Abstract A physics-assisted deep learning model is presented to facilitate transfer learning in unconventional reservoirs by integrating the complementary strengths of physics-based and data-driven predictive models. The developed model uses a deep learning architecture to map formation properties to their corresponding production responses using a low-dimensional feature space representation. Transfer learning is accomplished by first training the network weights using production data from a mature shale play and combining the learned weights with limited data from a new unconventional field to generate a predictive model. The simulated data provides approximate production predictions for input parameters of the target field, for which the source data may not provide a good prediction. The resulting model has a superior performance to simulation-based and data-driven predictions alone. The results indicate that (1) physics-based simulated data can facilitate production predictions when out-of-range (unseen) input parameters have to extrapolate from data, and (2) transferring the weights learned from the source field to the target field can add valuable information to enhance the prediction performance of the target field. Introduction There has been a significant increase in the development of unconventional reservoirs in recent decades, particularly tight oil reservoirs in the United States. These tight oil formations often have extremely low permeability and are not economically exploitable using traditional drilling and completion techniques used for conventional high permeability reservoirs. Fortunately, due to the advancement in key technologies, in particular hydraulic fracturing and horizontal drilling, these low permeability shale formations can be stimulated to induce the production of hydrocarbons (King, 2010). However, the traditional flow equations that are well documented and studied for conventional wells have not been applied to tight oil reservoirs. Traditionally, the development of conventional reservoirs relies heavily on numerical reservoir simulators (Altman et al., 2020, Aziz et al., 1979). These simulators provide a reliable prediction that assists in decision-making for resource development. They are built using well-studied and trusted physics-based analytical expressions that are solved during the simulation to provide prediction responses. Since these physics-based simulation models are based on analytical equations, they have the advantage of being able to produce a prediction response for any possible range of input variables, and this allows them to easily extrapolate to any range of data. Additionally, simulation models can be run for any combination of input variables and enable the collection of large amounts of simulated data for all possible scenarios. However, the equations used in simulation models are not able to provide dependable predictions for unconventional tight oil resources since the complex physical relation of flow from tight formations and fractures along with the fracture generation is poorly understood. Additionally, there are various components in the field that are not always accounted for when building simulation models (Fung et al., 2016). These limitations make simulation models unreliable to be used to develop unconventional reservoirs.

Abstract The ultimate recovery factor in tight and shale resources is limited and is usually in the range of 5-10%. Although high-intensity fracturing and refracturing can increase recovery, an enormous amount of oil will remain in place, hence the desirability of enhanced oil recovery methods. Many of the shale reservoirs are oil-wet with negligible water uptake. By altering the wettability, water can spontaneously imbibe into the formation matrix, creating a counter-current flow that forces the oil out. Here, we assess the likelihood of increased oil recovery by modifying the fracturing-fluid composition (salinity and ion concentration) that transforms the formation wettability into a more water-wet state. Oil wetting of tight formations is usually controlled by the adhesion of oil droplets on the surface of clay minerals. When clay minerals are not predominant, the oil attached to carbonate minerals can significantly control rock wettability. In this study, we first identify the primary reactions that define the initial wettability of the rock depending on the formation mineralogy, formation water composition, and oil type. Second, we build a geochemical model considering the surface complexation of various minerals to mimic the wettability state of the reservoir. Third, we validate our method on zeta-potential, contact angle, and imbibition data from a previously published study using different fluids with different salinities. Finally, we present a mechanistic approach to model the wettability-alteration impact on spontaneous imbibition and compute the incremental oil recovery contributed by different fracturing fluid compositions. Based on the studied case, the oil adhesion to clay can be reduced by tuning the fracturing fluid salinity. The ionic concentrations of 2.5 wt.% of NaCl and 5.0 wt.% of CaCl 2 induced the smallest contact angles of 44.7 and 51.2 degrees. We observe that further brine dilution increases the contact angle. For example, distilled water shows the most oil-wet condition with a contact angle to 93.4 degrees. We argue that the main factors that maximize water wetting for the reported optimum salinities are the contrast in the rock and oil surface potential and the sodium concentration. Spontaneous-imbibition simulations indicate that the low salinity fluids promote a change in water-oil capillary pressure, leading to increased water uptake in the cores and improved oil recovery compared to distilled water. The agreement between the developed model and experimental data implies that the wettability in shale and tight formations can be quantitatively predicted and regulated.

Abstract Water-oil displacement is an important process that occurs in a shale matrix after hydraulic fracturing and in water-based enhanced oil recovery. Current understanding of this displacement process is limited because of the complicated pore structure and surface properties in shale. In this work, this process and its controlling factors are investigated through a comparative study of three shale samples that have different types of pore systems and wettability. An integrated method of imbibition and multiscale imaging was applied, and a modified oleic tracer that can better represent oil flow was used in imbibition testing and micro-CT imaging. SEM pore characterization was then performed under high magnification with guidance from the micro-CT images showing the changes due to oil or water imbibition. New insights were obtained on the importance of both wettability and pore size effect on oil recovery and the distribution of residual oil after water-oil displacement. Connectivity of pores with different wettability is also discussed on the basis of 3D analysis and SEM pore characterization. Collectively, these new findings improve the understanding of the complicated process of water-oil displacement and the role of influencing factors. Important implications for improved oil recovery strategy in shale are discussed for different types of reservoir rocks. The integrated imaging and imbibition technique provides a new path for further investigation of improved oil recovery in shale. Introduction Water-oil displacement is an important process that occurs in hydraulic fracturing and enhanced oil recovery with water-based agents in shale oil reservoirs. It is also relevant to water-oil distribution during oil generation and expulsion and after hydraulic fracturing, and consequently affects water-oil two-phase flow during production. Because of the complicated surface properties of the pore systems, unclear connectivity of pore networks with different wetting properties, and the small pores and low porosity, an understanding of water-oil displacement in shale and its controlling factors are still limited.

Abstract Theia Energy discovered a prospective unconventional hydrocarbon resource in the Ordovician Lower Goldwyer Shale (LGS) located on the Broome Platform of the onshore Canning Basin. The collation, processing, analysis and interpretation of all available regional data culminated in a successful exploration well, Theia-1 (drilled in 2015), which intersected a 230 ft gross oil column between 4,910–5,140 ft (based on slim-hole log and core analyses). Theia-1 recovered continuous core and wireline log data required to analyse and assess the geological properties necessary for a commercially viable shale oil and gas resource and quantify the volumes of hydrocarbons in-place. Petroleum system modelling carried out in the basin confirmed the prospectivity of the LGS and a potential additional mature hydrocarbon resource in the deeper Nambeet Formation some 2,000 ft below the LGS. The Nambeet is believed to exhibit higher maturity and extensive thickness reaching over 2,000 ft across a significant region of the Broome Platform. The Nambeet Formation adds substantial wet gas prospective resources in addition to those already discovered in the LGS. This paper outlines Theia Energy's exploration strategy in the onshore Canning Basin utilizing shale specific play elements from which the early exploration program was designed and assessed following the drilling of Theia-1. Subsequent specialized testing of core and independent expert analysis have confirmed the reservoir quality, charge, completion quality and producibility of the LGS. In particular, proppant embedment and fracture conductivity tests indicate exceptional flow characteristics of the LGS. The abundance of oil and gas prospective resources in the onshore Canning Basin is significant to the northwest region of Western Australia. The establishment of an unconventional petroleum resources industry would bring pivotal increased activity to its vacillating economy which is currently relying on mining, offshore petroleum development and tourism. Whilst the development and export of the resource would substantially add to economic opportunity for the region, the establishment of much-needed infrastructure would kickstart further economic opportunity such as: mining, downstream manufacturing including fertilisers, irrigated agriculture and cattle farming. Plans are under development for the implementation of major renewable energy projects in the region which can co-exist with an unconventional resource project.

Abstract With the strategic developments of unconventional shale oil and gas development, robust reservoir characterization algorithm is extremely crucial for the field engineers. Accurate description of hydraulic fracture geometry provides critical guidance regarding optimal well spacing and estimated ultimate recovery (EUR) forecasting, that has significant uncertainties if performed using rate transient analysis (RTA). The objective of this work is to demonstrate a novel numerical algorithm to explicitly calibrate the large uncertainties such as hydraulic fracture half-length; fracture heights; fracture water saturation; heel bias effects; etc., using embedded discrete fracture model (EDFM). The methodology can be briefly summarized as following. First, the workflow generates a large ensemble of initial uncertainty samples using Latin Hypercube (LH) method assuming uniform prior distributions. Then, these uncertainty samples are modelled numerically by EDFM and then are non-intrusively coupled with a reservoir simulator. Next, simulation results are extracted, and an advanced machine learning model is trained using these input/output pairs. Markov Chain Monte Carlo (MCMC) algorithm is subsequently employed to find suitable candidate samples from the resulting posterior distributions, which are fed back into the simulator again. After this iterative process, solution ensembles based on the selected criteria is filtered, and subsequent studies such as EUR forecasting are performed, which forms the general workflow of EDFM-artificial intelligence (EDFM-AI). The conclusions of this work highlight that traditional history matching may over-estimate fracture height and fracture permeability decay. Furthermore, heel bias effect observed in reality could provide more pressure supports from top/bottom layers and thus larger long-term EUR. Our workflow is also reliable and efficient that it can be effortlessly adapted to numerical analysis/characterization of any other shale oil/gas plays. Introduction The U.S. Energy Information Administration (EIA, 2021) estimated that crude oil production from shale/tight formations reaches approximately 2.67 billion barrels (nearly 65% of all US crude oil production) in the year of 2019. With the expansive development of oil recovery in low permeability reservoirs, operators are paying more attention to how do the wells perform in the long term, and a potential correlation with configurations of hydraulic fractures so that optimizations on frac job design could increase net return. However, quantitatively accurate determination of long-term estimated ultimate recovery (EUR) can be challenging. Traditional methods such as rate transient analysis (RTA), pressure transient analysis (PTA), and decline curve analysis (DCA) are all possible solutions to quantify long term EUR (Mishra et al., 2012; Wei et al., 2014; Zhukovsky et al., 2016; Beohar et al., 2017). However, these methods are considered either too time-expensive, lacking enough physical supports, and extremely unreliable for three phase flow under the current fast-paced and accuracy driven shale development environment.

Abstract This study evaluates the effect of nanopores on the production performance of liquid-rich shale reservoirs, which have become one of the most significant energy sources globally. Liquid-rich shale reservoirs are characterized by complex heterogeneous fracture networks and numerous nanopores. Traditional fluid flow models do not entirely capture the fluid flow behavior in porous media when a substantial hydrocarbon volume is stored in the nanopores. The flow mechanism in nanopores is different from the flow in large pores present in conventional reservoirs. This deviation in flow behavior is caused by the significant interactions between the hydrocarbon molecules and the pore walls. The hydrocarbon-wall interaction should not be ignored and must be accounted for in numerical simulators. In this study, we use a compositional reservoir simulator to model the reservoirs with different fluid properties and the distribution of nanopores. We integrate the deviated phase behavior into the reservoir model by shifting the critical properties based on extensive simulated results for simple and long-chain alkanes from a prior study. We compare the estimated ultimate recovery (EUR) of reservoirs at the end of two years for reservoirs with different degrees of pore confinement effect. The gas-oil ratio (GOR) for the reservoirs is also presented to describe the impact of the nanopores. The producing GOR for reservoirs with nanopores showed a significant deviation from the standard behavior expected for a conventional reservoir. Sensitivity analysis showed that the presence of nanopores in an unconventional reservoir could have a substantial impact on (EUR). The combined effect of initial fluid composition and nanopores on EUR is also evaluated and presented. The procedure presented in this study provides an effective method to incorporate the nanopores for better production performance evaluation and reserves estimation. The results show that the modified phase behaviors caused by the nanopores significantly affect the EUR from liquid-rich shale reservoirs.

Abstract Production forecasting sits at the core of reservoir engineering workflows, required for critical tasks across the asset lifecycle. Machine learning (ML) offers distinct advantages for production forecasting, including improved accuracy and ease of automation. In this paper, we present the results of a ML-based forecasting pilot test. ML-based approach shows accuracy on par with or better than Modified Arps method for both pre-drill (future) and post-drill (existing) wells. We discuss the advantages of the ML-based approach as well as our vision for an integrated, automated production forecasting workflow. We compared traditional Arps-based Decline Curve Analysis ("DCA") forecasts against ML-based forecasts. The latter utilized autoregressive (AR) and extra randomized trees (ETR) algorithms. This multi-variate ML forecasting workflow utilizes geology, spacing, completion variables, and historic production to forecast future production. We independently hyper-parameterized and trained models to forecast at 5- or 30-day intervals, for pre-drill and post-drill wells, and for each production stream (oil, gas, water). The production data in post-drill wells before the budget cycle date was split into training and test datasets; these robust models were then used to generate a series of forecasts aligned with budget cycles. Back-testing with production acquired after the budget cycle date provided further evidence that ML-based forecast, on average, outperforms or is on par with Modified Arps method. On the asset level, in the pre-drill wells, the Modified Arps method shows better performance in the first sixty days, but ML-based forecasts show improved accuracy beyond that. In the post-drill wells, the ML-based models show better accuracy for oil forecasts. For gas and water, the ML forecasts are more accurate over the first 90-120 days, though less accurate at later time. The outperformance of ML-based forecast is largely attributed to its multi-variate approach, which DCA methods are incapable of. The results from this pilot provide confidence to integrate ML methods in the quarterly/annual forecasting and reserves estimation processes. The work shows that ML-based solutions for production forecasting can match, and in certain cases exceed, the performance of simple curve fitting DCA method. ML-based methods for production forecasting, pad optimization, and budget planning are promising not only for their accuracy, but also for their speed, automation and low cost. This provides an opportunity for reservoir engineers to maximize their time on optimization workflows, which are critical in reducing costs and improving profitability of shale assets, where thousands of wells will be drilled to fully develop a field. This interdisciplinary work includes contributions from reservoir engineering, geoscience, data science, and software engineering.

Abstract CO 2 injection is a promising method for enhanced oil recovery (EOR) in unconventional liquid reservoirs. Addition of surfactants, which have been shown to improve oil recovery during water-based EOR, could also improve CO 2 EOR. Because oil and CO 2 are not fully miscible, even at pressures up to 9000 psi, oil extraction by CO 2 can be improved through addition of surfactants. The three surfactants evaluated in this study included linear alkyl ethoxylate Indorama SURFONIC ® L12-6, branched alkyl ethoxylate Indorama SURFONIC ® TDA-9, and branched phenol ethoxylate Indorama SURFONIC ® N-100. Surfactants were evaluated for CO 2 solubility, shale wettability alteration, ability to increase oil recovery during core soaking, and effect on CO 2 -oil interfacial tension (IFT). All three surfactants exhibited good solubility in CO 2 (roughly 1 wt% at pressures of 2000-5000 psi). Contact angle experiments demonstrated that the CO 2 +surfactant solutions can shift the wettability of shale toward water-wet. CO 2 and CO 2 +surfactant huff 'n puff experiments using oil-saturated Eagle Ford and Mancos shale cores showed that CO 2 solutions of SURFONIC ® TDA-9 gave highest cumulative oil recoveries of 72% and 75% (0.01 wt.% and 0.1 wt.% of TDA-9 respectively) compared to 71% recovery for the core immersed in pure CO 2 . The addition of surfactant to the CO 2 phase did not have a significant effect on CO 2 -oil IFT, suggesting that CO 2 -soluble surfactants can increase oil recovery by wettability alteration rather than IFT reduction. Introduction Oil from unconventional reservoirs has dominated U.S. domestic energy production in the past decade. However, oil extraction from shale reservoirs typically remains below 10%. The petroleum industry has had to rapidly adapt technology developed for porous conventional reservoirs to extremely tight unconventional reservoirs. CO 2 EOR is one oil extraction technique that works well for conventional reservoirs but is still being optimized for unconventional reservoirs. CO 2 is a relatively good solvent for oil and, in conventional reservoirs, its extraction ability is increased even further as it moves through the porous matrix collecting hydrocarbons-a process referred to as multiple contact miscibility (MCM). During CO 2 EOR in unconventional reservoirs, however, CO 2 is injected through high-permeability fractures and then allowed to soak into the shale matrix. MCM cannot be achieved because CO 2 does not flow through the matrix.

Abstract Due to the existence of nano-scale porous media in shale reservoirs, there are substantial rock-fluid interactions that can lead to a heterogeneous distribution of molecules and an alteration of the fluid PVT in these reservoirs. Molecular simulation is one of the most accurate techniques to model these rock-fluid interactions and their effect on hydrocarbon phase behavior. Most relevant molecular simulations to date have considered single-component or binary hydrocarbon mixtures, which may not fully interpret the compositional distribution and phase behavior in a shale multi-scale system. In this work, we apply the modified NPT-GEMC method to investigate the confinement effect on the phase behavior of an 8-component reservoir fluid in a multi-scale (nano-scale + macro-scale) media. Two similar cubic boxes are utilized to simulate the macro-scale porous media while an extra gauge cell with fixed volume is added to the system to simulate the nano-scale porous media. The simulation process follows the constant composition expansion (CCE) experiment, in which the fluid composition and temperature are constant while the pressure reduces in the process. The results show that the compositional heterogeneity exists between the fluids in macro-scale and nano-scale porous media and the fluid in the nano-scale region is always leaner compared with the fluid in the macro-scale region. In addition, the saturation pressure "disappears" during the simulation process due to the significant differences between the macro-scale and nano-scale fluid compositions and their continuous change with pressure. Our results indicate that a large amount of intermediate to heavy components are trapped in the nanopores, which causes the condensate yield much lower than expected. Introduction Liquid-rich shale (LRS) reservoirs belong to ultra-tight unconventional formations with an significant amount of hydrocarbons (Kanfar and Clarkson 2016). In recent years, the wide application of horizontal wells and multi-stage hydraulic fracturing has largely enhanced the recovery of LRS (Alharthy et al. 2018). One common issue observed in LRS production is that the oil recovery of LRS is much less compared with the oil recovery of a conventional reservoir with the same drawdown (Whitson and Sunjerga 2012). To understand this phenomenon, the phase behavior of hydrocarbons in LRS should be investigated.

Abstract The premise of this work is twofold. First, we postulate that existing methods to forecast the production performance of wells in unconventional reservoirs fail to recognize the nature of the evolution of the fluids produced. Mo techniques which are applied are variations or extensions of methods used to account for production when the reserve consist of mobile fluids existing in the effective porosity of the rock. In unconventional reservoirs the bulk of the molecules are thought to be contained in separate systems, either adsorbed onto the rock or stored in a very (very low permeability matrix, where these fluids can only be accessed by a created hydraulic fracture. Second, using this "two-system" concept, we propose a methodology that enables an engineer to use a simple formulation for modeling flowrate that can be used to quickly and reliably interpret and project production rate for wells in such reservoir systems. Computations in this work are based solely on time-rate data. This simplistic approach eliminates the need for knowledge of the formation and fluid properties, nor does one need to know the well completion geometry o configuration. In this work we present the formulation in terms of typical decline curve relations and parameter Analysis/interpretation of several wells are presented to demonstrate the methodology — projections of production rates and estimated ultimate recovery (EUR) are provided for guidance and validation of the methodology. Introduction Base Relations : The relations in this work are based on the work of Arps (1945), where the Arps' relations are "derived" from the work of Johnson and Bollens (1927) and are given as follows: Hyperbolic Decline Equation: (derived for the case of b = constant) (equation 1) Exponential Decline Equation: (derived for the case of D i = constant and b = 0) (equation 2) Harmonic Decline Equation: (exception case where b = 1) (equation 3) We note that the Exponential Decline relation (Eq. 2) can also be derived from fundamental principles for the case of a black oil and for a dry gas — these derivations are provided in Appendix A (black oil) and Appendix B (dry gas).

Abstract Well interference events, which are commonly referred to as frac hits, can significantly affect the production in fracture-stimulated horizontal shale wells. Increased infill drilling in recent years places more emphasis on better understanding of the causes and effects of frac hits. This work aims to a better understand the prolonged production effects of sustained frac hits. The data analytics approach used extensive field datasets which included well trajectories, child well completions data, parent well pressure data and parent well production data. The methodology was to first characterize every induced pressure change in a parent well (i.e. to determine frac-hit intensity), and subsequently to estimate the average daily production change after each frac hit. This was accomplished by utilizing a frac hit auto-detection algorithm and a log-log analysis of daily production. To add the effect of interference time (i.e., the time of the pressure change caused by the frac hit), a metric named the "pressure-change integral (PCI)" was calculated for each parent well. Finally, the effects of frac-hit intensity and interference time on parent well production were quantified through correlation analysis. The datasets covered 50 horizontal hydraulically fractured wells in the Eagle Ford shale formation and the Bakken formations. The results show that parent well production is not only influenced by the total magnitude of frac hits, but also by the time duration of the overall pressure interference. For the Eagle Ford wells in our datasets, the analysis shows that a larger PCI resulted in a higher average production increase post-hit. This paper presents a novel data-analytics approach to quantify the effect of different types of frac hits on parent well production by introducing a new metric named the PCI, which combines the effects of frac-hit intensity and interference time. This approach can be used to better understand the frac hits detected in a parent well, and to more accurately estimate the possible impact these frac hits may bring to future production. Based on such information, it may be possible in the future to actively manipulate frac hits for production gain.

Abstract We report the design, drilling, completion, flow back and production of lacustrine shale oil deposit (LSOD) lateral well in Guangdong block, Kongdian Formation in the Cangdong Sag, China. Lacustrine shale oil from China features higher viscosity and density, low and fast varying GOR. The LSOD laterals are further complicated by the fast varying local geology. We compare the key deposit and development features of the shale oil reserve in Cangdong Sag with other lacustrine reserves in China as well as the predominately marine shale oil reserves in North America. The difference between marine shale deposit in north America and lacustrine shale deposit in China are compared from completion design to production planning as well. Key parameters for completion design including the lateral length, number of stages and cluster spacing, fracing fluids and proppants loading per unit length, and key resulting post-frac parameters, such initial bottom hole shut-in pressure and flowing pressure, initial production, GOR of production as well as production curve over a period of more than 500 days are presented and compared to typical shale oil wells in lacustrine shale oil deposits as well as marine shale oil deposits in North America. Introduction Just as lacustrine conventional oil dominates the oil reserves in China, lacustrine unconventional oil deposits dominate shale or tight oil. Unlike marine shale oil in the US, lacustrine shale oil deposits in China have quite unique geochemistry, and also greatly varying maturities and deposit environments. The deposit environment affect the kerogen significantly. For example, high salinity deposit usually have oil-prone type I kerogen with significantly delayed and suppressed gas window and therefore high oil with low gas yield. The oil generated also features high amounts of cyclic ring hydrocarbons, i.e. naphthenic content. This includes the heaviest shale oil ever reported, the Jimusaer Lower Permian deposit has density as high as 0.92g/cm 3 or API 20, and gas to oil ratio (GOR) as low as 15m 3 /m 3 or 90CF/bbl. The maturity characterized by vRo, Tmax and gas isotope measurement ranges from 0.80 to 1.05. The corresponding typical marine shale oil deposit usually have density at 0.84~0.8g/cm 3 or API 36~45, and GOR from 3,000CF/bbl to over 2,000CF/bbl.

Abstract Different chemical additives (friction reducer, scale inhibitor, surfactant, etc.) can be injected into shale reservoirs during hydraulic fracturing operations/production stages to help generate fracture networks, enhance well productivity, and mitigate production risks. This paper presents a comprehensive testing approach simulating field operating conditions to optimize the selection and application of chemical additives (especially friction reducer) in shale reservoirs and prevent severe solid deposition due to undesired chemical-fluid-rock interaction at subsurface conditions. The compatibility between friction reducer (FR) with multiphases (water, oil and formation rocks) and other chemicals are considered and evaluated under simulated field conditions. Systematic laboratory experiments are designed and conducted to understand the impact of various factors on chemical stability and compatibility. The impact of various factors including but not limited to the presence of crude oil, type of crude oil, presence of cuttings, type of cuttings, presence of ferric iron, ferric iron concentration, type of FRs, concentrations of FR, temperature, pH, and reaction time were investigated. The tested friction reducer products can be categorized into two types. For Type 1 chemicals, presence of ferric iron can lead to incompatibility with brine and oil. For Type 2 chemicals, presence of ferric iron or shale rock individually has limited effect on stability/compatibility. Testing results show the co-presence of shale and ferric iron can dramatically impact the friction reducer stability and compatibility under subsurface conditions. Shale rock and ferric ions are included in the evaluation tests of chemical additives to simulate real subsurface conditions. The application of this novel testing approach is expected to significantly optimize chemical additive selection/application, eliminate/minimize subsurface solid deposition and corresponding formation damage during hydraulic fracturing and shut-in stages due to chemical-fluid-rock interactions. This study integrates the aspects/considerations from both completion and production stages. Introduction Fracturing fluids have been widely used f or shale and tight assets developments. It has been applied for various unconventional field developments in North America such as Marcellus, Bakken, Permian and across the world's unconventional fields. (Kurdi, et al., 2020; Karadkar, et al., 2020; Jonhson, et al., 2018; Geri, et al., 2019). With growing maturity of hydraulic fracturing, it has also been used for deepwater assets to enhance production (Song, et al., 2020). In general, water-based fracture fluid along with proppant is introduced into the formation at high pressures and high flow rate to break formation rock and form artificial fractures. There are several components in fracturing fluid mixtures including but not limited to: friction reducer, biocide, scale inhibitor, and corrosion inhibitor. As the main component, friction reducer (FR) can reduce the drag force between fracturing fluid and piping surface which leads to energy saving and OPEX reduction. The goal of FR application is to use low dosage (500-2000 mg/L) to prevent 70-80% of frictional losses while carrying proppant to the newly formed fracture network.

Abstract Hydrocarbon production from shale reservoirs is inherently inefficient and challenging since these are low permeability plays. In addition, there is a limited understanding of the fundamentals and the controlling mechanisms, further complicating how to optimize these plays. Herein, we summarize our past and current efforts to reveal the shale fundamentals and devise development strategies to enhance extraction efficiency with a minimal environmental footprint. Integrating these fundamentals with machine learning, we outline a pathway to improve the predictive power of our models, enhancing the forecast quality of production, thereby improving the economics of operations in unconventional reservoirs. For instance, we have developed science-informed workflows and platforms for optimizing pressure-drawdown at a site, which allow operators to make reservoir-management decisions that optimize recovery in consideration of future production. Recently, our work relies on the hybridization of physics-based prediction and machine learning, whereby accurate synthetic data (combined with available site data) can enable the application of machine learning methods for rapid forecasting and optimization. Consequently, the workflow and platform are readily extendable to operations at other sites, plays, and basins. Introduction Shale reservoirs have redefined the energy landscape of the world (Melikoglu, 2014). Shale reservoirs currently contribute to 70% of U.S. natural gas production and 60% of U.S. oil production (Perrin & Geary, 2019). However, these reservoirs possess unique characteristics that entail custom-designed development plans (Mehana & El-monier, 2016). Currently, we recover less than 10% of the hydrocarbons in place from these low permeability plays with the most efficient development plan (Alharthy et al., 2015). This low efficiency is also due to the limited understanding of shale characteristics' effects on the fluid properties and transport through porous media. Therefore, a fundamental understanding and quantification of these effects are required for devising better development plans to improve the current recovery factors (Middleton et al., 2017).

Abstract The US shale revolution in the past decade has doubled the country's crude production, and tight clay-rich formations account for 60% of the US crude production. Nevertheless, our knowledge of petroleum system controls on the reservoir performance of these tight shale systems remains poor. We report on a methodology to utilize production data to compare and rank the US shale oil plays in conjunction with geologic descriptions of the plays. For this study, we have used 36,642 horizontal oil wells from 12 tight shale formations in the Rockies and Midwest basins with vertical depths ranging from 5,000-15,000 feet and with at least 24 months of production to rank these plays based on their performance. These formations include: Wolfcamp (Delaware and Midland/Permian), Bone Spring (Delaware/Permian), Eagle Ford (Gulf Coast West and Central/ Cretaceous), Niobrara (Denver and Powder River/Cretaceous), Bakken (Williston/Devonian), Austin Chalk (Gulf Coast West/Cretaceous), Woodford (Anadarko/Devonian), Spraberry (Midland/Permian), and Barnett (Fort Worth/Mississippian). Their production data were normalized for cumulative oil (STB/ft/month), and the results were then analyzed in light of geologic and geochemical data from the formations. Integrated production and geologic-geochemical data on the tight formations offer valuable insights into the control of petroleum system elements on production patterns. These comparative patterns were used for ranking the shale plays for various parameters. In terms of cumulative oil production (STB/ft/month), Wolfcamp (Delaware) ranks top (0.93 STB/ft/month) while Barnett (Fort Worth) has the lowest crude production (0.07 STB/ft/month). In terms of GOR changes during production, Wolfcamp (Delaware) shows the lowest change (1.38 times), while Barnet (Fort Worth) has the highest change (13.37 times). These may be related to the oil-prone nature of Wolfcamp and the gas-prone characteristics of Barnet. Overall, deeper shale plays yield more oil (per foot per month) than the shallower plays. Some plays exhibit intra-formational migration of hydrocarbons. The results and the methodology of this study provide a multi-disciplinary geo-engineering to characterize shale oil plays in various basins. Variability in the performance of shale oil plays given by production data can be reverse engineered to petroleum systems.

Abstract This study corresponds to the second part of a paper series detailing the evolving approaches developed to improve project economics in Vaca Muerta shale formation ( Lerza et al., 2020 ). The main objective of this work is to develop a methodology that overcomes the well-known limitations of currently existing workflows that attempt to optimize well spacing and well completion design for maximizing project value in shale and tight (S&T). Currently published workflows on this topic generally fail to capture the interactions existing between well spacing and completion design, therefore the results obtained from applying those work steps are only optimal under a given completion design or well spacing assumption. However, most of the times, we are interested in finding the unconstrained global optimal solution rather than the local optimal solution offered by current methodologies, for which a simultaneous study of well spacing and completion design is required. The proposed methodology consists of applying fracture and reservoir simulation to calibrate a base model by history matching fracture and production data. Next, Design of Experiments technique is leveraged to ensure most of the interaction across the previously selected parameters are correctly captured while minimizing the number of required runs. Lastly, a machine learning model is built and tested from the gathered reservoir simulation results, and later used to efficiently obtain the predicted results for 1 million possible development alternatives comprising different combinations of completion and well spacing parameters. Finally, from the vast number of scenarios run, the one that offers the desired value of the target value metric and is also operationally achievable gets selected. By following the suggested methodology, the asset development team was able to find and propose an operationally executable development alternative that is expected to almost double the current wells Expected Ultimate Recovery (EUR) per section, while reducing the current discounted unit development cost between 10 and 30%.

Abstract Field and laboratory observations to date indicate that the efficiency of hydraulic fracturing, as it relates to hydrocarbon recovery, depends significantly on geochemical alterations to rock surfaces that diminish accessibility by partial or total plugging of the pore and fracture networks. This is caused by mineral scale deposition such as coating of fracture surfaces with precipitates, particle migration, and/or crack closure due to dissolution under stress. In reactive flow-through experiments, mineral reactions in response to acidic fluid injection significantly reduced system porosity and core permeability. The present study focuses on changes to fluid chemistry and shale surfaces (inlet and fracture walls) resulting from shale-fluid interactions and integrating these findings for an improved estimate of transport-related consequences. The reacted shale surfaces were examined by spatially-resolved scanning electron microscopy - energy dispersive spectroscopy (SEM-EDS) analysis. Importantly, inductively coupled plasma - mass spectrometry/optical emission spectroscopy (ICP-MS/OES) was utilized to probe the chemical evolution of the core-flood effluents. The three study cores selected from the Marcellus formation represent different mineralogies and structural features. In flow-through experiments, lab-generated brine and HCl-based fracture fluid (pH=2) were injected sequentially under effective stress (up to 500 psi) at reservoir temperature (80°C). SEM-EDS results confirmed by the ICP concentration trends showed significant Fe hydroxide precipitates in clay- and pyrite-rich outcrop samples due to partial oxidation of Fe-bearing phases in the case of intrusion of low salinity water-based fluids. Porosity reduction in the MSEEL (Marcellus Shale Energy and Environmental Laboratory) carbonate-rich sample is related to compaction of cores under stress due to matrix softening with substantial dissolution, and pore-filling by hydroxides, as well as barite and salts. Despite the same fluid compositions and experimental conditions used for both MSEEL samples, barite precipitation was much more intense in the MSEEL clay-rich sample due to its greater sorption capacity and additional sulfate source as well as fissile nature with multiple lengthwise cracks. ICP tests revealed time-resolved concentration trends in produced brine and reactive fluids that in turn complemented the pre-/post-reaction SEM-EDS observations.

Abstract Objectives/Scope: Understanding drainage dynamics of hydraulically stimulated horizontal wells through time is among the most critical parameters for optimizing field development strategies in unconventional reservoirs. Several years ago, we established a novel geochemical-based methodology using production allocation of produced oils collected in time-series and calibrated to core extracted oils to cost effectively ascertain vertical drainage. While the core-calibrated methodology has proven successful, it requires time-consuming project-specific model building, uncontaminated proximal core, depth shifting and triangulation, and parameter restriction to the extracted fraction. To overcome these challenges, an innovative processing technology was developed that exploits data analytics and machine-learning of the existing Time-Lapse Geochemistry (TLG) database. Methods/Procedures/Process: We created a 3-D full field compositional model across ConocoPhillips' Eagle Ford acreage using high confidence produced oils. The best field-wide compositional parameters distinguishing drilling targets were determined by rigorous statistical filtering with principal components and cluster analyses on uniformly aligned and indexed gas chromatography data. Oil compositions per landing zone were back-calculated and matched to contributions by layer in each oil mixture that were known a priori from previously core-calibrated allocations. Compositions vertically in-between drilling targets were interpolated from the known landing zones. Multi-layer compositional depth profiles were generated at each sampled project and extrapolated across the field using correlation gridding with geological parameters derived from a calibrated 3-D basin model. Results/Observations/Conclusions: Mapped geochemical parameter grids were stacked and amalgamated to construct the 3-D compositional grid system, which forms a field wide 17-layer vertical profile of geochemical endmembers. An integrated platform was built to visualize the compositional model in combination with indexed TLG oils. Python was used to digitize the compositional model and explore multiple machine learning algorithms as a back check on grid system quality. Random TLG samples served as blind validation data. We identified a random forest ensemble regression algorithm from the Scikit-Learn package as the optimal model to forward predict the 3-D location of every produced oil. Production allocations were further assisted by a random forest classifier that calculated prediction probabilities for drainage contributions from each layer. The machine learning toolkit aids in selecting appropriate endmember depth profiles and provides optionality for running production allocations through either the numerical mixing optimizer or machine learning algorithm. Applications/Significance/Novelty: This is the first example of a 3-D geochemical model of produced oil compositions generated for an unconventional reservoir. This next-generation workflow couples a consistent and robust grid system across the field with machine learning assistance made possible by data analytics of a high density, high resolution TLG dataset. Access to a wider molecular range, expanded spatial reach away from core control, and dramatically reduced allocation run times have greatly increased production allocation versatility. Tremendous business value is created by having the ability to quickly ascertain vertical drainage dynamics across multiple well configurations, completions, enhanced recovery projects, and infill drilling as feedback for long-term strategy decisions in addition to incorporating geochemistry with any multivariate investigation of subsurface challenges.

The objective of this work is to predict the estimated ultimate recovery (EUR) and rank locations following application of the Digital Analogue Shale (DAS) model to exploration data types (character, geomechanics, quantity, quality, maturity, mineralogy) across shales worldwide. To address this objective, we developed the reduced order DAS model using a machine-learning (ML) approach for rapid determination of EUR with applications to independent data sets including the East Coast Basin, NZ and the Reconcavo Basin, BR. A priori knowledge of the EUR will save money and reduce time by prioritizing economic investments and returns in unconventional shale assets. Development of the DAS model follows a ML workflow - training, feature selection, testing, and uncertainty quantification. The unsupervised ML network involves competitive training and self-organization of publicly-available reservoir data from unconventional shale plays in the USA (e.g. Barnett, Bakken, Eagle Ford, Haynesville, Marcellus, Niobrara, others) that include character (depth and thickness), geomechanics (porosity, permeability, Poisson ratio, and Young’s modulus), quantity (free hydrocarbons, amount of hydrocarbons generated through thermal cracking, and total organic carbon, EUR), quality (hydrogen index), maturity (maximum temperature and vitrinite reflectance), and mineralogy (clay content, carbonate content, and silica content). Minimization of quantization and topographical error vectors provide EUR predictions for testing generalizability and quantifying uncertainty by stochastic cross-validation. The DAS model provides unbiased EUR predictions and their uncertainty estimates when applied to independent shale data. Differences in average and median EUR predictions reveal a nonlinear process underscoring the importance of using the unsupervised ML approach to develop the DAS model. Given the range of estimation uncertainty, the preferred DAS model predictions (closest to observations) are median EUR values. We successfully applied the DAS model to quantify amount and rank locations of EUR across three structural blocks in the Renconcavo basin; and by block, formation and member in the East Coast Basin. The DAS model represents an innovative step-change toward near real-time determination of EUR with quantifiable uncertainty anywhere that (sparse) exploration data are available. By ranking EUR predictions, the DAS model facilitates rapid prospect generation consistent with world-class shale plays. As new reservoir data become available, the DAS model can be refined and redeployed across unconventional shale assets. The DAS model can be applied using only sparse data from exploratory wells and can be applied at an existing field to identify targets for infill drilling. Preliminary tests demonstrate model ability to predict (in addition to EUR) economic considerations, such as well cost, operational well cost, recovery, volume of drilling water per well, volume of fracking water per well. Lastly, the rapid deployment of the DAS model can provide guidance to governments for future bid rounds, exploration planning and prioritization.

The research goal is the development of a novel integrated solution of formation water content and salinity determination for petrophysical characterization. The workflow relies on three techniques: evaporation method (EM) with isotopic composition analysis, analysis of water extracts, and cation exchange capacity (CEC) study. The EM offers a fast, efficient, and accurate measurement of the residual water content with breakdown into free and loosely clay-bound types. The isotopic composition reveals the origin and genesis of pore water. The chemical analysis of water extracts delivers a lower bound salinity in terms of NaCl. CEC describes rock-fluid interactions. The workflow is applicable for tight reservoir rock samples, including shales and source rocks. A representative collection of rock samples is formed based on the petrophysical interpretation of well logs for a complex tight gas reservoir rock of the the Bazhenov formation (West Siberia, Russia). The evaporation method employs the retort principle but delivers much more accurate and reliable results. The suite of auxiliary laboratory methods includes derivatography, Rock-Eval pyrolysis, XRD. Water extracts from the rock samples at natural humidity deliver a lower bound for mineralization (salinity) of formation water. Isotopic analysis of the evaporated water samples covered δ 18 O and δD. A modified alcoholic NH 4 Cl method provides CEC and exchangeable cations concentration of the rock samples with low carbonate content. The target rock samples contained residual formation water 0.11–4.27 wt.%, including free 0.04–3.92 wt.% and loosely clay-bound water 0.09–0.96 wt.%. The loosely bound water content correlates well to the clay mineral fraction. The estimated pore water salinity reached tens of grams per liter; the corresponding isotopic composition indicated the deep formation genesis and generally correlated to that of the deep stratal waters of West Siberia. The amount of chemically bound water fell in a range of 0–6.40 wt.% and exceeds that of free and loosely bound water. The isotopic composition proved the formation origin of the extracted pore water samples. CEC falls in 1.5–4.73 cmol/kg and depends on the clay content. The study made a qualitative step up towards the quantitative characterization of formation water in shale reservoir rocks. Research effort delivered an integrated workflow for reliable determination formation of water content, salinity lower bound, and water origin. The results fill the knowledge gaps in the petrophysical interpretation of well logs, as well as general reservoir characterization and reserves estimation. The research novelty is in using a unique suite of laboratory methods adapted for tight shale rocks with the initial water content of less than 1 wt.%.