Unconventional plays present a challenging case to design an optimized stimulation program and to maximize reservoir contact and hydrocarbon production. In this regard, conducting a reliable well spacing optimization study demands realistic and explicit fracture descriptions. This work applies an integrated technique to a multi-well case study in the Permian Basin to extract fracture dimensions based on microseismicity-derived behavioral fracture maps, while honoring the RTA-based estimates of the contributing fracture volume. The fracture dimensions are then used to conduct analytical and numerical studies to decide the optimal well spacing/placement design in the target formation. The numerical simulations in two stacked and staggered configurations confirm that although the staggered development causes a marginal decrease in the individual wells' performance, if successfully accomplished, it contributes to a higher vertical sweep efficiency from the section. Furthermore, comparing the approximations of failure planes, constructed based on the spatiotemporal analysis of microseismic events, with those achieved through seismic moment tensor inversion confirms that the collective behavior analysis gives fair estimates of fracture spatial evolutions.

Recently machine learning has being extensively deployed for oil and gas industry for improving result and expedite process. However, the black box models do not explain their prediction which considered as a barrier to adopt machine learning. This paper is about optimizing hydraulic fracture with machine learning methods and making informative decision with interpreting machine learning model. The solution can show that it could save over million dollars per well and improve well performance significantly. Interestingly, the machine leaning explainability approach was utilized to explain and measure the reason behind of why some wells are performing better than other and vice versa. Hydraulic fracturing modeling and optimization in tight oil and unconventional reservoir requires substantial geological modeling, fracture design, post-fracture production simulation with excessive sensitivity analysis due to complexity and uncertainty in the nature of data. These types of studies are computationally and monetarily expensive. Furthermore, digital oil technology has facilitated the process of data gathering enabled operators to have access to huge amount of data. Common approaches are no longer suitable to handle this pile of data but machine learning methods could be successfully utilized for this purpose. In this paper, a variety types of advanced machine learning methods including linear regression, Random forest, Gradient Boost, XGBoost, Bagging, ExtraTrees and neural network were employed to optimize well completion in Montney formation. The objective was to create a robust predictive model capturing all the effective operational well parameters (features) capable of optimizing the first 12 months cumulative of equivalent well production. Special Individual Conditional Expectation (ICE) plots and Partial Dependency plots(PDP) were used to depict how HF completion features influence the prediction of a machine learning model. Furthermore, a novel approach was employed to explain the model prediction of an existing well by computing the contribution of each feature to the prediction. Over 1838 hydraulically fractured (HF) wells producing from 2008 till 2019 in Montney formation have been considered for this analysis. The outcome of Explanatory Data Analysis (EDA) revealed that well production performance has not been improved despite of continues enhancement of hydraulic fracture parameters such as proppant injected volume, length of stimulated horizontal wells, and number of stages per well in the course of two years. This finding raises the concern of whether operators are properly optimizing completion design. After comparing all machine learning methods, Random Forest method was chosen as the most appropriate and accurate method to proceed for further analysis. ICE and PDP plots helped to understand the impacts of different fracturing features on production for individual well in addition to define optimum operation features on Montney Formation. Furthermore, quantifying of each feature’s impact on individual well production and linking it to an economic model, we were able to demonstrate potential profit and loss for each well. The model suggests that some wells could have achieved over $1 million extra profit during the first 12-months of production. In this study, not only a reliable predictive data-driven model has been built for hydraulically-fractured wells in Montney formation, but also a comprehensive workflow of sensitivity and explainatability analysis has been introduced to obtain an optimized fit-to-purpose well completion design.

This paper builds on Klenner et al. 2018 , which utilized machine learning to understand well-to-well communication ("Frac hits") or fracture-driven interaction (FDI) during hydraulic fracturing operations. This paper introduces an infrastructure that enahances the process for real-time enablement including data acquisition, visualization, and alerting for completion adjustments during operations. A use case is also presented which monitored, characterized, and provided understanding to the root causes of FDI during operations in the field. Machine learning in real-time provided insights and informed decisions regarding treating pressure, slurry rates, or proppant concentrations. The analysis included real-time streaming of time series completion data of the hydraulically completed infill wells and the time series pressure data of the offset wells. The processing of the data in real-time uses s a 3-step process: 1) Apply analytics to detect and characterize interference events during the active stimulation and offset monitoring 2) Utilize alert mechanisms to notify operators when a significant communication event is occurring and 3) Apply machine learning techniques to determine what completion parameters inform the occurrence of FDI events, and quantify those parameters in order to provide operators with suggestive actions to help control the event in real-time. The novelty of this approach is the application of analytics and machine learning to multiple streaming data sets in order to alert the operator of events, identify causality and consequently optimize completion design in real-time. The quantification of the completion parameters provides rules or recommendations that enable operators to make informed decisions to mitigate or manage communication. Furthermore, the methodology lends a fundamental analytics framework for visualization, continuous learning to create and optimize other real-time completion and stimulation operations at the asset level.

A revolution of unconventional reservoirs is a turning point in the global oil and gas industry since these resources have massive reserves with large potential in contributing to hydrocarbon production. Previous EOR laboratory experiments and simulation studies in the literature illustrated promising results in terms of recovery factor for different EOR applications, such as CO 2 , surfactant, and natural gas. However, pilot tests performance reported contrast behavior due to misleading predicting for the EOR physics processes. This paper presents the experimental work to evaluate the feasibility of CO 2 -EOR using the huff-n-puff (HNP) protocol in the Middle Bakken (MB) Formation, the Mountrail County, Williston Basin, ND. We evaluate the oil recovery from CO 2 -EOR under several scenarios of operational and well/reservoir conditions. The parameters considered in the sensitivity study include temperatures, pressure, soak time, and number of injection cycles to obtain optimum conditions under which the incremental oil recovery from the MB Formation is increased. The wettability alteration (i.e. contact angle) was also studied using rock-chip samples before and after the HNP experiment at the Bakken reservoir conditions (present for example P & T in psi/F). The outcomes indicated on the effect of the reservoir temperature and pressure on the performance of the CO 2 , where the recoverable oil increases as the temperature and pressure increase until reach the optimum. As a previous research outcome, the number of cycling and soaking time are crucial design parameters for the HNP experiment and on the field as well to let the CO 2 time to diffuse into the deep formation and swell more oil. In addition, the wettability alteration was changed by CO 2 -EOR as injection pressures increase and the wetting phase move from the oil-wet toward the water-wet system. As overall outcomes from this research, the CO 2 HNP process has a good potential in the lab, and could be succeeded economically in field applications that might reduce the need for refracturing stimulation or infill drilling.

Analytical rate transient analysis (RTA) techniques are widely adopted for analyzing production data obtained from hydraulically fractured horizontal wells in tight or shale reservoirs. However, for a detailed characterization of the uncertain distributions of those complex heterogeneous and multi-scale fractures, numerical simulation and assisted history-matching approaches are often preferred. Generally, RTA results can be used to constrain the initial distributions of fracture properties (e.g., transmissivity, aperture, or intensity). A set of initial models are then perturbed during the assisted history-matching step. Unfortunately, the final updated models may deviate substantially from the initial RTA estimates; moreover, specific information regarding the flow regimes is not incorporated directly into the history-matching step. In this paper, a new assisted history-matching workflow is presented, where RTA results are used to constrain not only the initial DFN models, the interpreted flow regimes are also used to formulate a localization scheme for more efficient updating of the pertinent DFN model parameters. The outcome is an ensemble of DFN realizations that are calibrated to both geologic and dynamic production data. First, RTA interpretations and other pertinent geological data are used to infer the prior probability distributions of the unknown fracture parameters, from which an ensemble of initial DFN models is sampled. Next, the DFN models are subjected to numerical multiphase flow simulation; the predicted production profiles are compared with the actual historical production data. Finally, the fracture parameters are adjusted following an indicator-based probability perturbation method, which is capable of minimizing the objective function and reducing the uncertainties in the unknown fracture parameters simultaneously. A key feature is that the flow regimes identified from RTA are used to formulate a localization strategy, where individual segments of the production data is used to tune only a specified subset of the unknown model parameters. The adoption of localization strategies in other settings has been demonstrated to improve the convergence behavior of such ill-posed inverse problems. In a case study, the method is applied to characterize the probability distributions of four parameters in a multifractured shale gas well: primary fracture transmissivity, aperture of the secondary fracture, transmissivity of the secondary induced fracture and global fracture intensity. Results of the sensitivity analysis reveal that the production performance is most sensitive to these particular parameters. Their probability distributions are updated following the proposed approach to match the production history. Multiple realizations of the DFN model are sampled. A probabilistic approach facilitates the representation of uncertainties in fracture parameters via multiple equally-probable DFN models and their corresponding upscaled flow-simulation models. A more comprehensive and robust approach is presented for integrating specific RTA interpretations and estimations into various steps of the history-matching process.

This paper provides the technical details of our latest progress of the hydraulic fracturing ball seat event recognition model. In our previous work ( Shen et al., 2020 ), the U-Net architecture was used to build models to recognize the ball seat event. In this paper, we train the Convolutional Neural Network (CNN) to solve the same problem. Comparing with U-Net approach, the CNN model requires less computation resource during training phase while it has slight better performance.

An Augmented AI approach has been developed to optimize completion design parameters and access the full potential of unconventional assets by leveraging big data sculpting, domain-induced feature engineering, and robust and explainable machine learning models with quantified uncertainty. This method unlocks the full potential of a well using completion design parameters optimization that considers all the factors that impact well performance, geological characteristics, well trajectory, spacing, etc. By leveraging basin-level knowledge captured by big data sculpting with the use of uncertainty quantification, Augmented AI can provide quick and science-based answers for completion optimization, and also assess the full potential of an asset in unconventional reservoirs. By leveraging computer vision and natural language processing techniques, unstructured data from various sources were deciphered, combined and organized into a structured database. Imputation techniques were used to fill the gaps of missing data. With the Augmented AI approach, the median accuracies of IP and EUR predictions for new drills is around 90%, which often outperforms industry-standard type curving methods. With the explainable machine learning (ML) model, the direct impact of completion design parameters on well performance is deconvoluted among other parameters, such as engineering and geological attributes. The prediction also comes with an 80% confidence interval to quantify the prediction uncertainties, which allows for better risk management and confident business decision making. With the ML model and given economic inputs and metrics, many sensitivity analyses are performed to evaluate optimized completion design parameters. The proposed Augmented AI approach has been deployed to Eagle Ford wells.

In this paper, a new workflow for data wrangling, cleansing, processing, automated identification and classification of drilling dysfunctions (e.g. bit balling, drillstring vibrations, etc.) for any number of wells is presented. Statistically-derived baseline rock strengths were used to improve drilling efficiency and transparency in completions design. The methodology leverages commonly available rig surface sensor data and mud motor parameters. Data handling and processing across multiple databases was automated. Time-series data was converted to depth-based data, and drilling dysfunction indicator metrics (mechanical specific energy - MSE, depth-of-cut, etc.) were computed. Subsequently, the combined data set was segmented based on formation tops, bottom hole assembly (BHA) used, and hole sections. A multivariate physics-based decision tree methodology was then applied for identification and classification of drilling dysfunction. A baseline MSE was statistically derived and a subsurface rock strength database was created for every formation in every well within a given prospect. Over 100 historical unconventional wells across two shale basins were analyzed using the new workflow, and 10 different dysfunction categories were classified. The most commonly encountered drilling dysfunctions were whirl, stick-slip, and bit balling. An algorithm generates a set of reports per well based on independent segments for individual drilling operations, including strip charts of drilling parameters and dysfunction indicators, with identification of potential (low/high) problem areas. Auto-generated scatter plots, identifying optimal control parameters for every geological formation at a given location, are used for planning future wells and improving drilling performance. The calculated MSE baselines were averaged to build a repository of baseline MSEs per formation, which in turn can be utilized as a diagnostic tool for real-time MSE surveillance. The MSE baseline information can also be used as a representation of unconfined compressive strength (UCS), and thus used to formulate a reliable rock hardness map to enhance completion design. Manual dysfunction analysis is time- and labor-intensive, and therefore often done either incompletely or bypassed altogether. Automating this process reduces time and effort significantly. The novelty of this work lies in the translation of the qualitative understanding of MSE to a reliable quantification of rock strength through the analysis of magnitudes and trends in drilling parameters and dysfunction indicators. Screening out drilling dysfunctions enables a much better rock strength approximation to aid completion optimization.

Aimed at sharing the unconventional wisdom gained from a hydraulic fracturing monitoring case study in the Montney tight gas play, the work showcases the ability of 4D modeling of collective behaviors of microseismic events to chase the frac fluid and navigate the spatiotemporal fracture evolution. Moreover, microseismicity-derived deformation fields are integrated with volumetric estimates made by rate transient analysis to calibrate spatially-constrained SRV models. Through the case study, we give evidence of fracture containment, evaluate the role of natural fractures and the use of diverting agents, estimate cluster efficiencies, conduct analytical well spacing optimization, model productivity decline induced by communication frac-hits from offsets, and provide contributing fracture dimensions and numerical production forecasts. To support the interpretations, we supplement the work by the results of 3D physics-based analytical modeling and multi-phase numerical simulations, and the findings are then validated using two extensive datasets: production profiles acquired by fiber optic DAS, and reservoir fluid fingerprints extracted from mud logs. Besides describing the evolution of seismicity during the treatment, the applied integrated fracture mapping process gives a more reliable and unique SRV structure that streamlines forward modeling and simulations in unconventional reservoirs as well as contributes to solving inverse problems more mechanistically.

The objective of the study is to develop an integrated multi-disciplinary workflow to screen wells for suitable candidates and to evaluate the feasibility of selected wells for refracturing using advanced geomechanical modeling. A four-step workflow to select and evaluate wells for refracturing was developed. First, hundreds of wells were screened based on age, spacing, and production performances, mainly IP90 and IP30 to IP360 ratio. Second, the selected wells from the first step were furthered narrowed down by analyzing their original completion and stimulation. Wells that were under-stimulated compared to historical average stimulation metrics, i.e. fewer stages/clusters and larger spacing, lower proppant loading, and less fluid volume pumped, were chosen as the final candidates for geomechanical modeling. Third, the wells were modeled with in-house advanced geomechanical fracture simulation and then depleted to obtain current stresses state. Finally, the technical feasibility of refracturing the wells was assessed. The workflow for best candidate selection has been improved considering volume of proppant placed, stress orientation, number of clusters / stages, reservoir quality, and production history. Due to the low horizontal stress contrast (1[MMPA]), fine scale stress determination was required. A 3D initial stress model was available from a previous study. From this stress state geomechanical hydraulic fracture mapping was performed. This geomechanical modeling utilizes Zero Thickness Element (ZTE) technology and the fractures only open if achieving the energy criteria. The stimulated volume and stress changes due to hydraulic fracturing were modeled with high resolution unstructured grids. Fracture driven interaction such as stress shadowing, cluster interference and fractures interconnection were observed. Production history was modeled with equivalent volume to determine the stress changes from the most accurate fracture geometry. The novelty is the systematic workflow integrating completion, production and geomechanics for selecting and evaluating the most suitable wells for refracturing, either stimulating new reservoir or reconnecting previous fractures. The advanced geomechanical techniques used to model the stresses at very fine scale and with unstructured grids are also state of the art.

This paper introduces a 3D hydraulic fracturing propagation model (3D-HFPM) for evaluating fracture extension, geometry, stress response, fracture spacing, and potential for refracturing. The model is illustrated with data of the Horn River shale of Canada. The model is developed using a combination of finite element method (FEM) and boundary element method (BEM) for evaluating fluid-flow, fracture deformation, and stress change in the reservoir. The model is calibrated using a limited amount of microseismic observations and recreate the fracture network when microseismic data are unavailable. An adapting meshing algorithm is incorporated to improve the capacity of the model to handle large and complex fracture networks such as the ones found in low permeability reservoirs. The continuity of fracture propagation and fluid leak-off during stimulation may be high enough to connect different production intervals and to create interference between stages, especially in wells with small path fracture spacing and multi-level completions. The comparison between the propagation model and microseismic data shows good agreement as the number of events increases as the fracture propagates into the reservoir. However, using only microseismic data to calculate the extension of the hydraulic fracture results in an overestimation of the fracture length. The model quantifies the altered stress zone, which is helpful to determine possible fracture reorientation and spacing. The evaluation of stress shadow and fracture reorientation reveals the advantages of refracturing using new over old perforations. The operation restores fracture conductivity and increases the fracture network as well as the drainage areas leading to an economic operation. The model improves the characterization of the Stimulated Reservoir Volume (SRV) in tight and shale reservoirs in those cases where microseismic data are scarce. Furthermore, the model is a viable tool for evaluating potential refracturing candidates.

For tight oil reservoirs, primary depletion using multistage hydraulically fractured horizontal wells typically only recovers between 5 to 15% of original oil in place, and can exhibit an exponential decline in production rates within a few months. Various enhanced recovery techniques exist, which offer great potential to increase the recovery in conventional reservoirs, however their feasibility within tight formations is not proven. This paper investigates the applicability of several different recovery techniques using a simulation-integrated workflow to both optimize the hydraulic fracture and well configuration design, and determine under which conditions these EOR methods are feasible. In addition to natural depletion, the EOR techniques investigated included immiscible and miscible gas flood and a cyclic Huff n Puff process. Using a simulation model representative of the Elm Coulee reservoir in the Bakken formation, a net present value optimization was performed with comprehensive economic analysis. Optimal operating constraints, fracture design and well configurations were found for each of the techniques investigated. Upon completion of the optimization, each optimal case was subjected to a sensitivity analysis on geological and economic parameters which cover the expected range of the Middle Bakken formation within the Elm Coulee region. Based on this uncertainty, a greater confidence can be achieved in each of the techniques as well as their versatility. From the simulation study, it was discovered that the optimal configuration was different for each of the recovery techniques evaluated. For natural depletion and immiscible flooding, the optimal well spacing was 400m combined with the largest fracture spacing of 125m. Conversely, the miscible flooding method achieved optimal NPV values at the closest well spacing. The cyclic process, which used CO 2 for the injected fluid, produced its optimal NPV at a closer fracture spacing when compared to the other 3 methods. Additionally, the Huff n Puff process required further optimization on cycle time to achieve the ideal conditions. The uncertainty analyses performed showed that for each of these recovery techniques the results varied significantly depending on the combination of reservoir, hydraulic fracture design and operating input parameters. Economically, the oil price was the most influential parameter, while geologically the connate water saturation and formation thickness carried the greatest impact. A change in conditions could greatly influence the application of the recovery techniques and possibly yield different optimum configurations. This paper provides an extensive workflow for optimizing EOR techniques and insight into their applicability in tight oil formations. It also demonstrates the importance of understanding the geological parameters prior to determining the ideal recovery technique and corresponding well configuration.

A requirement for appraisal of unconventional resource plays is the development of type wells: predictions of production results for future wells drilled within geologic subsets (areas with relatively homogeneous reservoir properties). Unfortunately, evaluators typically have a limited sample of wells available, so while the subset may be considered relatively homogeneous, the full range of reservoir properties may not have been sampled. Unconventional plays also have gradational changes in reservoir properties, so defining a "homogeneous" geologic subset implies an averaging of these gradients. This paper presents a workflow to deal with these issues. In this workflow, type wells are modified to account for the expected productivity difference between a sampled region and an unexplored region of a given area of interest. Since well productivity and key inputs (e.g. completion design, reservoir quality) are not perfectly correlated, statistical relationships are used to modify type wells while accounting for uncertainty. Combined with an economic model and mapped variability in geologic/reservoir properties, a regional map of break-even oil price is generated. This technique translates the gradational variation of multiple subsurface parameters into a continuous map of relative economic value, which can then be used to discuss a multitude of appraisal and development issues.

In this Paper, the methods developed by Clarkson et al. (2014) are used to analyze flowback data, which involves modeling flow both before and after the breakthrough of formation fluids. Despite the versatility of these techniques, achieving an optimal combination of parameters is often difficult with a single deterministic analysis. Because of the uncertainty in key model parameters, this problem is an ideal candidate for uncertainty quantification and advanced assisted history-matching techniques, including Monte Carlo (MC) simulation and genetic algorithms (GAs) amongst others. MC simulation, for example, can be used for both the purpose of assisted history-matching and uncertainty quantification of key fracture parameters. In this work, several techniques are tested including both single-objective (SO) and multi-objective (MO) algorithms for history-matching and uncertainty quantification, using a light tight oil (LTO) field case. The results of this analysis suggest that many different algorithms can be used to achieve similar optimization results, making these viable methods for developing an optimal set of key uncertain fracture parameters. An indication of uncertainty can also be achieved, which assists in understanding the range of parameters which can be used to successfully match the flowback data.

A semi-analytical formulation of the spontaneous capillary imbibition is used to analyze the liquid intake of six shale samples, linking imbibition capacity and rate to lab-scale measurements. Moreover, a data-driven approach is utilized to examine the effect of mineralogy and porosity on the macroscale wettability of shales. According to the results, the presence of connected organic sub-layers lowers the destructive impacts of spontaneous water imbibition on hydrocarbon permeability. Furthermore, the intrinsic permeability, tortuosity, wettability, and initial and residual saturations are among the most influential factors influencing the water uptake during shut-in periods after hydraulic fracturing operations.

This paper presents a novel approach in statistical analysis of well performance of multi-stage fractured wells completed in specific part of Montney. For the first time, both geological and hydraulic fracture parameters were considered in analysis with objective of finding most influential parameters that impact deliverability of wells. Results from single point statistical analysis were further extended to unsupervised discrete classification using neural-network to bin the reservoir. Binning classification was then used to design and optimize hydraulic fracture parameters.

The efficient utilization of automation systems necessitates a clear understanding of the interaction of the human operator, the automation system and any automated routines being run. If automated routines perform actions not desirable to the human operator, time is lost as the routine is interrupted and human control re-engaged. In addition, automatic handoff back to the human operator, both due to human intervention and due to exist conditions or anomalies must also be managed. Activity data from rigs across North America is analyzed to understand automation process utilization and interrupt timing. Realtime and historic data is tagged, either automatically, semi-automatically using machine learning, or manually, to create a minute-by-minute timeline of rig operations. Operations are then classified both by operation – steering, reaming, making hole, etc. – and well plan to understand how operational demands change automation system utilization. This results in a new set of metrics which can be used to precisely quantify the performance metrics of both the human and automated drilling systems. Performance of the automation system is found to be a strong function of hole deviation with the system outperforming during simple operations and in the vertical hole, but with reduced performance while in the curve and horizontal, due to high interruption of certain tasks. It is found that standard performance metrics, such as slip to slip or weight to weight are affected by standard practices and if these are used to grade system performance, these practices must be account for. This paper presents a detailed investigation of the interaction of the driller with an automated drilling automation system and lays out the utilization of the automation system as a function of rig operations and well path. It is specially noted that standard performance metrics must consider standard practices which may differ between operations.

Confidently establishing single well and/or aggregated production profiles, particularly estimated ultimate recovery (EUR), is both an important and challenging process in unconventional reservoirs. Numerous papers have proposed forecasting techniques, but four fundamental approaches dominate: Empirical decline curve analysis (DCA), such as multisegment Arps, the modified stretched exponential production decline (YM-SEPD) model, Duong's method, and power-law methods. Rate-transient analysis (RTA), which can include corrections for special dynamic mechanisms (e.g., stress-sensitivity, multiphase flow, adsorption/desorption). Numerical simulation for history-matching and forward modeling. Volumetrically determining in-place resources based on geological data and then applying a recovery factor deemed suitable for the reservoir system and depletion scheme. Each of these approaches can be implemented using either probabilistic or single-point estimates. Seidle et al. (2016) recently outlined recommended methodologies to accurately forecast time-series volumes and ultimate recovery in unconventional systems. A key recommendation in Seidle et al. (2016) is that the multiple approaches listed previously should be reconciled to establish confidence in forecasts and ultimate recovery estimates. However, Seidle et al. (2016) does not detail the means to achieve this goal. This paper establishes a methodology/workflow to reconcile the different types of empirical DCA methods, which should serve as a starting point for the ultimate goal of reconciling the four fundamental approaches listed. This paper compares and contrasts the DCA methods applied to various field cases in prominent Canadian [Western Canadian Sedimentary Basin (WCSB)] and US unconventional (oil and gas) plays (Bakken, Barnett, Cadomin, Eagle Ford, and Niobrara). Data sets are selectively chosen based on data quality and history length to increase confidence in the appraisal of the DCA approach. Hindcasting is applied to validate the results and conclusions. From analysis of a number of wells in different types of reservoirs, a new workflow (methodology) is proposed and validated with hindcasting that allows practically and accurately reconciling EURs based on various empirical methods. This manuscript is the first paper to discuss systematic reconciliation of EURs from varying approaches for horizontal wells in tight/shale reservoirs.

Standard laboratory techniques for determining petrophysical and geomechanical properties (e.g. porosity/permeability, Young's modulus, unconfined compressive strength) of tight rocks routinely require perfectly-cylindrical core plugs and/or large amounts (30+ g) of crushed-rock materials. Further, these measurements are time-consuming and expensive to perform. Therefore, indirect approaches for estimating rock petrophysical and geomechanical properties using small amounts of drill cuttings, which are usually the only reservoir samples available from multi-fractured horizontal wells (MFHWs), have recently received attention. Using an integrated, multidisciplinary approach that combines a customized image analysis software developed in-house with non-destructive microscopic techniques, practical laboratory-based workflows are generated to determine a variety of rock petrophysical properties including mineralogical composition, cementation, and porosity using drill cuttings. An "artificial" cuttings sample suite (core plugs crushed/sieved to 20/35 US mesh size), obtained from a prolific liquid-rich tight siltstone reservoir within the Montney Formation (Alberta, Canada), is analyzed in this study. Using a scanning electron microscope (SEM), back scattered electron (BSE) images, energy-dispersive X-ray spectroscopy (EDS) and cathodoluminescence (CL) images are collected to be used as inputs for an in-house image analysis software. Elemental maps obtained from EDS allow for the distribution of the mineral assemblage to be computed. Experimental observations indicate that rock petrophysical and geomechanical properties are partly controlled by the rock microstructure/microfabric in the studied Montney samples. Using a cathodoluminescence (CL) microscope and image processing, the detrital quartz grain is resolved from surrounding cement to determine the total percentage of cement in the samples. The cement content in the sixteen samples ranges from 14.5% to 23.7%, with an average overall cement content of 19.5%. For the analyzed samples, the porosity values estimated from microscopic images ranges from 3.5% to 10.4%, averaging 5.8%. The developed algorithms for indirect estimation of mineralogical composition, cementation and porosity from drill cuttings have significant practical applications for characterization of the Montney. Performing non-destructive microscopic observations using drill cuttings, these algorithms can be used as an alternative tool to provide quantitative estimates of rock fabric/texture and reservoir quality along any vertical/lateral intervals of interest within the Montney. In absence of core plugs and/or large amounts (30+ g) of crushed-rock material, the application of this integrated workflow could be of significant interest to the Montney operators, at least at the preliminary stage of stimulation treatment, to selectively target intervals along the vertical/lateral sections of the reservoir for improving performance.

Diagnostic Fracture Injection Test (DFIT) responses in some shale reservoirs, such as the Duvernay shale in western Canada, are not consistent with those interpreted through traditional analysis methods. Indeed, interpretation with traditional techniques may result in significantly incorrect estimates of closure pressure, pore pressure and formation permeability. The goal of this paper is to explain the observed DFIT behaviours for selected Duvernay shale wells in terms of low leakoff of fracturing fluid to the formation, activation of pre-existing fractures, and tip extension during the test. DFIT data in the Duvernay shale are analyzed using pressure transient analysis methods. Two scenarios are presented to explain the overall falloff behavior; moving-hinge closure with tip extension, and activation of secondary natural fractures. The validity of each scenario is examined using rigorous coupled flow-geomechanical simulation, geological information and geomechanical settings in the Duvernay Formation. Due to extremely low leakoff, the main mechanism affecting pressure falloff during the DFIT is pressure dissipation through the primary fracture created during injection. This results in significant tip extension or activation of secondary fractures. The fluctuations and spikes observed on G-function or pressure derivative plots are explained in the context of these scenarios. The leakoff rate varies with the pressure change, and the enhanced fracture surface area, during tip extension. Therefore, the assumption of Carter leakoff, and the traditional closure picks based on a straight-line tangent to the semi-log derivative on a G-function plot or 3/2 slope on Bourdet-derivative plot are not valid. Due to very low matrix permeability and the additional fracture length created through tip extension, it is unlikely that formation radial flow is established during the test, compromising the ability to obtain a valid pore pressure or formation permeability.