizontal wells’ construction in the Achimov deposits was launched at the Yamburgskoye field in 2019. Two such wells with the 1000 m horizontal interval were successfully drilled and completed. The results of the calculations and the obtained experience indicate a risk of the further horizontal interval increase. The implementation of managed pressure drilling (MPD) technology through the application of a low-density drilling mud and the precise control of the wellhead pressure can reduce the magnitude of repression during circulation and provide with an opportunity to drill longer intervals.\u0000 Applying MPD technology makes it possible to provide safe drilling operations, tripping, early kick detection and fluid losses, running and cementing liners. The next step in the Yamburgskoye field pilot project was the construction of a well with a 6500 m depth and the 2500 m horizontal interval in the Achimov strata. Calculations results demonstrated that with the application of conventional drilling technologies there is a high probability of hydraulic fracturing if the horizontal section is more than 1500 meters. In addition, to make possible more fracturing stages execution, the use of Plug & Perf technology is required, which demands the liner cementing. The liner cementing, in turn, is most likely impossible with the traditional approach. It was decided to drill an extended productive strata interval with the MPD technology application. It was required to optimize each element of the drilling system, including BHA, drilling muds, and drilling regimes to achieve the best results. The necessary regulations and procedures have also been developed which purpose was not only to provide more safety during drilling operations, but also to reduce the time of well construction.\u0000 MPD has become a new stage in the development of the Achimov program and made it possible to remove several geological restrictions. It also allowed to create conditions for the construction of more complex and efficient wells with a longer productive interval and many hydraulic fracturing stages. The experience obtained allows us to talk about the prospects of this solution for the construc","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128583025","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":"Potential and Possible Technological Solutions for Field Development of Unconventional Reservoirs: Bazhenov Formation","authors":"D. Bukharov, Y. Alekseev, A. Prodan, A. Nenko","doi":"10.2118/201818-ms","DOIUrl":"https://doi.org/10.2118/201818-ms","url":null,"abstract":"\u0000 The \"shale revolution\" experience on the example of the Bazhenov formation (BF) development required a serious development program review and implementation.\u0000 The modern approach to the Bazhenov formation development in the Russian Federation is approaching the final stage of the existing formation stimulation technologies improvement. The technical limit in completion is about to be reached. It becomes obvious that existing stimulation technologies do not allow creating more efficient \"reservoir factories\" in the Bazhenov formation. The paper's objective is an attempt to identify hypotheses and prerequisites, as well as possible prospects for the stimulation technologies development, to justify the use of new alternative systems for hydraulic fracturing (HF) with an assessment of their potential at the BF reservoirs.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115965544","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":"Machine Learning for Fracture Parameter Estimation in Fractured Reservoirs from Seismic Data","authors":"G. Sabinin, T. Chichinina, V. Tulchinsky, M. Romero-Salcedo","doi":"10.2118/201934-ms","DOIUrl":"https://doi.org/10.2118/201934-ms","url":null,"abstract":"\u0000 Nowadays, Machine Learning (ML) is actively used in geophysical prospecting including seismic exploration. This study focuses on the applicability and feasibility of Deep Learning for the inverse problem in seismic exploration that is the estimation of the rock-physics parameters for a fractured reservoir, from seismic data. The main goal of this paper is to prove the efficiency of a neural network in estimating fractured medium parameters, represented as anisotropy parameters of HTI model. (HTI is \"Horizontal Transverse Isotropy\".) As such fracture parameters, we consider the normal and tangential weaknesses of fractures ΔN and ΔT, Thomsen anisotropy parameters ε, δ, γ, as well as the crack density e and the crack aspect ratio α (fracture opening). In addition, we consider a fractured medium, in which there are two fracture networks, characterized by two pairs of weaknesses (ΔN1, ΔT1) and (ΔN2, ΔT2); this is the so-called orthorhombic model.\u0000 We validate the accuracy of our neural network by comparing the predicted parameter values with the a priori given. We use mathematic formulae, which relate the considered parameters estimation to different effective-medium anisotropy models of a fractured medium, such as Schoenberg's Linear Slip model, Hudson's model for penny-shaped cracks and Thomsen's model for aligned cracks in porous rock.\u0000 In our study, seismic signatures (seismograms of the reflected waves PP and PS) of both the vertical UZ-component and the horizontal one UX are the inputs for the neural network. At the output, the network predicts fracture parameters and anisotropy parameters. The neural network is trained on synthetic seismograms of reflected waves, which were generated using 2D-elastic numerical finite-difference modelling.\u0000 Thus we demonstrate the applicability of Deep Learning for estimation of the fractured medium parameters, by training the neural network on synthetic seismograms. The normal and tangential weaknesses of fractures ΔN and ΔT, the crack density and the crack aspect ratio (crack opening) are successfully estimated as well as the anisotropy parameters ε(V), δ(V) and γ(V). In the prediction of ΔN and ΔT, the relative error does not exceed 1.7% and 1.4%, respectively, and in the prediction of crack density e — from 0.9% to 1.4%. In predicting the anisotropy parameters ε(V), δ(V) and γ(V), the error does not exceed 1.6%, 1.7%, and 1.8%, respectively. However, in estimating the value of crack opening α, the result is an order of magnitude worse, an error of 14.2%.\u0000 For the orthorhombic model, the prediction results are slightly worse than for the HTI model, but still within the acceptable accuracy. In predicting the fracture parameters for the first fracture network (ΔN1, ΔT1 and e1) the error does not exceed 2.3%, 4.2%, and 2.3%, respectively, and for the second fracture network (ΔN2, ΔT2 and e1) — respectively 4.3%, 5.7%, and 3.7%. This slight deterioration in the results (in comparison with HTI) is explained by ","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129357757","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 Oscillation Rheology Method to Studying Fracturing Fluids","authors":"Tsimur Donalovich Hiliazitdzinau, Andrei Mikhailovich Valenkov, M. V. Kazak, S. Panin","doi":"10.2118/202063-ms","DOIUrl":"https://doi.org/10.2118/202063-ms","url":null,"abstract":"\u0000 In this paper, rheological properties of fracturing fluid samples on polymer and non-polymer basis are studied. The interdependence between effective viscosity, viscoelastic properties and the proppant carrying capacity of the studied composite systems is shown. The advantage of using the method of oscillatory rheometry in the amplitude sweep mode when predicting the sedimentation rate of proppant is observed. The possibility of using this method to study viscoelastic properties of fracturing fluids,both on the basis of classical and alternative gelling agents, was experimentally confirmed.","PeriodicalId":359083,"journal":{"name":"Day 2 Tue, October 27, 2020","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114275416","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}