Systems and methods for the detection of key seating wear in underwater drilling

Abstract

Systems and methods of detecting key seating wear during drilling operations include a sensor for detecting vibrations in underwater drilling structures and a controller connected to the sensor. The controller processes the data from the sensor to determine if key seating wear is occurring. There may also be a signal injector connected to the controller that is used to inject a signal into the underwater drilling structures. This signal can also be detected by the sensor for use by the controller in determining if key seating wear is present.

Description

SYSTEMS AND METHODS FOR THE DETECTION OF KEY SEATING WEAR IN UNDERWATER DRILLINGCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority of U.S. Patent Application No. 18 / 526,655, filed on December 1, 2023, which is hereby incorporated herein in its entirety by reference.BACKGROUND

[0002] This disclosure relates to the field of drilling technology. In particular, but not exclusively, this disclosure is related to the detection of key seating wear to prevent key seating in offshore drilling. Key seating wear is the result of unintentional wear caused by contact between a drill pipe and the inner diameter of a marine drilling riser and / or the inner diameter of the subsea well production casing.

[0003] Offshore drilling is the general term for drilling operations used herein that involve drilling a wellbore underwater, for example, into a seabed, into an ocean floor, or under any other body of water, including underwater in a river, dam, or lake. This process takes place in waters of varying depths. A drilling platform is maintained on the surface of the water, with a drilling riser extending from the drill platform to a wellhead that is fixed to the seabed below the surface. The drilling riser is a hollow structure that contains the drill pipe, which is the rotating element that drives the drill bit that forms the well in the seabed. The drill pipe is driven by machinery on the drilling platform.

[0004] There are generally two types of drilling platforms: (1) fixed or founded platforms that are fixed to the seabed, and (2) floating platforms that are not fixed to the seabed. Floating platforms can be anchored to the seabed in shallower waters, but they are more generally used in deeper water situations where fixing the drilling platform to the seabed is not feasible. In these situations, the drilling platform is maintained at a setpoint by dynamic positioning, which uses the engines and thrusters of the drilling platform to hold the drilling platform in the desired position on the surface.

[0005] Ideally, the drilling platform is positioned at an optimum setpoint to keep the drilling riser as vertical as possible. This ensures that the drill pipe that is carried within the riser does not contact the inner walls of the drilling riser. In reality, several factors can act to push the drilling platform, and / or drilling riser, out of alignment with the fixed wellhead,including sea conditions, wind, and underwater currents. When this happens, the drilling riser will flex into a curve. This curvature can, in turn, bring the drilling pipe, which is passed through the drilling riser, into contact with the inner walls of the riser and / or parts of the subsea well structure. This contact and rotation of the drill pipe can result in accelerated wear of the drilling riser and other components connected to the riser. This phenomenon is called key seating and can result in damage to the drilling riser / subsea well equipment and, ultimately, equipment or drill pipe failure. This problem is significant because repairing a drilling riser / subsea well structure at a deep-water drilling project can require a multi-week stop in drilling operations at very substantial costs. Ideally, key seating wear can be detected as it initially occurs, and drilling operations can be paused until the drilling platform is realigned.

[0006] Some drilling platforms include sensors that can determine the amount of misalignment that the drilling riser has with respect to the wellhead. This is typically expressed as an angle of the riser with respect to the fixed wellhead. However, this misalignment is only partially helpful in avoiding key seating because the relationship between the misalignment angle and the onset of key seating can vary depending on the parameters of the drilling operation. For example, in one drilling operation, key seating may not occur until the misalignment angle exceeds two degrees. In other drilling operations, the key seating may occur at angles below one degree. There is no certain way to predict what angle will result in key seating in advance with any accuracy. This uncertainty makes it difficult to address key seating concerns by stopping drilling operations based on the misalignment angle.

[0007] Some drilling platforms are also able to sample the lubricating fluid, or mud, which is passed through the drilling riser to determine if there are any metal particles present. These particles may indicate the presence of key seating. However, because continuous sampling is logistically challenging on an offshore drilling platform, and the presence of metal particles indicates that key seating has already occurred, which makes this technique reactive instead of preemptive.

[0008] Therefore, there is a need for systems and methods of detecting and addressing key seating as it occurs in drilling operations. Preferably, the system will be adaptable to various drilling operations and will not require substantial reworking of existing drilling infrastructure.BRIEF SUMMARY

[0009] The disclosed invention is set out in the accompanying claims. Below are some embodiments of the disclosed technology.

[0010] According to a first aspect of the disclosed technology, a system for detecting key seating wear during drilling operations includes a sensor configured to detect vibrations in an underwater drilling structure and a controller operably connected to the sensor, the controller configured to receive data from the sensor and to process the data to detect key seating wear.

[0011] In some embodiments, the data may be processed to predict key seating instead or additionally to being processing to detect key seating wear.

[0012] According to another, second, aspect of the disclosed technology, a method for detecting key seating wear in drilling operations includes recording vibration data in an underwater structure during drilling operations using a sensor disposed on the underwater structure, and processing the vibration data from the sensor at a controller operably connected to the sensors to detect key seating by filtering the data to detect an anomaly that corresponds to key seating.

[0013] In some embodiments, the anomaly that corresponds to key seating wear may instead or additionally comprise an anomaly which predicts key seating.

[0014] By way of example, if a drill string is resting against an inner wall of a riser in the same location and over time results in wear, then the wear may cause key seating as a "seat" or groove is formed by the wear which may cause future operations to "catch" the drill string in the same groove. This then leads to even more wear in the "seat". Accordingly, key seating wear is often self-amplifying and as it often leads to key seating, it is advantageous to prevent this from happening as early as possible. Accordingly, a technical benefit of the disclosed technology is that it may allow vibrations from the key seating wear to be detected at an early stage, well before or just before any key seating happens. Accordingly, the disclosed technology seeks to provide a way of detecting key seating wear, which may in some embodiments enable corrective actions to be implemented to prevent the key seating wear resulting in a key seating event.

[0015] In some embodiments, the sensor is one of a plurality of sensors, each of the plurality of sensors being configured to detect vibrations in an underwater drilling structure and operably connected to the controller.

[0016] In some embodiments, the controller is configured to send an alert to a user when key seating and / or key seating wear is detected.

[0017] In some embodiments of the system, the system further includes a display operably connected to the controller, the controller being configured to use the display to alert the user to key seating and / or key seating wear.

[0018] In some embodiments, the controller is configured to use the sensor to record a baseline data set when key seating and / or key seating wear is not present.

[0019] In some embodiments, the controller is configured to use the baseline data to determine if key seating and / or key seating wear is detected.

[0020] In some embodiments, the sensors record vibrational data at a location close to the position where the drilling operation is occurring.

[0021] In some embodiments, in addition the controller processing the vibrational data is at a location close to where the drilling operation is occurring and / or to the vibrational recorder.

[0022] The location of the sensor data recording and the controller close to each other and / or to the location of the drilling operation allows higher fidelity data to be captured and processed with reduced latency than if they were located further away and / or top-side onboard the drilling operation platform.

[0023] In some embodiments, previously recorded baseline sensor data and newly recorded sensor data generated during a drilling operation may be combined or otherwise form statistical data which is transmitted or otherwise provided to a different computational platform, for example, to a topside or on-board computer system for further processing.

[0024] In some embodiments, the baseline data is compared with newly recorded data by the controller using a machine learning model trained to filter newly recorded data using baseline data.

[0025] In some embodiments, the machine learning model is hosted on a topside computer system.

[0026] In some embodiments, the method further comprises inputting a time-series of recorded base-line vibration data from the sensor into a machine learning, ML, model, to train the ML model to output one or more predictions indicating the input sensor data does not comprise a key seating wear or key seating anomaly.

[0027] In some embodiments, rig process data is also fused or otherwise correlated with the timing information of the base-line vibrational sensor data and used to train the ML model.

[0028] In some embodiments, once trained, the ML model is configured to compare vibrational data recording input to the ML model during a drilling operation in real-time with recorded baseline vibrational data and to output one or more indications of an anomaly corresponding to key seating or key seating wear if an anomaly is inferred using the ML model.

[0029] In some embodiments, the indications of the anomaly output by the ML model are generated in real-time or near-real time as the vibrational data is sensed by the sensor and provided via the controller as input to the ML model.

[0030] In some embodiments, the method further comprises when the output of the ML model indicates a key seating or key seating wear anomaly, automatically causing the drilling operation to be paused or stopped.

[0031] In some embodiments, the underwater drilling structure is selected from the group consisting of a drilling riser and a lower marine riser package.

[0032] In some embodiments of the system, the system includes a signal injector controlled by the controller and configured to transmit a vibrational signal into a drill pipe.

[0033] In some embodiments, the signal injector is configured to inject the vibrational signal into the drill pipe at a plurality of locations. This allows, based on the sound travel velocity in steel or whatever material the drill pipe is composed of, to determine the location of any key seating wear and / or key seating.

[0034] In some embodiments, the ML model may be trained on injected signals to locate where any key seating wear or key seating is along the drill pipe.

[0035] In some embodiments, the controller is configured to detect the vibrational signal using the sensor to detect key seating wear.

[0036] In some embodiments, the controller is configured to determine a position of the key seating by comparing the time of injection of the signal with the time of detection of the signal by the sensor.

[0037] According to another, third, aspect of the disclosed technology, an offshore drilling system, includes a drill platform; and any of the systems for detecting key seating wear described above.

[0038] According to another, fourth, aspect of the disclosed technology, a method for detecting key seating wear in drilling operations, includes recording vibration data in an underwater structure during drilling operations using a sensor disposed on the underwater structure; and processing the vibration data from the sensor at a controller operably connected to the sensors to detect key seating wear by filtering the data to detect an anomaly that corresponds to key seating wear. In some embodiments, recording vibration data further includes receiving data from a plurality of sensors, each of the plurality of sensors being configured to detect vibrations in an underwater drilling structure and operably connected to the controller.

[0039] In some embodiments, the method includes using a display operably connected to the controller to alert a user to key seating wear.

[0040] In some embodiments, the method includes using the sensor to record a baseline data set when key seating wear is not present.

[0041] In some embodiments, the method includes using the baseline data as part of the processing the vibration data to determine if key seating wear is detected.

[0042] In some embodiments, the method includes using a signal injector controlled by the controller to inject a vibrational signal into a drill pipe.

[0043] In some embodiments, the method includes detecting the vibrational signal using the sensor.

[0044] In some embodiments, the method includes determining a position of the key seating wear by comparing the detected vibrational signal to the signal injected by the signal injector.

[0045] In some embodiments, the method further comprises using the travel velocity of the vibrations in steel, or the travel velocity of the vibrations in whatever material the drill pipe is composed of, to determine the location along the drill pipe of one or more or all of : key seating wear, any predicted key seating, and actual key seating.

[0046] According to another, fifth, aspect of the disclosed technology, a key seating wear detector for offshore drilling, includes a sensor configured to detect vibrations in an underwater drilling structure, a controller disposed on a drill platform and operably connected to the sensor, wherein the controller is configured to receive data from the sensor and to process the data to detect key seating, and an alert device operably connected to the controller, wherein the controller is configured to use the alert device to alert a user to key seating wear.

[0047] In some embodiments, the alert device comprises a display.

[0048] In another, sixth, aspect of the disclosed technology, a computer-implemented method for detecting key seating or key seating wear in drilling operations comprises receiving at a controller operably connected to a sensor, from the sensor, recorded vibration data of an underwater structure used for drilling operations, during the drilling operations, wherein the sensor is disposed on the underwater structure, processing the vibration data from the sensor using a trained machine learning, ML, model, where the trained ML model is configured to receive sensor data from the controller operably connected to the sensors as input, and where the trained ML model is configured to detect key seating by processing received sensor data to determine if the received sensor data meets one or more conditions to be classifiable as an anomaly that corresponds to key seating or key seating wear; and

[0049] outputting, by the trained ML model, an indication of whether the input sensor data indicates a key seating or key seating wear anomaly

[0050] The computer-implemented method may be used by or form part of any of the methods disclosed herein for detecting key seating or key seating wear in recordings of vibrational data obtained from sensors attached to underwater structures during underwater drilling operations in real-time or near real-time.

[0051] In some embodiments, the method further comprises when the output of the ML model indicates a key seating anomaly, causing the drilling operation to be paused or stopped, either by a user or automatically.

[0052] In some embodiments, the method further comprises causing a pause or stop in the drilling responsive to detecting a key seating or key seating wear anomaly.

[0053] In some embodiments, the method further comprises causing a realignment of the drill platform used in the drilling responsive to detecting a key seating or key seating wear anomaly.

[0054] In some embodiments, the ML model is trained using data that has been preprocessed by filtering the sensor data.

[0055] In some embodiments, the time-series of sensor data is first pre-processed by running a Fast Fourier Transform, FFT, and the resulting power spectrum is input to the ML model.

[0056] In some embodiments, one or more additional types of pre-processing are performed before the sensor data is input to the ML model.

[0057] In some embodiments, the additional pre-processing comprises one or more of: normalizing the FFT power spectra, segmenting sequences within the time-series sensor data into sliding windows, performing a convolutional pre-processing technique, and applying a noise-reduction technique.

[0058] In some embodiments, the noise reduction technique comprises pre-filtering to remove frequencies which lie in frequency bands known to not be of interest for detecting a key seating signature event.

[0059] According to another, seventh, aspect of the disclosed technology, an apparatus for detecting key seating wear in drilling operations is provided, the apparatus comprising means for recording vibration data in an underwater structure during drilling operations using a sensor disposed on the underwater structure, and means for processing the vibration data from the sensor at a controller operably connected to the sensors to detect key seating by filtering the data to detect an anomaly that corresponds to key seating or key seating wear.

[0060] According to another, eighth, aspect of the disclosed technology, a computer program product is provided, the computer program product comprising a computer program, which when loaded into or executed on a computer comprising a controller, causes the computer to perform any one of the methods for detecting key seating and / or key seating wear in drilling operations or one or more embodiments of any one of the methods for detecting key seating and / or key seating wear disclosed herein.

[0061] According to another, ninth, aspect of the disclosed technology, a non-transitory computer-readable medium is provided on which instructions are encoded for carrying out any one of the method aspects or one or more of the embodiments thereof disclosed herein.

[0062] In some embodiments, the computer program is configured to output an indication of a detected anomaly in a time-series of sensed vibrational data obtained during a drilling operation associated with key seating or key seating wear.

[0063] In some embodiments, the indication is output in real-time during the drilling operation responsive to the anomaly being detected.

[0064] The disclosed embodiments and preferred aspects of the technology may be suitably combined with each other in any manner apparent to one of ordinary skill in the art.BRIEF DESCRIPTION OF THE DRAWINGS / FIGURES

[0065] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate the present disclosure and, together with the description, further serve to explain the principles thereof and to enable a person skilled in the pertinent art to make and use the same.

[0066] FIG. 1 is a schematic diagram of an offshore drilling system according to at least one embodiment of the disclosed technology.

[0067] FIG. 2 is a schematic of a portion of an offshore drilling system adjacent to a seabed according to at least one embodiment of the disclosed technology.

[0068] FIG. 3 is cross section view of drilling riser and drill pipe in a first configuration according to at least one embodiment of the disclosed technology.

[0069] FIG. 4 is a graph of data received by a sensor according to at least one embodiment of the disclosed technology

[0070] FIG. 5 is a schematic of a portion of an offshore drilling system adjacent a seabed according to at least one embodiment of the disclosed technology.

[0071] FIG. 6 is a cross-section view of a drilling riser and drill pipe in a second configuration according to at least one embodiment of the disclosed technology.

[0072] FIG. 7 is a graph of data received by a sensor according to at least one embodiment of the disclosed technology.

[0073] FIG. 8 is a system diagram of a key seating wear detection system according to at least one embodiment of the disclosed technology.

[0074] FIG. 9 is a schematic flowchart for a method of detecting key seating wear in drilling operations according to at least one embodiment of the disclosed technology.

[0075] In the drawings, reference numbers generally indicate identical or similar elements. Additionally, generally, the left-most digit(s) of a reference number identifies the drawing in which the reference number first appears.DETAILED DESCRIPTION

[0076] Reference will now be made in detail to representative embodiments illustrated in the accompanying drawings. References to “one embodiment,” “an embodiment,” “an exemplary embodiment,” etc., indicate that the embodiment described may include aparticular feature, structure, or characteristic, but every embodiment may not necessarily include the particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such a feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.

[0077] Systems and methods of the disclosed technology address the issue of key seating wear detection, at least in part by using sensors to detect vibration or acoustic frequencies in the drilling riser and related equipment. The detected signals are processed and anomalies representing key seating wear can be identified in real time, allowing drilling operations to be paused immediately. In some embodiments, the position of the key seating contact along the drilling riser or well can also be determined. These systems and methods have several advantages, including, but not limited to, technical advantages, such as allowing for immediate and accurate detection of key seating wear, while allowing for drilling operations to be continued until and unless key seating wear is detected. This improves uptime of drilling operations, and also substantially reduces the risks of damage caused by key seating on drilling equipment.

[0078] As shown in FIG. 1, a drilling system 100 includes a drill platform 101 located on the surface of the water 102. Drill platform 101 can be any type of water-based drilling platform. A so-called floating type drill platform 101 is generally more susceptible to key seating issues. Drill platform 101 functions to provide the platform for all drilling operations.

[0079] Extending into water 102 from the bottom of drill platform 101 is a drilling riser 103. Drilling riser 103 is a hollow structure, typically formed as a tube, that contains a drill pipe 105 (see FIGS. 2, 3, 5, and 6). Drill pipe 105 is the element that is rotated by suitable machinery on drill platform 101 to provide the rotational power to the drill bit (not shown) that creates the well.

[0080] Drilling riser 103 extends towards the seabed 108 where it meets with a blowout preventer stack (“BOP stack”) 107 that is connected above wellhead 106. These elements are fixed to seabed 108 and are the surface end of the well. BOP stack 107 can include all suitable equipment for interfacing with the end of wellhead 106, including valves, fittings, and blowout preventers. Present at the interface between drilling riser 103 and BOP stack 107 is a flex joint 104. Flex joint 104 accommodates misalignment between drilling riser103 and drill pipe 105 and wellhead 106. As seen in FIG. 1, there may also be a flex joint104 at drill platform 101 to perform a similar function.

[0081] FIG. 1 demonstrates a floating drill platform 101 that is not anchored or fixed to seabed 108. As seen in FIG. 1, drill platform 101 is not directly above wellhead 106. Instead, drill platform 101 is offset to the right in FIG. 1, which creates the curve seen in drilling riser 103 that results in the angular misalignment of drill pipe 105 discussed in the background with elements such as flex joint 104. FIG. 5 shows a schematic of the elements of drilling system 100 adjacent seabed 108 during such a misaligned state. Contrast FIG. 5 with FIG. 2, which shows the elements of drilling system 100 adjacent seabed 108 when there is alignment. The type of misalignment shown in FIGS. 1 and 5 can result in key seating wear.

[0082] FIGS. 3 and 6 are cross-sections of drilling riser 103 that show drill pipe 105. Drill pipe 105 rotates within drilling riser 103 to provide rotational torque to the drill bit in wellhead 106. Also present, but not shown in FIGS. 3 and 6, is heavy fluid that is circulated inside riser 103. This heavy fluid, often called mud, is used to maintain the proper pressure balance inside the well, provide sufficient lubrication, cooling, and debris removal for drilling operations.

[0083] FIG. 3 corresponds to FIG. 2 and shows drilling riser 103 in a normal operating mode where no key seating is occurring because drill platform 101 is aligned with wellhead 106. As shown, drill pipe 105 is not in contact with the inner wall of drilling riser 103.

[0084] FIG. 5 corresponds to FIG. 4 and shows multiple key seating points 109, which may also be referred to herein as key seating wear points 109, occurring with drill pipe 105 in contact with the inner wall of drilling riser 103 and other components of the underwater drilling system. It should be understood that drill pipe 105 will be rotating (as shown by the arrows in FIG. 5) in normal operation and will essentially grind away at the inside of drilling riser 103 or other components in the situation shown in FIG. 5. It also should be understood that key seating can occur at a single location in drilling riser 103 or the well, or at multiple locations along the length of drilling riser 103 or the well (or around the interior of drilling riser 103) at any given time. For example, in FIG. 5 there are four separate key seating points 109 being created. It should be understood that riser 103 (or an equivalent structure) extends through BOP stack 107 to wellhead 106. Key seating points 109 may shift depending on the parameters of the misalignment and drilling operations. For example, as the misalignment angle increases or decreases, the curve in drilling riser103 will increase or decrease, bringing different points or point of drill pipe 105 into contact with drilling riser 103.

[0085] As shown in FIG. 8, components of a system 116 for detecting key seating wear include sensor 110 and controller 112. Sensor 110 is a sensor that is attached to suitable underwater structures, such as drilling riser 103 or BOP stack 107, and is configured to detect vibrational frequencies. Sensor 110 can be, for example, an accelerometer or acoustic sensor (e.g., microphone) that is sensitive enough to detect the frequencies and amplitudes associated with drilling operations and key seating (or any other type of suitable sensor). For example, sensor 110 can detect frequencies from zero to twenty kilohertz or higher, for example, 50 kilohertz. There may be more than one sensor 110 present, as shown in FIGS. 1, 2, and 5. For example, in some embodiments, 3-axis sensors may be used. Each sensor 110 may be placed on a different structure to improve data collection across different structures. As shown in the Figures, two sensors 110 may be present, one attached to drilling riser 103 and one attached to BOP stack 107. This arrangement can improve data collection because drilling riser 103 and BOP stack 107 are separate structures. There may be additional sensors 110 positioned at other structures.

[0086] Sensor 110 is operably connected to controller 112, which may be disposed on drill platform 101 or may be disposed in or on a suitable subsea structure such as BOP stack 107. In some embodiments, controller 112 could be located remotely from drill platform 101, such as in a remote server that is connected to drill platform 101 by a suitable data network. Controller 112 is configured to receive the readings from sensor 110 and to process the data to determine if key seating wear is present. Controller 112 can be any suitable arrangement of processors and memory needed to carry out the relevant algorithms discussed below. Controller 112 can be operably linked to alert devices that can be used to receive an alert. These devices can include, for example display elements needed to display data and alerts to a user. In some embodiments, controller 112 can be operably linked to a data network on drill platform 101, such as a wired or wireless data network. This can allow controller 112 to send and receive data and to alert users using devices via the data network. In some embodiments, controller 112 can also use the display elements or data network to transmit other data, including graphical representations of vibration data or other outputs of the algorithms discussed here.

[0087] As discussed above, sensor 110 detects vibrational frequencies from one or more underwater structures (drilling riser 103 and BOP stack 107, for example). FIGS. 4 and 7show examples of hypothetical frequency measurements recorded by sensor 110. The graph shows a frequency spectrum analysis (e.g., Fast Fourier Transform (“FFT”) or the like), where the horizontal axis represents frequency, while the vertical axis represents amplitude. In some embodiments, spectrograms may be used to show frequency and amplitude over time, e.g., to identify potential times of key seating events. Data 120 shows a data set for a drilling operation where key seating is not occurring. In data 120, there is a vibration 122 detected at certain frequencies that correspond to the ordinary machinery noise during drilling operations. Data 124 shows a hypotheticals data set that is detected during key seating. Data 124 is aligned with data 120, and there is vibration 122 present in data 124 that corresponds with the vibration 122 present in data 120. However there is an additional anomaly 126 that is present in data 124 that is not present in data 120. Anomaly 126 illustrates hypothetical sensor data that shows the presence of key seating because the physical contact between drill pipe 105 and drill riser 103 creates the noise that is being detected as anomaly 126. As will be explained in detail below, detecting key seating via sensor 110 is accomplished by examining the data for unknown anomalies that do not correspond to known drilling vibrational data — this may be done by a specially purposed computer. This can involve determining or identifying specific frequency ranges that correspond to key seating wear, or it can involve filtering out the known vibrational data of drilling operations to leave only potential anomalies behind.

[0088] A method of detecting key seating using the system discussed above begins with obtaining data using sensor 110 or sensors 110. The data is transmitted to controller 112, which processes the recorded data. Processing the data involves application of frequency filters to separate out the expected noise (such as vibration 122), as shown in the example of data 120. In some embodiments the processing accounts for different operating states of the drilling system. For example, if controller 112 is also programmed to receive the current drilling system configuration, such as machinery that is currently in use, this data may be used to apply specific frequency filters based on the drilling system configuration. Any remaining data is then analyzed by controller 112 to determine whether it may represent key seating. This analysis may simply be confirmation that the remaining noise is outside of the expected frequency range. In other embodiments, the remaining noise may be compared to known examples of key seating data to determine if there is sufficient overlap between the recorded data and known key seating data to warrant a key seating alert. Controller 112 will then alert a user that key seating is suspected. As discussed above,controller 112 may use displays to alert the user, and / or transmit the alert using a data network. The user can then determine what action to take, such as pausing drilling operations while drill platform 101 is repositioned.

[0089] In some embodiments, it can be desirable to record data using sensor 110 of ordinary drilling operations where key seating is not suspected to have a baseline to compare sensor readings to in controller 112. This can be accomplished by recording data from sensor 110 using controller 112 when, for example, the misalignment angle of drilling riser 103 is known to be very low. In general, key seating is rare when misalignment is less than half of a degree (from zero degrees to half a degree of misalignment). Misalignment values over one and a half degree correspond to increasing chances of key seating, with values over two degrees corresponding to high chances of key seating. The baseline data can be stored and used to help filter any newly-recorded data for anomalies by comparison. In some embodiments, this process can be accomplished using machine learning. In these embodiments, baseline data can be processed using machine learning techniques to develop the algorithm that identifies anomalies that show key seating wear is occurring. For example, so-called training data that shows key seating wear occurring in various situations and operational configurations can be compiled with data that does not show key seating wear occurring. This data is identified and analyzed by the machine learning technique to develop an algorithm that can be used for new, unidentified data to determine if key seating wear is occurring.

[0090] Various time-series machine learning model architectures may be adapted for use in the process of comparing baseline data comprising sensor data where key seating is not suspected with newly recording data detect event signatures for key seating wear in order to detect key seating.

[0091] . For example, in some embodiments, where key seating is to be detected during drilling, suitable time-series machine learning model architectures include model architectures comprising one or more long- short-term memory models, LSTMs, one dimensional convolutional neural network models, ID CNNs, and Variational Autoencoders (VAEs).

[0092] In some embodiments, a time-series of sensor data is first pre-processed by running a Fast Fourier Transform, FFT, and the resulting power spectrum is input to the ML model. In some embodiments, one or more additional types of pre-processing are performed before the sensor data is input to the ML model. Examples of such additional pre-processinginclude for example, normalizing the FFT power spectra, segmenting sequences within the time-series sensor data into sliding windows (or performing as similar convolutional type of pre-processing), and optionally applying noise-reduction techniques, for example prefiltering to remove frequencies which lie in frequency bands known to not be of interest for detecting a key seating signature event.

[0093] The ML model may be initialized for training using any suitable standard weight initialization strategy, for example, Xavier (also known as Glorot) or normalized Xavier weight initialization heuristics may be used for nodes that use a Sigmoid or Tanh activation function. A He initialization may be used to initialize weights in neural network layers and nodes that use a rectified linear (ReLU) activation function.

[0094] The ML model may be configured with learning rules in some embodiments that use backpropagation with any suitable stochastic gradient decent, SDG, optimizer technique, or, for example, an adaptive moment optimizer or “ADAM” optimizer may be used.

[0095] The input training data comprises representative examples of both normal and key seating conditions for embodiments which use a supervised ML models. Embodiments where the ML model comprises a variational autoencoder may be trained using solely normal data to detect anomalies.

[0096] The ML models are trained until one or more criteria for sufficient training are met, for example, loss convergence or validation metrics may plateau, and to prevent overfitting the model may use early stopping.

[0097] Once the ML model has been sufficiently trained, time-series data can be fed in when the model is performing in an operational model. The output generated by the operational model when it is fed time-series sensor data represents a prediction or inference comprising indicators for whether a key seating event signature is present in the sensor data. For example, if the ML model is a binary classification model, the model output may be provided in the form of a classification of the sensor data as either belonging to a class of sensor data associated with or presentative of a key seating event or a class of sensor data not associated with a key-seating event.

[0098] The accuracy of the prediction or inference may be enhanced by performing one or more types of post-processing in some embodiments. For example, the ML model output may be post-processed to determine thresholding classification probabilities or to determine interpreting reconstruction errors to flag key seating events. Additionally,smoothing techniques can ensure robust decision-making by reducing false positives or oscillations in the detection.

[0099] In a further embodiment, the system described above includes an additional component: a signal generator 114. Signal generator 114 is located on drill platform 101 and is configured to transmit or inject a known signal into drill pipe 105. Signal generator 114 generates a signal that is synchronized to a certain time. In some embodiments, signal generator 114 is controlled by controller 112, as shown in FIG. 8. This signal is any recognizable signal that can be induced in drill pipe 105 by suitable equipment, such as an acoustic transducer or other similar systems. For example, the signal may be a sinusoidal signal. The signal induced in drill pipe 105 is set at a frequency that is substantially higher than the frequencies that are typically detected during normal drilling operations. For example, the signal may have a frequency of greater than 1000 hertz. Because drill pipe 105 is a solid material, the signal is conducted throughout drill pipe 105 at a known speed. As would be understood, both drilling riser 103 and drill pipe 105 may be hollow structures (such as tubes), but they are still formed from a solid material (such as steel or other metal s / alloys).

[0100] In this embodiment, when drill pipe 105 contacts drilling riser 103 in a key seating situation, the physical contact (key seating) between these structures allows the generated signal in drill pipe 105 to be transferred to drilling riser 103. This signal can be detected by sensor 110, which along with controller 112 operates in the same manner discussed above to detect key seating and alert a user.

[0101] This embodiment has the advantage of having a known signal to detect when analyzing data from sensor 110. Because the signal is set at a high, known frequency, it can be detected more easily by controller 112 during processing. This also reduces the reliance on filtering out the known baseline noise, and improves detection of key seating because the high frequency signal is less likely to be obscured by other noise in the data.

[0102] The use of signal generator 114 also allows for the determination of where the key seating is occurring along the length of drilling riser 103 or other undersea structures. This is useful for maintenance and tracking purposes because potential damage to drilling riser 103 or other undersea structures can be tracked if the position of the key seating is known. The signal that signal generator 114 injects into drill pipe 105 has a known frequency and frequency phase set by controller 112. The physical position of sensor 110 is also known with reference to fixed element such as BOP stack 107 because sensor 110 is fixed tovarious structures at a known distance during drill system assembly. Because the frequency and timing of the signal is known, the timing delay between signal injection and reception can be determined by comparing when sensor 110 detects the signal versus the known signal parameters. As an example, the injected signal may be a periodically repeating signal that has a known frequency of transmission. This frequency may be such that the signal is transmitted at a known set of times, such as at the start of every second for one second with a second pause between signals. When sensor 110 detects the signal, some time will have passed because the signal travel is not instantaneous. The speed that the signal travels through drill pipe 105 and drilling riser 103 (and other undersea structures, if applicable) is a known variable based on the materials of these structures, and thus the distance to the key seating can be determined by controller 112 by using the timing delay and known travel speed. Thus, in the above example, if the signal is transmitted at the start of every second of a clock, and sensor 110 receives the signal at 1.2 seconds per the same clock, then the signal took 0.2 seconds to reach the sensor, which then can be used with the known speed to determine the distance.

[0103] FIG. 9 is a schematic flowchart for a method 900 for detecting key seating or key seating wear in drilling operations according to at least one embodiment of the disclosed technology. In Figure 9, the method 900 of operation according to this embodiment begins with injecting a signal into drill pipe 105 using signal generator 114 in 902, which is controlled by controller 112. Next, sensor 110 records data in 904. The data is transmitted in 906 to controller 112, which processes in 908 the recorded data to determine if the signal is present. As discussed above, because the frequency of the signal is known, simple frequency filtering can be used to determine the presence of the signal in 910. For example, key seating and / or key seating wear characteristics comprising a frequency signature may be detected in the received signal and used to determine key seating or key seating wear is occurring during a drilling operation. Controller 112 will then alert a user that key seating wear is suspected and / or may provide some other indication that key seating or key seating wear has been detected in 914. As discussed above, controller 112 may use alert devices, such as displays, to alert the user, and / or transmit the alert using a data network. The user can then determine what action to take, such as pausing drilling operations while drill platform 101 is repositioned in 916. Repositioning can be accomplished using the engines and thrusters of drill platform 101. This process is often automated whereby the repositioning system is continuously working to center drill platform 101 at the specifiedsetpoint. In some embodiments, controller 112 may automatically command repositioning system to correct the position of drill platform 101. In this embodiment, controller 112 can also determine where the key seating has occurred by comparing the known signal injection time to the detected signal time and calculating the distance based on that time delay. After realignment, the drilling may be restarted in 918, for example, it may be automatically restarted in some embodiments in 918 when it is detected that the drill platform 101 has been centered on the specified set point.

[0104] In some embodiments, a computer-implemented method for detecting key seating or key seating wear in drilling operations is used by one or more of the above methods for detecting key seating or key seating wear in vibrational data recorded during drilling operations. The computer-implemented method may comprise in some embodiments receiving at a controller operably connected to a sensor, from the sensor, recorded vibration data of an underwater structure used for drilling operations. The vibration data may be received in real-time or near real-time during the drilling operations from the sensor. The sensor is preferably disposed on the underwater structure, or other underwater location where it is capable of detecting vibrations, which may be at acoustic frequencies or at other frequencies, generated by the drilling operation. The vibration data from the sensor is then processed using a trained machine learning, ML, model, wherein the trained ML model is configured to receive sensor data from a controller operably connected to the sensors. The trained ML model is trained using any of the above mentioned techniques suitable for the type of ML model architecture it is configured with to detect key seating by processing the input sensor data, for example, to classify it either as an anomaly that corresponds to key seating or key seating wear or not as an anomaly. For example, the trained ML model may be configured to detect key seating by processing the input sensor data to determine if the input sensor data meets one or more conditions to be classifiable as an anomaly that corresponds to key seating or key seating wear. The method further comprises the trained ML model outputting an indication, for example, an alert, whenever the ML model infers the input sensor data includes a key seating or key seating wear anomaly.

[0105] The claims and summary statements set out herein may be interpreted using the above detailed description. The Summary sections set forth one or more but not all exemplary embodiments of the present invention as contemplated by the inventor(s), and thus, are not intended to limit the present invention and the appended claims in any way. Moreover, the examples described above do not limit the present disclosure to what hasbeen particularly shown and described hereinabove. Rather, the scope of the present disclosure includes both combinations and sub-combinations of the various features described hereinabove, as well as variations and modifications thereof which would occur to persons skilled in the art upon reading the foregoing description and which are not disclosed in the prior art.

[0106] The foregoing description of the specific embodiments provides a disclosure of the claimed technology over the scope of the claims to fully reveal the general nature of the claimed technology so that others can, by applying knowledge within the skill of the art, readily modify and / or adapt for various applications the disclosed specific embodiments, without undue experimentation, without departing from the general concept of the invention. Therefore, such adaptations and modifications are intended to be within the meaning and range of equivalents of the disclosed embodiments, based on the teaching and guidance presented herein. It is to be understood that the phraseology or terminology herein is for the purpose of description and not of limitation, such that the terminology or phraseology of the present specification is to be interpreted by the skilled artisan in light of the teachings and guidance.

[0107] Various features of the invention which are, for clarity, described in the contexts of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features of the disclosed technology which are, for brevity, described in the context of a single embodiment may also be provided separately or in any suitable sub-combination in other embodiments of the disclosed technology.

[0108] Certain aspects of the disclosure may have other steps or elements in addition to or in place of those mentioned above. The steps or elements will become apparent to those skilled in the art from a reading of the above detailed description when taken with reference to the accompanying drawings.

[0109] The breadth and scope of the present invention should not be limited by any of the above-described exemplary embodiments but should be defined only in accordance with the following claims and their equivalents.

Claims

WHAT IS CLAIMED IS:

1. A system for detecting key seating wear during drilling operations, comprising: a sensor configured to detect vibrations in an underwater drilling structure; and a controller operably connected to the sensor, the controller configured to receive data from the sensor and to process the data to detect key seating wear.

2. The system of claim 1, wherein the sensor is one of a plurality of sensors, each of the plurality of sensors being configured to detect vibrations in an underwater drilling structure and operably connected to the controller.

3. The system of claim 1 or 2, wherein the controller is configured to send an alert to a user when key seating wear is detected.

4. The system of claim 3, further comprising a display operably connected to the controller, the controller being configured to use the display to alert the user to key seating wear.

5. The system of any one of the previous claims, wherein the controller is configured to use the sensor to record a baseline data set when key seating wear is not present.

6. The system of claim 5, wherein the controller configured to use the baseline data to determine if key seating wear is detected.

7. The system of any one of the previous claims, wherein the underwater drilling structure is one of: a drilling riser; and a lower marine riser package.

8. The system of any one of the previous claims, further comprising a signal injector controlled by the controller and configured to transmit a vibrational signal into a drill pipe.

9. The system of claim 8, wherein the controller is configured to detect the vibrational signal using the sensor to detect key seating wear.

10. The system of claim 9, wherein the controller is configured to determine a position of the key seating wear by comparing a time of injection of the signal with a time of detection of the signal by the sensor.

11. An offshore drilling system, comprising: a drill platform; and the system for detecting key seating wear of claim 1.

12. A method for detecting key seating wear in drilling operations, comprising: recording vibration data in an underwater structure during drilling operations using a sensor disposed on the underwater structure; and processing the vibration data from the sensor at a controller operably connected to the sensors to detect key seating wear by filtering the data to detect an anomaly that corresponds to key seating wear.

13. The method of claim 12, wherein recording vibration data further comprises receiving data from a plurality of sensors, each of the plurality of sensors being configured to detect vibrations in an underwater drilling structure and operably connected to the controller.

14. The method of claim 12 or 13, further comprising using a display operably connected to the controller to alert a user to key seating wear.

15. The method of any one of claims 12 to 14, further comprising using the sensor to record a baseline data set when key seating wear is not present.

16. The method of claim 15, further comprising using the baseline data as part of the processing the vibration data to determine if key seating wear is detected.

17. The method of any one of claims 12 to 16, further comprising using a signal injector controlled by the controller to inject a vibrational signal into a drill pipe.

18. The method of claim 17, further comprising detecting the vibrational signal using the sensor.

19. The method of claim 18, determining a position of the key seating wear by comparing the detected vibrational signal to the signal injected by the signal injector.

20. A key seating detector for detecting key seating during offshore drilling, comprising: a sensor configured to detect vibrations in an underwater drilling structure; and a controller disposed on a drill platform and operably connected to the sensor, the controller configured to receive data from the sensor and to process the data to detect key seating wear, and an alert device operably connected to the controller, the controller being configured to cause the alert device to generate a user alert for key seating wear.

21. The key seating wear detector of claim 20, wherein the alert device comprises a display.

22. A computer-implemented method for detecting key seating or key seating wear in drilling operations, the method comprising: receiving at a controller operably connected to a sensor, from the sensor, recorded vibration data of an underwater structure used for drilling operations, during the drilling operations, wherein the sensor is disposed on the underwater structure; processing the vibration data from the sensor using a trained machine learning, ML, model, wherein the trained ML model is configured to receive sensor data from the controller operably connected to the sensor, wherein the trained ML model is configured to detect key seating by processing received sensor data to determine if the received sensor data meets one or more conditions to be classifiable as an anomaly that corresponds to key seating or key seating wear; and outputting, by the trained ML model, an indication of whether the input sensor data indicates a key seating or key seating wear anomaly.

Images