The new-generation oil-base mud (OBM) microresistivity imagers provide photorealistic high-resolution quantified formation imaging. One of the existing interpretation methods is based on composite processing providing an apparent resistivity image largely free of the standoff effect. Another one is the inversion-based workflow, which is an alternative quantitative interpretation, providing a higher quality resistivity image, button standoff, and formation permittivities at two frequencies. In this work, a workflow based on artificial neural networks (NNs) is developed for quantitative interpretation of OBM imager data as an alternative to inversion-based workflow.
The machine learning approach aims to achieve at least the inversion-level quality in formation resistivity, permittivity, and standoff images an order of magnitude faster, making it suitable for implementation on automated interpretation services as well as integration with other machine learning based algorithms. The major challenge is the underdetermined problem since OBM imager provides only four measurements per button, and eight model parameters related to formation, mud properties, and standoff need to be predicted. The corresponding nonlinear regression problem was extensively studied to determine tool sensitivities and the combination of inputs required to predict each unknown parameter most accurately and robustly. This study led to the design of cascaded feed-forward neural networks, where one or more model parameters are predicted at each stage and then passed on to following steps in the workflow as inputs until all unknowns are accurately obtained.
Both inverted field data sets and synthetic data from finite-element electromagnetic modeling were used in multiple training scenarios. In the first strategy, field data from few buttons and existing inversion results were used to train a single NN to reproduce standoff and resistivity images for all other buttons. Although the generated images are comparable to images coming from inversion, the method is dependent on the availability of field data for variable mud properties, which at the moment limits the generalization of the NNs to diverse mud and formation properties.
In the second strategy, we utilized the synthetic responses from a finite element model (FEM) simulator for a wide range of standoffs, formation, and mud properties to develop a cascaded workflow, where each stage predicts one or more model parameters. Early stages of the workflow predict the mud properties from low formation resistivity data sections. NNs then feed the estimated mud angle and permittivities at two frequencies into next stages of the workflow to finally predict standoff, formation resistivity, and formation permittivities. Knowledge of measurement sensitivities was critical to design the efficient parameterization and robust cascaded neural networks not only due mathematically underdetermined nature of the problem but also the wide dynamic range of mud and formation properties variation and the measurements. Results for processed resistivity, standoff, and permittivity images are presented, demonstrating very good agreement and consistency with inversion-generated images. The combination of two strategies, training on both synthetic and field data, can lead to further improvement of robustness allowing customization of interpretation applications for specific formations, muds, or applications.
Relative permeability (kr) functions are among the essential data required for the simulation of multiphase flow in hydrocarbon reservoirs. These functions can be measured in the laboratory using different techniques including the steady state displacement technique. However, relative permeability measurement of shale rocks is extremely difficult mainly because of the low/ultralow matrix permeability and porosity, dominant capillary pressure and stress-dependent permeability of these formations.
In this study, the impacts of stress and capillary end effects (CEE) on the measured relative permeability data were investigated. The steady state relative permeability (SS-kr) measurements were performed on Eagle Ford and Pierre shale samples. To overcome the difficulties regarding the kr measurements of shale rocks, a special setup equipped with a high-pressure visual separator (with an accuracy of 0.07 cc) was used. The kr data were measured at different total injection rates and liquid gas ratios (LGR). In addition, to evaluate the impacts of effective stress, the kr data of an Eagle Ford shale sample were measured at two different effective stresses of 1000 and 3000 psi.
From the experimental data, it was observed that the measured SS-kr data of the shale samples have been influenced by the capillary end effects as the data showed significant variation when measured at different injection rates (with the same LGR). This suggested that the liquid hold-up (i.e. capillary end effects) depends on the competition of capillary and viscous forces. In addition, it was shown that it is more necessary to correct the experimental kr data measured at the lower LGRs. Furthermore, different relative permeability curves were obtained when the kr data were measured at different effective stresses. This behavior was explained as the capillary pressure was expected to be more dominant at the higher effective stress.
The results from this study improve our understanding of unconventional mechanisms in shale reservoirs. It is evident that the behavior of unconventional reservoirs can be better predicted when more reliable and accurate relative permeability data are available. The outcomes of this study will be useful for accurate determination of such kr data.
Recent studies have indicated that Huff-n-Puff (HNP) gas injection has the potential to recover an additional 30-70% oil from multi-fractured horizontal wells in shale reservoirs. Nonetheless, this technique is very sensitive to production constraints and is impacted by uncertainty related to measurement quality (particularly frequency and resolution), and lack of constraining data. In this paper, a Bayesian workflow is provided to optimize the HNP process under uncertainty using a Duvernay shale well as an example.
Compositional simulations are conducted which incorporate a tuned PVT model and a set of measured cyclic injection/compaction pressure-sensitive permeability data. Markov chain Monte Carlo (McMC) is used to estimate the posterior distributions of the model uncertain variables by matching the primary production data. The McMC process is accelerated by employing an accurate proxy model (kriging) which is updated using a highly adaptive sampling algorithm. Gaussian Processes are then used to optimize the HNP control variables by maximizing the lower confidence interval (μ-σ) of cumulative oil production (after 10 years) across a fixed ensemble of uncertain variables sampled from posterior distributions.
The uncertain variable space includes several parameters representing reservoir and fracture properties. The posterior distributions for some parameters, such as primary fracture permeability and effective half-length, are narrower, while wider distributions are obtained for other parameters. The results indicate that the impact of uncertain variables on HNP performance is nonlinear. Some uncertain variables (such as molecular diffusion) that do not show strong sensitivity during the primary production strongly impact gas injection HNP performance. The results of optimization under uncertainty confirm that the lower confidence interval of cumulative oil production can be maximized by an injection time of around 1.5 months, a production time of around 2.5 months, and very short soaking times. In addition, a maximum injection rate and a flowing bottomhole pressure around the bubble point are required to ensure maximum incremental recovery. Analysis of the objective function surface highlights some other sets of production constraints with competitive results. Finally, the optimal set of production constraints, in combination with an ensemble of uncertain variables, results in a median HNP cumulative oil production that is 30% greater than that for primary production.
The application of a Bayesian framework for optimizing the HNP performance in a real shale reservoir is introduced for the first time. This work provides practical guidelines for the efficient application of advanced machine learning techniques for optimization under uncertainty, resulting in better decision making.
The present study provides a comprehensive set of new analytical expressions to help understand and quantify well interference due to competition for flow space between the hydraulic fractures of parent and child wells. Determination of the optimum fracture spacing is a key factor to improve the economic performance of unconventional oil and gas resources developed with multi-well pads. Analytical and numerical model results are combined in our study to identify, analyze, and visualize the streamline patterns near hydraulic fractures, using physical parameters that control the flow process, such as matrix permeability, hydraulic fracture dimensions and assuming infinite fracture conductivity. The algorithms provided can quantify the effect of changes in fracture spacing on the production performance of both parent and child wells. All results are based on benchmarked analytical methods which allow for fast computation, making use of Excel-based spreadsheets and Matlab-coded scripts. Such practical tools can support petroleum engineers in the planning of field development operations. The theory is presented with examples of its practical application using field data from parent and child wells in the Eagle Ford shale (Brazos County, East Texas). Based on our improved understanding of the mechanism and intensity of production interference, the fracture spacing (this study) and inter-well spacing (companion study) of multifractured horizontal laterals can be optimized to effectively stimulate the reservoir volume to increase the overall recovery factor and improve the economic performance of unconventional oil and gas properties.
This study is based on the premise that most of the trapped hydrocarbons can be produced, if we substitute them with another ‘acrificial’ fluid that has amplified interactions with organic pore walls, such as CO2. For the presented study, a downhole shale sample is analyzed in the laboratory to predict gas storage properties such as pore-volume, pore compressibility, and gas adsorption capacity. Then a series of pressure pulse decay measurements are performed to delineate transport mechanisms and predict stress-sensitive permeability. These coefficients are obtained as the calibration parameters of a simulation-based optimization for injection and production. Simulation model considers compositional gas flow in a deformable porous media and includes a multi-continuum porosity, with organic and inorganic pores, and micro-fractures. The experimental and simulation results show that most of the injected CO2 is adsorbed in the organic matrix and are not produced back. This is because CO2 molecules have significantly larger adsorption capacity when compared to methane. The strong adsorption of CO2 improves the release of natural gas from kerogen pores. This indicates that the separation of produced CO2 will be a minimal cost. Transport in kerogen has significant pore wall effects, and includes large mass fluxes of the adsorbed molecules by the walls due to surface diffusion. In essence, the adsorbed CO2 molecules significantly influence transport of methane. The results also show core-plug permeability is stress-sensitive due to presence of micro-fractures. Forward simulation results using optimum parameters indicate that closure stress developing near the fractures could significantly control the volume of CO2 injected. This raises operational issues on when to start injecting, and how to inject CO2. Using a simulation study of a production well with single-fracture, we show that fracture closure stress develops rapidly and production rate becomes a slave of the fracture geo-mechanics, e.g., strength of the proppants and the level of proppant embedment.
Lin, Ran (Southwest Petroleum University) | Ren, Lan (Southwest Petroleum University) | Zhao, Jinzhou (Southwest Petroleum University) | Tao, Yongfu (Exploration and Development Research Institute, Yumen Oilfield Company) | Tan, Xiucheng (Southwest Petroleum University) | Zhao, Jiangyu (Southwest Petroleum University)
Multi-stage & multi-cluster fracturing in horizontal well drilling is the core technology in for commercial exploitation of shale gas resevoir. According to vast field data, there is remarkable positive correlation relationship between stimulated reservoir volume (SRV) and shale gas production. Hence, estimating the SRV is essential for both pre-fracturing design and post-fracturing evaluation. However, the forming process of SRV involves with many complex mechanisms, making it is difficult to be simulated.
In this paper, we establish a mathematical model to estimate the SRV by simulating multiple hydraulic fractures propagate, formation stress change and reservoir pressure rise; consequently, the stress and pressure change might make natural fractures occur tensile failure or shear failure, generating a high-conductivity zone (i.e., SRV) in the shale reservoir.
To solve the model, displacement discontinuity method (DDM) is applied to simulate non-planar propagation of multiple hydraulic fractures and calculate formation stress change. Finite difference method (FDM) is used to compute reservoir pressure rise. The natural fractures failure state is determined by tensor formulae derived from Warpinski's failure theory. This SRV estimation method involves a variety of complex but crucial physical mechanisms during shale fracturing process which include unequal flow-rate distribution in different hydraulic fractures, non-planar hydraulic fractures propagation under stress interference, reservoir permeability increases with SRV expanding, two types of natural fracture failure and so on.
A field case study was performed to show the dynamic processes of hydraulic fractures propagation, reservoir permeability increase, and the SRV expansion during shale gas fracturing. Then we compared the simulation results with analytical solution, published papers and on-site microseismic monitoring data to verify our model. Finally, the influence of geological condition and engineering parameters on SRV was investigated by sensitivity analysis.
The reporting of potential resources is essential to assess the future development plan and profitability of a petroleum discovery, but if the project is under appraised and production data are absent, analysts often use analogs for preliminary estimates of technically recoverable volumes. To address this, a workflow is presented for selecting appropriate analogs for unconventional plays and using them to estimate the target play's potential. The proposed technique is demonstrated with a case study of the as-yet undeveloped Bowland Shale, which is the most prominent of the shale plays in the United Kingdom (UK) and is at the early stage of its assessment. The paper describes the current shale gas activity in the UK, highlighting the enviromental constraints placed on would-be Bowland Shale developers, which impact on drilling and production operations and stem from the geographic proximity of urban developments, infrastructure and nature, which limit the size of well pad footprint in the UK where land use is high. Studies have estimated the play's in-place resources for possible future development, but there are few estimates of its corresponding recoverable volumes due to lack of production history. At the outset, a database is created with published minimum-average-maximum ranges of key parameters such as total organic carbon, maturity level, gas filled porosity, permeability, etc. that play a major role in resources estimation and recovery potential for all unconventional plays. A comparison of triangular distributions, key parameter by key parameter, between the target shale play and the analog database, is then carried out using novel graphical and statistical methods to establish a "confidence factor" relating to the analog's viability. The most appropriate analog for the Bowland Shale is chosen from an exhaustive list of North American shale gas plays. Analytical approaches are then used to transform a model of the published type well performance of the selected analog by exchanging key model parameters with those of the target shale play. The paper shows how UK operational constraints can be statistically incorporated into the workflow and have a marked effect on the estimated recovery from the Bowland Shale.
Berawala, Dhruvit Satishchandra (Department of Energy and Petroleum Technology, University of Stavanger, Norway and The National IOR Centre of Norway) | Østebø Andersen, Pål (Department of Energy Resources, University of Stavanger, Norway and The National IOR Centre of Norway)
Only 3-10 % of gas from tight shale is recovered economically through natural depletion, demonstrating a significant potential for enhanced shale gas recovery (ESGR). Experimental studies have demonstrated that shale kerogen/organic matter has higher affinity for CO2 than methane, CH4, which opens possibilities for carbon storage and new production strategies.
This paper presents a new multicomponent adsorption isotherm which is coupled with a flow model for evaluation of injection-production scenarios. The isotherm is based on the assumption that different gas species compete for adsorbing on a limited specific surface area. Rather than assuming a capacity of a fixed number of sites or moles this finite surface area is filled with species taking different amount of space per mole. The final form is a generalized multicomponent Langmuir isotherm. Experimental adsorption data for CO2 and CH4 on Marcellus shale are matched with the proposed isotherm using relevant fitting parameters. The isotherm is first applied in static examples to calculate gas in place reserves, recovery factors and enhanced gas recovery potential based on contributions from free gas and adsorbed gas components. The isotherm is further coupled with a dynamic flow model with application to CO2-CH4 substitution for CO2-ESGR. We study the feasibility and effectiveness of CO2 injection in tight shale formations in an injection-production setting representative of lab and field implementation and compare with regular pressure depletion.
The production scenario we consider is a 1D shale core or matrix system intitally saturated with free and adsorbed CH4 gas with only left side (well) boundary open. During primary depletion, gas is produced from the shale to the well by advection and desorption. This process tends to give low recovery and is entirely dependent on the well pressure. Stopping production and then injecting CO2 into the shale leads to increase in pressure where CO2 gets preferentially adsorbed over CH4. The injected CO2 displaces, but also mixes with the in situ CH4. Restarting production from the well then allows CH4 gas to be produced in the gas mixture. Diffusion allows the CO2 to travel further into the matrix while keeping CH4 accessible to the well. Surface substitution further reduces the CO2 content and increases the CH4 content in the gas mixture that is produced to the well. A result of the isotherm and its application of Marcellus experimental data is that adsorption of CO2 with resulting desorption of CH4 will lead to a reduction in total pressure if the CO2 content in the gas composition is increased. That is in itself an important drive mechanism since the pressure gradient driving fluid flow is maintained (pressure buildup is avoided). This is a result of CO2 being found to take ~24 times less space per mol than CH4.
The objective of this study is to visualize the drained rock volume (DRV) and pressure depletion in hydraulically and naturally fractured reservoirs, using a high-resolution simulator to plot streamlines and time-of-flight contours that outline the DRV, based on computationally efficient complex potentials. A recently developed expression based on fast, grid-less Complex Analysis Methods (CAM) is applied to model the flow through discrete natural fractures with variable hydraulic conductivity. The impact of natural fractures on the local development of DRV contours and streamline patterns is analyzed. A sensitivity analysis of various permeability contrasts between natural fractures and the matrix is included. The results show that the DRV near hydraulic fractures is significantly affected by the presence of nearby natural fractures. The DRV location shifts according to the orientations, permeability and the density of the natural fractures. Reservoirs with numerous natural fractures result in highly distorted DRV shapes as compared to reservoirs without any discernable natural fractures. Additionally, the DRV shift due to natural fractures may contribute to enhanced well-interference by flow channeling via the natural fractures, as well as the creation of undrained rock volumes between the natural fractures. Complementary pressure depletion plots for each case show how the local pressure field changes, in a heterogeneous reservoir, due to the presence of natural fractures. The results from this study offer insights on how natural fractures affect the DRV and pressure contour plots. This study uses a fast grid-less and meshless high-resolution flow simulation tool based on CAM to simulate the flow in heterogeneous naturally fractured porous media. The CAM tool provides a practical/efficient simulation platform, complementary to grid-based reservoir simulators.
Data Analytics is progressively gaining traction as a viable resource to improve forecasts and reserve estimations in most prospective US shale plays. Part of those learnings has been tested for the reserves and resources estimation of the next worldwide top-class shale play, Vaca Muerta formation in Argentina. In this work, we rely on advanced artificial intelligence methods to automate workflows for production forecasting and reserve estimation in the Vaca Muerta formation. To achieve this goal, we develop a computational platform capable of integrating several sequential operations into a single automated workflow: (1) data gathering; (2) data preparation; (3) model fitting and forecasting and, (4) EUR estimation. As new data becomes available, each of these steps is performed automatically. The proposed platform also integrates with advanced business intelligence tools that aid at facilitating graphical interpretation and communication among specialists and decision makers. Hence, the suggested workflow can deliver production forecasts several magnitudes faster than traditional workflows while maintaining accurate and engineering sound results. Having fast and reliable forecast turnarounds allow for timely tracking key differences and commonalities among multiple shale plays to facilitate informed decision strategies in unconventional field evaluation and development.