Wellsite pump system framework

Abstract

A method can include receiving by a computational device at a wellsite, real-time, time series data from a pump system operating at the wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; using the computational device, processing a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issuing a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

Description

WELLSITE PUMP SYSTEM FRAMEWORKRELATED APPLICATIONS

[0001] This application claims priority to and the benefit of a U.S. Provisional Application having Serial No. 63 / 626,337, filed 29 January 2024, which is incorporated by reference herein in its entirety.BACKGROUND

[0002] A reservoir can be a subsurface formation that can be characterized at least in part by its porosity and fluid permeability. As an example, a reservoir may be part of a basin such as a sedimentary basin. A basin can be a depression (e.g., caused by plate tectonic activity, subsidence, etc.) in which sediments accumulate. As an example, where hydrocarbon source rocks occur in combination with appropriate depth and duration of burial, a petroleum system may develop within a basin, which may form a reservoir that includes hydrocarbon fluids (e.g., oil, gas, etc.). Various operations may be performed in the field to access such hydrocarbon fluids and / or produce such hydrocarbon fluids. For example, consider equipment operations where equipment may be controlled to perform one or more operations.SUMMARY

[0003] A method can include receiving by a computational device at a wellsite, real-time, time series data from a pump system operating at the wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; using the computational device, processing a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issuing a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system. A system can include a processor; a memory accessible to the processor; and processor-executable instructions stored in the memory to instruct the system to: receive real-time, time series data from a pump system operating at a wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of aforthcoming performance issue of the pump system; and issue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system. One or more computer-readable storage media can include processor-executable instructions to instruct a wellsite computing system to: receive real-time, time series data from a pump system operating at a wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system. 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 as an aid in limiting the scope of the claimed subject matter.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Features and advantages of the described implementations can be more readily understood by reference to the following description taken in conjunction with the accompanying drawings.

[0006] FIG. 1 illustrates an example system that includes various framework components associated with one or more geologic environments;

[0007] FIG. 2 illustrates examples of equipment, an example of a network and an example of a system;

[0008] FIG. 3 illustrates an example of an electric submersible pump system;

[0009] FIG. 4 illustrates examples of portions of an electric submersible pump system;

[0010] FIG. 5 illustrates an example of a components of an electric submersible pump system;

[0011] FIG. 6 illustrates an example of a system;

[0012] FIG. 7 illustrates graphics of an example of a workflow;

[0013] FIG. 8 illustrates an example of a graphical user interface;

[0014] FIG. 9 illustrates an example of a system;

[0015] FIG. 10 illustrates an example of a framework;

[0016] FIG. 11 illustrates an example of a plot;

[0017] FIG. 12 illustrates an example of a plot;

[0018] FIG. 13 illustrates an example of a plot;

[0019] FIG. 14 illustrates an example of a plot;

[0020] FIG. 15 illustrates an example of a method and an example of a system; and

[0021] FIG. 16 illustrates examples of computer and network equipment.DETAILED DESCRIPTION

[0022] This description is not to be taken in a limiting sense, but rather is made merely for the purpose of describing the general principles of the implementations. The scope of the described implementations should be ascertained with reference to the issued claims.Example System

[0023] FIG. 1 shows an example of a system 100 that includes a workspace framework 110 that can provide for instantiation of, rendering of, interactions with, etc., a graphical user interface (GUI) 120. In the example of FIG. 1 , the GU1 120 can include graphical controls for computational frameworks (e.g., applications) 121 , projects 122, visualization 123, one or more other features 124, data access 125, and data storage 126.

[0024] In the example of FIG. 1 , the workspace framework 110 may be tailored to a particular geologic environment such as an example geologic environment 150. For example, the geologic environment 150 may include layers (e.g. , stratification) that include a reservoir 151 and that may be intersected by a fault 153. As an example, the geologic environment 150 may be outfitted with a variety of sensors, detectors, actuators, etc. For example, equipment 152 may include communication circuitry to receive and to transmit information with respect to one or more networks 155. Such information may include information associated with downhole equipment 154, which may be equipment to acquire information, to assist with resource recovery, etc. Other equipment 156 may be located remote from a wellsite and include sensing, detecting, emitting or other circuitry. Such equipment may include storage and communicationcircuitry to store and to communicate data, instructions, etc. As an example, one or more satellites may be provided for purposes of communications, data acquisition, etc. For example, FIG. 1 shows a satellite 170 in communication with the network 155 that may be configured for communications, noting that the satellite may additionally or alternatively include circuitry for imagery (e.g., spatial, spectral, temporal, radiometric, etc.).

[0025] FIG. 1 also shows the geologic environment 150 as optionally including equipment 157 and 158 associated with a well that includes a substantially horizontal portion that may intersect with one or more fractures 159. For example, consider a well in a shale formation that may include natural fractures, artificial fractures (e.g., hydraulic fractures) or a combination of natural and artificial fractures. As an example, a well may be drilled for a reservoir that is laterally extensive. In such an example, lateral variations in properties, stresses, etc. may exist where an assessment of such variations may assist with planning, operations, etc. to develop a laterally extensive reservoir (e.g., via fracturing, injecting, extracting, etc.). As an example, the equipment 157 and / or 158 may include components, a system, systems, etc. for fracturing, seismic sensing, analysis of seismic data, assessment of one or more fractures, etc.

[0026] In the example of FIG. 1 , the GUI 120 shows some examples of computational frameworks, including the DRILLPLAN, PETREL, TECHLOG, DRILLOPS, PETROMOD, ECLIPSE, and INTERSECT frameworks (SLB, Houston, Texas); noting that one or more other frameworks may be included, additionally or alternatively.

[0027] The DRILLPLAN framework provides for digital well construction planning and includes features for automation of repetitive tasks and validation workflows, enabling improved quality drilling programs (e.g., digital drilling plans, etc.) to be produced quickly with assured coherency.

[0028] The DRILLOPS framework may execute a digital drilling plan and ensures plan adherence, while delivering goal-based automation. The DRILLOPS framework may generate activity plans automatically individual operations, whether they are monitored and / or controlled on the rig or in town. Automation may utilize data analysis and learning systems to assist and optimize tasks, such as, for example, setting ROP to drilling a stand. A preset menu of automatable drilling tasks may be rendered, and, using data analysis and models, a plan may be executed in a mannerto achieve a specified goal, where, for example, measurements may be utilized for calibration. The DRILLOPS framework provides flexibility to modify and replan activities dynamically, for example, based on a live appraisal of various factors (e.g., equipment, personnel, and supplies). Well construction activities (e.g., tripping, drilling, cementing, etc.) may be continually monitored and dynamically updated using feedback from operational activities. The DRILLOPS framework may provide for various levels of automation based on planning and / or re-planning (e.g., via the DRILLPLAN framework), feedback, etc.

[0029] The PETREL framework can be part of the DELFI cognitive exploration and production (E&P) environment (SLB, Houston, Texas, referred to as the DELFI environment) for utilization in geosciences and geoengineering, for example, to analyze subsurface data from exploration to production of fluid from a reservoir.

[0030] One or more types of frameworks may be implemented within or in a manner operatively coupled to the DELFI environment, which is a secure, cognitive, cloud-based collaborative environment that integrates data and workflows with digital technologies, such as artificial intelligence (Al) and machine learning (ML). As an example, such an environment can provide for operations that involve one or more frameworks. The DELFI environment may be referred to as the DELFI framework, which may be a framework of frameworks. As an example, the DELFI environment can include various other frameworks, which can include, for example, one or more types of models (e.g., simulation models, etc.).

[0031] The TECH LOG framework can handle and process field and laboratory data for a variety of geologic environments (e.g., deepwater exploration, shale, etc.). The TECHLOG framework can structure wellbore data for analyses, planning, etc.

[0032] The PIPESIM simulator includes solvers that may provide simulation results such as, for example, multiphase flow results (e.g., from a reservoir to a wellhead and beyond, etc.), flowline and surface facility performance, etc. The PIPESIM simulator may be integrated, for example, with the AVOCET production operations framework (SLB, Houston Texas). As an example, a reservoir or reservoirs may be simulated with respect to one or more enhanced recovery techniques (e.g., consider a thermal process such as steam-assisted gravity drainage (SAGD), etc.). As an example, the PIPESIM simulator may be an optimizer that can optimize one or more operational scenarios at least in part via simulation of physical phenomena.

[0033] The ECLIPSE framework provides a reservoir simulator (e.g., as a computational framework) with numerical solutions for fast and accurate prediction of dynamic behavior for various types of reservoirs and development schemes.

[0034] The INTERSECT framework provides a high-resolution reservoir simulator for simulation of detailed geological features and quantification of uncertainties, for example, by creating accurate production scenarios and, with the integration of precise models of the surface facilities and field operations, the INTERSECT framework can produce reliable results, which may be continuously updated by real-time data exchanges (e.g., from one or more types of data acquisition equipment in the field that can acquire data during one or more types of field operations, etc.). The INTERSECT framework, as with the other example frameworks, may be utilized as part of the DELFI cognitive E&P environment, for example, for rapid simulation of multiple concurrent cases. For example, a workflow may utilize one or more of the DELFI on demand reservoir simulation features.

[0035] The aforementioned DELFI environment provides various features for workflows as to subsurface analysis, planning, construction and production, for example, as illustrated in the workspace framework 110. As shown in FIG. 1 , outputs from the workspace framework 110 can be utilized for directing, controlling, etc., one or more processes in the geologic environment 150 and, feedback 160, can be received via one or more interfaces in one or more forms (e.g., acquired data as to operational conditions, equipment conditions, environment conditions, etc.). While several simulators are illustrated in the example of FIG. 1 , one or more other simulators may be utilized, additionally or alternatively.

[0036] As an example, a platform, such as, for example, the LUMI platform (SLB, Houston, Texas) may be utilized. The LUMI platform includes features that provide for artificial intelligence solutions as may be integrated with data management capabilities. The LUMI platform provides for flexible deployment options and an open, secure, and modular architecture, for example, to empower data-driven decisionmaking. The LUMI platform is operable with the DELFI environment and, hence, one or more of various frameworks. While various platforms, environments, frameworks, libraries, etc., are mentioned, a wellsite pump system framework may be operable in an agnostic manner, for example, to be compatible with one or more other platforms, environments, frameworks, libraries, technologies, etc.

[0037] In the example of FIG. 1 , the visualization features 123 may be implemented via the workspace framework 110, for example, to perform tasks as associated with one or more of subsurface regions, planning operations, constructing wells and / or surface fluid networks, injecting fluid into a reservoir, and producing fluid from a reservoir.Example Geologic Environment

[0038] FIG. 2 shows an example of a geologic environment 210 that includes reservoirs 211-1 and 211-2, which may be faulted by faults 212-1 and 212-2, an example of a network of equipment 230, an enlarged view of a portion of the network of equipment 230, referred to as network 240, and an example of a system 250. FIG. 2 shows some examples of offshore equipment 214 for oil and gas operations related to the reservoir 21 1-2 and onshore equipment 216 for oil and gas operations related to the reservoir 211 -1 . In the example of Fig 2, the geologic environment 210 can include fluids such as oil (o), water (w) and gas (g), which may be stratified in the reservoirs 211-1 and 211-2.

[0039] In the example of FIG. 2, the equipment 214 and 216 can include one or more of drilling equipment, wireline equipment, production equipment, etc. For example, consider the equipment 214 as including a drilling rig that can drill into a formation to reach a reservoir target where a well can be completed for production of hydrocarbons. As an example, the equipment 216 can include production equipment such as wellheads, valves, pump equipment, gas handling equipment, etc. As an example, one or more features of the system 100 of FIG. 1 may be utilized for operations in the geologic environment 210. For example, consider utilizing a drilling or well plan framework, a drilling execution framework, a production framework, etc., to plan, execute, etc., one or more drilling operations, production operations, etc.

[0040] In FIG. 2, the network 240 can be an example of a relatively small production system network. As shown, the network 240 forms somewhat of a tree like structure where flowlines represent branches (e.g., segments) and junctions represent nodes. As shown in FIG. 2, the network 240 provides for transportation of fluid (e.g., oil, water and / or gas) from well locations along flowlines interconnected at junctions with final delivery at a central processing facility (CPF). Where fluid includes solids(e.g. , sand, etc.), one or more pieces of equipment may provide for solids removal, collection, etc.

[0041] In the example of FIG. 2, various portions of the network 240 may include conduit. For example, consider a perspective view of a geologic environment that includes two conduits which may be a conduit to Mani and a conduit to Man3 in the network 240, where Mani , Man2 and Man3 are manifolds.

[0042] As shown in FIG. 2, the example system 250 includes one or more information storage devices 252, one or more computers 254, one or more networks 260 and instructions 270 (e.g., organized as one or more sets of instructions). As to the one or more computers 254, each computer may include one or more processors (e.g., or processing cores) 256 and memory 258 for storing the instructions 270 (e.g., one or more sets of instructions), for example, executable by at least one of the one or more processors.

[0043] As an example, the instructions 270 can include instructions (e.g., stored in the memory 258) executable by at least one of the one or more processors 256 to instruct the system 250 to perform various actions.

[0044] As an example, various graphics in FIG. 2 may be part of a graphical user interface (GUI) that can be generated using executable instructions that may be executable locally and / or remotely using local and / or remote display devices (e.g., a mobile device, a workstation, etc ).

[0045] In various scenarios, one or more pumps may be utilized to inject fluid into a reservoir and / or to produce fluid from a reservoir. For example, consider artificial-lift technology that may be utilized for movement of fluid. While the term lift is utilized, various artificial-technologies may be operated in a manner to move fluid downhole or to lift fluid uphole. For example, various electric submersible pumps (ESP) may be operable in a lift-mode or in an inject-mode.

[0046] Artificial-lift equipment can add energy to a fluid column in a wellbore with the objective of initiating and / or improving production from a well. Artificial lift systems can utilize a range of operating principles (e.g., rod pumping, gas lift, electric submersible pumps, etc.). As such, artificial lift equipment can operate through utilization of one or more resources (e.g., fuel, electricity, gas, etc.).

[0047] Gas lift is an artificial-lift method in which gas is injected into production tubing to reduce hydrostatic pressure of a fluid column. The resulting reduction inbottomhole pressure allows reservoir liquids to enter a wellbore at a higher flow rate. In gas lift, injection gas can be conveyed down a tubing-casing annulus and enter a production train through a series of gas-lift valves. In such an approach, a gas-lift valve position, operating pressure and gas injection rate may be operational parameters (e.g., determined by specific well conditions, etc.).

[0048] A sucker rod pump is an artificial-lift pumping system that uses a surface power source to drive a downhole pump assembly. For example, a beam and crank assembly can create reciprocating motion in a sucker rod string that connects to a downhole pump assembly. In such an example, the pump can include a plunger and valve assembly to convert the reciprocating motion to vertical fluid movement. As an example, a sucker rod pump may be driven using electricity and / or fuel. For example, a prime mover of a sucker rod pump can be an electric motor or an internal combustion engine.

[0049] An ESP is an artificial-lift system that utilizes a downhole pumping system that is electrically driven. In such an example, the pump can include staged centrifugal pump sections that can be specifically configured to suit production and wellbore characteristics of a given application. ESP systems may provide flexibility over a range of sizes and output flow capacities.

[0050] A PCP is a type of a sucker rod-pumping unit that uses a rotor and a stator. In such an approach, rotation of a rod by means of an electric motor at surface causes fluid contained in a cavity to flow upward. A PCP may be referred to as a rotary positive-displacement unit.

[0051] As an example, one or more sensors may be included in an artificial-lift system. For example, consider a gauge coupled to a downhole end of an ESP where signals from sensors of the gauge can be transmitted to surface equipment using a power cable and / or a dedicated gauge cable. For example, consider the PHOENIX gauge (SLB, Houston, Texas), which includes sensors that can measure intake pressure, temperature, motor oil temperature, winding temperature, vibration, current leakage and / or pump discharge pressure. A gauge may be operatively coupled to a controller, which may, for example, provide controls for backspin of an ESP, sanding of an ESP, flux of an ESP and gas related issues of an ESP. For example, during operation where sand is present (e.g., suspended solid matter, etc.), sand mayaccumulate in one or more stages of an ESP where a control scheme may act to rid the ESP of at least a portion of the sand.

[0052] As an example, a PCP may be suitable for use in production for wells characterized by highly viscous fluid and high sand cut where the PCP has some sandlifting capability. However, sand may accumulate where a control scheme may be utilized to rid the PCP of at least a portion of the sand.

[0053] As an example, a sucker rod pump may be operable via as a stroke- through pump to release sand and other material. In such an example, to minimize damage to a plunger and barrel, a grooved-body plunger may be used to catch and carry the sand away from those components.Example ESP System

[0054] FIG. 3 shows an example of an ESP system 300 that includes an ESP 310 as an example of equipment that may be placed in a geologic environment. As an example, an ESP may be expected to function in an environment over an extended period of time (e.g., optionally of the order of years).

[0055] In the example of FIG. 3, the ESP system 300 includes a network 301 , a well 303 disposed in a geologic environment (e.g., with surface equipment, etc.), a power supply 305, the ESP 310, a controller 330, a motor controller 350 and a variable speed drive (VSD) unit 370. The power supply 305 may receive power from a power grid, an onsite generator (e.g., natural gas driven turbine), or other source. The power supply 305 may supply a voltage, for example, of about 4.16 kV.

[0056] As shown, the well 303 includes a wellhead that can include a choke (e.g., a choke valve). For example, the well 303 can include a choke valve to control various operations such as to reduce pressure of a fluid from high pressure in a closed wellbore to atmospheric pressure. A wellhead may include one or more sensors such as a temperature sensor, a pressure sensor, a solids sensor, etc. As an example, a wellhead can include a temperature sensor and a pressure sensor.

[0057] As to the ESP 310, it is shown as including cables 311 (e.g., or a cable), a pump 312, gas handling features 313, a pump intake 314, a motor 315, one or more sensors 316 (e.g., temperature, pressure, strain, current leakage, vibration, etc.) and a protector 317.

[0058] As an example, an ESP may include a REDA HOTLINE high- temperature ESP motor. As an example, an ESP motor can include a three-phase squirrel cage with two-pole induction. As an example, an ESP motor may include steel stator laminations that can help focus magnetic forces on rotors, for example, to help reduce energy loss. As an example, stator windings can include copper and insulation.

[0059] As an example, the one or more sensors 316 of the ESP 310 may be part of a digital downhole monitoring system. For example, consider the PHOENIX MULTISENSOR XT150 system (SLB, Houston, Texas). A monitoring system may include a base unit that operatively couples to an ESP motor (see, e.g., the motor 315), for example, directly, via a motor-base crossover, etc. As an example, such a base unit (e g., base gauge) may measure intake pressure, intake temperature, motor oil temperature, motor winding temperature, vibration, currently leakage, etc. As an example, a base unit may transmit information via a power cable that provides power to an ESP motor and may receive power via such a cable as well.

[0060] As shown in the example of FIG. 3, the one or more sensors 316 can include circuitry 360. As an example, such circuitry 360 can include one or more processors and memory that can store processor-executable instructions. As an example, such instructions can include instructions for one or more monitoring and / or control features. As an example, the circuitry 360 may be utilized as an edge device and / or as part of an edge device (see, e.g., FIG. 5).

[0061] As an example, a remote unit may be provided that may be located at a pump discharge (e.g., located at an end opposite the pump intake 314). As an example, a base unit and a remote unit may, in combination, measure intake and discharge pressures across a pump (see, e.g, the pump 312), for example, for analysis of a pump curve. As an example, alarms may be set for one or more parameters (e.g., measurements, parameters based on measurements, etc.).

[0062] Where a system includes a base unit and a remote unit, such as those of the PHOENIX MULTISENSOR XT150 system, the units may be linked via wires. Such an arrangement provide power from the base unit to the remote unit and allows for communication between the base unit and the remote unit (e.g., at least transmission of information from the remote unit to the base unit). As an example, a remote unit is powered via a wired interface to a base unit such that one or more sensors of the remote unit can sense physical phenomena. In such an example, theremote unit can then transmit sensed information to the base unit, which, in turn, may transmit such information to a surface unit via a power cable configured to provide power to an ESP motor.

[0063] In the example of FIG. 3, the well 303 may include one or more well sensors 320, for example, such as the OPTICLINE sensors or WELLWATCHER BRITEBLUE sensors (SLB, Houston, Texas). Such sensors are fiber-optic based and can provide for real time sensing of temperature, for example, in SAGD or other operations. As shown in the example of FIG. 1 , a well can include a relatively horizontal portion. Such a portion may collect heated heavy oil responsive to steam injection. Measurements of temperature along the length of the well can provide for feedback, for example, to understand conditions downhole of an ESP. Well sensors may extend a considerable distance into a well and possibly beyond a position of an ESP.

[0064] In the example of FIG. 3, the controller 330 can include one or more interfaces, for example, for receipt, transmission or receipt and transmission of information with the motor controller 350, a VSD unit 370, the power supply 305 (e.g., a gas fueled turbine generator, a power company, etc.), the network 301 , equipment in the well 303, equipment in another well, etc.

[0065] As an example, the controller 330 may include features of an ESP motor controller and optionally supplant the ESP motor controller 350. For example, the controller 330 may include features of the INSTRUCT motor controller (SLB, Houston, Texas) and / or features of the UNICONN motor controller (SLB, Houston, Texas), which may connect to a SCADA system, the ESPWATCHER surveillance system (SLB, Houston, Texas), the LIFTWATCHER system (SLB, Houston, Texas), LIFTIQ system (SLB, Houston, Texas), etc. The UNICONN motor controller and / or the INSTRUCT motor controller can perform some control and data acquisition tasks for ESPs, surface pumps or other monitored wells. As an example, a motor controller can interface with the aforementioned PHOENIX monitoring system, for example, to access pressure, temperature and vibration data and various protection parameters as well as to provide direct current power to downhole sensors. As an example, a motor controller can interface with fixed speed drive (FSD) controllers or a VSD unit, for example, such as the VSD unit 370.

[0066] For FSD controllers, a motor controller can monitor ESP system three- phase currents, three-phase surface voltage, supply voltage and frequency, ESP spinning frequency and leg ground, power factor and motor load. For VSD units, a motor controller can monitor VSD output current, ESP running current, VSD output voltage, supply voltage, VSD input and VSD output power, VSD output frequency, drive loading, motor load, three-phase ESP running current, three-phase VSD input or output voltage, ESP spinning frequency, and leg-ground.

[0067] In the example of FIG. 3, the ESP motor controller 350 includes various modules to handle, for example, virtual flow estimations, backspin of an ESP, sanding of an ESP, flux of an ESP, gas issues of an ESP, emulsion presence, emulsion formation, etc. The motor controller 350 may include any of a variety of features, additionally, alternatively, etc.

[0068] In the example of FIG. 3, the VSD unit 370 may be a low voltage drive (LVD) unit, a medium voltage drive (MVD) unit or other type of unit (e.g., a high voltage drive, which may provide a voltage in excess of about 4.16 kV). As an example, the VSD unit 370 may receive power with a voltage of about 4.16 kV and control a motor as a load with a voltage from about 0 V to about 4.16 kV. The VSD unit 370 may include control circuitry such as the SPEEDSTAR MVD control circuitry (SLB, Houston, Texas).Example Cut-Away Views of Equipment

[0069] FIG. 4 shows cut-away views of examples of equipment such as, for example, a portion of a pump 420, a protector 470, a motor 450 of an ESP and a sensor unit 460. In the examples of FIG. 4, each of the pieces of equipment may be considered to be assemblies that, for example, can be operatively coupled to form a system (e.g., an ESP or ESP system). In FIG. 4, the pump 420, the protector 470, the motor 450 and the sensor unit 460 are shown with respect to cylindrical coordinate systems (e.g., r, z, 0). Various features of equipment may be described, defined, etc. with respect to a cylindrical coordinate system. As an example, a lower end of the pump 420 may be coupled to an upper end of the protector 470, a lower end of the protector 470 may be coupled to an upper end of the motor 450 and a lower end of the motor 450 may be coupled to an upper end of the sensor unit 460 (e.g., via a bridge or other suitable coupling).

[0070] As shown in FIG. 4, a shaft segment of the pump 420 may be coupled via a connector to a shaft segment of the protector 470 and the shaft segment of the protector 470 may be coupled via a connector to a shaft segment of the motor 450. As an example, an ESP may be oriented in a desired direction, which may be vertical, horizontal or other angle (e.g., as may be defined with respect to gravity, etc.). Orientation of an ESP with respect to gravity may be considered as a factor, for example, to determine ESP features, operation, etc.

[0071] As shown in FIG. 4, the motor 450 is an electric motor that includes a connector 452, for example, to operatively couple the electric motor to a multiphase power cable, for example, optionally via one or more motor lead extensions. Power supplied to the motor 450 via the connector 452 may be further supplied to the sensor unit 460, for example, via a wye point of the motor 450 (e.g., a wye point of a multiphase motor).

[0072] As an example, a connector may include features to connect one or more transmission lines, optionally dedicated to a monitoring system. For example, the connector 452 may include a socket, a pin, etc., that can couple to a transmission line dedicated to the sensor unit 460. As an example, the sensor unit 460 can include a connector that can connect the sensor unit 460 to a dedicated transmission line or lines, for example, directly and / or indirectly.

[0073] As an example, the motor 450 may include a transmission line jumper that extends from the connector 452 to a connector that can couple to the sensor unit 460. Such a transmission line jumper may be, for example, one or more conductors, twisted conductors, an optical fiber, optical fibers, a waveguide, waveguides, etc. As an example, the motor 450 may include a high-temperature optical material that can transmit information. In such an example, the optical material may couple to one or more optical transmission lines and / or to one or more electrical-to-optical and / or optical-to-electrical signal converters.

[0074] In the examples of FIG. 4, one or more coated electrical conductors may be present. For example, the pump 420 may include one or more coated electrical conductors operatively coupled to and / or part of sensor circuitry and / or another type of circuitry; the protector 470 may include one or more coated electrical conductors operatively coupled to and / or part of sensor circuitry and / or another type of circuitry; the motor 450 may include one or more coated electrical conductors operativelycoupled to and / or part of sensor circuitry, electric motor circuitry and / or another type of circuitry; and the unit 460 may include one or more coated electrical conductors operatively coupled to and / or part of sensor circuitry and / or another type of circuitry.

[0075] In the examples of FIG. 4, the pump 420 can include a housing 424, the protector 470 can include a housing 474, the motor 450 can include a housing 454 and the unit 460 can include a housing 464. In such examples, a housing can include opposing ends, a longitudinal axis, an axial length defined between the opposing ends, a maximum transverse dimension that is less than the length and an interior space. As an example, circuitry may be disposed at least in part in the interior space. As an example, a coated electrical conductor can be electrically coupled to such circuitry where the coated electrical conductor includes an electrical conductor that includes copper and a length defined by opposing ends, a polymeric electrical insulation layer disposed about at least a portion of the length of the electrical conductor, and a barrier layer disposed about at least a portion of the polymeric electrical insulation layer.

[0076] As an example, an interior space of an assembly may be sealed via one or more seal elements, joints, etc. As to the pump 420, the motor 450, the unit 460 and the protector 470 of FIG. 4, these may be individual assemblies that include a coated electrical conductor electrically coupled to circuitry where the coated electrical conductor includes an electrical conductor that includes copper and a length defined by opposing ends, a polymeric electrical insulation layer disposed about at least a portion of the length of the electrical conductor, and a barrier layer disposed about at least a portion of the polymeric electrical insulation layer. As an example, one or more of such assemblies can include one or more sealed interior spaces, for example, consider a housing that includes one or more seal elements, one or more joints, etc. that aim to protect circuitry, etc., in the interior space or spaces from fluid in a downhole environment. As an example, an assembly can include an encapsulant or encapsulating material in an interior space. As an example, an assembly can include a specialized fluid in an interior space (e.g., a dielectric oil, etc.).

[0077] As an example, where water and / or gas (e.g., CO2, H2S, etc.) penetrates a housing and enters an interior space, a coated electrical conductor can include an electrical conductor that includes copper and a length defined by opposing ends, a polymeric electrical insulation layer disposed about at least a portion of the length of the electrical conductor, and a barrier layer disposed about at least a portion of thepolymeric electrical insulation layer where the barrier layer acts to protect the polymeric electrical insulation layer from the water and / or gas. In such an example, the barrier layer may prolong the useful life (e.g., operational life) of an assembly.

[0078] In various instances, an ESP system may experience current leakage, which may result in a ground fault. For example, a ground fault may involve current leakage where electric current supplied to a pump motor leaks to ground. Currently leakage may be due to an insulation issue and / or one or more other types of issues. As an example, intrusion of fluid may result in insulation damage, which may result in current leakage. Depending on type or types of telemetry utilized for one or more sensors, a current leakage issue may impact an ability to transmit sensor data. In various instances, a pump may still be operable upon experiencing current leakage. For example, consider a multiphase cable where a conductor for one phase experiences an issue and where conductors for remaining phases may be sufficient for supplying current to a motor.Example Diagram of a System Including Power Source and Data

[0079] FIG. 5 shows a block diagram of an example of a system 500 that includes a power source 501 as well as data 502 (e.g., information). The power source 501 provides power to a VSD block 570 while the data 502 may be provided to a communication block 530. The data 502 may include instructions, for example, to instruct circuitry of the circuitry block 550, one or more sensors of the sensor block 560, etc. The data 502 may be or include data communicated, for example, from the circuitry block 550, the sensor block 560, etc. In the example of FIG. 5, a choke block 540 can provide for transmission of data signals via a power cable 511 (e.g., including motor lead extensions “MLEs”). A power cable may be provided in a format such as a round format or a flat format with multiple conductors. MLEs may be spliced onto a power cable to allow each of the conductors to physically connect to an appropriate corresponding connector of an electric motor (see, e.g., the connector 452 of FIG. 4). As an example, MLEs may be bundled within an outer casing (e.g., a layer of armor, etc.).

[0080] As shown, the power cable 511 connects to a motor block 515, which may be a motor (or motors) of an ESP and be controllable via the VSD block 570. In the example of FIG. 5, the conductors of the power cable 511 electrically connect at awye point 525. The circuitry block 550 may derive power via the wye point 525 and may optionally transmit, receive or transmit and receive data via the wye point 525. As shown, the circuitry block 550 may be grounded.

[0081] As explained, the sensor block 560 may be operatively coupled to the circuitry block 550 for supply of power and / or for telemetry (e.g., receipt of instructions, transmission of sensor data, etc.). In a scenario where current leakage occurs (e.g., ground fault, etc.), depending on characteristics of circuitry block 550 and / or the sensor block 560, operation and / or telemetry may be interrupted. As mentioned, an ESP system may include one or more memory devices that may be configured to store sensor data that may be later accessed (e.g. , at surface, upon restoration of telemetry, etc.).Example System

[0082] FIG. 6 shows an example of a system 600 and an example of an architecture 601 where the system 600 can include various local components that can be in communication with one or more remote components. As shown in the example of FIG. 6, the architecture 601 can provide for one or more security components 602, one or more machine learning models 603, data 604, objects 605, detection techniques 606 (e.g., classification, regression, recognition, prediction, etc.), analysis techniques 607 and output(s) 608. As an example, the system 600 may be operatively coupled to one or more pumps, which can include one or more ESPs. As an example, the system 600 may operate as a controller, a motor controller, etc., and / or provide information to a controller, a motor controller, etc.

[0083] As shown, the system 600 can include a power source 613 (e.g., solar, generator, batter, grid, etc.) that can provide power to an edge framework gateway 610 that can include one or more computing cores 612 and one or more media interfaces 614 that can, for example, receive a computer-readable medium 640 that may include one or more data structures such as an operating system (OS) image 642, a framework 644 and data 646. In such an example, the OS image 642 may cause one or more of the one or more cores 612 to establish an operating system environment that is suitable for execution of one or more applications. For example, the framework 644 may be an application suitable for execution in an established operating system in the edge framework gateway 610.

[0084] In the example of FIG. 6, the edge framework gateway 610 (“EF”) can include one or more types of interfaces suitable for receipt and / or transmission of information. For example, consider one or more wireless interfaces that may provide for local communications at a site such as to one or more pieces of local equipment, which can include equipment 632, equipment 634 and equipment 636 and / or remote communications to one or more remote sites 652 and 654. In such an example, lesser or more equipment may be included.

[0085] As an example, the circuitry block 550 and / or the sensor block 560 of the example of FIG. 5 can be utilized as an edge device and / or as part of an edge device. As an example, the circuitry block 550 can include and / or host a framework such as the framework 644. As an example, the circuitry block 550 can include and / or host containerized instructions. As an example, the circuitry block 550 may be operatively coupled to one or more pieces of surface equipment such as, for example, the edge framework gateway 610 of FIG. 6. As an example, an ESP may be equipped with its own edge computing resources that can, at least in part, operate downhole for monitoring and / or control of the ESP. In various examples, one or more downhole sensors may acquire one or more pressures, one or more temperatures, a drive frequency, etc., which may be inputs to one or more models, monitoring and / or control components, etc. As an example, the equipment 632, 634 and 636 may include one or more types of equipment. As an example, equipment may include non-artificial lift equipment and / or artificial lift equipment.

[0086] As an example, the edge framework gateway 610 may be installed at a site where the site is some distance from a city, a town, etc. In such an example, the edge framework gateway 610 may be accessible via a satellite communication network and / or one or more other networks where data, control instructions, etc., may be transmitted, received, etc.

[0087] As an example, one or more pieces of equipment at a site may be controllable locally and / or remotely. For example, a local controller may be an edge framework-based controller that can issue control instructions to local equipment via a local network and a remote controller may be a cloud-based controller or other type of remote controller that can issue control instructions to local equipment via one or more networks that reach beyond the site. As an example, a site may include features for implementation of local and / or remote control. As an example, a controller mayinclude an architecture such as a supervisory control and data acquisition (SCADA) architecture.

[0088] Satellite communication tends to be slower and more costly than other types of electronic communication due to factors such as distance, equipment, deployment and maintenance. For wellsites that do not have other forms of communication, satellite communication can be limiting in one or more aspects. For example, where a controller is to operate in real-time or near real-time, a cloud-based approach to control may introduce too much latency.

[0089] As shown in the example of FIG. 6, the edge framework gateway 610 may be deployed where it can operate locally with the one or more pieces of equipment 632, 634 and 636, etc. As an example, the edge framework gateway 610 may include switching and / or communication capabilities, for example, for information transmission between equipment, etc.

[0090] As desired, from time to time, communication may occur between the edge framework gateway 610 and one or more remote sites 652, 654, etc., which may be via satellite communication where latency and costs are tolerable. As an example, the CRM 640 may be a removable drive that can be brought to a site via one or more modes of transport. For example, consider an air drop, a human via helicopter, plane, boat, etc.

[0091] As explained with respect to FIG. 6, a framework may execute within a gateway such as, for example, an AGORA gateway (e.g., consider one or more processors, memory, etc., which may be deployed as a “box” that can be locally powered and that can communicate locally with other equipment via one or more interfaces). As an example, one or more pieces of equipment may include computational resources that can be akin to those of an AGORA gateway or more or less than those of an AGORA gateway. As an example, an AGORA gateway may be a network device with various networking capabilities.

[0092] As an example, a gateway can include one or more features of an AGORA gateway (e.g., v.202, v.402, etc.) and / or another gateway. For example, consider features such as an INTEL ATOM E3930 or E3950 dual core with DRAM and an eMMC and / or SSD. Such a gateway may include a trusted platform module (TPM), which can provide for secure and measured boot support (e.g., via hashes, etc.). A gateway may include one or more interfaces (e.g., Ethernet, RS485 / 422, RS232, etc.).As to power, a gateway may consume less than about 100 W (e.g., consider less than 10 W or less than 20 W). As an example, a gateway may include an operating system (e.g., consider LINUX DEBIAN LTS or another operating system). As an example, a gateway may include a cellular interface (e.g., 4G LTE with global modem / GPS, 5G, etc.). As an example, a gateway may include a WIFI interface (e.g., 802.11 a / b / g / n). As an example, a gateway may be operable using AC 100-240 V, 50 / 60 Hz or 24 VDC. As to dimensions, consider a gateway that has a protective box with dimensions of approximately 10 in x 8 in x 4 in (e.g., 25 cm x 20.3 cm x 10.1 cm).

[0093] As an example, a system may include one or more components, features, etc., of a SENSIA system (SENSIA LLC, Houston, Texas), such as, for example, the SENSIA AVALON lift surveillance application and associated hardware. For example, such a system may include one or more edge gateways that may be operatively coupled to a surveillance and / or control system. As an example, a system may provide for implementation of Artificial Intelligence Response Prioritization (AiRP) for one or more pumps, which may include one or more ESPs. As an example, one or more machine learning models, which may include one or more trained and / or trainable machine learning models, may be implemented for detection of one or more unwanted events. In such an example, a system may provide for handling a number of wells for instantaneous detection of one or more types of issues. In such an example, a system may be operable to help prevent premature ESP failures field-wide while, for example, reducing or eliminating false alarms.

[0094] As an example, a system may include one or more features of the QRATE HCC2 controller (SENSIA LLC), which may include a dedicated ARM microcontroller, embedded I / O, serial communications unit(s), Ethernet unit(s), one or more serial ports, one or more GPS units (e.g., consider a GNSS receiver for time synchronization), a video port (e.g., consider an HDMI port for edge application touch interface, etc.), one or more wireless option features, one or more modems (e.g., 5G, 4G LTE, WIFI, etc.), firmware, operating system, etc. As an example, such a controller may provide for running DOCKER containers (Docker, Inc, Palo Alto, California), etc.

[0095] As an example, a system may include one or more VSDs that may include one or more integrated components for implementation of one or more machine learning technologies. For example, consider a VSD that includes a machine learning processing component operatively coupled to one or more sensor and / or dataacquisition components whereby the machine learning processing component may provide for local and / or remote machine learning implementations, which, in turn, may provide for VSD control to control one or more electric motors of a pump or pumps. As to data acquisition, consider acquisition of data from one or more of the types of equipment as shown in the example of FIG. 3. As explained, sensor data may include ESP sensor data (e.g., from a gauge, etc.) and / or well-based sensor data (e.g., from fiber-optic sensors, wellhead sensors, etc.). As an example, a VSD unit may include one or more types of sensors that may provide for generating sensor data as to one or more aspects of power supply, power quality, power utilization, data transmission via power cable(s), etc. For example, consider a VSD unit with an integrated or addon data sensing and acquisition assembly that may provide for acquiring sensor data and transmitting such sensor data to other hardware, which may be processor-based hardware configured to execute one or more machine learning models to generate output that may be relevant to pump performance, VSD unit performance, well performance, environmental conditions, reservoir performance, etc. Such output may, for example, provide for improved control of one or more pieces of field equipment, as may relate to production and / or injection of fluid from or into a well, respectively.

[0096] As an example, a system may provide for utilization of polling frequencies, such as, for example, on a minute-by-minute basis, which may provide for assessing and controlling ESP equipment and / or other well equipment responsive to dynamically changing operating environments.

[0097] As an example, a system may provide for hybrid pump monitoring and control. For example, consider a field that may include ESPs and one or more other types of pumps, such as, for example, sucker rod pumps (SRPs), progressing cavity pumps, (PCPs), etc.

[0098] As an example, a system may provide an architecture that may be secure, flexible, extensible, etc., which may, for example, provide for operatively coupling to one or more other frameworks. As an example, a system may provide for utilization of one or more types of communication protocols, messaging protocols, etc. As an example, one or more application programming interfaces (APIs) may be implemented, which may provide for issuance of API calls and responses thereto. For example, consider a local piece of equipment that may acquire sensor data and make an API call that includes a portion of the sensor data and / or one or more types of dataderived therefrom. In such an example, the local piece of equipment may receive a response to the API call that may include one or more executable instructions, commands, etc., to thereby cause the piece of equipment to perform an action and / or not perform an action. For example, consider a pump related action, which may, for example, aim to improve survival of the pump responsive to acquisition of sensor data that may indicate an issue that may impact pump survival (e.g., longevity, service schedule, etc.).

[0099] As an example, security within a system may be achieved using one or more techniques, such as, for example, one or more encryption techniques; noting that one or more secure communication channels may be utilized, which may be dedicated and thereby provide some assurances as to security. As an example, a system may provide data back-fill, which may be beneficial in various locations that may experience connectivity issues. As an example, a system may provide for redundancy, which may be at double or triple or more.

[0100] As an example, a gateway may be part of a drone. For example, consider a mobile gateway that can take off and land where it may land to operatively couple with equipment to thereby provide for control of such equipment. In such an example, the equipment may include a landing pad. For example, a drone may be directed to a landing pad where it can interact with equipment to control the equipment. As an example, a wellhead can include a landing pad where the wellhead can include one or more sensors (e.g., temperature and pressure) and where a mobile gateway can include features for generating fluid flow values using information from the one or more sensors. In such an example, the mobile gateway may issue one or more control instructions (e.g., to a choke, a pump, etc ).

[0101] As an example, a gateway itself may include one or more cameras such that the gateway can record conditions. For example, consider a motion detection camera that can detect the presence of an object. In such an example, an image of the object and / or an analysis (e.g., image recognition) signal thereof may be transmitted (e.g., via a satellite communication link) such that a risk may be assessed at a site that is distant from the gateway.

[0102] As an example, a gateway may include one or more accelerometers, gyroscopes, etc. As an example, a gateway may include circuitry that can performseismic sensing that indicates ground movements. Such circuitry may be suitable for detecting and recording equipment movements and / or movement of the gateway itself.

[0103] As explained, a gateway can include features that enhance its operation at a remote site that may be distant from a city, a town, etc., such that travel to the site and / or communication with equipment at the site is problematic and / or costly. As explained, a gateway can include an operating system and memory that can store one or more types of applications that may be executable in an operating system environment. Such applications can include one or more security applications, one or more control applications, one or more simulation applications, etc.

[0104] As an example, various types of data may be available, for example, consider real-time data from equipment and ad hoc data. In various examples, data from sources connected to a gateway may be real-time, ad hoc data, sporadic data, etc. As an example, lab test data may be available that can be used to fine tune one or more models (e.g., locally, etc.). As an example, data from a framework such as the AVOCET framework may be utilized where results and / or data thereof can be sent to the edge. As an example, one or more types of ad hoc data may be stored in a database and sent to the edge.

[0105] As to real-time data, it can include data that are acquired via one or more sensors at a site and then transmitted after acquisition, for example, to a framework, which may be local, remote or part local and part remote. Such transmissions may be as streams (e.g., streaming data) and / or as batches. As to batches, a buffer may be utilized where an amount of data may be stored and then transmitted as a batch. In various instances, real-time data may be characterized using a sampling rate or sampling frequency. For example, consider 1 Hz as a sampling frequency that is adequate to track various types of physical phenomena that can occur during well operations. As an example, a sensor and / or a framework may provide for adjustment of sampling (e.g., at the sensor and / or at the framework). In various instances, data from multiple sensors may be at the same sampling rate or at one or more sampling rates. As an example, data sampling can be at a rate sufficient to provide for detection, prediction, etc., as to a probability of occurrence of an event at a future time. In such an example, the sooner data are analyzed, the sooner such detection, prediction, etc., can occur. For example, consider a system where advance notice of a risk of an eventcan be greater than 10 minutes, greater than 30 minutes, greater than 1 hour, etc., such that one or more control actions can be taken to mitigate the risk of the event.

[0106] As explained, various systems may operate in a local manner, optionally without access to a network such as the Internet. For example, a site may be relatively remote where satellite communication exists as a main mode of communication, which may be costly and / or low bandwidth. In such scenarios, security may resort to local features rather than a remote feature such as a remote authentication server.

[0107] An authentication server can provide a network service that applications use to authenticate credentials, which may be or include account names and passwords of users (e.g., human and / or machine). When a client submits a valid credential or credentials to an authentication server, the authentication server can generate a cryptographic ticket that the client can subsequently use to access one or more services.

[0108] In the example of FIG. 6, the framework 644 can be an edge-enabled data processing framework. As an example, such a framework can include features to perform one or more of the followings tasks: real-time data cleansing to synchronize information from existing well metrology (e.g., wellhead, tubing, flow, ESP, etc.); executing one or more machine learning (e.g., including self-learning) models in realtime (e.g., one or more ML models that can identify one or more issues, etc.); and conveying a control set point and / or another instruction to a controller (e.g., an actuatable valve, etc.) and / or one or more other pieces of equipment. As mentioned, an edge framework may be deployable using downhole circuitry (see, e.g., the circuitry block 550 of FIG. 5), which may be downhole circuitry operatively coupled to surface circuitry, etc.

[0109] The system 600 can be part of an infrastructure that serves as a secure gateway to transmit surveillance into an operator’s surveillance station or its own surveillance platform. The presence of such a gateway can also support an operator for introduction of one or more additional HOT (industrial internet of things) implementations.

[0110] As explained, an ESP can be implemented at a site for pumping fluid, whether for injection and / or production. For example, an ESP may be utilized in a stimulation treatment to inject fluid that includes various chemicals and an ESP may be utilized as an artificial lift technology to assist production of fluid from a reservoir.

[0111] As ESPs find various uses in various environments, knowledge as to operation, performance, etc., can be spread amongst various domains where each domain may have its own experts. One type of issue that can arise in ESP operation pertains to presence of an emulsion, which may be formed prior to an inlet to an ESP, near an inlet to an ESP and / or within an ESP. An emulsion is a mixture of two or more fluids that can be immiscible (e.g., unmixable or unblendable) owing to liquid-liquid phase separation where phases include a dispersed phase and a continuous phase. For example, consider an oil in water emulsion where oil is dispersed in a continuous water phase and a water in oil emulsion where water is dispersed in a continuous oil phase. Emulsions can form, be stabilized, and / or be destabilized via mechanical, thermal, pressure and / or chemical means. For example, mechanical mixing can result in emulsion formation where stability of the emulsion can depend on factors such as temperature, pressure, surface active agents (e.g., surfactants), salt concentrations, etc.

[0112] Another type of issue that can arise in ESP operation relates to gas, which can be ingested by an ESP and degrade ESP pumping performance. Gas issues can include, for example, gas degradation and gas lock, which can be due to various types of phenomena, behaviors, etc. Gas degradation may be evident in time series data as one or more signatures which may be prior to or include a low or no flow period. As to gas lock, it is a condition in pumping and processing equipment caused by the induction of free gas where compressible gas can interfere with proper operation of valves and / or other pump components, which may be preventing intake of fluid, pumping of fluid, etc.

[0113] Yet another type of issue that can arise in ESP operation relates to scaling. For example, scale can buildup on one or more components of an ESP, such as, for example, on a motor housing, which may affect cooling efficiency and raise intake pressure on the ESP. For example, a trend associated with scaling may include a rise in intake pressure, a rise in motor temperature, and a drop in current. In various scenarios, if scale continues to build unchecked, threatening events such as high pressure or motor overheating can occur, in which case an ESP controller may shuts down the well as a protective measure. When an ESP is not operating properly in a production scenario, production can suffer, which may result in nonproductive time (NPT) and make it challenging to meet production goals. If an ESP needs to be pulledout of hole and replaced or repaired and run back in hole, additional NPT can result, which may be substantial for an ESP that may be hundreds of meters or a thousand meters or more in a hole (e.g., a wellbore). In general, unnecessary downtime begins as soon as an ESP shuts down. In various instances, one or more types of sensors and associated circuitry may store information locally that can be retrieved once an ESP is pulled to surface. For example, for various installations, the purpose of a downhole gauge may be to enable diagnostics of a shutdown cause after occurrence of an event. In such an example, an operator may mobilize a field crew, download sensor data from the gauge, analyze well performance, and begin remediation, which is a process that may last a number of days.

[0114] As an example, a framework may provide for detection of conditions indicative of scaling and / or one or more other types of issues in advance of a need for shutdown. For example, consider a ML model trained to detect signatures in one or more types of time series data prior to a scaling event that demands shutdown. In such an example, the ML model may be a trained classification model or classifier. As to types of time series data, as mentioned, scaling may affect intake pressure, motor temperature, and current. Thus, one or more of pressure sensor, temperature sensor, and current sensor data, where available, may be utilized by a trained classifier to detect scaling before a scaling event occurs (e.g., a shutdown due to scaling). As an example, one or more types of sensors may be utilized, which may include types of sensors for sensing pressure, temperature, flow, viscosity, density, fluid composition (e.g., liquid, gas, solids, etc.), rotational speed, reciprocating speed (e.g., for a linear motor), drive frequency, vibration, shock, wear, proximity of components, etc. As an example, a sensor may be part of a pump system or part of another system (e.g., fiberbased sensor disposed in a well, a wellhead sensor, etc.).

[0115] As an example, a fiber-optic-based sensor system may provide for sensing one or more types of phenomena. For example, consider thermal (e.g., temperature), flow, equipment vibration, seismic activity, etc. As an example, one or more of such phenomena may be sensed where data may be utilized for one or more purposes, such as, for example, training a machine learning model, implementing a machine learning model, assessing a pump, controlling a pump, decision-making, etc.

[0116] Issues such as, for example, emulsion formation issues, gas issues, and scaling issues, may occur at various levels of frequency and cause damage to long-term lifespan of equipment as well as reduction in fluid movement (e.g., injection or production). Existing operational workflows to detect occurrence of such issues tend to be reactive and can be susceptible to human error. As an example, a system can include one or more ML models that can be trained such that they learn behaviors that may be exhibited in one or more types of data where, upon detection of a learned behavior or learned behaviors, a trained ML model can output a result, which may be an indicator as to the likelihood of an issue or issues arising (or not arising) during operation of equipment such as an ESP. For example, a trained ML model can receive data as input and output a likelihood of occurrence of a particular issue or issues, where such output may indicate a time frame associated with the likelihood (e.g., or likelihoods with respect to time, etc.).

[0117] As an example, a method can include building one or more ML models that can detect if flow constraints like emulsion formation, gas degradation, or scaling are developing at a pump in real-time by utilizing suitable ESP sensor data. In such an example, an ML model can be embedded in a data science workflow or workflows. As an example, a system can provide for output generation, which may be directed to a controller, controllers, a dashboard, a network interface, etc. As an example, a system can output one or more alarms, which may be directed to humans and / or machines such that one or more control decisions can be taken, within an appropriate time frame (e.g., in real-time, prior to prediction of occurrence, etc.), which can thereby help to prolong pump life and improve pump operation (e.g., injection, production, etc.). While emulsion formation, gas degradation, and scaling are mentioned, one or more other issues may be detectable, such as, for example, current leakage. As an example, a current leakage issue may be detectable through one or more types of data, such as, for example, telemetry data, power efficiency data, etc. For example, intermittent issues in telemetry may indicate an imbalance at a wye point of an ESP motor that may interfere with telemetry. As an example, such intermittent issues may be a result of insulation degradation, etc., which, may become more frequent leading up to a complete failure of one or more phases of an ESP cable (e.g., ground faulting, etc.).

[0118] In machine learning, data are required, which can include actual data and / or synthetic data. In supervised learning, data can be labeled and referred to as labeled data. In unsupervised learning, data may be labeled and / or unlabeled. As totraining a ML model, data may be split into one or more groups, which can include training data and testing data. For example, a portion of a dataset can be utilized for training to generate a trained ML model that can then be tested using another portion of the dataset.

[0119] In machine learning, overfitting and underfitting can cause poor performance of a machine learning model. In statistics, a fit can refer to how well a target function is approximated. In supervised machine learning, training aims to have a ML model approximate an unknown underlying mapping function for output variables given input variables. Statistics can be utilized to describe goodness of fit which refers to measures used to estimate how well the approximation of the function matches the target function. In machine learning, training can include calculating residual errors, etc.; however, some statistical techniques do not readily as the form of a target function to be approximate may not be known. For example, machine learning can be utilized to train a ML model to approximate an unknown function or functions, which can be referred to as a behavior or behaviors. In various instances, a behavior of equipment interacting with fluid can be challenging to elaborate using a physics-based approach; whereas, through machine learning with appropriate data, a ML model can be trained to learn and model that behavior, at times without a detailed understanding of physics underlying the behavior. As explained, various physics-based simulators can be utilized to simulate behavior; however, these can demand substantial computational resources, which can include large matrixes that can be of orders that can challenge even sophisticated simulators.

[0120] As an example, a trained ML model may be built that is relatively lightweight for implementation locally such as via an edge framework. In such an example, local monitoring and / or control may be improved when compared to an approach that relies on sophisticated, physics-based simulation as performed using highly parallelized computational resources.

[0121] As mentioned, however, overfitting and underfitting can occur during machine learning such that a trained ML model is either overfit or underfit. An overfit ML model may lack an ability to handle various real-world scenarios (e.g., scenarios on the margins, scenarios that may occur in certain situations, etc.) while an underfit ML model may lack an ability to robustly generate output for generally occurring real-world scenarios. As training demands data, cleanliness, noise, etc., in data can also be issues that are additional to volume of data.

[0122] As explained, an issue may occur with an expected frequency, which may be, for example, relatively frequent or relatively infrequent. As to relatively frequent issues, sensor data may be more balanced in that over a given span of time, the number of occurrences of an issue may be sufficient to consider time series data over that given span of time to be balanced; whereas, for relatively infrequent issues, that time series data may be unbalanced. Where an issue occurs quite infrequently, it may be considered an anomaly. Various ML techniques may be applied on the basis of data balance. For example, some ML techniques may be suitable for balanced data; whereas, other ML techniques may be suitable for unbalanced data, which may include data that includes one or more anomalies.

[0123] In ML, an approach referred to as classification may attempt to categorize data into different classes. In an unbalanced dataset, one class may make up a large portion of a training dataset (e.g., the majority class), while another class is underrepresented in the training dataset (e.g., a minority class). For example, consider a majority class as corresponding to no scaling or normal operation and a minority class as corresponding to scaling or abnormal operation. As to classification for unbalanced data, a problem can arise with a model trained on the unbalanced data whereby the model learns that it can achieve high accuracy by consistently predicting the majority class, even if recognizing the minority class is equal or more important when applying the model to a real-world scenario.

[0124] As an example, an approach to ML can include processing series data (e.g., time series and / or other series data) prior to training (e.g., learning). For example, consider a method that processes series data to identify events and to then extract a window of the series data that includes data leading up to an identified event. In such an approach, a number of windows may be extracted from series data where the number of windows may be suitable utilized for training and / or testing of one or more ML models.

[0125] As an example, a workflow may employ series homogeneity variationbased event classification. Such an approach can provide for identifying a mechanism of failure present in equipment (e.g., an ESP, etc.) by analyzing time series sensor data using a ML model classifier.

[0126] As an example, a classifier may be selected based on amount of training and / or testing data available and / or based on one or more other criteria (e.g., computational demand, etc.). ML model classifiers may include, for example, perceptron models, naive Bayes models, decision tree models, logistic regression models, k-nearest neighbor (KNN) models, artificial neural network (ANN) models, deep learning (DL) ANN models, support vector machines (SVMs), etc. As an example, a classifier may be implemented using one or more types of ensemble techniques, such as, for example, random forest, bagging, AdaBoost, XGBoost, CATBoost, etc.

[0127] As to generation of suitable training and / or testing data, consider a workflow that applies a standard normal homogeneity test (SNHT) to series data with a tailored windowing technique that is applied through an entire series of the data to identify and extract change event locations inside each window. In such an approach, if the SNHT is positive inside a window, a change event index ID may be used to establish a second window end point. In such an approach, the second window may be used to compute statistical properties of the series of data inside the second window. For example, consider implementing one or more statistical techniques to compute one or more statistical properties, which may be features, as in machine learning feature engineering. As to some examples of statistical properties or features, consider, for example, average, median, standard deviation, minimum, maximum, and slope. These characteristics, as present inside each window, may be used as features for a machine learning model classifier relating these statistical properties with the pump failure modes. As explained, a workflow may provide for correlating features and events, for example, by creating a training dataset with time intervals matching previous known logged events and identified homogeneity changes.

[0128] As explained, an SNHT may be utilized; noting that one or more other types of homogeneity tests may be utilized. As an example, a workflow may include accessing a test that is available in one or more libraries. For example, consider the pyHomgeneity library as written in Python. As an example, a homogeneity test can be a statistical test technique that checks if two or more datasets come from the same distribution or not. For example, in time series data, a homogeneity test may be applied to detect one or more changes in the time series data. In such an example, achange, which may be referred to as a breakpoint, can occurs where the time series data exhibits a change in its distribution.

[0129] As to the pyHomogeneity library or package, it provides various tests that can check the homogeneity of time series data. For example, the pyHomogeneity package can perform the following tests: Pettitt test (pettitt_test); Standard Normal Homogeinity Test (SNHT) Test (snht_test); Buishand Q Test (buishand_q_test); Buishand’s Range Test (buishand_range_test); Buishand’s Likelihood Ration Test (buishand_likelihood_ratio_test); and Buishand’s U Test (buishand_u_test).

[0130] A homogeneity test can include various input parameters. For example, consider one or more of the following input parameters: x (a vector (e.g., a list, a numpy array or a pandas series) data); alpha (significance level (e.g., 0.05 default)); and sim (number of Monte Carlo simulations for p-value computation). A homogeneity test can provide various output parameters. For example, consider a homogeneity test that provides a tuple that can include one or more of the following output parameters: h (e.g., true (if data is nonhomogeneous) or false (if data is homogeneous)); cp (e.g., probable change point location); p (e.g., p value of the significance test); one or more test statistics (e.g., which may depend on test technique); and avg (e.g., mean values at before and after a change point).

[0131] As to the SNHT, an article by Alexandersson is incorporated by reference herein in its entirety (Alexandersson, “A homogeneity test applied to precipitation data”, Journal of Climatology, 6(6), pp. 661-675 (1986) (doi: 10.1002 / joc.3370060607)). The SNHT, as developed by Alexandersson to detect a change in a series of rainfall data, was applied to a series of ratios that compare the observations of rainfall at a measuring station with the average of several stations. The ratios may then be standardized by subtracting their means and dividing by their standard deviations. The SNHT can compute the mean of the data on the previous period and on the following period where the test statistic at each observation may then be computed. As an example, a technique may involve comparing the means of these two periods and normalizing by the standard deviation. As an example, an approach as described in an article by Haimberger may be implemented, which is incorporated by reference herein in its entirety (Haimberger, “Homogenization of radiosonde temperature time series using innovation statistics”, Journal of Climate 20 (7), pp. 1377-1403 (2007)).

[0132] As explained, the SNHT is a statistical test that can be used to detect a change in a time series data. The SNHT is based on the premise that if two time series are homogeneous, then their means should be equal. The SNHT compares the means of data of two time series and computes a test statistic that follows a standard normal distribution under the null hypothesis of homogeneity. The test statistic can then be used to determine the significance level and the p-value of the test.

[0133] As an example, a workflow can include identifying where are statistically significant changes in a series. For example, consider evaluating series variations using the SNHT (e.g., as available via the pyHomogeneity library). In such an example, a series can be divided into constant windows with or without overlapping and where each window (e.g., a homogeneity window scan) is then tested for homogeneity variations. In such an example, windows with a p value less than a particular value (e.g., less than approximately 0.05) may be labeled with a statistically significant difference indicator where an event change ID can be tracked.

[0134] To provide for tracking the most significant changes, the pi and p2 absolute difference may be computed for each window, which may be referred to as abs(A|j). Next, a threshold may be utilized to only track the larger difference. For this case, the threshold used may be a trimmed absolute difference standard deviation per sensor (e.g., consider using percentiles P15-P85).Example Graphics

[0135] FIG. 7 shows various example graphics 710, 720, and 730 associated with an example of a workflow that implements the SNHT. As shown in the graphic 710, constant windows may be subjected to the SNHT for homogeneity. In the example of FIG. 7, the window spans a period of time greater than one day, noting that a window may be sized in accordance with one or more types of behavior that may be associated with one or more issues. As shown in the graphic 720, mean values can be determined along with an absolute value difference, which may be compared to a threshold, to detect a change, denoted by a change event ID (eventID). As shown in the graphic 730, the change event ID may be a point in time that can be utilized to define another window, referred to as a second window or an event window. Such a window can provide a lookback period of time in which series data include indicia of a future event (e.g., as indicated by the change event ID).

[0136] The approach described with respect to FIG. 7, can allow for identification of events that include stronger statistical changes, which, as explained, can be then utilized to create a second window (e.g., an event window). In the example of FIG. 7, the change event ID is utilized as the end of the second window. The objective of the second window can be to select a portion of a series before the identified change as detected by application of the SNHT. These data prior to the change event can represent behavior before a change occurs.

[0137] In the example of FIG. 7, the second window (e.g., the event window) may be utilized for computation of one or more statistical characteristics of a portion of a series before occurrence of a change event. For example, consider computation of one or more of: average, median, standard deviation, minimum, maximum, and slope. For A training dataset, change points can be matched to one or more logged events categories (e.g., as may be determined by an interpreter, etc.) such that the ground truth can be established and utilized for the training of one or more ML models (e.g., one or more classification models). As an example, results of classification of each window may be used to identify one or more detected events and classify them, which may be part of a monitoring operation and / or a control operation.

[0138] As explained, a classifier may be a tree type of classifier. For example, consider one or more of an XGBoost classifier, a random forest classifier, a light GBM classifier, a gradient boosted trees classifier, etc.Example Graphical User Interface

[0139] FIG. 8 shows an example of a graphical user interface (GUI) 800 that includes channels of sensor data with respect to time for a period of months along with event windows for multiple types of events, denoted event type-1 and event type-2. In the example GUI 800, markers are indicated for historical events as identified manually from sensor data logs. As shown in the example of FIG. 8, the homogeneity test approach provides for generation of training and / or testing data for training one or more ML models where such one or more trained ML models can detect a change event prior to its occurrence. As an example, a framework may implement one or more ML models for purposes of monitoring and / or control of field operations in realtime. In such an example, the one or more ML models may be relatively lightweight such that they may be applied expeditiously to provide for real-time performance.

[0140] As explained, a framework may be suitable for detection of change events in advance of occurrence or expected occurrence with respect to an ESP system. As mentioned, an ESP system may include one or more types of sensors and / or one or more other sources of series data. For example, consider a framework that may utilize one or more of pressure data (e.g., intake, outtake, etc.), voltage, current, vibration, temperature, etc., as input for purposes of detecting behaviors indicative of one or more events prior to occurrence or likely occurrence.

[0141] As an example, a framework may provide for early identification and classification of anomalous behavior of series data information acquired by surface and / or downhole sensors in an ESP system that can potentially be addressed in a manner that can help to avoid pump failure and to reduce NPT.

[0142] As explained, a framework may provide for testing homogeneity in a series within various windows and classifying anomalous events with a corresponding failure mode. While various examples are explained with respect to ESPs, one or more other types of field equipment may be monitored and / or controlled via a framework or frameworks. As explained, a framework may provide for integrating anomaly identification using a classification technique (e.g., a classifier). As explained, a framework can provide for identifying anomalies prior a change point (e.g., occurrence of an event, etc.).

[0143] As an example, a method for ML model selection and building can generate a trained ML model suitable for use in various pumping scenarios. Such a ML model may be, for example, a nonparametric ML model (e.g., a decision tree model, etc.). As mentioned, a pump such as an ESP can be utilized in a wide variety of pumping scenarios in locations that can include remote locations and in locations where conditions may change. A combination of remote location and possible changing conditions can be an exceptionally challenging pumping scenario to monitor and control.

[0144] As explained, a lack of adequate and accurate monitoring of pumps can lead to low efficiency, high lifting costs, and frequent repair and replacements. Analysis of sensor data can demand multiple experts to remain alert. Moreover, given a number of wells with a number of pumps, an analysis demands a well-to-well assessment, which may take hours for a human to perform.

[0145] As explained, a ML model-based approach can help to reduce workover time and improve production (e g., or injection) while extending lifespan of equipment. For example, consider a system that includes components that allow for rapid processing of sensor data via use of one or more feature engineering techniques and ML models to identify signatures for one or more issues such as, for example, emulsion formation, gas degradation, scaling, current leakage, etc.

[0146] As explained, a method can involve supervised and / or unsupervised learning. As explained labeled data may be utilized where data can be labeled by domain experts to capture human knowledge and understanding of one or more types of pump issues that may occur in one or more of various scenarios.

[0147] As an example, information may be presented in the form of a dashboard, whether local and / or remote. For example, consider a visual dashboard of current and past issues and / or anomalous activities detected along with additional pump performance indicators. As explained, an alert system can provide for automatically sending notifications and alert messages to one or more destinations responsive to detection of one or more issues, anomalies, etc.Example System

[0148] FIG. 9 shows an example of a system 900 that can be implemented locally at a field site. As shown, the system 900 can be operatively coupled to one or more data sources 904 and 908, which can include pump data and / or sporadic data. As shown, an edge device 910 (e.g., an edge gateway, etc.) can include components germane to monitoring and / or control of pump equipment. For example, the edge device 910 can include a pump suite component 920 and a classifier component 940. The edge device 910 can provide for generation of information regarding operation of an ESP where such information can include information as to scaling 922, emulsion 924, gas degradation (GD) 926, and gas lock (GLK) 928. As shown in the example of FIG. 9, the system may also provide information as to wear 961 , productivity index (PI) drop 962, leakage 963, operational condition 964, and efficiency 965. In FIG. 9, the blocks 922, 924, 926, 928, 961 , 962, 963, 964, and 965 can be alarm and / or control blocks that provide for issuance of one or more of an alarm and a control signal (e.g., a control command, etc.).

[0149] In the example of FIG. 9, the pump suite 920 can be a suite of specialized components for real-time ESP alarms and / or control, which may be enabled through data analytics and an edge-based framework. For example, the system 900 can provide for continuous monitoring of one or more ESPs to help ensure optimal pump working conditions, which can help to avoid deferred oil production. As wells around the world age and experience decreased production (e.g., due to decreasing reservoir pressure, etc.), artificial lift can be increasingly utilized to assist production from such wells. With the increased population of ESPs deployed worldwide, the system 900 can be a comprehensive alarm triggering and / or control system that is part of an oilfield production surveillance strategy. As explained, issues may be identified using data analytics, which can generate one or more types of models for issue detection. In such an approach, one or more physics-based simulators may be utilized in conjunction with one or more of such models. For example, consider a relatively lightweight simulator that can be deployed at the edge where such a light-weight simulator may include a honed-down version of an actual simulator and / or one or more proxies (e.g., trained ML model proxies, regression model proxies, etc.).

[0150] As an example, a suite of components may include, for example, one or more of a gas degradation component, a GLK component, an emulsion component, a scaling component, a pump upthrusting / downthrusting component, a pump wear component, a productivity index (PI) drop component, a tubing leak component, and a current leak component. In such an example, each component can be tailored to target a specific potential suboptimal pump working condition.

[0151] Various features can provide for early-stage detection of anomalies in operation of a pump and allow a human and / or a machine (e.g., a controller) a sufficient amount of time to perform one or more control actions, which may help to reduce unnecessary shutdowns (e.g., non-productive time, etc.).Example Framework

[0152] FIG. 10 shows an example of a framework 1000 that includes various components that may be implemented to generate one or more outputs as to equipment condition. For example, the framework 1000 may generate output for a number of pieces of equipment or equipment systems where each output may be specific to condition of a particular piece of equipment or condition of a particularequipment system. In such an example, output may be specific to an ESP or another type of pump, whether as an individual piece of pump equipment or a pump system.

[0153] As an example, the framework 1000 may provide for model training utilizing data for a number of pieces of equipment and / or equipment systems. For example, consider model training that is performed using historical data for a number of pumps installed in a field where the number of pumps may be greater than three, greater than 10, greater than 20, etc. In such an example, input to a trained model may be specific to a particular one of the pumps such that output is generated that is specific to that particular pump.

[0154] As an example, a framework such as, for example, the framework 1000 may provide for estimating a survival function of an ESP based on a multidimensional feature vector where the framework 1000 may be trained using data for a number of ESPs and where the output may be generated for any chosen ESP. Hence, training may be for many and output may be for any particular one.

[0155] As shown in the example of FIG. 10, the framework 1000 can include a number of inputs such as, for example, an event log dataset 1022, sensor data 1024 from one or more types of sensors, output of one or more anomaly detection selfregressor models 1026, output as to one or more SNHT events 1028, and output as to one or more rule-based alerts 1030.

[0156] As to an anomaly detection self-regressor model (e.g., a regression model), consider using one or more of the techniques described in a published U.S. patent application having a Publication No. US 2024 / 0003242 A1 , as published 4 January 2024, entitled “Field Pump Equipment System”, which is incorporated by reference herein in its entirety and referred to as the ‘242 application. For example, consider a remaining useful life (RUL) model and / or one or more other machine learning (ML) models that may provide for generation of a prediction responsive to receipt of input. As an example, a model may be a neural network model, a tree model, or another type of model. As an example, by training a ML model using normal behavior data, it may not suitably reconstruct abnormal behavior data (e.g., anomalous behavior data), which can thereby be an indication that abnormal (e.g., anomalous) behavior exists upon receipt of particular input data. As an example, reconstruction error can be utilized as a metric to quantitatively and / or qualitatively assess a trainedML model’s ability and inability to reconstruct input, which can be utilized as a proxy for the existence of abnormal (e.g., anomalous) behavior.

[0157] As an example, an unsupervised isolation forest model may be utilized to directly detect anomalies using isolation (e.g., how far a data point is to the rest of the data). Such an approach may run in a linear time complexity akin to distance- related models such as k-nearest neighbors (KNN), which may also be utilized for anomaly detection. An isolation forest can provide for pivoting on attributes of an outlier such as there will be few outliers and that outliers will be different characteristically than non-outliers. An isolation forest can introduce an ensemble of binary trees that recursively generate partitions by randomly selecting a feature and then randomly selecting a split value for the feature. The partitioning process can continue until it separates data points from the rest of the samples. In an isolation forest, an outlier can be expected to demand fewer partitions on average to get isolated compared to normal samples. Each data point can then receive a score based on how easily they are isolated after a number of rounds such that data points that have abnormal scores can be detected as anomalies.

[0158] As an example, a random survival forest (RSF) may be utilized in a system where a workflow can include computing input features using both normal model deviations and input features from equipment data (e.g., pump data, etc.) and metadata and building RUL model(s) using a RSF for survival curve generation.

[0159] As an example, a time-dependent Cox model may be utilized for purposes of survival (e.g., RUL). A time-dependent Cox model can be a timedependent Cox regression model (TDCM), which quantifies the effect of repeated measures of covariates in an analysis of time to event data. As an example, one or more of a pooled logistic regression model (PLRM), a cross sectional pooling model (CSPM), a Kaplan-Meier (KM) survival model and a log-rank test model may be implemented (e.g., for output, for comparisons, for additional output, etc.). As an example, a survival model that accounts (e.g., statistically) for times at which time dependent covariates are measured may provide more reliable estimates compared to an unadjusted approach.

[0160] Referring again to the example framework 1000 of FIG. 10, input may be received for a particular period of time, which may be of the order of days, weeks, etc. For example, the framework 1000 may include a window component 1040 thatprovides for receiving input over a particular window or number of windows (e.g., consider a number of weekly windows such as 1 week to 10 weeks or more). As shown, the window component 1040 can provide output to a multi-task logistic regression (MTLR) model component 1060.

[0161] As mentioned, remaining useful life (RUL) types of models may provide for output curves as to survival. Such types of models find use in fields such as medical prognosis. For example, coefficients in a Cox regression model relate to hazard where a positive coefficient may indicate a worse prognosis and a negative coefficient may indicate a protective effect of the variable with which it is associated. In the context of medical prognosis, an MTLR model differs from the Cox regression model in that the MTLR model may provide for improved output as to particular individuals.

[0162] An article by Yu et al., “Learning Patient-Specific Cancer Survival Distributions as a Sequence of Dependent Regressors”, NIPS’11 : Proceedings of the 24th International Conference on Neural Information Processing Systems, December 2011 , pp. 1845-1853, is incorporated by reference herein in its entirety. The article by Yu et al., describes a local regression method for learning patient-specific survival time distribution based on patient attributes such as blood tests and clinical assessments that generates survival time predictions using patient-specific attributes rather than using cancer site and stage only. In particular, the article by Yu et al., describes an MTLR model that directly models survival function rather than the hazard function (conditional rate of failure / death) where, by modeling a survival distribution as the joint output of a sequence of dependent local regressors, the MTLR model can capture time-varying effects of features.

[0163] As an example, a framework may provide for early decision-making when replacing one or more types of equipment (e.g., ESPs, etc.) by predicting behavior at a the time before a failure occurs. As explained, a ML-based approach may be implemented using one or more data-driven ML models such as, for example, one or more survival analysis models. In such an approach, a framework may utilize a specific set of parameters to feed one or more models. As an example, one or more techniques may be utilized to evaluate model performance, for example, consider evaluation of model performance based on a set of classification metrics andcomparison with a base model such as, for example, a Kaplan-Meier model (e.g., a KM estimator).

[0164] As explained, an MTLR model may be used to approximate a survival function where, for example, data to train the MTLR model may include windowed data such as, for example, a number of weekly windows. In an example trial, seven weekly windows were utilized as an endpoint that included the average and the first-order derivative of sensor data from pumps along with a sum of alerts as output from three framework components, which are shown in the example framework 1000 of FIG. 10 as anomaly detection alerts from the anomaly detection self-regressor model(s) 1026, the SNHT events 1028 (e.g., SNHT anomalous events), and a number of rule-based alerts 1030.

[0165] Survival analysis is a type of regression problem (e.g., to predict a continuous value) but with a particular difference from traditional regression as to the fact that parts of the training data may only be partially observed, which may be deemed as training data that are censored.

[0166] A blog post by Polsterl, “Survival Analysis for Deep Learning”, July 2019, (https: / / k-d-w.org / blog / 2019 / 07 / survival-analysis-for-deep-learning / ), is incorporated by reference herein in its entirety. The blog post describes use of generated censored data and an estimate of a survival function to see what the risk scores actually mean in terms of survival. The blog post describes building, training, and evaluating a convolutional neural network for survival analysis on a particular dataset. In particular, the blog post describes stratifying the training data by class label, and estimating the corresponding survival function using the non-parametric Kaplan-Meier estimator.

[0167] As described in the Polsterl blog post, censored data may be generated and utilized. As an example, a framework may model survival time as a continuous non-negative random variable T, from which the survival function can be derived. As an example, a survival function S(t) may be defined as a basic quantity for time-to- event that returns a probability of survival (e.g., consider ESP survival, etc.) beyond time t, which may be defined as S(t) = P(T > t).Example Plot Using Kaplan-Meier Model

[0168] FIG. 11 shows an example of a plot 1100 of estimated probability of survival versus time in days with a vertical line indicating a current time (e.g., “today”),a line indicating a base model survival probability and an envelope indicating a base model confidence interval. The base model survival probability line is for the Kaplan- Meier model, which is an empirical function. As an example, a base model may be utilized to measure behavior of another model (e.g., a MTLR model, etc.).

[0169] As an example, a Kaplan-Meier approach to estimate a survival function may be utilized without considering any difference between behavior of different equipment (e.g., pumps, etc.). In such an example, a decision may be to bring to the equation a suitable set of multiple input features (e.g., an input vector x).

[0170] As explained, to approximate a survival function for a piece of equipment or equipment system considering a multidimensional feature vector, a model such as the MTLR model may be utilized. As an example, such an approach may utilize a series of logistic regression models for each value in a vector of time points (t) where the probability of surviving more than ti days may be modeled, for example, as follows:

[0171] In such an approach, while the input features (vector x or x) stay the same for each / (equivalent to one classification task in each iteration), the binary labels can change depending on the threshold ti.Example Plot Using MTLR Model

[0172] FIG. 12 shows an example of a plot 1200 of estimated probability of survival versus time in days using an MTLR model, which, in the plot 1200, can be compared to the aforementioned Kaplan-Meier model.Examples of Equipment Data

[0173] FIG. 13 shows an example of a plot 1300 of equipment data for a number of pieces of equipment (e.g., pumps) where some of the pieces of equipment experienced failure and where others did not experience failure.

[0174] As an example, an approach may be implemented where data may be relatively limited. In various instances, a machine learning model may demand a substantial amount of training data. As shown in the plot 1300, data that are actually available may be limited. As an example, an approach that utilizes a MTLR model,which may be trained via an optimization technique, may provide for generating output as to survival in a manner that demands less data than a neural network model (e.g., consider deep learning, etc.).

[0175] As an example, a framework may utilize an MTLR model-based approach (e.g., or other relatively “low” data model) where there can be a limitation of counting with a relatively small amount of data (see, e.g., FIG. 13). Such an approach can be improved compared to an approach that utilizes a traditional survival analysis model such as, for example, a Cox model, which may restrict the effect of each feature on survival to be constant over time. In an example trial using a Cox model, resulting survival plots were too squared such that survival drops quickly over a relatively short period of time. In contrast, by running a logistic regressor model on each time ti, resulting survival curves were smoother; thereby, providing an increased range of survival probabilities over an increased span of time.

[0176] As explained, an MTLR approach may be applied as a core engine (see, e.g., the MTLR component 1060 of FIG. 10) of a framework to estimate a survival function for a number of pumps (e.g., ESPs, etc.), individually, by incorporating an MTLR model into a workflow that can include additional input components such as, for example, one or more of an anomaly detection alert component, an SNHT anomalous events component, and one or more known rule-based alerts components (see, e.g., the components 1026, 1028, and 1030 of FIG. 10).

[0177] As an example, anomaly detection self-regressor models may integrate a number of machine learning models designed to leverage normal operational characteristics observed in ESP pumps. In such an example, each model may be trained to perform a specific task. For example, consider three models where one is to predict motor temperature, another is to predict delta pressure, and yet another is to predict electric current measured in amperes; noting that one or more additional or alternative self-regressor models may be utilized to evaluate data from one or more other sensors, controllers, etc.

[0178] As an example, regression model(s) predictions may be compared against corresponding actual sensor values to generate residuals (e.g., differences between predicted and actual values), which may be utilized to then identify one possible ESP pump anomalous behavior. For example, consider implementation of a statistical approach where a residual may be assessed against a predefined threshold,for example, consider a threshold derived from a number of standard deviations (e.g., consider + / - 2 std. deviations, + / - 3 std. deviations, etc.). As an example, where a magnitude of a residual exceeds this threshold, that can provide an indication as to the presence of a possible anomaly.

[0179] In various example trials, statistical hypothesis tests were run on the outputs from a framework to determine a suitable feature vector that can be used to train a model. For example, the inputs to the MTLR model component 1060 in the example of FIG. 10 may be considered elements of a feature vector. In such an example, it was determined that the vector exhibited a satisfactory correlation with the status of the pumps in the time (t > 0) when a failure occurred to a pump or the time (c > 0) of censoring (e g., that indicates that is unknown whether the pump had or did not have a failure in the given time).

[0180] As an example, a set of features may be composed by including a number of events reported in a log that a pump has experienced during its lifetime, which may be shown to be suitable on the basis of statistical hypothesis testing. In such an example, this set of features may be utilized in combination with one or more other features (e.g., one or more other sets of features, etc.).

[0181] As an example, a set of features may be composed by the average within the last seven weekly windows of: the output of the pump’s drive frequency sensor; the residuals of the electric current self-regressor model that was created to determine anomaly detection in that sensor; the residuals of the delta pressure self-regressor model that was created to determine anomaly detection in the differential on pump’s intake pressure versus discharge pressure sensors; the residuals of the motor temperature self-regressor model that was created to determine anomaly detection in that sensor; and the output of the rule-based alert that was created to show the differential of intake pressure versus discharge pressure given the condition of a scenario where the drive frequency would be 60 Hz constant.

[0182] As an example, a set of features may be composed by the first-order derivative calculated on each change on the last seven weekly windows of: the average output of the electric current sensor (in amperes); the average output of the output voltage sensor; and the sum of alerts / anomalous events from the Standard Normal Homogeneity Test (SNHT) and classified as failure in the Variable Frequency Drive (VFD).

[0183] As an example, a set of features may be composed by the sum within the last seven weekly windows of: the motor temperature alerts detected on the selfregressor model that was created to determine anomaly detection on that sensor (it is considered one alert if the difference between the predicted and real values in the sample are greater than three times the standard deviation); the total alerts detected on the three self-regressor models that were created to determine anomaly detection on the respective sensors; the alerts / anomalous events from the Standard Normal Homogeneity Test (SNHT) and classified as failure in the Variable Frequency Drive (VFD); and the alerts / anomalous events from the Standard Normal Homogeneity Test (SNHT) and classified as failure due to scales.

[0184] As an example, a set of features may be composed of the cumulative (e.g., since time-series start date) sum within the last seven weekly windows of the electric current alerts detected on the self-regressor model that was created to determine anomaly detection on that sensor.

[0185] As an example, a set of features may be composed of the cumulative residuals (absolute value) of the motor temperature self-regressor model that was created to determine anomaly detection in that sensor where the maximum value on each one of the last seven weekly windows was taken.

[0186] As an example, various sets of features may be combined to provide input features. For example, consider concatenated the various aforementioned sets of features to create a multidimensional feature vector with 99 dimensions used for an MTLR: 7 features (equivalent to the 7 weekly windows) per aggregation (14 aggregations) plus one feature representing the events reported counts (from the event log dataset).

[0187] As explained, an MTLR model may be utilized by a framework to generate the survival function of one specific ESP pump based on a vector x. As an example, a method may include creating an index to measure the behavior of an MTLR model as implemented against, for example, a Kaplan-Meier model. Such an index may be computed as a relation between survival probability computed using a framework model (e.g., MTLR(ti)) to each t> days and survival probability computed through a Kaplan-Meier model (e.g., KM(ti)): Index = (MTLR(ti) - KM(ti)) / KM(ti).Example Model Comparison Plot

[0188] FIG. 14 shows an example of a plot 1400 generated by a comparison of an MTLR model-based approach and a K-M model-based approach to provide a survival probability index versus time in days. As shown, the values of the index fall below zero, as the model KM(ti) generates survival probabilities that are greater than those of the model MTLR(ti). In such an approach, the index can indicate if the survival probability detected by the MTLR model for any particular single instance of equipment is constantly greater or constantly lower than the survival probability of the entire field (calculated with the Kaplan-Meier model) in one window of time. For example, the survival probability from a pump labeled A22 as in the plot 1400 goes less than the field estimation around December 2022 and continues going further to the base during the next 3 months until the failure date.

[0189] As an example, a framework can provide for early identification of failure probability in ESP pumps; which can potentially predict the failure date to reduce nonproductive time (NPT) in replacement acquisition.

[0190] As an example, a method may include testing an MTLR model by training it with a diverse set of features and implementing one strategy to test and measure the model’s effectiveness.

[0191] As an example, a framework may provide for generating survival information for one or more types of equipment (e.g., pumps or other devices that can generate time-series data to feed a framework).

[0192] As explained, a framework may implement one or more types of components for one or more purposes. For example, consider a framework that may include an SNHT component and / or an MTLR component. As explained, an SNHT component may be utilized for purposes of generating input for an MTLR component, noting that one or more other components may be utilized, additionally or alternatively. As explained, a framework may utilize one or more anomaly components and / or rulebased components.

[0193] As an example, a framework may provide for generating output that may be utilized for decision-making, which may include control. As an example, by estimating the residuary lifetime of one or more ESP pumps, a framework may operate to provide NPT reduction (e.g., in spare parts, pump replacement acquisition, etc.).

[0194] As explained, a system can provide for an understanding of operational conditions of ESP assets in real-time, which can facilitate proper operation, particularlyat remote field sites. In such an approach, demand for human intervention by travel to a field site may be reduced while also providing assurances as to proper operation to meet various goals. As explained, edge-based computing resources may be utilized for real-time computations, which may be utilized for monitoring and / or control. As explained, executable instructions can be stored in memory for deployment on a gateway. As explained, one or more ML model frameworks may be implemented using a gateway or other edge-based computational resources.

[0195] As an example, a data science framework may be implemented (DATAIKU) along with a container framework (DOCKER). Such a container framework may provide for construction of a unit of software that packages up executable code and its dependencies such that an application can execute quickly and reliably from one computing environment to another. As an example, a container can be an image that is a lightweight, standalone, executable package of software that includes code, runtime, system tools, system libraries and settings. A container image becomes a “container” at runtime, for example, when run on a suitable engine (e.g., DOCKER engine for a DOCKER container image). As an example, an edge implementation may utilize a framework such as, for example, a lightweight machine learning framework such as the TENSORFLOW LITE (TFL) framework (GOOGLE LLC, Mountain View, California).

[0196] As explained, while various technologies are mentioned with respect to vendors as to data science, artificial intelligence, machine learning, etc., a wellsite system framework may be operable in an agnostic manner or other manner compatible with one or more other vendor technologies. For example, consider the Cognite Data Fusion (CDF) platform (Cognite AS, Oslo, Norway), which includes features for streaming data into a CDF data model where the data may be processed (e.g., normalized and enriched), for example, by adding connections between data resources of different types and then storing in an industrial knowledge graph (e.g., in the cloud). With data in the cloud, various CDF services and tools may be implemented to build solutions and applications. As an example, such services may be applied for refactoring and, for example, to be able to host one or more models within the CDF platform. As an example, a system may provide for use of one or more vendors and / or vendor technologies in wellsite pump monitoring, control, etc.

[0197] As an example, a framework may be implemented using one or more computational devices, systems, etc., which may be operatively coupled via one or more networks (e.g., wired, wireless, etc.). As an example, one or more application programming interfaces (APIs) may be utilized where, for example, a call may be made according to an API where, in response, information is received. For example, consider one or more local and / or remote resources that may provide for use of one or more models that can be called according to one or more APIs. As an example, one or more APIs may be utilized to acquire data, control instructions, etc. As an example, a framework may be implemented using local and / or remote resources where, for example, local resources may include one or more types of edge devices installed locally at a wellsite for one or more wells. As an example, a framework may be implemented using one or more types of circuitry, which may include, for example, embedded circuitry. For example, consider embedding one or more components of a framework in a device, which may be a downhole device or a surface device. As an example, a framework may be embedded in circuitry of a downhole pump such as, for example, in a gauge of a downhole pump (e.g., consider the PHOENIX gauge, etc.).

[0198] As an example, a system, a method, etc., may utilize one or more machine learning features, which can be implemented using one or more machine learning models. As to types of machine learning models, consider one or more of a support vector machine (SVM) model, a k-nearest neighbors (KNN) model, an ensemble classifier model, a neural network (NN) model, etc. As an example, a machine learning model can be a deep learning model (e.g., deep Boltzmann machine, deep belief network, convolutional neural network, stacked auto-encoder, etc.), an ensemble model (e.g., random forest, gradient boosting machine, bootstrapped aggregation, AdaBoost, stacked generalization, gradient boosted regression tree, CATBoost, etc.), a neural network model (e.g., radial basis function network, perceptron, back-propagation, Hopfield network, etc.), a regularization model (e.g., ridge regression, least absolute shrinkage and selection operator, elastic net, least angle regression), a rule system model (e.g., cubist, one rule, zero rule, repeated incremental pruning to produce error reduction), a regression model (e.g., linear regression, ordinary least squares regression, stepwise regression, multivariate adaptive regression splines, locally estimated scatterplot smoothing, logistic regression, etc.), a Bayesian model (e.g., naive Bayes, average on-dependenceestimators, Bayesian belief network, Gaussian naive Bayes, multinomial naive Bayes, Bayesian network), a decision tree model (e.g., classification and regression tree, iterative dichotomiser 3, C4.5, C5.0, chi-squared automatic interaction detection, decision stump, conditional decision tree, M5), a dimensionality reduction model (e.g., principal component analysis, partial least squares regression, Sammon mapping, multidimensional scaling, projection pursuit, principal component regression, partial least squares discriminant analysis, mixture discriminant analysis, quadratic discriminant analysis, regularized discriminant analysis, flexible discriminant analysis, linear discriminant analysis, etc.), an instance model (e.g., k-nearest neighbor, learning vector quantization, self-organizing map, locally weighted learning, etc.), a clustering model (e.g., k-means, k-medians, expectation maximization, hierarchical clustering, etc.), etc.

[0199] As an example, a machine model may be built using a computational framework with a library, a toolbox, etc., such as, for example, those of the MATLAB framework (MathWorks, Inc., Natick, Massachusetts). The MATLAB framework includes a toolbox that provides supervised and unsupervised machine learning algorithms, including support vector machines (SVMs), boosted and bagged decision trees, k-nearest neighbor (KNN), k-means, k-medoids, hierarchical clustering, Gaussian mixture models, and hidden Markov models. Another MATLAB framework toolbox is the Deep Learning T oolbox (DLT), which provides a framework for designing and implementing deep neural networks with algorithms, pretrained models, and apps. The DLT provides convolutional neural networks (ConvNets, CNNs) and long shortterm memory (LSTM) networks to perform classification and regression on image, time-series, and text data. The DLT includes features to build network architectures such as generative adversarial networks (GANs) and Siamese networks using custom training loops, shared weights, and automatic differentiation. The DLT provides for model exchange various other frameworks.

[0200] As an example, a system may utilize one or more recurrent neural networks (RNNs). One type of RNN is referred to as long short-term memory (LSTM), which can be a unit or component (e.g., of one or more units) that can be in a layer or layers. A LSTM component can be a type of artificial neural network (ANN) designed to recognize patterns in sequences of data, such as time series data. When provided with time series data, LSTMs take time and sequence into account such that an LSTMcan include a temporal dimension. For example, consider utilization of one or more RNNs for processing temporal data from one or more sources, optionally in combination with spatial data. Such an approach may recognize temporal patterns, which may be utilized for making predictions (e.g., as to a pattern or patterns for future times, etc.).

[0201] As an example, the TENSORFLOW framework (Google LLC, Mountain View, CA) may be implemented, which is an open-source software library for dataflow programming that includes a symbolic math library, which can be implemented for machine learning applications that can include neural networks. As an example, the CAFFE framework may be implemented, which is a DL framework developed by Berkeley Al Research (BAIR) (University of California, Berkeley, California). As another example, consider the SCIKIT platform (e.g., scikit-learn), which utilizes the PYTHON programming language. As an example, a framework such as the APOLLO Al framework may be utilized (APOLLO.AI GmbH, Germany). As an example, a framework such as the PYTORCH framework may be utilized (Facebook Al Research Lab (FAIR), Facebook, Inc., Menlo Park, California).

[0202] As an example, a training method can include various actions that can operate on a dataset to train a ML model. As an example, a dataset can be split into training data and test data where test data can provide for evaluation. A method can include cross-validation of parameters and best parameters, which can be provided for model training.

[0203] The TENSORFLOW framework can run on multiple CPUs and GPUs (with optional CUDA (NVIDIA Corp., Santa Clara, California) and SYCL (The Khronos Group Inc., Beaverton, Oregon) extensions for general-purpose computing on graphics processing units (GPUs)). TENSORFLOW is available on 64-bit LINUX, MACOS (Apple Inc., Cupertino, California), WINDOWS (Microsoft Corp., Redmond, Washington), and mobile computing platforms including ANDROID (Google LLC, Mountain View, California) and IOS (Apple Inc.) operating system-based platforms.

[0204] TENSORFLOW computations can be expressed as stateful dataflow graphs; noting that the name TENSORFLOW derives from the operations that such neural networks perform on multidimensional data arrays. Such arrays can be referred to as “tensors”.

[0205] As an example, a device and / or distributed devices may utilize TENSORFLOW LITE (TEL) or another type of lightweight framework. TFL is a set of tools that enables on-device machine learning where models may run on mobile, embedded, and loT devices. TFL is optimized for on-device machine learning, by addressing latency (no round-trip to a server), privacy (no personal data leaves the device), connectivity (Internet connectivity is demanded), size (reduced model and binary size) and power consumption (e.g., efficient inference and a lack of network connections). TFL can provide multiple platform support, covering ANDROID and iOS devices, embedded LINUX, and microcontrollers. TFL can provide diverse language support, which includes JAVA, SWIFT, Objective-C, C++, and PYTHON. TFL can provide high performance, with hardware acceleration and model optimization. As an example, the system 600 of FIG. 6 may utilize one or more features of the TFL framework.Example Method and System

[0206] FIG. 15 shows an example of a method 1500 and an example of a system 1590. As shown, the method 1500 can include a reception block 1510 for receiving by a computational device at a wellsite, real-time, time series data from a pump system operating at the wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; a process block 1520 for, using the computational device, processing a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and an issuance block 1530 for issuing a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system. As an example, the signal may be a control signal that controls one or more pieces of equipment at the wellsite.

[0207] The method 1500 is shown in FIG. 15 in association with various computer-readable media (CRM) blocks 1511 , 1521 , and 1531 . Such blocks generally include instructions suitable for execution by one or more processors (or processor cores) to instruct a computing device or system to perform one or more actions. While various blocks are shown, a single medium may be configured with instructions to allow for, at least in part, performance of various actions of the method 1500. As an example, a computer-readable medium (CRM) may be a computer-readable storagemedium that is non-transitory and that is not a carrier wave. As an example, one or more of the blocks 1511 , 1521 , and 1531 may be in the form processor-executable instructions.

[0208] In the example of FIG. 15, the system 1590, which may be a wellsite system, can include one or more information storage devices 1591 , one or more computers 1592, one or more networks 1595 and instructions 1596. As to the one or more computers 1592, each computer may include one or more processors (e.g., or processing cores) 1593 and memory 1594 for storing the instructions 1596, for example, executable by at least one of the one or more processors 1593 (see, e.g., the blocks 1511 , 1521 , and 1531 ). As an example, a computer may include one or more network interfaces (e.g., wired orwireless), one or more graphics cards, a display interface (e.g., wired or wireless), etc.

[0209] As an example, a method can include receiving by a computational device at a wellsite, real-time, time series data from a pump system operating at the wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; using the computational device, processing a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issuing a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

[0210] As an example, a portion of the time series data can be a windowed portion. For example, consider a windowed portion defined in part by a homogeneity test applied to historic pump system time series data for change event identification. In such an example, a homogeneity test may be a standard normal homogeneity test.

[0211] As an example, a method can include training a machine learning model to generate a trained machine learning model. In such an example, training can include utilizing training data generated via application of a homogeneity test to historic time series data for change event identification. In such an example, application of the homogeneity test may utilize a window and means of the historic time series data within the window (e.g., average values). In such an example, the method may include defining an event window for windowing a portion of the historic time series data responsive to identification of a change event within the window. In such an example, the method may include generating training data feature values for the event window.In such an example, training can include utilizing the training data feature values in association with an identification of a change event to generate a trained machine learning model as a classifier.

[0212] As an example, feature values may include values for one or more statistical features. For example, consider statistical features such as one or more of average, median, standard deviation, minimum, maximum, and slope.

[0213] As an example, a trained machine learning model can be or can include a classifier. In such an example, the classifier may include at least one decision tree.

[0214] As an example, a forthcoming performance issue may include one or more of a scaling issue, an emulsion issue, a gas issue, and a current issue; noting that one or more other issues may be considered, additionally or alternatively.

[0215] As an example, a pump system may include at least one electric submersible pump (ESP). As an example, a pump system may include a pump that operates in a manner that rotates one or more components about an axis or in a manner that reciprocates one or more components along an axis. As an example, a motor may be an electric motor with a rotor and a stator where the rotor rotates or an electric motor with a linear shaft and a stator where the linear shaft reciprocates.

[0216] As an example, time series data can include sensor data and / or control parameter data.

[0217] As an example, a trained machine learning model may be or include a survival probability model. In such an example, consider a survival probability model that is or includes a multi-task logistic regression (MTLR) model.

[0218] As an example, a method can include training a machine learning model to generate the trained machine learning model, where the training includes utilizing data for a number of pump systems, and where the trained machine learning model generates output for a selected single pump system. In such an example, the machine learning model may be a train for many and output for individual type of machine learning model.

[0219] As an example, feature values may include feature values generated by one or more anomaly detectors. For example, consider one or more anomaly detectors such as one or more of a SNHT model-based anomaly detector and a regression model anomaly detector.

[0220] As an example, feature values may be windowed over a period of time using a number of windows. In such an example, the number of windows may include a number of weekly windows.

[0221] As an example, a method may include implementing a Kaplan-Meier model and / or one or more other models to generate output and comparing the output to output generated by a trained machine learning model. In such an example, output of the Kaplan-Meier model may be output for a group of pump systems and the output of the trained machine learning model may be output for a single pump system.

[0222] As an example, a method may include training a machine learning model by solving an optimization problem to generate a trained machine learning model. For example, an MTLR model may be a type of model that can be trained via solving an optimization problem.

[0223] As an example, feature values may include sets of feature values from a number of sets of features. In such an example, the number of sets of features can be greater than two. As an example, feature values may include (e.g., for each set), windowed feature values with respect to a predetermined time window. As an example, a workflow may include feature engineering where such feature engineering may tailor features for a particular type of equipment, equipment system, etc.

[0224] As an example, a system can include a processor; a memory accessible to the processor; and processor-executable instructions stored in the memory to instruct the system to: receive real-time, time series data from a pump system operating at a wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

[0225] As an example, one or more computer-readable storage media can include processor-executable instructions to instruct a wellsite computing system to: receive real-time, time series data from a pump system operating at a wellsite, where the wellsite includes a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issueof the pump system; and issue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

[0226] As an example, a computer program product can include one or more computer-readable storage media that can include processor-executable instructions to instruct a computing system to perform one or more methods and / or one or more portions of a method. Various example methods may be performed in various combinations.Example System

[0227] In some embodiments, a method or methods may be executed by a computing system. FIG. 16 shows an example of a system 1600 that can include one or more computing systems 1601-1 , 1601-2, 1601-3, and 1601-4, which may be operatively coupled via one or more networks 1609, which may include wired and / or wireless networks. As shown, the system 1600 can include one or more other components 1608.

[0228] As an example, a system can include an individual computer system or an arrangement of distributed computer systems. In the example of FIG. 16, the computer system 1601 -1 can include one or more modules 1602, which may be or include processor-executable instructions, for example, executable to perform various tasks (e.g., receiving information, requesting information, processing information, simulation, outputting information, etc.).

[0229] As an example, a module may be executed independently, or in coordination with, one or more processors 1604, which is (or are) operatively coupled to one or more storage media 1606 (e.g., via wire, wirelessly, etc.). As an example, one or more of the one or more processors 1604 can be operatively coupled to at least one of one or more network interface 1607. In such an example, the computer system 1601-1 can transmit and / or receive information, for example, via the one or more networks 1609 (e.g., consider one or more of the Internet, a private network, a cellular network, a satellite network, etc.).

[0230] As an example, the computer system 1601-1 may receive from and / or transmit information to one or more other devices, which may be or include, for example, one or more of the computer systems 1601-2, etc. A device may be located in a physical location that differs from that of the computer system 1601-1. As anexample, a location may be, for example, a processing facility location, a data center location (e.g., serverfarm, etc.), a rig location, a wellsite location, a downhole location, etc.

[0231] As an example, a processor may be or include a microprocessor, microcontroller, processor module or subsystem, programmable integrated circuit, programmable gate array, or another control or computing device.

[0232] As an example, the storage media 1606 may be implemented as one or more computer-readable or machine-readable storage media. As an example, storage may be distributed within and / or across multiple internal and / or external enclosures of a computing system and / or additional computing systems.

[0233] As an example, a storage medium or storage media may include one or more different forms of memory including semiconductor memory devices such as dynamic or static random access memories (DRAMs or SRAMs), erasable and programmable read-only memories (EPROMs), electrically erasable and programmable read-only memories (EEPROMs) and flash memories, magnetic disks such as fixed, floppy and removable disks, other magnetic media including tape, optical media such as compact disks (CDs) or digital video disks (DVDs), BLUERAY disks, or other types of optical storage, or other types of storage devices.

[0234] As an example, a storage medium or media may be located in a machine running machine-readable instructions, or located at a remote site from which machine-readable instructions may be downloaded over a network for execution.

[0235] As an example, various components of a system such as, for example, a computer system, may be implemented in hardware, software, or a combination of both hardware and software (e.g., including firmware), including one or more signal processing and / or application specific integrated circuits.

[0236] As an example, a system may include a processing apparatus that may be or include a general-purpose processors or application specific chips (e.g., or chipsets), such as ASICs, FPGAs, PLDs, or other appropriate devices.

[0237] As an example, a device may be a mobile device that includes one or more network interfaces for communication of information. For example, a mobile device may include a wireless network interface (e.g., operable via IEEE 802.11 , ETSI GSM, BLUETOOTH, satellite, etc.). As an example, a mobile device may include components such as a main processor, memory, a display, display graphics circuitry(e.g. , optionally including touch and gesture circuitry), a SIM slot, audio / video circuitry, motion processing circuitry (e.g., accelerometer, gyroscope), wireless LAN circuitry, smart card circuitry, transmitter circuitry, GPS circuitry, and a battery. As an example, a mobile device may be configured as a cell phone, a tablet, etc. As an example, a method may be implemented (e.g., wholly or in part) using a mobile device. As an example, a system may include one or more mobile devices.

[0238] As an example, a system may be a distributed environment, for example, a so-called “cloud” environment where various devices, components, etc. interact for purposes of data storage, communications, computing, etc. As an example, a device or a system may include one or more components for communication of information via one or more of the Internet (e.g, where communication occurs via one or more Internet protocols), a cellular network, a satellite network, etc. As an example, a method may be implemented in a distributed environment (e.g., wholly or in part as a cloud-based service).

[0239] Although only a few example embodiments have been described in detail above, those skilled in the art will readily appreciate that many modifications are possible in the example embodiments. Accordingly, all such modifications are intended to be included within the scope of this disclosure as defined in the following claims. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents, but also equivalent structures. Thus, although a nail and a screw may not be structural equivalents in that a nail employs a cylindrical surface to secure wooden parts together, whereas a screw employs a helical surface, in the environment of fastening wooden parts, a nail and a screw may be equivalent structures.

Claims

CLAIMSWhat is claimed is:

1. A method comprising: receiving by a computational device at a wellsite, real-time, time series data from a pump system operating at the wellsite, wherein the wellsite comprises a wellbore in contact with a fluid reservoir; using the computational device, processing a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issuing a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

2. The method of Claim 1 , wherein the portion of the time series data comprises a windowed portion.

3. The method of Claim 2, wherein the windowed portion is defined in part by a homogeneity test applied to historic pump system time series data for change event identification.

4. The method of Claim 1 , comprising training a machine learning model to generate the trained machine learning model.

5. The method of Claim 4, wherein the training comprises utilizing training data generated via application of a homogeneity test to historic time series data for change event identification, and wherein the application of the homogeneity test utilizes a window and means of the historic time series data within the window.

6. The method of Claim 5, comprising defining an event window for windowing a portion of the historic time series data responsive to identification of a change event within the window.

7. The method of Claim 6, comprising generating training data feature values for the event window, and wherein the training comprises utilizing the training data feature values in association with the identification of the change event to generate the trained machine learning model as a classifier.

8. The method of Claim 1 , wherein the feature values comprise values for one or more statistical features, and wherein the statistical features comprise one or more of average, median, standard deviation, minimum, maximum, and slope.

9. The method of Claim 1 , wherein the trained machine learning model comprises a classifier.

10. The method of Claim 9, wherein the classifier comprises at least one decision tree.11 . The method of Claim 1 , wherein the forthcoming performance issue comprises one or more of a scaling issue, an emulsion issue, a gas issue, and a current issue.

12. The method of Claim 1 , wherein the pump system comprises at least one electric submersible pump.

13. The method of Claim 1 , wherein the time series data comprise one or more of sensor data and control parameter data.

14. The method of Claim 1 , wherein the trained machine learning model comprises a survival probability model.

15. The method of Claim 1 , wherein the feature values comprise feature values generated by one or more anomaly detectors.

16. The method of Claim 1 , further comprising implementing a Kaplan-Meier model to generate output and comparing the output to output generated by the trained machine learning model, wherein the output of the Kaplan-Meier model comprises output for agroup of pump systems and wherein the output of the trained machine learning model comprises output for a single pump system.

17. The method of Claim 1 , comprising training a machine learning model by solving an optimization problem to generate the trained machine learning model.

18. The method of Claim 1 , wherein the feature values comprise sets of feature values from a number of sets of features, wherein the number of sets of features is greater than two, and wherein the feature values comprise, for each of the sets, windowed feature values with respect to a predetermined time window.

19. A system comprising: a processor; a memory accessible to the processor; and processor-executable instructions stored in the memory to instruct the system to: receive real-time, time series data from a pump system operating at a wellsite, wherein the wellsite comprises a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; and issue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

20. One or more computer-readable storage media comprising processor-executable instructions to instruct a wellsite computing system to: receive real-time, time series data from a pump system operating at a wellsite, wherein the wellsite comprises a wellbore in contact with a fluid reservoir; process a portion of the time series data to generate feature values as input to a trained machine learning model to detect pump system behavior indicative of a forthcoming performance issue of the pump system; andissue a signal responsive to detection of the pump system behavior to mitigate the forthcoming performance issue of the pump system.

Images