Offset Wells Research Articles

Data-driven prediction of drilling strength ahead of the bit

This paper compares the performance of two data-driven methods, Signal-Matching Predictor (SMP) and Long Short-Term Memory (LSTM), for predicting drilling strength (Es) ahead of the bit based on drilling data from nearby offset wells. The comparison is based on the accuracy, applicability, complexity, and computational cost of the methods with the objective of suggesting the most appropriate tool for look-ahead drilling strength prediction. The methods were tested using data from offshore wells in Newfoundland. The SMP used a fixed-size sliding-window of real-time Es data from the target well to find a match in the offset well within similar geological formations and chose the scaled value from the offset well as the prediction. In the second approach, twelve LSTM models were trained using the drilling data of twelve offset wells, and the drilling data of the thirteenth well was used for blind testing. Results showed that the SMP achieved a coefficient of determination (R2) of 0.92, 0.92, and 0.79 for predicting 1.5, 3, and 5 feet ahead of the bit, respectively, while the LSTM reached an R2 of 0.95, 0.92, and 0.80 for the respective prediction intervals. The R2 of the LSTM models was further increased to 0.96, 0.94, and 0.83 after retraining it with weighted samples in formation transition zones. Also, a post-processing technique was proposed that further enhanced the R2 of the LSTM-based approach to 0.98, 0.97, and 0.93, respectively. The strength of the LSTM-based approach was to use measurable drilling parameters as the only inputs and not the Es itself. According to the results, the LSTM-based method can be reliably used to predict the Es ahead of the bit allowing drillers to identify upcoming drilling dysfunctions.

Hess’ ‘Drainage’ Wells Show Strong Production but Face Cost and Spacing Challenges

A new kind of horizontal well has emerged from an experiment in North Dakota’s Bakken and Three Forks shales. It uses an uncemented slotted liner to produce hydrocarbons by intercepting fractures that originate from offset wells. Hess Corp. revealed the concept over 2 years ago, confirming at an SPE conference that crude was flowing from the pilot well. The New York City-headquartered oil company also announced that additional trials were in progress, though it withheld specific production figures. The wait for those key details is now over. At the recent Unconventional Resources Technology Conference (URTeC) in Houston, Hess presented its first technical paper on the augmented drainage development (ADD) pilots which it claims have proven the approach to be, at the very least, a technical success. In URTeC 4044110, the company’s authors underlined their optimism in noting that ADD wells “could lead to a step change in productivity and recovery in unconventional developments.” Hess acknowledges that economic and technical hurdles currently stand in the way of that vision. But the upbeat assessment is backed by impressive production data from its inaugural pilot launched in 2021. Placed about 240 ft between two actively stimulated Middle Bakken wells, the first drainage well exceeded expectations, averaging nearly 1,700 B/D of crude in the first 8 days of flowback. Output fell to about 1,500 B/D over the next few days as the nearest actively stimulated well—called a “standard well” by Hess—began its flowback. This was one of several confirmations that the wells were in direct communication. “A lot of bets were lost on this well—maybe not by me, but by a lot of people,” remarked Craig Cipolla, a principle completions engineering advisor for Hess. Cipolla presented the company’s findings at the conference this past summer. He said the ADD well also pumped out about as much frac sand as is seen in a typical shale well, which in addition to tracer data, verified that it was well propped. During the initial production period, the first ADD well achieved about 85% of the output of the offsetting standard wells on the pad. Although this figure eventually dropped to just over 40% over the next 2 years, it was enough for Hess to consider the concept validated. Based on the results from all three pilots, Hess estimates ADD wells in the Bakken can be expected to achieve 10 to 45% of the output of typical hydraulically fractured horizontal wells. On the low end, the oil recovered from distant and low-conductivity fractures likely won’t justify drilling even a wellbore as barebones as the ADD concept calls for. Conversely, if a drainage well produces too much oil it is a sign that well spacing is too tight and that it is overly competing for resources with more-expensive offsets. This is likely the story of why the first ADD well came on so strong.

Multilayer 3D seismic attribute analysis and fault framework modeling of shallow extensional fault systems in the Delaware Basin

Shallow extensional fault systems within the southern Delaware Basin have gained the interest of Permian operators since published works noted them in 2019. These shallow extensional fault systems, referred to as “lineaments” by operators, are seismogenic hazards, pose significant hazards to drilling operations, and are associated with elevated water-oil ratios and H2S concentrations. Previous work using seismic attribute analysis has resulted in the successful mapping of extensional fault systems at the Ochoan/Guadalupian interface, whereas fault geometry, origin, and depth of penetration have remained elusive. First- and second-order ant tracking of the variance cube is rendered on multiple seismic horizons throughout an 856 km2 (approximately 330 mi2) depth-migrated 3D seismic reflection volume to expand the interpretation of the deformation zone at depth and guide a robust 3D fault model that forms the framework for the predictive model. A total of 60 previously undocumented features are identified in southwestern Reeves County — 56 penetrate the Ochoan/Guadalupian interface, 60 penetrate the middle Guadalupian, 29 penetrate the early Guadalupian, and 21 penetrate the early Leonardian. Structural modeling returns an average strike orientation of 318° north-northwest–south-southeast and an average dip of 84°. Spatial and temporal analysis of the model constrains the timing of extension to the Cenozoic. In contrast to previous studies, the improved depth of investigation of the 3D fault model allows for the correlation of documented drilling hazards with offset wastewater disposal wells along shared lineaments. Communication is indicated by H2S risking flags from mass spectrometry data along the vertical and horizontal sections of producing and injecting wells. Drilling data validates the seismic characterization and prediction of the 3D faulted architecture within the largest contiguous study of shallow extensional fault systems within the Delaware Basin and establishes a repeatable workflow for mitigating the negative impacts of shallow lineament geohazards on drilling operations and production profiles.

Quantifying the Effects of Interwell Communication Using Dynamic Fluid-in-Place Calculations

Summary Multifractured horizontal wells (MFHWs) completed in the same reservoir layer, or different reservoir layers, commonly experience interwell communication through hydraulic fractures. For example, after a well is placed on production, its production performance can be impacted by communication with an offsetting well placed on production after it. The degree of communication between wells is important to quantify for the purposes of well production forecasting, reserves estimation, completions, and well spacing design optimization. In this study, dynamic fluid-in-place calculations, performed using the impacted producing well rates and flowing pressures, are applied to quantify the effect of communication with an offset-producing well on the impacted well-contacted fluid-in-place estimates. Agarwal (2010) demonstrated that pressure transient analysis theory can be used to derive the volume of fluid in place contacted by a well (CFIP) over time during constant rate, transient production. The method was later extended to variable-rate/pressure scenarios. However, all previous applications of Agarwal’s method were for single, isolated wells and assumed single-phase flow of oil and gas. To evaluate the usefulness of the method for modern development scenarios, it is extended to allow for quantification of interwell communication during flowback, for which single-phase flow of water before the breakthrough of formation fluids may precede multiphase flow of formation and fracturing fluids, and for analysis of multiphase data. Analysis of flowback data enables early-time identification and quantification of interference effects. Multiple numerical simulation cases are generated to simulate different degrees of communication for the case of a two-phase flow of oil and water. Wells are assumed to be communicating through a hydraulic fracture with a specified transmissibility multiplier (Tmult) used to adjust the amount of interwell communication. Corrections for multiphase flow in the CFIP method are performed using two different methods—the total volumetric flow rate (combined phase) approach and the multiphase pseudovariable approach. The CFIP diagnostic plot (i.e., log-log plot of CFIP vs. material balance time) is applied to the impacted producing well to evaluate the CFIP trend before and after offset well production and the magnitude of CFIP change. The practical application of the method is demonstrated with field cases. From the simulation cases, it is observed that, after the offset well is placed on production, a reduction of CFIP for the impacted producing well occurs (rapidly decreasing at first and then stabilizing after a transition period) proportional to productivity index reduction. The loss in CFIP for the impacted producing well can be determined simply by estimating its CFIP immediately before and after offset well production. For high connectivity (Tmult> 0.25) scenarios, application of the combined phase approach resulted in estimates of the impacted well CFIP reduction of ~46–50%, whereas application of the modified pseudovariable approach resulted in estimates of ~49–51%. For the low connectivity case (Tmult = 0.001), these estimates were ~11% and ~9%, respectively, for the two approaches. Therefore, for the simulation cases studied herein, the two approaches agreed within acceptable error. Numerical simulation was also used to verify the absolute change in CFIP using these two approaches for correcting for multiphase flow. The practical application of the modified CFIP method was demonstrated using two field cases with early-time production. Both field cases demonstrated that changes in CFIP for the impacted well can be unambiguously interpreted. In the first field case corresponding to early-time production data (gas and water) associated with Well 23 of the SPE data repository, the reduction in CFIP of the impacted producing well was estimated to be ~37% using the combined phase approach. In the second field case, for which a producing well completed in a low-permeability gas condensate reservoir is impacted by placing multiple offset wells on production at the same time, the reduction in CFIP of the impacted well was estimated to be ~20% using the combined phase approach. In this study, we demonstrate for the first time that CFIP calculations can be applied to quantify interwell communication between two wells during flowback or early-time production when multiphase flow occurs in the reservoir.

Ponding and compartmentalization of a giant Paleogene slope fan by mass transport complexes, offshore Newfoundland, Canada

The Ephesus fan is a giant channel-lobe complex that extends across 460 km2 in the lower structured slope of the West Orphan Basin, offshore Newfoundland, Canada. Based on biostratigraphic dating of offset wells, the fan was likely deposited during the Oligocene. High resolution 3D seismic data reveal that the fan overlies a succession of mass transport deposits that initially created a flatter topographic surface, with later slumping events building elongated ridges. Subsequent high density turbidity currents coursed into the area through a single feeder channel, depositing as lobes onto the flatter area via a network of distributary channels. The lobes were circumferentially ponded and internally compartmentalized by the slump ridges, and the ponding forced the fan to assume its unique, 46 km-wide shape. Distributary channels were initially directed to the south side of the fan, due to the higher relief of the mass transport deposits to the north. After the infill of the accommodation space to the south, subsequent sediment input levelled the remaining rugose topography, and some distributary channels were able to avulse north. Throughout deposition of the fan, distributary channels also continued down-dip from the lobe complex, indicating some sediment bypass continuing down the lower slope. Overlying laminated mudrock displays pervasive polygonal faulting, signaling dewatering processes and the end of deposition by muddy submarine landslides and high density turbidity currents.

Real-time chemostratigraphy at wellsite; removing drilling uncertainties

Real-time chemostratigraphy is a workflow deployed at wellsites to remove stratigraphic uncertainty and support drilling operations. This study presents the chemostratigraphic wellsite workflow which, during drilling operations, support either the placement of casing points (using predefined geochemical markers to target a depth above a specific stratigraphic feature) or aid geosteering. Geosteering is typically employed during the drilling of a horizontal well and is done to keep the drill within a certain zone or horizon to maximise production. Wellsite chemostratigraphy has been effectively deployed to wellsites within multiple basins around the globe and has recently been utilised in the Perth Basin. The wellsite workflow starts before the deployment, with offset wells analysed by inductively coupled plasma (ICP) spectroscopy within a laboratory-based setting. This builds the correlative element-based framework and identifies the trends required to assist with either core placement, casing point, or geosteering. Samples are analysed by energy dispersive x-ray fluorescence units (these data are then used to confirm the zone previously defined) and can be identified at wellsite using the onsite tool, the selected elements of which are those less likely to be affected by drilling mud contamination and loss of circulation material (LCM). However, during drilling operation, there may be instances where LCM must be added to the drilling muds, which may have an effect on the data quality for certain elements. With a dedicated ICP spectrometry elemental database behind the wellsite operation, machine learning tools may be employed to ‘repair’ correlation critical elements.

Novel Resin-Coated Sand Placement Design Guidelines for Controlling Proppant Flowback Post-Slickwater Hydraulic Fracturing Treatments

Summary Resin-coated sand (RCS) is an effective way for controlling post-stimulation proppant flowback. However, with the shift to slickwater treatment fluids, the “tail-in” placement approach has proved to be less efficient for complete flowback control due to the proppant settling characteristics of using low-viscosity fluids. A new RCS placement approach was developed based on the results of several flowback studies. Trial wells were completed in different US basins with successful results. Proppant flowback samples were collected during different stages of drillout and production from wells using slickwater fluid systems. Thirty-five wells, completed by thirteen different operators, within the Permian and MidCon basins were evaluated. All wells were completed using multiple proppant mesh sizes. A total of 375 flowback samples were collected during the drillout and production phases. The samples were sieved, and the results were fed into an in-house material balance model to determine the percentages of different mesh sizes in the flowback samples. The conclusions were used as guidelines for a new placement approach implemented in multiple new wells to control proppant flowback. The flowback samples ranged from predominantly lead proppant to a similar proportion of the pumped mesh sizes. Not one of the 35 wells had flowback samples containing the majority tail-in mesh size. This supports the early sand dune assumption, suggesting that the early proppant forms dunes near the wellbore and late sand settles over the existing proppant beds. The use of late RCS appears to have a minimal effect on preventing flowback of the early proppant within a stage utilizing slickwater fracturing. Therefore, RCS efficiency to control proppant flowback with the tail-in method is reduced when used in such slickwater stimulations. To seal the different proppant beds, the new approach recommends pumping multiple RCS steps within a stage. The first RCS step is recommended within the first 10–20%, the second sequence within the first 40–60% of proppant volume, and the third as a tail-in. The exact percentages and step design were based on the results of flowback samples from neighboring wells. The implementation of this approach in more than 30 wells resulted in superior flowback control compared to offset control wells. In all trials, the proppant flowback completely stopped within 1 to 7 days of starting production. In this paper, we discuss the drawbacks of the current RCS placement practice while suggesting a new practical approach supported by data. RCS tail-in showed successful flowback control with viscous fracturing fluids and hybrid systems. For slickwater systems, an optimized placement design for RCS throughout the pump schedule provided enhanced flowback control compared to RCS tail-in. Finally, we illustrate the results of field trials in which utilizing the new RCS placement approach successfully reduced flowback.

Crosswell Strain and Production-Interference Testing Integrated in the Permian Basin

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 212336, “How and Where Wells Talk: Integrating Crosswell Strain and Production-Profile-Interference Testing,” by Matthew Lawrence, SPE, Ziebel; Janel Andersen, SPE, PDC Energy; and Yuanshan Zhang, Ziebel. The paper has not been peer reviewed. _ In Permian Basin fracturing operations, azimuths and intensities recorded by single-use distributed fiber optics in two wells are integrated and compared with post-flowback production-allocation and interference testing acquired with intervention fiber optics to identify areas of hydraulically conductive fractures and communication between offset wells. The production-allocation/interference testing method appears to reveal hydraulically connected fractures even in the absence of strain data. Introduction Intervention-based fiber optics have proved a valuable and cost-effective tool in understanding during-fracturing and post-flowback well and section performance. During-fracturing diagnostics, such as crosswell strain measurements from an offset monitor well, describe where fractures intersect the monitor well and reveal details such as azimuth, fracture-propagation velocity, and activation and reactivation statistics. These critical data points help inform cluster, stage, and well-spacing decision-making. While the absence of observed strain activity does not indicate the absence of created fractures, the existence of a strain activity indicating a fracture does not guarantee hydraulic functionality. The hydraulically functional fracture network can be understood with post-flowback production profiling. Particularly when combined with interference testing, fiber-optic production profiling identifies and differentiates productive and unproductive sections of unconventional laterals. Combination of during-fracturing and post-flowback technologies reveals the clearest picture of the functional fracture network. On an eight-well pad in the Permian Basin, the operator acquired fiber-optic crosswell strain data on two wells using single-use fiber, revealing details about the created fracture network. Sixty days after the wells went on production, a combination production-allocation and interference sensing program was executed with intervention fiber optics on two of the wells. Designing a Diagnostic Project in the Delaware Basin The operator completed an eight-well pad in the Delaware Basin on the western slope of the Permian. The wells were landed in three rows across the upper and lower Wolfcamp, which had a formation thickness of approximately 600 ft true vertical depth. Four of the laterals were completed from south to north (Wells 11N, 14N, 21N, and 31N), and five of the laterals were completed north to south (Wells 12S, 13S, 22S, 23S, and 31S). Two of the north-to-south wells, one in the middle (23S) and one in the lower bench (31S), were designated as monitor wells for during-fracturing diagnostic data acquisition (Fig. 1). Zipper operations were conducted in three groups, with the laterals on the outside edges of the pad fractured first (11N, 14N, and 21N) in a south-to-north order, followed by the inner laterals in the second zipper group (12S, 13S, and 22S) in a north-to-south order.

Characterize fracture geometry and propagation with low frequency distributed acoustic sensing strain data

Low-frequency distributed acoustic sensing (LF-DAS) strain data from offset wells are widely used to characterize hydraulic fracture geometry. In this paper, we propose a new optimization model for DAS interpretation. The model quantifies the fracture propagation trajectory and fracture geometry variation with time from low-frequency DAS strain data of offset wells. First, the optimization model is developed by combining the Green function and the Sneddon fracture width function to correlate parameters such as fracture width distribution, fracture length and fracture location with the far-field strain distribution. Second, the optimization model is solved by the Levenberg-Marquardt nonlinear least squares algorithm. The model was validated by generating cases. In addition, the model was applied to the interpretation of fracture propagation characteristics for the fiber optic well case of HFTS-2 offset wells. The interpretation results can visually determine the fracture modification effect and help to improve the production efficiency.

Fiber-Optics Approach Monitors Drainage Profile of Multistage Stimulation

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 212364, “Can You Feel the Pressure? Strain-Based Pressure Estimates,” by Kyle Haustveit, SPE, and Jackson Haffener, SPE, Devon Energy. The paper has not been peer reviewed. _ The combination of horizontal wells and hydraulic stimulation has been key in unlocking vast unconventional resources across North America and elsewhere, but much remains to be understood regarding how the reservoir is drained across the miles of laterals accessing a resource. The density of pressure gauges required to accurately measure the drainage pattern from a horizontal multistage stimulation is currently not realistic economically or technologically. In this case study, the authors describe a method of monitoring the drainage profile of a horizontal multistage stimulation using optical fiber. Introduction An advanced subsurface diagnostic project was deployed as part of the Hydraulic Fracture Test Site I, Phase 3, funded by the Department of Energy. A lateral observation well was positioned approximately 225 ft away from the nearest two stimulated and producing horizontal wells. Using optical fiber and Raleigh frequency-shift distributed strain sensing (RFS-DSS), strain change was measured in the far field as the offset well was stimulated and first produced. RFS-DSS provides a spatial resolution of 20 cm, allowing for the monitoring of strain changes at finer spatial resolution than can be sensed accurately with low-frequency distributed acoustic sensing. The strain change during the interference test correlated with both the strain change during the crosswell refracturing monitoring and the pressure-gauge change along the monitor well during the interference test. The strong correlation between crosswell refracturing and interference strain monitoring suggests that the magnitude of crosswell strain responses is directly proportional to the intensity of strain change when the offsetting wells are put online. The comparison of data recorded from RFS-DSS and six reservoir-sensing pressure gauges displayed a strong correlation between the total strain change and total pressure drawdown. Three gauges positioned in regions of large negative strain change during first production, interpreted as drawdown, showed large pressure declines; the other three gauges, positioned in areas of small strain change, saw small pressure declines. The correlation was applied to the total length of the wellbore based on the strain change to estimate a pressure along the entire lateral, including areas without pressure gauges. Pilot Design Located in DeWitt County, Texas, the project consisted of a combination of production infill wells (12H and 13H), pre-existing primary wells (2H, 4H, 6H, and 8H), liner refracturing wells (3H and 5H) and one unstimulated observation well (14H). The wells are landed in the lower Eagle Ford formation. While drilling the observation lateral, horizontal core was collected and used to investigate the frequency of hydraulic and propped fractures created from the 4H and 5H original stimulations. The 14H lateral was equipped with permanent optical fiber, one internal pressure-sensing gauge used for sealed wellbore pressure monitoring, and nine reservoir-sensing pressure gauges to monitor fracture interactions and pressure drawdown from the offset wells. Three of the gauges failed to show connection to the reservoir and were discarded for this analysis.

Comments: Unconventional Shale Reservoirs Meet Cancer Genome Sequencing

The cross-industry development and application of technologies is the epitome of innovation and ingenuity. The more divergent the industries are, the more the unique ways of thinking strike us as remarkable. Technology transfer between the oil and gas industry and other sectors isn’t new. Many examples can be found in medical science, space exploration, and renewable and sustainable technologies. In this JPT issue, the technologies of Oceaneering Space Systems and Impossible Sensing Energy are featured. A case study presented at the 2023 SPE Annual Technical Conference and Exhibition took medical science to a new low—into the depths of the Permian Basin for downhole reservoir drainage diagnostics (RDD). The coauthors, a subsurface and wells manager and a chapter manager, surveillance, analysis, and optimization and pilots at Chevron Technology Ventures (CTV), described using genome sequencing for cancer detection and treatment in a successful proof-of-concept (POC) test conducted from 2017 to 2019. Nampetch Yamali and Daniel Emery explained in SPE 215052 how in 2017 CTV was given the task of identifying technology to assess vertical drainage dimensions to estimate drainage volumes from fractured wells. The goal was to minimize interference from offset wells and co-developments to optimize reservoir recovery in unconventional fields. At that time, the technologies and their capabilities were limited. Although subsurface microbial DNA sequencing for upstream assets has found applications in enhanced biocide and corrosion inhibition, reservoir sweet spot indicators, and tracer technology, use of the technique was thought also to have potential in horizontal well development planning via RDD. Learning of a genomic sequencing technique for cancer cell identification in humans at the biology and biochemistry department at the University of Houston, the wheels started turning at CTV with thoughts it could be developed and adopted for RDD. Similarities were seen because RDD uses in-situ subsurface microbial DNA to infer the depths from which fluids drain after fracturing a horizontal well and optimizes well spacing on a pad, especially in stacked reservoirs. The predictive analytics platform mapped the hydrocarbon footprints of geologic subzones by using noninvasive DNA testing tools. The unique DNA markers for RDD were extracted from mud and cuttings from the horizontal and vertical drilling sections. DNA from produced fluids was also collected. The authors wrote, “Produced fluids collected at two-to-four-week intervals at the wellhead are mapped to the RDD framework. They are ‘lineage traced’ back to shale and tight rock. Drainage heights and percentage of contribution from these locations is computed adapting DNA sequencing and data analytics pipelines which were developed for lineage tracing cancer cells, breaking off from the primary tumor and metastasizing to distant sites back to the tissue of origin.” After validating the DNA extractions, CTV conducted a blind test using only the DNA data to determine the landing zones for each of the four wells in the pad. Using clustering methods, the results of the blind test accurately identified all landing zones within an acceptable margin, according to Yamali and Emery. The next step in the trial analyzed the DNA of the produced fluids. “Preliminary results showed drainage height estimates for each horizontal well.” In 2021, the study continued with an additional four new field trials in the Permian Basin for application of the technique to zonal production identification. The authors wrote, “Although the results from the POC were positive, subsurface DNA RDD technology is a novel technique within the oil and gas industry, and advanced applications are still evolving. This combination naturally gives rise to more general skepticism and a need to thoroughly evaluate the technology from operational, economic, and accuracy perspectives.” The paper describes the ongoing cross-disciplinary work being done by petroleum engineers, geologists, microbiologists, and stratigraphers to fully evaluate this method of RDD. It further details CTV’s wider approach to “accelerate the innovation life cycle, from pilot stages to widespread adoption at scale,” emphasizing the need for leaders’ support for nonconventional thinking and approaches. As the close of 2023 approaches, I’d like to take this opportunity to wish you the best in 2024 on behalf of the JPT Editorial Review Board and the JPT staff. We’ll be kicking off the new year with the commemoration of JPT’s 75th anniversary. Each issue will include an article dedicated to the evolution of technology and industry practices over the seven and a half decades JPT has covered the upstream industry. In January, Trent Jacobs and Stephen Rassenfoss will revisit technological advancements since JPT’s 50th anniversary in 1999. Since then, over 25 million B/D have been added to the global supply thanks in very large part to advancements in drilling, completions, and reservoir technologies. Join us as we explore these along with some predictions made 25 years ago and the surprising realities of where the industry stands today.

Maximising offset well information in unravelling onshore geohazards indicators: Case study of the Gale field

An attempt is made in this study towards maximising offset wells information in unravelling onshore geohazards indicators in the Gale field. The Gale field is located about 100km north-west of Port Harcourt, Nigeria. The field consists of a highly faulted and elongated rollover anticline, bounded to the north by a regional growth fault. The data used for this study integrates the quadrature and reflectivity amplitude attributes from seismic data, with offset well data. The conventional reflectivity seismic data was 90° phase rotated to derive the quadrature volume. The quadrature seismic was considered a more appropriate reflectivity seismic attribute for use in shallow geohazard analysis as it is known for its characteristic preservation of high frequency spectrum inherent in the data. Offset wells (GALE-01, GALE-03, GALE-04, GALE-05, GALE-06, and GALE-08) analysis revealed mud losses, stuck pipe, overpull and gas cut as gathered from the daily drilling reports. These could translate to potential triggers of some geohazards where poorly managed. Review of field geotechnical report did not reveal any geohazards issues. Based on the geohazards assessment carried out for these wells; chances of encountering shallow gas for all the units as shown in the well summary is rated low. Results from a geohazards analysis indicate the presence of possible shallow gas within the area of interest and particularly along the shallow section of planned well trajectory. This is further supported by the presence of faults within the vicinity of gas bearing reservoirs at deeper level and a potential for these faults extending to the shallower interval. These faults are likely to serve as migration pathways for gas to seep to the shallower section, hence forming a potential geohazard. In addition, some of the offset wells targeting deeper gas reservoirs penetrated pockets of gas at the shallower interval that stratigraphically correlates with the shallow section that would be penetrated by the planned wells. The results of this work were used to move the proposed drilling location of the Gale planned wells to a nearby area free of shallow gas signatures.

Real-Time Gamma Ray Log Generation from Drilling Parameters of Offset Wells Using Physics-Informed Machine Learning

Summary By 2026, USD 5.05 billion will be spent per year on logging while drilling (LWD) according to the market report from Fortune Business Insights (2020). Logging tools and wireline tools are costly services for operators to pay for, and the companies providing the services also have a high cost of service delivery. They are, however, an essential service for drilling wells efficiently. The ability to computationally generate logs in real time using known relationships between the rock formations and drilling parameters, namely, rate of penetration (ROP), rotations per minute (RPM), surface weight on bit (SWOB), surface torque (TQX), standpipe pressure (SPPA), and hookload (HKLD), provides an alternative method to generate formation evaluation information (analysis of the subsurface formation characteristics such as lithology, porosity, permeability, and saturation). This paper describes an approach to creating a digital formation evaluation log generator using a novel physics-informed machine learning (PIML) approach that combines physics-based approaches with machine learning (ML) algorithms. The designed approach consists of blocks that calculate mechanical specific energy (MSE), physical estimates of gamma ray (GR) using physical and empirical models, and formation information. All this information and the drilling parameters are used to build a classification model to predict the formations, followed by formation-based regression models to get the final estimate of GR log. The designed PIML approach learns the relationships between drilling parameters and the GR logs using the data from the offset wells. The decomposition of the model into multiple stages enables the model to learn the relationship between drilling parameters data and formation evaluation data. It makes it easier for the model to generate GR measurements consistent with the rock formations of the subject well being drilled. Because the computationally generated GR by the model is not just dependent on the relationships between drilling parameters and GR logs, this model is also generalizable and capable of being deployed into the application with only retraining on the offset wells and no change in the model structure or complexity. For this paper, the drilling of the horizontal section will not be discussed, as this was done as a separate body of work. Historically collected data from the US Land Permian Basin wells are used as the primary data set for this work. Results from the experiments based on the data collected from five different wells have been presented. Leave-one-out validation for each of the wells was performed. In the leave-one-out validation process, four of the wells represent the set of offset wells and the remaining one becomes the subject well. The same process is repeated for each of the wells as they are in turn defined as a subject well. Results show that the framework can infer and generate logs such as GR logs in real time. The average root-mean-squared error (RMSE) observed from the experiments is 27.25 api, representing about 10% error. This error value is calculated based on the mean estimate and does not consider the predicted confidence interval. Considering the confidence interval helps further reduce the error margin.

Integration of Seismic Poisson Impedance Data with Real-Time Geomechanics and Real-Time 3D Ultradeep Resistivity Inversion Enabled New Opportunities in Developed Field: Case Study from Kuwait

Summary 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, referenced as W7–W1. Three of these sequences (W6, W5, and W3) were expected in the case study well. The objective was to set a 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. 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 (PI). Historically, targeting the Wara formation occasionally resulted in multiple sidetracks due to drilling challenges. A real-time geomechanics service was used to overcome drilling challenges, and real-time 3D ultradeep resistivity inversion was implemented to optimize the well placement. An extensive predrilling study for geomechanical and ultradeep resistivity inversion modeling helped to develop a road map for an optimal and safe well-construction process. The study showed that use of real-time 3D ultradeep azimuthal resistivity (UDAR) inversion would help to optimize the well placement and maximize the 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 used 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 ultradeep 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 PI analysis. Formation evaluation of the lateral section showed an average porosity of 24 p.u., water saturation of 11%, and up to 3 darcies/cp mobility. The application of real-time 3D ultradeep resistivity inversion helped to triple the planned formation exposure and 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 in the field. The integrated approach described above was recommended to be utilized for all future Wara wells.

Combining Magnetic and Gyroscopic Surveys Provides the Best Possible Accuracy

Summary A survey program is designed for every well drilled to meet the well objective of penetrating the target reservoir and avoiding a collision with nearby offset wells. The selection of the wellbore survey tools within the survey program is limited in number and accuracy by the current surveying technologies available in the industry. This article demonstrates how a higher level of accuracy can be achieved to meet challenging well objectives when the accuracy of the most accurate wellbore surveying tools and technologies taken individually is insufficient. This high level of wellbore positioning accuracy is achieved by combining two independent wellbore positions of the same wellbore trajectory. The first wellbore position is calculated using the latest technology of magnetic measurement-while-drilling (MWD) definitive dynamic surveys (DDS). The accuracy of the MWD DDS can be further improved by minimizing error sources such as misalignment of the survey package from the borehole, drillstring magnetic interference, the use of localized geomagnetic reference, using high-accuracy accelerometer sensors, and a high-accuracy gravity reference. Furthermore, the MWD DDS inclination accuracy is improved using an independent inclination measurement from the rotary steerable system. A first wellbore position is calculated from the magnetic MWD DDS after applying in-field referencing (IFR), multistation analysis (MSA), bottomhole assembly (BHA), sag correction (SAG), and dual-inclination (DI) corrections to improve both azimuth and inclination accuracy. A second wellbore position is calculated using gyro-MWD (GWD) technology. The results and comparisons of multiple combined survey runs are presented. The highest accuracy of wellbore positioning had been proved in this successful case study by penetrating a very small reservoir target on an extended-reach well that was unfeasible using either the most accurate enhanced MWD DDS or GWD technology individually. The presented case study shows how the wellbore objectives of penetrating a very small reservoir target had been confirmed by logging-while-drilling images and the reservoir mapping interpretation of the client subsurface team. This gave a high-accuracy wellbore position during drilling and provided higher confidence in wellbore placement to maximize reservoir production without colliding with nearby offset wells. Wellbore survey accuracy limits a borehole’s lateral and true vertical depth (TVD) spacing, constraining reservoir production in those sections. In the top and intermediate sections, wellbore survey accuracy limits how close the wellbore can be drilled to other offset wells due to collision concerns. This directly impacts the complexity of the directional work and the cost per section. Combining independent wellbore surveys unlocks the potential to improve the wellbore positioning accuracy significantly. It demonstrates the highest wellbore positioning accuracy that can be achieved to date compared with the latest magnetic MWD surveys after correcting all known errors or compared with GWD.

Machine Learning, Numerical Simulation Integrated To Estimate Child-Well Depletion

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper URTeC 3719366, “Integration of Machine Learning and Numerical Simulation To Estimate Child-Well Depletion,” by Edward Wolfram, SPE, James Cassanelli, and Soodabeh Esmaili, SPE, Occidental Petroleum, et al. The paper has not been peer reviewed. _ In the complete paper, the authors analyzed a robust, well-distributed parent/child well data set of the Delaware Basin Wolfcamp formation using a combination of available empirical data and numerical simulation outputs, which was used to develop a predictive machine-learning model (consisting of a multiple linear regression model and a simple neural network). This model has been implemented successfully in field developments to optimize child-well placement and has enabled improvements in performance predictions and net present value. Introduction Pervasive parent/child well pairs have complicated the development of the Delaware Basin Wolfcamp formation by introducing the need to forecast child-well performance reliably. This problem is made more difficult by the complex nature of the physical processes involved in parent/child well interactions and the variety of geometrical configurations that can be realized. In broad terms, the following three classifications of child wells can be recognized based on their spatial relationship to the associated parent well and other offset wells (Fig. 1): - Type 1—Child wells completed in between two parent wells - Type 2—Child wells completed adjacent to a parent well along with concurrently completed or codeveloped infill wells - Type 3—Child wells completed adjacent to a single parent well To narrow the range of complexities in the study, the authors focused on Type 2 child wells because this configuration will be used most often in future development activities and because it had the most existing field examples. The principal objective of this assessment was to generate accurate quantitative predictions of the diminished production performance of child wells because of pre-existing parent wells. In this work, a novel, hybrid approach is detailed involving a combination of machine-learning techniques and numerical simulations. This hybrid approach is an attempt to overcome the individual limitations of machine-learning techniques, which need abundant empirical data, and numerical simulations, which are too costly to simulate the full range of conditions of interest. By combining these two methodologies, the final model can leverage the individual strengths of machine-learning techniques (extracting information from the available empirical data) and numerical simulations (providing rigorous, theory-based physical modeling). To facilitate the approach, the authors constructed a robust, well-distributed parent/child well data set using both available empirical data and numerical simulation outputs, which were used to develop a predictive machine-learning model. Much of the complete paper is dedicated to the critical data-collection and -cleaning process, including parent/child well identification and data-set construction, response variable definition, and selection of features and predictor variables. These topics are not included in this synopsis.

Reservoir and Fracture Characterization for Enhanced Geothermal Systems: A Case Study Using Multifractured Wells at the Utah Frontier Observatory for Research in Geothermal Energy Site

Summary For enhanced geothermal systems (EGS), multistage hydraulic fracturing along a deviated or horizontal well is a key technology used to create a high-conductivity fracture network between injection and production wells in deep, low-permeability geothermal reservoirs. The purpose of the created fracture network is to allow for the efficient transfer of fluid, heated by the geothermal reservoir, from the injection to the production well; therefore, well spacing (between injection and production wells) and hydraulic fracturing must be designed not only to promote connectivity between well pairs but also to mitigate thermal short-circuiting and thermal breakthrough. Analysis of post-fracture pressure decay (PFPD) data measured after each stage of a hydraulic fracturing treatment can be used to provide critical reservoir and fracture parameters required for well and hydraulic fracturing design optimization; this method provides a low-cost alternative and complementary approach to in-situ observation techniques, such as core-through experiments, fiber optics, or image logs in offset wells. Until now, PFPD has primarily been applied to multifractured horizontal wells (MFHWs) completed in low-permeability hydrocarbon reservoirs. The goal of this study is therefore to develop a methodology to estimate fracture and reservoir parameters using stage-by-stage PFPD data associated with EGS projects. An analytical model is proposed herein to estimate fracturing fluid efficiency, fracture length, average fracture aperture, average fracture conductivity, and reservoir permeability for different possible fracture geometries in EGS reservoirs. PFPD data collected for three hydraulic fracture stages in the injection well at the Utah Frontier Observatory for Research in Geothermal Energy (FORGE) site were analyzed to demonstrate the practical application of the proposed method. The results of this study indicate that, due to the presence of natural fractures in the target (granitic) reservoir, the hydraulic fracturing treatment (using slickwater) in the openhole section resulted in low fracturing fluid efficiency and small hydraulic fractures. In contrast, hydraulic fracturing treatments conducted in the perforated casedhole wellbore resulted in higher fracturing fluid efficiency and created larger hydraulic fractures even with smaller injected volumes. The results of the PFPD analysis were confirmed using a Formation MicroScanner image log and microseismic data collected during each stage of hydraulic fracturing.

Multidisciplinary Integrated Approach Unlocks Fracture-Characterization Ambiguity

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 213771, “Unlocking Fracture-Characterization Ambiguity: A Multidisciplinary Integrated Approach,” by Ali Abughneej, Mohamed Al-Naqeeb, and Mohammed Al-Ostath, SPE, Kuwait Oil Company, et al. The paper has not been peer reviewed. _ Fracture systems play a significant role in production in the case of tight, low-porosity complex reservoirs. An exploration well in North Kuwait that targeted the Middle Marrat formation encountered uncertainty regarding reservoir compartments in an underexplored field, with minimal offset wells to properly evaluate its potential. A multidomain approach is proposed leveraging different types of expertise to tackle the petrophysical and geomechanical aspects of the fracture system governing the carbonate layer and assess its producibility and qualification for stimulation through hydraulic fracturing. Introduction A comprehensive study was performed on Well A targeting the Marrat formation, a tight, low-permeability carbonate Jurassic reservoir. During the last two decades, the operator has drilled many vertical and deviated wells through this formation. Hydrocarbon production from the closely spaced wells shows a major variation in flow rates, which has been attributed to the presence of uncharacterized fracture clusters as well as the limited understanding of reservoir compartmentalization. To address the challenges of production discrepancy, an innovative multidomain approach was proposed as an integrated solution in an attempt to understand the fracture systems governing the reservoir setting and its compartments and further assess their connectivity and extent to form a perspective on their possible contribution or hindrance to flow. An end-to-end comprehensive work flow was created combining the results of advanced formation evaluation, high-resolution borehole imaging, and state-of-the-art sonic imaging [the 3D far-field sonic (3DFFS) approach], followed by a fracture-stability analysis for the purposes of completion optimization and potential future hydraulic fracture design. Geological Setting Middle Marrat is targeted as the hydrocarbon-bearing zone, consisting of dolomitized oolitic grainstone to packstone facies, reflecting the high-energy inner-midramp shoal depositional environment.

After Years of Ever-Longer Fracturing Stages, Some Engineers Say Shorter Can Be Better

The trend in fracturing designs has been longer stages with more perforation clusters, which save time and money. Based on papers at this year’s SPE Hydraulic Fracturing Technology Conference and Exhibition, the thinking that has sharply reduced the number of perforations per cluster and improved the effectiveness of fracturing is becoming the industry norm. But there are a couple of companies questioning the consensus by fracturing wells with shorter stages and fewer clusters to see if they deliver more oil and gas production. ConocoPhillips has been considering “going backward to fewer clusters per stage,” said Dave Cramer, a senior engineering fellow at the company. “In some areas, we are testing out fewer clusters per stage based on fiber-based observations in offset wells that (indicate) far-field treatment uniformity is improved as a result,” he said, adding “another advantage of reducing stage length is that injection rate into the hydraulic fractures is increased, which leads to increased fracture width and improved proppant transport.” The offset-well observations mentioned were in a recent paper by Devon Energy that reported on a test which concluded that stages with fewer clusters were more efficiently fractured than longer stages with more of them (SPE 212340). The 36-page paper was based on extensive testing at Phase 3 of the Hydraulic Fracturing Test Site 1 in the Eagle Ford, where a well fractured using a wide range of clusters per stage was used to find a way to most effectively refracture productive rock missed by an old fracture design. The data were shared with companies backing the private-public partnership with the US Department of Energy, which included ConocoPhillips. The paper’s authors wrote, “Generally, cluster efficiencies are higher for stage designs with fewer clusters.” Cluster efficiency depends on the entry holes in a cluster getting enough fluid at a high enough rate to create a productive fracture. Fracture length is the result of other choices, including the pump rate, the entry-hole numbers, diameter, and placement. The thinking behind these designs—as with most fracturing nowadays—is based on the limited-entry method. This technique ensures the pumping rate and fluid volumes are enough to properly stimulate all the perforations, which are sized and placed to ensure all the entry holes have a chance of developing a fracture. The industry’s focus on equalizing treatment goes back to early studies using fiber optics that revealed that the first clusters passed nearest the heel side of the lateral were taking in the lion’s share of the fluid and developing dominant fractures, leaving many later clusters under stimulated. Based on recent papers from Devon and Hess, limited-entry ensures most clusters are stimulated, but dominant clusters are still getting more than their share because the high flow rates that favor them with more fluid also ensure they erode faster, allowing them to take in more fluid.

Multi-stage hydraulic fracture monitoring at the lab scale

We introduce a novel experimental and modeling approach to perform lab-scale multistage hydraulic fracturing in Poly-Methyl-Meth-Acrylate (PMMA) and study pressure-based methods to monitor hydraulic fracturing. The approach combines an internal circular V-notch method for fracture initiation with a method for creating hydraulic isolation between subsequent stages. Using high-resolution models of fracturing experiments with injection and monitoring wells, we present novel results of displacement, pressure, and stress evolution around the wells. We identify injection-induced rotational deformation in the domain and asymmetric growth of the hydraulic fracture. We quantify the role of the monitoring well on the rotational deformation and the fracture asymmetry. We analyze the role of undrained deformation in generating the monitor well’s pressure response to individual fracture stages at the injection well. Our results show that lab-scale multistage hydraulic fracturing can be a useful tool to estimate the geometry and property of a hydraulic fracture using pressure data from offset wells in the field.