- Home
- Conferences
- Conference Proceedings
- Conferences
82nd EAGE Annual Conference & Exhibition
- Conference date: October 18-21, 2021
- Location: Amsterdam, The Netherlands
- Published: 18 October 2021
1121 - 1137 of 1137 results
-
-
Correction of Terminal Saturations for Capillary End Effects
More LessSummaryOil displacement experiments constitute a major component of special core analyses for both improved and enhanced oil recovery (IOR and EOR) operations. Capillary end effects (CEE) is a well-recognized phenomena that affects the accuracy of unsteady-state (USS) oil-displacement results. Huang and Honarpour presented the procedure of correcting for CEE in 1998. They derived the equations and demonstrated the procedure for correcting water terminal saturation in drainage of water-wet media. However, since their pioneering work, their methodology has not been extended to correction of oil terminal saturation for imbibition in oil-wet media—despite the well-recognized and impact of CEE on oil-displacement studies. Therefore, in this work, we revisit Huang and Honarpour (HH) pioneering method and extend it to the correction of CEE of terminal saturations for imbibition in oil-wet media. The extended method was also applied and validated using special core analysis data for a Middle Eastern carbonate. In the process, we present the base and extended methods in a concise and clear fashion that is of utility to the practicing reservoir engineer and petrophysicist.
-
-
-
Deep learning-based dealiasing for estimated surface-related multiples from limited sources
Authors D. Zhang and E. VerschuurSummaryThe main prediction engine in surface-related multiple elimination (SRME) is the multidimensional convolution process, where data sampling plays an essential role for accurate surface multiple prediction. Therefore, fully sampled sources and receivers are preferred. If especially the source sampling is far from ideal, the estimated multiples will suffer from the severe aliasing effect. Consequently, this can lead to poorly estimated primaries. Interpolation of coarsely sampled sources is not a trivial task and computation intensive. Dealiasing on the estimated multiples from limited sources might provide a potential solution. In theory, this dealiasing problem is highly non-linear, which suits well for deep learning (DL)-based methods. Therefore, we propose a U-Net-based approach to dealiase the estimated surface multiples from limited sources. Applications on two subsets of the field data demonstrate the effective performance of the proposed method.
-
-
-
Multi-method pore space analysis of the Portland sandstone
Authors B. Mehalli and K. TitovSummaryIn this study, we investigate the pore space of a sample of the Portland sandstone using various methods with different resolution, and based on different physical principles (Mercury injection capillary pressure, MICP, Micro-computed tomography, µ-CT, and Spectral induced polarization, SIP). According to their physical principles, each method provides either the pore body size (µ-CT) or the pore throat size (MICP and SIP). Moreover MICP characterizes the pore-throat radii, whereas SIP rather gives the pore-throat lengths. The objective of this work was to compare results provided by each technique in a common parameter framework.
The comparison approach between the different techniques is based on the relationships of ‘the incremental porosity – pore size’. Our results show that the recovered porosity can increase when using a method with higher resolution. We believe each method gives a specific attribute of the pore space topology similar to reflector attributes obtained in exploration seismology. We also believe that extensive works must be done to improve our understanding of these attributes to better characterize the pores space topology, and, consequently to better predict the transport and storage properties of soils, rocks and sediments.
-
-
-
Best Practices-design for Scale Reduction During Produced Water Reinjection (PWRI)
Authors S.A. Mostafavi, S. Riahi, M. Mavaddat and A. BigdeliSummaryWater disposal treatment is one of the most important environmental challenges in oil industry. Produced water is hazardous because of its impurities, such as different type of salts, natural inorganic and organic materials, injected chemical and high salinity concentration. In spite of injectivity risks, environmental concerns force industry from surface discharge of produced water to produced water reinjection (PWRI), which is known as the method for produced water treatment. During PWRI, the most common problem is scaling which is highly affected by thermodynamic condition and water chemical composition. Thereafter, it is critical to identify influence of different variables and potential problems related to scales. In this study the OLI ScaleChem software was used to investigate the scaling problem. Results reveal that the common scales are strontium sulfide and calcite carbonate, and If EOR scenario is considered, one-order diluted refinery water is helpful while refinery water is a better candidate as IOR injection fluid.
-
-
-
Evolution of the Orange Basin; Cretaceous Deepwater Fold-and-Thrust Belts to Cenozoic Mass Transport Systems
More LessSummaryThe focus of this study is on the deepwater Orange Basin, offshore SW Africa, in which several DWFTB systems are found. Previous studies have mainly focused on the 2D seismic interpretation of the Orange Basin, which is naturally limited. In this study, the availability of high-resolution, 3D seismic reflection data will allow us to constrain the strato-structural architecture of the deep-water Orange Basin from a Cretaceous DWFTB system to the overlying Cenozoic deposits using Schlumberger’s Petrel E & P software package for seismic interpretation. Understanding the architectural elements of southern Africa’s passive margin, and the tectonic evolution of the DWFTB systems contained within, is important in building on the scientific knowledge known of what occurs in these settings worldwide and in further constraining prospective sites for petroleum exploration in similar settings.
-
-
-
Natural fractures prediction in a lacustrine carbonate reservoir integrating 3D structural restoration and seismic AVAz techniques
Authors P.H. Silvany, M.N.C. Araujo, R.B. Bunevich, J. Almeida, C.E.B.D.S. Abreu and C.E.L. PereiraSummaryLow permeability carbonate reservoirs constitutes significant reserves of oil and gas for Petrobras’ E&P sector. Micro-porosities above ~ 10% in these sedimentary sequences allow the accumulation of significant volumes of hydrocarbon. However, the predominance of pore throats smaller than 10 microns produces low permeability in this type of rock, making it difficult for the flow of interstitial fluid during production. This constitutes the main challenge for making production feasible. The shortage of static and dynamic data in most reservoirs of low permeability makes it extremely difficult to understand the spatial distribution of the different scales of heterogeneities and, consequently, influence in obtaining realistic flow scenarios. In this work, an integrated supervised methodology is proposed for the characterization of natural fractures in a low permeability lacustrine carbonate reservoir in Santos basin, Southeastern Brazilian margin. To achieve the objectives of the study, a workflow was developed that involves the actions, briefly described below: (i) descriptive and kinematic analysis at multiple scales of the brittle structures; (ii) understanding of lithological control in the deformation process; (iii) quantification of the deformation in different phases of movement by 2D / 3D structural restoration techniques and (iv) analysis of seismic anisotropy.
-
-
-
Remote sensing, geological and geophysical integration – Study case in Valinho de Fátima, Portugal
More LessSummaryThe work intends to establish a methodology for the characterization, evaluation of the potential and choice of sites in rock masses, for the implantation of ornamental rock explorations.
A first phase was carried out with lithological and structural assessment at a regional scale, identifying in the Geological Map of Portugal (1: 50k). The next phase consisted of thematic geological mapping, with lithological and structural assessment at the local level.
In parallel, geophysical exploration was carried out using the Electromagnetic Method in the Time Domain [TDEM], to map the variation and distribution of resistivity, underground, and its correlation with the existing and modern cartography.
This document presents the results of remote sensing, using Copernicus Sentinel-2 composite images, in which the use of automatic assisted classification allows a quick thematic mapping of the surface, created from field knowledge and bibliographic knowledge of the region, for evaluation radiometric signatures for the objects of interest in the designated area.
-
-
-
Water Saturation Prediction in the Reservoir Zone of a Gas Field using SVR Method
Authors S.A. Afzali Fatatouei and M. BagheriSummaryEstimation of reservoir parameters is one of the most important factors in oil and gas reservoir investigations. One of the most important parameters for modeling reservoir is water saturation, which any errors in evaluating it can cause sever financial damages. By core analysis, calculating water saturation is possible, but there isn’t core in every place. Therefor there is a need to estimate water saturation. In this investigation for estimating water saturation, support vector regression was used, which is one of the applications of machine learning and can solve the curse of dimensionality. The well’s data was used for training in the method and water saturation was considered as labelled data. In the end for obtaining the best estimation two different kernels was used.
-
-
-
Quantitatively Evaluating the Preservation of Deep-water Channel Architecture using 3D Synthetic Seismic from Outcrop
Authors T. Langenkamp, L. Stright, S. Hubbard and B. RomansSummaryForward seismic reflectivity models can be used to interpret depositional architecture and stratal surfaces. However, such studies often stop short at a qualitative assessment of the link between underlying depositional architecture and seismic resolvability, lacking a quantitative assessment. This study addresses this gap with a direct quantitative comparison of 3-dimensional facies architecture predicted from seismic with a “ground truth” to quantify heterogeneity facies associations and architecture preserved in inverted seismic data. The primary goal is to quantify how facies architecture information is preserved in and predicted from inverted seismic reflectivity data. The objective is to explore what the variables are that impact correct vs incorrect facies classification. With increasing seismic frequency, channel axis becomes harder to predict while mass transport deposits became easier to predict. Facies in shallow reservoirs are easier to predict than in deep reservoirs. Disorganized channel systems show greater facies predictability than organized systems due to greater AI contrasts. This study highlights what architectural information is preserved in 3-dimensional inverted seismic data, built from outcrop data of a deep-water system, which can aid directly in interpretation, reservoir prediction, and modelling.
-
-
-
Entropy-driven particle swarm optimization for reservoir modelling under geological uncertainty – application to a fractured reservoir
Authors B. Steffens, V. Demyanov, D. Arnold and H. LewisSummaryIn this work we introduce a novel reservoir modelling workflow where modelling is assisted by an entropy-driven particle swarm optimizer. Producing a representative range of reservoir models that cover geological uncertainties in an effective way is a challenging task. We therefore make use of entropy to ensure that the ensemble of generated models adequately reflects the available information and provides diversity that reflects the associated variability in fluid flow behavior. The workflow is tested on a synthetic case study of a fractured reservoir.
The results indicate that the entropy-driven PSO is able to prevent the diversity of the ensemble of models from collapsing whilst staying within the bounds of a predefined expected dynamic flow response. It is also shown that the entropy-driven PSO outperforms a standard PSO in this task. Secondary outcomes from the workflow, such as a spatial entropy map, provide a great tool for further uncertainty assessment and can be used to identify swept or unswept reservoir regions and the regions where more information is needed to reduce the uncertainty.
-
-
-
Segmentation of Digital Rock Images Guided by Edge Feature Using Deep Learning
More LessSummarySegmentation of digital rock images is a crucial and basic step in digital rock process, and equivalent elastic parameter and fluid properties calculated from the digital rock can be affected by the result of segmentation. Conventional segmentation algorithm based on thresholding algorithm cannot perform a satisfying result in small structure due to noise impact. To address issues, a modified guided by prior information, edge feature, is proposed to improve accuracy of small structure. Edge feature reflects information of the effect of transport, weathered, and eroded in the deposition process, but the shape of noise and artifacts can’t reflect these information, rather show regularity due to the influence of instruments, hence boundary feature can improve the discrimination of noise. Furthermore, conventional SegNet was used to compare with modified SegNet, the former obtains 90.21% accuracy using 38-layers network, proposed approach using prior information achieves 93.07% accuracy using 30-layers network, which demonstrates less computational time and better anti-noise property. In addition, connectivity was used to evaluate segmentation result, modified SegNet shows a better similarity with origin image.
-
-
-
Characterizing Subsurface Damage Zones From 3D Seismic Data Using Artificial Neural Network Approach
More LessSummaryTo further improve the quality and efficiency of subsurface fault zone image and study its geometry. Herein we adopted post-stack seismic data conditioning and a combination of seismic multi-attribute for producing a new hybrid attribute through a supervised multilayer perceptron (MLP) neural network in the Jurassic formation of Cai36 3D prospect located in the eastern part of the Junggar Basin. We first conditioned original seismic data by using the dip-steering cube extracted from the original seismic data. Secondly, we extracted conventional seismic attributes from the conditioned data sensitive to fault zone signatures. Thirdly, we selected a set of “picks” at a time slice representing the presence or absence of fault zones. Then we adopted the supervised MLP neural network to train over the selected seismic attributes extracted at the fault zone and non-fault zone positions. We obtained a new fault probability cube as new attributes. Finally, we analyzed a typical strike-slip fault zone using the new attributes. This study provides an effective way of fault zone imaging from seismic data and adds new insights into its geometry. Therefore, the workflows used here could be widely applied to other 3D surveys.
-
-
-
Projection-based autoregressive neural network for model-reduced adjoint-based variational data assimilation
More LessSummaryThe adjoint method has been used very often for computation of analytical gradient, however, the generic implementation of the adjoint model often needs both significant programming efforts and computational cost. We proposed a novel projection-based autoregressive neural network (aNN) where the model-reduced adjoint is efficiently produced with the help of an easy-to-use auto-differentiation tool in deep-learning frameworks. This study restricts focus to propel orthogonal decomposition (POD) due to their physical interpretation and high scalability. Analogy to reduced-order tangent linear model, a projection-based aNN (POD-aNN) structure is proposed to accelerate the construction of adjoint model based on the reduced subspace. The POD-aNN consists of a dimensionality reduction and an intermediate non-linear transition unit which are used to produce the low-representation of the state system and approximate the time-varying dynamic, respectively. Thus we can derive a model-reduced adjoint model very efficiently. We demonstrate the performance of proposed methodology with many representative data assimilation experiments on a synthetic 2D subsurface flow model characterized by random spatially dependent parameters. The results have shown that this proposed POD-aNN obtains satisfactory results with significantly reduced computational cost and therefore demonstrates promising applicability to practical cases.
-
-
-
Higher-order mass-lumped wave propagators on variable resolution, triangular meshes
Authors K.J. Roberts, A.F.G. Olender, R.D.S. Gioria and B.S. CarmoSummaryUsing the finite element method leads to a sparse system of equation that are relatively computationally expensive to solve for as compared to finite difference methods. However, by using higher-order mass-lumped triangular finite elements that lead to diagonal mass matrices, the computational cost is dramatically reduced enabling the use of unstructured triangular meshes. However, to efficiently use these elements, a cost-effective distribution of degrees-of-freedom needs to be chosen. This abstract shows results using different spatial polynomial orders of mass-lumped triangular elements while varying mesh sizes for a 2D homogeneous case and applies those results in a heterogeneous 2D case.
-
-
-
Rejuvenation of the T40-T45 paleogeography model in the Flett Sub-Basin (West of Shetlands)
Authors R. Julien, H. Olsen, C. Climent and H. CromieSummaryA new paleogeographic model for the Late Paleocene to Early Eocene T40-T45 sequences ( Ebdon et al., 1995 chronostratigraphic nomenclature) for the Flett Sub-Basin (West of Shetlands Basin) has been constructed. Seismic interpretation (on 2D and 3D data) was integrated with wells information to constrain the spatial extension of the different gross depositional environments. During this period, the Flett Sub-Basin was semi-confined between the West of Shetland platform and the massive Faroe basalt plateau. A continental clastic wedge was deposited in the south west of the sub-basin whereas in the north east (at Bunnehaven and Tobbermory well location), starved deepwater settings prevailed. The basin was also marked by a multi-scale interaction of sedimentation with volcanism linked with the opening of the North Atlantic Ocean and the activity of the proto Icelandic mantle plume. Different scenario regarding sand paths (from the coastal domain to the deepwater basin) have been investigated. Despite all the efforts, no scenario prevailed on the others and Bunnehaven sands’ provenance remains ambiguous.
-
-
-
Early tectonic structures concealed by the latest deformation phase might reveal exploration opportunities. An outcrop example.
Authors D. Casabianca, A. Barrier, A. Ricciato, R. Di Cuia and S. BorelloSummaryThe geological evolution of most sedimentary basins currently explored for hydrocarbon resources, involved more than one tectonic deformation phase. The earlier structures, concealed by the easily recognized latest deformation, often remain overlooked by interpreters. This might result in poor exploration decisions such as unnecessary exploratory wellbores or missed opportunities. The Maiella mountain outcrop, in the Southern Apennines of Italy, is an example where a Pliocene to Pleistocene aged compression anticline has folded carbonate sequences containing Cretaceous aged extension faults. The study of this outcrop provides insights for the deliberate search of exploration opportunities hidden within early structures in basins where multiple deformation phases are recorded.
-
-
-
Lithology segmentation using deep neural network
More LessSummaryThis paper avoids the difficulties in using conventional methods in lithology segmentation task by putting the tasks in the frame of computer vision.
First, we setup a lithology dataset which contains paired topology, satellite and lithology images; Second, two heated neural networks HyperNet and UNet are introduced and applied in lithology segmentation task.
The experiments show that both HyperNet and UNet are efficient and promising for the application in lithology segmentation.
%
Neural networks can increase the predicted accuracy three times than random guess, that greatly reduce the workload of professional lithology geologist.
-