Faster production declines than initially forecast were observed in numerous deep-water assets. These wells were completed as Cased Hole Frac-Pack (CHFP) completions (
Seven key damage mechanisms were identified as forming the basis for PI degradation: 1) off-plane perforation stability, 2) fines migration, 3) fracture conductivity, 4) fracture connectivity, 5) fluid invasion, 6) non-Darcy flow and 7) creep effects. A near wellbore production model incorporating the completion, fracture geometry and reservoir is coupled with a geomechanics model to assess each mechanism. A Design of Experiment setup varies the input ranges associated with each of the seven damage mechanisms. Input parameters for the model are risked and rely on ranges from standard and newly developed well and lab tests. The model assesses well performance and driving mechanisms at different points in time within the production life.
Primarily the study focused on high permeability and highly over pressured reservoirs. For the types of wells/fields assessed in the study, the results indicated three phases of decline based on the interaction between the formation properties, the completion components and the operating parameters. The three phases breakdown into: (1) a pre-rock failure stage where declines are relatively small, (2) an ongoing rock failure stage where declines are rapid and (3) a post failure stage where declines are again moderate. In each of these stages different parameters and damage mechanisms were assessed to be impactful. The workflow was also utilized to match pre and post acidizing treatments. A comparison for varying rock types was included looking at the impact of rock strength and formation permeability on the ranking of the damage mechanisms. The impact of operating parameters such as drawdown can also be assessed with the tool showing that increased drawdowns may not always be beneficial to the long-term production of the well.
The paper presents the underlying drivers for PI Decline for deep-water assets of a specific attribute set. Through accurate representation of reservoir and completion, the workflow highlights the impact and combined impact of different damage mechanisms. The paper also shows a direct link between the mechanical properties (moduli and strength) and boundary conditions (pore pressure and stress) and the well performance and productivity. The workflow provides a methodology by which lab and field tests can be transformed into assessments of future well performance without strictly relying on analogs that may or may not be appropriate.
Williams, Ryan (Schlumberger) | Artola, Pedro (Schlumberger) | Salinas, Javier (Schlumberger) | Mirakyan, Andrey (Schlumberger) | MacKay, Bruce (Schlumberger) | Hoefer, Ann (Schlumberger) | Kraemer, Chad (Wisconsin Proppants) | Reese, Harrison (PRI Operating) | Roybal, Zack (PRI Operating) | Williamson, Brant (PRI Operating)
Use of regional sand in the Permian Basin dramatically increased in 2018. Regional or in-basin sand is often perceived as lower quality compared to northern white sand (NWS); however, its use is fairly new, and production data has not been available to determine if, or in what cases, higher quality matters. This paper presents the results from a production comparison of Permian Basin wells that were hydraulically fractured with NWS and regional sand or both.
A dataset consisting of approximately 450 wells completed with NWS or regional sand or both within the Delaware and Midland Basins was studied to determine the relationship between production performance and sand type (or quality). To evaluate the effect of sand quality in well production, the dataset was divided in smaller groups of wells with similar reservoir characteristics and completion practices. The initial phase of the study was completed using public domain production data, while the second phase focused on the development of regional reservoir models to forecast production of wells using NWS or regional sand or both.
When analyzing an area containing sufficient wells for a reliable comparison, the survey revealed no statistically significant difference in production for wells that used NWS versus regionally sourced sand. Models were built to predict differences in the production performance of each sand type. These models take into account and demonstrate the effects of differences in sand properties, as well as the impact of the favorable economics associated with regional sands. It was confirmed with the study that the sand type is not a critical factor in regards to production performance when completing wells that are hydraulically fractured in ultralow-permeability nonconductivity-limited reservoirs.
This paper presents an early look at the production numbers of West Texas wells completed with regionally sourced sand in the Permian Basin. The results of the study will encourage operators to further contemplate the use of regional sand when completing wells in ultralow-permeability shale reservoirs. This dataset will continue to evolve and reveal the effects of regional sand over the life of the well; this will be presented in a future paper.
Maximizing economic performance in shale requires optimal selection of well and cluster spacing, among other parameters. Reservoir engineering calculations can be used to optimize spacing, but these calculations are impacted by uncertainties in input parameters. System permeability is particularly important and difficult to measure. Diagnostic Fracture Injection Tests (DFIT’s) are often used to estimate permeability because they provide a direct, in-situ measurement. However, in recent work, it was shown that conventional DFIT interpretation techniques can overestimate permeability in gas shale by two orders of magnitude. In this study, the impact of the permeability estimate is demonstrated using a dataset from the Utica/Point Pleasant. Production data is history matched with models assuming high and low permeability. It is possible to history match both models because of non-uniqueness between fracture area and permeability. Sensitivity analysis simulations are performed to assess the impact of well and cluster spacing on net present value. Relative to the high permeability model, the low permeability model has a greater optimal well spacing and a tighter optimal cluster spacing. The comparison shows that improved accuracy in the permeability estimate significantly improves economic performance. The low permeability model has much earlier production interference than the high permeability model because the low permeability model requires greater effective fracture length to match production. This is consistent with the operator’s experience that outer wells outproduce inner wells within weeks or months from the start of production.
Seth, Puneet (The University of Texas at Austin) | Manchanda, Ripudaman (The University of Texas at Austin) | Elliott, Brendan (Devon Energy) | Zheng, Shuang (The University of Texas at Austin) | Sharma, Mukul (The University of Texas at Austin)
During stimulation of unconventional reservoirs, offset well pressure measurements are often used to estimate hydraulic fracture geometry. These measurements can also be used to make a quantitative estimate of the created fracture network area and the permeability of the stimulated rock volume (SRV) around the hydraulic fractures. Offset well pressure measurements recorded in the field clearly show a change in the pressure response of the monitor well when the injection rate in a nearby fracture treated well is changed. The shut-in period between two frac stages in the treatment well corresponds to a distinct pressure fall-off in the monitor well. We present a workflow where we analyze and match this pressure fall-off in an offset monitor well in response to fluid leak-off from a hydraulic fracture in the treatment well to estimate SRV permeability and the created fracture network area. The workflow and model are applied to field data from the Permian Basin.
A fully-coupled, 3-D, poroelastic reservoir-fracture simulator has been used to simulate pressure fall-off in the offset monitor well. Field data and simulation results are presented to show that during shut-in between two frac stages in the treatment well, a decrease in the injection rate causes the monitored offset well pressure to fall-off. We find that this fall-off in pressure is influenced by leak-off from the treatment well fracture. During the shut-in period, fluid leak-off from the treatment well fracture into the SRV region decreases the width of the fracture which consequently affects the stress-shadow and the poroelastic pressure fall-off in the offset monitor well. The pressure fall-off in the monitor well is, therefore, shown to be caused by 1) the fluid leak-off from the monitor well fracture and 2) stress-shadow relaxation around the monitor well fracture as fluid leaks-off from the nearby treatment well fracture into the formation.
We present a new method to estimate the permeability of the stimulated region around the created fractures. We show that, along with the permeability of the SRV region, the stress-shadow of the treatment well fracture on the monitor well fracture also has a significant impact on the pressure fall-off in the monitor well. We use a conceptual model to estimate the created fracture network area which can be used as a metric to identify the effectiveness of a frac job and provide insights into the generated fracture complexity during the frac job. In addition, the estimated SRV permeability and fracture network area are critical inputs in production forecast simulations that can guide an operator to make better economic decisions in a relatively inexpensive manner.
Karazincir, Oya (Chevron) | Li, Yan (Chevron) | Zaki, Karim (Chevron) | Williams, Wade (Chevron) | Wu, Ruiting (Chevron) | Tan, Yunhui (Chevron) | Rijken, Peggy (Chevron) | Rickards, Allan (Proptester Inc.)
Fracture face permeability loss related to proppant embedment under depletion conditions is one of the factors that contribute to PI decline in hydraulically fractured reservoirs. The damage is a result of proppant embedment and proppant/core crushing that also generates fines and plugs the pores in the embedment zone. A recently developed test method was used to measure the effects of embedment on permeability reduction at the fracture face as a function of depletion, core porosity and permeability, core UCS and injected fluid types. A range of stresses was applied to the cores and the permeability values across the fracture face were tracked and compared.
Standard proppant conductivity tests only measure permeability / conductivity losses within the proppant pack due to frac gel damage and compaction and do not measure the damage at the fracture face.
Recently, we developed a new test method that can directly measure fracture-face permeability under depletion. The test flows fluid from the matrix into the fracture thus coupling permeability measurements within the proppant pack with those across the fracture face. Initial tests were conducted in intermediate permeability, intermediate strength rock. The resulting core permeabilities were compared to the values from conventional core permeability tests. The proppant conductivity values were compared to those measured using a modified API conductivity test set-up. During the second part of this study, a lower permeability, higher UCS rock system was tested and permeability decline was measured as a function of applied stress and flow rate. Cores saturated with different fluid systems were tested to mimic injected fluid imbibition into the fracture face and their effect on proppant embedment and permeability loss. A numerical model was built to calculate fracture face permeability reduction as a function of depletion and injected fluid systems in different types of fractured formations. By incorporating embedment induced permeability reduction into the detailed reservoir geomechanics model, we are able to evaluate the contribution of proppant embedment on overall PI decline. During post-test analysis, proppant embedment percentage and porosity reduction across the fracture face were measured, using Micro CT-scanning, and the damage was also studied using thin section analysis.
Objectives/Scope: In order to maximize the recovery of hydrocarbons from liquids rich shale reservoir systems, the cause and effect relationships between production and the stimulation methods need to be clearly understood. In this study, we utilize multivariate regression models to narrow down the variables in flow simulation models and their range. We then use the flow simulation model to understand the fractured well production behavior and field wide well performance in a liquids rich petroleum system in the Duvernay Basin.
Methods, Procedures, Process: Statistical models assume no physical relationship between the model parameters and the response variable, which in this case is produced volumes over a period of time. On the other hand, simulation studies incorporate physical mechanisms of flow to model and predict the production behavior. The simulation models, however, fall short of incorporating all the mechanisms contributing to the production behavior in the complex shale gas reservoir. Thus there is a need for integration of statistical approaches of understanding production behavior along with physics based model and simulation approach. We use the statistical methods to identify the important physical mechanisms that control the production.
Results, Observations, Conclusions: Multivariate linear regression analysis of the 6 month produced volume and its relationship with parameters such as fracture fluid volumes used, proppant weight placed, number of stages fractured provides a model with reasonably good correlation. The 6 month produced volumes correlate with large proppant weights, lower fluid placements and greater density of fracture stages. Use of Random Forests machine learning algorithm on the dataset confirms that the total proppant placed, well length completed with fractures have high importance coefficients. In order to examine the well performance using full physical models, fractured well simulations are performed on particular wells using the trilinear model. The trilinear model predictions are then compared against other production analyses and the regression model results for consistency. The models showed that in the absence of stress dependent permeability, the production forecast was much higher. Thus, stress dependent permeability appears to be an important factor in the modeling and prediction of production from liquids rich shale reservoirs.
Novel/Additive Information: In this study we describe a method to understand the production data from a liquids rich shale reservoir, by integrating multivariate linear regression analysis, machine learning algorithms along with physical model simulations. The results are novel and offer a method to validate either approach to understand cause and effect relationships. This approach may be classified as a new hybrid modeling workflow that may potentially be used to optimize stimulation techniques in liquids rich shale reservoirs.
Tabatabaei, Maryam (Pennsylvani State University) | Dahi Taleghani, Arash (Pennsylvani State University) | Cai, Yuzhe (Pennsylvani State University) | Yang Santos, Livio (Pennsylvani State University) | Alem, Nasim (Pennsylvani State University)
Proppant bed plays a critical role in enhancing oil and gas production in stimulated wells. In the last two decades, there have been consistent efforts to improve shape characteristics and mechanical strength properties to guarantee high permeability in the resultant propped fracture. However tuning wettability of proppants have not yet engineered considerably maybe because natural sand has been a typical raw material for proppant manufacturing. However, water wet proppants may not only limit production due to reduced hydrocarbon relative permeability but also facilitate fine migration through the proppant bed. Fine migration and increasing water saturation may deteriorate oil production over time. Intrinsic hydrophobicity of graphitic surfaces and their two-dimensional geometries made them a promising candidate for coating proppant to alter its wettability. In this paper, we present a methodology for treating proppant surfaces with graphite nanoplatelets. Standard laboratory tests following modified API RP61 have conducted to show the effectiveness of the proposed methodology.
Tight gas is the term commonly used to refer to low permeability reservoirs that produce mainly dry natural gas. Many of the low permeability reservoirs that have been developed in the past are sandstone, but significant quantities of gas are also produced from low permeability carbonates, shales, and coal seams. Production of gas from coal seams is covered in a separate chapter in this handbook. In this chapter, production of gas from tight sandstones is the predominant theme. However, much of the same technology applies to tight carbonate and to gas shale reservoirs. Tight gas reservoirs have one thing in common--a vertical well drilled and completed in the tight gas reservoir must be successfully stimulated to produce at commercial gas flow rates and produce commercial gas volumes. Normally, a large hydraulic fracture treatment is required to produce gas economically.
Kumar, Ashish (The University of Texas at Austin) | Shrivastava, Kaustubh (The University of Texas at Austin) | Manchanda, Ripudaman (The University of Texas at Austin) | Sharma, Mukul (The University of Texas at Austin)
Production from naturally and hydraulically fractured reservoirs is highly dependent on the complex fracture geometry of the fracture network. It is computationally very expensive to model the mechanics, closure and flow of each individual fracture in a large domain with thousands of fractures. We propose a workflow to convert the discrete fracture network (DFN) of fractures into an effective permeability tensor that can be used to simulate flow in such complicated fracture networks.
A discrete fracture network (DFN) of natural fractures is stochastically generated and the displacement discontinuity method based hydraulic fracturing simulator (Multi-Frac-NF) is used to model the hydraulic fracture propagation. This created fracture network along with induced unpropped (IU) fractures are imported into a geomechanical reservoir simulator. During flowback, the permeability tensor for the stimulated reservoir volume (SRV) is calculated. The effect of fracture height and natural fracture orientation on effective permeability tensor of SRV is systematically investigated.
We show that the propagating hydraulic fracture can generate enough stress perturbations to allow hydraulically disconnected natural fractures to fail in its vicinity. These disconnected IU fractures can also increase the effective permeability of the reservoir close to the hydraulically connected fracture. Simulation results indicate that the effective permeability of the SRV is a strong function of the natural fracture orientation and hydraulic fracture height.
We propose a workflow which includes the coupled effect of geomechanics and reservoir flow on the estimation of the effective permeability tensor for the SRV. The workflow presented in this paper provides a novel method to generate the reactivated natural fracture network around propagating hydraulic fractures and to capture the behavior of complex fracture networks in simplified reservoir simulation models using an effective permeability tensor for the SRV.
In ultra-low permeability reservoirs, hydraulic fracturing stimulation is performed to maximize the surface area available for production of hydrocarbons. In the presence of natural fractures, hydraulic fracturing stimulation can create complex fracture networks (Fisher et al., 2002; Weng et al., 2011; Shrivastava et al., 2018b). Stress perturbations caused by hydraulic fracture propagation can lead to shear failure of natural fractures far away from the hydraulic fractures (Agrawal et al., 2019). These shear slippage events are registered as microseismic events. In the case of highly fractured reservoirs such as the Barnett a very complex pattern of microseismic events is observed (Fisher et al., 2002; Cipolla and Wallace, 2014). To capture the impact of this complex stimulation behavior on production and well performance in a simplified model, the concept of “Stimulated Reservoir Volume” (SRV) was introduced by Fisher et al., 2004. This has enabled traditional reservoir simulation models to use SRV as a proxy for complex fracture networks to mimic the actual production behavior observed in the field (Mayerhofer et al., 2010). However, often the SRV parameters are used as calibration parameters in history matching and are disassociated from the fracture modeling. This type of workflow can mimic the early production trends but can lead to erroneous predictions of well performance (Cipolla and Wallace, 2014).
During the hydraulic fracturing process, the fracturing fluid may cause water blockage, if the nearby secondary fractures subsequently close and get disconnected due to changes in effective stress distribution during flowback and production. The fluid inside the fractures could also get squeezed out upon fracture closure. The circumstances and detailed mechanisms associated with this phenomenon are still poorly understood. In this work, a coupling scheme for incorporating a pressure-dependent apparent permeability model in reservoir simulation is implemented. The numerical models are subsequently used to investigate the impacts of water blockage and apparent permeability modeling on gas production and water flowback.
A high-resolution 3D reservoir model is constructed based on the field data obtained from the Horn River shale gas reservoir. Stochastic 3D discrete fracture network (DFN) model is upscaled into equivalent continuum dual-porosity dual-permeability (DPDK) model by analytical techniques. A realistic DFN configuration is examined to simulate the potential scenarios of water blocking. An apparent permeability (Kapp) model that accounts for contributions of Knudsen diffusion, slip flow and surface pore roughness is introduced. In order to capture the pressure dependency, a novel coupling scheme is developed to facilitate the updating of Kapp and effective stress after a certain designated time interval. In addition, a novel method involving rock-type indicators is introduced to represent the open and closed states of secondary fractures, facilitating the modeling of stress-dependent closure of the secondary fracture system.
Fracture closure and the resulting water blockage would impact the gas production and water recovery, particularly if the near-well fractures are disconnected. Neglecting the effects of Kapp could essentially overestimate the contribution of hydraulic fracture for a certain observed gas production. The existence of secondary fractures could also enhance water loss, which is contrary to some conclusions in previous research where Kapp modeling and disconnected fractures are ignored. The impacts of shut-in duration and matrix multiphase flow functions are systematically studied. It is concluded that gas and water production would increase if less water is imbibed into the matrix during the shut-in period in the presence of disconnected secondary fractures. It is also observed that a shorter shut-in period may be beneficial to both water and gas recovery, where previous studies have reported no observable increase in gas production when secondary fracture closure was not considered.
This work presents a set of detailed simulation studies to examine the scenarios or conditions that may be responsible for water blockage, particularly in the presence of disconnected secondary fractures. A novel, yet practical, scheme is implemented to couple stress-dependent matrix apparent permeability and fluid flow, as well as to model pressure-dependent fracture closure. The modeling scheme can be readily integrated in most commercial reservoir simulation packages. The results have revealed several potential scenarios of water loss, along with the associated implications on optimal operational strategies and estimation of stimulated reservoir volume.