Hydraulic fracturing framework

Abstract

A method can include receiving input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predicting production data for the group of wells using a machine learning model; and outputting the predicted production data.

Description

HYDRAULIC FRACTURING FRAMEWORKRELATED APPLICATION

[0001] This application claims priority to and the benefit of a US Provisional Application having Serial No. 63 / 605,541 , filed 3 December 2023, which is incorporated by reference herein in its entirety.BACKGROUND

[0002] Hydraulic fracturing may be employed to enhance fluid flow from a reservoir. In various instances, hydraulic fracturing may provide for unlocking fluid reserves from otherwise uneconomical or challenging geological formations. In deciding whether or not to employ hydraulic fracturing, various factors may be taken into consideration. For example, efficiency and efficacy of hydraulic fracturing operations can weigh in favor of employing hydraulic fracturing. In various instances, efficiency and / or efficacy may be constrained, for example, by geological complexities, operational constraints, and / or economic considerations. To assess such constraints with a view of optimizing hydraulic fracturing, one or more numerical or analytical techniques may be employed, which tend to be computationally intensive, timeconsuming, and often, unable to capture the realm of multi-dimensional complexities of real-world reservoir conditions. Moreover, such techniques might not effectively handle the vast amount of data generated during oil and gas operations; thereby, missing out on valuable insights that could be harnessed for improved decisionmaking and optimization, along with actual field implementation.

[0003] Various technologies, techniques, etc., described herein pertain to hydraulic fracturing operations, and optimization thereof, using data-driven modeling. Such an approach can improve hydraulic fracturing operations and, for example, production of fluid from a reservoir.SUMMARY

[0004] A method can include receiving input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predicting production data for the group of wells using a machine learning model; and outputting the predicted production data.

[0005] A system can include a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: receive input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predict production data for the group of wells using a machine learning model; and output the predicted production data.

[0006] One or more computer-readable storage media including processorexecutable instructions executable by a system to instruct the system to: receive input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predict production data for the group of wells using a machine learning model; and output the predicted production data.

[0007] Various other apparatuses, systems, methods, etc., are also disclosed. This summary is provided to introduce a selection of concepts that are further described below in the detailed description. This summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to be used as an aid in limiting the scope of the claimed subject matter.BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Features and advantages of the described implementations can be more readily understood by reference to the following description taken in conjunction with the accompanying drawings.

[0009] FIG. 1 illustrates an example system that includes various components for simulating a geological environment;

[0010] FIG. 2 illustrates an example of a system;

[0011] FIG. 3 illustrates examples of subsurface geologic regions and wells;

[0012] FIG. 4 illustrates an example of a hydraulically fractured subsurface volume and an example of a system for hydraulic fracturing;

[0013] FIG. 5A and FIG. 5B illustrate an example of a method;

[0014] FIG. 6 illustrates an example of a workflow;

[0015] FIG. 7 illustrates examples of graphics renderable via a graphical user interface (GUI);

[0016] FIG. 8 illustrates examples of graphics renderable via a graphical user interface (GUI);

[0017] FIG. 9A and FIG. 9B illustrate examples of graphics renderable via a GUI;

[0018] FIG. 10 illustrates examples of graphics renderable via a GUI;

[0019] FIG. 11 illustrates examples of graphics renderable via a GUI;

[0020] FIG. 12 illustrates examples of graphics renderable via a GUI;

[0021] FIG. 13 illustrates examples of graphics renderable via a GUI;

[0022] FIG. 14 illustrates examples of graphics renderable via a GUI;

[0023] FIG. 15 illustrates examples of graphics renderable via a GUI;

[0024] FIG. 16 illustrates examples of graphics renderable via a GUI;

[0025] FIG. 17 illustrates examples of graphics renderable via a GUI;

[0026] FIG. 18 illustrates examples of graphics renderable via a GUI;

[0027] FIG. 19 illustrates examples of graphics renderable via a GUI;

[0028] FIG. 20 illustrates examples of graphics renderable via a GUI;

[0029] FIG. 21 illustrates examples of graphics renderable via a GUI;

[0030] FIG. 22 illustrates examples of graphics renderable via a GUI;

[0031] FIG. 23 illustrates examples of graphics renderable via a GUI;

[0032] FIG. 24 illustrates examples of graphics renderable via a GUI;

[0033] FIG. 25 illustrates examples of graphics renderable via a GUI;

[0034] FIG. 26 illustrates examples of graphics renderable via a GUI;

[0035] FIG. 27 illustrates examples of graphics renderable via a GUI;

[0036] FIG. 28 illustrates examples of graphics renderable via a GUI;

[0037] FIG. 29 illustrates examples of graphics renderable via a GUI;

[0038] FIG. 30 illustrates examples of graphics renderable via a GUI;

[0039] FIG. 31 illustrates examples of graphics renderable via a GUI;

[0040] FIG. 32 illustrates examples of graphics renderable via a GUI;

[0041] FIG. 33 illustrates examples of graphics renderable via a GUI;

[0042] FIG. 34 illustrates examples of graphics renderable via a GUI;

[0043] FIG. 35 illustrates an example of a method and an example of a system; and

[0044] FIG. 36 illustrates example components of a system and a networked system.DETAILED DESCRIPTION

[0045] The following description includes the best mode presently contemplated for practicing the described implementations. This description is not to be taken in a limiting sense, but rather is made merely for the purpose of describing the general principles of the implementations. The scope of the described implementations should be ascertained with reference to the issued claims.

[0046] Hydraulic fracturing operations to enhance fluid production from a reservoir can depend on various factors, which may include constraints associated with geological complexities, operations, resources expended, resources produced, etc. As an example, a framework may implement one or more ML techniques where training (e.g., learning) depends on data, which can include actual field data. In various instances, data may be historic data and / or real-time data.

[0047] As an example, in a design phase of hydraulic fracturing operations, one or more ML models may be implemented that are trained on historic data, which may be acquired from field operations at one or more wells offset to a target well. As an example, in a control phase of hydraulic fracturing operations, real-time data may be utilized for one or more purposes, which may include additional training, input, etc. As an example, one or more ML models may be executable to generate output in an amount of time that is substantially less than that of a numerical model-based simulation. As an example, a reduction in time to generate output may be leveraged for optimization, particularly for iterative optimization.

[0048] Below, various types of environments, frameworks, workflows, data acquisition techniques, field equipment, field operations, etc., are described, which may involve use of a framework or frameworks, optionally during one or more field operations (e.g., hydraulic fracturing, etc.).

[0049] FIG. 1 shows an example of a system 100 that includes a workspace framework 110 that can provide for instantiation of, rendering of, interactions with, etc., a graphical user interface (GUI) 120. In the example of FIG. 1 , the GU1 120 can include graphical controls for computational frameworks (e.g., applications) 121 , projects 122, visualization 123, one or more other features 124, data access 125, and data storage 126.

[0050] In the example of FIG. 1 , the workspace framework 110 may be tailored to a particular geologic environment such as an example geologic environment 150.For example, the geologic environment 150 may include layers (e.g., stratification) that include a reservoir 151 and that may be intersected by a fault 153. A geologic environment 150 may be outfitted with a variety of sensors, detectors, actuators, etc. In such an environment, various types of equipment such as, for example, equipment 152 may include communication circuitry to receive and to transmit information, optionally with respect to one or more networks 155. Such information may include information associated with downhole equipment 154, which may be equipment to acquire information, to assist with resource recovery, etc. Other equipment 156 may be located remote from a wellsite and include sensing, detecting, emitting, or other circuitry. Such equipment may include storage and communication circuitry to store and to communicate data, instructions, etc. One or more satellites may be provided for purposes of communications, data acquisition, etc. For example, FIG. 1 shows a satellite 170 in communication with the network 155 that may be configured for communications, noting that the satellite may additionally or alternatively include circuitry for imagery (e.g., spatial, spectral, temporal, radiometric, etc.).

[0051] FIG. 1 also shows the geologic environment 150 as optionally including equipment 157 and 158 associated with a well that includes a substantially horizontal portion that may intersect with one or more fractures 159. For example, consider a well in a formation that may include natural fractures, artificial fractures (e.g., hydraulic fractures) or a combination of natural and artificial fractures. As an example, a well may be drilled for a reservoir that is laterally extensive. In such an example, lateral variations in properties, stresses, etc., may exist where an assessment of such variations may assist with planning, operations, etc., to develop a laterally extensive reservoir (e.g., via fracturing, injecting, extracting, etc.). As an example, the equipment 157 and / or 158 may include components, a system, systems, etc. for fracturing, seismic sensing, analysis of seismic data, assessment of one or more fractures, etc.

[0052] In the example of FIG. 1 , the GUI 120 shows some examples of computational frameworks, including the DRILLPLAN, PETREL, TECHLOG, PETROMOD, ECLIPSE, INTERSECT, KINETIXA / ISAGE, and PI PESIM frameworks (SLB, Houston, Texas). One or more types of frameworks may be implemented within or in a manner operatively coupled to the DELFI environment, which is a secure, cognitive, cloud-based collaborative environment that integrates data and workflows with digital technologies, such as artificial intelligence (Al) and machine learning (ML).Such an environment can provide for operations that involve one or more frameworks. The DELFI environment may be referred to as the DELFI framework, which may be a framework of frameworks. The DELFI environment can include various other frameworks, which may operate using one or more types of models (e.g., simulation models, etc.).

[0053] As an example, a framework, system, etc., may provide for utilization with one or more types of other frameworks, systems, environments, etc. For example, consider an agnostic approach and / or a tailored approach that may provide for operative coupling via one or more technologies, techniques, etc. For example, consider an application programming interface (API) implementation that provides for receipt of an API call and transmission of response to the API call. In such an example, an API may be agnostic and / or tailored. While the DELFI environment is mentioned, as explained, one or more other types of environments (e.g., platforms, etc.) may be utilized. In various instances, one or more American Petroleum Institute (API) conventions may be utilized, for example, as to characterizing equipment, wells, etc., where the American Petroleum Institute acronym API may be distinguished from application programming interface acronym API through attention to context. Further, while one or more American Petroleum Institute (API) conventions may be utilized to specify, characterize, etc., one or more wells, wells may be characterized using one or more other conventions, additionally or alternatively. In various figures, fields, wells, etc., are provided as examples, noting that methods, frameworks, systems, etc., described herein may be applied to such fields, wells, etc., and / or one or more other fields, wells, etc.

[0054] The DRILLPLAN framework provides for digital well construction planning and includes features for automation of repetitive tasks and validation workflows, enabling improved quality drilling programs (e.g., digital drilling plans, etc.) to be produced quickly with assured coherency.

[0055] The DRILLOPS framework, which may be included in the system 100 of FIG. 1 , may execute a digital drilling plan and ensures plan adherence, while delivering goal-based automation. The DRILLOPS framework may generate activity plans automatically individual operations, whether they are monitored and / or controlled on the rig or in town. Automation may utilize data analysis and learning systems to assist and optimize tasks, such as, for example, setting ROP to drilling a stand. A presetmenu of automatable drilling tasks may be rendered, and, using data analysis and models, a plan may be executed in a manner to achieve a specified goal, where, for example, measurements may be utilized for calibration. The DRILLOPS framework provides flexibility to modify and replan activities dynamically, for example, based on a live appraisal of various factors (e.g., equipment, personnel, and supplies). Well construction activities (e.g., tripping, drilling, cementing, etc.) may be continually monitored and dynamically updated using feedback from operational activities. The DRILLOPS framework may provide for various levels of automation based on planning and / or re-planning (e.g., via the DRILLPLAN framework), feedback, etc.

[0056] The PETREL framework can be part of the DELFI cognitive exploration and production (E&P) environment (SLB, Houston, Texas, referred to as the DELFI environment) for utilization in geosciences and geoengineering, for example, to analyze subsurface data from exploration to production of fluid from a reservoir.

[0057] The TECHLOG framework can handle and process field and laboratory data for a variety of geologic environments (e.g., deepwater exploration, shale, etc.). The TECHLOG framework can structure wellbore data for analyses, planning, etc. As an example, the TECHLOG framework may be coupled to one or more ML models for purposes of generation of output, training, etc.

[0058] The PETROMOD framework provides petroleum systems modeling capabilities that can combine one or more of seismic, well, and geological information to model the evolution of a sedimentary basin. The PETROMOD framework can predict if, and how, a reservoir has been charged with hydrocarbons, including the source and timing of hydrocarbon generation, migration routes, quantities, and hydrocarbon type in the subsurface or at surface conditions.

[0059] The ECLIPSE framework provides a reservoir simulator with numerical solvers for prediction of dynamic behavior for various types of reservoirs and development schemes.

[0060] The INTERSECT framework provides a high-resolution reservoir simulator for simulation of geological features and quantification of uncertainties, for example, by creating production scenarios and, with the integration of precise models of the surface facilities and field operations, the INTERSECT framework can produce results, which may be continuously updated by real-time data exchanges (e.g., from one or more types of data acquisition equipment in the field that can acquire dataduring one or more types of field operations, etc.). The INTERSECT framework can provide completion configurations for complex wells where such configurations can be built in the field, can provide detailed chemical-enhanced-oil-recovery (EOR) formulations where such formulations can be implemented in the field, can analyze application of steam injection and other thermal EOR techniques for implementation in the field, advanced production controls in terms of reservoir coupling and flexible field management, and flexibility to script customized solutions for improved modeling and field management control. The INTERSECT framework, as with the other example frameworks, may be utilized as part of the DELFI environment, for example, for rapid simulation of multiple concurrent cases.

[0061] The KINETIX framework provides for reservoir-centric stimulation-to- production analyses that can integrate geology, petrophysics, completion engineering, reservoir engineering, and geomechanics, for example, to provide for optimized completion and fracturing designs for a well, a pad, or a field. The KINETIX framework can be operatively coupled to and / or integrated with features of the PETREL framework (e.g., within the DELFI environment). As to the VISAGE framework it can be part of or otherwise operatively coupled to the KINETIX framework.

[0062] The VISAGE framework includes finite element numerical solvers that may provide simulation results such as, for example, results as to compaction and subsidence of a geologic environment, well and completion integrity in a geologic environment, cap-rock and fault-seal integrity in a geologic environment, fracture behavior in a geologic environment, thermal recovery in a geologic environment, CO2 disposal, etc.

[0063] As an example, the KINETIX framework can provide for analyses from 1 D logs and simple geometric completions to 3D mechanical and petrophysical models coupled with the INTERSECT framework high-resolution reservoir simulator and VISAGE framework finite-element geomechanics simulator. The KINETIX framework can provide automated parallel processing using cloud platform resources and can provide for rapid assessment of well spacing, completion, and treatment design choices, enabling exploration of many scenarios in a relatively rapid manner (e.g., via provisioning of cloud platform resources). The KINETIX framework may be operatively coupled to the MANGROVE simulator (SLB, Houston, Texas), which canprovide for optimization of stimulation design (e.g., stimulation treatment operations such as hydraulic fracturing) in a reservoir-centric environment.

[0064] The MANGROVE framework can combine scientific and experimental work to predict geomechanical propagation of hydraulic fractures, reactivation of natural fractures, etc., along with production forecasts within 3D reservoir models (e.g., production from a drainage area of a reservoir where fluid moves via one or more types of fractures to a well and / or from a well). The MANGROVE framework can provide results pertaining to heterogeneous interactions between hydraulic and natural fracture networks, which may assist with optimization of the number and location of fracture treatment stages (e.g., stimulation treatment(s)), for example, to increased perforation efficiency and recovery.

[0065] The PIPESIM simulator includes solvers that may provide simulation results such as, for example, multiphase flow results (e.g., from a reservoir to a wellhead and beyond, etc.), flowline and surface facility performance, etc. The PIPESIM simulator may be integrated, for example, with the AVOCET production operations framework (SLB, Houston Texas). The PIPESIM simulator may be an optimizer that can optimize one or more operational scenarios at least in part via simulation of physical phenomena.

[0066] The aforementioned DELFI environment provides various features for workflows as to subsurface analysis, planning, construction and production, for example, as illustrated in the workspace framework 110. As shown in FIG. 1 , outputs from the workspace framework 110 can be utilized for directing, controlling, etc., one or more processes in the geologic environment 150, and feedback 160 can be received via one or more interfaces in one or more forms (e.g., acquired data as to operational conditions, equipment conditions, environment conditions, etc.).

[0067] In the example of FIG. 1 , the visualization features 123 may be implemented via the workspace framework 110, for example, to perform tasks as associated with one or more of subsurface regions, planning operations, constructing wells and / or surface fluid networks, and producing from a reservoir.

[0068] Visualization features may provide for visualization of various earth models, properties, etc., in one or more dimensions. As an example, visualization features may include one or more control features for control of equipment, which can include, for example, field equipment that can perform one or more field operations.A workflow may utilize one or more frameworks to generate information that can be utilized to control one or more types of field equipment (e.g., drilling equipment, wireline equipment, fracturing equipment, etc.).

[0069] FIG. 2 shows an example of a system 200 that can be operatively coupled to one or more databases, data streams, etc. For example, one or more pieces of field equipment, laboratory equipment, computing equipment (e.g., local and / or remote), etc., can provide and / or generate data that may be utilized in the system 200.

[0070] As shown, the system 200 can include a geological / geophysical data block 210, a surface models block 220 (e.g., for one or more structural models), a volume modules block 230, an applications block 240, a numerical processing block 250 and an operational decision block 260. As shown in the example of FIG. 2, the geological / geophysical data block 210 can include data from well tops or drill holes 212, data from seismic interpretation 214, data from outcrop interpretation and optionally data from geological knowledge. As an example, the geological / geophysical data block 210 can include data from digital images, which can include digital images of cores, cuttings, cavings, outcrops, etc. As to the surface models block 220, it may provide for creation, editing, etc. of one or more surface models based on, for example, one or more of fault surfaces 222, horizon surfaces 224 and optionally topological relationships 226. As to the volume models block 230, it may provide for creation, editing, etc. of one or more volume models based on, for example, one or more of boundary representations 232 (e.g., to form a watertight model), structured grids 234 and unstructured meshes 236.

[0071] As shown in the example of FIG. 2, the system 200 may allow for implementing one or more workflows, for example, where data of the data block 210 are used to create, edit, etc. one or more surface models of the surface models block 220, which may be used to create, edit, etc. one or more volume models of the volume models block 230. As indicated in the example of FIG. 2, the surface models block 220 may provide one or more structural models, which may be input to the applications block 240. For example, such a structural model may be provided to one or more applications, optionally without performing one or more processes of the volume models block 230 (e.g., for purposes of numerical processing by the numerical processing block 250). Accordingly, the system 200 may be suitable for one or moreworkflows for structural modeling (e.g., optionally without performing numerical processing per the numerical processing block 250).

[0072] As to the applications block 240, it may include applications such as a well prognosis application 242, a reserve calculation application 244 and a well stability assessment application 246. As to the numerical processing block 250, it may include a process for seismic velocity modeling 251 followed by seismic processing 252, a process for facies and petrophysical property interpolation 253 followed by flow simulation 254, and a process for geomechanical simulation 255 followed by geochemical simulation 256. As indicated, as an example, a workflow may proceed from the volume models block 230 to the numerical processing block 250 and then to the applications block 240 and / or to the operational decision block 260. As another example, a workflow may proceed from the surface models block 220 to the applications block 240 and then to the operational decisions block 260 (e.g., consider an application that operates using a structural model).

[0073] In the example of FIG. 2, the operational decisions block 260 may include a seismic survey design process 261 , a well rate adjustment process 252, a well trajectory planning process 263, a well completion planning process 264 and a process for one or more prospects, for example, to decide whether to explore, develop, abandon, etc. a prospect.

[0074] Referring again to the data block 210, the well tops or drill hole data 212 may include spatial localization, and optionally surface dip, of an interface between two geological formations or of a subsurface discontinuity such as a geological fault; the seismic interpretation data 214 may include a set of points, lines or surface patches interpreted from seismic reflection data, and representing interfaces between media (e.g., geological formations in which seismic wave velocity differs) or subsurface discontinuities; the outcrop interpretation data 216 may include a set of lines or points, optionally associated with measured dip, representing boundaries between geological formations or geological faults, as interpreted on the earth surface; and the geological knowledge data 218 may include, for example knowledge of the paleo-tectonic and sedimentary evolution of a region.

[0075] As to a structural model, it may be, for example, a set of gridded or meshed surfaces representing one or more interfaces between geological formations (e.g., horizon surfaces) or mechanical discontinuities (fault surfaces) in thesubsurface. As an example, a structural model may include some information about one or more topological relationships between surfaces (e.g., fault A truncates fault B, fault B intersects fault C, etc.).

[0076] As to the one or more boundary representations 232, they may include a numerical representation in which a subsurface model is partitioned into various closed units representing geological layers and fault blocks where an individual unit may be defined by its boundary and, optionally, by a set of internal boundaries such as fault surfaces.

[0077] As to the one or more structured grids 234, it may include a grid that partitions a volume of interest into different elementary volumes (cells), for example, that may be indexed according to a pre-defined, repeating pattern. As to the one or more unstructured meshes 236, it may include a mesh that partitions a volume of interest into different elementary volumes, for example, that may not be readily indexed following a pre-defined, repeating pattern (e.g., consider a Cartesian cube with indexes I, J, and K, along x, y, and z axes).

[0078] As to the seismic velocity modeling 251 , it may include calculation of velocity of propagation of seismic waves (e.g., where seismic velocity depends on type of seismic wave and on direction of propagation of the wave). As to the seismic processing 252, it may include a set of processes allowing identification of localization of seismic reflectors in space, physical characteristics of the rocks in between these reflectors, etc.

[0079] As to the facies and petrophysical property interpolation 253, it may include an assessment of type of rocks and of their petrophysical properties (e.g., porosity, permeability), for example, optionally in areas not sampled by well logs or coring. As an example, such an interpolation may be constrained by interpretations from log and core data, and by prior geological knowledge.

[0080] As to the flow simulation 254, as an example, it may include simulation of flow of hydro-carbons in the subsurface, for example, through geological times (e.g., in the context of petroleum systems modeling, when trying to predict the presence and quality of oil in an un-drilled formation) or during the exploitation of a hydrocarbon reservoir (e.g., when some fluids are pumped from or into the reservoir).

[0081] As to geomechanical simulation 255, it may include simulation of the deformation of rocks under boundary conditions. Such a simulation may be used, forexample, to assess compaction of a reservoir (e.g., associated with its depletion, when hydrocarbons are pumped from the porous and deformable rock that composes the reservoir). As an example, a geomechanical simulation may be used for a variety of purposes such as, for example, prediction of fracturing, reconstruction of the paleogeometries of the reservoir as they were prior to tectonic deformations, etc.

[0082] As to geochemical simulation 256, such a simulation may simulate evolution of hydrocarbon formation and composition through geological history (e.g., to assess the likelihood of oil accumulation in a particular subterranean formation while exploring new prospects).

[0083] As to the various applications of the applications block 240, the well prognosis application 242 may include predicting type and characteristics of geological formations that may be encountered by a drill bit, and location where such rocks may be encountered (e.g., before a well is drilled); the reserve calculations application 244 may include assessing total amount of hydrocarbons or ore material present in a subsurface environment (e.g., and estimates of which proportion can be recovered, given a set of economic and technical constraints); and the well stability assessment application 246 may include estimating risk that a well, already drilled or to-be-drilled, will collapse or be damaged due underground stress.

[0084] As to the operational decision block 260, the seismic survey design process 261 may include deciding where to place seismic sources and receivers to optimize the coverage and quality of the collected seismic information while minimizing cost of acquisition; the well rate adjustment process 262 may include controlling injection and production well schedules and rates (e.g., to maximize recovery and production); the well trajectory planning process 263 may include designing a well trajectory to maximize potential recovery and production while minimizing drilling risks and costs; the well trajectory planning process 264 may include selecting proper well tubing, casing and completion (e.g., to meet expected production or injection targets in specified reservoir formations); and the prospect process 265 may include decision making, in an exploration context, to continue exploring, start producing or abandon prospects (e.g., based on an integrated assessment of technical and financial risks against expected benefits).

[0085] The system 200 can include and / or can be operatively coupled to a system such as the system 100 of FIG. 1. For example, the workspace framework110 may provide for instantiation of, rendering of, interactions with, etc., the graphical user interface (GUI) 120 to perform one or more actions as to the system 200. In such an example, access may be provided to one or more frameworks (e.g., DRILLPLAN, DRILLOPS, PETREL, TECHLOG, PETROMOD, ECLIPSE, INTERSECT, KINETIX / VISAGE, PIPESIM, etc.). One or more frameworks may provide for geo data acquisition as in block 210, for structural modeling as in block 220, for volume modeling as in block 230, for running an application as in block 240, for numerical processing as in block 250, for operational decision making as in block 260, etc.

[0086] As an example, the system 200 may provide for monitoring data, which can include geo data per the geo data block 210. In various examples, geo data may be acquired during one or more operations. For example, consider acquiring geo data during drilling operations via downhole equipment and / or surface equipment. As an example, the operational decision block 260 can include capabilities for monitoring, analyzing, etc., such data for purposes of making one or more operational decisions, which may include controlling equipment, revising operations, revising a plan, etc. In such an example, data may be fed into the system 200 at one or more points where the quality of the data may be of particular interest. For example, data quality may be characterized by one or more metrics where data quality may provide indications as to trust, probabilities, etc., which may be germane to operational decision making and / or other decision making. As an example, the system 200 of FIG. 2 may include one or more ML models, for example, consider one or more ML models for use in one or more types of hydraulic fracturing workflows. As an example, the system 200 may be operatively coupled to one or more frameworks that may implement one or more ML models (e.g., for production prediction, etc.). As an example, the system 200 may provide for generation of data that may be utilized to train one or more ML models, which, in turn, may be utilized for one or more purposes (e.g., predicting production, optimization, geoscreening, etc.).

[0087] FIG. 3 shows an example of a portion of a field 300 that may include one or more boundaries 302 where, within the one or more boundaries 302, a number of wells may be drilled from a surface pad 304. For example, three or more wells may be drilled from the surface pad 304 where the wells may extend in one or more directions. As an example, wells and planned wells may be within one or more areas defined with respect to one or more constraints, which may be surface constraints,subsurface constraints, etc. For example, roads, land rights, etc. may be a few examples of some types of constraints. As an example, wells, whether planned or drilled, can be part of a field development plan. As an example, where one or more wells are to be drilled from a pad, a pad placement framework may be utilized to position pads in a field (e.g., surface of a geologic environment). As an example, such a framework may aim to place pads and / or wells in a field development plan to efficiently extract resources from the geologic environment (e.g., oil, gas, etc.).

[0088] FIG. 3 also includes a three-dimensional rendering of a portion of a geologic environment 350 that includes a surface level 351 , a rig 352 at a pad 354 and a reservoir level 353 where a plurality of lateral portions of wells 355 are to be drilled and developed to extract one or more resources from a reservoir 356. In such an example, a framework such as, for example, a pad placement framework, may generate a plan that includes pad positions and approximate positions of wells that can be developed and utilized to extract one or more resources from a geologic environment.

[0089] FIG. 3 further shows an example of a two-dimensional rendering of a portion of a geologic environment 370 that illustrates some parameters that may be associated with a well that may include a lateral portion along with some examples of layers 375 within the geologic environment 370. For example, consider the Permian basin, which can include layers such as Leonard Shale (LS), Upper Spraberry Shale (USS), Middle Spraberry Shale (MSS), Lower Spraberry Shale (LSS), Jo Mill (JM), Spraberry Shale (SS), Dean Shale, Wolfcamp A Shale, Wolfcamp B Shale, Wolfcamp C Shale, and Wolfcamp D Shale. As an example, various layers in a geologic environment may be amenable to production of hydrocarbons via drilling of wells where production may be enhanced via hydraulic fracturing.

[0090] As an example, a geologic environment may include multiple distinct flow units, which may be referred to as benches. As an example, consider an interval that includes separations formed by materials such as hard limestone baffles, or barriers, that may limit vertical growth of fractures during stimulation. In such an example, wells that have been drilled and completed may be draining only a portion of the interval if they are within a distinct flow unit (e.g., a distinct bench). To enhance production, field operations may involve drilling wells into one or more other benches and performing hydraulic fracturing where such new wells and / or old wells may define a spacinggeometry such as, for example, a wine rack geometry. For example, in FIG. 3, a series of cross-sectional views 377 of the layers 375 shows geometries of wells with respect to the layers 375 where one or more of the geometries may be considered a wine rack or wine rack geometries; noting that while a wine rack may be a structure for holding a number of bottles of wine in a horizontal position, herein, a wine rack geometry or geometric representation refers to wells as may be positioned in a subsurface region where well positions may be viewed in cross-section to provide a view akin to bottles of wine horizontally positioned and viewed on end.

[0091] As an example, a region may be undeveloped or developed where a number of new wells may be planned, drilled, hydraulically fractured and produced and / or where a number of old wells may be subjected to hydraulic fracturing, which may be or include re-fracturing. As an example, a wine rack geometry may be generated as part of a plan where, for example, the wine rack geometry includes at least three wells, which may be old wells, new wells, or a combination of old and new wells.

[0092] In the example of FIG. 3, consider a scenario where the Lower Spraberry Shale layer includes three benches (e.g., lower, middle and upper). In such an example, a wine rack geometry may include wells in one or more of these three benches.

[0093] As an example, a pad placement workflow may include planning for wells in an environment (e.g., for shale gas producers, oil sand producers, etc.). Such a workflow may be applied to one or more environments of interest. As an example, a pad placement workflow may include well placement for wells that extend from a pad. For example, consider the pad 354 and the wells 355 in the example of FIG. 3, which may be specified according to a plan. As an example, a plan may specify one or more geometries for wells within a subsurface environment.

[0094] As explained, when developing a regional field of reservoirs, operators may consider drilling multiple wells from a well pad location in an effort to maximize a return on investment. As an example, wells drilled at a pad may follow one of a plurality of configurations. For example, a well head configuration can include a row of three or more producer wells where such a row may be at a common depth and / or at a varying depth (e.g., consider dipping, in one or more benches, etc.). As explained, multiple rows may be specified. As an example, an operator may choose well padlocations based on a combination of constraints at a ground level, such as roads, rivers, buildings, etc., and constraints at a reservoir level, such as lease boundary. A concern of the operators can be selection of pad locations and configurations to achieve more reservoir coverage, which may be characterized based at least in part on drainage area of pads and associated wells. As an example, a stimulated reservoir volume (SRV) may be a metric that is related to drainage. As an example, an SRV may be defined within a volume such as a six-sided volume; noting that one or more curved shapes may be utilized to define a volume (e.g., an ellipsoid, a sphere, etc.). As an example, a parallelepiped, which may be defined as a three-dimensional figure formed by six parallelograms, may be utilized to define a volume. In such an example, angles may be 90 degree and / or other angles.

[0095] As an example, a pad placement process may operate in conjunction with a pad well design process, which may be a plug-in for creation of proposed wells on regular configurations (e.g., to be repeated at each pad location), to produce well designs. Applications for such a process can include reservoirs with high well density, such as, for example, shale gas, heavy oil, etc. Such a process may seek to control or define well length, vertical and horizontal spacing, orientation, etc. (see, e.g., the geologic environment 370 of FIG. 3).

[0096] As an example, a plan may be implemented in the field using various field operations, which may involve operating controllable equipment. For example, a controller can be operatively coupled to one or more pieces of equipment to control one or more actions thereof. As an example, a controller can provide for control of pumping equipment and, for example, measurement equipment, which can include one or more sensors. As an example, field operations may include one or more of fracturing operations, wireline operations, maintenance operations, monitoring operations, etc. As an example, re-planning may occur during one or more field operations and / or between one or more field operations.

[0097] As to pumping fluid, consider, as an example, hydraulic fracturing operations that can include pumping fluid into a borehole in a formation to generate fractures in the formation. Such pumping can utilize a pump driven by an internal combustion engine where a drive shaft of the internal combustion engine can be operatively coupled to a transmission, which can include various gears that can gear- up or gear-down rotational speed of the drive shaft of the internal combustion enginein a manner that aims to effectively control a pump shaft to achieve one or more desirable pumping parameters (e.g., pump pressure, pump flow rate, etc.). While a single pump is mentioned, a field operation can involve a fleet of pumps where each pump may be mounted on a trailer along with an internal combustion engine and a transmission. A fleet operation can pump fluid to a manifold or manifolds, mixing equipment, etc.

[0098] FIG. 4 shows an example of a system 410 that may be utilized to perform field operations to generate an SRV for a group of wells 404 that may include wells arranged according to a plan that includes one or more geometries 408. As shown, the geometries 408, in a cross-sectional view, may be defined by spacing parameters such as h as a height between wells and x as a width between wells. As an example, a plan may call for hydraulic fracturing of a number of wells prior to producing fluid from the wells. In such an example, the hydraulic fracturing may be referred to as simultaneous. In various instances, one set of equipment may be utilized to hydraulically fracture each well where, for example, fracturing fluid is directed to each well individually, for example, according to a sequence. As to each well, hydraulic fracturing may occur in stages. For example, a well may be hydraulically fractured to create a number of spaced fractures, which may be spaced and created in stages.

[0099] In the example of FIG. 4, as explained, the system 410 may be utilized to generate the SRV as shown in the environment for the group of wells 404, which may be referred to as an effective SRV. As shown, the system 410 can include water tankers 412, a precision continuous mixer (PCM) 420, one or more sand chiefs 430, an optional acid and / or other chemical supply 440, a blender 450, a missile manifold 460, and a fleet of pump systems 470. As shown, the pump systems 470 are operatively coupled to the missile manifold 460, which is supplied with fluid via at least the PCM 420 and the blender 450, which may receive fluid from one or more of the water tankers 412, which can include conduits operatively coupled via a manifold or manifolds. As shown, the system 410 can provide for output of blended fluid, optionally with solids (e.g., sand as proppant, etc.) and optionally with chemicals (e.g., surfactant, acid, etc.), to a wellhead, which is a wellhead 480 to at least a partially completed well (e.g., with one or more completion components). As an example, hydraulic fracturing can be performed using the system 410. At the wellhead 480, various types of equipment may be present such as a wireline truck 492, a crane truck 494 andmonitoring and / or control (M&C) equipment 496. As an example, the system 410 may be operated at least in part according to a plan, which may be or may include a digital plan that includes digital instructions that may be executable by one or more controllers (e.g., consider the M&C equipment 496, etc.).

[0100] FIG. 5A and FIG. 5B show an example of a method 500 that includes generating fractures (e.g., hydraulic or artificial fractures). As shown, the method 500 can include various operational blocks such as one or more of the blocks 501 , 502, 503, 504, 505, and 506. The block 501 may be a drilling block that includes drilling into a formation 510 that includes layers 512, 514, and 516 to form a bore 530 with a kickoff 532 to a portion defined by a heel 534 and a toe 536, for example, within the layer 514.

[0101] As illustrated with respect to the block 502, the bore 530 may be at least partially cased with casing 540 into which a string or line 550 may be introduced that carries a perforator 560. As shown, the perforator 560 can include a distal end 562 and charge positions 565 associated with activatable charges that can perforate the casing 540 and form channels 515-1 in the layer 514. Next, per the block 503, fluid may be introduced into the bore 530 between the heel 534 and the toe 536 where the fluid passes through the perforations in the casing 540 and into the channels 515-1. Where such fluid is under pressure, the pressure may be sufficient to fracture the layer 514, for example, to form fractures 517-1. In the block 503, the fractures 517-1 may be first stage fractures, for example, of a multistage fracturing operation.

[0102] Per the block 504, additional operations are performed for further fracturing of the layer 514. For example, a plug 570 may be introduced into the bore 530 between the heel 534 and the toe 536 and positioned, for example, in a region between first stage perforations of the casing 540 and the heel 534. Per the block 505, the perforator 560 may be activated to form additional perforations in the casing 540 (e.g., second stage perforations) as well as channels 515-2 in the layer 514 (e.g., second stage channels). Per the block 506, fluid may be introduced while the plug 570 is disposed in the bore 530, for example, to isolate a portion of the bore 530 such that fluid pressure may build to a level sufficient to form fractures 517-2 in the layer 514 (e.g., second stage fractures).

[0103] In a method such as the method 500 of FIG. 5A and FIG. 5B, it may be desirable that a plug (e.g., the plug 570) includes properties suited to one or moreoperations. Properties of a plug may include mechanical properties (e.g., sufficient strength to withstand pressure associated with fracture generation, etc.) and may include one or more other types of properties (e.g., chemical, electrical, etc.). As an example, it may be desirable that a plug degrades, that a plug seat degrades, that at least a portion of a borehole tool degrades, etc. For example, a plug may be manufactured with properties such that the plug withstands, for a period of time, conditions associated with an operation and then degrades (e.g., when exposed to one or more conditions). In such an example, where the plug acts to block a passage for an operation, upon degradation, the passage may become unblocked, which may allow for one or more subsequent operations.

[0104] As shown in the example of FIG. 5A and FIG. 5B, a fracture may extend to a boundary or boundaries of a layer such as the boundary between the layers 512 and 514 and the boundary between the layers 514 and 516. As explained, a particular layer may include benches as producible units that may be bound such that a fracture within one bench does not extend to another bench. While the example of FIG. 5A and FIG. 5B shows various layers as being substantially parallel, layers may dip. Dip may be defined as the angle between a planar feature, such as a sedimentary bed or a fault, and a horizontal plane. True dip may be defined as the angle a plane makes with a horizontal plane, the angle being measured in a direction perpendicular to the strike of the plane. Apparent dip may be defined as the angle measured in any direction other than perpendicular to the strike of the plane. Given the apparent dip and the strike, or two apparent dips, the true dip may be computed.

[0105] As explained, a volume may be defined for a number of wells where the volume may be an SRV. As an example, such a volume may be referred to as a tank or a rock cube. As to the term cube, in this context, it is not restricted to a geometric cube; rather, it may be a volume defined by one or more surfaces, sides, etc.

[0106] As an example, a rock cube may be a group of horizontal wells which are spatially and / or temporally close to each other and can be considered be as a drilling unit. As an example, they may be within the same formation and similar in horizontal drilling length. As explained, a formation may include distinct producible units, which may be referred to as benches. As an example, a rock cube may include wells that are within one or more benches.Rock cube development (RCD) is a technique that involves drilling multiple horizontal wells in different layers of an unconventional reservoir, such as shale or tight oil, and then fracturing them simultaneously to create a large stimulated rock volume (SRV) (see, e.g., Jacobs, T. (2019), “Dominator project raises key questions about future of cube drilling”, Journal of Petroleum Technology, 71 (10), 40-42 (doi: 10.2118 / 1019- 0040-JPT), which is incorporated by reference herein in its entirety). This technique aims to maximize the recovery of hydrocarbons from low-permeability formations by increasing the contact area between the wellbore and the reservoir (see, e.g., Salama, A. &.-A. (2017), “Flow and Transport in Tight and Shale Formations: A Review”, Geofluids, 10, 1-21 (doi: 10.1155 / 2017 / 4251209), which is incorporated by reference herein in its entirety). Optimal well spacing and fracture design may present challenging issues that depend on the reservoir properties, fluid characteristics, and operational parameters (see, e.g., Zhi-dong Yang, Y. W.-y.-w.-h. (2020), “Numerical Simulation of a Horizontal Well With Multi-Stage Oval Hydraulic Fractures in Tight Oil Reservoir Based on an Embedded Discrete Fracture Model”, Frontiers in Energy Research, 8 (doi: 10.3389 / fenrg.2020.601 107), which is incorporated by reference herein in its entirety).

[0107] As an example, a computational framework may provide for rock cube development, which may include planning and / or execution (e.g., control) of one or more field operations to construct one or more producing wells within a subsurface environment. As an example, such a framework can implement one or more data- driven techniques, which may include machine learning (ML). Such an approach may accelerate development planning by optimizing cube wells to achieve the best technical recoverable resource (TRR) per drill spacing unit (DSU) while maximizing the return of investment (ROI). As an example, a framework may be implemented to optimize efficiency using one or more artificial intelligence technologies that can be applied to rock cubes (e.g., rock cube data, etc.).

[0108] As an example, a framework may be implemented for one or more basins. For example, consider a framework that may be implemented for a drillingunit level of analysis and forecasting for the Permian basin where such a framework may expeditiously evaluate potential areas for development.

[0109] As an example, a framework may provide for implementation of one or more workflows for forecasting unconventional reservoirs. Such a framework may aimto maximize data utilization to provide comprehensive unconventional reservoir analyses.

[0110] FIG. 6 shows an example of a workflow 600 that may be implemented by a framework. As shown, a scenario block 610 may be presented for rock cube design and production forecasting to enhance production. As shown, a data block 620 may be utilized to access data germane to the scenario of the scenario block 610. For example, consider utilizing one or more data science framework (e.g., DATAIKU, SNOWFLAKE, etc.) to analyze data for wells in a basin where such data may include well location and position data, well production data, formation data, etc. As shown, pre-processing, clustering, and analysis block 630 may be utilized. For example, consider data formatting for implementation of one or more techniques such as, for example, spatial-temporal (ST) techniques and / or object detection techniques (e.g., edge detection, etc.). In such an example, consider implementation of ST-DBSCAN to identify wells in a spatial and temporal manner as belonging to a rock cube where a plurality of rock cubes may be identified. ST-DBSCAN may be accessed via one or more libraries as a spatial-temporal clustering technique (e.g., consider a PYTHON library, etc.). ST-DBSCAN is described in an article by Derya Birant, A.K. (2007), “ST- DBSCAN: An algorithm for clustering spatial-temporal data”, Data & Knowledge Engineering, Volume 60, Issue 1 , pp. 208-221 , ISSN 0169-023X (https: / / doi.Org / 10.1016 / j.datak.2006.01.013), which is incorporated by reference herein in its entirety.

[0111] As to object detection, consider, for example, implementation of one or more edge detection techniques for identification of one or more layers (e.g., layer boundary or layer boundaries). As an example, a Hough transform approach may be utilized, for example, to identify layers in one or more cross-sections of an identified rock cube. In such an example, the rock cube may be further defined with respect to one or more layers (e.g., one or more layer boundaries). In such an example, a layer or layer boundary may be identified as being substantially flat (e.g., horizontal) and / or dipping (e.g., according to a dip angle). As an example, clustering may be utilized to generate various clusters based on one or more parameters to provide for rock cube identification and characterization, which, as explained, may include one or more production characteristics. As shown, a model building and training and evaluation block 640 may be utilized using output of the block 630. For example, the block 640may aim to train one or more ML models using output of the block 630 to generate production predictions for input parameters associated with one or more rock cubes. For example, consider a decision tree model that may be a boosted model that can receive input for a rock cube and prediction production for the rock cube. As another example, consider a neural network approach where, for example, a convolution neural network (CNN) may be implemented to receive input for a rock cube and predict production of the rock cube. In the example of FIG. 6, the workflow 600 may utilize output from the block 630 as training data to train one or more ML models to generate one or more trained ML models via the block 640. In such an example, a trained ML model may be utilized in one or more workflows that may be driven by one or more features of a framework portal per a framework portal block 650 for output generation. For example, consider a scenario where one or more locations may be identified for a new rock cube and / or enhancement of an existing rock cube. In such an example, one or more inputs may be utilized to generate output where the inputs may be rock cube specifications and the output may be predicted production for the rock cube specifications. Such an approach may be implemented within one or more optimization loops such that, for example, various inputs may be considered to arrive at an optimized set of inputs for a rock cube that may aim to maximize production (e.g., via an achievable SRV, etc.).

[0112] As an example, the workflow 600 may be data-driven, with or without use of one or more physics-based models. As an example, the workflow 600 may generate production predictions for rock cubes on a month-by-month and / or one or more other bases. As an example, consider a workflow that generates production predictions for a number of months where such output may be utilized in one or more other models (e.g., Arp’s model, etc.) to predict further production (e.g., post-peak production, etc.).

[0113] As explained, a framework may be driven by an interactive interface (e.g., one or more graphical user interfaces (GUIs), etc.). As an example, a framework may be operatively coupled to one or more databases, which may be or include one or more real-time databases. As an example, a framework may provide for automated workflows, for example, to automatically generate rock cubes as potential for production improvements in a field or fields. As an example, a framework may be integrated into an environment such as, for example, the DELFI environment. As anexample, a framework may be compatible for interoperation with the PETREL framework.

[0114] As to the block 610, consider use of horizontal well and production meta data for analysis and modeling, which may be accessed from one or more databases. For example, consider the WellDatabase, which can be accessed by the SNOWFLAKE connector in the PYTHON language; noting that the SNOWFLAKE connector can also be applied in the DATAIKU framework. As an example, input data may include Well Header, Deviation, Production, and Stimulation. As an example, Formation Tops may also be loaded. As an example, a schema may be flexible to accept one or more other features such as, for example, Geological Features, Stage data, etc.

[0115] As an example, data may be supplemented and / or augmented using one or more techniques. For example, consider using one or more simulation techniques to generate synthetic and / or augmented data. As an example, one or more frameworks may be utilized to generate simulation data, which may be utilized for training, assessing, etc., one or more groups of wells, which may be or may include one or more groups of hydraulically fractured wells. As an example, a group of wells may be a production unit of hydraulically fractured wells, which may be already hydraulically fractured, to be hydraulically fractured, to be hydraulically re-fractured, etc.

[0116] FIG. 7 and FIG. 8 shows map views 700 and 800, respectively, of various wells as based on well data, including well data for horizontal well path distribution in west Texas. Such data may be accessed from one or more databases. While the WellDatabase is mentioned, an S&P (IHS) database may be accessed, which includes a greater number of wells than WellDatabase, particularly for deviation surveys. As an example, a workflow may be mapped an IHS table into the format of WellDatabase and feed them into a framework.

[0117] As to the block 630, it may be implemented to generate a well object with each well including cultural, geometry, production merged from file Header, Directional, and Production. For example, input data files from the WellDatabase may include: Header, Directional, Production / Production Volume, Stimulation, and Formation. As an example, a workflow may include extracting various headers from a file Header, which may include, for example, one or more of: api, wellid, wellname,welltype, basin, play, primaryformation, currentoperator, reportedoriginaloperator, field, county, state, lease, latitude, longitude, bhlatitude, bhlongitude, measureddepth, trueverticaldepth, laterallength, groundelevation, kellybushingelevation, trajectory, wellboreprofile, koplatitude, koplongitude, heellatitude, heellongitude, spuddate, completiondate, firstproddate, cumulativegas, cumulativeoil, cumulativewater. In such an example, these may then be stored into a well object.

[0118] As an example, a workflow may convert time-related headers “spuddate, completiondate, and firstproddate” to timestamp and store them into the well object as three new headers “spuddate_timestamp, completion_timestamp, and first_prod_timestamp”.

[0119] As an example, a workflow may group file Directional by “wellid” and sorted by “measureddepth”, then extract headers “measureddepth, incline, azimuth”, then combined these headers into one and stored it into the well object with 1 new header “deviation”.

[0120] As an example, a workflow may then create survey trajectory attributes for each well within the well object by computing and storing the following new headers: “Proyectedto, latitude_dms, longitude_dms, head_x, head_y, kop_x, kop_y, heel_x, heel_y, bh_x, bh_y, interpolateddeviation, survey, tvd_avg, easting_avg, northing_avg, survey_H_easting_resampled, survey_H_northing_resampled, survey_H_depth_resampled, survey_H_lenght”.

[0121] As an example, PYTHON packages of geopandas and wellpathpy may be utilized to build a survey for each well. In such an example, a framework may provide for converting a survey from UTM X & Y coordinates to Longitude & Latitude, and adding more headers: “surveyjat, surveyjon, surveyjinestring, survey_H_deg_resampled, survey_lat_avg, survey_lon_avg”.

[0122] As an example, a workflow can include loading production data followed by computing and merging new headers to the well object: “prod_reportdate, prod_days, prod_welloil, prod_wellgas, prod_wellwater, prod_welldailyoil”.

[0123] As an example, a workflow can include computing and merging new headers to the well object after loading stimulation data: “stimulation, totalproppantvolbbl, totalproppantvolcft, totalfracfluidvolume, totalbasewatermass, totalbasewatervolumegallons, totalfracfluidmass, totalproppantmass,acidtreatmentpresent, fluidvolperft, watermassperft, fluidmassperft, proppantmassperft”.

[0124] As an example, a well object may include a number of features. For example, in the foregoing example workflow, a well object may include approximately 84 features after loading Header, Directional, Production, and Stimulation.

[0125] FIG. 9A and FIG. 9B show example plots and graphics 902, 904, and 910 for data analysis per well. As shown, various types of metrics may be plotted, analyzed, etc., for a well, which may be presented in a multi-dimensional view (e.g., for a horizontal portion of the well). Such plots and / or graphics may be generated using pre-processed data. As explained, clustering and analysis may be applied as part of a workflow using pre-processed data. In the example of FIG. 9A and FIG. 9B, various graphics 910 illustrate a clustering technique that may be a spatial-temporal technique (e.g., ST-DBSCAN, etc.).

[0126] As an example, a goal of clustering may be to generate a rock cube object (e.g., an object that may be part of an object-oriented programming platform, etc.) including the same attributes as the previous well object such as cultural, geometry, and production, etc., where such attributes may then be feed into a rock cube data analysis for a model for production prediction.

[0127] As an example, for spatial and temporal analysis, which may relate wells in space and in time (e.g., time of fracturing, time of production, etc.), a technique such as, for example, ST-DBSCAN may be implemented as an unsupervised ML technique based on DBSCAN for clustering spatial and temporal data (see, e.g., Derya Birant, 2007) (see also, e.g., the graphics 910 of FIG. 9B).

[0128] DBSCAN computes the Euclidean distance dist(i, j) between points i and j that tells how far points i and j are, in order to determine the set of points are similar enough to be considered as a cluster or group. dist(i, j) =(xil — xjl)2+ (xi2 — xj2)2+ — I- (xin — xjn)2i = (xil. xi2, ... , xin) and j = (xjl. xj2, ... , xjn)

[0129] ST-DBSAN utilizes several parameters Epsl, Eps2, MinPts, where Epsl is the spatial distance, Eps2 is the temporal distance, and MinPts is the minimumnumber of points within Epsl and Eps2 distance of a point (see, e.g., Derya Birant, 2007):Epsl = / (xl - yl)2+ (x2 — y2)2Eps2 =(tl - t2)2

[0130] An example trial workflow used an average of horizontal section coordinates and first production date of each well for clustering wells.

[0131] FIG. 10 shows a map view 1000 of clustering results where numbers indicate identifiers for each of the groups. Specifically, the map view 1000 is a map of cubes after clustering wells and generating a 2D area using a ConvexHull technique in the south of New Mexico and the north of Texas.

[0132] In the example of FIG. 10, the clustering technique added new headers to the well object: “st_db, kmeans”. K-Means was hired as an extra method to group wells. As to clustering, one or more K-Means approaches may be implemented where, for example, an optimal k-value (e.g., K or k) may be determined using one or more techniques, which may include, for example, an elbow technique.

[0133] As an example, a generated cube object may include attributes: “rc_id, geometry, geometry_xy, acres, data_3D, wellsids, wellsname, n_wells, sum_gas, sum_oil, sum_water, ini_firstprod_date, last_firstprod_date, n_months_prod_avg”. As indicated, attributes may include geometry attributes and production attributes where, for example, production attributes may include production for one or more fluids (e.g., consider oil, water, and gas).

[0134] As explained, data analysis may be performed on a per cube basis (e.g., per rock cube basis) using one or more object detection techniques such as, for example, the Hough transform technique.

[0135] FIG. 11 shows various plots and graphics 1100 associated with application of the Hough transform technique, which may be used to extract features of a particular shape in an image. For example, consider implementing the Hough transform of R. Fisher (2003) to detect one or more lines in a cross section of a well location within a rock cube (see, e.g., Fisher, R., (2003), “Image Transforms - Hough Transform” (https: / / homepages.inf.ed.ac.uk / rbf / HIPR2 / hough.htm), which is incorporated by reference herein in its entirety).

[0136] In the example of FIG. 11 , a close view of wells including production and bench scanning using the Hough transform technique within a rock cube is shown. As shown in the lower right, lines for benches can be detected, which may be detected along with angle such as, for example, dip angle in a cross-section.

[0137] FIG. 12 shows an example of a wine rack geometry 1200 that may be utilized to characterize wells in a rock cube where the wine rack geometry 1200 may be generated on the basis of output from application of a Hough transform technique. As shown, the particular rock cube has an identifier RC-30 that can be characterized by a grid in total vertical depth below sea level (TVDSS) and tank development width (X), which may be specified as indexes, as may be based on length (e.g., feet or meters).

[0138] FIG. 13 shows example plots 1300 for production and bench comparison between two rock cubes where cross-sectional views of the two rock cubes (RC-30 or ID-30 and RC-67 or ID-67) are overlain. As explained, each of the cross-sectional views may be defined using a geometry or geometries, which may include one or more wine rack geometries. As shown in FIG. 13, the production for the rock cube ID-67 exceeds that for the rock cube ID-30. Such comparisons may provide for assessing one or more aspects of a rock cube. For example, consider assessing efficiency, opportunity to re-fracture and / or apply one or more other stimulation techniques that may improve production. In the examples of FIG. 13, the production time line extends to 18 months for oil, gas, and water production.

[0139] FIG. 14 shows example plots for an average production comparison by month between rock cubes within Lea County along with a tank analysis in Loving County, noting that a rock cube may be referred to as a tank. As shown, the tank analysis results indicate wells per tank, average horizontal length, acres, etc., which may be metrics that can be utilized for planning one or more field operations.

[0140] As explained, a workflow can include generating various types of data for rock cubes where such data may be utilized as a basis for modeling and forecasting. As an example, a decision tree approach that is boosted (e.g., XGBoost, etc.) may be utilized to predict production for given input. As an example, a Deep Learning CNN approach may be utilized, additionally or alternatively.

[0141] As to XGBoost, it is a boosting technique designed to train weak learners that learn from their predecessors’ mistakes, and reduce bias. XGboost are sequentiallearners: models are trained sequentially so as to learn from predecessors’ mistakes. At each iteration XGBoost fits a weak learner to the opposite of the gradient of the current fitting error with respect to the current ensemble model, in order to find an optimal model after building a weighted sum of weak learners (see, e.g., Rocca, J. (2019), “Ensemble methods: bagging, boosting and stacking”, Towards Data Science (https: / / towardsdatascience.com / ensemble-methods-bagging-boosting-and-stacking- c9214a10a205), which is incorporated by reference herein in its entirety).

[0142] FIG. 15 shows an example graphic of boosted trees 1510 that may be part of an ML technique for generating one or more ML models that can predict production based on given input in a per month or another basis. As an example, a mask may be designed to keep true and predicted in a same shape by ignoring nan values in the true values.

[0143] As an example, historical data may be used to build a regression model per month, such as, for example, the monthly models 1520 as shown in FIG. 15. The selected geometry and stimulation features of neighbor cubes may be initial features to predict the 1stmonth production using a 1stmodel. Then the 1stmonth production and the neighbor data may be used to predict the 2ndmonth using a 2ndmodel, and so on. In such an example, the (n-1 )th month production may be used to predict the nth month except the 1stmonth.

[0144] As to input features, consider, as an example, features per tank used for the forecasting per month in XGBoost as follows: Average horizontal length (Wells navigation), Lateral length (wide), Vertical extension (height), Number of wells, Number of benches, Proppant per foot, and Frac fluid volume per foot. Such input features include various aspects of a rock cube and aspects of field operations. For example, field operational parameters such as proppant per unit distance and / or frac fluid volume per unit distance may be included, which may provide for understanding how to control field operations for a new well or wells to be drilled and hydraulically fractured. In machine learning, features may be a result of feature engineering. For example, feature engineering may involve selecting, generating, etc., one or more features that may provide for improved training, performance, etc., of one or more machine learning models. In various instances, feature engineering may provide for effective training and performance. As an example, a feature may be a physics-based feature, a user generated feature, a parameter feature, etc. In various instances, anappropriate number and appropriate types of features may be engineered, etc., to provide for appropriate training, performance, etc. As an example, feature engineering may provide for conservation of resources, such as, for example, computational resources. As an example, feature engineering may provide for improved hyperparameter tuning. As mentioned, a hyperparameter may be a parameter of a machine learning model, such as, for example, the parameter “K” (or k) in K-Means clustering. As mentioned, an elbow technique may be employed to determine an appropriate or optimal value of the parameter K. As an example, one or more machine learning models may be generated, tuned, etc., using one or more techniques, which may involve one or more of feature engineering and hyperparameter tuning.

[0145] As an example, a workflow may include ranking. For example, consider a workflow that can include ranking the rock cube performance in a selected area based on historical data, which may provide insight for rock cube design optimization. As to ranking, it may be based on one or more parameters such as, for example, one or more of oil, gas, water, etc. As an example, a ranking may be based at least in part on one or more operational parameters (e.g., one or more hydraulic fracturing parameters, etc.).

[0146] FIG. 16 shows example plots and an example table 1600 as to forecasts. In particular, FIG. 16 shows oil, gas, water production prediction on rock cube attributes for a particular project based on neighbor rock cube historical data. As shown, the ML-based approach to prediction is superior to neighbor-based prediction. In particular, the neighbor-based predictions for oil and gas per unit distance is too high in comparison to the actual values where the ML-based predictions are quite close to the actual values. Hence, the ML-based approach to predictions can be superior to a neighbor-based approach. FIG. 16 also shows a table of real versus simulated results that include various metrics, which include field operations metrics. In such an approach, one or more field operations may be tailored, optimized, assessed, etc. The results in FIG. 16 are generated using the aforementioned XGBoost ML technique where, for example, score metrics included R2, MAE and RMSE.

[0147] As mentioned, a neural network-based approach may be implemented for prediction of production and / or one or more optimizations. As an example, an artificial neural network (ANN) approach may improve an ability to account for well griddistribution (e.g., geometry such as, for example, wine rack geometry, etc.), which may be more challenging to implement using an XGBoost model where, for example, bench distribution (e.g., arrangement type, symmetry and proximity) may not necessarily be fully represented in geometric features generated for forecast per rock unit.

[0148] FIG. 17 shows an example of a geometry 1700 that may be utilized in an ANN-based approach to prediction of production. For example, the geometry 1700 can be a bench grid represented by ellipses on each of the wells plotted in a local bench coordinate considering the horizontal distance between wells as well as the vertical distance. In FIG. 17, a rasterized version 1710 of the geometry 1700 is shown, which may be a rasterized version of a polygon plot that results in a 2D array that includes ellipse shapes. In such an example, the raster grid may be populated with one or more parameters such as, for example, average length of a tank, average proppant per foot, and average frac fluid volume per area, thereby encoding these features in a number of channels (e.g., consider RGB channels) of the 2D raster. As an example, three channels may be utilized where an ANN-based approach may leverage one or more ANNs that are tailored for images (e.g., image analysis, image recognition, image generation, etc.), which may be for three-channel images (e.g., red, green and blue and / or one or more other color scale types of images). While three channels are mentioned, as an example, a number of channels may be appropriately selected, engineered, etc. For example, consider an approach that may provide for a number of channels from 1 to N where N may depend on one or more aspects of input, model-type, engineered features, etc. In such an example, N may be two or greater, three or greater, etc. In FIG. 17, a graphic 1720 is shown that includes some example operations as to feature learning and regression for an ANN-based approach. As shown, output may be a vector of values of production on a basis such as a monthly basis.

[0149] As an example, for memory optimization purposes, a raster grid may be scaled to a 64x256x3 matrix. With rasterized benches images, a CNN model may be trained using the rasters as inputs and monthly production per fluid as target (e.g., outputs). Such an approach may be utilized additionally or alternatively to a decision tree-based approach. As explained, an image-based approach may provide for conserving spatial information that may be germane to a rock cube and that may be amenable to rendering to a display in a manner that allows for human understanding,which may be for quality control and / or one or more other purposes. For example, a human can visualize the raster 1710 and make some sense of its content in relationship to an arrangement of wells. An ANN-based approach may provide for generating output akin to that of the boosted tree approach where the ANN-based approach has a grid distribution inherit to raster images, which may provide valuable information not considered by the linear features.

[0150] FIG. 18 shows an example model architecture 1800 of an ANN-based approach that includes three convolutional layers for image encoding and a regression layer for a production forecast, using 1 neuron per month as target.

[0151] As explained, a framework may provide for rendering one or more GUIs for user interactions. As an example, a framework may include a web-based interface that may be a portal that may provide for rendering of one or more dashboard style GUIs to one or more displays.

[0152] FIG. 19 shows examples of graphics and code, for example, for one or more of DATAIKU and stand-alone PYTHON implementations. As an example, a framework may include sections: Home, Data, Tanks and Designer.

[0153] As to Home, it may be a landing page or start page for various options. As to Data, there may be multiple ways in a drop-down menu to select how to input data. For example, a user may directly access WellDatabase by user defined polygon(s). As an example, a user may access a stored WellDatabase or IHS data previously download and pre-processed.

[0154] FIG. 20 shows an example GUI 2000 for an area-based approach, which may utilize a circle, a polygon, etc. In such an example, an area may be drawn, selected, etc., such that data for wells within the area are accessed from one or more databases.

[0155] FIG. 21 shows an example GUI 2100 for wells within the area of the GUI 2000 where the GUI 2100 may indicate horizontal paths of wells.

[0156] FIG. 22, FIG. 23, FIG. 24, and FIG. 25 shows example GUIs 2200, 2300, 2400, and 2500 for tabs such as Create tanks, Review tanks, Tanks performance, and Tanks comparison, respectively. In such an example, Create tanks may cluster wells together to create cubes; Review tanks may provide for review of well and cube attributes such as geometry and production curves for each cube; Tanks performance may provide for ranking and review the tank performance by user selected metrics;and Tanks comparison may provide for comparing production, bench, and other attributes between multiple cubes, the comparison could be per tank or per well.

[0157] FIG. 26 and FIG. 27 show example GUIs 2600 and 2700 as part of a designer workflow. In the example of FIG. 26, a tab Area of interest (AOI) lets a user select an AOI and then display production and statistics of neighbor cubes and not- clustered wells. The data can be displayed by tank or by well. In the example of FIG. 27, the tab New development can provide for display of an optimal rock cube design.

[0158] FIG. 28 and FIG. 29 show example GUIs 2800 and 2900 that may provide for an overview of front-end and back-end workflow interactions within a data science framework (e.g., consider the DATAIKU framework).

[0159] As an example, a framework may provide for data export where, for example, pre-processed data can be directly imported into a framework such as, for example, the PETREL framework where data may be written into an ASCII file in the PETREL format, such as well headers in text including Longitude & Latitude and X&Y, and deviation in CSV. As an example, created rock cubes may be written into a shape file format of the PETREL framework.

[0160] FIG. 30, FIG. 31 , FIG. 32, and FIG. 33 show example GUIs 3000, 3100, 3200, and 3300 for the Loving County well path, cubes, inside a cube after importing into the PETREL framework.

[0161] As an example, a framework may provide for geoscreening. For example, consider a framework that may provide for interaction with a plug-in for the PETREL framework for selection of more representative geologic models that capture a range of static and dynamic variability. As an example, such an approach may enable the study of dynamic connectivity across different geologic realizations in tandem with volumetric calculations. As an example, a framework may provide for use of one or more of spatial-temporal clustering and / or layer detection that may be part of a geoscreening workflow.

[0162] As an example, a workflow may include generating realizations as part of an optimization process where, for example, sampling may be performed from the realizations to generate output that may be utilized to select an optimal realization and / or for realization ranking. In such an example, realizations may be generated using one or more parameters as variables to generate a space from which sampling may be performed. As explained, optimization may be based on one or more metrics,which may be, for example, one or more production metrics, one or more operational metrics, etc. As an example, an ML-based approach may provide for more rapid optimization compared to an approach that depends on execution of a simulator using a multidimensional numerical simulation mesh where physics-based equations are utilized.

[0163] As explained, a framework may provide for predicting production of drilling units, optimizing selection of drilling units, optimizing parameters for field operations for drilling unit, etc. As explained, unconventional drilling units (e.g., rock cubes) may be assessed, planned, created, improved, etc., using one or more ML- based techniques for forecasting, which may include use of a 2D CNN, for example, using images in a wine rack geometry with imbedded features.

[0164] As explained, one or more ML techniques may provide for forecasting production streams in unconventional wells using tabular features such as proppant load and well depth and focused on single-well performance. With movement to large multiwell development programs, one or more ML models may provide for taking into account relative spatial position of wells, for example, in a wine rack cross-sectional view (e.g., gun barrel cross-section, etc.). As an example, an ML model may depend on engineered features such as, for example, staggered offset and vertical offset. As explained, a 2D CNN may encode geometric aspects of a wine rack geometry. As an example, one or more ANNs may be utilized, which may include, for example, one or more features of a ResNet ANN, etc. A ResNet ANN may be a deep learning model in which the weight layers learn residual functions with reference to the layer input.

[0165] As an example, a framework may provide for performing spatial- temporal density base clustering to create the drilling units (also referred as rock cubes or tanks), engineering of relevant features at each drilling unit such as geometry and stimulation, creating a rasterized 2D matrix of the wine rack geometry of each drilling unit, embedding stimulation features to the matrix to create a multidimensional tensor compatible with a CNN workflow, training a CNN model for regression using the wine rack rasters as inputs and monthly production of each drilling unit as targets, and evaluating optimization with real data and creating one or more forecasting scenarios.

[0166] In an example trial, a framework was applied to the Wolfcamp play in the northern Midland Basin. Monthly production per completed foot of multiwell drilling units were compared to the forecasted results of the CNN regression modeldemonstrating good correlation and the applicability of CNN for forecasting methodology. Through the training process, filter activations of the 2D CNN learn the relative importance of well offset and spacing across a normalized wine rack crosssection. A trained model from the northern Midland Basin indicates as much as 30% improved production performance when stacked benches are developed vertically (zero offset) as opposed to a staggered scheme.

[0167] As an example, a framework may provide for generation of an rendering of one or more types of GUIs. As an example, a GUI may be interactive and include one or more graphical controls that may be actuatable via one or more humanmachine interfaces and / or machine-machine interfaces. As an example, a GUI may transition automatically from one state to another state responsive to an interaction or interactions. For example, consider a GUI with a panel that allows for generation, modification, etc., of one or more geometrical representations of a subsurface region, which may include one or more subsurface features as related to existing structures and / or possible structures. As an example, a GUI panel may provide for wine rack generation, modification, etc. For example, consider a GUI panel that includes a cross-section view, which may be a cut-away view, a transparent view, etc., of a portion of a subsurface region. In such an example, a wine rack type of geometric representation of wells and / or surfaces may be generated, modified, etc. For example, consider one or more of the wine rack representations of one or more of the various figures where a GUI may provide for dragging and dropping, moving, shaping, etc., renderable objects to generate a desirable arrangement of wells with respect to one or more subsurface features (e.g., surfaces, layers, etc.).

[0168] As an example, a GUI may provide for design of a wine rack and use of one or more trained models (e.g., one or more ML models, etc.) plus one or more features, such as, for example, average length of a tank, average proppant per foot, average frac fluid volume per area, etc., to estimate an expected production on the designed wine rack by the user. In such an example, a framework may provide for dynamic responses to GUI interactions, for example, where a dynamic option allows a user to change one or more parameters, allows for automated optimization, etc. In such an approach, a framework tool may provide for generation of suggested parameter values that may aim to maximize production for a given design, machine optimized design, etc. As an example, a framework may provide for generation of aninitial design followed by optimization of the design where a GUI may dynamically render an optimized design and provide for generation of predicted production. In such an example, a GUI may allow for receipt of input whereby an optimized design may be tailored (e.g., modified, etc.), which may be part of an iterative process (e.g., a loop, etc.). In such an example, a user may evaluate and modified and / or accept one or more designs, for example, in accordance with one or more criteria, etc. As explained, a framework may provide for dynamic generation of a wine rack, which may be an optimized wine rack, via one or more GUIs that may provide for a human-in-the- loop (HITL) approach to achieving one or more goals (e.g., maximizing production, balancing cost, etc.).

[0169] FIG. 34 shows various example graphics and plots 3400 from the aforementioned example trial. In FIG. 34, the upper graphic and plot are for a test dataset of a real case and the lower graphic and plot are for synthetic scenarios with eight wells and two different horizontal lengths.

[0170] As an example, a framework may be applied in one or more scenarios where drilling units may include irregular well spacing where manual geometric feature generation is challenging to implement. As an example, a visual representation of 2D CNN filter activations from the training process may be rendered to a display to provide insights into model explainability and geometric controls on the production forecast. As an example, a trained model can be used to create different scenarios and / or evaluate proposed developments based on wine rack geometries and stimulation schemes. As an example, a workflow may be applied for asset development optimization and planning and assessment of infill drilling or refracturing opportunities.

[0171] As an example, a workflow may provide for analysis of multi-well historical data from nearby areas as baseline, selection of a ranking metric which defines the optimal production scheme (e.g., total production of oil and / or gas per drilling unit, production per foot per drilling unit, etc.), implementation of an optimization scheme for rock cube development by drill unit based on fluid (e.g., maximize oil production, maximize gas production, minimize water ratio, combined optimization such as maximize oil and gas while minimizing water, etc.), and provide for time reduction for decision making in large areas.

[0172] As an example, a framework may be utilized to output an optimized recommended geometry and completions for a rock cube based on a ranking metric,which may be a default metric, a user selectable metric, or defined by a user. As an example, a framework may generate intermediate results that may be passed to a component for ML-based Decline Curve Analysis (DCA-ML), for example, to estimate production forecast of a rock cube.

[0173] As an example, a framework may provide for an assessment of well interference. As explained, fractures may extend outwardly from a well and a risk may exist of fractures of one well connecting fluidly with fractures from another well (e.g., short-circuiting). As an example, short-circuiting may be generally viewed as detrimental; noting that in some instances, such as, for example, geothermal energy extraction, caging techniques may benefit from short-circuiting (e.g., fracture connections).

[0174] As an example, a framework may be implemented to enhance hydraulic fracturing operations, achieve better production rates, and reduce operational costs through data-driven decisions and optimized fracturing parameters. As explained, a framework may be integrated with one or more other frameworks, environments, systems, etc. For example, consider integration with one or more products for reservoir management and monitoring to create a more robust and comprehensive operational platform.

[0175] As explained, a framework may provide for enhanced real-time monitoring and / or control of field operations. For example, such a framework may be integrated with one or more real-time data acquisition technologies, which may thereby provide for real-time monitoring and optimization of hydraulic fracturing operations to enhance responsiveness of field operations to changing reservoir conditions.

[0176] As to types of machine learning (ML) models, consider one or more of a support vector machine (SVM) model, a k-nearest neighbors (KNN) model, an ensemble classifier model, a neural network (NN) model, etc. As an example, a machine learning model can be a deep learning model (e.g., deep Boltzmann machine, deep belief network, convolutional neural network, stacked auto-encoder, etc.), an ensemble model (e.g., random forest, gradient boosting machine, bootstrapped aggregation, AdaBoost, stacked generalization, gradient boosted regression tree, etc.), a neural network model (e.g., radial basis function network, perceptron, back-propagation, Hopfield network, etc.), a regularization model (e.g., ridge regression, least absolute shrinkage and selection operator, elastic net, leastangle regression), a rule system model (e.g., cubist, one rule, zero rule, repeated incremental pruning to produce error reduction), a regression model (e.g., linear regression, ordinary least squares regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing, logistic regression, etc.), a Bayesian model (e.g., naive Bayes, average on-dependence estimators, Bayesian belief network, Gaussian naive Bayes, multinomial naive Bayes, Bayesian network), a decision tree model (e.g., classification and regression tree, iterative dichotomiser 3, C4.5, C5.0, chi-squared automatic interaction detection, decision stump, conditional decision tree, M5), a dimensionality reduction model (e.g., principal component analysis, partial least squares regression, Sammon mapping, multidimensional scaling, projection pursuit, principal component regression, partial least squares discriminant analysis, mixture discriminant analysis, quadratic discriminant analysis, regularized discriminant analysis, flexible discriminant analysis, linear discriminant analysis, etc.), an instance model (e.g., k-nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, etc.), a clustering model (e.g., k-means, k-medians, expectation maximization, hierarchical clustering, etc.), etc.

[0177] As an example, a machine model may be built using a computational framework with a library, a toolbox, etc., such as, for example, those of the MATLAB framework (Math Works, Inc., Natick, Massachusetts). The MATLAB framework includes a toolbox that provides supervised and unsupervised machine learning algorithms, including support vector machines (SVMs), boosted and bagged decision trees, k-nearest neighbor (KNN), k-means, k-medoids, hierarchical clustering, Gaussian mixture models, and hidden Markov models. Another MATLAB framework toolbox is the Deep Learning Toolbox (DLT), which provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. The DLT provides convolutional neural networks (ConvNets, CNNs) and long shortterm memory (LSTM) networks to perform classification and regression on image, time-series, and text data. The DLT includes features to build network architectures such as generative adversarial networks (GANs) and Siamese networks using custom training loops, shared weights, and automatic differentiation. The DLT provides for model exchange various other frameworks.

[0178] As an example, the TENSORFLOW framework (Google LLC, Mountain View, CA) may be implemented, which is an open-source software library for dataflow programming that includes a symbolic math library, which can be implemented for machine learning applications that can include neural networks. As an example, the CAFFE framework may be implemented, which is a DL framework developed by Berkeley Al Research (BAIR) (University of California, Berkeley, California). As another example, consider the SCIKIT platform (e.g., scikit-learn), which utilizes the PYTHON programming language. As an example, a framework such as the APOLLO Al framework may be utilized (APOLLO. Al GmbH, Germany). As an example, a framework such as the PYTORCH framework may be utilized (Facebook Al Research Lab (FAIR), Facebook, Inc., Menlo Park, California).

[0179] As an example, a training method can include various actions that can operate on a dataset to train a ML model. As an example, a dataset can be split into training data and test data where test data can provide for evaluation. A method can include cross-validation of parameters and best parameters, which can be provided for model training.

[0180] The TENSORFLOW framework can run on multiple CPUs and GPUs (with optional CUDA (NVIDIA Corp., Santa Clara, California) and SYCL (The Khronos Group Inc., Beaverton, Oregon) extensions for general-purpose computing on graphics processing units (GPUs)). TENSORFLOW is available on 64-bit LINUX, MACOS (Apple Inc., Cupertino, California), WINDOWS (Microsoft Corp., Redmond, Washington), and mobile computing platforms including ANDROID (Google LLC, Mountain View, California) and IOS (Apple Inc.) operating system-based platforms.

[0181] TENSORFLOW computations can be expressed as stateful dataflow graphs; noting that the name TENSORFLOW derives from the operations that such neural networks perform on multidimensional data arrays. Such arrays can be referred to as “tensors”.

[0182] As an example, one or more features of the KERAS library may be utilized. The KERAS library is an open-source library that provides a PYTHON interface for artificial neural networks (ANNs). The KERAS library can act as an interface for the TENSORFLOW library.

[0183] As an example, a device may utilize TENSORFLOW LITE (TFL) or another type of lightweight framework. TFL is a set of tools that enables on-devicemachine learning where models may run on mobile, embedded, and loT devices. TFL is optimized for on-device machine learning, by addressing latency (no round-trip to a server), privacy (no personal data leaves the device), connectivity (Internet connectivity is demanded), size (reduced model and binary size) and power consumption (e.g., efficient inference and a lack of network connections). TFL includes multiple platform support, covering ANDROID and iOS devices, embedded LINUX, and microcontrollers. TLF provides diverse language support, which includes JAVA, SWIFT, Objective-C, C++, and PYTHON. TFL provides high performance, with hardware acceleration and model optimization. As an example, one or more machine learning tasks may include, for example, classification, regression, object detection, pose estimation, question answering, text classification, etc., on one or more of multiple platforms.

[0184] FIG. 35 shows an example of a method 3500 and an example of a system 3590. As shown, the method 3500 can include a reception block 3510 for receiving input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; a prediction block 3520 for predicting production data for the group of wells using a machine learning model; and an output block 3530 for outputting the predicted production data.

[0185] The method 3500 is shown in FIG. 35 in association with various computer-readable media (CRM) blocks 3511 , 3521 , and 3531 . Such blocks generally include instructions suitable for execution by one or more processors (or processor cores) to instruct a computing device or system to perform one or more actions. While various blocks are shown, a single medium may be configured with instructions to allow for, at least in part, performance of various actions of the method 3500. As an example, a computer-readable medium (CRM) may be a computer-readable storage medium that is non-transitory and that is not a carrier wave. As an example, one or more of the blocks 3511 , 3521 , and 3531 may be in the form processor-executable instructions.

[0186] In the example of FIG. 35, the system 3590 includes one or more information storage devices 3591 , one or more computers 3592, one or more networks 3595 and instructions 3596. As to the one or more computers 3592, each computer may include one or more processors (e.g., or processing cores) 3593 and memory 3594 for storing the instructions 3596, for example, executable by at least one of theone or more processors 3593 (see, e.g., the blocks 3511 , 3521 , and 3531 ). As an example, a computer may include one or more network interfaces (e.g., wired or wireless), one or more graphics cards, a display interface (e.g., wired or wireless), etc.

[0187] As an example, a method can include receiving input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predicting production data for the group of wells using a machine learning model; and outputting the predicted production data. In such an example, the group of wells can define a rock cube as the hydraulically fractured production unit.

[0188] As an example, a machine learning model may include a tree-based model where, for example, the tree-based model may be a boosted model.

[0189] As an example, a machine learning model may include an artificial neural network-based model.

[0190] As an example, a machine learning model may be trained using raster images where, for example, the raster images represent cross-sectional views of offset wells.

[0191] As an example, a machine learning model may be trained using offset well data. For example, consider a machine learning model that is trained using operational parameters of hydraulic fracturing operations performed at the offset wells.

[0192] As an example, offset well data may be processed using clustering where, for example, the clustering includes spatial-temporal clustering.

[0193] As an example, offset well data may be processed using object detection where, for example, the object detection includes Hough transform-based object detection. As an example, object detection may detect one or more layer boundaries (e.g., consider formation and / or bench boundaries).

[0194] As an example, a group of wells may include at least three wells. As an example, a group of wells may include wells at different depths.

[0195] As an example, a group of wells may be represented using a wine rack geometry where, for example, the wine rack geometry is defined with respect to at least one bench.

[0196] As an example, a system can include a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: receive input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit;predict production data for the group of wells using a machine learning model; and output the predicted production data.

[0197] As an example, one or more computer-readable storage media including processor-executable instructions executable by a system to instruct the system to: receive input for a group of wells in a subsurface region, where the group of wells defines a hydraulically fractured production unit; predict production data for the group of wells using a machine learning model; and output the predicted production data.

[0198] As an example, a computer program product can include one or more computer-readable storage media that can include processor-executable instructions to instruct a computing system to perform one or more methods and / or one or more portions of a method.

[0199] In some embodiments, a method or methods may be executed by a computing system. FIG. 36 shows an example of a system 3600 that can include one or more computing systems 3601-1 , 3601-2, 3601-3 and 3601-4, which may be operatively coupled via one or more networks 3609, which may include wired and / or wireless networks.

[0200] As an example, a system can include an individual computer system or an arrangement of distributed computer systems. In the example of FIG. 36, the computer system 3601-1 can include one or more modules 3602, which may be or include processor-executable instructions, for example, executable to perform various tasks (e.g., receiving information, requesting information, processing information, simulation, outputting information, etc.).

[0201] As an example, a module may be executed independently, or in coordination with, one or more processors 3604, which is (or are) operatively coupled to one or more storage media 3606 (e.g., via wire, wirelessly, etc.). As an example, one or more of the one or more processors 3604 can be operatively coupled to at least one of one or more network interfaces 3607; noting that one or more other components 3608 may also be included. In such an example, the computer system 3601-1 can transmit and / or receive information, for example, via the one or more networks 3609 (e.g., consider one or more of the Internet, a private network, a cellular network, a satellite network, etc.).

[0202] As an example, the computer system 3601-1 may receive from and / or transmit information to one or more other devices, which may be or include, forexample, one or more of the computer systems 3601-2, etc. A device may be located in a physical location that differs from that of the computer system 3601 -1. As an example, a location may be, for example, a processing facility location, a data center location (e.g., serverfarm, etc.), a rig location, a wellsite location, a downhole location, etc.

[0203] As an example, a processor may be or include a microprocessor, microcontroller, processor module or subsystem, programmable integrated circuit, programmable gate array, or another control or computing device.

[0204] As an example, the storage media 3606 may be implemented as one or more computer-readable or machine-readable storage media. As an example, storage may be distributed within and / or across multiple internal and / or external enclosures of a computing system and / or additional computing systems.

[0205] As an example, a storage medium or storage media may include one or more different forms of memory including semiconductor memory devices such as dynamic or static random access memories (DRAMs or SRAMs), erasable and programmable read-only memories (EPROMs), electrically erasable and programmable read-only memories (EEPROMs) and flash memories, magnetic disks such as fixed, floppy and removable disks, other magnetic media including tape, optical media such as compact disks (CDs) or digital video disks (DVDs), BLUERAY disks, or other types of optical storage, or other types of storage devices.

[0206] As an example, a storage medium or media may be located in a machine running machine-readable instructions, or located at a remote site from which machine-readable instructions may be downloaded over a network for execution. As an example, various components of a system such as, for example, a computer system, may be implemented in hardware, software, or a combination of both hardware and software (e.g., including firmware), including one or more signal processing and / or application specific integrated circuits.

[0207] As an example, a system may include a processing apparatus that may be or include a general-purpose processors or application specific chips (e.g., or chipsets), such as ASICs, FPGAs, PLDs, or other appropriate devices.

[0208] As an example, a device may be a mobile device that includes one or more network interfaces for communication of information. For example, a mobile device may include a wireless network interface (e.g., operable via IEEE 802.11 , ETSIGSM, BLUETOOTH, satellite, etc.). As an example, a mobile device may include components such as a main processor, memory, a display, display graphics circuitry (e.g., optionally including touch and gesture circuitry), a SIM slot, audio / video circuitry, motion processing circuitry (e.g., accelerometer, gyroscope), wireless LAN circuitry, smart card circuitry, transmitter circuitry, GPS circuitry, and a battery. As an example, a mobile device may be configured as a cell phone, a tablet, etc. As an example, a method may be implemented (e.g., wholly or in part) using a mobile device. As an example, a system may include one or more mobile devices.

[0209] As an example, a system may be a distributed environment, for example, a so-called “cloud” environment where various devices, components, etc. interact for purposes of data storage, communications, computing, etc. As an example, a device or a system may include one or more components for communication of information via one or more of the Internet (e.g., where communication occurs via one or more Internet protocols), a cellular network, a satellite network, etc. As an example, a method may be implemented in a distributed environment (e.g., wholly or in part as a cloud-based service).

[0210] As an example, information may be input from a display (e.g., consider a touchscreen), output to a display or both. As an example, information may be output to a projector, a laser device, a printer, etc. such that the information may be viewed. As an example, information may be output stereographically or holographically. As to a printer, consider a 2D or a 3D printer. As an example, a 3D printer may include one or more substances that can be output to construct a 3D object. For example, data may be provided to a 3D printer to construct a 3D representation of a subterranean formation. As an example, layers may be constructed in 3D (e.g, horizons, etc.), geobodies constructed in 3D, etc. As an example, holes, fractures, etc., may be constructed in 3D (e.g., as positive structures, as negative structures, etc.).

[0211] Although only a few example embodiments have been described in detail above, those skilled in the art will readily appreciate that many modifications are possible in the example embodiments. Accordingly, all such modifications are intended to be included within the scope of this disclosure as defined in the following claims. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents, but also equivalent structures. Thus, although a nail and a screw maynot be structural equivalents in that a nail employs a cylindrical surface to secure wooden parts together, whereas a screw employs a helical surface, in the environment of fastening wooden parts, a nail and a screw may be equivalent structures.

Claims

CLAIMSWhat is claimed is:

1. A method comprising: receiving input for a group of wells in a subsurface region, wherein the group of wells defines a hydraulically fractured production unit; predicting production data for the group of wells using a machine learning model; and outputting the predicted production data.

2. The method of claim 1 , wherein the group of wells define a rock cube as the hydraulically fractured production unit.

3. The method of claim 1 , wherein the machine learning model comprises a tree-based model.

4. The method of claim 3, wherein the tree-based model comprises a boosted model.

5. The method of claim 1 , wherein the machine learning model comprises an artificial neural network-based model.

6. The method of claim 1 , wherein the machine learning model is trained using raster images.

7. The method of claim 6, wherein the raster images represent cross-sectional views of offset wells.

8. The method of claim 1 , wherein the machine learning model is trained using offset well data.

9. The method of claim 8, wherein the machine learning model is trained using operational parameters of hydraulic fracturing operations performed at the offset wells.

10. The method of claim 8, wherein the offset well data are processed using clustering.

11. The method of claim 9, wherein the clustering comprises spatial-temporal clustering.

12. The method of claim 8, wherein the offset well data are processed using object detection.

13. The method of claim 12, wherein the object detection comprises Hough transformbased object detection.

14. The method of claim 12, wherein the object detection detects one or more layer boundaries.

15. The method of claim 1 , wherein the group of wells comprises at least three wells.

16. The method of claim 1 , wherein the group of wells comprises wells at different depths.

17. The method of claim 1 , wherein the group of wells are represented using a wine rack geometry.

18. The method of claim 1 , wherein the wine rack geometry is defined with respect to at least one bench.

19. A system comprising: a processor; a memory operatively coupled to the processor; processor-executable instructions stored in the memory and executable to instruct the system to: receive input for a group of wells in a subsurface region, wherein the group of wells defines a hydraulically fractured production unit;predict production data for the group of wells using a machine learning model; and output the predicted production data.

20. One or more computer-readable storage media comprising processor-executable instructions executable by a system to instruct the system to: receive input for a group of wells in a subsurface region, wherein the group of wells defines a hydraulically fractured production unit; predict production data for the group of wells using a machine learning model; and output the predicted production data.

Images