Unconventional tight reservoirs currently make up more than 60% of domestic oil and gas production in the United States. However, developing unconventional formations requires intensive drilling and completion campaigns to maintain steady production of a field. Therefore, the prediction of estimated ultimate recovery, which measures the producible reserve from a well, is demanding, particularly as operators becomes more rational under the current volatile market conditions. Despite unconventional reservoirs being considered a resource play with low geological risks, their economic appraisal is challenged by unknown stimulation outcomes and intricate producing mechanisms. Therefore, this work aimed to leverage machine-learning techniques with big data to analyze the multivariant relationship of geological and engineering parameters with unconventional reservoir production and to improve the prediction of estimated ultimate recovery in unconventional formations. In this case study, a multiscale machine-learning workflow was deliberated and applied to a big data set from the Eagle Ford shale. First, quality control and feature selection were performed on a data set consisting of 4,067 wells with 30+ geophysical, petrophysical, drilling and completion, and production features. Then, a regional inferencing model, based on a K -nearest neighbor with bagging algorithm, was trained to obtain the spatial trend of estimated ultimate recovery across the Eagle Ford formation. The last part of the analysis was to build a local-scale prediction model. With the study area confined to East Texas, a random forest regression was performed to rigorously predict oil and gas estimated ultimate recoveries. The selected training features were finalized based on the results of a higher-dimension regression, as well as domain knowledge. Overall, the data-driven model trained with physically controlled data captured the production behavior of the Eagle Ford shale. The application of the proposed workflow on the Eagle Ford shale demonstrates a progressive building of the machine-learning model. The quality control of data allows global inspection of the data set and, more importantly, confirms the statistical distribution of training data. This study emphasizes the philosophy of multiscale data analytics. The large-scale model portraying sweet spots using location variables grants direct guidance for acreage acquisition and development across the basin; the small-scale model trained with reduced dimensionality generates quantitative prediction of oil and gas estimated ultimate recoveries for an area of interest. Compared with our previous work using higher dimensionality and extensive spatial interest, this progressive learning maintains similar explained variance in out-bag model check, but grants 26% and 52% reductions in mean square error for predicting oil and gas estimated ultimate recoveries in the Eagle Ford formation. In the end, prediction validation is performed by revisiting the data set. Overall, the proposed workflow demonstrates successful application in the Eagle Ford formation such that it can be directly implemented for other unconventional resource plays.

Lost circulation (LC) is extremely costly to many drilling operations due to loss of mud and productive time. Significant industry resources have been devoted to providing recommendations for both preventing and curing losses. These comprise bridging particles, settable pills and a focus on the better understanding, measurement and control of downhole pressures. The industry has also started to examine the relationship between downhole losses and fracture geometry. However, there has been very limited verification and utilization of such effort. Knowledge of the loss mechanism(s), location of the thief zone and fracture geometry is fundamental for understanding and treating losses effectively. The proposed method introduces a systematic approach for diagnosing lost circulation into potential root cause(s) by utilizing up to twenty-two diagnostic indicators. This allows the most appropriate treatments to be mapped to the suspected root cause. This technique supports the creation of preventative strategies during well planning. For example, if induced fractures are identified as the most likely root cause in the planning stage, then a wellbore strengthening formulation based on the fracture width estimation is recommended. A proper treatment can be designed based on the fracture width analysis as the data becomes invaluable in determining optimum sealing and effective treatment for losses into conductive natural fractures/faults. In case of loss events, the fracture width can be estimated based upon loss rate, the mud rheology, bottom hole pressure and pore pressure. Field experiences demonstrate that natural and induced fractures are the two most common root causes of losses. For induced fractures, fracture width usually develops beyond 3000-micron (µm) once losses have occurred and the preventive lost circulation materials (LCM) have not then worked effectively. If the bottom hole pressures can be kept below fracture propagation limits (or far-field minimum horizontal stress), then induced fractures will not cause significant volume of losses. The drilling induced fractures can be modelled by using in-situ stresses, well trajectory, bottom hole pressure, geothermal and mud temperature information.

In the age of artificial intelligence, digitalization, rising energy demand, falling prices of barrel of oil and increasing difficulty in oil & gas recovery we need to have an integrated approach based on physics, artificial intelligence and rock mechanics to reduce the non-productive time in drilling and -in parallel- enhance well production. The integrated approach should help in reducing cost, minimize human intervention, reduce drilling associated risks, minimize the negative impact on near wellbore rock behavior due to stimulation and enhance the recovery of hydrocarbons. Stuck pipe is a major stake holder in "non-productive time" and is estimated to cost the oil and gas industry around $250 to $300 Million a year. Stuck pipe due to wellbore stability issues is a regular phenomenon while drilling weak zones in minimum stress direction especially in Middle East. William Lyon's in 2010, estimated that cost of stuck pipe in deep oil and gas wells is around 25% of overall budget. To counter stuck pipe, for instance, drilling engineer may decide to increase the mud weight inorder to minimize the wellbore stability issues, and this could enhance challenges to a stimulation engineer associated with potential damage. Simliarly, an improper acidizing could soften the rock and negatively impact mechanical response of near wellbore rock during production. Two simple examples demonstrate the value of an engineering holistic approach based on wellbore stability integration into hydraulic fracturing treatment design considering the complexity involved around geomechanics. This study introduces a workflow that holistically integrates a rock mechanics approach to optimize drilling performance and characterize the stresses around the wellbore with the completion design, combining the geomechanical and petrophysical properties to optimize the completion and stimulation design. This engineering workflow will enable to design and customize a particulate diverter system for effective fluid diversion and wellbore coverage by uniformly distributing the stimulation fluid with an aim to create fracture network complexities, enhancing the production. Additionally, this paper showcases the learnings from various field case histories including but not limited to drilling across weak bedding plane from Asia, wellbore stability issues in Middle East that resulted in high non-productive time from drilling, Uniform Fracture Growth from Horizontal Wells and re-fracturing strategies from North and South America. This approach will enable optimizing well performance from drilling to production, minimize risks and optimize intervention by retro alimenting each phase of the process to the next. This workflow provides a innovative strategic approach optimizing drilling, completion and stimulation mitigating challenges in unconventional formations that can be extrapolated to conventional reservoirs as well.

Reliable characterization of mechanical behavior (e.g., elastic properties) in anisotropic and heterogeneous formations require advanced methods for understanding the impacts of spatial distribution of rock components, pore structure, and pore pressure on mechanical properties. However, the existing methods for assessment of mechanical properties (e.g., effective elastic properties) such as effective medium models, assume constant stiffness values and idealized shapes for rock constituents and pores. These models also do not take into account coupled hydraulic and mechanical (HM) processes, which cause significant uncertainties in geomechanical evaluation. The objective of this paper is to investigate the effects of realistic spatial distribution of minerals, pore pressure, and pore structure on the effective elastic properties of rock-fluid systems. In order to pursue this objective, we developed a pore-scale numerical simulator by satisfying conservation equations and considering the coupling among relevant HM processes. We adopted peridynamic theory to discretize the micro-/nano-scale medium. The inputs to the numerical modeling include pore-scale images of rock samples as well as mechanical and hydraulic properties of each rock constituent. We used micro-computed tomography (micro-CT) scan and focused ion beam (FIB) scanning electron microscope (SEM) images of rock samples to obtain a realistic micro-/nano-scale structure of both rock matrix and pore space. We then assigned realistic mechanical and hydraulic properties to each rock constituent within the pore-scale medium. The outcomes of numerical modeling include the variation of effective stress and the evolution of corresponding strain by honoring the variability in mechanical properties of rock components caused by their spatial distribution, size, pore pressure, and pore structure at the micro-/nano-scale level. We successfully tested the reliability of the developed framework using results of an analytical solution for the case of consolidation. We then performed sensitivity analyses to quantify the effects of concentration and spatial distribution of rock components, divergence in mechanical properties of minerals, and pore structure on variations in effective elastic properties of rock components. For instance, the deformation of clay minerals dispersed in between the quartz minerals was approximately 60% less than that in clay minerals colonized next to the quartz under the same load. In the next step, we compared these mechanical characterizations with estimates obtained from the effective medium models. We observed measurable uncertainties (more than 15% depending on mineral content and distribution) in elastic properties of rock components estimated by the effective medium models such as self-consistent approximation. These uncertainties are associated with spatial distribution, shape, and size of minerals, which are not considered in those models. Such effective medium models also overlook the effects of pore structure and pore pressure on the mechanical properties. The results of coupled HM analysis for cases with the same mineral concentration but different pore structure revealed more than 12% error in estimates of effective mechanical properties.

Since 2010, hydrocarbon production from long horizontal wells targeting shales has become the norm for industry leaders. Because of the steep decline rates, it is vital to understand the reservoir and its properties before going through with a full-scale fracture stimulation. Through the application of Diagnostic Fracture Injection Tests (DFIT), one can determine accurate estimates of closure pressure, net pressure, pore pressure, formation permeability, and induced fracture geometry. The Utica shale is among the most promising reservoirs of the future, but there is limited information available discussing its properties. In SPE-196149-MS, we analyzed a DFIT from one horizontal well targeting the Utica. However, in order to fully understand the Utica shale at scale, further analysis is required. In this study, we will present three additional horizontal wells targeting the Utica, and analyze the pressure and its derivative to accurately estimate the properties mentioned above. DFIT analysis is an advanced technique to accurately predict stress regimes and reservoir properties. However, interpretation of DFIT data is challenging, especially in shale formations. In this study, we overview the geologic properties of the Utica shale, discuss the development of DFIT analysis and its governing equations, then present the three data sets and resulting conclusions. We specifically discuss the Tangent Line Method, the Compliance Method, and the Variable Compliance method in detail, while comparing their underlying equations and assumptions to determine closure pressure. After-Closure analysis is then performed in order to verify fracture closure and identify flow regimes. Through linear regression of this data, pore pressure from a linear flow regime is extrapolated, and through numerical simulation, key reservoir properties, such as permeability and fracture geometry, are estimated for the Utica shale. The DFIT interpretation and simulation results from this study are very insightful. Interpreting the GdP/dG function, the closure pressure ranges from 4,943 psi to 6,141 psi, contributing to a closure pressure gradient of 0.797 to 0.891 psi/ft for the Utica shale. Based on the pressure transient analysis, the pore pressure ranges from 3,238 psi to 4,064 psi, contributing to a pore pressure gradient of 0.486 to 0.616 psi/ft for the Utica shale. Additionally, field wide ranges of reservoirs properties are presented, allowing industry to further optimize their drilling and fracing techniques in the Utica shale. Two of the wells in this study are close in proximity and show very similar results both in After-Closure analysis and in pressure response curves. The third well displays a different GdP/dG response, leak off characteristics, pressure transient behaviors, formation permeability, and fracture geometry. This variance in results can be attributed to regional differences in geology, stresses, and pressures. Therefore, operators need to consider regional differences in reservoir properties in order to enhance the development of Utica shale and unlock all potential recoverable hydrocarbons.

This study demonstrates the application of an interpretable (or explainable) machine learning workflow using surface drilling data to identify fracable, brittle and productive rock intervals along horizontal laterals in the Marcellus shale. The results are supported by a thorough model-agnostic interpretation of the input-output relationships to make the model explainable to users. The methodology described here can easily be generalized to real-time processing of surface drilling data for optimal landing of laterals, placing of fracture stages, optimizing production and minimizing frac hits. In practice, this information is rarely available in real-time and requires tedious and time-consuming processing of logs (including image logs), core, microseismic data and fiber optic sensor data to provide post-job validation of frac- and well-placement. Post-completion analyses are generally too late for corrective action leading to wells with a low probability of success and increasing risk of frac hits. Our workflow involves identifying geomechanical facies from core- and well-log data. We verify that the geomechanical facies derived using core- and well-log data have characteristically different brittleness, fracability and production characteristics. We test and investigate several different supervised classifiers to relate surface drilling data to the geomechanical facies. The data was divided into training and test datasets, with supervised classification techniques being able to accurately predict the geomechanical facies with 75% accuracy on the test dataset. The clusters predicted on test well (unseen data) were qualitatively verified using the microseismic interpretation. The use of Shapley Additive Explanations (SHAP) helps explain the predictive models, rank the importance of various inputs in the prediction of the facies and provides both local and global sensitivities. Our study demonstrates that pre-existing natural fracture networks control both the hydraulic fracture geometry as wells as the production. Natural fractures promote the formation of complex fracture networks with shorter half-lengths which increase well productivity while minimizing frac hits and neighboring well interactions. The natural fracture network is itself controlled by the geomechanical properties of the rock. The ability of the surface drilling data to reliably predict the geomechanical rock facies provides a powerful tool for real-time optimization of wellbore trajectory and completions.

Mini frac, or diagnostic fracture injection test (DFIT), is a short hydraulic fracturing test that provides formation break-down pressure, minimum horizontal stress, and reliable value for formation permeability of shale reservoirs. The calculated formation permeability is particularly more reliable from the analysis of the second cycle of the mini frac test because the fracture is already created. In this paper, we present a simple technique to analyze and interpret DFIT data similar to the analysis of the classic drillstem test (DST) data in vertical wells. The only difference is that in DFIT pressure-time characteristics approximate the linear flow regime while in DST, pressure-time behavior follows the radial flow regime. In general, DFIT analysis provides matrix permeability while the analysis of long-term pressure decline of the production data yields stimulated formation permeability , which can be attributed to microfracture permeability . Thus, we will show how we utilize the permeability from DFIT, and the permeability calculated from the production decline data of production wells (i.e., multistage hydraulically fractured wells) to construct a viable dual-porosity model to assess the performance of wells under primary production and gas injection EOR . We also compare our results with those of Nolte G-function . The paper includes numerical modeling of two-phase nonlinear flow , analytical solution methodology from multi-phase flow, experimental data, and field data to illustrate the viability of our interpretation method. Furthermore, our analysis technique is simple because it only uses pressure falloff data points during the shut-in period of the DFIT. Not only the method is confirmed by numerical modeling , but it is also verified by pressure falloff from laboratory data.

Current experimental methods of assessing rock fracture toughness require a large sample size (i.e., Brazillian test, semi-circular bending, three-point bending), which cannot be extracted at greater depths and along horizontal well. Alternatively, fracture toughness can be evaluated at a finer scale using scratch testing. This study investigates the rock failure mechanism using micro-scratch testing and phase field modeling on rock fragments. The phase field approach models the crack growth and initiation based on energy minimization principles and by portraying the crack surface as a diffused entity. The method has been known for its robustness in preventing numerical singularities due to sharp crack discontinuities in complex crack topologies. In the phase-field scheme, a regularization scalar order parameter is used to indicate the material’s state (from undamaged to damage) during fracture formation. The associated loss of stiffness in rock during fracture formations is captured by the coupling of selected energy degradation function with embedded scalar order parameter and partial differential equations defining the deformation, history, and phase-field evolutions. In doing so, information on stress strain development is needed to evaluate the change in free energy during cracks formation. In this study, scratch testing is used to obtain load-displacement data related to stress strain history. During the test, an indenter scratches the surface of the rock under increasing load. The critical loads where the crack initiates and the chipping spallation occurs are identified based on the microscopic observations, acoustic emission signals, recorded tangential force and recorded depth. The critical loads are used to determine the crack length associated with chipping formation, while the recorded force displacement data are used to obtain the dimensionless stress-strain curve. Both the crack lengths and the dimensionless stress strain curve are then used as the input to the phase field model developed to approximate the fracture toughness of the rock tested. Scratch tests are performed on samples obtained from Eagle Ford formation. The experiments are conducted in short transverse and divider orientations. The crack formations are studied under the progressive load application. The critical loads where crack initiates and chips form are identified mainly based on the panoramic picture obtained after the test and the spikes seen on the Acoustic Emission signals. Preliminary results show that the fracture toughness is lower for samples tested in parallel to the bedding orientation (i.e., divider). Fracture toughness of rocks has attracted wide attention in the last few years in the design and analysis of hydraulic fracturing for hydrocarbon and geothermal recovery. The currently proposed methodology allows for a quicker and more reliable way of approximating rock fracture toughness from small rock samples. The incorporation of the phase field model allows better prediction of rock fracture toughness as the method is capable of overcoming classical model limitations of quantifying crack initiation and crack propagation in the complex fracture networks.

Numerical flow simulation models for CO 2 injection, transport and trapping mechanisms have been used to optimize geologic CO 2 storage (GCS) by improving trapping efficiency and injection strategies. Injection of CO 2 into geologic formations can cause geomechanical changes and lead to reservoir expansion, ground surface uplift, and induced seismicity. Such deformations create geomechanical risks associated with GCS, which can compromise the safety and storage capacity of CO 2 . The geomechanical risks of CO 2 sequestration have drawn significant attention in recent years due to considerable ground uplifting and induced microseismic activities that have been reported in several field cases. However, studies in this area have not considered the effect of geomechanical risks in optimizing the storage performance. We present an optimization framework to strategically place CO 2 injection wells to maximize the storage capacity and minimize the associated geomechanical risks. To this end, we perform the optimization using three-dimensional coupled flow and geomechanical models, by applying the Mohr-Coulomb plastic failure criterion to model mechanical rock failure risk. Additionally, the geomechanical simulation results are used to quantify the risk associated with ground surface displacement and plastic strain, which are extracted from the simulation outputs. A multi-objective optimization problem is formulated to maximize CO 2 storage while minimizing the two forms of geomechanical risks. The injection well locations are defined as decision variables and Genetic Algorithm is implemented to solve the multi-objective optimization problem. The solution of the proposed framework leads to nontrivial optimal decisions, which are different from the case where geomechanical risks associated with CO 2 injection are ignored. We find that the wells may not necessarily be concentrated in areas with the highest storage capacity because that may lead to rock failure and unacceptable levels of ground surface uplift. Instead, the maximum storage is limited to keep the reservoir in the elastic state based on the specified reservoir geomechanical and flow properties. We also observe that well interactions during CO 2 injection can cause shear failure of the formation between the wells. These observations suggest that accounting for geomechanical risks associated with CO 2 injection is necessary to design a sustainable plan for CO 2 injection and storage. The paper ends with a discussion about the importance of accounting for the uncertainty in rock properties and the resulting model predictions in optimizing the well locations.

While diagnostic fracture injection tests (DFIT) data is relatively rare, most hydraulic fracture treatment stages in multiple transverse fracture horizontal wells (MTFHWs) follow pressure and rate data during pumping with several minutes of pressure falloff data after the end of pumping. Recent papers have shown value in applying analysis developed for DFIT data to hydraulic fracture treatment falloff (HFTF) data. As is often the case with routinely acquired operational data, the time required for analyzing HFTF datasets for each of the treatment stages in a long horizontal well may be prohibitive. This paper offers an automated analysis procedure that starts from standard treatment pressure and rate data and produces estimates for wellbore and perforation friction loss, near wellbore tortuosity friction loss, and the instantaneous shut in pressure (ISIP). Steps in the automated procedure include isolating the HFTF data from the rest of the hydraulic fracture treatment data (which is typically subject to hydraulic hammer), applying an optimized low-pass filter (LPF), and computing the friction and ISIP estimates by automating a previously published graphic procedure. We employ the automated analysis on field data that was previously analyzed by hand. Then we compare and contrast between the two analyses. The comparison between manual and automatically analyzed fracture treatment falloffs demonstrates that the automated procedure reproduces the previous analyses except in a few cases that pose special challenges to any analysis. The new field dataset results demonstrate the approach is practical for field application. Variations in friction loss and ISIP estimates along MTFHWs provide data useful for well and pad completion design decisions related to perforating strategies, and fracture and well spacing.

This paper presents an optimum way to produce down to depletion a compartmentalized reservoir in offshore deep environment by considering geomechanical stress-deformation mechanisms and associated problems. The case study is for a faulted reservoir zone of the Aphrodite field, located in the Eastern Mediterranean. The study is based on finite element modelling using 2D plane strain analysis that incorporates pore pressure and elastoplastic deformation of reservoir and overburden rock formations using the Drucker-Prager plasticity model. The mechanical properties of the reservoir sandstones were derived from calibration of data obtained from triaxial tests and for the overburden shale layers from acoustic velocities and correlation functions. The compartmentalized geometry was constructed based on seismic data and logging data obtained at the exploration and appraisal phases. The estimated insitu stress field was transformed and applied on the boundaries of the model blocky geometry. Four different initial and equilibrium depletion scenarios were examined and the obtained results in terms of deformation and effective stresses are compared. The first scenario reflects the initial stress state, the next two intermediate scenarios present non-uniform depletion cases for each fault block, and the fourth scenario presents the case of a uniform depletion. It was found that the uniform depletion of the reservoir compartments creates the least stress contrast in the field and consequently, ensures better control of stress-related impacts during the production. The analysis highlights the local regions of a fault blocks system that potentially suffer by high shear strains that can cause fault reactivation or induced fractured zones but the over-all risk remains low. Furthermore, the analysis establishes relationships between the mean effective stress, volumetric strain, and permeability changes in order to predict the regions with improved transmissibility characteristics or the less permeable compacted rock regions of the reservoir. Overall, the analysis can provide an appreciation of the stress/strain-driven characteristics of the reservoir showing the area of rock compaction tendencies of the faulted blocks and the further deformation in depletion conditions. The presented work demonstrates clearly that a properly calibrated reservoir geomechanical model can be used as a screening tool for examining depletion scenarios of compartmentalized reservoirs, highlighting areas of potential problems such as fault activation, wellbore shearing, reservoir compaction, permeability changes and fault sealing.

The most common stimulation technique for shale production is multistage hydraulic fracturing. Estimating fracture geometry is a focal parameter to judge the fracture operation and predict the well performance. Different direct and indirect techniques can be used for fracture diagnostics to estimates fracture geometries. The current study combines fracture measurements and pressure transient analysis to estimate fracture surface area on each stage and to estimate production as a pseudo production log. The numbers and kinds of fractures were calculated as a function of treating pressures, injection rates, proppant concentrations, and formation properties to compute fracture surface area (FSA). Pressure transient analyses were then conducted with the leak-off data upon completion of each frac stage to estimate the producing surface (PSA). The fall-off data was processed first to remove the noise and water hammering effects. The PTA diagnostic plots were used to define the flow regime and the data were matched with an analytical model to calculate producing surface area. Tensile and shear fractures are both created during the injection of frac fluids. Shear fractures are caused by movement in already existing natural (fluid expulsion) fractures found in all shale source rocks. Shear fractures form a pressure below the minimum horizontal stress. These shear fractures take advantage of the rock fabric and develop higher surface area than tensile fractures for the same given volumes of water and sand. FSA is a measure of permeability enhanced area due to hydraulic fracturing. Producing surface area is the resulting effective flow areaconnected to the wellbore. Diagnostic plots showed a linear and radial flow regime depending on the formation and the completion design. Good correlations were found between PSA and FSA results. In general, higher FSA produces higher PSA. In cases where producing surface area was higher than expected from fracture surface area, communication was found with offset wells. When FSA higher than PSA were found, it was usually caused by increased stress from too close offset wells. Combining FSA and PSA measurements provides forecasts of production for each stage and helps to optimize well spacing at the end of each frac stage.

This study provides a systematic method to study the gas/water transport behavior and water retention behavior in gas shale reservoir using analytical and numerical modelling approaches. The critical water saturation determined from the relative permeability curve is regarded as the turning point between two gas production stages: Stage I and Stage II. At Stage I, the gas /water two phase flow dominates shale gas production period, while at Stage II the single gas flow and the flow of water in adsorbed phase are dominative. The improved permeability evolution model incorporates both the gas and water sorption-induced swelling strains. The permeability ratio is a time- and location-dependent parameter and its change can be divided into three stage, namely, the increase at early time, the decrease at mid-term time, and the increase at later production time. With gas depletion, gas pressure drawdown significantly dominates the evolution of permeability ratio. Continuously, the effect of sorption induced matrix shrinkage strain will become dominative with gas pressure decreases to low values. Due to the gas source supplement induced by gas desorption at Stage I, the gas production rate will temporarily increase but then will continuously decrease till to the lowest level at critical water saturation. At Stage II, the gas production rate is significantly influenced by residual water content in shale controlled by the flow of water in adsorbed phase. In addition, the effects of elastic properties ( i.e . Young's modulus and Poisson's ratio), initial permeability and diffusion time on permeability ratio and gas/water production profiles were discussed. These results provide a first rational method for analyzing the gas/water transport behavior and water retention behavior with potential application of shale gas recovery.

Recently new Distributed Acoustic Sensing (DAS) data have been collected during hydraulic fracturing in shale. Low frequency DAS signals show patterns that are intuitively consistent with the understanding of the strain field around hydraulic fractures. This study utilizes a fracture simulator combined with a finite element solver to further understand the various patterns of the strain field caused by hydraulic fracturing. The results can serve as a "type-curve" template for the further interpretation of cross well strain field plots. Incorporating detailed pump schedule and frac fluid/proppant properties, we use a hydraulic fracture simulator to generate fracture geometries, which are then passed to a finite element (FE) solver as boundary conditions for elastic-static calculation of the strain field. Since the FE calculated strain is a tensor, it needs to be projected along the monitoring well trajectory to be comparable with the fiber strain, which is uniaxial. Moreover, the calculated strain field is transformed into time domain using constant fracture propagation velocity. Strain rate is further derived from the simulated strain field using differentiation along fracture length. Scenarios including a single planar hydraulic fracture, a single fracture with a discrete fracture network (DFN), and multiple planar hydraulic fractures, in both vertical and horizontal directions were studied. The scenarios can be differentiated in the strain patterns based on the finite element simulation results. In general, there is a tensile heart shaped zone in front of the propagating fracture tip. On the sides there are compressional zones parallel to the fracture. Multiple planar fracture show polarity reversals in horizontal fiber due to interactions between fractures. Strain field/strain rate show consistent patterns with what is observed from field cross well strain data. The application of the study is to provide a template to better interpret hydraulic fracture characteristics using low frequency fiber strain monitoring. To the author's understanding, there are no comprehensive templates for engineers to understand the strain signals from cross well fiber monitoring. The results of this study will guide engineers toward better optimization of well spacing and frac design to minimize well interference and improve efficiency.

Well productivity reduction over time is one of the critical issues for deep-water wells with huge implications on expected recoveries from these wells. It is important to account for this uncertainty accurately in order to generate reliable production forecasts. Typically, reservoir simulation engineers utilize an expression (e.g. linear or exponential) for modeling changes in skin or Productivity Index (PI) over time as a function of either time or cumulative liquid production or pressure depletion. These commonly used single variable-based PI degradation modeling methods are easy to implement with a flow simulator, but they do not address the multi-dimensional nature of PI degradation which is a result of multiple subsurface effects combined with operational conditions. As a result, these single variable-based modeling methods generally do not have good reliability for predicting PI degradation trend. This article proposes a method for predicting reduction in producing well's PI by integrating a few key well operating variables (drawdown, borehole depletion, and water cut) into a single mathematical formulation. One of the important assumptions of IPDM is that certain critical drawdown pressure exists for each well in the field. When a well is operated below the critical drawdown pressure, negligible to no PI reduction is observed; however, when drawdown exceeds the critical drawdown value, noticeable PI reduction is seen. The reduction level depends on the ratio of current drawdown pressure to critical drawdown pressure, as well as on pressure depletion and water cut levels. By doing so, the global trend and local granularity of PI reduction are well captured. The concept of critical drawdown used in IPDM is aligned well with the awareness of safe drawdown among Reservoir Management practitioners. Critical drawdown ranges derived from IPDM form a useful analog data set while deciding safe drawdown limits or forecasting PI degradation trends for future wells. The predictability and generality of IPDM were assessed with historical well test (or PTA – Pressure Transient Analysis) data from more than 50 wells across seven deep-water fields. A detailed workflow of implementing IPDM was demonstrated through one of the field applications. From practical implementation point of view, IPDM method can be easily written in Excel Macro or using a PYTHON script or any other programming language. Once the methodology is coded (or implemented), it can then be plugged in a dynamic flow simulator (e.g. INTERSECT™). A similar approach is usually taken while incorporating single variable-based method for forecasting PI degradation into a flow simulator. Compared to the coupled modeling of geo-mechanical simulator - dynamic flow simulator, the IPDM is a much simpler approach to use and also enables engineers to evaluate the impact of key operating conditions on individual well's PI.

Diagnostic fracture injection tests (DFITs) have been widely studied and implemented in unconventional reservoirs to derive properties such as closure stress, pore pressure, and permeability. During a DFIT, a small volume of water is pumped into a formation to create a small-sized crack. Formation permeability is typically obtained by means of modeling fluid leakoff during the shut-in period. Early studies have assumed a constant fluid pressure boundary condition on the fracture walls or a constant leakoff rate into the formation. However, the results deduced based on these assumptions may introduce significant errors because the fluid pressure inside a fracture dissipates quickly as the fluid leaks off into the formation. In this study, we propose a material balance approach to obtain formation permeability using DFIT data. The proposed analysis takes into account fluid leakoff during both fracture propagation and well shut-in periods. To model fluid leakoff during fracture propagation, we adopt the superposition principle to decompose the problem into two separate problems; we then obtain the analytical solution. Two synthetic cases are presented to validate the proposed analysis. The results suggest that the proposed approach provides a good estimation of formation permeability. This approach has broad field application potential, as it can be used even when pressure data contains significant levels of noise. In addition, the solution is more accurate than those provided in available studies of formation permeability estimation using DFITs data, especially when formation permeability is not extremely tight.

Maximizing stimulated natural and hydraulic fracture network is one of the primary hydraulic fracturing concerns for economic production from a horizontal shale gas well. Geomechanical facies and preexisting fractures in each stage are identified based on similarities in formation characteristics to optimize the locations of perforation clusters. This often requires analyzing large volumes of drilling, Logging While Drilling (LWD) and Measurement While Drilling (MWD) data. In this paper, we develop a methodology that calculates the mechanical specific energy (MSE) using real-time drill string acceleration signals directly from its definition. High resolution vibration signals have been collected using a tri-axial accerlometer, which was an auxiliary tool included in acoustic borehole imager. This technique provides a cost-efficient solution for engineered completion design. Furthermore, we adopt deep Convolutional Neural Network (CNN) with signal processing to build a data pipeline that effectively extracts patterns from dynamic acceleration signals for rock lateral MSE classification. First, we apply discrete wavelet transform and Short-Time Fourier Transform (STFT) for signal denoising and pattern recognition. Then we construct an image dataset using multi-scale image fusion at pixel level from 3 sensor channels, including axial, lateral acceleration spectrograms and zero-padded revolutions per minute (RPM). The resulted RGB image dataset includes 4,000 images of 5 MSE ranges with various rock strength conditions. Our results demonstrate that the proposed deep learning model can achieve more than 90% classification accuracy. The deep learning results, as a reference source, were applied in selected Marcellus Shale Energy and Environmental Lab (MSEEL) wells engineered completion located in the Marcellus shale gas site.

We use a high-quality dataset in the Bakken Shale to calibrate a numerical model to a complex and diverse set of parent/child observations. Two vertical wells (V1 and V2) were drilled 1000 ft and 1200 ft away from a legacy well with 10 years of production, H1. A DFIT was performed in the V1, followed by a 24 hour low-rate injection in the H1 (a microseismic depletion delineation, MDD, test). Subsequently, a small frac job was performed in the V1, followed by DFITs in the V1 and V2. The dataset yields a diversity of data to calibrate a numerical model: historical production of the H1, pressure response in the H1 from the MDD injection and the V1 fracture treatment, production rate uplift in the H1 following the V1 frac, microseismic, and pressure response during the three DFITs. The entire dataset was history matched in a single continuous simulation with a numerical simulator that fully integrates hydraulic fracture and reservoir simulation. The simulation was set up to closely match a geologic model that was built in prior work. The integrated simulation allows simulation of the fractures reopening around the H1 as a consequence of the MDD, the transport of proppant from the V1 to the H1 well, and the subsequent communication and poroelastic stress response. The Biot coefficient was calibrated to match the observed change in stress at the H1 well after ten years of depletion. The fracture toughness was calibrated to match the observed fracture geometry from the microseismic around the V1 well during fracturing. A proppant transport parameter called ‘maximum immobilized proppant’ was tuned to the production and DFIT data. The match to the V2 DFIT suggests that it is not directly in contact with the V1 fracture, even though the wells are relatively close together along fracture strike. The initial V1 DFIT suggests that it has, at most, weak contact with the H1. The second V1 DFIT, performed after the fracturing treatment, demonstrates communication with the H1, and consequently, depletion. The observations demonstrate that the H1 was able to produce from the previously undepleted rock around the V1, even though it was 1000 ft away. Overall, the results indicate that Bakken wells can achieve substantial (at least 1000 ft) effective half-length, that frac hits on parent wells in the Bakken do not necessarily result in production degradation and can even increase production, that the apparent Biot coefficient is relatively low (∼0.34), that the amount of proppant trapping due to localized screenout is relatively low (but nonzero), and this entire, complex dataset can be explained using a planar fracture modeling approach.

Knowledge of pore pressure, in-situ stress, and lithology in unconventional reservoirs is important for safe and economic drilling, hydrocarbon production, and geomechanics applications such as wellbore stability analysis and hydraulic fracturing. Reliable predrill predictions of pore pressure, in-situ stress, and lithology are thus required for safe drilling and optimal development in such reservoirs. In the Permian Basin, changes in lithology occur over vertical depths that cannot be resolved by seismic velocities obtained by kinematic analysis, as these have poor vertical resolution. To obtain improved vertical resolution, seismic prestack depth-migrated (PSDM) data are input to amplitude variation with offset (AVO) inversion, for an area in the Delaware Basin where wide-offset 3D seismic data are available. AVO inversion provides estimates of both P- and S-impedance. The results are used to build a 3D mechanical earth model, which is employed to predict pore pressure, in-situ stress, and geomechanical properties. The model enables integrating the results of seismic inversion with drilling data, measurements on cores, wireline logs, formation and fracture closure pressures, and other data. By employing P- and S-impedance, and their ratio, pore pressure, in-situ stress, and lithology derived from seismic prestack inversion provides greater resolution than estimates obtained using seismic velocities from kinematic analysis. Examples from the Permian Basin illustrate the importance of the results for unconventional reservoir development.

The depletion of producing layers leads to significant stress changes in adjacent targets, especially when complex natural fractures are present. Due to weak bedding interfaces or small stress barriers, hydraulic fractures can easily penetrate neighboring layers, and this further increases the chance of multilayer stress disturbance. In this work, we investigate the effects of vertical fracture complexity, i.e., hydraulic fracture penetration and interlayer natural fracture existence, on stress interference between different layers using a data set from a typical shale gas well in the Sichuan Basin. Two geological contexts, i.e., without interlayer natural fractures (w/o INF) and with interlayer natural fractures (w/ INF), are considered under different degrees of fracture penetration and interlayer connectivity. An in-house iteratively coupled geomechanics and compositional reservoir simulator is used to model the three-dimensional pressure and stress changes. The non-uniform hydraulic fractures and stochastic natural fractures are incorporated in our coupled simulation with an embedded discrete fracture model (EDFM). Comprehensive spatial-temporal stress analysis quantifies the approximate range of orientation change of S Hmax and magnitude change of S hmin under various reservoir conditions. Numerical results indicate that the presence of natural fractures in the interlayer upgrades the risk of stress interference between different pay zones. A larger hydraulic fracture penetration increases gas production, but also exerts a significant impact on stress reorientation and redistribution in the upper potential layer. The orientation change of S Hmax along the prospective infill location is below 10 degrees in the w/o INF cases but up to 30 degrees in the w/ INF cases with a moderate number of interlayer natural fractures. The average magnitude change of S hmin is within 3.5 MPa along the prospective infill location for most w/o INF cases, whereas that in the w/ INF cases is above 10 MPa at most times. Moreover, the existence of natural fractures in the interlayer brings forward the occurrence of maximum orientation change in the upper layer by around one year. While inducing non-negligible stress drop in the upper layer, higher interlayer matrix permeability does not significantly reorient the horizontal principal stresses. Varying the density of interlayer natural fractures not only affects the stress magnitude but also causes considerable orientation change in the upper layer. The findings from this work help understand the extent of stress interference in the upper potential layer of the Sichuan Basin under different vertical fracture complexities. It is of guiding significance for future hydraulic fracture design and child-well operations in similar highly fractured tight formations.