Directional drilling frame

Abstract

A method may include receiving digital well plan data for directional drilling of a borehole of a well in a subterranean geological environment (1510); generating candidate trajectories for the borehole using the digital well plan data (1520); generating an ordering of the candidate trajectories using historical observations of one or more other boreholes, using values derived from inverse reinforcement learning (1530); and outputting one or more of the candidate trajectories for directional drilling of the borehole based on the ranking (1540).

Description

Background Art

[0001] A reservoir may be a subsurface formation that can be characterized, at least in part, by its porosity and fluid permeability. For example, a reservoir may be part of a basin, such as a sedimentary basin. A basin may be a depression in which sediments accumulate (e.g., caused by plate tectonic activity, subsidence, etc.). For example, where source rocks are present with an appropriate combination of burial depth and duration, a petroleum system may develop within the basin, which may form a reservoir containing hydrocarbon fluids (e.g., oil, gas, etc.).

[0002] In oil and gas exploration, interpretation involves analyzing data to identify and locate various subsurface structures (e.g., horizons, faults, geological bodies, etc.) within a geological environment. Various types of structures (e.g., layered formations) may indicate hydrocarbon traps or flow pathways, such as those associated with one or more reservoirs (e.g., fluid reservoirs). In the field of resource extraction, enhanced interpretation can allow for the construction of more accurate models of subsurface regions, which in turn can improve the characterization of the subsurface region for resource extraction purposes. The characterization of one or more subsurface regions within a geological environment can, for example, guide the execution of one or more operations (e.g., field operations, etc.). As an example, a more accurate model of the subsurface region can enable drilling operations to more accurately determine the trajectory of a borehole, such as the trajectory of a borehole penetrating a reservoir, and the trajectory of fluids produced through the borehole (e.g., as a completion, etc.). As an example, one or more workflows can be executed using one or more computing frameworks and / or one or more pieces of equipment that include features for planning, analysis, acquisition, modeling, control, etc., for exploration, interpretation, drilling, fracturing, production, etc. Summary of the Invention

[0003] A method may include: receiving digital well planning data for directional drilling of a well in a subsurface geological environment; generating candidate trajectories for the borehole using the digital well planning data; generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and outputting one or more candidate trajectories for the directional drilling of the borehole based on the ranking. A system may include one or more processors; a memory accessible to at least one of the one or more processors; processor-executable instructions stored in the memory and executable to instruct the system to: receive digital well planning data for directional drilling of a well in a subsurface geological environment; generating candidate trajectories for the borehole using the digital well planning data; generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and outputting one or more candidate trajectories for the directional drilling of the borehole based on the ranking. One or more non-transitory computer-readable storage media may include processor-executable instructions to instruct a computing system to: receive digital well plan data for directional drilling of a well in a subsurface geological environment; generate candidate trajectories for the borehole using the digital well plan data; generate a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and output one or more candidate trajectories from the candidate trajectories for directional drilling of the borehole based on the ranking. Various other apparatuses, systems, methods, etc. are also disclosed.

[0004] 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 to help limit the scope of the claimed subject matter. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The following detailed description refers to the accompanying drawings. Wherever possible, convenient features and advantages of the described embodiments may be more readily understood by referring to the following description taken in conjunction with the accompanying drawings.

[0006] Figure 1 An example of a system is shown;

[0007] Figure 2 An example of a system is shown;

[0008] Figure 3 An example of a system is shown;

[0009] Figure 4 An example of a system is shown;

[0010] Figure 5An example of a system is shown;

[0011] Figure 6 An example of a system is shown;

[0012] Figure 7 An example of a system is shown;

[0013] Figure 8 An example of a workflow is shown;

[0014] Figure 9 An example of a system is shown;

[0015] Figure 10 An example of a reinforcement learning process and an example of an inverse reinforcement learning process are shown;

[0016] Figure 11 An example of a method is shown;

[0017] Figure 12 An example of a table is shown;

[0018] Figure 13 An example of a method is shown;

[0019] Figure 14 An example of a method is shown; and

[0020] Figure 15 An example of a method is shown. DETAILED DESCRIPTION

[0021] This description should not be taken in a limiting sense and is made merely for the purpose of describing the general principles of the embodiments. Reference should be made to the issued claims to determine the scope of the described embodiments.

[0022] Figure 1 An example of a system 100 is shown that includes a workspace framework 110 that can provide instantiation, presentation, interaction with, etc., a graphical user interface (GUI) 120. Figure 1 In the example of , GUI 120 may include graphical controls for a computing framework (eg, application, etc.) 121 , projects 122 , visualization 123 , one or more other features 124 , data access 125 , and data storage 126 .

[0023] exist Figure 1In the example of , the workspace framework 110 may be customized for a particular geological environment, such as the exemplary geological environment 150. For example, the geological environment 150 may include layers (e.g., strata) that include a reservoir 151 and may intersect a fault 153. As an example, the geological environment 150 may be equipped with various sensors, detectors, actuators, and the like. For example, equipment 152 may include communication circuitry for receiving and sending information relative to one or more networks 155. Such information may include information associated with downhole equipment 154, which may be equipment that acquires information, aids in resource recovery, and the like. Other equipment 156 may be located remotely from the well site and include sensing, detection, transmission, or other circuitry. Such equipment may include storage and communication circuitry to store and transmit data, instructions, and the like. As an example, one or more satellites may be provided for purposes of communication, data acquisition, and the like. For example, Figure 1 A satellite is shown in communication with a network 155, which may be configured for communications, noting that the satellite may additionally or alternatively include circuitry for imaging (eg, spatial, spectral, temporal, radiometric, etc.).

[0024] Figure 1 The geological environment 150 is also shown as optionally including equipment 157 and 158 associated with a well, the well including a substantially horizontal portion that may intersect one or more fractures 159. For example, consider a well in a shale formation, which 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 laterally extending reservoir. In such an example, there may be lateral variations in properties, stresses, etc., where assessment of such variations may assist in planning, operations, etc., to develop the laterally extending reservoir (e.g., via fracturing, injection, extraction, etc.). As an example, equipment 157 and / or 158 may include components, a system, multiple systems, etc., for fracturing, seismic sensing, analysis of seismic data, assessment of one or more fractures, etc.

[0025] exist Figure 1 In the example of , the GUI 120 shows some examples of computational frameworks, including DRILLPLAN, DRILLOPS, PETREL, TECHLOG, PETROMOD, ECLIPSE, PIPESIM, and INTERSECT frameworks (SLB, Houston, TX).

[0026] The DRILLPLAN framework provides digital well construction planning, and includes features for automation of repetitive tasks and validation workflows, enabling rapid generation of drilling programs (eg, digital well plans, etc.) of improved quality while ensuring consistency.

[0027] The DRILLOPS framework can execute digital well plans and ensure plan adherence while providing goal-based automation. The DRILLOPS framework can automatically generate activity plans for individual operations, whether they are monitored and / or controlled on the drilling rig or in town. Automation can utilize data analytics and learning systems to assist and optimize tasks, such as setting the ROP to drill a stand. A preset menu of automatable drilling tasks can be presented, and using data analytics and models, plans can be executed in a manner that achieves specified goals, where, for example, measurements can be used for calibration. The DRILLOPS framework provides the flexibility to dynamically modify and replan activities, for example, based on a real-time assessment of various factors (e.g., equipment, personnel, and supplies). Well construction activities (e.g., tripping, drilling, cementing, etc.) can be continuously monitored and dynamically updated using feedback from operational activities. The DRILLOPS framework can provide various levels of automation based on planning and / or replanning (e.g., via the DRILLPLAN framework), feedback, etc.

[0028] The PETREL framework can be part of the DELFI environment for geoscience and geoengineering, for example, to analyze subsurface data from exploration to production of fluids from reservoirs. The DELFI Cognitive Exploration and Production (E&P) Environment (SLB, Houston, TX), referred to herein as the DELFI environment or DELFI framework, is a secure, cognitive, cloud-based collaborative environment that integrates data and workflows with digital technologies such as artificial intelligence and machine learning.

[0029] The PETREL framework provides components that allow for the optimization of various exploration, development, and production operations. The PETREL framework includes a seismic-to-simulation software component that can output information used to increase reservoir performance, for example, by improving asset team productivity. Using such a framework, various professionals (e.g., geophysicists, geologists, and reservoir engineers) can develop collaborative workflows and integrate operations to streamline processes (e.g., across one or more geological environments, etc.). Such a framework can be considered an application (e.g., executable using one or more devices) as well as a data-driven application (e.g., where data is input for modeling, simulation, etc.).

[0030] The TECHLOG framework can handle and process field and laboratory data from various geological environments (e.g., deepwater exploration, shale, etc.). The TECHLOG framework can structure wellbore data for analysis, planning, etc.

[0031] The PETROMOD framework provides petroleum system modeling capabilities that can combine one or more of seismic, well, and geological information to model the evolution of sedimentary basins. The PETROMOD framework can predict whether and how reservoirs will fill with hydrocarbons, including the source and timing of hydrocarbon generation, migration routes, quantities, and types of hydrocarbons in subsurface or surface conditions.

[0032] The ECLIPSE framework provides numerical solutions for reservoir simulators (eg, as a computational framework) for quickly and accurately predicting the dynamic behavior of various reservoir types and development scenarios.

[0033] The INTERSECT framework provides a high-resolution reservoir simulator for simulating detailed geological features and quantifying uncertainties. For example, by creating accurate production scenarios and integrating precise models of surface facilities and field operations, the INTERSECT framework produces reliable results that can be continuously updated through real-time data exchange (e.g., from one or more types of data acquisition equipment in the field, which can acquire data during one or more types of field operations). The INTERSECT framework also provides completion configurations for complex wells that can be constructed in the field, detailed enhanced oil recovery (EOR) recipes that can be implemented in the field, analysis of the application of steam injection and other thermal EOR techniques for implementation in the field, advanced production control with reservoir connectivity and flexible field management, and the flexibility to script customized solutions for improved modeling and field management control. Like other example frameworks, the INTERSECT framework can be used as part of the DELFI environment, for example, to rapidly simulate multiple concurrent scenarios. For example, workflows can leverage one or more of the DELFI environment's on-demand reservoir simulation features.

[0034] The DELFI environment described above provides various features related to the workflow of subsurface analysis, planning, construction, and production, as shown, for example, in the workspace frame 110. Figure 1 As shown, output from the workspace framework 110 may be used to direct, control, etc., one or more processes in the geological environment 150, and feedback 160 may be received in one or more forms (e.g., acquired data regarding operating conditions, equipment conditions, environmental conditions, etc.) via one or more interfaces.

[0035] As an example, a workflow may proceed to a geological and geophysical ("G&G") service provider, which may generate well trajectories, which may involve executing one or more G&G frameworks (eg, consider the PETREL framework, etc.).

[0036] exist Figure 1In the example of FIG. 1 , visualization features 123 can be implemented via workspace framework 110 , for example, to perform tasks associated with one or more of a subsurface region, planning operations, constructing wells and / or surface fluid networks, and producing from a reservoir.

[0037] As an example, the visualization process can implement one or more of various features that may be suitable for one or more web applications. For example, the template may involve the use of JavaScript Object Notation (JSON) and / or one or more other languages / formats. As an example, the framework may include one or more converters. For example, consider a JSON to Python converter and / or a Python to JSON converter. Such an approach can provide compatibility of devices, frameworks, etc. with respect to one or more instruction sets.

[0038] As an example, a visualization feature can provide visualizations of various earth models, features, and the like in one or more dimensions. As an example, a visualization feature can provide a presentation of information in multiple dimensions, which can optionally include a multi-resolution presentation. In this example, the information being presented can be associated with one or more frameworks and / or one or more data stores. As an example, a visualization feature can include one or more control features for controlling equipment, which can include, for example, field equipment that can perform one or more field operations. As an example, a workflow can utilize one or more frameworks to generate information that can be used to control one or more types of field equipment (e.g., drilling equipment, cable equipment, fracturing equipment, etc.).

[0039] With respect to reservoir models that may be suitable for use by simulators, consider the acquisition of seismic data acquired via reflection seismology, which can be used in geophysics, for example, to estimate the properties of subsurface formations. As an example, reflection seismology can provide seismic data representing waves of elastic energy (e.g., as transmitted by P- and S-waves, in a frequency range of approximately 1 Hz to approximately 100 Hz). Seismic data can be processed and interpreted, for example, to better understand the composition, fluid content, extent, and geometry of subsurface rock. Such interpretations can be used to plan, simulate, execute, etc., one or more operations for producing fluids from a reservoir (e.g., reservoir rock, etc.).

[0040] Field acquisition equipment can be used to acquire seismic data, which can be in the form of traces, where the traces can include values ​​related to time and / or depth organization (e.g., considering 1D, 2D, 3D, or 4D seismic data). For example, consider an acquisition device that acquires digital samples at a rate of approximately one sample every 4 milliseconds. Given the speed of sound in one or more media, the sampling rate can be converted to an approximate distance. For example, the speed of sound in rock can be approximately 5 kilometers per second. Therefore, a sample time interval of approximately 4 milliseconds would correspond to a sample "depth" interval of approximately 10 meters (e.g., assuming a path length from source to boundary and boundary to sensor). As an example, the duration of a trace can be approximately 4 seconds; therefore, for a sampling rate of one sample every 4 milliseconds, such a trace would include approximately 1000 samples, with later samples corresponding to deeper reflecting boundaries. If the 4-second trace duration of the aforementioned example is divided by two (e.g., to account for reflections), the deepest boundary depth can be estimated to be approximately 10 kilometers for a vertically aligned source and sensor (e.g., assuming a sound speed of approximately 5 kilometers per second).

[0041] As an example, a model can be a simulated version of a geological environment. As an example, a simulator can include features for simulating physical phenomena within a geological environment based, at least in part, on one or more models. A simulator, such as a reservoir simulator, can simulate fluid flow within a geological environment based, at least in part, on a model that can be generated via a framework that receives seismic data. A simulator can be a computerized system (e.g., a computing system) that can use one or more processors to execute instructions to solve a system of equations describing physical phenomena subject to various constraints. In such an example, the system of equations can be spatially defined (e.g., numerically discretized) based on a spatial model that includes rock layers, geological body layers, etc., with corresponding locations that can be interpreted based on seismic and / or other data. The spatial model can be a cell-based model, where cells are defined by a grid (e.g., a mesh). Cells in a cell-based model can represent physical regions or volumes within the geological environment, where physical properties (e.g., permeability, fluid properties, etc.) can be assigned to the cells, which can be closely related to one or more physical phenomena (e.g., fluid volume, fluid flow rate, pressure, etc.). A reservoir simulation model can be a cell-based spatial model.

[0042] Simulators can be used to simulate production from real reservoirs, for example, to examine different production scenarios to find the optimal one before production or further production occurs. Reservoir simulators, at least in part, do not provide an exact replica of flow in and production from a reservoir because the reservoir description and the boundary conditions of the equations for flow in porous rocks are typically known with significant uncertainty. Certain types of physical phenomena can occur at relatively small spatial scales compared to the size of the field. A balance can be struck between model scale and computational resources, resulting in model cell sizes on the order of meters, rather than smaller dimensions (e.g., pore detail). Modeling and simulation workflows for multiphase flow in porous media (e.g., reservoir rocks, etc.) can involve generalizing real microscale data from macroscale observations (e.g., seismic and well data) and scaling them up to a manageable scale and problem size. Uncertainties may exist in the input data and solution process, making simulation results somewhat uncertain. A process known as history matching can involve comparing simulation results with actual field data acquired during production of fluids from the field. Information gleaned from history matching can inform adjustments to models, data, and other aspects, which can help improve simulation accuracy.

[0043] As an example, a simulator can utilize various types of constructs, which can be referred to as entities. Entities can include earth entities or geological objects, such as wells, surfaces, reservoirs, etc. Entities can include virtual representations of actual physical entities that can be reconstructed for simulation purposes. Entities can include entities based on data acquired via sensing, observation, etc. (e.g., consider entities based at least in part on seismic data and / or other information). As an example, an entity can be characterized by one or more properties (e.g., a geometric column grid entity of an earth model can be characterized by a porosity property, etc.). Such properties can represent one or more measurements (e.g., acquired data), calculations, etc.

[0044] As an example, a simulator can utilize an object-based software framework that can include entities based on predefined classes to facilitate modeling and simulation. As an example, an object class can encapsulate reusable code and associated data structures. An object class can be used to instantiate object instances for use by programs, scripts, etc. For example, a borehole class can define objects for representing boreholes based on well data. A model of a basin, reservoir, etc. can include one or more boreholes, where a borehole can be used, for example, for measurement, injection, production, etc. As an example, a borehole can be the borehole of a well, which can be a completed well (e.g., for producing resources from a reservoir, for injecting material, etc.).

[0045] Although Figure 1Several simulators are shown in the examples, but one or more other simulators may additionally or alternatively be utilized. For example, consider the VISAGE geomechanical simulator (SLB, Houston, TX) or the PIPESIM network simulator (SLB, Houston, TX). The VISAGE simulator includes a finite element numerical solver that can provide simulation results, such as results on compaction and subsidence within the geological environment, well and completion integrity within the geological environment, cap rock and fault-seal integrity within the geological environment, fracture behavior within the geological environment, thermal recovery within the geological environment, CO2 processing, and so on. The PIPESIM simulator includes a solver that can provide simulation results, such as multiphase flow results (e.g., from the reservoir to the wellhead and beyond), flowlines, and surface facility performance. The PIPESIM simulator can be integrated with the AVOCET production operations framework (SLB, Houston, TX). As an example, one or more reservoirs can be simulated with respect to one or more enhanced recovery techniques (e.g., considering thermal processes such as steam-assisted gravity drainage (SAGD)). As an example, the PIPESIM simulator can be an optimizer that can optimize one or more operational contexts at least in part via simulation of physical phenomena. The MANGROVE simulator (SLB, Houston, TX) provides optimization of stimulation designs (e.g., stimulation treatment operations such as hydraulic fracturing) in a reservoir-centric environment. The MANGROVE framework can combine scientific and experimental work to predict the geomechanical propagation of hydraulic fractures, the reactivation of natural fractures, etc., as well as production predictions within a 3D reservoir model (e.g., production from drainage zones of a reservoir where fluids move to and / or from a well via one or more types of fractures). The MANGROVE framework can provide results related to the heterogeneous interactions between hydraulic fracture networks and natural fracture networks, which can help optimize the number and location of fracture treatment stages (e.g., stimulation treatments), for example to improve perforation efficiency and recovery.

[0046] As an example, the data may include geochemical data. For example, consider data acquired using X-ray fluorescence (XRF), Fourier transform infrared spectroscopy (FTIR), and / or wireline geochemistry techniques.

[0047] As an example, one or more probes may be deployed in a borehole via one or more cables. As an example, the probes may transmit and receive energy, wherein such energy may be analyzed to help determine the mineralogical composition of the rock surrounding the borehole. As an example, nuclear magnetic resonance may be performed (e.g., via a cable, a downhole NMR probe, etc.), for example, to obtain data regarding the nuclear magnetic properties of elements (e.g., hydrogen, carbon, phosphorus, etc.) in the formation.

[0048] As an example, lithologic scanning technology can be used to acquire and analyze data. For example, consider the LITHO SCANNER technology (SLB, Houston, Texas). As an example, the LITHO SCANNER tool can be or include a gamma ray spectroscopy tool.

[0049] As an example, a tool can be positioned to acquire information from a portion of a borehole. Analysis of this information can reveal voids, dissolution planes (e.g., dissolution along bedding planes), stress-related features, dip events, and the like. As an example, a tool can acquire information that can help characterize a fractured reservoir, where the fractures can be natural and / or artificial (e.g., hydraulic fractures). This information can assist with well completion, stimulation treatments, and the like. As an example, a framework such as the aforementioned TECHLOG framework can be used to analyze the information acquired by the tool.

[0050] As an example, a workflow may utilize one or more types of data for one or more processes (e.g., formation modeling, basin modeling, completion design, drilling, production, injection, etc.). As an example, one or more tools may provide data that may be used in one or more workflows that may implement one or more frameworks (e.g., PETREL, TECHLOG, PETROMOD, ECLIPSE, etc.).

[0051] exist Figure 1 In the example of , drilling can be performed in a geological environment 150, for example, to access a reservoir 151, which can be accessed from land or sea. Figure 1 In this embodiment, downhole equipment 154 may be, for example, part of a bottom hole assembly (BHA). The BHA can be used for drilling a well. Downhole equipment 154 can transmit information to equipment at the surface and receive instructions and information from equipment at the surface. During the well construction process, various operations (such as cementing, wireline evaluation, testing, etc.) may be performed. In such an embodiment, data collected by tools and sensors and used for purposes such as reservoir characterization may be collected and transmitted.

[0052] A well may include a substantially horizontal portion (e.g., a lateral portion) that may intersect one or more fractures. For example, a well in a shale formation may intersect natural fractures, artificial fractures (e.g., hydraulic fractures), or a combination thereof. Such a well may be constructed using directional drilling techniques as described herein. However, these same techniques may be used in conjunction with other types of directional wells (e.g., deviated wells, S-shaped wells, deep deviated wells, etc.) and are not limited to horizontal wells.

[0053] Figure 2An example of a wellsite system 200 is shown (e.g., at a wellsite that may be onshore or offshore). As shown, the wellsite system 200 may include a mud tank 201 for holding mud and other materials (e.g., where the mud may be a drilling fluid), a suction line 203 serving as an inlet for a mud pump 204 for pumping mud from the mud tank 201 to a vibrating hose 206, a winch 207 for winding one or more drilling lines 212, a riser 208 for receiving mud from the vibrating hose 206, and a kelly hose 206 for receiving mud from the riser 208. Pipe 209, gooseneck 210, traveling block 211, crown block 213 for carrying traveling block 211 via one or more drilling pipelines 212, derrick 214, kelly 218 or top drive 240, kelly bushing 219, rotary table 220, drill floor 221, bell joint 222, one or more blowout preventers (BOPs) 223, drill string 225, drill bit 226, casing head 227, and flow tube 228 for carrying mud and other materials to, for example, mud tank 201.

[0054] exist Figure 2 In the exemplary system of FIG. , a borehole 232 is formed in a subsurface formation 230 by rotary drilling; note that various exemplary embodiments may also utilize one or more directional drilling techniques, equipment, etc.

[0055] like Figure 2 As shown in the example of FIG, a drill string 225 is suspended within a borehole 232 and has a drill string assembly 250 including a drill bit 226 at its lower end. As an example, the drill string assembly 250 may be a bottom hole assembly (BHA).

[0056] Wellsite system 200 can provide for the operation of drill string 225 and other operations. As shown, wellsite system 200 includes traveling block 211 and derrick 214 positioned above borehole 232. As described above, wellsite system 200 can include rotary table 220, with drill string 225 passing through an opening in rotary table 220.

[0057] like Figure 2As shown in the example of , the wellsite system 200 may include a kelly 218 and associated components, or a top drive 240 and associated components. Regarding the kelly example, the kelly 218 may be a square or hexagonal metal / alloy rod with a hole drilled therein to serve as a mud flow path. The kelly 218 may be used to transmit rotational motion from a rotary table 220 to a drill string 225 via a kelly drive bushing 219, while allowing the drill string 225 to be lowered or raised during rotation. The kelly 218 may pass through the kelly drive bushing 219, which may be driven by the rotary table 220. As an example, the rotary table 220 may include a main bushing operably coupled to the kelly drive bushing 219, such that rotation of the rotary table 220 may rotate the kelly drive bushing 219 and, thereby, the kelly 218. The kelly drive bushing 219 may include an interior profile that matches the exterior profile of the kelly 218 (eg, square, hexagonal, etc.); however, with slightly larger dimensions so that the kelly 218 can move freely up and down inside the kelly driver bushing 219 .

[0058] Regarding the top drive example, top drive 240 can provide the functions previously performed by a kelly and rotary table. Top drive 240 can rotate drill string 225. For example, top drive 240 can include one or more motors (e.g., electric and / or hydraulic motors) connected via an appropriate transmission to a short section of tubing called a casing shaft, which in turn can be threaded into a saver sub or the drill string 225 itself. Top drive 240 can be suspended from traveling block 211 so that the rotating mechanism can freely travel up and down the derrick 214. For example, top drive 240 can allow drilling to be performed using more joint stands than with a kelly / rotary table approach.

[0059] exist Figure 2 In the example of , mud tank 201 can hold mud, which can be one or more types of drilling fluids. As an example, a wellbore can be drilled to produce fluids, inject fluids, or both (e.g., hydrocarbons, minerals, water, etc.).

[0060] exist Figure 2In the example shown, drill string 225 (e.g., including one or more downhole tools) may be composed of a series of pipes threaded together to form a long tube, with drill bit 226 located at the lower end. As drill string 225 is advanced into a wellbore for drilling, mud may be pumped from mud tank 201 (e.g., or other source) via pump 204, via pipelines 206, 208, and 209, to ports of kelly 218, or, for example, to ports of top drive 240, at some point prior to or concurrently with drilling. The mud may then flow through a channel (e.g., or multiple channels) in drill string 225 and out of a port located on drill bit 226 (see, e.g., directional arrows). Once the mud exits drill string 225 via the port in drill bit 226, it may then circulate upward through the annular region between the outer surface of drill string 225 and the surrounding wall (e.g., an open borehole, casing, etc.), as indicated by the directional arrows. In this manner, the mud lubricates the drill bit 226 and carries heat energy (e.g., friction or other energy) and formation cuttings to the surface, where the mud (e.g., and cuttings) can be returned to the mud tank 201, e.g., for recirculation (e.g., by processing to remove cuttings, etc.).

[0061] The mud pumped into drill string 225 by pump 204 can form a mudcake after exiting drill string 225, lining the wellbore. Among other functions, this mudcake can reduce friction between drill string 225 and the surrounding walls (e.g., borehole, casing, etc.). This reduced friction can facilitate advancement or retraction of drill string 225. During drilling operations, the entire drill string 225 can be pulled out of the wellbore and optionally replaced, for example, with a new or sharpened drill bit, a smaller diameter drill string, etc. As described above, the act of pulling the drill string out of the hole or replacing it in the hole is called tripping. Depending on the direction of tripping, tripping can be referred to as tripping up or tripping out, or tripping down or tripping in.

[0062] As an example, consider drilling down, where, when the drill bit 226 of the drill string 225 reaches the bottom of the wellbore, the pumping of mud begins to lubricate the drill bit 226 for drilling to enlarge the wellbore. As described above, the mud can be pumped by the pump 204 into the channel of the drill string 225, and when filling the channel, the mud can be used as a transmission medium to transmit energy, for example, energy that can encode information as in mud pulse telemetry.

[0063] As an example, a mud pulse telemetry device may include a downhole device configured to implement pressure changes in the mud to generate one or more acoustic waves that can modulate information. In such an example, information from the downhole device (e.g., one or more modules of the drill string 225) can be transmitted uphole to the uphole device, which can relay such information to other equipment for processing, control, etc.

[0064] As an example, the telemetry device may operate via energy transmission via the drill string 225 itself. For example, consider a signal generator that transmits a coded energy signal to the drill string 225 and a repeater that can receive this energy and repeat it to further transmit the coded energy signal (e.g., information, etc.).

[0065] As an example, the drill string 225 may be equipped with a telemetry device 252, the telemetry device 252 including a rotatable drive shaft, a turbine wheel mechanically coupled to the drive shaft so that mud can cause the turbine wheel to rotate, a modulation rotor mechanically coupled to the drive shaft so that rotation of the turbine wheel causes the modulation rotor to rotate, a modulation stator mounted adjacent to or proximate to the modulation rotor so that rotation of the modulation rotor relative to the modulation stator generates pressure pulses in the mud, and a controllable brake for selectively braking rotation of the modulation rotor to modulate the pressure pulses. In such an example, an alternator may be coupled to the drive shaft, wherein the alternator includes at least one stator winding electrically coupled to a control circuit for selectively short-circuiting the at least one stator winding to electromagnetically brake the alternator, thereby selectively braking rotation of the modulator rotor to modulate the pressure pulses in the mud.

[0066] exist Figure 2 In the example of , the uphole control and / or data acquisition system 262 may include circuitry for sensing pressure pulses generated by the telemetry device 252 and, for example, transmitting the sensed pressure pulses or information derived therefrom for processing, control, etc.

[0067] The assembly 250 of the illustrated example includes a logging while drilling (LWD) module 254, a measurement while drilling (MWD) module 256, optional modules 258, a rotary steering system (RSS) and / or motor 260, and a drill bit 226. Such components or modules may be referred to as tools, where a drill string may include multiple tools.

[0068] With respect to RSS, it relates to techniques used for directional drilling. Directional drilling involves drilling into the earth to form a deviated borehole such that the trajectory of the borehole is not vertical; rather, the trajectory deviates from vertical along one or more portions of the borehole. As an example, consider a target located at a lateral distance from a surface location where a drilling rig may be parked. In such an example, the borehole may begin at the vertical portion and then deviate from the vertical portion so that the borehole is aimed at the target and ultimately reaches the target. Directional drilling may be practiced where the target may not be accessible from a vertical position at the earth's surface, where there are materials in the earth that may impede drilling or be otherwise harmful (e.g., consider salt domes, etc.), where formations extend laterally (e.g., consider relatively thin but laterally extending reservoirs), where multiple boreholes are to be drilled from a single surface borehole, where relief wells are required, etc.

[0069] One method of directional drilling involves a mud motor; however, mud motors can present challenges depending on factors such as rate of penetration (ROP), weight transfer to the drill bit due to friction (e.g., weight on bit, WOB), and so on. The mud motor can be a positive displacement motor (PDM), which operates to drive the drill bit (e.g., during directional drilling, etc.). The PDM operates when drilling fluid is pumped through it, where the PDM converts the hydraulic power of the drilling fluid into mechanical power to rotate the drill bit.

[0070] For example, a mud motor (e.g., a PDM) can operate in different modes, including rotary and sliding modes. Sliding mode involves drilling using the mud motor to rotate the drill bit downhole without rotating the drill string from the surface. This operation is possible when the BHA is equipped with a bent sub, a bent housing mud motor, or both for directional drilling. Sliding mode can be used to establish, control, or adjust the hole angle. In directional drilling, the drill bit's pointing direction is achieved using a bent sub, which can have a relatively small angular offset relative to the axis of the drill string, and a measuring device used to determine the direction of the offset. Without rotating the drill string, the drill bit can be rotated as mud flows through the mud motor to drill in the direction it is pointed. With a steerable motor, once the desired wellbore direction is achieved, the entire drill string can be rotated to drill in a straight line rather than at an angle. By controlling the amount of hole drilled in sliding mode versus rotary mode, the wellbore trajectory can be controlled with considerable precision.

[0071] As an example, the PDM may be operated in a combined rotation mode, wherein surface equipment is used to rotate the drill bit of the drill string by rotating the entire drill string (e.g., a rotary table, a top drive, etc.), and wherein drilling fluid is used to rotate the drill bit of the drill string. In such an example, the surface RPM (SRPM) may be determined using the surface equipment, while the downhole RPM of the mud motor may be determined using various factors related to the flow of drilling fluid, the type of mud motor, etc. As an example, in the combined rotation mode, assuming that the SRPM and the mud motor RPM are in the same direction, the drill bit RPM may be determined or estimated as the sum of the SRPM and the mud motor RPM.

[0072] As an example, when the drill string is not rotating from the surface, the PDM mud motor can be operated in a so-called sliding mode. In such an example, the drill bit RPM can be determined or estimated based on the RPM of the mud motor.

[0073] An RSS can be deployed when drilling directionally (e.g., in a deviated, horizontal, or extended-reach wellbore) while continuously rotating the drill bit. The RSS can be designed to minimize interaction with the borehole wall, which can help maintain borehole quality. The RSS can be designed to apply a relatively consistent lateral force similar to a stabilizer that rotates with the drill string or orients the drill bit in the desired direction while continuously rotating at the same RPM as the drill string.

[0074] LWD module 254 can be housed in a suitable type of drill collar and can include one or more selected types of logging tools. It will also be understood that more than one LWD and / or MWD module can be employed, for example, as represented by module 256 of drill string assembly 250. Where reference is made to the location of an LWD module, this may refer to a module at the location of LWD module 254, module 256, etc., by way of example. The LWD module may include capabilities for measuring, processing, and storing information, as well as capabilities for communicating with surface equipment. In the example shown, LWD module 254 may include a seismic measurement device.

[0075] The MWD module 256 can be housed in a suitable type of drill collar and can include one or more devices for measuring characteristics of the drill string 225 and the drill bit 226. By way of example, the MWD tool 254 can include a device for generating electrical power, e.g., to power various components of the drill string 225. By way of example, the MWD tool 254 can include telemetry equipment 252, e.g., where a turbine wheel can generate electrical power from the flow of mud; it will be appreciated that other power sources and / or battery systems can be employed to power the various components. By way of example, the MWD module 256 can include one or more of the following types of measurement devices: a weight-on-bit measurement device, a torque measurement device, a vibration measurement device, a shock measurement device, a stick-slip measurement device, a direction measurement device, and an inclination measurement device.

[0076] Figure 2 Also shown are some examples of the types of holes that can be drilled. For example, consider an angled hole 272, an S-shaped hole 274, a deep angled hole 276, and a horizontal hole 278.

[0077] As an example, the drilling operation may include directional drilling, where, for example, at least a portion of the well includes a curved axis. For example, consider a radius defining a curvature, where the inclination relative to the vertical may vary until reaching an angle between about 30 degrees and about 60 degrees, or, for example, up to about 90 degrees or possibly greater than about 90 degrees.

[0078] As an example, a directional well can include several shapes, each of which can be designed to meet specific operational requirements. As an example, when information is relayed to a drilling engineer, a drilling process can be performed based on the information. As an example, the inclination and / or direction can be modified based on the information received during the drilling process.

[0079] As an example, deviation of the bore may be achieved in part through the use of downhole motors and / or turbines.With respect to motors, for example, the drill string may include a positive displacement motor (PDM).

[0080] As an example, the system may be a steerable system and include equipment for performing methods such as geosteering. As described above, the steerable system may be or include an RSS. As an example, the steerable system may include a PDM or turbine on the lower portion of the drill string, with a bent sub mounted directly above the drill bit. As an example, above the PDM, MWD equipment and / or LWD equipment may be mounted, with the MWD equipment providing real-time or near-real-time data of interest (e.g., inclination, direction, pressure, temperature, actual weight on bit, torque stress, etc.). With respect to the latter, the LWD equipment may transmit various types of data of interest to the surface, including, for example, geological data (e.g., gamma-ray logs, resistivity, density, and sonic logs, etc.).

[0081] The coupling of sensors that provide real-time or near real-time information about the course of the wellbore trajectory with, for example, one or more well logs that geologically characterize the formation can enable geosteering methods. Such methods can include navigating a subsurface environment, for example, to follow a desired route to reach one or more desired targets.

[0082] As an example, a drill string may include an azimuthal density neutron (ADN) tool for measuring density and porosity; an MWD tool for measuring inclination, azimuth, and shock; a compensated dual resistivity (CDR) tool for measuring resistivity and gamma-ray related phenomena; one or more variable gauge stabilizers; one or more bend joints; and a geosteering tool, which may include a motor and, optionally, equipment for measuring and / or responding to one or more of inclination, resistivity, and gamma-ray related phenomena.

[0083] As an example, geosteering may include intentionally directionally controlling a wellbore based on results of downhole geological logging measurements in a manner intended to maintain the directional wellbore within a desired area, zone (e.g., a pay zone), etc. As an example, geosteering may include directing the wellbore to maintain the wellbore within a specific section of a reservoir, e.g., to minimize gas and / or water breakthrough, and e.g., to maximize the economic yield from a well including the wellbore.

[0084] Reference again Figure 2, the wellsite system 200 may include one or more sensors 264 operably coupled to the control and / or data acquisition system 262. As an example, the one or more sensors may be at a surface location. As an example, the one or more sensors may be at a downhole location. As an example, the one or more sensors may be located at one or more remote locations that are not within a distance of approximately one hundred meters from the wellsite system 200. As an example, the one or more sensors may be at an offset wellsite in a common field (e.g., an oil and / or gas field) between the wellsite system 200 and the offset wellsite.

[0085] As an example, one or more of sensors 264 may be provided for tracking pipes, tracking movement of at least a portion of a drill string, etc.

[0086] As an example, system 200 may include one or more sensors 266 that can sense signals and / or transmit signals to a fluid conduit, such as a drilling fluid conduit (e.g., a drilling mud conduit). For example, in system 200, one or more sensors 266 may be operably connected to a portion of riser 208 through which mud flows. As an example, a downhole tool may generate pulses that travel through the mud and are sensed by one or more of the one or more sensors 266. In such an example, the downhole tool may include associated circuitry, such as encoding circuitry that can encode the signals, e.g., to reduce the need for transmission. As an example, the circuitry at the surface may include decoding circuitry to decode the encoded information transmitted at least in part via mud pulse telemetry. As an example, the circuitry at the surface may include encoder circuitry and / or decoder circuitry, and the downhole circuitry may include encoder circuitry and / or decoder circuitry. As an example, system 200 may include a transmitter that can generate signals that can be transmitted downhole via mud (e.g., drilling fluid) as a transmission medium.

[0087] As an example, one or more portions of the drill string may be stuck. The term stuck may refer to varying degrees of inability to move or remove one or more portions of the drill string from the borehole. As an example, in a stuck state, the pipe may be rotated or returned to the borehole, or, for example, in a stuck state, the drill string may not be able to be moved axially in the borehole, although some rotation may be possible. As an example, in a stuck state, at least a portion of the drill string may not be able to be moved axially and rotationally.

[0088] The term "stuck pipe" can refer to a portion of the drill string that is unable to rotate or move axially. For example, a condition known as "differential pressure sticking" can occur when the drill string is unable to move (e.g., rotate or reciprocate) along the axis of the hole. Differential pressure sticking can occur when high contact forces, caused by low reservoir pressure, high wellbore pressure, or both, are applied to a sufficiently large area of ​​the drill string. Differential pressure sticking can have both time and financial costs.

[0089] As an example, the sticking force can be the product of the pressure differential between the wellbore and the reservoir and the area over which the pressure differential acts. This means that a relatively low pressure differential (Δp) applied over a large working area can be just as effective at sticking pipe as a high pressure differential applied over a small area.

[0090] As an example, a condition referred to as "mechanical stuck pipe" may be a condition that occurs when movement of the drill string is restricted or prevented by a mechanism other than differential pressure sticking. Mechanical stuck pipe may be caused, for example, by one or more of trash in the hole, abnormal wellbore geometry, cement, key seat, or accumulation of drill cuttings in the annulus.

[0091] Figure 3 A schematic diagram of a computing or processor system 300 according to one embodiment is shown. Processor system 300 may include one or more processors 302 having different core configurations (including multiple cores) and clock frequencies. One or more processors 302 may be operable to execute instructions, application logic, etc. It should be understood that these functions may be provided by multiple processors or multiple cores on a single chip operating in parallel and / or communicatively linked together. In at least one embodiment, one or more processors 302 may be or include one or more GPUs.

[0092] The processor system 300 may also include a memory system, which may be or include one or more memory devices and / or computer-readable media 304 of varying physical size, accessibility, storage capacity, etc., such as a flash drive, hard drive, magnetic disk, random access memory, etc., for storing data, such as images, files, and program instructions executed by the processor 302. In one embodiment, the computer-readable media 304 may store instructions that, when executed by the processor 302, are configured to cause the processor system 300 to perform operations. For example, execution of such instructions may cause the processor system 300 to implement one or more portions and / or embodiments of the methods described above.

[0093] The processor system 300 may also include one or more network interfaces 306. The network interfaces 306 may include any hardware, applications, and / or other software. Thus, the network interfaces 306 may include an Ethernet adapter, a wireless transceiver, a PCI interface, and / or a serial network component for communicating over a wired or wireless medium using protocols such as Ethernet, wireless Ethernet, and the like.

[0094] As an example, processor system 300 may be a mobile device that includes one or more network interfaces for information communication. For example, the mobile device may include a wireless network interface (e.g., operable via one or more IEEE 802.11 protocols, ETSI GSM, Bluetooth, satellite, etc.). As an example, the 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 card 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, the mobile device may be configured as a cellular phone, a tablet computer, etc. As an example, the method may be implemented using (e.g., in whole or in part) a mobile device. As an example, the system may include one or more mobile devices.

[0095] The processor system 300 may also include one or more peripheral interfaces 308 for communicating with a display, projector, keyboard, mouse, touchpad, sensors, other types of input and / or output peripherals, and the like. In some embodiments, the components of the processor system 300 need not be enclosed within a single housing or even positioned in close proximity to one another, but in other embodiments, the components and / or other components may be provided within a single housing. By way of example, the system may be a distributed environment, such as a so-called "cloud" environment, in which various devices, components, and the like interact for purposes of data storage, communication, computation, and the like. By way of example, the method may be implemented in a distributed environment (e.g., in whole or in part as a cloud-based service).

[0096] As an example, information can be input from a display (e.g., a touchscreen), output to a display, or both. As an example, information can be output to a projector, laser device, printer, etc. so that the information can be viewed. As an example, information can be output stereoscopically or holographically. Regarding printers, consider 2D or 3D printers. As an example, a 3D printer can include one or more substances that can be output to construct a 3D object. For example, data can be provided to a 3D printer to construct a 3D representation of a subsurface formation. As an example, layers can be constructed in 3D (e.g., horizons, etc.), geological bodies can be constructed in 3D, etc. As an example, pores, fractures, etc. can be constructed in 3D (e.g., as positive structures, as negative structures, etc.).

[0097] The memory device 304 can be physically or logically arranged or configured to store data on one or more storage devices 310. The storage device 310 can include one or more file systems or databases in any suitable format. The storage device 310 can also include one or more software programs 312, which can contain interpretable and / or executable instructions for performing one or more of the disclosed processes (e.g., processor-executable instructions that can be stored in the memory 304 and executed to instruct the system 300 to perform one or more actions). Upon request by the processor 302, one or more of the software programs 312, or portions thereof, can be loaded from the storage device 310 to the memory device 304 for execution by the processor 302.

[0098] Those skilled in the art will appreciate that the above components are merely one example of a hardware configuration, as the processor system 300 may include any type of hardware components for executing the disclosed embodiments, including any accompanying firmware or software. The processor system 300 may also be implemented partially or entirely by electronic circuit components or processors, such as application specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs).

[0099] Processor system 300 can be configured to receive a directional drilling plan 320 (e.g., and / or generate a directional drilling plan). As described above, a well plan is a description of a proposed wellbore to be used by a drilling crew when drilling a well. A well plan typically includes information regarding the shape, orientation, depth, completion, and evaluation of the well, as well as information regarding the equipment to be used, actions to be taken at different points in the well construction process, and other information the team planning the well deems relevant / helpful to the team drilling the well. A directional drilling plan may also include information regarding how to steer and manage the well's direction.

[0100] The processor system 300 may be configured to receive drilling data 322. The drilling data 322 may include data collected by one or more sensors associated with surface equipment or downhole equipment. The drilling data 322 may include information such as data related to the position of the BHA (such as survey data or continuous position data), drilling parameters (such as weight on bit (WOB), rate of penetration (ROP), torque, or other parameters), text information entered by individuals working at the wellsite, or other data collected during the construction of the well.

[0101] In one embodiment, processor system 300 is part of a rig control system (RCS) of a drilling rig (e.g., including downhole equipment operatively coupled to the drilling rig). In another embodiment, processor system 300 is a separately installed computing unit that includes a display installed at the drilling rig site and receives data from the RCS. In such an embodiment, the software on processor system 300 can be installed on the computing unit, brought to the wellsite, and installed and communicatively coupled to the drilling rig control system in preparation for constructing a well or a portion thereof.

[0102] In another embodiment, the processor system 300 may be located at a location remote from the well site and receive the drilling data 322 via a communication medium using a protocol such as the Well Site Information Transfer Specification or Standard (WITS) and Markup Language (WITSML). In such an embodiment, the software on the processor system 300 may be a web-native application accessed by a user using a web browser. In such an embodiment, the processor system 300 may be remote from the well site where the well is being constructed, and the user may be at the well site or at a location remote from the well site.

[0103] Well plan 320 typically includes information about the direction and shape of the well to be drilled. It can also include information about the parameters and tools used to achieve the desired shape and location. However, while drilling, the actual trajectory may deviate from the plan, or unexpected circumstances may be encountered. In this and other situations, the plan may need to be adjusted to account for changing conditions and circumstances. For example, consider methods that may require replanning to generate a revised drilling plan.

[0104] In one embodiment, a system includes a drilling planning component for monitoring and updating a drilling plan, where the drilling plan may be in digital format, e.g., as digital data structured and stored in a memory of a computing device, computing system, or the like. For example, consider a controller including a memory capable of storing a drilling plan as one or more digital files. As the team conducts a survey or otherwise determines the location of a well, the drilling planning component may derive a work plan. In one embodiment, the work plan is effectively a spatial trajectory in multiple dimensions that constructs a path from the current drill head position (e.g., the bottom of the hole) to the next position, which may be referred to as a target, which may be an intermediate target or a final target. The path is constructed based on a variety of considerations. These may include, but are not limited to: the target; the allowable deviation from the original plan in terms of position and / or angular deviation; the maximum dogleg capability of the steering assembly; constraints set by the user based on initial preferences; allowed tortuosity, risk metrics, hole quality, confidence level, etc.; and others.

[0105] Generating a work plan can involve generating a set of trajectory candidates that meet a number of different specified criteria and evaluating the candidates based on trajectory context, different characteristics, constraint violations, and ranking. Trajectories can be ranked according to different optimization objectives. The user can cause the system to present one or more available candidates for selection.

[0106] As an example, the path of a wellbore, or its trajectory, can be determined by obtaining direction and inclination (D&I) measurements at various points along the wellbore, which can be referred to as survey measurements or survey information. With respect to survey, the position of a wellbore (e.g., a borehole, etc.) can be referenced to some vertical and / or horizontal reference (e.g., a wellhead location and an elevation reference). The position can be obtained using one or more inertial measurement techniques. With respect to azimuth, it can be considered as, for example, the directional angular heading relative to a reference direction (such as north) at the measurement location. With respect to inclination, it can be considered as the angular deviation of the borehole relative to a vertical direction (e.g., a reference to gravity). With respect to measured depth, it can be considered as the distance measured from a surface location along the wellbore (e.g., a borehole, etc.). The measured depth can include the driller's depth, and it can also include a depth correction algorithm that accounts for elastic stretching and compression of the drill string along its length.

[0107] A directional wellbore can be drilled through a formation along a selected trajectory. Various factors can combine to unpredictably affect the trajectory of the wellbore. It is desirable to accurately measure the wellbore trajectory in order to guide the wellbore to its geological and / or locational target. Therefore, it is desirable to measure the wellbore's inclination, azimuth, and depth during wellbore operations to assess whether the selected trajectory is being maintained.

[0108] The drilling trajectory of a wellbore can be estimated by acquiring a wellbore or a directional survey, which may be referred to herein as a survey other than an extended surface-based seismic survey that acquires a seismic data cube of an Earth volume. A wellbore survey may consist of a series or set of survey stations. A survey station may be generated by taking measurements at a single location in a wellbore (e.g., a borehole, etc.) to estimate the position and / or orientation of the wellbore. The act of performing these measurements and generating survey data may be referred to as performing a survey or surveying the wellbore or borehole.

[0109] A wellbore survey can be performed using downhole survey instruments (e.g., drill string equipment). Such instruments may include, for example, one or more orthogonal accelerometers, magnetometers, and / or gyroscopes. The survey instruments can be used to measure the direction and magnitude of the local gravity field, magnetic field, and / or Earth's spin rate vector (e.g., collectively referred to as the Earth vector), respectively. These various measurements can correspond to the instrument's position and orientation within the wellbore relative to the Earth vector. By way of example, the wellbore's position, inclination, and / or azimuth can be estimated based on the instrument measurements.

[0110] One or more survey stations can be generated using either discrete or continuous measurement modes. Discrete or static wellbore surveys can be performed by creating survey stations along the wellbore when drilling is stopped or interrupted, for example, to add additional drill pipe joints or pipe stands to the drill string at the surface. Continuous wellbore surveys can involve various measurements of earth vectors and / or angular velocity of the downhole tool obtained for each wellbore section using one or more survey instruments. Successive measurements of these vectors may be separated by fractions of a meter during the drilling operation, and depending on the rate of change of the vectors in the borehole, these measurements can be considered continuous or they can be considered discrete.

[0111] Figure 4 and Figure 5 One embodiment of a system 400 for work plan generation (e.g., a work plan generator (WPG)) is shown. As shown, the system 400 may include an application layer 410, a data service 412, a state estimation component 414, a trajectory generator component 416, a user runtime parameter component 418, a drilling command scheduler component 420, a sequencing system component 422, an activity plan management component 424, and a state manager component 426.

[0112] In one embodiment, the system 400 includes a state estimation component 414 for inferring the state of the system using incoming data from the data service 412. As an example, the state manager component 426 can be configured to process multiple parallel states for determining the effects of certain actions that a user (e.g., or a machine) might want to take or might take, and infer states from these "what-if" contexts. Figure 4 In the example of FIG, the track manager component 416 can interface with multiple track creation sources / states and manage multiple track candidates.

[0113] Regarding the drilling command scheduler component 420, it can generate a hardware-specific drilling command sequence for each trajectory (e.g., a command schedule for each trajectory). Regarding the sequencing system component 422, it can evaluate constraint violations and feature classification candidates according to user-defined optimization goals. Figure 4 In the example of , the activity plan manager component 424 can monitor ongoing actions and make / suggest corrections, trigger rescheduling of actions, and request user intervention.

[0114] In one embodiment, Figure 5 The trajectory evaluation and ranking method shown can evaluate the constraint violations of the generated candidates; note that it can also evaluate the cost function of multiple candidate characteristics. In some embodiments, the user can prioritize and / or weight the parameters to be optimized. The component can also generate a prioritized list of candidates for the user to select from.

[0115] In one embodiment, the trajectory evaluation and ranking solver takes as input the trajectory and the drilling command schedule from the command scheduler; note that it may also receive constraint configurations, constraint violation penalties, and candidate characteristic weights. In one embodiment, the output may be a prioritized list of candidates.

[0116] Figure 6 An example of a system 600 is shown, including an off-site device 601 (e.g., remote) and a field device 602 (e.g., local). As shown, the off-site device 601 may include a drilling operations framework 610, a drilling planning framework 620, and a database 630, while the field device 602 may include a controller 640 that can receive real-time data and output recommendations, such as control instructions, to control the field device. In such an example, the drilling operations framework 610 may provide steering tables, execution parameters, and the like, while the drilling planning framework 620 may provide steering response and statistical evaluations. As shown, the controller 640 may output information to the drilling operations framework 610 and receive information from the drilling planning framework 620. The system 600 may include a plan generation feature for real-time plan generation during the execution phase of drilling operations and / or plan generation during the planning phase. The system 600 may be used for one or more types of drilling (e.g., rotary, mud motor, RSS, ABSS, etc.). The system 600 may operate in cycles that may include at least one real-time cycle for providing control of equipment to perform drilling operations.

[0117] A system such as system 600 may utilize various functions and penalties to generate a plan that may provide for single or multiple target targeting. As explained, a plan may be generated that is intended to provide for drilling operations that simultaneously target multiple targets. As an example, system 600 may include Figure 4 and Figure 5 One or more features of system 400, one or more other systems described herein, and the like.

[0118] Figure 7 An example of a wellsite system 700 is shown, specifically, Figure 7 Wellsite system 700 is shown in approximate side view and approximate plan view, as well as a block diagram of system 770 .

[0119] exist Figure 7In the example of FIG. 7 , wellsite system 700 may include cabin 710 , rotary table 722 , drawworks 724 , mast 726 (e.g., optionally carrying a top drive, etc.), mud tank 730 (e.g., having one or more pumps, one or more vibrators, etc.), one or more pump buildings 740 , boiler building 742 , HPU building 744 (e.g., having a rig fuel tank, etc.), combination building 748 (e.g., having one or more generators, etc.), tubing barrel 762 , catwalk 764 , flare 768 , etc. Such equipment may include one or more associated functions and / or one or more associated operational risks, which may be risks to time, resources, and / or personnel.

[0120] like Figure 7 As shown in the example of , the wellsite system 700 may include a system 770 that includes one or more processors 772, a memory 774 operably coupled to at least one of the one or more processors 772, instructions 776 that may be stored, for example, in the memory 774, and one or more interfaces 778. As an example, the system 770 may include one or more processor-readable media that include processor-executable instructions that are executable by at least one of the one or more processors 772 to cause the system 770 to control one or more aspects of the wellsite system 700. In such an example, the memory 774 may be or include one or more processor-readable media, wherein the processor-executable instructions may be or include instructions. As an example, the processor-readable medium may be a computer-readable storage medium that is not a signal and is not a carrier wave.

[0121] Figure 7 Also shown is a battery 780 that can be operably coupled to the system 770, for example, to power the system 770. As an example, the battery 780 can be a backup battery that operates when another power source is unavailable to power the system 770. As an example, the battery 780 can be operably coupled to a network, which can be a cloud network. As an example, the battery 780 can include smart battery circuitry and can be operably coupled to one or more devices via an SMBus or other type of bus.

[0122] exist Figure 7 In the example of FIG, service 790 is shown as being available, for example, via a cloud platform. Such services may include data services 792, query services 794, and drilling services 796. As an example, service 790 may be such as Figure 6 The services 790 may include one or more services for directional drilling, such as system 600, another system described herein, etc. As an example, the services 790 may include one or more services for directional drilling, which may include, for example, one or more steering trending services (e.g., considering a computing framework that may provide one or more services for utilizing survey information to estimate one or more steering response parameters, etc.).

[0123] As an example, the system 770 may be used to generate one or more rate-of-penetration drilling parameter values ​​that may be used, for example, to control one or more drilling operations.

[0124] As an example, a method may include automating the operation of one or more types of downhole tools. For example, consider automating the operation of one or more of a mud motor, a rotary steering system (RSS), and an ABS. As an example, a method may include automating the operation of one or more of a mud motor, a rotary steering system (RSS), and an ABS. Figure 2 One or more features of the system 200 may be implemented to implement one or more of these types of devices, systems, and the like.

[0125] As an example, an ABSS may include an actuator with dropout range, hold-in-tilt and azimuth (HIA), and dual downlink capability. As an example, an ABSS may include an onboard near-bit sensor that can acquire continuous six-axis inclination and azimuth measurements with a 6-ft range, and optionally natural gamma ray and azimuth imagery with a 9-ft range. As an example, an ABSS may include one or more features of one or more of the NEOSEER family of ABSS (SLB, Houston, TX).

[0126] As an example, the framework can provide information from plans with deviation analysis and adjust hybrid models for real-time execution; consider various downhole automation possibilities at a given time to optimize the execution of recommended work trajectories, thereby utilizing tool capabilities and minimizing unnecessary surface actions; consider real-time data information and derived tool health and tool state estimates at a given point in time to optimize the current ongoing recommendations, and recommend real-time corrections to address deviations as they occur. This approach can involve different calculations and actions based on drilling constraints and drilling context, for example, outputting the best recommendation for the fastest path with minimized risk.

[0127] Regarding the best path recommendation, the derivation can be done via WPG, for example, as described in relation to Figure 4 and Figure 5 As explained, it may provide a single-target approach and / or a multi-target approach, and, for example, it may take into account various factors, which may include energy, emission, etc.

[0128] Figure 8An example of a workflow 800 including data acquisition, state calculation, work plan generation, sequencing and command scheduling, and user selection is shown. As shown, such a workflow can be implemented during drilling operations at a site. For example, the acquired data can be real-time data, and the state calculation can calculate the state of the BHA or other tool located in the wellbore. Regarding work plan generation, it can answer the question of where to proceed, which can be based on the current plan, constraints, and context. As explained, sequencing and command scheduling can be performed, where commands can be suitable for automatic and / or manual execution. Recommendations from such a workflow can be presented to one or more displays, where, for example, one or more GUIs can provide user interaction.

[0129] As explained, a work plan can be or include a trajectory for constructing a path from the current drill head position (bottom of hole (HB) point) to the next target. This path construction can be performed based on an aiming target, where, for example, various trajectory constraints can be considered. For example, one or more of the following constraints can be considered: allowable deviations from the original plan in terms of position and angular deviation; maximum dogleg capability of the steering assembly; constraints recommended by automated plan analysis that can be adjusted manually and / or automatically (e.g., based on user preferences, etc.); and allowable tortuosity, risk metrics, hole quality, confidence level, etc.

[0130] As explained, work plan generation can be performed using a trajectory generator (e.g., for generating multiple trajectory candidates with different conditions) and a ranking system (e.g., for evaluating each of the generated candidates based on trajectory context, different characteristics, constraint violations, etc.). In such an approach, the WPG can output a single best candidate or several top-ranked candidates.

[0131] With respect to the ranking system, it can operate based on a list of classification items that define the characteristics selected in order to rank the candidates. For example, consider the candidate characteristics, where some examples of these characteristics may include: trajectory length, ROP, total steering length, tool face (TF) orientation, maximum steering ratio, average deviation from plan, risk level, target constraint, angular deviation, tortuosity and one or more other torque and drag constraints, DDI, hole quality, tool wear, number of downlinks, geomechanics, confidence level, and production level index. For each characteristic i utilized (e.g., from i=1 to I, where I is the total number of characteristics), a weight can be defined based on trajectory context, basin location, client type, rig type, etc. In such an example, the weights can be associated with the candidate's cost function to produce a total cost for each candidate trajectory, such as:

[0132]

[0133] In the above, each feature denoted as i may have a cost C i and the associated weight W i , weight W i can be the context-dependent weights, where the total cost of the candidate features of the candidate trajectory can be the sum C prop . Such a ranking system can be modular and extensible. Defining the ranking system in this manner can facilitate the addition of additional candidate characteristics (e.g., according to mathematical and / or logical descriptions). For example, using machine learning methods, the system can map the ease or difficulty of drilling paths from historical data and automatically define a drilling difficulty index from the surface and add it to the aforementioned equation. In such an example, one or more cost functions can be optimized in a manner that can take into account the ease and / or difficulty of drilling. As an example, the system can optionally include a cost function that estimates greenhouse gas (GHG) emissions for each path (e.g., candidate). In such an example, when the cost function is associated with carbon emissions or GHG emissions, the ranking system can operate in a mode (e.g., emissions mode) in which weights can be selected to minimize emissions first (e.g., before other minimizations).

[0134] As an example, once an optimal path has been selected and steering commands for achieving that path have been defined, there may be an additional optimization problem of defining linkages with surface automation equipment to efficiently execute the optimal path. This approach can be performed iteratively, for example, by deriving optimal drilling parameters after a work plan has been generated, or it can be performed concurrently by incorporating surface automation constraints within a sequencing system (e.g., in a scalable manner, etc.). As an example, the system can use an artificial intelligence-based planner to determine a set of possible drilling parameters to use once a work plan has been generated.

[0135] As mentioned, the intelligent actuator can control the steering command, whether it is for one or more of motor steering, RSS steering, or ABSS steering. In certain embodiments, a control layer can be provided when running RSS tools or ABSS tools (e.g., regarding the availability of direct downhole trajectory automation).

[0136] Regarding various aspects of drilling, consider the following terms.

[0137] Rotary Steerable System (RSS): A specialized downhole tool that can deviate the wellbore via electronic command.

[0138] Dogleg Score (DLS): A measure of the change in direction of a wellbore over a defined length, usually measured in degrees per 100 feet of length.

[0139] Yield (Y): Maximum DLS capacity.

[0140] Real-time flexion rate (RT): The flexion rate at a specific time.

[0141] Tool Face (TF): The angle measured in a plane perpendicular to the drill string axis, between a reference direction on the drill string and a fixed reference.

[0142] Expected Command: The expected command to be sent to the downhole tool.

[0143] Actual command: The resulting command executed by the downhole tool.

[0144] Tool face offset (TF O ): The angle measured between the desired TF and the actual TF.

[0145] Build Rate (BR): The DLS projected in the vertical plane at the location of the attached tool. It is also the rate of change of tilt.

[0146] Turn Rate (TR): The DLS projected into the horizontal plane. It is also the rate of change of azimuth angle.

[0147] Travel Rate (WR): The rate of change of the azimuth angle attached to the tool reference axes.

[0148] Steering Ratio (SR): The percentage of time the RSS tool spends steering (biased) versus trying to go neutral (no bias).

[0149] BR O and WR O : BR and WR during the neutral phase.

[0150] As an example, an automated method can estimate steering trend parameters for RSS during well construction execution. In such an example, the parameters can include buckling rate (maximum DLS capability of the steering tool), neutral build (trend of build or decline angle during the neutral phase when no particular direction is preferred), neutral turn (tendency to turn during the neutral phase), and toolface offset (the angular difference between where we were aiming and where we ended up).

[0151] Figure 9 An example of a system 900 is shown that includes a trajectory generator 910 and a ranking system 920. As shown, the trajectory generator 910 can generate candidates and command schedules, while the ranking system 920 includes a weight and penalty block 930, a constraint evaluator block 940, a cost function block 950, and a ranking block 960. The system 900 can generate an output for a sorted list of candidates (e.g., ranked candidates), as indicated by output block 970. As an example, Figure 4The system 400 may include a sorting system 920, such as the sorting system 422, and may include a trajectory generator 910, such as the trajectory generator 416. Figure 5 As shown, ranking system 422 can be a system for handling ranking and constraint violations. As explained, ranking can be performed using weights, for example, weights can be associated with a cost function for a candidate to produce a total cost for each candidate trajectory. For example, in system 900, weight and penalty block 930 can provide and / or generate appropriate weights, penalties (e.g., constraints), etc.

[0152] As an example, the system may include one or more components (e.g., blocks, etc.) that can automatically derive weight values ​​for a ranking system. As explained, such a ranking system may be part of, or otherwise operatively coupled to, a work plan generator (WPG). As explained, such a WPG may be used in field operations, such as directional drilling. As an example, the WPG may be part of a directional drilling advisory system.

[0153] As an example, the ranking system can utilize a technique known as inverse reinforcement learning (IRL). IRL can operate by learning observed behavior from, for example, actual directional drilling (DD) operations performed by one or more humans. IRL methods can be used to map human DD behavior relative to one or more factors, such as considering one or more of a specific region, a specific customer, a specific formation in the subsurface, and a specific type of drilling rig.

[0154] IRL is a paradigm that can exploit Markov Decision Processes (MDPs), where the goal of an apprentice agent can be to find a reward function from expert demonstrations that can explain the expert's behavior.

[0155] In IRL, the goal can be to model an agent taking actions in a given environment. For example, consider a state space S (e.g., the set of states the agent and the environment can be in), an action space A (e.g., the set of actions the agent can take), and a transition function T(s'|s, a), which gives the probability of moving from state s to state s' when taking action a. For example, for an AI learning to control a car, the state space can be the car's possible positions and orientations, the action space can be the set of control signals the AI ​​can send to the car, and the transition function can be a model of the car's dynamics. In IRL, a tuple (S, A, T) can be called an MDP∖R, which is a Markov decision process without a reward function. As an example, an MDP∖R can have a known horizon or discount rate γ.

[0156] The reasoning problem in IRL can be to find the optimal policy for an MDP\R given In such a problem, the system can learn a policy from a sample of states (s, a) and according to corresponding actions (e.g., which can be random). Such samples can come from routes, which record the history of the agent's states and actions in a single episode. For example, consider the following: (a0, s0), (a1, s1), . . . (a N , s N ).

[0157] In the car example above, these may correspond to actions taken by an expert human driver exhibiting desired driving behavior (eg, where the actions may be recorded as signals of the steering wheel, brakes, etc.).

[0158] Given an MDP\R and an observed path, the goal can be to infer a reward function R. In a Bayesian framework, if the system specifies a prior on R:

[0159]

[0160] In such an example, the likelihood yes ,That where is the optimal policy under the reward function R. Note that computing the optimal policy given the reward can be non-trivial (e.g., except in very simple cases). As an example, the system can use reinforcement learning to approximate the policy. Due to the challenges in specifying priors, computing the optimal policy, and integrating the reward function, IRL can achieve an approximation to the Bayesian objective. Piot et al., “Bridging the Gap Between Imitation Learning and Inverse Reinforcement Learning,” IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(8), pp. 1814-1826, is incorporated herein by reference.

[0161] As an example, the ranking system may utilize one or more machine learning techniques to derive appropriate values ​​for the weights to use based on behavior observed from actual directional drilling (DD) operations (e.g., as performed by one or more humans and / or one or more automated system settings on a drilling rig).

[0162] As explained, the system may implement an automated method to derive weight values ​​for a ranking process that may be included as part of work plan generation for a Directional Drilling Advisor (DDA).

[0163] Due to the complexity of the various wells drilled, there are numerous risk and reward contexts for the proper planning and execution of DD for a well. The system can be designed to so-called de-skill and de-manualize the DD process while ensuring efficiency and consistency. As an example, a DD framework can provide various levels of de-skill and de-manualization of the DD process while ensuring efficiency and consistency. In such an example, the DDA framework can provide optimal decisions in real time (e.g., perform real-time optimal decision making). As an example, the DDA framework can provide outputs for one or more of motors, trajectories, commands, and downlink recommendations. For example, consider a DDA framework that provides outputs for an RSS tool for directional drilling. Such a framework can provide outputs for various types of wells, for example, providing outputs from start to finish by automatically providing the next series of actions at each survey point.

[0164] The DDA framework can enable well construction with minimal human intervention, as well as provide local and / or remote (e.g., at the drilling rig or from a town) monitoring and / or intervention. As explained, the DDA framework can include a WPG system that can determine how to construct the best path from the current state to one or more target objective states.

[0165] The WPG may be implemented as a system within a DDA framework (e.g., consider the RSS advisor framework), where the WPG is responsible for deriving a work plan to be followed, e.g., in response to conducting surveys and updating the location of the well (e.g., bottom hole location).

[0166] In practice, the work plan may be a trajectory that constructs a path from the current drill head position (e.g., bottom hole position) to the next target. This path construction may be achieved based on an aiming target and optionally also by taking into account one or more trajectory constraints, such as, for example: allowable deviations from the original plan in terms of position and angular deviation; maximum dogleg capability of the steering assembly; constraints set by the user at initiation (e.g., based on his / her preferences); allowed tortuosity; risk metric; hole quality; confidence level; ROP; carbon emissions; ecologically optimal path (e.g., with minimum energy consumption); and the like.

[0167] As explained, the generation of the work plan may include trajectory generation, which may aim to generate an appropriate number of trajectory candidates under different conditions, and ranking, which may aim to evaluate each of the candidates based on one or more of trajectory context, different characteristics, constraint violations, etc., to rank the available candidates according to a defined optimization goal (e.g., optionally user-defined).

[0168] With respect to the output from the ranking process (e.g., a ranking system), the output can be a single best candidate, or several best candidates that can be exposed to a user for selection, for example, optionally using one or more techniques, such as, for example, a Pareto technique, to select a candidate when no single one of the several best candidates is substantially optimally better than another one of the several best candidates (e.g., a trade-off selection process).

[0169] Reference again Figure 9 An example system 900 is provided, which can be used in RSS applications, mud motor applications, and / or one or more other types of DD applications. As shown, ranking system 920 includes a constraint evaluator block 940, which can provide for evaluating trajectory constraints and applying penalties for each violation, and a cost function block 950, which can provide for evaluating candidate characteristics and applying costs to each characteristic. Ranking block 960, for example, can receive candidates and costs and sort the candidates, for example, in order of increasing cost (e.g., sorted by cost).

[0170] About can be Figure 9 The weight generation performed by the weight and penalty block 930 of the ranking system 920 may generate weights including feature weights that may be used to classify a number of candidates.

[0171] For an example of candidate characteristics and constraints, consider Table 1 below.

[0172] Table 1 Example candidate characteristics and constraints.

[0173]

[0174] As an example, for each feature, the system may associate default and / or user-defined weights based on the trajectory context. Such weights can be associated with the candidate's cost function to produce a total cost for each candidate. As an example, each constraint with an associated constraint violation k can be represented as The constraint value is associated with, and in various examples, represented as W k The weight can be used to constrain the value (See e.g. Figure 14 ), and denoted as P k The weight of can be used for the violation of constraint k, where the cost of constraint violation is denoted as V k (See e.g. Figure 11 ). Furthermore, in drilling engineering (DE), for planning purposes (e.g., generation of well plans), the Z i The weight pair is expressed as KPI iThe performance indicators are weighted (see e.g. Figure 11 and Figure 13 ). This weighting approach can be part of an inverse reinforcement learning (IRL) workflow where the weight values ​​can be determined.

[0175] Regarding the weighting factors, it is considered possible to utilize user-defined weights indicating the importance of the features. For example, As an example, with respect to the weight factors, the method may perform normalization so that the weights are within a desired range (e.g., and )Inside.

[0176] Regarding the cost function, consider that it can produce a single scalar cost value Example of ; normalize so that and and , where values ​​can exceed 1 but can be normalized to keep them below 1; where lower indicates a lower cost for the characteristic; and where 0 indicates an ideal solution, which may not necessarily be achievable. Such an approach can also be applied to constraint violations.

[0177] As an example, each violated constraint k (eg, k=1 to K) may have an associated bound, where a penalty may be applied to each violated constraint k. (e.g., the weight of violating constraint k). In such an example, each violated constraint may have The cost of which the corresponding weight P can be applied k In such an example, the total feature cost for each candidate can be calculated as:

[0178]

[0179] For trajectory context, consider the example of Table 2 below.

[0180] Table 2 Examples of trajectory backgrounds.

[0181]

[0182] As an example, the members of the track context may be dynamic. For example, contexts may be added, deleted, changed, etc., which may be based on system feedback and / or user input.

[0183] Regarding examples of weights, for a set of defined contexts, the system can associate a set of default weights. For example, such weights can be adjusted by the user if desired or appropriate; note that using the default weights can be recommended. Table 3 below shows examples of characteristics of landing contexts and example weight values.

[0184] Table 3. Examples of features and weight values ​​(landing context).

[0185]

[0186] As explained, Table 3 shows an example of weight definitions for features in a landing context. Such weights can be based on the importance attached to each feature in the landing context. The weights can be different for different contexts.

[0187] Table 4 below shows examples of constraint violation types and associated penalty weight values.

[0188] Table 4 Example violation types and penalty weight values ​​(landing context).

[0189]

[0190] As explained, the system can act to derive an optimal path from the current hole position to one or more targets. In such an example, a user may want to know the optimal path to one or more targets given the current hole position and the constraints associated with the current context. Regarding constraints, one or more of, for example, flexion rate, allowable deviation from the original plan, pose constraints, etc. may be considered.

[0191] As an example, a function may consume a bottom hole location and one or more targets as inputs, and develop an optimal path to the one or more targets as output. In such an example, the inputs may include, for example, one or more of the HBE, a trend, and one or more targets. In such an example, the outputs may include one or more work plans. As an example, configuration may include using the original plan along with trajectory constraints and tool type (e.g., RSS or mud motor).

[0192] Figure 10 An example of a reinforcement learning process 1000 and an example of an inverse reinforcement learning (IRL) process 1060 are shown. As explained, rewards can drive the reinforcement learning process to generate control actions. As shown, the reinforcement learning process 1000 includes a reinforcement learning block 1012 that receives a dynamics model 1004 and a reward function 1008 to generate a control action 1016. In contrast, the IRL process 1060 includes an inverse reinforcement learning block 1072 that receives a control action 1064 and a dynamics model 1068 to generate a reward function 1076.

[0193] As explained, the IRL method can learn from human behavior by utilizing the IRL process. For example, consider the following definition of the reward function:

[0194]

[0195] In the preceding equation for the reward function, and Can be an unknown value of the weight to be derived by the IRL process. , if the candidate does not violate the constraint ,but = 0 (i.e., the cost of constraint violation is zero). However, if the candidate trajectory violates the constraint ,but The value depends on the degree of constraint violation. For example, the greater the degree of constraint violation, the more likely the candidate will have a very large (e.g., huge) cost. The larger the value of ; therefore, such a candidate will be very unlikely to be ultimately selected. As explained, may be a weight that can be applied to the constraint violation k, such as, for example, by Multiply by cost .

[0196] With respect to the DDA framework, expert DD decisions may be available for multiple wells successfully drilled in a directionally controlled manner, including, for example, steering commands and drilling parameters used. This information may be stored in a database, which may be a behavior database, where the behaviors may be those associated with successful DD operations. Thus, historical data may be utilized to improve frameworks (e.g., DDA frameworks) that include automated directional drilling capabilities.

[0197] Once appropriate weights are defined, a ranking system can utilize them, which can also allow for improved understanding of the behavior of automated DDA. As an example, the framework can provide for mapping human DD behavior and / or machine DD behavior in a specific area, particularly a formation subsurface, for a specific type of drilling rig, etc., in relation to a specific client.

[0198] While directional drilling is mentioned, the framework may alternatively or additionally provide improved planning, which may be a task performed by drilling engineers. For example, one or more drilling engineers may perform drilling engineering (DE) by generating a digital well plan. For example, the framework may provide learning based on the actions of one or more drilling engineers and / or one or more automated or semi-automated planners (e.g., considering a well planning framework that provides some level of automation).

[0199] Figure 11 An example of a method 1100 that may be implemented by the framework for directional drilling (DD) and / or drilling engineering (DE) purposes is shown. Figure 11In the example of , example reward equations for DD and DE are shown, which include weights. One or more of such equations can be used as part of the IRL process to determine the weights (e.g., the values ​​of the weights).

[0200] like Figure 11 As shown in the example of , method 1100 includes a receiving block 1112 for receiving a well plan (e.g., a digital well plan) for a subject well, an acquisition block 1114 for acquiring data from one or more offset wells, an execution block 1116 for performing a data quality control check, and a decision block 1118 for determining whether the data quality is acceptable. As shown, if decision block 1118 determines that the data quality is unacceptable, method 1100 may continue to decision block 1120 for determining whether the data can be refined, wherein if the data is not suitable for sufficient refinement, method 1100 may proceed to default weight block 1122 for using default weights (e.g., default values ​​for the weights); however, if the data is suitable for refinement, refinement block 1124 may continue to refine the data and then return to blocks 1112 and / or 1114, as may be appropriate.

[0201] like Figure 11 As shown in the example of FIG, if decision block 1118 determines that the data quality is acceptable, method 1100 may proceed to a filtering block 1130 for filtering the data by trajectory context, which may then proceed to a plotting block 1132 for plotting each run using a planned steering tool for directional drilling of a planned well. As shown, a decision block 1134 may provide a determination of whether reward weights are available. If it is determined that they are not, method 1100 may proceed to a plotting block 1136 for plotting human DD and / or human DE and / or autonomous system behavior (e.g., for DD, DE, etc.), and then to a generating block 1138 for generating weights using IRL techniques. As shown, if decision block 1134 determines that reward weights are available, method 1100 may proceed to another decision block 1140 for determining whether information on one or more additional wells is available since the last reward generation. If decision block 1140 determines that data is not available for one or more additional wells, method 1100 may proceed to use block 1142 for using weights, as previously generated; however, if decision block 1140 determines that data is available for one or more additional wells (e.g., due to activity, availability of data in one or more data stores, etc.), method 1100 may proceed to plot block 1144 to plot the data for the one or more additional wells and then proceed to generate block 1146 to generate weights using IRL techniques.

[0202] exist Figure 11 In the example of , an example of a reward function is shown, which includes and , which may be an unknown value (e.g., a weight value) to be derived by an IRL technique (e.g., an IRL process):

[0203]

[0204] As explained, IRL techniques (see e.g. Figure 10 The IRL process 1060 can leverage knowledge of control actions, such as the types of actions that can be taken in certain contexts (e.g., historical observations of actions, etc.). This knowledge can come from the current well and / or from one or more offset wells. As explained, methods such as method 1100 can utilize IRL techniques to determine weights for characteristic costs and constraint violation costs. As explained, this method can be used for drilling (e.g., directional drilling (DD)).

[0205] exist Figure 11 In the example above, another equation for the reward is as follows:

[0206]

[0207] In this example, a performance indicator (e.g., a KPI) can be expressed as and may have corresponding values ​​and may also have corresponding weights, which are expressed in the above equation as As explained, methods such as method 1100 can utilize IRL techniques to determine weights for performance indicators. As explained, such methods can be used for planning (eg, drilling engineering (DE)) to generate digital well plans, etc.

[0208] Regarding observations of control actions that are observed behaviors of directional drilling of a well, consider, for example, steering commands as a type of control action (e.g., observed behaviors). In such an example, for a mud motor application, the data may indicate the sequence of slides and rotations, and for slides, the tool face may be used along with the slide length; however, for an RSS application, the data may indicate the mode used and the mode parameters. As an example, the method may include extracting direct and / or indirect indicators of one or more drilling parameters, such as ROP, flow rate, surface RPM, differential pressure (e.g., when using a mud motor), etc. As explained, the observations may be used for decision making, whether control decisions or other decisions, which may be used for DD and / or DE. As an example, the observations may be used in decisions made by drilling engineers and / or planning systems, where such observations may be utilized to improve the planning process (e.g., planning framework, etc.).

[0209] Figure 12Table 1200 is shown, which includes examples of available modes and mode parameters for one or more types of RSS tools. As shown, the modes may include manual, vertical, tilt hold, hold tilt and azimuth, and auto bend, where these modes may be abbreviated and categorized as manual or automatic. In such an example, operating rules may be defined along with firmware constraints, mandatory inputs, and optional additional inputs.

[0210] As an example, the framework can provide for automatically deriving feature weights and violation weights for a ranking system that derives an optimal path from a current bottom hole location to one or more desired targets. As an example, the framework can provide for learning and mapping human and / or machine behavior during directional drilling (DD) applications and / or during drilling engineering (DE) applications.

[0211] As an example, a framework may be provided for deriving acceptable directional drilling practices (e.g., for DD and / or for DE) associated with a particular basin. For example, considering an analysis of weights, etc., the analysis may determine which practices are acceptable and / or improve directional drilling for one or more contexts. This approach may provide for deriving good directional drilling practices associated with one or more of a particular region, a particular customer's preferences, a particular formation, a particular type of drilling rig, and / or a particular type of well profile.

[0212] As an example, the framework may be operable to evaluate and / or improve one or more practices. As an example, the framework may remove one or more practices, such as based on lack of contribution, inefficiency, etc.

[0213] As an example, the framework may provide for deriving acceptable directional drilling practices associated with a particular type of well in line-of-sight mode or blind mode (e.g., optionally without measurement data from an RSS tool) and / or associated with a particular type of BHA including particular drill bit behavior.

[0214] As an example, the framework can provide for characterizing human behavior, particularly human decision making regarding control actions for directional drilling (e.g., generated by a weighted ranking system based on human behavior). As an example, the framework can provide for using human behavior to derive a reward function in the presence of data scarcity.

[0215] Figure 13An example of a method 1300 that can be implemented by the framework is shown. As shown, method 1300 includes a receiving block 1310 for receiving a new trajectory design task. In response, method 1300 may proceed to one or more blocks 1312, 1314, and 1316, where block 1312 provides automatic trajectory generation (ATD) to generate multiple trajectories that satisfy given constraints and calculate performance metrics (e.g., KPIs) for each trajectory, block 1314 may extract data from one or more relevant offset wells in one or more databases, and block 1316 may extract data from one or more previous versions of the well plan (if available). As shown, method 1300 may proceed to a decision block 1320, which may determine whether the data quality is acceptable (e.g., whether the data quality is appropriate or inappropriate). If the quality is acceptable, method 1300 may proceed to a generating block 1322 for generating weights using an IRL technique, where method 1300 may then proceed to an initialization block 1324 to initialize a counter, such as Counter X, which may be set to an appropriate value (e.g., 1, etc.). At block 1326, which may be part of a loop with blocks 1328 and 1332, method 1300 may recommend X candidates based on the weights (e.g., note that if X equals 1, one candidate may be recommended). Next, in the loop, block 1328 (decision block) may determine whether to accept the recommended candidates. If decision block 1328 indicates acceptance (yes branch), method 1330 may save the X candidates and weights in a database; otherwise, for rejection (no branch), method 1300 may increment a counter by an appropriate number (e.g., 10, etc.) and proceed to recommendation block 1326.

[0216] Referring again to decision block 1320, if decision block 1320 indicates that the data quality is unacceptable, method 1300 may proceed to block 1340 where a trajectory analysis recommender may be implemented, which may recommend ten or more candidates according to recommendation block 1342. This may be followed by decision block 1344 where it is determined whether the candidate is acceptable, wherein if so, save block 1346 may provide for saving the recommender parameters; however, if the candidate is unacceptable, method 1300 may proceed to true recommender block 1348 for true parameters of recommendation block 1340.

[0217] like Figure 13 As shown in the example of , there are various customization options, including customization options for customers, sites, locations, human DE preferences, track types, formations, etc.

[0218] exist Figure 13 In the example, the reward equation is as follows:

[0219]

[0220] In such an example, a performance indicator (e.g., a KPI) may be represented as and may have corresponding values ​​and may also have corresponding weights (e.g., in the above equation ,like Figure 13 As explained, methods such as method 1300 may utilize IRL techniques to determine weights for performance indicators.

[0221] Figure 14 An example of a method 1400 is shown that includes a receive block 1410 for receiving a request for a new trajectory design and / or work plan generation task. In response to such a request, the method 1400 may automatically proceed to extraction blocks 1412 and 1414 for extracting data from relevant offset wells in a database and / or for extracting data for one or more previous versions of one or more drilling plans, if available. For example, the request may relate to an existing drilling plan, such that data may be extracted from a previous version of the drilling plan (e.g., as a digital well plan that may be stored in a database). As shown, the method 1400 may continue to an extraction block 1420 for extracting constraints from historical data, followed by a determination block 1422 that may determine whether a reward weight is available. In Figure 14 In the example of FIG, if the decision block 1422 determines that the reward weights are available, the method 1400 may continue to the use block 1424 to use the previously generated available weights, followed by the recommendation block 1426 to recommend constraints. However, if the decision block 1422 determines that the reward weights are not available, the method 1400 may proceed to the generation block 1428, which may analytically generate constraints, and then continue to the recommendation block 1426.

[0222] As shown, given the constraints, method 1400 can proceed to decision block 1430, which can determine whether to accept the recommended constraints. In the event of acceptance at decision block 1430, method 1400 can proceed to save block 1432 to save the constraints in a database, and then proceed to generate block 1440 for generating weights using IRL techniques. As shown, if decision block 1430 determines that the constraints are unacceptable (e.g., unacceptable), method 1400 can proceed to manual input block 1434 for manually entering one or more constraints, wherein method 1400 can then proceed to save block 1432 and generate block 1440 (e.g., using the one or more constraints entered in accordance with block 1434 to generate weights).

[0223] As an example, a constraint may be an allowable deviation from the plan (eg, in one or more directions) with respect to the execution of the plan (eg, directional drilling). As an example, a constraint may be a DLS limit for one or more portions of the trajectory.

[0224] exist Figure 14 In the example, the reward equation is as follows:

[0225]

[0226] As explained, constraints can be associated with constraint values. is denoted as k, and may also have corresponding weights (e.g., in the above equation ,like Figure 14 As explained, methods such as method 1400 may utilize IRL techniques to determine the weights of constraints.

[0227] As explained, P k is the weight (e.g., for k = 1 to K, where K is the total number of constraints or constraint violations), and V k is the cost of constraint violation (e.g., k = 1 to K). When a constraint is defined as having value T k When the following background exists: (1) If V k <T k , then no penalty is applied, and the violation cost of the constraint violation k can be zero; and (2) if V k > T k , then the cost of violating the characteristic will be the weight P k Multiply by V k and T k The difference between (for example, V k -T k ). This approach helps ensure that the greater the deviation from the constraint, the greater the cost of the feature.

[0228] As an example, the framework can take sustainability issues into account. For example, the carbon emissions and / or GHG emissions of each candidate are taken into account as part of the ranking system. As explained, the ranking system can have an eco-path mode that can select the most eco-friendly path or prioritize low-emission paths over other paths. As an example, the framework can include a performance mode where the ranking system can focus more on the highest ROP candidates than other ROP candidates to complete the well as quickly as possible. As an example, the ranking system can be evaluated using different drilling parameters, where human behavior is mapped by taking these parameters into account. As explained, Table 1 lists various examples of constraints.

[0229] As an example, one or more machine learning and / or artificial intelligence techniques can be utilized to derive weights for the ranking system by modeling at least some human behavior. While IRL is mentioned, the framework can utilize optimization techniques that can implement one or more genetic algorithms. As an example, the framework can generate weights based on an iterative technique that can start with default weights and gradually derive weights that best mimic the behavior of one or more people (e.g., considering projection techniques).

[0230] As explained, the values ​​of the weights may be applied to determine a cost associated with each of a plurality of candidate trajectories for drilling execution and / or drilling planning.

[0231] Figure 15 An example of method 1500 is shown, which may include a receiving box 1510 for receiving digital well planning data for directional drilling of a well in a subsurface geological environment; a generating box 1520 for generating candidate trajectories for the borehole using the digital well planning data; a generating box 1530 for generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and an output box 1540 for outputting one or more candidate trajectories from the candidate trajectories for the directional drilling of the well based on the ranking.

[0232] like Figure 15 As shown, method 1500 may be implemented in accordance with blocks 1511, 1521, 1531, and 1541 via one or more computer readable media (CRMs), which may be implemented using, for example, a computing system (see, for example, Figure 3 Example system 300, Figure 7 Such blocks may include processor-executable instructions.

[0233] As an example, a computational framework may include a solver, which may be implemented via executable instructions. For example, consider a computational framework comprising a processor and a memory accessible to the processor, wherein the executable instructions may be stored in the memory and accessed for execution by the processor to cause the computational framework to perform one or more actions. Such a computational framework may include one or more interfaces for receiving information such as survey information and / or for outputting information, which may include parameter values, instructions, etc. As an example, the computational framework may be part of a controller. As an example, the computational framework may be part of a system (see, e.g., Figure 3 and Figure 7 computing systems).

[0234] As explained, various systems, methods, etc. can implement one or more ML models. Regarding the type of ML model, consider one or more of a support vector machine (SVM) model, a k-nearest neighbor (KNN) model, an ensemble classifier model, a neural network (NN) model, incremental learning, Q-learning, etc. As examples, the machine learning model can be a deep learning model (e.g., a deep Boltzmann machine, a deep belief network, a convolutional neural network, a stacked autoencoder, etc.), an ensemble model (e.g., a random forest, a gradient boosting machine, bootstrap aggregation, adaptive boosting (AdaBoost), stacked generalization, a gradient boosted regression tree, etc.), a neural network model (e.g., a radial basis function network, a perceptron, backpropagation, a Hopfield network, etc.), a regularization model (e.g., a ridge regression, a least absolute shrinkage and selection operator, an elastic net, a least angle regression), a rule system model (e.g., a cube, a one rule, a zero rule, repeated incremental pruning to produce error reduction), a regression model (e.g., a linear regression, an ordinary least squares regression, a stepwise regression, a multivariate adaptive regression spline, a local estimation scatter plot smoothing, a logistic regression, etc.). ), Bayesian models (e.g., naive Bayes, average dependency estimator, Bayesian belief network, Gaussian naive Bayes, multinomial naive Bayes, Bayesian network), decision tree models (e.g., classification and regression trees, iterative bisection3, C4.5, C5.0, chi-square automatic interaction detection, decision stump, conditional decision tree, M5), dimensionality reduction models (e.g., principal component analysis, partial least squares regression, Sammon map, 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.), instance models (e.g., k-nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, etc.), clustering models (e.g., k-means, k-median, expectation maximization, hierarchical clustering, etc.), etc.

[0235] As an example, the system can utilize one or more recurrent neural networks (RNNs). One type of RNN is called a long short-term memory (LSTM), which can be a unit or component (e.g., one or more units) that can be in one or more layers. The LSTM component can be an artificial neural network (ANN) designed to recognize patterns in data sequences (such as time series data). When provided with time series data, the LSTM considers time and sequence, so that the LSTM can include a time dimension. For example, consider utilizing one or more RNNs to process temporal data from one or more sources, optionally combined with spatial data. This approach can identify temporal patterns that can be used to make predictions (e.g., one or more patterns about future times, etc.).

[0236] As an example, the Tensorflow framework (Google LLC, Mountain View, California) can be implemented. It is an open source software library for dataflow programming, including a symbolic mathematics library that can be implemented for machine learning applications that may include neural networks. As an example, the Caffe framework, a DL framework developed by Berkeley AI Research (BAIR) (University of California, Berkeley), can be implemented. 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 AI framework (APOLLO.AI GmbH, Germany) can be utilized. As mentioned above, a framework such as the Pytorch framework can also be used.

[0237] As an example, a training method can include various actions that can be performed on a dataset to train an ML model. As an example, a dataset can be split into training data and test data, where the test data can be provided for evaluation. The method can include cross-validation of parameters and optimal parameters, which can be provided for model training.

[0238] The Tensorflow framework can be run on multiple CPUs and GPUs (with optional CUDA (NVIDIA Corporation, Santa Clara, CA) and SYCL (Khronos Group Inc., Beaverton, OR) extensions for general-purpose computing on graphics processing units (GPUs). Tensorflow is available for 64-bit Linux, macOS (Apple Inc., Cupertino, CA), Windows (Microsoft Corporation, Redmond, WA), and mobile computing platforms including platforms based on the Android (Google LLC, Mountain View, CA) and iOS (Apple Inc.) operating systems.

[0239] Tensorflow computations can be represented as stateful data flow graphs; note that the name Tensorflow comes from the operations that this neural network performs on multidimensional arrays of data. Such arrays can be called "tensors."

[0240] As an example, the ML model can be run online using cloud computing resources, followed by a target well delivery method that can automatically feed data to the ML model, and the ML model can be updated at a given frequency. As an example, the ML model can be run in an offline manner, where one or more results can be transmitted to a planned workflow.

[0241] As an example, a method may include receiving digital well plan data for directional drilling of a well in a subsurface geological environment; generating candidate trajectories for the borehole using the digital well plan data; generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and outputting one or more of the candidate trajectories for directional drilling of the borehole based on the ranking. In such an example, the values ​​derived from inverse reinforcement learning may include values ​​of weights, wherein, for example, the values ​​of the weights are applied to determine a cost associated with each of the candidate trajectories for directional drilling after directional drilling begins, and / or the values ​​of the weights are applied to determine a cost associated with each of the candidate trajectories to generate the digital well plan for directional drilling.

[0242] As an example, historical observations may include observations of one or more of human decisions used to control the directional drilling and human decisions used to plan the directional drilling.

[0243] As an example, the historical observations may include observations of autonomous control system decisions used to control directional drilling and / or autonomous planning system decisions used to generate a plan for directional drilling. For example, the autonomous control system may be a computational control framework that can make at least some automated decisions, and the autonomous planning system may be a computational planning framework that can make at least some automated decisions.

[0244] As examples, values ​​derived from inverse reinforcement learning may be derived from historical observations of mud motor-based directional drilling and / or from historical observations of rotary steering system-based directional drilling.

[0245] As an example, a method may include classifying historical observations by context from a set of predefined contexts. In such an example, the predefined contexts may include, for example, one or more of a mud motor context and a rotary steering system context, a landing context, and at least one nudge context.

[0246] As an example, a method may include updating a value in response to receiving one or more additional historical observations. For example, the computational framework may be an online framework that can optionally access and / or receive observations from one or more operations in real time. In such an example, the observations may be used for directional drilling and / or for drilling engineering. As an example, the updated value may be used to control directional drilling and / or plan directional drilling (e.g., to generate a digital well plan, etc.).

[0247] As an example, values ​​derived via inverse reinforcement learning (IRL) may include values ​​that weight context-dependent costs, and values ​​that weight constraint violations.

[0248] As an example, inverse reinforcement learning can include utilizing a reward function defined by a sum of a context-dependent cost and a sum of a constraint violation cost for each of a plurality of candidate trajectories.

[0249] As an example, a method may include using a candidate trajectory from one or more candidate trajectories to perform directional drilling from a current bottom hole location of a borehole to a target location of the borehole. In such an example, the method may include, in response to receiving survey data from a downhole survey performed by a downhole survey device in the borehole, repeating generating a new candidate trajectory for the borehole and generating a ranking of the new candidate trajectories.

[0250] As an example, the generation of candidate trajectories for a borehole and the generation of a ranking of the candidate trajectories can occur automatically in response to receiving survey data from a downhole survey performed by downhole survey equipment in the wellbore. For example, consider a downhole survey that generates survey data, which can be transmitted from a downhole location to a surface location via telemetry, where a computing framework at the surface can receive the survey data and automatically respond to receipt of the survey data to generate and rank candidate trajectories for directional drilling from a survey location determined using the survey data to a target location, which can be specified, for example, in a digital well plan.

[0251] As an example, a system may include one or more processors; a memory accessible to at least one of the one or more processors; processor-executable instructions stored in the memory and executable to instruct the system to: receive digital well planning data for directional drilling of a well in a subsurface geological environment; generate candidate trajectories for the borehole using the digital well planning data; generate a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and output one or more candidate trajectories from the candidate trajectories for directional drilling of the borehole based on the ranking.

[0252] As an example, one or more non-transitory computer-readable storage media may include processor-executable instructions to instruct a computing system to: receive digital well planning data for directional drilling of a well in a subsurface geological environment; generate candidate trajectories for the borehole using the digital well planning data; generate a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning using historical observations of one or more other boreholes; and output one or more candidate trajectories from the candidate trajectories for directional drilling of the borehole based on the ranking.

[0253] As an example, a computer program product may include computer-executable instructions to instruct a computing system to perform one or more methods, such as one or more of the methods described herein (eg, in part, in whole, and / or in various combinations).

[0254] The embodiments disclosed in this disclosure are intended to help explain the concepts described herein. This description is not exhaustive and does not limit the claims to the precise embodiments disclosed. Modifications and variations from the precise embodiments in this disclosure may still fall within the scope of the claims.

[0255] Likewise, the steps described need not be performed in the same order or with the same degree of separation as discussed. Various steps may be omitted, repeated, combined, or divided as appropriate. Accordingly, the present disclosure is not limited to the embodiments described above, but is defined by the appended claims in accordance with their full scope of equivalents. Throughout the foregoing description and the following claims, unless otherwise indicated, the term "execution" and its variations are to be construed as referring to any operation on program code or instructions on a device, whether compiled, interpreted, or run using other techniques.

[0256] Some of the following claims may include numbered lists. These numbers are provided as an organizational tool to aid readability. The numbers themselves do not indicate an intended order of configuration or execution or otherwise have substantive meaning. For U.S. applications, the appended claims do not invoke section 112(f) unless the phrase "means for..." is expressly used in conjunction with an associated feature.

Claims

1. A method (1500) comprising: receiving digital well planning data (1510) for directional drilling of a well in a subsurface geological environment; generating candidate trajectories for the borehole using the digital well plan data (1520); Using historical observations of one or more other boreholes, generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning (1530); and Based on the ranking, one or more candidate trajectories for directional drilling of the borehole are output (1540).

2. The method of claim 1 , wherein the values ​​derived from inverse reinforcement learning include values ​​of weights, wherein the values ​​of the weights are applied to determine a cost associated with each of the candidate trajectories for the directional drilling after the directional drilling begins, and / or wherein the values ​​of the weights are applied to determine a cost associated with each of the candidate trajectories to generate a digital well plan for the directional drilling.

3. The method of claim 1 or 2, wherein the historical observations include observations of one or more of human decisions used to control directional drilling and human decisions used to plan directional drilling.

4. A method according to any preceding claim, wherein the historical observations include observations of autonomous control system decisions used to control directional drilling.

5. A method according to any preceding claim, wherein the value derived from inverse reinforcement learning is derived from historical observations of mud motor based directional drilling.

6. A method according to any preceding claim, wherein the value derived from inverse reinforcement learning is derived from historical observations of directional drilling based on a rotational steering system.

7. A method according to any preceding claim, comprising classifying the historical observations by context from a set of predefined contexts, optionally wherein the predefined contexts include a mud motor context and a rotary steering system context.

8. The method according to any preceding claim, wherein the predetermined background comprises a landing background, and / or wherein the predetermined background comprises at least one nudging background.

9. A method according to any preceding claim, comprising updating the value in response to receiving one or more additional historical observations.

10. A method according to any preceding claim, wherein the values ​​include values ​​for weighting applied to context-related costs and values ​​for weighting applied to constraint violation costs.

11. A method according to any preceding claim, wherein the inverse reinforcement learning utilizes a reward function defined by the sum of context-dependent costs and the sum of constraint violation costs for each of the candidate trajectories.

12. A method according to any preceding claim, comprising using one of the one or more candidate trajectories to carry out directional drilling from a current bottom hole position of the borehole to a target position of the borehole, optionally comprising repeating the generation of new candidate trajectories for the borehole and generating a ranking of new candidate trajectories in response to receiving survey data from a downhole survey performed by downhole survey equipment in the borehole.

13. A method according to any preceding claim, wherein the generating of candidate trajectories for the borehole and the ranking of the candidate trajectories occur automatically in response to receiving survey data from a downhole survey performed by downhole survey equipment in the borehole.

14. A system (300), comprising: one or more processors (302); a memory (304) accessible by at least one of the one or more processors; Processor-executable instructions (312) stored in the memory and executable to instruct the system to: receiving digital well planning data (1511) for directional drilling of a well in a subsurface geological environment; generating candidate trajectories for the drill holes using the digital well plan data (1521); Using historical observations of one or more other boreholes, generating a ranking of the candidate trajectories using values ​​derived from inverse reinforcement learning (1531); and Based on the ranking, one or more candidate trajectories for directional drilling of the borehole are output (1541).

15. A computer program product comprising computer executable instructions for instructing a computing system to perform the method according to any one of claims 1 to 13.

Images