Drilling into the big data analytics engine

Abstract

The present invention relates to systems, methods and apparatus for determining a rig state of a drilling rig and detecting and mitigating drilling faults during a wellbore drilling operation. These systems, methods and apparatus provide a computer having a memory and a processor, a plurality of sensors associated with the wellbore drilling operation for acquiring time series data, wherein the data is formatted for sample and bandwidth regularization and is corrected in time to provide substantially time synchronized data, a processing graph of data stream networked mathematical operators, the processing graph applying a continuous analysis method to the data at least as fast as the data is acquired to determine a dynamic state of a plurality of rig conditions associated with the wellbore drilling operation, and determining a rig state based on the plurality of rig conditions.

Description

[0001] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0002] none. Technical Field

[0003] The present invention generally relates to detecting, quantifying and mitigating failures in hydrocarbon drilling. More specifically, but not by way of limitation, embodiments of the present invention include applying analytics to real-time data collected from wellbore drilling operations in order to mitigate drilling failures. Background Art

[0004] Hydrocarbon reserves are developed through drilling operations using a drill bit associated with a drill string that is rotated from the surface, or using a downhole motor, or both using a downhole motor and rotating the drill string from the surface. The bottom hole assembly (BHA) at the end of the drill string may include components such as drill collars, stabilizers, drilling motors and logging tools, and measurement tools. The BHA is also capable of sending various drilling and geological parameter telemetry to surface facilities.

[0005] The drag encountered by the drill string in the wellbore during drilling results in significant wear of the drill string, and particularly, the drill bit and BHA. It is important to understand how the geometry of the wellbore affects the drag on the drill string and BHA, and to manage dynamic conditions that potentially lead to downhole equipment failures, in order to enhance drilling efficiency and minimize costs. Various conditions, known as drilling faults, that may lead to component failures include excessive torque, impact, bit bounce, induced vibration, bit whirl, stick-slip, bit bounce, and the like. These conditions must be detected quickly so that efforts can be made to mitigate them as quickly as possible, as some faults can quickly lead to tool failure.

[0006] Rapidly aggregating and analyzing data from multiple sources related to wellbore drilling operations facilitates efficient drilling operations by providing timely responses to drilling failures. Accurate timing information of borehole or drill string time series data collected by downhole sensors is important for aggregating information from surface and downhole sensors. However, each sensor may have its own internal clock, or data from many sensors may be collected and recorded relative to multiple clocks that are not synchronized. This asynchronicity of timing information creates problems when merging and processing data from different sensors. In addition, it is known that sensor timing is sometimes affected by various environmental factors that cause variable timing drifts that may affect various sensors differently. Many factors may make the timing of independent sensors inaccurate, and the sensors then need to be corrected or adjusted so that the data can be correctly assimilated when all sensor information is consistent in time, so as to accurately inform the drilling operations center of the dynamic state of the well being drilled.

[0007] Downhole drilling faults can cause serious operational problems that are difficult to detect or predict. The faster and more efficiently a drilling fault can be identified, the faster the fault can be mitigated. Therefore, efficient methods, systems and devices are needed to quickly identify and mitigate faults during drilling operations. Summary of the invention

[0008] It should be understood that although illustrative implementations of one or more embodiments are provided below, various specific embodiments may be implemented using any number of techniques known to those of ordinary skill in the art. This disclosure should in no way be limited to the illustrative embodiments, drawings, and / or techniques presented below, including the exemplary designs and implementations presented and described herein. In addition, this disclosure may be modified within the scope of the appended claims and all equivalents thereof.

[0009] In a non-limiting embodiment, the present invention more specifically includes a system for determining a rig state of a drilling rig during a wellbore drilling operation, the system comprising: a computer, the computer including a memory and a processor; a plurality of sensors, the plurality of sensors associated with the wellbore drilling operation, for acquiring time series data, wherein the data is formatted for sample and bandwidth regularization and is time-corrected to provide substantially time-synchronized data; a processing graph of data stream networked mathematical operators, the processing graph applying a continuous analysis method to the data at least as fast as the data is acquired to determine the dynamic status of a plurality of rig conditions associated with the wellbore drilling operation; and determining the rig state based on the plurality of rig conditions.

[0010] In another non-limiting embodiment, a method for determining a rig state of a drilling rig includes: acquiring data from a plurality of sensors associated with a wellbore; formatting the acquired data for sample and bandwidth normalization; time-correcting the data to provide substantially isochronously sampled data from the plurality of sensors; processing the acquired data through a processing graph of networked mathematical operators that applies a continuous analysis method to the data at least as fast as the data is acquired to determine dynamic conditions of a plurality of drilling rig operations associated with the wellbore; and determining a rig state based on the plurality of drilling rig operation conditions.

[0011] In still another non-limiting embodiment, a drilling apparatus for mitigating drilling failures comprises: a drilling rig associated with a plurality of sensors that provide time series data to a surface-based aggregator, wherein the data is formatted for sample and bandwidth regularization and is time-corrected to provide substantially time-synchronized data, a computer comprising a memory and a processor; a processing graph of data stream networked mathematical operators that applies continuous analysis at least as fast as the time series is acquired to determine dynamic conditions of a plurality of rig conditions associated with a wellbore drilling operation; and detecting drilling failures based on the plurality of rig conditions.

[0012] In yet another non-limiting embodiment, a computer program product is implemented in a non-transitory computer-readable medium, the computer program product being adapted to perform a method for mitigating drilling failures during a wellbore drilling operation, the method comprising: acquiring data from a plurality of sensors associated with the wellbore drilling operation; formatting the acquired data for sample and bandwidth regularization; time-correcting the data to provide substantially synchronously sampled data from the plurality of sensors; processing the acquired data through a processing graph of networked mathematical operators, the processing graph applying a continuous analysis method to the data at least as fast as the data is acquired to determine dynamic conditions of a plurality of rig operations associated with the wellbore; detecting drilling failures based on the plurality of rig operation conditions; and outputting rig control instructions to mitigate the detected drilling failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] A more complete understanding of the present invention and its advantages may be obtained by referring to the following description taken in conjunction with the accompanying drawings, in which:

[0014] Figure 1 Examples of drilling operations according to various embodiments of the present disclosure are shown;

[0015] Figure 2 Schematically illustrates a process diagram according to various embodiments of the present disclosure;

[0016] Figure 3 The parameters related to the geometric meander bending function are presented;

[0017] Figure 4 The forces relative to the bent drill pipe are shown;

[0018] Figure 5 Methods for determining real-time drilling faults by measuring power loss of signal propagation associated with a drill string are presented in accordance with various embodiments of the present disclosure;

[0019] Figure 6A system associated with a drill string in a wellbore for acquiring a time series according to various embodiments of the present disclosure is shown;

[0020] Figure 7 Demonstrating the use of a drilling apparatus for multi-well drilling according to various embodiments of the present disclosure;

[0021] Figure 8 shows examples of time series data before and after temporal correction of the data according to various embodiments of the present disclosure;

[0022] Fig. 9 Examples of before and after clock offset correction of downhole data according to various embodiments of the present disclosure are shown;

[0023] Fig.10 Examples are shown before and after linear moveout correction of data acquired from a downhole transducer according to various embodiments of the present disclosure;

[0024] Fig.11 Methods according to embodiments of the present disclosure are presented, the methods according to various embodiments of the present disclosure being used to adjust time series data relative to a reference time;

[0025] Fig.12 Methods according to alternative embodiments of the present disclosure are presented for adjusting time series data relative to a reference time according to various embodiments of the present disclosure;

[0026] Fig.13 Methods according to further embodiments of the present disclosure are presented, the methods according to various embodiments of the present disclosure being used to automatically adjust time series data relative to a reference time;

[0027] Fig.14 Schematic diagrams showing embodiments of systems that may correspond to a computer or may be part of a computer according to various embodiments of the present disclosure;

[0028] Fig.15 A system for determining a drilling rig status of a drilling rig during a wellbore drilling operation according to various embodiments of the present disclosure is presented; and

[0029] Fig.16 Methods for determining a drilling rig status of a drilling rig during a wellbore drilling operation are presented according to various embodiments of the present disclosure. DETAILED DESCRIPTION

[0030] Turning now to the detailed description of the preferred arrangements of the invention, it should be understood that the features and concepts of the invention can be embodied in other arrangements and that the scope of the invention is not limited to the embodiments described or illustrated. The scope of the invention is intended to be limited only by the scope of the following claims.

[0031] The following examples of certain embodiments of the present invention are given. Each example is provided by way of explanation of the present invention, being one of many embodiments of the present invention, and the following examples should not be interpreted as limiting or defining the scope of the present invention.

[0032] Mitigating drilling failures in oilfield drilling operations is a priority in the industry. Low-frequency surface data, such as RPM, torque, and acceleration data, are routinely used to mitigate drilling failures. Recent advances in recording high-frequency surface and downhole data provide better detection and analysis and lead to faster mitigation of drilling failures. Complex event processing (CEP) is provided by the data acquisition and processing capabilities contained in the embodiments disclosed herein. Real-time analytics are feasible when tool motion and fault indicators are analyzed during drilling operations using signal processing, vibration analysis, CEP, and feedback loops (including instructions to the rig control unit for mitigating faults). This results in efficient acquisition of downhole tool wear indicators through monitoring and prediction, which allows optimized preventive maintenance on all parts of the drill string. This also allows for an effective and continuous understanding of downhole conditions, resulting in a wellbore optimized for completion. Indicators useful for the analytics engine include fault indicators such as stick-slip index (SSI), bit bounce index (BBI), bit whirl index (BWI), and specific mechanical energy (MSE).

[0033] The continuous real-time acquisition of multiple data streams of conventional and new data types as quickly and efficiently as possible allows for analytical means for each directly acquired time series as well as the combination of these data together. How these measurements change over time and how each measurement data stream changes relative to the other data streams provides new analytical tools for understanding drill string dynamics and wellbore conditions as the data is acquired.

[0034] Proper data consolidation and analysis is integral to understanding CEP and creating rules or learnings that can be applied to current drilling operations. This data consolidation and comparison is affected when different data types are not synchronized to the same reference time.

[0035] During drilling operations, sensors are placed at various wellbore locations, drill string locations, and time intervals to provide real-time measurements such as revolutions per minute (RPM), torque, weight on bit (WOB), and acceleration. The data collected with the sensors may be irregularly distributed and subject to transmission losses due to absorption, scattering, and leakage caused by the curvature of the well trajectory. The nonlinear combination of these effects causes significant attenuation or power loss to the signal amplitude, which may compromise integrity and fault prediction at multiple sections of the drill string.

[0036] Figure 1 An example of drilling a formation using a first wellbore and a second wellbore according to various embodiments of the present disclosure is shown. Various embodiments disclosed herein are used in Figure 1 In the illustrated drilling environment, a wellbore 102 is drilled from a surface drilling facility 101, which includes a drilling rig, a drill string associated with surface-based sensors 103 (e.g., electro-acoustic receivers attached to a Kelly or BOP) for obtaining data from within the wellbore, and associated control and support facilities 105, which may include data aggregation, data processing infrastructure (including computer systems), and a drilling control system. During drilling operations, the wellbore 102 includes a drill string, which includes an associated bottom hole assembly (BHA), which may include a mud motor 112, an adjustable bending housing or 'BHA dynamic subgroup (Dynamic Sub)' 114 (including various sensors, transducers and electronic components), and a drill bit 116. The BHA dynamic subgroup collects time series data such as RPM, torque, bending moment, tension, pressure (ECS) and vibration data. In addition, the BHA collects measurement while drilling and logging while drilling (MWD / LWD) data such as inclination, azimuth, gamma ray, resistivity and other advanced LWD data in high fidelity or standard mode. Any data collected with the BHA can be transmitted to the drilling rig 101 as time series data via drill string telemetry or via mud pulse telemetry.

[0037] The drill string may also include associated sensors, such as a mid-column dynamics subgroup 110 that collects high fidelity time series data (such as RPM, torque, bending moment, tension, and vibration data), and these instrumented subgroups may telemetrically transmit signals representing the measurements up the drill string, which are also recorded on or near the drilling rig 101.

[0038] In various embodiments, the results acquired while drilling the first wellbore 102 are provided to the second wellbore (eg, Figure 1The drilling of a wellbore 104 in the second wellbore 104 can improve the efficiency of subsequent drilling. The model parameters determined from drilling the first wellbore 102, combined with the geometric information and other time series data received by telemetry from the BHA associated with the drill string of the second wellbore 104, can be used to determine the downhole dynamic characteristics associated with the drilling operation, so that faults can be quickly detected and effectively mitigated.

[0039] Embodiments disclosed herein provide data-driven drilling performance optimization. Performance optimization of drilling operations is a big data problem, requiring rapid (real-time) integration and analysis of a wide variety of high-volume data streams. Analysis of relevant data is as fast as data collection. Performance optimization translates into safer, more efficient, and lower-cost drilling.

[0040] Embodiments of the drilling big data analytics engine according to the present disclosure provide a stream computing paradigm to flow data samples through a processing graph of networked mathematical operators interconnected by data streams. Figure 2 An example processing graph 200 is schematically illustrated in FIG. Data ingestion is performed continuously using many input data streams so that time-ordered and time-registered data samples flow through the graph to apply continuous analytics. The analytics engine (processing graph) is provided with access to data from all drilling operation sensors, communication systems, and hardware that requires streaming data in and out. Each mathematical operator is designed, positioned, and connected in the data stream so that the aggregate action of the processing graph produces the desired analysis and output flow at each exit point of the processing graph. Figure 2 The structure of the input data flow to the processing diagram is schematically depicted, as well as the analysis method and rig data control flow of the output to the outlet 250, although it will be understood that the processing diagram is not limited to one outlet. The output data flow includes fault indicators, specific mechanical energy, rig status, wellbore condition, tool wear and rig control feedback.

[0041] This rig big data analytics engine coordinates the flow and analysis of all data streams from wellbore drilling operations in "real time." For historical analysis of recorded data, "real time" means that operations are performed at least as fast as the sampling rate of the data.

[0042] As a pair Figure 2As a non-limiting example of inputs to the processing diagram example, input 210 may include conventional means for monitoring drilling operations, such as hook load and torque. Input 220 may be high-fidelity surface measurements, such as RPM, torque, bending, tension, pressure, and acceleration. Input 230 may include downhole measurements sent telemetrically to the surface, such as MWD / LWD, temperature, directional information, data stored in memory, and other "wired pipe" information stored and sent telemetrically. Input 240, as another example, may include other data associated with drilling operations, such as quality assurance / quality control data, WellView data, and other commercially available information or data associated with well operations acquired on-site from third-party vendors.

[0043] Non-limiting examples include: Data flow coordination and analysis through the processing graph, including real-time ingestion of diagnostic drilling data, feedback, and drilling control parameters into the processing graph of networked mathematical operators. Ingestion preserves the temporal order of input data streams and registers, with each sample in the stream having the correct clock time. Data flow coordination and analysis includes data formatting for flow through the processing graph and for visualization at strategic locations within and out of the graph 250.

[0044] Sampling regularization takes into account the variable sampling rate of the sensor. This regularization is included in the signal processing algorithms and visualization software. For downsampled data streams, anti-aliasing filters can be applied. Bandwidth regularization takes into account the variable resolution. Resolution often varies between sensor types and can affect the results of data mining and analysis if not accounted for.

[0045] Data flow coordination and analysis also includes "static" and "dynamic" time corrections to account for time stamp differences between different sensors. "Static" errors are independent of the time stamp coordinates, and "dynamic" errors vary over time. Static errors are most commonly introduced by human error, when the initial timestamps are inaccurately set for the data stream from the sensor. Dynamic timing errors are usually associated with clock drift, when the clocks on all sensors associated with the rig time series measurements are not running at the same rate.

[0046] Dynamic "time difference" corrections can be used to account for signal travel times between sensors located at different physical locations. The reference time associated with a particular sensor or sensor group can be selected arbitrarily, such as selecting one at the surface and all travel times corrected to the reference time, which may involve one or more corrections, including time difference, static or dynamic corrections. More information on time synchronization across sensor data is provided below.

[0047] Data flow coordination and analysis through process graphs are applied to the signal processing of the measured data streams to remove non-informative signal components. For example, an acceleration sensor attached to a rotating device may contain non-informative signal components due to rotation. These contributions are removed in real time to reveal drilling faults. Vibrationally non-informative components can be targeted to mitigation algorithms. Real-time signal processing also maps data from a local rotating coordinate system to a global static coordinate system.

[0048] Embodiments described herein provide for the computation of an output data stream of "rig status." Rig status is a sample-by-sample automated classification of ongoing rig operations computed from a diagnostic input data stream. Important categories include rotary drilling, sliding, reaming, back reaming, tripping, etc. Data mining and analysis are supported by incorporating knowledge of the rig status for each data sample.

[0049] Data flow coordination and analysis by the process graph also includes computing an output data stream characterizing wellbore tortuosity from wellbore position data measured while drilling. Multiple points of high tortuosity in the wellbore create large contact forces on drill string components, thereby generating undesirable vibrations.

[0050] Coordination and analysis of data flow through the processing graph also includes: Energy loss correction of data streams obtained from the surface to account for the attenuation of diagnostic signals traveling from a source point downhole to the surface or vice versa from the surface to the downhole sensor. Weight-on-bit correction accounts for the incomplete transfer of weight applied to the surface to the weight on the drill bit. Detection and quantification of drilling faults (typically unwanted vibrations and / or weight-on-bit fluctuations) can be measured from diagnostic data from sensors deployed anywhere (from surface equipment, through the drill string, to the bottom hole assembly and drill bit). Torque, acceleration, and tension measurements constitute typical diagnostic data streams used to detect and quantify unwanted vibration modes during drilling operations. Fault mitigation algorithms and rig feedback control can focus on minimizing fault metrics calculated from input data streams, rather than focusing directly on the input streams.

[0051] The coordination and analysis of data flow through the process graph also includes output data streams of drill rig control commands for changing rpm, applied bit weight at the surface, pump pressure and other controllable drilling parameters for the purpose of mitigating drilling failures.

[0052] Embodiments disclosed herein provide a drilling big data analytics engine that coordinates the streams and performs analysis of measured data streams from wellbore drilling operations. Embodiments of the analytics engine include data-driven drilling performance optimization. Performance optimization in the context of drilling includes reducing undesirable mechanical vibrations generated by drilling operations, thereby reducing downtime. Drilling performance metrics also include optimal penetration rates.

[0053] Optimization is achieved by the analytics engine through real-time acquisition and analysis of incoming diagnostic data streams. Driven by online data analysis, commands can be issued in real time to the rig control to change the RPM of the rotating drill string or, for example, change the weight applied to the drill bit. Real-time automated control of RPM or weight-on-bit parameters requires high-density (e.g., 100 samples / second) diagnostic data streams. Consistent with what is known as the Big Data problem, such an analytics engine is capable of acquiring and analyzing hundreds of high-density data streams simultaneously. These diagnostic data streams can be generated simultaneously from any part of the drilling operation, including sensors deployed downhole or from equipment and operations on the surface.

[0054] Drilling optimization can also be achieved through historical analysis of recorded data to improve wellbore design, quantify new drilling techniques, and establish data-driven best practices for future drilling. That is, the analytics engine is agnostic about its data source. It can be operated on-site in real time from direct sensor input at the drilling location, or after the fact from data recorded in memory during the drilling operation and transmitted to an off-site operations center, or a combination of the two. When deployed in an operations center to receive information from several concurrent drilling operation locations, the drilling big data analytics engine is able to analyze high-density data from an entire fleet of rigs simultaneously.

[0055] The Drilling Big Data Analytics Engine provides an integrated platform for big data analysis and drilling operation performance optimization. It performs a wide range of cross-correlated analyses (such as signal processing, fault detection / characterization / mitigation, and data mining) in real time. It simultaneously analyzes streaming data from all sensors within the drilling system. The Analytics Engine can be deployed in an off-site operations center or directly on the rig.

[0056] Embodiments disclosed herein also provide for predicting real-time drilling failures at any location in the drill string. Advantages provided by the various embodiments disclosed herein include: (a) simplicity of detecting and modeling a wide range of possible power losses with only three parameters; (b) determination of downhole conditions that is well-set and amenable to testing of stable estimation parameters at different scales; (c) flexibility in use with different bending functions and signal representations (e.g., averages, envelope values); (d) effectiveness in predicting failures by determining power loss at any point in time / depth, and thus useful for measuring and understanding dynamic downhole conditions through measurements associated with the drill string collected at the surface drilling facility, so that similarly positioned wells can be drilled without using dynamic subsets in the string and only using data collected from the surface to characterize the dynamic downhole environment during the drilling operation.

[0057] During drilling operations, sensors are placed at various wellbore locations, drill string locations, and time / depth intervals to provide real-time measurements such as revolutions per minute (RPM), torque, weight on bit (WOB), and acceleration, etc. The data collected with the sensors may be irregularly distributed and subject to transmission losses due to absorption, scattering, and leakage caused by the curvature of the well trajectory. The nonlinear combination of these geometrically related effects results in significant attenuation of signal amplitude or power loss, which may compromise integrity and fault prediction at multiple sections of the drill string along the wellbore.

[0058] Understanding the laws governing power loss along the wellbore enables detection and rig control mechanisms that can mitigate undesirable vibrations or other conditions to prevent or delay actual drill bit or BHA failures. The disclosed embodiments provide a simple but powerful power loss model that predicts the signal capability degradation due to wellbore geometry under arbitrary bending effects. The understanding of power loss due to wellbore geometry provided by this power loss model facilitates the understanding of dynamic downhole conditions (including failures) while drilling.

[0059] The power loss model depends on a set of three parameters: one parameter, alpha (α), which controls the control loss along the vertical section (i.e., without considering bending effects), and two parameters, beta (β) and optionally gamma (γ), which control the trade-off between exponential and hyperbolic signal attenuation for a given bending function or wellbore geometry.

[0060] The power loss model incorporates similar terms for both sheet (rigid) and fibrous (soft) model losses, similar to the models presented in Optics [Hunsperger, 2009] and Photonics [Pollock, 2003]. The presently disclosed embodiments include, but are not limited to, three different bending functions relative to the wellbore geometry, which can be described by mathematical relationships using α, β, and γ: 1) geometric tortuosity, 2) accumulated dog-legs, and 3) clamping efficiency.

[0061] Wellbore tortuosity is inherent to drilling and is fluctuation or variation from the planned wellbore trajectory, such as spiraling in a vertical section or sliding-rotating behavior in a horizontal section. A bend is a turn in the wellbore where the trajectory of the wellbore deviates from a straight path. A bend may be intentionally created in directional drilling to divert the wellbore to a horizontal path, such as in the case of unconventional shale wells. A standard calculation of bend severity is expressed in two-dimensional numbers per 100 feet (or every 30 meters) of wellbore length.

[0062] The increased use of sensors in real-time downhole operations is useful for investigating the wellbore environment during drilling and measuring the actual geometry of the wellbore. The possibility of modeling the power loss of the signal traveling up the drill string as a result of the wellbore geometry can now be addressed with instrumented drilling practices. Models are typically governed by exponential decay functions. These functions can take different forms to accommodate different material types, to capture other sources of loss on curved geometries, such as those produced by microbends and sudden or relatively rapid changes in curvature.

[0063] Advantages of the warp function model disclosed herein include: (a) simplicity of using a combination of three model parameters (designated herein as α, β, and γ) to accommodate a wide range of possible losses through different mathematical descriptions; (b) a well-set model or set of models that can withstand stable estimation of its parameters at different scales; (c) flexibility in use with different warp functions and signal representations (e.g., mean, envelope values); and (d) effectiveness in predicting faults using power loss at any point in time / depth along the drill string, resulting in efficient and timely fault mitigation.

[0064] Low frequency surface data, such as RPM, weight on bit (WOB), torque (TOB), and acceleration data, are routinely used to discover and mitigate drilling failures. However, recent developments in recording high frequency surface and downhole data have added a new dimension to better understand drilling failures. Wave optics and photonics literature provide similar content that can be used to understand transmission losses such as absorption, scattering, and leakage through different materials that are subject to bending effects (such as caused by geometry in a wellbore).

[0065] In general, the losses due to curvature and other geometric considerations in the wellbore can be described as: P(z) = P(0)·e -az , where P is the power loss, z is the depth and a is the signal propagation strength in the drill string, so that

[0066] Assuming that all propagation constants can be combined and phase effects can be ignored, the signal propagation a can be expressed as a = α·e -β·R (in the case of sheet models, this can be used to model at relatively short distances) and (For fibrous models, this can be used to model over longer distances) where R is the radius of curvature, α is a constant whose magnitude depends on the situation, and β and γ are parameters that relate to the curvature or radius in an exponential or hyperbolic sense.

[0067] Various embodiments of the present disclosure provide a hybrid sheet / fiber model of power loss. The disclosed model includes exponential coefficients that decay as a mix of exponential and hyperbolic trends from a bending model, where

[0068]

[0069] where τ≡ is the clamping efficiency. Note that For , it is an attenuation model on the vertical domain, such as the initial vertical section of the wellbore construction.

[0070] Two-step parameter estimation: (1) For i = 1, 2, ..., N z ; j = 1, 2, ..., N s , there is ln(P 0,j / P i,j )+a i z i = 0, and (2) is a three-parameter problem to take into account the combined sheet / fibrous effects, where i is an index in depth and j refers to the observation station.

[0071] Implementations of different preferred embodiments for characterizing or modeling power loss faults include the option of selecting a bending function or modeling a selected bending function (i.e., geometric tortuosity, bending, and clamping efficiency). Also, the option of deriving experiments with different fitting options can be used with these model parameters. In addition, it is possible to define the fitting geometry from any given starting depth. There are also definitions provided by applying model parameters for different smoothing and filtering options. Both lamellar and fibrous models can be used to estimate power loss by inversion using a combination of surface sensor time series data compared to equivalent downhole sensor time series data. Regression can be performed on data from any sensor or aggregated data from some or all sensors.

[0072] Geometric tortuosity bending function Can be Given, where l k is the idealized length from one subsurface observation station location to the next subsurface observation station location, and z k is the actual distance along the actual geometric length of the wellbore being drilled. The numerator and denominator of the last term in the equation are shown in Figure 3 The cumulative bending function δ is given by:

[0073] As in Figure 3 The geometric tortuosity bending function from station 1 to station 2 is shown in is measured bidirectionally, including the numerator ‖TVD k ,NS k ,E.W. k ‖2 and denominator MD kThe denominator is the actual geometry measured along the wellbore between stations 1 and 2, for example using data acquired from the BHA, while the numerator is based on the vertical distance (TVD k ), north-south distance (NS k ) and east-west distance (EW k ), taking into account the azimuth Az1 and inclination I1 of the drill string at observation station 1, and the azimuth Az2 and inclination I2 of the drill string at observation station 2.

[0074] To further analyze the bending function in the wellbore, the clamping efficiency parameter can be described in a physics-based equation, where the forces acting on the drill pipe 400 are expressed as Figure 4 The places shown in the table are designated as For the tension or lateral force on the drill string in the direction of the wellbore trajectory, the force along the trajectory of the drill string is F t , and the force perpendicular to the wellbore trajectory at this point is F n The force in the other direction from the trajectory of the drill string at the bend is F t +ΔF t , due to bending, the force is in the direction as The weight of the drill string is designated as W. With these parameters, the forces can be combined to describe the clamping efficiency in a form similar to the resistance of the wellbore to drilling operations due to the interaction of the drill string with the wellbore geometry:

[0075]

[0076] Figure 5A method for determining drilling faults in real time by measuring power loss of signal propagation associated with a drill string is presented. A (first) well is drilled with an instrumented drill string, wherein the drill string includes a mid-string drilling subunit (e.g., torque and tension subunits) to acquire, store, and telemetrically transmit time series data to the surface 501. A first time series is acquired from sensors associated with the mid-string drilling subunit in the wellbore, wherein the sensors are below the surface of the earth 503. A second time series is acquired from sensors associated with the drill string, which is in the wellbore, wherein the sensors associated with the drill string are at or near the surface of the earth, such as associated with acoustic receivers attached to a Kelly or other drilling rig component for acquiring signals. The geometry of the wellbore is determined from data from the bottom hole assembly that is telemetrically transmitted to the surface 505. Model parameters describing the power loss of the propagation of the wellbore signal due to the geometry effect are determined using the first time series, the second time series and the geometry of the wellbore to derive model parameters alpha and beta that characterize the power loss of the signal propagation for the signal traveling through the drill string based on the attenuation caused by the geometry of the wellbore 509 and other dynamic effects. Differential power loss between different sensors at different locations can assist in the characterization. Analysis of the differential power loss effects of the different time series comparisons allows for the detection and then mitigation of drilling failures. The second well can be drilled as follows, wherein the drill string does not include a mid-string drilling subunit that collects and sends time series data to the drill string 511. The dynamic state of the second well drill string in the second wellbore can be determined based on third time series data collected by sensors associated with the drill string in the wellbore, wherein the sensors are on or near the surface of the earth (i.e., associated with the acoustic sensors on Kelly), and the third time series data is combined with the data telemetered by the BHA and the model parameters determined from the first well 513. Drilling failures while drilling the second well can be detected and mitigated 515 using the third time series, with model parameters derived from the geometry of the first wellbore and the second wellbore.

[0077] Figure 6A system is shown including an in-string drill sub-sensor (110) associated with a drill string in a wellbore in a first well to acquire a first time series 601. The sensor associated with the first well drill string is used to acquire a second time series, wherein the sensor is on a drilling rig or near the surface of the earth 603. A bottom hole assembly 112, 114, 116 associated with the drill string in the wellbore 102 provides data to determine a geometry of the first wellbore 102 605. A first computer program module determines model parameters using the first time series, the second time series, and the wellbore geometry to derive model parameters alpha and beta that characterize power losses of signal propagation of signals traveling through the drill string 607. Optionally, the system may also include a second well drill string in the wellbore 104, wherein the drill string does not include an in-string drill sub-unit 609 that acquires and transmits the time series data to the drill string. Optionally, the system may further include a sensor 103 associated with the second well drill string, wherein the sensor is on or near the surface of the earth (e.g., an acoustic sensor associated with Kelly) to provide data to determine the dynamic state of the second well drill string in the wellbore based on a third time series acquired by the sensor combined with the model parameters determined from the first well 611. The system may also include a second computer model to determine drilling failures when drilling the second well, the failures being determined using the model parameters determined from the first well, the third time series, and the geometry of the second wellbore (such as derived from BHA data associated with the second drill string) 613. The system may also include a third computer third computer program module for mitigating drilling failures when drilling the second well 615.

[0078] Figure 7The use of a drilling apparatus for drilling a plurality of wells 701 is shown, the drilling apparatus comprising a drilling rig 101 with a first drill string in a wellbore 102 for drilling a first well, the first drill string having a string neutron sensor 110 associated with the drill string for acquiring a first time series 703. A second sensor 103 is associated with the drill string in the wellbore 102, wherein the second sensor is on or near the drilling rig 101 at the surface of the earth, the second sensor for acquiring a second time series 705. A bottom hole assembly 112, 114, 116 is associated with the drill string to provide data to determine a geometry of the wellbore associated with the drill string in the wellbore 102. The apparatus comprises a first computer program module for determining model parameters (alpha, beta and gamma) using the first time series, the second time series and the geometry of the wellbore to derive model parameters alpha and beta 709 characterizing power losses of signal propagation of a signal traveling through the drill string in the wellbore. The second well may be drilled as follows, wherein the drill string in the wellbore 104 does not include a mid-string drilling subunit 711. The bottom hole assembly 112, 114, 116 may be associated with the second drill string in the wellbore 104 to provide data to determine the geometry of the second wellbore 713 and to provide time series data for comparison with sensors associated with the drill string on the surface 103, thereby providing a third time series to derive signal power loss along the drill string in the wellbore and to determine drilling failures while the well is being drilled 715. After deriving the parameters alpha and beta, these parameters may be used in drilling the second well, wherein the geometry data of the second well, the third time series data (such as from the sensor 103) combined with the time series data provided by the BHA to derive power loss information about the second wellbore may be inverted to detect and then mitigate drilling failures in the drilling operation. Additionally, the second computer program module may determine a parameter gamma, which along with alpha and beta may be used to characterize the power loss of signal propagation of a signal traveling in the first or second drill string. Using a combination of these parameters, the fault detection computer program module can determine the dynamic state of the second drill string in the wellbore.When a drilling fault is detected, measures for mitigating the fault can be taken.

[0079] Embodiments disclosed herein also include synchronizing time between many different sensors and the types of data that may be acquired by the analytics engine, such as in Figure 2 Processing diagram 200 in FIG. The drilling industry needs to optimize downhole data acquisition operations that properly synchronize or correct for timing differences between different time series measurements. In the field, laborious manual operations are used to synchronize or adjust the time differences between surface and downhole sensors. However, these manual time adjustment operations are not only slow, but are also known to be prone to potential human errors during the field data acquisition phase.

[0080] For example, each sensor may have its own internal clock. In an ideal world, field operations would be able to synchronize the clocks of all surface and downhole sensors simultaneously to ensure that each clock starts at the same time and / or that all time differences are known. However, in reality, synchronization is not accomplished during field operations. Surface sensors are often not synchronized or cannot be synchronized with downhole sensors, or the clocks of downhole sensors start at different times. This non-synchronization of the clocks creates a time misalignment between the surface and downhole measurements. This timing error may range from a few minutes to several hours.

[0081] Another major source of timing errors is related to clock drift at each sensor, where a clock or timer associated with a sensor does not run at the same speed as another clock. That is, after a period of time, the clock "drifts away" in time from the other clock. The timing of different sensors can drift relative to other timing devices for any number of reasons, including physical composition, temperature, pressure, power changes, and timer quality. Timing drift can vary arbitrarily at different rates. Timing errors from clock drift can range from a few seconds to several minutes.

[0082] In order to correct the timing errors caused by the desynchronization of the clocks, the drilling industry usually uses manual methods to correlate the downhole data with the surface data, assuming that the surface data is a reference signal, because the surface data is always available, usually convenient to use and synchronized with the master clock, so it is usually most convenient to use the clock associated with the surface as the reference signal. However, the manual method is laborious, prone to error, and depends on personal judgment and preference and is less accurate. Because the clock offset is difficult to determine manually, the drilling industry often ignores this correction or approximates it.

[0083] In order to avoid manual correction of timing errors, embodiments disclosed herein provide one or more automated methods to correct time misalignment between surface and downhole data. After correction, all measurements are correctly represented relative to a reference clock, and therefore all measurements are substantially synchronized in time. Substantially synchronized in time will be understood to mean within one or two standard deviations of measurement errors. This facilitates simple and accurate comparisons between all sensors and data sets. The application of time adjustment consists of three key corrections: 1) correction of clock asynchrony based on a cross-correlation method, 2) correction of clock offset based on a dynamic cross-correlation method or a dynamic time warping method, and 3) travel time path correction between surface and downhole sensors based on a "linear time difference correction". The benefits of this multi-step application produce accurate corrections to timing errors and greatly speed up processing time, which avoids the laborious and error-prone methods currently used in the drilling industry. After correction, all measurements are correctly represented relative to a reference clock.

[0084] The following outlines a framework for automatic correction of timing errors required to compensate downhole data. There are many time migration methods that can be applied to calculate time corrections. For example, it may be preferred to initially use a time migration method based on cross-correlation. In signal processing, cross-correlation is a measure of the similarity of two waveforms as a function of time delay, which gives a measure of the time adjustment that can be applied to one of them. For discrete real-time series of f(t) and g(t), the cross-correlation is defined as (Oppenheim and Schafer, 1989; Telford et al., 1976): where C(τ) refers to the cross-correlation function, τ is the displacement of g(t) relative to f(t), called the time delay, Δt is the time sampling rate, and n is the time sampling index.

[0085] In some embodiments, the data segments use time intervals for drilling at least 2 drill pipe positions. Each drill pipe is approximately 90 feet. It typically takes 3 to 5 hours to complete drilling 2 drill pipe positions. When there is a newly added drill pipe, the value of the time series usually decreases to zero, thereby generating a step function. The cross-correlation of the time series including those step functions produces an accurate and robust estimate of the time correction.

[0086] For example, f(t) may correspond to surface data and g(t) represents downhole data. The time offset is obtained by maximizing the cross-correlation function C(τ). The time offset is applied to all data to correct for the desynchronization of all clocks with a reference clock (typically a surface clock). As an example, the length of data (NΔt) included in the cross-correlation process may sometimes be about 3 to 5 hours, but of course varies depending on the situation. This process is repeated until the end of the data set.

[0087] Figure 8 Examples of time series data before and after time correction of the data are shown, with a surface clock as a reference. Time series 801 is transducer data representing revolutions per minute (RPM) measured at the surface and associated with a surface reference clock. Time series 803 is transducer data obtained from a sensor in the wellbore, associated with a drill string, also measuring RPM. The addition of drill pipe occurred at around 75 minutes, showing an illustrative example of a step function. After applying the cross correlation as described, a time offset is obtained to be applied to adjust the time of the wellbore sensor RPM data relative to the surface time series associated with the reference time. Time series 811 is the same transducer RPM time series data 801 associated with a surface reference clock, and time series 813 is the wellbore sensor RPM data after the time adjustment determined from the cross correlation has been applied.

[0088] Another method that can be used to correct clock offsets additionally uses a dynamic cross-correlation method that is similar to the cross-correlation method. The key difference is that a smaller overlapping time window is used to calculate the time offset. For example, a typical window size for dynamic cross-correlation is 30 minutes with a 50 percent overlapping window; however, the overlap will depend on the situation and amount of clock offset.

[0089] Another method for correcting clock drift uses a dynamic time warping method (Hale, 2013), which calculates a sample-by-sample time drift. This method can produce a good match between surface and downhole measurements. Fig. 9 Examples are shown before and after clock offset correction of downhole data relative to a surface reference clock by a dynamic time warping method. Time series 901 is transducer data representing RPM measured at the surface, associated with a surface reference clock. Time series 903 is transducer data obtained from a sensor in the wellbore, associated with a drill string, also measuring RPM. After applying dynamic time warping as described, an adjusted time offset sequence is obtained to be applied to adjust the time of the wellbore sensor RPM data. Time series 911 is the same transducer RPM time series data 901 associated with a surface reference clock, and time series 913 is the wellbore sensor RPM data after the time adjustment determined from the cross correlation has been applied.

[0090] Another time adjustment may be added because the downhole sensor locations vary in depth. For sensors associated with the drill string, the linear time difference correction takes into account the travel time of the signal from one sensor's depth location to the next sensor and / or the surface. The correction ΔT is calculated as follows: ΔT = Z / V, where Z is the distance from the downhole sensor location to the surface, and V is the signal propagation speed, which may be the speed of the steel pipe, the drill string, or the speed of the signal through the conductor of the wiring pipe. The ΔT correction is dynamic and increases with the depth of the sensor.

[0091] Fig.10 Examples are shown before and after linear moveout correction of data acquired from a downhole transducer, in this case an accelerometer. Time series 1001, 1003, and 1005 are downhole acceleration time series data acquired from a sensor in a wellbore, such as in or on a drill string. After applying a linear moveout time adjustment correction as described, time series data 1011, 1013, and 1015 are shown such that the data is significantly closer to being synchronized in time relative to a reference time associated with, for example, the surface. Other time adjustments, such as cross-correlation or time warping methods, may be added after this linear moveout correction.

[0092] Fig.11A method according to an embodiment of the present disclosure is presented for automatically adjusting time series data relative to a reference time. A first time series is acquired from a downhole sensor 1101. A reference time series is acquired, which may be acquired using a time series associated with a surface transducer (having a known relationship to the reference time) 1103. A linear time difference time series is determined to adjust the first time series due to the variable depth of the downhole sensor. The linear time difference time offset adjustment is equal to the depth of the downhole sensor divided by the signal propagation velocity 1105. Then, a linear time difference offset correction may be applied to the first time series 1107. The first time series and the reference time series may be cross-correlated to determine a cross-correlation time offset correction to apply to the first time series 1109, and the cross-correlation time offset correction is applied to obtain a cross-correlation corrected time series 1111.

[0093] Fig.12 A method according to an alternative embodiment of the present disclosure is presented for automatically adjusting time series data relative to a reference time. A first time series is acquired from a downhole sensor 1201. A reference time series is acquired, which may be acquired using a time series associated with a surface transducer (having a known relationship to the reference time) 1203. A linear time difference time series bias adjustment is determined to adjust the first time series due to the variable depth of the downhole sensor. The amount of the linear time difference time bias adjustment is equal to the depth of the downhole sensor divided by the signal propagation velocity or drill string 1205. The linear time difference time bias adjustment is applied to the first time series to obtain a time difference corrected time series 1207. The first time series and the reference time series are cross correlated to determine a cross correlation time correction to apply to the first time series 1209. The cross correlation time correction is applied to the first time series 1211 to obtain a cross correlation corrected time series. In order to correct for clock-sensor offset, a dynamic cross correlation may be applied to the first time series with the reference time series to obtain a dynamic cross correlation time bias adjustment to apply to the first time series 1213. Alternatively, a dynamic time warping process may be used to determine an adjustment value to the data for clock offset.A dynamic cross-correlation time offset adjustment is applied to the cross-correlation corrected time series to obtain a dynamically adjusted time series 1215. Where a dynamic time warping adjustment has been determined, it may be applied to the first time series.

[0094] Fig.13A method according to another embodiment of the present disclosure is presented for automatically adjusting time series data relative to a reference time. A first time series is acquired from a sensor in a wellbore 1301. A reference time series is acquired, which may be acquired using a time series associated with a surface transducer (having a known relationship to the reference time) 1303. A linear time difference time series bias adjustment is determined to adjust the first time series due to a variable depth of the downhole sensor. The amount of the linear time difference time bias adjustment is equal to the depth of the downhole sensor divided by the signal propagation velocity or drill string 1305. The linear time difference time bias adjustment is applied to the first time series to obtain a time difference corrected time series 1307. Dynamic time warping may be applied to the first time series relative to the reference time series to determine a series of dynamic time warping bias adjustments to apply to the first time series 1309. The sequence of dynamic time warping bias adjustments is then applied to the first time series to obtain a dynamically adjusted time series 1311.

[0095] Fig.14 A schematic diagram of an embodiment of a system 1400 is shown, which may correspond to or may be a part of a computer and / or any other computing device (such as a workstation, server, mainframe, supercomputer, process map and / or database). The system 1400 includes a processor 1402, which may also be referred to as a central processing unit (CPU). The processor 1402 may communicate with and / or provide instructions to other components within the system 1400, such as an input interface 1404, an output interface 1406 and / or a memory 1408. In other embodiments, the processor 1402 may include one or more multi-core processors and / or a memory (e.g., a cache memory) that operates as a buffer and / or data storage. In alternative embodiments, the processor 1402 may be part of one or more other processing components, such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), and / or a digital signal processor (DSP). Although Fig.14 It is shown that processor 1402 can be a single processor, but it will be understood that processor 802 is not limited thereto and can represent multiple processors, including massively parallel implementations and processing graphs including mathematical operators connected by data flows distributed across multiple platforms (including cloud-based resources). Processor 1402 can be configured to implement any of the methods described herein.

[0096] Fig.14It is shown that the memory 1408 can be operatively coupled to the processor 1402. The memory 1408 can be a non-transitory medium configured to store different types of data. For example, the memory 1408 can include one or more memory devices, including auxiliary storage, read-only memory (ROM) and / or random access memory (RAM). The auxiliary storage typically includes one or more disk drives, optical drives, solid-state drives (SSDs) and / or tape drives, and is used for non-volatile data storage. In some instances, if the allocated RAM is not large enough to accommodate all working data, the auxiliary storage can be used to store overflow data. The auxiliary storage can also be used to store such programs loaded into the RAM when the program is selected for execution. ROM is used to store instructions and perhaps data read during program execution. ROM is a non-volatile memory device, typically having a smaller storage capacity relative to the larger storage capacity of the auxiliary storage. RAM is used to store volatile data and perhaps for storing instructions.

[0097] As in Fig.14 As shown in , the memory 1408 can be used to hold instructions for executing the various embodiments described herein. In an embodiment, the memory 1408 can include a computer program module 1410 that can implement a computer program product that can be accessed and implemented by the processor 1402. Alternatively, an application interface 1412 can be stored within the memory and accessed by the processor 1402. In particular, the program module or application interface can perform signal processing and / or conditioning and apply the analysis method to the time series data described herein.

[0098] Programming and / or loading executable instructions onto the memory 1408 and the processor 1402 to convert the system 1400 into a specific machine or device that operates on time series data is known in the art. Instructions, real-time monitoring, and other functions can be converted into hardware implementations by loading executable software into a computer. For example, the choice between implementing an idea in software or hardware can depend on the number of design options, including the stability of the design, the number of units to be produced, and the problems involved in the transition from the software domain to the hardware domain. Typically, the design can be developed and tested in software form and then converted into an equivalent hardware implementation in an ASIC or a dedicated hardware that hard-wires the software instructions by known design rules. In the same way that a machine controlled by a new ASIC is a specific machine or device, similarly, a computer that has been programmed and / or loaded with executable instructions can be considered a specific machine or device.

[0099] in addition, Fig.14It is shown that the processor 1402 can be operatively coupled to an input interface 1404 configured to obtain time series data and an output interface 1406 configured to output and / or display results or transmit the results to other processing processes. The input interface 1404 can be configured to obtain time series data via sensors, cables, connectors and / or communication protocols. In one embodiment, the input interface 1404 can be a network interface including multiple ports, which are configured to receive and / or send time series data via a network. Specifically, the network can send the collected time series data via a wired link, a wireless link and / or a logical link. Other examples of the input interface 1404 can be a universal serial bus (USB) interface, a CD-ROM, a DVD-ROM. The input interface 1406 can include, but is not limited to, one or more connections for an image display (e.g., a monitor) and / or a printing device that produces a hard copy of the generated results.

[0100] As in Fig.15 As shown in , according to a non-limiting embodiment of the present disclosure, a system 1501 for determining a drilling rig status during a wellbore drilling operation is provided, the system comprising: a computer (1400) comprising a memory (1408) and a processor (1402) 1503; a plurality of sensors (103, 110) associated with the wellbore drilling operation 102, 104 for acquiring time series data, wherein the data is formatted for sample and bandwidth regularization and is corrected in time to provide substantially time synchronized data 1505, a processing diagram of a networked mathematical operator (1506) for processing the data stream; Figure 2 ), the processing diagram applies a continuous analysis method to the data at least as fast as the data is acquired to determine a dynamic state 1507 of a plurality of rig conditions associated with the wellbore drilling operation, and determines a rig state 1509 based on the plurality of rig conditions.

[0101] Other aspects of the system may include a data stream 1511 of the output of the processing map that detects and quantifies drilling failures. The processing map may also output a data stream 1513 of rig control instructions for the purpose of mitigating drilling failures. The rig control instructions may be to change RPM, change weight on bit, change pump pressure, or change top drive rotation parameters 1515. The process may also output a characterization of the tortuosity of the wellbore 1517. The acquired time series data input to the processing map may be rotary drilling measurements, sliding measurements, reaming measurements, reverse reaming measurements, or tripping related measurements 1519. The processing map output may be an energy loss correction or a weight on bit correction 1521 of the measurements obtained from the surface.

[0102] In other non-limiting embodiments, a portion of Fig.16In the present invention, a method 1601 for determining a rig state of a drilling rig during a wellbore drilling operation includes: collecting data from a plurality of sensors associated with the wellbore 1603, formatting the collected data for sample and bandwidth regularization 1605, time-correcting the data to provide substantially isochronously sampled data from the plurality of sensors 1607, processing the collected data through a processing graph of networked mathematical operators, the processing graph applying a continuous analysis method to the data at least as fast as the data is collected to determine a dynamic condition of a plurality of rig operations associated with the wellbore 1609, and determining a rig state based on the plurality of rig operation conditions 1611.

[0103] In other aspects, determining the rig status also includes detecting and quantifying drilling failures 1613. The processing map can output a data stream of rig control instructions to change the rig operation for the purpose of mitigating drilling failures 1615. The rig control instructions can be to change RPM, change WOB, change pump pressure, or change top drive rotation parameters 1617. The output of the processing map can also be a characterization of the tortuosity of the wellbore 1619. The acquired data input to the processing map can be rotary drilling measurements, drill string sliding measurements, reaming measurements, reverse reaming measurements, or tripping related measurements 1621. Other output data streams from the processing map can be energy loss corrections or bit pressure corrections 1623 of the input data streams obtained from the surface.

[0104] In still another non-limiting embodiment, a drilling rig apparatus for mitigating drilling failures includes a drilling rig associated with a plurality of sensors that provide time series data to a surface-based aggregator, wherein the data is formatted for sample and bandwidth normalization and is temporally corrected to provide substantially time-synchronized data; a computer comprising a memory and a processor; a processing graph of data stream networked mathematical operators that applies a continuous analysis method at least as fast as the time series is acquired to determine the dynamic conditions of a plurality of drilling rig conditions associated with a wellbore drilling operation; and detecting drilling failures based on the plurality of drilling rig conditions.

[0105] In other non-limiting aspects of the apparatus, the processing map outputs a data stream of rig control instructions for the purpose of mitigating the detected drilling fault. The rig control instructions may be for changing RPM, changing bit weight, changing pump pressure, or changing top drive rotation parameters. The output data stream of the processing map may quantify the drilling fault.

[0106] In yet another non-limiting embodiment, a computer program product is implemented in a non-transitory computer-readable medium, the computer program product being adapted to perform a method for mitigating drilling failures during a wellbore drilling operation, the method comprising: acquiring data from a plurality of sensors associated with the wellbore drilling operation; formatting the acquired data for sample and bandwidth regularization; time-correcting the data to provide substantially synchronously sampled data from the plurality of sensors; processing the acquired data through a processing graph of networked mathematical operators, the processing graph applying a continuous analysis method to the data at least as fast as the data is acquired to determine dynamic conditions of a plurality of rig operations associated with the wellbore; detecting drilling failures based on the plurality of rig operation conditions; and outputting rig control instructions to mitigate the detected drilling failures.

[0107] In summary, it should be noted that the discussion of any reference is not an admission that it is prior art to the present invention, especially any reference that may have a publication date after the priority date of this application. At the same time, each claim below is hereby incorporated into the detailed description or specification as an additional embodiment of the present invention.

[0108] Although the systems and methods described herein have been described in detail, it should be understood that various changes, substitutions and modifications may be made without departing from the spirit and scope of the present invention as defined in the following claims. Those skilled in the art may be able to learn preferred embodiments and discern other ways to practice the present invention that are not identical to those described herein. It is the intention of the inventor that variants and equivalents of the present invention fall within the scope of the claims, and the description, abstract and drawings should not be used to limit the scope of the present invention. The scope of the present invention is particularly intended to be as broad as the scope of the following claims and their equivalents.

Claims

1. A system for determining a drilling rig status of a drilling rig during a wellbore drilling operation, the system comprising: a plurality of surface sensors at the surface of the earth and associated with the wellbore drilling operation, the plurality of surface sensors acquiring time series data, wherein the time series data is formatted for sample and bandwidth normalization and is corrected in time to provide time synchronized data; a processing graph of data stream networked mathematical operators that applies a continuous analysis method to the time synchronized data at least as fast as the time series data is acquired by the plurality of surface sensors to determine parameters of a power loss model of the data of the plurality of surface sensors and further determine downhole dynamic conditions of a plurality of rig conditions associated with the wellbore drilling operation and drilling faults detectable from the downhole dynamic conditions of the plurality of rig conditions, wherein the parameters of the power loss model are determined based on attenuation of signals from the plurality of surface sensors to a source point downhole, wherein the parameters of the power loss model include parameters for controlling control losses along vertical sections and parameters for controlling a tradeoff between exponential and hyperbolic signal attenuation for a given bend function or wellbore geometry; as well as A drilling rig having a drilling rig status determined based on the plurality of drilling rig conditions.

2. The system of claim 1, wherein the processing graph outputs a data stream of drill rig control instructions and mitigates drilling failures.

3. The system of claim 2, wherein the rig control command is at least one selected from the group consisting of: i) changing RPM, ii) changing weight on bit, iii) changing pump pressure, and iv) changing top drive rotation parameters.

4. The system of claim 1, wherein the process map outputs a data stream characterizing the tortuosity of the wellbore.

5. The system of claim 1, wherein the time series data input to the processing map includes at least one item from the group consisting of: i) rotary drilling measurements, ii) sliding measurements, iii) reaming measurements, iv) reverse reaming measurements, and v) tripping-related measurements.

6. The system of claim 1, wherein the processing map outputs a data stream selected from at least one of the group consisting of: i) an energy loss correction of an input data stream derived from the surface, and ii) a weight-on-bit correction.

7. A method for determining a drilling rig status of a drilling rig during a wellbore drilling operation, the method comprising: collecting surface data from a plurality of surface sensors associated with the wellbore; formatting the surface data for sample and bandwidth regularization; temporally correcting the surface data to provide isochronously sampled data from the plurality of surface sensors; processing the isochronously sampled data from the plurality of surface sensors through a processing graph of networked mathematical operators, the processing graph applying a continuous analysis method to the isochronously sampled data in real time to determine parameters of a power loss model of the data from the plurality of surface sensors, and further determine downhole dynamic conditions for a plurality of drilling rig operating conditions associated with the wellbore, wherein the parameters of the power loss model are determined based on attenuation of signals from the plurality of surface sensors to a source point downhole, wherein the parameters of the power loss model include parameters for controlling control losses along vertical sections and parameters for controlling a tradeoff between exponential and hyperbolic signal attenuation for a given bend function or wellbore geometry; as well as The drilling rig status is determined based on the drilling rig operating conditions, wherein determining the drilling rig status includes detecting and quantifying drilling faults.

8. The method of claim 7, wherein the process map outputs a data stream of drill rig control instructions to change drill rig operation and mitigate the drilling fault.

9. The method of claim 8, wherein the rig control command is at least one of the group consisting of: i) changing RPM, ii) changing weight on bit, iii) changing pump pressure, and iv) changing a top drive rotation parameter.

10. The method of claim 7, wherein the process map outputs a data stream characterizing the tortuosity of the wellbore.

11. The method of claim 7, wherein the acquired data input to the processing map is at least one selected from the group consisting of: i) rotary drilling measurements, ii) drill string sliding measurements, iii) reaming measurements, iv) reverse reaming measurements, and v) tripping-related measurements.

12. The method of claim 7, wherein the processing map outputs a data stream selected from at least one of the group consisting of: i) an energy loss correction of an input data stream derived from the surface, and ii) a weight-on-bit correction.

13. A drilling rig device for mitigating drilling failures, wherein the device comprises: a drilling rig associated with a plurality of surface sensors that provide time series data to a surface-based aggregator, the time series data being formatted for sample and bandwidth normalization and time-corrected to provide time-synchronized data; as well as A processing graph of data stream networked mathematical operators that applies a continuous analysis method at least as fast as the time series data is acquired by the plurality of surface sensors to determine parameters of a power loss model of the data of the plurality of surface sensors, and further determine downhole dynamic conditions of a plurality of rig conditions associated with the wellbore drilling operation and drilling faults that can be determined from the plurality of rig conditions, wherein the parameters of the power loss model are determined based on the attenuation of signals from the plurality of surface sensors to a source point downhole, wherein the parameters of the power loss model include parameters for controlling control losses along vertical sections and parameters for controlling the tradeoff between exponential and hyperbolic signal attenuation for a given bend function or wellbore geometry.

14. The apparatus of claim 13, wherein the processing graph outputs a data stream of drill rig control instructions and mitigates drilling failures.

15. The apparatus of claim 14, wherein the rig control command is at least one of the group consisting of: i) changing RPM, ii) changing weight on bit, iii) changing pump pressure, and iv) changing a top drive rotation parameter.

16. The apparatus of claim 13, wherein the output data stream of the processing graph includes quantifying the drilling failure.

17. A computer program product embodied in a non-transitory computer readable medium, the computer program product being adapted to perform a method for mitigating drilling failures during a wellbore drilling operation, the method comprising: collecting surface data from a plurality of surface sensors associated with the wellbore drilling operation; formatting the surface data for sample and bandwidth regularization; temporally correcting the surface data to provide synchronously sampled data from the plurality of surface sensors; processing the isochronously sampled data from the plurality of surface sensors through a processing graph of networked mathematical operators, the processing graph applying a continuous analysis method to the isochronously sampled data in real time to determine parameters of a power loss model of the data from the plurality of surface sensors, and further determine downhole dynamic conditions for a plurality of rig operating conditions associated with the wellbore drilling operation, wherein the parameters of the power loss model are determined based on attenuation of signals from the plurality of surface sensors to a source point downhole, wherein the parameters of the power loss model include parameters for controlling control losses along vertical sections and parameters for controlling a tradeoff between exponential and hyperbolic signal attenuation for a given bend function or wellbore geometry; detecting a drilling fault based on the plurality of drilling rig operating conditions; and outputting drill rig control instructions to mitigate the drilling failure.

Images