While many factors in the reservoir cannot be controlled, there are three controllable factors in field development that make a significant impact. More reservoir contact leads to more oil produced. Controlling sand and water means lower treatment costs, and in-situ reservoir management leads to higher cumulative production. While the underlying technologies have been around for up to 20 years, it is only recently that their synergies and true value are understood. This paper will demonstrate the effect each of these technologies has on increasing overall production rates, improving recovery, and reducing the cost per Barrel of Oil Equivalent (BOE).
The successful implementation of multilaterals in the North Sea will be analyzed. Since 1996, over 300 multilateral junctions have been installed on the Norwegian continental shelf fields with currently approximately 30 junctions completed each year.
Additionally, simulations will be used to demonstrate the incremental improvements in oil recovery that can be obtained by using properly designed advanced completions that include multilaterals, sensors, and passive/active flow control equipment.
The paper will evaluate production performance of a vertical well field development base case against scenarios using horizontal and multilateral wells. It will show how fields can be optimized, leading to increased oil and decreased water production.
Production rates can be significantly improved by combining multilaterals with other advanced completion techniques, such as intelligent completions and inflow control devices. The subject field simulation can be further optimized to manage gas and water production.
With a tailored multilateral field design, combined with properly designed advanced completions systems, the simulation succeeds in terms of achieving maximum contact with the oil reservoir and meeting improved ultimate recovery objectives.
It can be concluded that as reservoir contact is increased, a reduced decline in production rate is observed leading to both a higher Estimated Ultimate Recovery (EUR) and optimized drawdown profile distributions. Additionally, results will be presented that have considered oil production and a method to lower production of unwanted fluids or gas.
This paper also demonstrates the value of field development design from the perspective of reservoir simulation. It is through reservoir insight that a level of understanding is created that can help define the optimum well and completion design to meet field expectations.
Advanced multilaterals continue to grow in popularity with many operators, and it therefore becomes important to evaluate the value of different field development methods. This knowledge can aid operators in unlocking new reservoir targets and optimizing field development, and ultimately will improve recovery factors and overall field economics.
Intelligent multilateral well completions provide downhole flow rate, pressure, and temperature measurements at multiple well segments which allows for a continuous spatiotemporal data stream. Such an extensive data input poses a challenging task to decide on the optimal strategy of manipulating the inflow control valve (ICV) settings over time for best performance. This study investigated the use of machine learning to analyze and predict well performance given different ICV settings to ultimately maximize the well output.
A commercial reservoir simulator was used to generate two synthetic reservoir models: homogeneous (Case A) and heterogenous (Case B). These synthetic data were used to train, validate, and test machine learning models. The reservoir cases were generated based on a segmented, trilateral producer completed with three ICV devices installed at tie-in segments. The data used were measurements of wellhead and downhole flow rates across ICV segments over a period of 4,000 days. A total of 1,330 experiments were conducted with an eight-day timestep, generating a total of 667,660 sample data points for each of Case A and Case B. Fully connected neural networks were used to fit the data while model generalizability was enhanced using regularization techniques, namely L2 regularization and early stopping.
Both random sampling and Latin Hypercube Sampling (LHS) methods were evaluated in constructing the training, validation, and testing splits. Trained with different sample sizes drawn from the 1,330 simulated data histories for the two reservoir models, the proposed neural network showed excellent results. Given only ten simulated choices of ICV settings for training, the network proved capable of predicting oil and water production profiles at surface for both homogeneous and heterogeneous reservoir models with over 0.95 coefficient of determination (R2) when evaluated at unseen, test ICV settings. Extending the problem to downhole flow performance prediction, about 40 training simulated settings were necessary to achieve 0.95 R2. We observed that LHS was superior to random sampling in both R2 average and confidence interval. We also found that increasing the training and validation sample sizes increased the test R2 when testing against unseen cases. Study results suggest the applicability of machine reinforcement learning to estimate the well output at different ICV settings, where the neural network model depends fully on the real-time well feedback and production measurements.
By using a machine learning approach during the operation of a well with multiple ICV settings, it would be feasible to estimate the lateral-by-lateral output at unseen scenarios. Hence, it becomes possible to maximize the well output by using an optimization algorithm to determine the optimal ICV settings.
Monitoring and reevaluation of petrophysical attributes in a mature field under production for many decades is crucial for optimizing production and further development planning. In this case study, a multidisciplinary approach is deployed for formation evaluation and reservoir characterization using logging-while-drilling (LWD) sensors spanning formation volumetrics, fluid analysis, high-resolution image interpretation, and geomechanics to confirm remaining oil saturations and help identify recompletion intervals. LWD technologies were used in four wells in Sahmah field of Oman to provide an integrated petrophysical and geomechanical field study using a bottomhole assembly (BHA) including gamma ray, resistivity, formation bulk density, thermal neutron, acoustic, high-resolution imaging, and formation pressure testing sensors. A deterministic multimineral petrophysical model was used to derive formation volumetrics and fluid analysis. Geomechanical interpretation used high-resolution microresistivity imaging, acoustic slownesses, and formation pressure data to verify principal stress orientations and to quantify pore pressure and horizontal minimum and maximum stress magnitudes. These data were then correlated with historical data to evaluate sweep efficiency and residual fluid saturations. LWD sensors have proven to provide robust geological, petrophysical, and geomechanical data compared to previous traditional wireline data acquisition.
Li, Feng (Southwest Petroleum University) | Xie, Xiong (CNOOC-Shenzhen) | Huang, Li (CNOOC-Shenzhen) | Zhou, Luyao (CNOOC-Shenzhen) | Chang, Botao (Schlumberger) | Wang, Chao (Schlumberger) | Wang, Fei (Schlumberger) | He, Chengwen (Schlumberger)
In China, the main sandstone reservoir M of the LF oilfield entered the mature development stage with high water cut (average 93%) and 66.1% recovery. Remaining oil exists vertically in the H layer at the top section of this massive bottomwater reservoir and laterally at margins of current development area with less well control. The H layer consists of several thin (0.5 to 2 m) sand sublayers interbedded with calcareous tight sublayers with low permeability; the effective oil drainage radius of single borehole is 100 to 150 m. Maximum reservoir contact (MRC) technology was employed to increase drainage area and volumetric sweep efficiency for optimal production and recovery to rejuvenate this mature reservoir.
In an original hole with 98 to 99.9% water cut targeted for a workover operation, two new laterals were sidetracked to comprise a three-lateral MRC configuration with openhole completion to develop the SL1 target sublayer of the H layer. The success of MRC wells depends on an efficient openhole sidetrack and azimuth turning. Moreover, multilaterals need to precisely chase the sweet zone in the reservoir. Drilling into overlying shale causes borehole collapse, and penetrating the underlying tight zone causes fast bottom water breakthrough. Low resistivity contrast increases the difficulty of distinguishing the target zone from the shoulders. Sparse well control and limited seismic resolution bring high structural and stratigraphic uncertainties. Accordingly, effective services were equipped to overcome these challenges to achieve the required engineering and reservoir objectives. The new-generation hybrid rotary steerable system (RSS) tool provides stable, rapid, and accurate steering control, even with high dogleg severity, to achieve engineering objectives. With a balance between resolution and depth of investigation (DOI), high-definition deep-looking resistivity inversion uses the Metropolis coupled Markov chain Monte Carlo method to clearly identify multiple layers (more than three) within an approximately 6 m DOI, formation resistivity distribution, anisotropy, and dip, even in this low-resistivity-contrast environment. Reservoir details could be clearly unveiled to help MRC lateral steering along the thin target. Furthermore, a wide-range-displacement electrical submersible pump (ESP) helps optimize openhole performance.
Six new laterals were drilled in three MRC wells. Hybrid RSS tools provided 100% openhole sidetrack success rate, and laterals were turned laterally with 15 to 70° azimuth change and 200- to 570-m displacement to maximize the drainage area. Deep-looking inversion revealed high-definition reservoir details by delineating three key boundaries and four adjacent layers' profiles simultaneously and identifying target zone's thickness and property variation. The target sand is 0.5 to 2 m thick with resistivity of 2 to 9 ohm-m, surrounded by interbeds with resistivity 8 to 10 ohm-m. Within the refined 3D reservoir model, the horizontal laterals efficiently chased the top section of effective target sand while avoiding high-risk shoulders. Total 4298-m horizontal footage was achieved in six laterals with net-to-gross 91% in the SL1 thin, low-permeability reservoir. With the proper ESP configuration, approximately 688,500 bbl of oil have been produced as of December 2018. Especially in two workover MRC wells, after approximately 2.5 years of production, the current water cut is 96 to 97%, lower than water cut (98 to 99.9%) before the workover operation, and daily oil production increased significantly.
Integrated drilling, logging, and production services provided MRC efficiency to rejuvenate this thin, low-permeability and low-resistivity mature reservoir.
Mustafa, Ayyaz (King Fahd University of Petroleum and Minerals) | Abdulraheem, Abdulazeez (King Fahd University of Petroleum and Minerals) | Abouelresh, Mohamed Ibrahim (King Fahd University of Petroleum and Minerals) | Sahin, Ali (King Fahd University of Petroleum and Minerals)
The lower Silurian Qusaiba Shale is one of the major source rocks for Paleozoic petroleum reservoirs in Saudi Arabia and is considered a potential shale gas resource. The study aims to evaluate the prospectivity and improve the production potential of Qusaiba shale by defining the lithofacies and mineralogy as controlling factors for brittleness and other mechanical parameters.
The continuous 30 feet subsurface cores and log data of Qusaiba Shale from Rub’ Al-Khali Basin were utilized for the study. Geological characteristics on the core were fully demonstrated in terms of size, mineralogy, color, primary structures and diagenetic features to identify lithofacies. In addition, 30 thin sections were used to study micro scale geological characteristics. The powder X-ray diffraction (XRD) was used to determined the mineralogical compositions. Surface morphology visualization and elemental analysis were performed using the scanning electron microscope supplemented with energy dispersive spectroscopy (SEM-EDS). Acoustic velocity measurements and compressive strength tests were performed on 15 core plugs (5 from each lithofacies).
Based on the above-mentioned analyses, three lithofacies were identified: (1) Micaceous laminated organic-rich mudstone facies (Lithofacies-I), (2) Laminated clay-rich mudstone facies (Lithofacies-II), and (3) Massive siliceous mudstone facies (Lithofacies-III). Mineralogical composition resulted in variable amounts of quartz ranging from 39 to 40, 45-55 and 60 to 78% for Lithofacies-I, II and III, respectively. Lithofacies-I having relatively lower quartz and higher clay percentage and total organic content (12% by volume) exhibited low stiffness. Mineralogy- and elastic parameters-based brittleness indices exhibited ductile behavior of this lithofacies. Lithofacies-II with relatively higher quartz (45 to 55%) and lower clay contents and TOC (3-5%) than Lithofacies-I resulted in relatively higher stiffness and brittleness. The brittleness index exhibited brittle behavior for silica rich Lithofacies-III (low TOC< 3%) as reflected by Young's modulus (average 32 GPa) and low Poisson's ratio (average 0.25). Hence, it is concluded that mineralogy and geological characteristics are the main controlling factors on mechanical properties and brittleness. The integration of three essential disciplines i.e. geology, mineralogy and geomechanics, plays the key role to better evaluate the production potential by highlighting the sweet spots within the heterogeneous shale gas reservoirs.
Reservoir evaluation of source rock is still a challenge because the geochemical assessment of the kerogen content is complicated and time consuming. Existing traditional methods to characterize kerogen involves the removal of inorganic minerals which is a critical preliminary step. The incomplete isolation of kerogen may introduce some errors and uncertainties in kerogen content estimation. The alteration of kerogen microstructure during this process has also been documented. The current approach still requires input from geochemical measurement of total organic carbon (TOC) while the conversion of TOC to kerogen volume requires the precise value of a conversion factor and kerogen density. Overall, there is yet a standard lab or field scale approach to characterize kerogen content. These difficulties and uncertainties prompt the motivation to attempt a new methodology to quantify the kerogen content of unconventional shale from porosity measurements.
Porosity is the basic rock property that is related to the volumetric average of pore space. The distinction between the total and effective porosity is meaningless for shale and this characteristic property has enabled the preservation of its organic content. The recent popularity and growth of different measurement techniques is in part closely tied to the near zero porosity of shale. Two special cases of practical interest are NMR and density porosity measurements which can both be measured in the rock physics lab and well logs. NMR porosity is sensitive to 1H which is naturally enriched in kerogen whereas density porosity must be calibrated to the mineral matrix.
Based on porosity measurements, the emerging aproach is that the kerogen volume fraction is the contrast between NMR and density porosity. Although, the theoretical basis of this approach is not satisfactory, it is straightforward and far less complicated than the existing approaches to quantify kerogen content. We investigate this concept further based on laboratory measurement. We conducted laboratory measurements of NMR porosity, bulk density, grain density and TOC on Qusaiba shale to characterize its kerogen content. In our approach, we conducted the NMR experiment on the shale samples in the dry state without fluid saturation.
By International Petroleum Technology Conference (IPTC) Monday, 25 March 0900-1600 hours Instructors: Olivier Dubrule and Lukas Mosser, Imperial College London Deep Learning (DL) is already bringing game-changing applications to the petroleum industry, and this is certainly the beginning of an enduring trend. Many petroleum engineers and geoscientists are interested to know more about DL but are not sure where to start. This one-day course aims to provide this introduction. The first half of the course presents the formalism of Logistic Regression, Neural Networks and Convolutional Neural Networks and some of their applications. Much of the standard terminology used in DL applications is also presented. In the afternoon, the online environment associated with DL is discussed, from Python libraries to software repositories, including useful websites and big datasets. The last part of the course is spent discussing the most promising subsurface applications of DL.
Late Cretaceous plate collision and subsequent ophiolite emplacement at the eastern margin of the Arabian Plate led to compressional events, responsible for the formation of the structural traps of the giant fields onshore Abu Dhabi. In addition, the onset of this structuration during the Turonian caused the configuration of some hence-to-forth overlooked features (pinch-outs and siliciclastic sand deposits). The objective of the present work is to analyze the origin and distribution of these geometries and their potential to constitute stratigraphic traps.
To understand the genesis and the distribution of these geometries which formed during the Late Cretaceous, we used a combination of large scale regional stratigraphic well correlations and seismic lines interpretation, together with age dating, core description, and well data information. The methodology consisted in using this data for detailed mapping of relevant time stratigraphic intervals, placing the mapped architecture in the context of the global eustatic sea levels and major geodynamic events of the Arabian Plate.
The ensuing plate collision during the Turonian in eastern plate margin was felt hundreds of kilometers into the plate over Abu Dhabi area. Buckling and uplifting created paleo-relief which caused exposure and erosion of Wasia Group sediments in northern and eastern areas of Abu Dhabi Emirate. This led to the configuration of some overlooked stratigraphic features: eroded rims and lateral facies change against structural dip (Mishrif Formation); onlap pinch-outs onto flanks of major structures (Ruwaydhah Formations) and the deposition of siliciclastic sand deposits of the Tuwayil Formation. The features follow low relief areas along contemporaneous synclines in onshore Abu Dhabi and salt withdrawal synclines in offshore Abu Dhabi.
With further advance of the obducting ophiolites, a foredeep developed leading to drowning of the previously exposed structures. Shales and interbedded limestones of the Laffan Formation were unconformably deposited over the eroded Wasia Group during the Coniacian transgression associated with the generation of this foredeep. They are now forming an extensive regional seal for these deposits forming potentially stratigraphic traps.
We postulate that the rejuvenation of the Shilaif intrashelf basin during the Late Turonian and the deposition of the (Ruwaydhah Formation) was aborted at its early stages by periods of uplift, erosion and their successive erosional unconformities, features that are confirmed on the crest of several eastern area structures. This provided the context for the generation of pinch-out geometries that constitute potential stratigraphic traps downdip of major structures in Abu Dhabi.
Very little has been published about the outline and architecture of these stratigraphic traps in Abu Dhabi and the detailed circumstances that led to their genesis, topics that are comprehensively analyzed in the present work.
El Hawy, Ahmed (Schlumberger) | Al Busaidi, Adil (Schlumberger) | Vasquez Bautista, Ramiro Oswaldo (Schlumberger) | Awadallah, Muhannad (Schlumberger) | R. Heidari, Mohammad (Schlumberger) | Saidi, Khaled (Schlumberger) | Escamilla, Barton (Schlumberger) | Al Abri, Zahran (Petroleum Development Oman) | deBoehmler, Guy (Petroleum Development Oman) | Al Harthi, Said (Petroleum Development Oman) | Haeser, Patrick (Petroleum Development Oman) | Al Riyami, Khaleel (Petroleum Development Oman) | Picha, Mahesh (Petroleum Development Oman)
As one of the worst oil & gas business downturns struck, the need for a revolutionary approach of drilling was needed. Optimization was the key word during that period, it was about time to look back at drilling fundamentals, review and learn from previous failures and lessons while establishing new foundation for a transformed yet successful process that ensured an all-time historical success.
While many trials of drilling optimization initiatives were executed over the years, inconsistent drilling performance delivery and repetitive failures continued to raise a red flag each time for variety of reasons.
From challenges achieving required performance levels and dog legs in the top sections with increased risks of axial and lateral vibrations, to the difficulties faced in the landing section drilling through unconsolidated and reactive shales in the north, and through fragile weak formations in the south to the difficulties transferring weight to the bit at deeper depths in the horizontal laterals drilling highly porous zones of sticky limestones.
While cost optimization was the trend during the downturn, there was no better option to achieve desired financial results for both operator and service provider than the inclusion of the drilling optimization in action initiative into every well drilling program, it was proven to be an ultimate win-win technical and business solution.
The task of reliable characterization of complex reservoirs is tightly coupled to studying their microstructure at a variety of scales, which requires a departure from traditional petrophysical approaches and delving into the world of nanoscale. A promising method of representatively retaining a large volume of a rock sample while achieving nanoscale resolution is based on multiscale digital rock technology. The smallest scale of this approach is often realized in the form of working with several 3D focused-ion-beam–scanning-electron-microscopy (FIB-SEM) models, registration of these models to a greater volume of rock sample, and estimation and scaling up of model local properties to the volume of the entire sample. However, a justified and automated selection of representative regions for building FIB-SEM models poses a big challenge to a researcher. In this work, our objective was to integrate modern SEM and mineral-mapping technologies to drive a justified decision on location of representative zones for FIB-SEM analysis of a rock sample. The procedure is based on two experimental methods. The first method is automated mapping of sample surface area with the use of backscattered electrons (BSEs) and secondary electrons (SEs); this method has resolution down to nanometers and spatial coverage up to centimeters, also referred to as large-area high-resolution SEM imaging. The second method is automated quantitative mineralogy and petrography scanning that allows covering sample’s cross section with a mineral map, with resolution down to 1 µm/pixel. Data gathered with both methods on millimeter-sized cross sections of rock samples were registered and integrated in the paradigm of joint-data interpretation, augmented with computer-based image-processing techniques, to provide a reliable classification of nanoscale and microscale features on sample cross sections. The superimposed SEM and mineral-map images were combined with physics-based selection criteria for reasonable selection of FIB-SEM candidates out of a great number of potential sites. In the result, a semiautomated work flow was developed and tested. Demonstration of the work flow is made on one of Russia’s most promising tight gas formations, where the characteristic dimension of void-space objects spans from a single nanometer to millimeters. An example of an optimized site selection for FIB-SEM operations is discussed.